The Routledge Handbook of Indian Buddhist Philosophy

‘A splendid and comprehensive study of Buddhist philosophy in its country of origin.
The
Routledge Handbook of Indian Buddhist Philosophy is an important addition to the global
philosophy movement.’
– Owen Flanagan, Duke University, USA
‘With an emphasis on the creative originality of key historical philosophers, and with a line-up
of the finest scholars in Buddhist philosophy, this volume is sure to become an indispensable
resource for teachers and scholars alike, and also serve as a magnificent introduction to the
dynamism and potentiality of the field.’
– Jonardon Ganeri, University of Toronto, Canada
‘This carefully edited volume provides an understanding of Indian Buddhist philosophy

through the study of the authorial figures that shaped its schools and doctrines. From the early
discourses to the Fourteenth Dalai Lama, a picture emerges of an active intellectual enterprise
devoted to the exploration of some of the most profound issues in philosophy.’
– Anita Avramides, University of Oxford, UK
‘An amazing resource for scholar and student alike, the chapters in this book, written by the
very best scholars in the field, masterfully introduce the life and thought of the major figures
of Indian Buddhist philosophy in clear and readable prose. There is no book quite like it. The
next time I teach a course on this subject, this book will be atop the reading list.’
– José Ignacio Cabezón, University of California, Santa Barbara, USA
‘This Handbook is a landmark contribution to the field, offering superb coverage of the
principal Indian Buddhist philosophers and showcasing the breadth and depth of one of the
world’s richest philosophical traditions. It will be an essential reference work for scholars and
students alike for years to come.’
– Evan Thompson, University of British Columbia, Canada
‘This is a great resource for all those who want to learn more about Indian Buddhism and its
philosophical traditions. It explores in great depth many different aspects of these rich traditions
and will be an invaluable resource for those interested in further exploring Indian Buddhist
philosophy.’
– Georges Dreyfus, Williams College, USA
THE ROUTLEDGE HANDBOOK OF
INDIAN BUDDHIST PHILOSOPHY
The Routledge Handbook of Indian Buddhist Philosophy is the first scholarly reference
volume to highlight the diversity and individuality of a large number of the most influential
philosophers to have contributed to the evolution of Buddhist thought in India. By placing the
author at the center of inquiry, the volume highlights the often unrecognized innovation and
multiplicity of India’s Buddhist thinkers, whose unique contributions are commonly subsumed
in more general doctrinal presentations of philosophical schools. Here, instead, the reader is
invited to explore the works and ideas of India’s most important Buddhist philosophers in a
manner that takes seriously the weight of their philosophical thought.
The forty chapters by an international and interdisciplinary team of renowned contributors
each seek to offer both a wide-ranging overview and a philosophically astute reading of the
works of the most seminal Indian Buddhist authors from the earliest writings to the twentieth
century. The volume thus also provides thorough coverage of all the main figures, texts,
traditions, and debates animating Indian Buddhist thought, and as such can serve as an in-depth
introduction to Buddhist philosophy in India for those new to the field.
Essential reading for students and researchers in Asian and comparative philosophy, The
Routledge Handbook of Indian Buddhist Philosophy is also an excellent resource for specialists
in Buddhist philosophy, as well as for contemporary philosophers interested in learning about
the rigorous and rich traditions of Buddhist philosophy in India.
William Edelglass is Associate Professor of Philosophy at Emerson College and Director of

Studies at the Barre Center for Buddhist Studies, USA.
Pierre-Julien Harter is Assistant Professor of Philosophy and The Robert H.N. Ho Family
Foundation Professor of Philosophy in Buddhist Studies at University of New Mexico, USA.
Sara McClintock is Associate Professor of Religion at Emory University, USA.

ROUTLEDGE HANDBOOKS IN PHILOSOPHY
Routledge Handbooks in Philosophy are state-of-the-art surveys of emerging, newly refreshed,

and important fields in philosophy, providing accessible yet thorough assessments of key
problems, themes, thinkers, and recent developments in research.
All chapters for each volume are specially commissioned, and written by leading scholars
in the field. Carefully edited and organized, Routledge Handbooks in Philosophy provide
indispensable reference tools for students and researchers seeking a comprehensive overview
of new and exciting topics in philosophy. They are also valuable teaching resources as
accompaniments to textbooks, anthologies, and research-orientated publications.
Also available:
THE ROUTLEDGE HANDBOOK OF LIBERAL NATURALISM
Edited by Mario De Caro and David Macarthur
THE ROUTLEDGE HANDBOOK OF PHILOSOPHY, POLITICS, AND

ECONOMICS
Edited by C.M. Melenovsky
THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF SEX AND SEXUALITY

Edited by Brian D. Earp, Clare Chambers, and Lori Watson
THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF FRIENDSHIP

Edited by Diane Jeske
THE ROUTLEDGE HANDBOOK OF INDIAN BUDDHIST PHILOSOPHY

Edited by William Edelglass, Pierre-Julien Harter, and Sara McClintock
For more information about this series, please visit: www.routledge.com/Routledge-Hand

books-in-Philosophy/book-series/RHP
THE ROUTLEDGE
HANDBOOK OF INDIAN
BUDDHIST PHILOSOPHY
Edited by William Edelglass, Pierre-Julien Harter

and Sara McClintock
Cover image: © Getty Images
First published 2023
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
605 Third Avenue, New York, NY 10158
Routledge is an imprint of the Taylor & Francis Group, an informa
business
© 2023 selection and editorial matter William Edelglass, Pierre-Julien
Harter, and Sara McClintock; individual chapters, the contributors
The right of William Edelglass, Pierre-Julien Harter, and Sara McClintock
to be identified as the authors of the editorial material, and of the authors
for their individual chapters, has been asserted in accordance with
sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in
any information storage or retrieval system, without permission in writing
from the publishers.
Trademark notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and explanation
without intent to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Names: McClintock, Sara L., editor. | Edelglass, William, editor. | Harter,
Pierre-Julien, editor.
Title: The Routledge handbook of Indian Buddhist philosophy/edited by
William Edelglass, Pierre-Julien Harter, and Sara McClintock.
Description: 1. | New York City: Routledge, 2022. | Series: Routledge
handbooks in philosophy | Includes bibliographical references and index.
Identifiers: LCCN 2022001712 (print) | LCCN 2022001713 (ebook) |
ISBN 9781138492257 (hardback) | ISBN 9781032301952 (paperback) |
ISBN 9781351030908 (ebook)
Subjects: LCSH: Buddhism – Philosophy. | Buddhism – India. |
Buddhist philosophy.
Classification: LCC B162. R68 2022 (print) | LCC B162 (ebook) |
DDC 294.3 – dc23/eng/20220202
LC record available at https://lccn.loc.gov/2022001712
LC ebook record available at https://lccn.loc.gov/2022001713
ISBN: 978-1-138-49225-7 (hbk)
ISBN: 978-1-032-30195-2 (pbk)
ISBN: 978-1-351-03090-8 (ebk)
DOI: 10.4324/9781351030908
Typeset in Times New Roman
by Apex CoVantage, LLC
I think those with wisdom drink something else, something that gives strength, is
delicious and irresistible, they drink like travelers who gulp rain just fallen from a
dark cloud.
– Sukka, from the Therīgāthā,
translated by Charles Hallisey
Thought is not calmed without speaking words.

Scholars will know which words have weight and which are drivel.
Through long habituation, those who are attached to their own positions, having
lapsed into hostile intent, do not believe others even when they speak what is true
and beneficial. But scholars, remaining unbiased toward both the trustworthy and
the untrustworthy, understand through intellectual examination the meaning of
both the proponent’s and the opponent’s position. Having analyzed whether this or
that word has weight or not, they should speak accordingly. . . . Fittingly, scholars
will amuse themselves by analyzing your words and our words, saying “this has
weight” and “this is drivel.”
– Bhāviveka, the Heart of the Middle, verse 4.74,
with commentary from the Flame of Logic
For all the myriad embodied hearts and minds who have conceptualized,
composed, taught, remembered, copied, promulgated, edited, translated,
debated, explained, wrestled with, and provided for the texts and ideas that
form the ground from which this book has grown.
CONTENTS
Acknowledgmentsxvi
Notes on Contributors xvii
Introduction 1
William Edelglass, Pierre-Julien Harter, and Sara McClintock
PART 1
Buddhas as Philosophers 7
Introduction to Part 1 9
1 Gotama Buddha: His Quest and His Teachings 11

Bhikkhu Anālayo
2 Siddhārtha Gautama: Beyond the Historical Figure 27

Richard F. Nance
3 The Tantric Buddha: Primordial Buddhas as Philosophical Authors 46

Vesna A. Wallace
4 Maitreya: The Future Buddha as an Author 64

Klaus-Dieter Mathes
PART 2
Poet Philosophers 79
xi
Contents
5 Aśvaghoṣa: The Dawn of Indian Buddhist Philosophy 83

Vincent Eltschinger
6 The Milindapañha: How to Use a Philosophical Resource and Find a

Literary Gem 97
Sonam Kachru
7 Cāttaṉār: Poet-Philosopher in Tamiḻ 113

Anne E. Monius
8 Saraha: The Anti-Philosopher as Philosopher 124

Roger R. Jackson
PART 3
Abhidharma Philosophers 139
9 The Dhammasaṅgaṇī and Vibhaṅga: The Perfectly Awakened

Buddha and the First Abhidhammikas 143
Maria Heim
10 Moggaliputta Tissa’s Points of Discussion (Kathāvatthu): Reasoning

and Debate in Early Buddhist Thought 160
Rupert Gethin
11 Kātyāyanīputra and the Large Commentary (Mahāvibhāṣā): The

Development of Abhidharma Literature and of a Sarvāstivāda Self-identity 172
Bart Dessein
12 The Vātsīputrīya/Sāṃmitīya: Buddhist Personalism as a Mainstream

School of Thought 188
Peter Skilling
PART 4
Philosophical Founders 205
13 Nāgārjuna: Dependent Arising Without Any Thing Arising 211

Amber D. Carpenter
14 Āryadeva: Quietism and Buddhist Ethics 236

Tom J. F . Tillemans
xii
Contents
15 Asaṅga: Great Systematizer of Yogācāra Thought 252

Jowita Kramer
16 Vasubandhu: Mainstream and Mahāyāna 266

Jonathan C. Gold
17 Dignāga: Early Innovator in Buddhist Epistemology 284

Kei Kataoka
18 Dharmakīrti: Philosopher and Defender of the Faith 303

John Taber
PART 5
Early-Period Commentators (Fifth–Seventh Century) 323
19 Buddhaghosa: Phenomenology, Hermeneutics, and Understanding 328

Chakravarthi Ram-Prasad
20 Bhāviveka: Madhyamaka Dialectic, Doxography, and Soteriology 346

Karl-Stéphan Bouthillette
21 Dharmapāla: A Janus-Faced Interpreter of Yogācāra? 361

Ching Keng
22 Sthiramati: A Yogācāra Commentator and Innovator 376

Roy Tzohar and Jowita Kramer
23 Devendrabuddhi and Śākyabuddhi: Dharmakīrti’s First Commentators 393

Alexander Yiannopoulos
24 Candrakīrti: Gardener of Sky-Flowers 404

Mattia Salvini
PART 6
Middle-Period Commentators (Eighth–Ninth Century) 421
25 Śubhagupta: An Externalist Outsider Within the Dharmakīrtian

Tradition 427
Serena Saccone
xiii
Contents
26 Dharmottara: Systematic and Innovative Commentator 439

David Nowakowski
27 Jñānagarbha: Two Truths Theory, Gradualism, and Mādhyamika

Philosophy 450
Ryusei Keira
28 Śāntarakṣita: Climbing the Ladder to the Ultimate Truth 463

Allison Aitken
29 Kamalaśīla: Mādhyamika Champion of Magical Reason 480

Sara McClintock
30 Haribhadra: The Voice of Perfect Wisdom 497

Pierre-Julien Harter
31 Śāntideva: Virtue on the Empty Path of the Bodhisattva 511

Stephen Harris
32 Prajñākaragupta: Buddhist Epistemology as the Path to the Wisdom

of Non-Duality 528
Shinya Moriyama
PART 7
Late-Period Commentators (Tenth–Twelfth Century) 541
33 Jitāri: A Later Buddhist Master of Debate 547

Junjie Chu
34 Jñānaśrīmitra: Variegated Non-Duality 558

Lawrence McCrea
35 Ratnakīrti: Aligning Everyday Experience with Momentariness

and Idealism 573
Patrick McAllister
36 Ratnākaraśānti: The Illumination of False Forms 587

Gregory Max Seton
37 Atiśa: The Great Middle Way of Mere Appearance 601

James B. Apple
xiv
Contents
38 Abhayākaragupta: A Last Great Paṇḍita 615

Kazuo Kano
PART 8
Modern Philosophers 631
39 B. R. Ambedkar: Justice, Religion, and Buddhist Political Philosophy 635

William Edelglass
40 The Dalai Lama XIV: A Modern Indian Philosopher 650

Jay L. Garfield
Index663
xv
ACKNOWLEDGMENTS
We are deeply grateful to our colleagues for the excellent chapters they contributed and
for their collaboration and patience. We are especially thankful to Sarah Fleming and Eva
Seligman, graduate interns from Harvard Divinity School at the Barre Center for Buddhist
Studies, who assisted us during the editing process; their tremendous skill as editors contrib-
uted greatly to this volume. We would also like to acknowledge the institutions whose support
directly benefited our work: the Barre Center for Buddhist Studies, Emory University, Harvard
University, and the University of New Mexico. Finally, we are profoundly appreciative of
our family members, with whom we are looking forward to spending more time: Leo, Sarah,
Jasmine, Jake, Amit, Kirstin, and Win.
xvi
NOTES ON CONTRIBUTORS
Allison Aitken is Assistant Professor of Philosophy at Columbia University, USA. Her

research centers on metaphysics in the history of philosophy. She has a forthcoming mono-
graph titled Introduction to Reality: Śrīgupta’s Tattvāvatāravṛtti, and her articles on topics in
Indian and Tibetan Buddhist philosophy, as well as Early Modern European philosophy, have
appeared in such publications as Philosophers’ Imprint, Analysis, Philosophy East and West,
and Journal of South Asian Intellectual History.
Bhikkhu Anālayo completed a Ph.D. thesis on the Satipaṭṭhāna-sutta at the University of Perad-
eniya in the year 2000 (published by Windhorse in the UK) and a habilitation thesis at the Univer-
sity of Marburg in the year 2007, comparing the Majjhima-nikāya discourses with their Chinese,
Sanskrit, and Tibetan counterparts (published by Dharma Drum in Taiwan). The main focus of
his more than 400 publications is on comparative studies of early Buddhist texts. He recently
retired from a position as a professor at the University of Hamburg and currently resides at the
Barre Center for Buddhist Studies in the USA, where he spends most of his time in meditation.
James B. Apple is Professor of Buddhist Studies at the University of Calgary, Canada. His
research focuses on the critical analysis of Mahāyāna sūtras and topics within Indian and
Tibetan forms of Buddhism. His books include Jewels of the Middle Way: The Madhyamaka
Legacy of Atiśa and His Early Tibetan Followers (Wisdom, 2019), A Stairway taken by
the Lucid: Tsong kha pa’s Study of Noble Beings (Aditya Prakashan, 2013) and Stairway to
Nirvāṇa (SUNY, 2008).
Karl-Stéphan Bouthillette, born in Montréal, received his first M.A. in Sciences of Reli-
gions at Laval University, Canada (2011), and his second one in Sanskrit Studies at Jawaharlal
Nehru University, India (2013). He obtained his Ph.D. (2018) from the Institute for Indology
and Tibetology of the Ludwig-Maximilians-Universität, Germany. In 2020, he published
his most recent book, at Routledge, titled Dialogue and Doxography in Indian Philosophy:
Points of View in Buddhist, Jaina, and Advaita Vedānta Traditions. Now at Ghent University,
Belgium, as an FWO post-doctoral fellow, his research focuses on lists and list-making as
contemplative devices in Indian philosophy.
xvii
Notes on Contributors
Amber D. Carpenter is Associate Professor at Yale-NUS College, Singapore. She publishes

in ancient Greek philosophy, especially the ethics, epistemology, and metaphysics of Plato.
Her book, Indian Buddhist Philosophy (Routledge) appeared in 2014. Her current project,
Buddhist Platonist Dialogues (buddhistplatonistdialogues.com) joins the two traditions in
philosophical inquiry. She has taught at Oxford, Cornell, St. Andrews, and York, and held
research fellowships and visiting appointments at the Einstein Forum (Potsdam), University of
Melbourne, Yale University, and with the Templeton Religious Trust. Her co-edited Portraits
of Integrity, arising from the work of The Integrity Project (integrityproject.org), appeared
in 2021.
Junjie Chu is a research fellow in the Department of Indology and Central Asian Studies at
the University of Leipzig, Germany. His current research centers on the Yogācāra system,
especially its epistemological theories, as well as Sanskrit manuscript studies. He is the author
of “On Dignāga’s Theory of the Object of Cognition as Presented in PS(V) 1” (Journal of the
International Association of Buddhist Studies, 29 (2), 2006 (2008) pp. 211–53.); and “Sanskrit
fragments of Dharmakīrti’s Santānāntarasiddhi” (Religion and Logic in Buddhist Philosophi-
cal Analysis, ed. by Krasser Helmut et al., Vienna, 2012, pp. 33–42).
Bart Dessein is full Professor at the Department of Languages and Cultures of Ghent Uni-
versity, Belgium and Member of the Royal Academy for Overseas Sciences, Belgium.
He obtained his Ph.D. from Ghent University in 1994 with a translation and study of
Dharmatrāta’s *Saṃyuktābhidharmahṛdaya, a Sarvāstivāda Abhidharma treatise. He has pub-
lished extensively on the philosophy of early Buddhism, mainly on the Sarvāstivādins and the
Mahāsāṃghika schools, as well as on early Chinese Buddhism.
William Edelglass is Associate Professor of Philosophy at Emerson College, USA, and

Director of Studies at the Barre Center for Buddhist Studies. He publishes broadly in Bud-
dhist studies, environmental humanities, and philosophy. William is co-editor of the journal,
Environmental Philosophy, as well as Buddhist Philosophy: Essential Readings (2009), The
Oxford Handbook of World Philosophy (2011), and Facing Nature: Levinas and Environmen-
tal Thought (2012). William lives on a homestead in southern Vermont.
Vincent Eltschinger is Professor of Indian Buddhism at the École Pratique des Hautes Études,
PSL University, France. His research focuses on the genealogy and the early history of Bud-
dhist philosophy. He is one of the editors of Brill’s Encyclopedia of Buddhism. His numerous
publications include Caste and Buddhist Philosophy (2012); Penser l’autorité des Écritures
(2007), Can the Veda Speak? (2012, with H. Krasser and J. Taber), Self, No-Self, and Salvation
(2013, with I. Ratié), Buddhist Epistemology as Apologetics (2014), and Dharmakīrti’s Theory
of Exclusion (2018, with J. Taber, M.T. Much and I. Ratié).
Jay L. Garfield is Doris Silbert Professor in the Humanities and Professor of Philosophy,
Logic, and Buddhist Studies at Smith College, USA; Visiting Professor of Buddhist Philoso-
phy at the Harvard Divinity School, USA; Professor of Philosophy at Melbourne University,
Australia; and Adjunct Professor of Philosophy at the Central Institute of Higher Tibetan Stud-
ies, India. His research addresses topics in the philosophy of mind and cognitive science, eth-
ics, epistemology, the history of Western philosophy, Indo-Tibetan Buddhist Philosophy, and
the recent history of Indian philosophy.
xviii
Rupert Gethin is Professor of Buddhist Studies at the University of Bristol, UK, where he
has taught Indian religions since 1987. His primary research interest is in the history and devel-
opment of Indian Buddhist thought in the Nikāya-Āgamas and exegetical literature. He is the
author of The Buddhist Path to Awakening (Brill 1992; Oneworld 2001), The Foundations of
Buddhism (1998), Sayings of the Buddha: A Selection of Suttas from the Pali Nikāyas (2008)
and, with R.P. Wijeratne, Summary of the Topics of Abhidhamma and Exposition of the Topics
of Abhidhamma (2002). He has also published a number of scholarly articles on the theory of
Buddhist meditation and Abhidharma. He has been president of the Pali Text Society since 2003.
Jonathan C. Gold is Professor in the Department of Religion and Director of the Center for
Culture, Society and Religion at Princeton University, USA. A scholar of Indian and Tibetan
Buddhist philosophy, he is especially interested in Buddhist approaches to meaning, ethics,
language, and learning. He is the author of The Dharma’s Gatekeepers: Sakya Paṇḍita on
Buddhist Scholarship in Tibet (2007) and Paving the Great Way: Vasubandhu’s Unifying
Buddhist Philosophy (2015), and is co-editor, with Douglas S. Duckworth, of Readings of
Śāntideva’s Guide to Bodhisattva Practice (Bodhicaryāvatāra) (2019).
Stephen Harris is Assistant Professor (Universitair Docent) at Leiden University’s Institute

for Philosophy, the Netherlands. He specializes in cross-cultural and Indian philosophy, with a
particular interest in Buddhist ethical texts. He has published articles in a number of academic
journals, including Journal of Buddhist Ethics, Journal of Indian Philosophy, Philosophy East
and West, and Sophia.
Pierre-Julien Harter is Assistant Professor of Philosophy and The Robert H.N. Ho Fam-
ily Foundation Professor of Philosophy in Buddhist Studies at University of New Mexico,
USA. He specializes in Buddhist philosophy in India and Tibet. His research on the Buddhist
concept of the path has nurtured his wide-ranging interests in different aspects of Buddhist
thought, such as metaphysics and ontology, epistemology, and ethics. He works also on Indian
philosophy more broadly, ancient Greek philosophy, and continental philosophy, framing
his research in the larger context of philosophy by fostering conversations between different
philosophical traditions and texts.
Maria Heim is George Lyman Crosby 1896 & Stanley Warfield Crosby Professor in Religion
at Amherst College and a Guggenheim fellow. Her most recent books are Words for the Heart:
A Treasury of Emotions from Classical India (Princeton 2022) and Voice of the Buddha: Bud-
dhaghosa on the Immeasurable Words (Oxford 2018). She is currently working on a transla-
tion of the Milindapañha for the Murty Classical Library of India.
Roger R. Jackson is Professor Emeritus at Carleton College, USA, where he taught South
Asian and Tibetan religions for nearly three decades. He has a B.A. from Wesleyan University,
USA, and a Ph.D. in Buddhist studies from the University of Wisconsin, USA. His scholarship
ranges over Indic and Tibetan Buddhist ritual, meditative practice, philosophy, and literature,
as well as modern Buddhist thought. Books he has either authored or co-edited include: Is
Enlightenment Possible? (Snow Lion, 1993), Tibetan Literature (Snow Lion, 1996), Buddhist
Theology (Curzon Press, 2000), Tantric Treasures (Oxford, 2004), and Mind Seeing Mind:
Mahāmudrā and the Geluk Tradition of Tibetan Buddhism (Wisdom, 2019). His most recent
book is Rebirth: A Guide to Mind, Karma, and Cosmos in the Buddhist World. (Wisdom, 2022).
xix
Sonam Kachru is Assistant Professor in the Department of Religious Studies, University of

Virginia, USA. He studies the history of philosophy with particular emphasis on the history
of Buddhist philosophy in South Asia; topics of particular salience for him include philosophy
of mind (intentionality, consciousness, attention), metaphysics, and practices of self. His first
book, Other Lives: Mind and World in Indian Buddhism (2021), a study of Vasubandhu’s phi-
losophy of mind, was published by Columbia University Press.
Kazuo Kano is an Associate Professor at Komazawa University, Japan. He is working on

Buddha-nature teaching in India and Tibet, and published a monograph on this issue (Buddha-
nature and Emptiness: rNgog Blo-ldan-shes-rab and a Transmission of the Ratnagotravibhāga
from India to Tibet. Vienna: Vienna Series for Tibetan and Buddhist Studies, 2016). His
research interest also includes Sanskrit manuscripts preserved in Tibet and the history of their
transmission.
Kei Kataoka is Associate Professor of Indian philosophy at Kyushu University, Japan. His
publications include critical editions of Śabara’s commentary, Kumārila’s Ślokavārttika and
Tantravārttika, Sucarita’s Kāśikā, Jayanta’s Nyāyamañjarī and Nyāyakalikā, and Aghoraśiva’s
Tattvasaṃgrahalaghuṭīkā. His book, Kumārila on Truth, Omniscience, and Killing (Austrian
Academy of Sciences Press, 2011) deals with a Brahmanical defense of the authoritativeness
of the Vedic scripture.
Ryusei Keira is Professor of Logic and Philosophy at Hosei University, Japan. After study-
ing at Keio University (B.A. 1988) and at Tokyo University (M.A. 1991), he received his
D.Litt. (2003) from Lausanne University, Switzerland. His research focuses on Mādhyamika
philosophy. He is the author of Mādhyamika and Epistemology (2004). His works include
“Kamalaśīla’s Madhyamakāloka, Part of Uttarapakṣa (Replies), Chapter One: Tibetan Text
and Annotated Japanese Translation (1)” in Acta Tibetica et Buddhica 9 (2016) and “The
description of niḥsvabhāvatā and its intentional meaning: Kamalaśīla’s solution for the doc-
trinal conflict between Mādhyamika and Yogācāra” in Acta Tibetica et Buddhica 2 (2009).
Ching Keng is Associate Professor in the Department of Philosophy, National Taiwan Uni-
versity, Taiwan. His fields of research include Yogācāra, Tathāgatagarbha, and Abhidharma
thought in India and China. His current research explores how, under various Buddhist cogni-
tive frameworks, mental consciousness and sensory consciousness work together. He is the
author of Toward a New Image of Paramartha: Yogacara and Tathagatagarbha Buddhism
Revisited (Bloomsbury 2022) and a co-editor of Buddhist Philosophy of Consciousness: Tra-
dition and Dialogue (Brill 2020).
Jowita Kramer is professor of Indology at Leipzig University. She specializes in Indian and
Indo-Tibetan Buddhism, with particular focus on the psychological concepts of the Yogācāra
tradition. Her research interests also include aspects of authorship and intertextuality in
Buddhist literature. She is the author of a monograph on the Yogācāra concept of the “five
categories” (vastu) of reality and numerous publications on the Pañcaskandhakavibhāṣā, a
sixth-century commentary by the Indian scholar Sthiramati. Before joining the University
of Leipzig, she held positions at the Universities of Heidelberg and Munich, the University
of California, Berkeley, the University of Oxford, and at the École Pratique des Hautes
Études.
xx
Klaus-Dieter Mathes is the Head of the Department of South Asian, Tibetan and Buddhist
Studies at the University of Vienna, Austria. His current research deals with “emptiness of
other” (gzhan stong) in the early Jonang tradition. He obtained a Ph.D. from Marburg Univer-
sity, Germany, with a translation and study of the Yogācāra text Dharmadharmatāvibhāga (pub-
lished in 1996 in the series Indica et Tibetica). His habilitation thesis was published by Wisdom
Publications under the title A Direct Path to the Buddha Within: Gö Lotsāwa´s Mahāmudrā
Interpretation of the Ratnagotravibhāga (Boston, 2008) and his latest work, Maitrīpa: India’s
Yogi of Nondual Bliss (Boulder 2021) was published by Shambhala Publications in the Series
Lives of the Masters. He is also a regular contributor to Journal of the International Associa-
tion of Buddhist Studies.
Patrick McAllister received an M.A. in philosophy in 2005 and a Ph.D. in Buddhist studies in
2011 (supervised by Helmut Krasser) at the University of Vienna. He has been working at the
Institute for the Cultural and Intellectual History of Asia since July 2016. His primary research
interest is the development of Buddhist epistemological theories during the ninth–eleventh
centuries (primarily in the works of Prajñākaragupta, Jñānaśrīmitra, and Ratnakīrti). He is
also engaged in projects of the digital humanities, in particular EAST (http://east.uni-hd.de), a
tool to collect bibliographical and prosopographical information on South Asian and Tibetan
literature, and SARIT (http://sarit.indology.info), a library of electronic Indic texts.
Sara McClintock is Associate Professor of Religion at Emory University, USA. Their

research interests focus on questions of truth, ethics, and the production of knowledge through
philosophy and narrative. They are the author of Omniscience and the Rhetoric of Reason:
Śāntarakṣita and Kamalaśīla on Rationality, Argumentation, and Religious Authority (2010),
and the co-editor with Georges Dreyfus of The Svātantrika-Prāsaṅgika Distinction: What
Difference Does a Difference Make? (2003). They have published in such journals as Argu-
mentation, Journal of the American Academy of Religion, Journal of Indian Philosophy, and
Journal of the International Association of Buddhist Studies; they are currently working on a
book on truth.
Lawrence McCrea is the author of numerous papers on traditional Indian poetry, poetics, lan-
guage theory, and hermeneutics. He has taught Sanskrit at the University of Chicago, Harvard
University, and Cornell University, all USA, where he is currently Professor in the Depart-
ment of Asian Studies. He is the author of The Teleology of Poetics in Medieval Kashmir
(Harvard, 2008) and co-author (with Parimal Patil) of Buddhist Philosophy of Language in
India: Jñānaśrīmitra on Exclusion (Columbia, 2010). He is currently working on a Reader on
Mīmāṃsā: A Historical Sourcebook in Indian Hermeneutical Theory, under preparation for
Columbia University Press.
Anne E. Monius was Professor of South Asian Religions at Harvard Divinity School, USA.
Her research interests lay in examining the practices and products of literary culture to recon-
struct the history of religions in South Asia. Her book, Imagining a Place for Buddhism:
Literary Culture and Religious Community in Tamil-Speaking South India (Oxford, 2001),
examines the two extant Buddhist texts composed in Tamil. At the time of her unexpected
passing in 2019, she was deeply engaged in a research project focused on the role of aesthetics
and moral vision in the articulation of a distinctly Śaiva religious identity in twelfth-century
South India.
xxi
Shinya Moriyama is Professor in the Department of Philosophy at Shinshu University, Japan.

His research interests lie in Buddhist epistemology in India and China, and comparative philoso-
phy. He is the author of Omniscience and Religious Philosophy: A Study on Prajñākaragupta’s
Pramāṇavārttikālaṅkārabhāṣya ad Pramāṇavārttika II 8–10 and 29–33 (LIT Verlag, 2014)
and a number of articles on the theory of perception and religious philosophy of Dharmakīrti
and his followers.
Richard F. Nance is Associate Professor in the Department of Religious Studies at Indiana

University, USA, where he teaches courses on Buddhist philosophy, rhetoric, ritual, and visual
culture. He is the author of Speaking for Buddhas: Scriptural Commentary in Indian Buddhism
(Columbia University Press, 2012), and has published work in various journals, including
Journal of the American Academy of Religion, Journal of Indian Philosophy, Journal of the
International Association of Buddhist Studies, and Revue d’Etudes Tibétaines. He is currently
at work on a translation of Vasubandhu’s Vyākhyāyukti and a monograph on the transmission
of Buddhist traditions in India.
David Nowakowski is a public philosopher, community educator, and independent scholar.

He serves as an adviser for Merlin CCC, a non-profit educational organization in Helena,
Montana, USA, and as senior mentor for the Merlin Fellowship Program. He earned his Ph.D.
in philosophy from Princeton University, USA, and his work has appeared in Philosophy East
and West, Asian Philosophy, and Journal of Indian Philosophy.
Chakravarthi Ram-Prasad is Distinguished Professor of Comparative Religion and Philoso-

phy at Lancaster University, UK, and Fellow of the British Academy. He has written more than
fifty papers on a variety of topics and seven books so far, including Knowledge and Libera-
tion in Classical Indian Thought (Palgrave, 2001), Divine Self, Human Self: The Philosophy
of Being in Two Gītā Commentaries (Bloomsbury, 2018), and Human Being, Bodily Being:
Phenomenology from Classical India (Oxford, 2018).
Serena Saccone is Associate Professor at the University of Naples “L’Orientale.” She holds
a Ph.D. in Indology and Tibetology from the University of Turin, and was a research fellow
at the Austrian Academy of Sciences from 2015–2021. Her main area of research is the intel-
lectual history of Buddhism, focusing on South Asian authors from the early medieval period,
with a specific interest in epistemology, logic, and soteriology, as well as their interconnec-
tions. Saccone’s monograph, On the Nature of Things (University of Vienna, 2018), concerns
the internal Buddhist debate on cognitions and their object in the eighth century. Her second
book, Tantra and Pramāṇa. Studies in the Sāramañjarī (University of Naples, 2022),
co-authored with Péter-Dániel Szántó, deals with the interrelationship between Tantric Bud-
dhism and the Dignāga-Dharmakīrtian tradition of logic and epistemology.
Mattia Salvini studied Sanskrit and Indian Philosophy mostly in India, obtaining a B.A.
and M.A. Sanskrit from RKM Vivekananda College (Madras University), India, and read-
ing Buddhist texts with Prof. Ramshankar Tripathi (Sarnath, CIHTS). He then obtained a
Ph.D. from SOAS University of London, under the guidance of Dr. Tadeusz Skorupski.
He has published articles on Madhyamaka philosophy with the Journal of Indian Philoso-
phy, Thai International Journal of Buddhist Studies, and the Bulletin de l’École française
d’Extrême-Orient. Presently, Mattia is the Dean of Scriptural Languages and the Dean
of the Faculty of Liberal Arts, International Buddhist College, Sadao Campus, Thailand,
xxii
and his research continues to focus on Buddhist philosophical texts and other aspects of
Buddhism in India.
Gregory Max Seton has been Senior Lecturer in the Religion Department at Dartmouth Col-
lege, USA, since 2017. He was previously a professor of Buddhist Studies at Mahidol Univer-
sity in Thailand. He received his D.Phil. in South Asian Buddhist Studies from the University
of Oxford, UK; a M.A. in Indo-Tibetan Buddhist studies from Naropa University, USA; a
M.A. in Religious Studies from the University of California, Santa Barbara, USA, and was a
DAAD research fellow at the University of Hamburg, Germany. His Sanskrit and Tibetan criti-
cal edition of Ratnākaraśānti’s Sāratamā is in press, and his monograph on Ratnākaraśānti’s
system is forthcoming.
Peter Skilling is Special Lecturer at Chulalongkorn University, Thailand, and an Honorary

Associate, Department of Indian Sub-Continental Studies, University of Sydney, Australia.
He retired as Professor of the French School of Asian Studies (EFEO) in 2016. His main field
of research is the epigraphy, archaeology, history, and literature of Buddhism in South and
Southeast Asia according to Sanskrit, Pali, Thai, and Tibetan sources. In 2009, he was awarded
the Ikuo Hirayama Prize by the Académie des Inscriptions et Belle-Lettres, Paris. In 2010, he
was elected honorary member of the Siam Society, Bangkok. In 2012, he was appointed fellow
of the Khyentse Foundation. In 2017, he was elected honorary fellow of the Asiatic Society of
Mumbai. Recently he published translations and studies of twenty-five sutras from the Tibetan
Kanjur under the title Questioning the Buddha: A Selection of Twenty-Five Sutras (Wisdom
Publications, 2021). Forthcoming is a new collection of translations from Pali and Tibetan,
Buddha’s Words for Tough Times: An Anthology (Wisdom Publications, 2023).
John Taber is Regents’ Professor Emeritus of Philosophy at the University of New Mexico,
USA. His publications include Kumārila on Perception: A Hindu Critique of Buddhist Epistemol-
ogy (RoutledgeCurzon, 2005), Can the Veda Speak? Dharmakīrti Against Mīmāṃsā Exegetics
and Vedic Authority: An Annotated Translation of PVSV 164,24–176,16 (with Vincent Eltsch-
inger and Helmut Krasser; Austrian Academy of Sciences Press, 2012), Dharmakīrti’s Theory
of Exclusion (apoha): An Annotated Translation of Pramāṇavārttikasvavṛtti 24,16–93,5,
Part I (with Vincent Eltschinger, Michael Torsten Much, and Isabelle Ratié; International Insti-
tute of Buddhist Studies, 2018), and diverse articles on Advaita Vedānta, Mīmāṃsā, Indian logic,
and Buddhist epistemology.
Tom J. F. Tillemans is a Canadian and Dutch citizen. Educated as a philosopher and

philologist, he made his living as a professor of Buddhist Studies at the University of Laus-
anne, Switzerland. He was, for several years, editor in chief of a large project to translate Bud-
dhist canonical texts from Tibetan and Sanskrit (www.84000.co). He lives on an island off the
west coast of Canada.
Roy Tzohar specializes in the history of philosophy with a focus on Buddhist and Brahmani-
cal philosophical traditions in India. He is a currently an associate professor in the East and
South Asian Studies Department at Tel Aviv University, Israel. He holds a Ph.D. from the
Religion Department at Columbia University, USA (2011), and an M.A. in philosophy from
Tel Aviv University’s Interdisciplinary Program for Outstanding Students (2004). His recent
monograph, A Buddhist Yogācāra Theory of Metaphor (Oxford University Press, 2018), deals
with Indian philosophy of language and experience.
xxiii
Vesna A. Wallace is Professor of Buddhist Studies in the Department of Religious Stud-

ies at the University of California, Santa Barbara, USA, where she teaches courses in South
Asian and Mongolian Buddhist traditions, and advanced Sanskrit language courses. She has
authored, translated, and edited six books and authored numerous articles on different aspects
of Indian and Mongolian Buddhism. Her most recent book is a co-authored volume titled Text,
Image, and Ritual in Mongolian Buddhism, to be published by Columbia University Press.
Alexander Yiannopoulos has been studying and practicing Buddhadharma since 2005, when
he took refuge under the Bodhi Tree with Chökyi Nyima Rinpoche. Dr. Yiannopoulos is a
graduate of the M.A. program in Buddhist Studies and Himalayan Languages at the Rangjung
Yeshe Institute in Kathmandu, Nepal, where he studied for six years, and completed his first
Fulbright research fellowship. He obtained his Ph.D. in 2020 from Emory University, USA.
xxiv
INTRODUCTION
William Edelglass, Pierre-Julien Harter, and Sara McClintock
The topic of this book, Indian Buddhist philosophy, is “vast and profound,” to borrow a tra-
ditional Buddhist expression. Buddhists employ this phrase to refer both to the vastness of
reality itself and to a buddha’s profound and unerring realization and teaching of that reality.
Ranging over approximately two and a half millennia, composed and preserved in diverse
languages both in oral and written form, manifesting in myriad genres from lyric to scholastic,
Buddhist philosophy in India is a field whose borders are yet to be found. Numberless manu-
scripts remain untranslated and even unedited, some located in India and some in monasteries,
libraries, and private collections elsewhere in the world. Other works survive only in Chinese
or Tibetan translations made centuries ago and are still mainly known to contemporary schol-
ars through catalogs and by reputation. Other texts have nearly entirely disappeared and can
only be glimpsed in fragmentary quotations or elusive references. No doubt, many others have
vanished without a trace.
The expanse of Indian Buddhist philosophy is indeed vast. But it is also profound, as those
who have explored Indian Buddhist philosophy very well know. Like the rest of Indian phi-
losophy, Indian Buddhist philosophy touches on nearly every aspect of human experience,
including the nature of suffering, reality, truth, goodness, compassion, love, freedom, death,
the body, the self, matter, consciousness, and divinity. Furthermore, Indian Buddhist philoso-
phy explores these aspects of human experience with a relentless ambition to get to the heart of
the matter, interrogating everything from the instruments by which knowledge can be certified
to the status of knowledge itself to the problems of language, perception, inference, argument,
objectivity, properties, kinds, agency, causality, morality, transformation, illusion, and much
more. As the chapters that follow demonstrate, Indian Buddhist philosophers have come to a
variety of conclusions about these diverse areas of inquiry.
The present volume, with its forty chapters, is a window onto the vast and profound world
of Indian Buddhist philosophy. It is a multifaceted stained-glass window with numerous panes,
each offering a view of complex philosophical topics colored by the perspective of a particular
Indian Buddhist philosopher as read and presented by a particular highly qualified contempo-
rary scholar. Our decision to place the author at the center of inquiry represents a deliberate
choice to highlight the philosophical creativity and innovation of diverse Indian Buddhist
philosophers, who too often are seen as mere exegetes working in the service of pre‑existing
schools. But categorizing thinkers according to schools, as if every thinker associated with that
DOI: 10.4324/9781351030908-1 1
William Edelglass, et al.
school had agreed to a stable set of doctrinal commitments, obscures the reality that schools
are themselves amorphous and ever-changing streams made up of the thought of many indi-
viduals over periods of time. Even as each author reflects and refracts the innovations and
ideas of their predecessors and contemporaries, it is individual thinkers who make up the
schools, both in terms of comprising them and, more literally, inventing them. While it is true
that Indian Buddhist philosophers do generally understand themselves as working from within
a tradition to which they unabashedly belong, this is not in itself anything out of the ordinary
for philosophers more generally. Indeed, it is important to remember that Western philoso-
phers, too, have undertaken their work within a series of inherited philosophical traditions,
with all the resources and blinders that come along with being so situated. The tendency to
see Indian philosophers – whether Buddhist or not – as so deeply enmeshed in their particular
schools as to be compromised in terms of their rational capacity is an Orientalist bias. The
decision to prioritize authors over schools, then, is in part a decision to resist this bias. By pre-
senting major figures who have contributed to the evolution of philosophical thought in Indian
Buddhist traditions and in Indian philosophy more broadly, this volume thus implicitly makes
a case for an enlarged conception of philosophy not restricted to its European embodiment.
To be clear, we are not suggesting that the study of Indian Buddhist philosophy can or
should abandon reference to schools. It is important to recognize that doxographical classifi-
cation, identification with a particular school, and refutation of opposing Buddhist and non-
Buddhist schools were all key elements of Indian Buddhist philosophical practices, especially
in later periods. Moreover, as these practices spread beyond India, so did the role of schools
in providing philosophical frameworks, authority, and identity for Buddhist philosophers. In
Tibetan Buddhist scholastic curricula, school affiliation is still central to the study of Indian
Buddhist philosophy. Leaving schools out of the equation is therefore not an option, and the
reader will find that nearly every contribution in this volume makes reference to the various
schools and sub-schools of the rich Indian Buddhist doxographical tradition.
Organizing a study of Indian philosophers by schools is thus not without justification.
Indeed, many scholars of Indian Buddhism, both traditional and modern, have chosen to do
so for a variety of reasons. Jan Westerhoff, for example, partially chooses this route in The
Golden Age of Indian Buddhist Philosophy, though he also acknowledges that
much of the seemingly clear division between the different schools is an ex post facto
arrangement, and that the individual thinkers concerned would have been unlikely
to ascribe themselves to the specific schools they are supposed to have belonged to
quite so readily.
(2018, 10)
To make up for some of what is lost through an organization based on schools, Westerhoff
introduces sections dedicated to individual philosophers, where he also notes some of the
ways such thinkers transcend any easy identification with a single school. Amber D. Car-
penter, in her book Indian Buddhist Philosophy (2014), also takes something of a hybrid
approach. While the book narrates the progression of Indian Buddhist philosophical thought
mainly through themes, important chapters are also dedicated to the Madhyamaka, Yogācāra,
and epistemological schools, especially as exemplified by their most iconic representatives:
Nāgārjuna, Vasubandhu, Dignāga, and Dharmakīrti.
Thus, even as we recognize that some reference to schools is necessary in the study of
Indian Buddhist philosophy, we need to be clear about what we mean by the term “school,”
since the word can have several valences ranging from a concrete institutional reality to an
2
Introduction
abstract doxographical construct (see McClintock 2018). Many Indian Buddhist philosophers
were engaged in the construction and reconstruction of schools as doxographical constructs,
and some were no doubt interested in forming institutional structures to maintain particular
philosophical views and lineages. But to structure our study of Indian Buddhist philosophy
around these schools, however we understand them, risks giving the constructs more concrete
reality than may be warranted. Doing so carries with it significant hermeneutic consequences,
as it would obscure the dynamic philosophizing and original contributions of individual
authors. Our focus on authors pushes back against the tendency to default to a history of
reified schools – even as the individual authors themselves often make claims on behalf of a
particular school affiliation.
An alternative way to structure this Handbook might have been through themes, with contri-
butions organized in sections such as metaphysics, language, epistemology, ethics, philosophy of
mind, political philosophy, argumentation, and so on. Organizing the volume according to these
contemporary subfields of the academic discipline of philosophy would, perhaps, have made the
volume more accessible to academic philosophers, who are one of the primary audiences for
this book. There are indeed advantages to this tactic, despite the challenge of figuring out which
authors should be situated in which categories, as so many of the philosophers in this volume
span multiple areas of philosophical thought. But those advantages would likely have a better
chance at being realized had the scope of the inquiry been expanded to Buddhist philosophy
beyond only India. Indeed, such a thematic approach is employed in A Companion to Buddhist
Philosophy (Emmanuel 2013), which treats philosophical topics across the whole development
of Buddhism in Asia. Thus, while that book does contain some short contributions on regional
schools (i.e., Theravāda, Indian Mahāyāna, Tibetan Mahāyāna and Vajrayāna, and East Asian
Buddhism), the bulk of the volume consists of chapters that take up specific themes in one or
more of these regions so that the audience can productively read them against one another. Our
decision to organize this Handbook by authors resembles more closely the approach taken by
Jonardon Ganeri in The Oxford Handbook of Indian Philosophy (2017), which also focuses on
figures in an effort to better highlight both chronology and innovation and to make space for
nonconformists whose role in the development of philosophical thinking should not be under-
estimated. While that work treats Indian philosophy more generally, this Handbook somewhat
artificially and with more focus highlights just those Indian philosophers who identify as Bud-
dhists. We do so not because we imagine that Indian Buddhist philosophy can be understood
apart from the broader study of philosophy in India (it most emphatically cannot!). Rather, we
want to make space for our readers to engage in a philosophically rich encounter with a greater
variety of Indian Buddhist philosophers in a single volume than has been possible to date.
The decision to place the author at the center of inquiry in these contributions further raises
the quite serious question: What is an author? While the notion of an author at first appears
self-evident, closer examination reveals a complex and shifting ground. Foucault’s proposal of
an author function is instructive here, as it reminds us that an author’s name is not the same as
a proper name that points to an individual. Rather, an author’s name performs a set of functions
that serve to classify, define, authenticate, and establish relationships among diverse texts. By
identifying a text as the product of a particular author, one also indicates something about how
that text is to be read and received. Thus, in specifying that the contributions in the Handbook
are to be organized by authors, we are pointing not only to individual historical figures but also
to powerful authorial figures whose name and reputation have been associated with particular
ideas, texts, practices, lineages, schools, and spiritual attainments. We believe that by attend-
ing to these authorial figures we gain a much richer understanding of the development and
practice of Indian Buddhist intellectual traditions.
3
William Edelglass, et al.
Having made the decision to organize this Handbook by authors, we were still confronted
by how to choose and how to arrange the philosophers. Although we lack concrete dates for
the majority of figures in this volume, there is general agreement about a relative chronology
based on the ways prior texts and thinkers seem to be presupposed by a given author. We
therefore decided to proceed chronologically when possible, with the understanding that each
subsequent philosopher could be seen, in part, as responding to philosophers who had come
before. At the same time, there are some authors who we believed needed to be placed into
special groups based less on schools or time periods but more on the genres in which their
philosophical material appears. The book opens with three parts organized by genre before
moving on in the next five parts to a more strictly chronological approach employing a some-
what arbitrary periodization. Here again, however, we continue to prioritize the author as the
central organizing principle.
We begin, perhaps somewhat unusually, with a part on buddhas as philosophers, based on
our contention that buddhas should be considered philosophical author figures insofar as they
may be credited with initially propounding – at least in our cosmic era – the foundational doc-
trines and ideas that serve as springboards for later Buddhist thinkers. This is true despite the
fact that the various buddhas may appear not to agree with each other (or, for that matter, with
themselves, given the diversity of texts attributed to the same buddha) and may also appear
not to be historical persons. The ideas of these buddha philosophers often emerge embedded
in dialogical scriptures which typically have significant narrative elements. Thus, the first part
of the book focuses on the genre of buddhavacana (“speech of the/a buddha”) as a modality
through which Buddhist authors – in this case, buddhas themselves – make known their philo-
sophical ideas.
We next move on to another genre with a part on poet philosophers. Indian Buddhist philos-
ophers composed their works in a wide variety of literary genres, with many of them employ-
ing poetry, narratives, and dialogues as well as treatises and commentaries. For some, the
poetry, narrative, and dialogue do important philosophical work, as might also be said of the
way that literary form has a philosophical function in Plato’s dialogues, Montaigne’s essays,
and Nietzsche’s aphorisms. In this part, we focus on poet philosophers whose philosophy can
be said to be embodied in their chosen literary form and whose philosophically significant
meaning would be lost if it were articulated in the more abstract language of a treatise.
The final part organized primarily on genre is concerned with abhidharma philosophers.
Here we are often thwarted in highlighting the specific contributions of individual authors due
to the fact that the texts themselves appear to have multiple authors. In this chapter, then, indi-
vidual texts often take on the author function. But by resisting the urge to talk about “Abhid-
harma” or “Abhidhamma” as a school and focusing instead on the individual texts or textual
traditions, we come as close as we can to highlighting the contributions of particular authors
even when we are lacking a name.
At this point, the parts shift and begin to focus more obviously on the iconic founding fig-
ures of Indian Buddhist philosophy. We start with a unit on philosophical founders, by which
we mean philosophers whose names have become associated with the founding of a particu-
lar stream or school. The part begins with Nāgārjuna and his student Āryadeva, and is filled
out with chapters on the remaining four figures making up the “six ornaments” according to
Tibetan Buddhist scholasticism: Asaṅga, Vasubandhu, Dignāga, and Dharmakīrti.
Following these formative figures, we consider a host of commentator philosophers, by
which we mean authors who in many cases literally and in other cases more figuratively posi-
tion themselves as commentators on the thought of one or more of the authors treated earlier
in the volume. Because there were a great number of these, we broke them into three major
4
Introduction
groups: early period, middle period, and late period commentators. In this way, we made room
for thinkers from the fifth–twelfth centuries, bringing us close to the end of the flourishing of
Buddhist philosophy in India. Throughout these chapters, we find authors who are bending,
contesting, reformulating, and repudiating various doxographical labels while also promoting
new ideas in the domains we mentioned earlier: metaphysics, language, epistemology, ethics,
philosophy of mind, political philosophy, argumentation, and so on.
The volume ends with a short part that is something of an outlier in that it treats modern
Indian Buddhist philosophers. Here we have just two figures, both of whom, however, have
played an outsized role in revitalizing Buddhist philosophy in India and beyond. The first is B.
R. Ambedkar, whose philosophical writings on Buddhism, religion, and politics have become
influential among practicing Buddhists – especially in India – and scholars alike. The second
is His Holiness, Tenzin Gyatso, the Fourteenth Dalai Lama of Tibet. While not an Indian by
birth, the Dalai Lama has spent more than half a century living in exile in India, where he has
consistently engaged with and emphasized the Indian heritage of Tibetan Buddhist philosophy.
He has also been a key figure in the promotion of what he calls “the Nālandā tradition,” a style
of doing philosophy with deep roots in the thought of many of the figures found in this volume.
The chapters that follow illustrate the broad diversity of Indian Buddhist philosophy. They
also exemplify the many ways in which contemporary scholars are engaging with Indian Bud-
dhist philosophers and their texts, for our contemporary authors are as diverse as the Indian
Buddhist authors treated in the volume, and their chapters showcase a range of approaches,
some more historical, some more philosophical, some more literary, some more technical,
some more playful. Like the Indian Buddhist authors who are the subject of this volume, our
contributors have also inherited intellectual practices and views that are manifest in their own
writings. Still, despite their differences, each chapter can serve in its own way as a window
onto the major philosophical innovations and ideas of the figures we have chosen for this vol-
ume as seen through the eyes of our particular contributors.
References
Carpenter, Amber D. 2014. Indian Buddhist Philosophy. London: Routledge.
Emmanuel, Steven M., ed. 2013. A Companion to Buddhist Philosophy. Blackwell Companions to Phi-
losophy. New York: Wiley-Blackwell.
Ganeri, Jonardon, ed. 2017. The Oxford Handbook of Indian Philosophy. Oxford: Oxford University
Press.
McClintock, Sara. 2018. “Schools, Schools, Schools – Or, Must a Philosopher Be Like a Fish?” In Bud-
dhist Spiritual Practices: Thinking with Pierre Hadot on Buddhism, Philosophy, and the Path, edited
by David Fiordalis, 71–103. Berkeley: Mangalam Press.
Westerhoff, Jan. 2018. The Golden Age of Indian Buddhist Philosophy. Oxford History of Philosophy.
Oxford: Oxford University Press.
5
PART 1
Buddhas as Philosophers
BUDDHAS AS PHILOSOPHERS
Introduction to Part 1
The opening part of this Handbook proceeds on the premise that considering buddhas as phi-
losophers is a worthy and interesting endeavor. This volume also argues that approaching
Buddhist philosophy through authors yields insights not available through other approaches.
Yet buddhas are often not recognized as authors in the usual sense of the term. Rather, they
are teachers whose discourse survives through the recollections of their disciples as preserved
in written and oral texts. These recollections, moreover, do not consist merely of the verbatim
contents of the buddhas’ teachings. Rather, they typically narrate a scene of instruction in
which a buddha appears as one actor among several, responding to the questions and cir-
cumstances of various interlocutors. Thus, whatever philosophy a buddha has taught must be
accessed indirectly, much as what we learn of Socrates comes through witnessing his conver-
sations with sundry characters in the dialogues of Plato.
For this part, we have chosen buddhas who are primordial, of this age, or of a future age.
Their status as authors might not be obvious. However, the word “author” is etymologically
related to “authority” (they are both derived from auctoritas), and in Buddhist traditions, bud-
dhas are precisely those who have the most authority. Such authority is grounded in their long
moral training and profound wisdom, their collection of merit, and their vision of reality as it
truly is. While they themselves are not authors composing texts, their speech is recorded by
others who attest to its accuracy and authenticity on the basis of their authority and thus make
the buddhas authors as sources of the teaching. The words of these buddhas are embedded in
narrative and doctrinal tracts full of rich philosophical content. As such, these buddhas – both
those for whom we have seeming historical evidence and those whose identities come into
greater focus only in literary or ritual milieus – can be considered foundational authors whose
works generate and are justified by philosophical thinking.
We start with Gotama Buddha, the buddha of the early discourses, oral texts that were even-
tually written down and that today form important root sources for the Theravāda tradition.
Bhikkhu Anālayo, in his chapter, pieces together philological evidence from these discourses
in Pāli and their parallels preserved in other languages to paint a picture that, as he writes,
“takes us as close as possible to the historical Buddha” while acknowledging that this picture
depends not on historical facts but rather on representations of the teacher Gotama Buddha as
found in the surviving texts. Philosophically, this buddha appears most interested in how to
attain the “deathless,” a term that for Gotama indicates not a state of eternal life but rather
DOI: 10.4324/9781351030908-3 9
Buddhas as Philosophers
“a supreme condition of freedom of the mind such that even the terror of mortality has com-
pletely lost its sting.” The picture painted here is of a philosopher of deep pragmatic concern,
whose teachings therefore involve explicit instructions for undertaking the contemplative
practices designed to lead one to this state of radical freedom.
Next, we move to Siddhārtha Gautama, a figure who might be considered the same as
Gotama Buddha but whose representations in the Sanskrit discourses associated with the
Mahāyāna stream of Indian Buddhism yield a rather different personage. Richard F. Nance
tackles the vexed question of the authorship of these Sanskrit discourses through a sustained
analysis of their understanding of what counts as buddhavacana, the “word of the/a buddha.”
In contrast to the previous chapter, which sought to bring us as close as possible to the histori-
cal buddha, this chapter urges us to look “beyond the historical figure” to recognize that bud-
dhavacana may, in the end, have very little to do with the question of provenance and much
more to do with the effects that such inspired speech may have on particular audiences. In a
manner quite different from that of Bhikkhu Anālayo, then, Nance arrives at a similar conclu-
sion in urging us to read Siddhārtha Gautama’s philosophy more in terms of its enactment in
embodied practice than as a set of philosophical doctrines.
The third buddha under consideration is the tantric Buddha, the quintessential “primordial
buddha” (ādibuddha) who appears in innumerable peaceful and wrathful forms and with a
great plethora of names and attributes within the tantric streams of Indian Buddhism. Taking
us still further away from “historical fact” toward something transcendent, Vesna A. Wallace
explains in her chapter that the buddhas who author the texts known as tantras are traditionally
understood as “nothing other than the ultimate truth” which reveals itself “in the conventional
and at times symbolic terms that constitute a tantric text.” The collapse of author and text is
here complete, a situation that itself raises a complex of philosophical questions. Yet, these
texts also contain explicitly philosophical passages, and Wallace helpfully guides us through
some of them. The chapter reveals the tantric Buddha as engaging in philosophical disputation
and offering arguments to refute non-Buddhists and Buddhists alike. Yet, it is also clear that
this is but one small portion of a buddha’s activity.
Finally, we turn to the future Buddha, Maitreya, whose role as an author is particularly
complex due to his relationship with a human author, Asaṅga (who is also treated later in
this volume). Although he is technically a bodhisattva, we include Maitreya in this first part
due to his transcendent and celestial status. Maitreya is a figure of great optimism for many
Buddhists, as it is foretold that he currently resides in a heavenly realm, Tus․ita (“Joyful”),
waiting for the moment when all traces of the previous Buddha’s teachings have disappeared
from the earth. At that point, Maitreya will take birth in human form, attain nirvāṇa, as did
Śākyamuni Buddha before him, and reintroduce the Dharma to the world. Meanwhile, though,
it is possible for some advanced meditation practitioners to meet Maitreya – as we find in
the case of Asaṅga. In his chapter, Klaus-Dieter Mathes relates how Asaṅga is said to have
received oral texts from Maitreya and then written them down. These texts include analyses
of buddha-nature and buddhahood, as well as the path of transformation from ordinary mind
to insight into the true nature of phenomena. While they have sometimes been interpreted as
consistent with Madhyamaka philosophy, they generally present what came to be identified
as a Yogācāra Buddhist view. Mathes carefully examines one popular corpus of five Maitreya
texts to conclude that there is sufficient evidence to hold that these works were all composed
“by a single hand.”
10
1
GOTAMA BUDDHA
His Quest and His Teachings
Bhikkhu Anālayo
Introduction
The main purpose of the present chapter is to delineate the figure of Gotama Buddha as a
seminal ancient Indian thinker and practitioner in the way he is depicted in the “early dis-
courses,” an expression whose implications I will explore in more detail in the next section.
The information that can be gathered from a comparative study of these textual materials takes
us as close as possible to the historical Buddha. Nevertheless, it is necessary to keep in mind
that this kind of evidence does not enable reconstructing with certainty historical facts about
an individual who lived in ancient India.1 To do so is not possible, given the limitations of the
sources at our disposal. For this reason, the reconstruction presented here is only a reflection of
early layers of textual depictions of Gotama Buddha. References given in the present chapter
to his thought or practice should for this reason invariably be understood to intend descriptions
or representations of his thought or practice in the textual sources without any implicit claim
to be reporting definite facts.
At the same time, it also needs to be borne in mind that these textual depictions have had
a lasting impact on later generations of Buddhists, and to that extent, they are themselves
historical facts. Even though the actual person who would have walked on the soil of India
some 2,500 years ago is beyond our reach, textual descriptions of Gotama Buddha, which
have made him a source of lasting inspiration for Buddhists ancient and modern, are within
our reach. Their historical-critical study enables us to discern layers in the development of
such descriptions and thereby understand the growth of Buddhist thought and practice. In the
end, this is perhaps even more historically significant than a quest for certainty about historical
facts regarding the person who lived in ancient India.
A convenient way of approaching the early textual depictions of the Buddha’s thought and
practice is through his quest for, and eventual attainment of, awakening, examined from the
viewpoint of how these events relate to key aspects of early Buddhist teachings. The received
biography of the Buddha, however, is of rather limited use for this purpose, as it comes inexo-
rably intertwined with later hagiography.2 Hence, in what follows, I will employ only material
from the early discourses in order to position my presentation within the earliest period of
Buddhism still accessible to us in the present time.
DOI: 10.4324/9781351030908-4 11
Bhikkhu Anālayo
The overall concern of these early discourses is what the Buddha taught rather than what he
did as an individual. This type of textual material does not even offer a continuous biography
but only biographical snippets here and there, used in order to illustrate or clarify some aspect
of the doctrine. Thus, the material employed here is not really biographical and its main pur-
pose in its original setting is to clarify philosophical points, which is precisely the way it will
be used in the following pages.
I begin with a brief introduction to the source material to be used in this chapter. Then
I examine passages that have a bearing on the Buddha’s progress to awakening and hence the
unfolding of his philosophical and soteriological contributions, in the way this is reported by
the texts. Precise dating and historical accuracy remain perennial problems for ancient Indian
texts. For this reason, what I present here is confined to an emic perspective, in the sense of
gathering together what the early Buddhist texts considered to have been key aspects of the
Buddha’s realization, and in what ways, according to the same texts, these key aspects differed
from related conceptions held by other thinkers and practitioners of his time.
My reference to the thought and practice of Gotama Buddha is deliberate. It is intended to
signal that my exploration needs to do justice to the fact that the Buddha’s thought, as a con-
tribution to ancient Indian philosophy, is inseparably connected to a corresponding mode of
living and practicing. It is, in a way, neither strictly philosophical nor strictly religious. Instead
of fitting smoothly into one of these two categories, it can perhaps best be seen to combine,
in a pragmatic manner, dimensions of both under the overarching aim of providing a viable
answer to the human predicament. From the emic perspective, such an answer needs to be both
realistically practicable and philosophically coherent.
As a starting point for exploring the Buddha’s quest, I consider his motivation to go forth
from lay life to become a renunciant, followed by examining depictions of his temporary over-
coming of mental defilements in order to gain deep concentration experiences. Then I turn to
his pursuit of ascetic practices. Next, I take up three knowledges he reportedly attained on the
night of his awakening and examine the textual records of his decision to teach, together with
exploring the content of what tradition regards as his first teaching.
The passages from the early discourses examined in this way bring out central teachings of
early Buddhist philosophy-cum-soteriology. These include a redefinition of the “deathless”;
an analytical attitude toward deep concentration experiences; a re-evaluation of the notion
of Nirvana and of the pleasure/pain paradigm by introducing the crucial ethical distinction
between what is wholesome/skillful and its opposite; insight into conditionality and not self
(placed within the context of rebirth); the importance of mindfulness and the awakening fac-
tors for meditation practice; the role of compassion in relation to the teaching activity of a
Buddha; and the apparent adoption of a medical scheme of diagnosis to formulate the core
teaching of the four noble truths. In conjunction, these various aspects converge on the role of
the Buddha as a remarkable teacher who has remained a source of inspiration for generations
upon generations of Buddhists from ancient India to contemporary times.
Source Material
The “early discourses” that are the source material in the ensuing exploration are the final results of
centuries of oral transmission of texts believed to have been spoken originally by the Buddha and
his chief disciples in their various encounters with followers, visitors, and opponents in the ancient
setting. These orally transmitted discourses eventually reached Sri Lanka, where they were subse-
quently committed to writing in an Indian language called Pāli. According to the traditional report
in the Ceylonese (Sri Lankan) chronicles, the Pāli discourses were written down shortly before
12
Gotama Buddha
the beginning of the Common Era (Adikaram 1946/1994, 79). By then, the main philosophical
content of the Pāli discourses appears to have been fairly closed, leaving aside the minor variants
that tend to occur naturally in written transmission. Hence, these materials can be reckoned as
testimonies for the developments in Buddhist thought up to that time (Anālayo 2012a).
Other Buddhist oral transmission lineages in India also preserved their records of the Bud-
dha’s teachings (Salomon 2018, 52f), and the texts passed on orally by these lineages were
also written down around the same time as the writing down of the Pāli discourses in Sri Lanka
or soon thereafter.
Due to climatic conditions in most of the Indian subcontinent, written records of the teach-
ings, made on such fragile material as palm leaves, are in constant need of being recopied.
With the eventual disappearance of Buddhism from India, the process of copying manuscripts
came to an end, and much of the early material that had been committed to writing there was
lost. In Sri Lanka, however, the Pāli manuscript tradition continued to be maintained.
Early Buddhist discourse material has also been preserved in manuscripts that mostly come
from Central Asia. Due to the dry climatic conditions in Central Asia, texts written on material
like birch bark endure much longer than palm-leaf manuscripts in Sri Lanka or India.
As far as we know, the discourse collections also reached Tibet, but the resultant trans-
lations into Tibetan appear to have been lost during a period in which Buddhism declined.
Hence, the Buddhist canon preserved in Tibetan translation no longer has counterparts to the
four main discourse collections, but only to some selected discourses preserved either on their
own or else as citations in other works.
Collections of discourses from different Buddhist oral transmission lineages had also been
brought to China and translated into Chinese. The large body of Buddhist texts translated into
Chinese contains counterparts to each of the four Pāli discourse collections. These are the
Chinese Āgamas (Anālayo 2015a), each of which stems from one of the various oral transmis-
sion lineages that attempted to preserve the teaching of the Buddha and that served as the basis
for the canonical textual collections of Buddhist monastic traditions. In the case of discourses
extant in the Chinese Āgamas, there is, of course, the problem of translation errors. Rendering
an Indic text into Chinese involves working with two languages that are substantially different
from each other, making it a rather demanding task to transpose content from one language to
the other without a change or even a loss of the meaning. Nevertheless, the discourses in the
Chinese Āgamas are not themselves products of Chinese culture but are instead testimonies of
early Indian Buddhism. Recourse to the Chinese Āgamas enables comparing different versions
of a particular discourse extant in Indic language(s) and thereby identifying and potentially
rectifying transmission errors. Such comparison of sources provides the main foundation for
my exploration of the thought and practice of Gotama Buddha.
In Quest of the Deathless

A background to the future Buddha Gotama’s decision to go forth in quest of the deathless
emerges from a reflection attributed to him in a Pāli discourse and its Chinese Āgama parallel
(AN 3.38 and MĀ 117; trans. in Bodhi 2012, 240 and Anālayo 2017d, 6, respectively).3 The
two separately transmitted versions of this discourse agree in reporting that Gotama was pro-
foundly affected by seeing others experiencing sickness, old age, and death: these sights made
him realize that he was subject to the same predicament himself. The texts explicitly present
this as a type of reaction distinct from the usual response by ordinary people, who tend to turn
away in disgust when encountering others being sick, growing old, and passing away, thereby
overlooking their own situation.
13
Bhikkhu Anālayo
The different outlook that emerges in this way is a departure not so much from the ancient
Indian philosophical perspective but from ordinary ways of thinking. For this reason, it
remains quite relevant to any time and society, including our own. The challenge of facing dis-
ease, old age, and death is a perennial one for humanity and has motivated a range of responses
in philosophy and religion. The distinct early Buddhist perspective on the matter, in the way it
emerges from the passage under discussion, is the need to face squarely these predicaments of
human life in full recognition that they apply to all of us (Anālayo 2016b).
The need to address the existential question posed by these fundamental aspects of human
life recurs in another Pāli discourse and its Chinese Āgama parallel. According to this passage,
the Buddha described to his disciples what had motivated him to go forth: the realization that –
being himself subject to disease, old age, and death – it was not appropriate to be spending
his life seeking what is also subject to these. For this reason, he decided to dedicate his life to
the quest for that which leads beyond disease, old age, and death (MN 26 and MĀ 204; trans.
in Ñāṇamoli 1995/2005, 256 and Anālayo 2012b, 25, respectively). Both texts mark out that
Gotama was still in the prime of his youth at the time of taking this decision and setting out
on such a quest.
The two texts in question also agree in reporting that Gotama went forth even though this
made his parents weep. This serves to clarify an element in later hagiography, according to
which his going forth took the form of a secret departure at night. This tale is the outcome of
later narrative developments and is at odds with the testimony of the early discourses. It might
be the result of a mode of storytelling in ancient India that involved the employment of a kind
of canvas to illustrate key aspects of a narrative during oral performance. In such a setting, a
pictorial depiction of Gotama’s decision to go forth could have been the starting point for these
hagiographic developments, in the sense that the graphic illustration was taken literally and
then influenced textual accounts.
The two early discourses just mentioned, which record the Buddha describing to his dis-
ciples what had motivated him to go forth, also report that, upon attaining awakening, he
proclaimed to have attained the “deathless” (amata/amṛta). The usage of this expression to
denote the successful completion of his quest is significant in view of its connotations in the
ancient Indian setting as a referent to a form of immortality or eternal existence (Collins 1982,
43; Vetter 1995, 217). The attainment of the deathless in early Buddhist thought is not the
achievement of a state of eternal life. Instead, it signifies a liberating insight that results in the
complete conquest of any fear of death. With full awakening reached, the Buddha and those of
his disciples who had also reached complete liberation from defilements of the mind are con-
sidered to have reached a condition of the mind that will remain completely composed in the
face of their own passing away (or that of others). In this way, their conquest of death neither
avoids the passing away of their physical body nor leads to gaining a state of immortality in a
heavenly realm. Instead, it involves a supreme condition of freedom of the mind such that even
the terror of mortality has completely lost its sting.
Analysis of Absorption
Returning to the account of the future Buddha Gotama’s quest for the deathless, the relevant
material for reconstructing an early record of his progress to awakening occurs in various early
discourses that unfortunately do not stand in an explicit chronological order to one another.
For this reason, at times it is uncertain whether a particular episode should be placed earlier
or later in relation to other episodes. This holds for a description of Gotama’s struggle with a
range of mental obstructions that prevent the attainment of deep concentration in the form of
14
Gotama Buddha
absorption ( jhāna/dhyāna). Although the commentarial tradition considers this description to

report events that took place only during the night of the Buddha’s awakening, the relevant
discourses contradict this directly: they describe a mode of practice involving several days and
nights, which cannot be confined to what happened in a single night (Anālayo 2011, 741 n.
270). This makes it probable that the episode in question should be placed at an earlier time,
well before the actual event of awakening.
The episode itself takes its occasion from a visit paid by the Buddha to a group of seri-
ously meditating monastics. During their practice, these monastics had been encountering
inner experiences of light, which subsequently disappeared again. The Buddha is on record for
informing them that the same had happened to him during the time before his awakening. At
that time in the past, he had identified various mental obstructions that needed to be overcome
in order to prevent the inner light disappearing. Overcoming these obstructions eventually led
to his successful attainment of absorption (MN 128 and MĀ 72; trans. in Ñāṇamoli 1995/2005,
1015 and Anālayo 2017d, 35, respectively).
The description of the attainment of absorption, given in the Pāli and Chinese versions of
the discourse in question, does not employ the standard pericope for such description found
elsewhere among the early discourses. Instead of listing the four levels of absorption men-
tioned in the standard pericope, it presents three. The resultant two modes of description,
enumerating either three or four levels of absorption, both involve an analysis of the mental
factors present during absorption attainment.4
The main concern behind the threefold model found in the description of Gotama’s own
pre-awakening struggle with obstacles to the gaining of deep concentration appears to be the
degree of effort required for attaining absorption and remaining in it. This degree of effort
gradually diminishes as concentrative absorption deepens. At first, more effort is required for
directing the mind and keeping it in its collected condition. Then directing becomes effortless
and the mind only needs to be kept where it is. Eventually, even that is no longer necessary.
The standard pericope instead presents a fourfold analysis of absorption. The first level
involves an experience of joy and happiness born of seclusion from sensual distraction and
unwholesome states of mind. The second level concerns an experience of joy and happiness
born of concentration proper. With the third level of absorption, joy is left behind and only
happiness remains. Happiness is then also transcended with the fourth level of absorption, an
experience that is characterized by deep equanimity.
The differences between these two modalities of analysis, which survey what in actual prac-
tice would be the same process of gradually deepening concentration, appear to reflect different
concerns: whereas the threefold analysis places more emphasis on the gradual reduction of
effort, the fourfold analysis gives more room to the hedonic dimension of absorption experience.
Now, an emphasis on the degree of effort required to enter deep concentration would have
been a natural way of analyzing absorption attainment as part of Gotama’s progress to awak-
ening, particularly in view of his own experience of needing to overcome a range of mental
obstructions. Several of these, in fact, require a fine-tuning of effort such that the mind is
neither too lax nor too tense.
Giving priority to the type of joy and happiness present or absent during different levels of
absorption instead seems to relate to a crucial understanding gained by Gotama subsequently,
after he had given up an attempt to reach awakening through the pursuit of ascetic practices
(a topic to which I will return in what follows). It seems that the realization of the important
contribution offered by wholesome types of joy and happiness to progress to awakening may
have informed the mode of presentation found regularly in teachings given on other occasions,
which take the form of the fourfold analysis.
15
Bhikkhu Anālayo
Both schemes mark a departure from the way absorption attainment appears to have been
conceived by some meditators in the ancient Indian setting. Judging from the testimony of
other discourses, experiences of absorption led some non-Buddhist practitioners to a range
of speculative views and metaphysical conclusions about the nature of the individual and the
world (DN 1 and DĀ 21, as well as other parallels; trans. in Walshe 1987, 73–87 and Anālayo
2017b, 122–70, respectively). In contrast, the two modes of analyzing absorption under dis-
cussion eschew any metaphysical conclusions and instead focus squarely on the condition of
the mind itself. This condition is to be examined closely in order to discern the principal men-
tal factors responsible for absorption to occur. Such analysis offers clear directives on how this
experience can be repeated and also points directly to its conditioned nature. In this way, it can
stimulate the cultivation of two meditative qualities that in early Buddhist thought collaborate
(Anālayo 2003, 88–91): tranquility (samatha/śamatha) and insight (vipassanā/vipaśyanā).
Immaterial Attainments and Nirvana

Presumably subsequent to having successfully attained absorption in the way described in the
episode concerning a range of mental obstructions to inner experiences of light, Gotama con-
tinued to deepen his cultivation of the mind under the tutelage of two contemporary teachers,
Āḷāra Kālāma and Uddaka Rāmaputta, whose teaching involved the meditative experiences of
“nothingness” and of “neither-perception-nor-non-perception.” The relevant Pāli and Chinese
discourses report that he was soon able to reach a personal experience of the meditative attain-
ments on which the teachings of Āḷāra Kālāma and Uddaka Rāmaputta were based (MN 26
and MĀ 204; trans. in Ñāṇamoli 1995/2005, 257 and Anālayo 2012b, 27, respectively).
In the standard accounts of meditation practice in other discourses, these meditative attain-
ments find a placing as the third and fourth of four “immaterial” (arūpa) experiences. The
whole set of four immaterial attainments requires leaving behind any perceptions related to
the material world. They take as their respective objects the perceptual experience of infinite
space, infinite consciousness, nothingness, and neither-perception-nor-non-perception. Their
cultivation requires the concentrative strength of the fourth absorption (in the fourfold scheme
discussed previously). The teaching of Āḷāra Kālāma was concerned with the immaterial
attainment of nothingness; the teaching of Uddaka Rāmaputta was concerned with neither-
perception-nor-non-perception (an attainment of such subtlety that one is neither really per-
ceptive nor actually impercipient).
The discourses report that both teachers acknowledged Gotama’s success to the extent of
inviting him to take on a teacher’s role himself. Gotama is on record as declining the offer,
as he did not consider these attainments to provide the answer to his existential quest. This
thereby marks another departure from the ancient Indian setting, in as much as the notion of
the final goal is concerned. Although the respective attainments must at first have seemed
promising enough to motivate Gotama to choose Āḷāra Kālāma and Uddaka Rāmaputta as his
teachers, their actual experience failed to provide him with what he was seeking: a definitive
solution to the human predicament of being subject to old age, disease, and death.
The solution he eventually found, on the night of his awakening, involves a reappraisal of
the notion of Nirvana. The survey of metaphysical interpretations of absorption experiences,
mentioned previously, covers some relevant notions, presumably held by other practitioners in
the ancient Indian setting. These are on record for identifying sensory enjoyment or absorption
experiences as forms of realizing Nirvana (DN 1 and DĀ 21, as well as other parallels; trans. in
Walshe 1987, 85–86 and Anālayo 2017b, 168–70, respectively). Another passage even reports
that bodily health was identified with Nirvana (Anālayo 2011, 410).
16
Gotama Buddha
The implications of the same term in early Buddhist discourse differ, as the central idea
here is a complete transcendence of all modes of being and existing. The realization of Nirvana
attained with full awakening, besides being a realization of the deathless, can also be seen as
a realization of the “birthless,” so to say. This is meant in the sense that it leads beyond the
prospect of any future rebirth in saṃsāra, the endless cycle of births and deaths that – accord-
ing to ancient Indian thought – holds in bondage ordinary humans and other living beings. The
novelty of the idea of a middle-way approach by way of a complete transcendence of saṃsāra
that neither involves an alternative form of existence nor is mere annihilation forms a recurrent
topic in the discourses. Time and again, these discourses show the Buddha and his disciples in
discussion with contemporary practitioners unable to understand what the final goal of early
Buddhist practice was about.
The attainments of nothingness and neither-perception-nor-non-perception, in spite of their
profound nature, were for this reason insufficient to provide an answer to Gotama’s quest.
They only led to rebirth in a corresponding heavenly realm and hence failed to be the “birth-
less” solution that Gotama was seeking.
Ascetic Practices and the Pleasure/Pain Paradigm

Whereas the chronology of the account of Gotama’s absorption practice in relation to other
episodes relevant to his progress to awakening is open to different interpretations, the early
discourses provide clear indications regarding what happened after he left his two teachers,
Āḷāra Kālāma and Uddaka Rāmaputta. In his continued quest for what leads beyond old age,
disease, and death, Gotama is on record for engaging in some of the ascetic practices that were
in vogue in ancient India.
Closer inspection of relevant textual material brings to light that some of these descriptions
appear to refer to experiences he had in a previous life, spent as an ascetic (Anālayo 2011,
116). Hence, of direct relevance to the account of his progress to awakening in his final life,
after having left his two teachers, appear to be only three ascetic practices: forceful control of
the mind, breath retention, and fasting. An account of these three modalities of asceticism can
be found in a Pāli discourse and a parallel extant in Sanskrit fragments (MN 36 and Liu 2010;
trans. in Ñāṇamoli 1995/2005, 337 and Anālayo 2017d, 56, respectively).
The attempt to control the mind by sheer force takes the form of firmly clenching the teeth,
pressing the tongue against the roof of the mouth, and then trying to beat down and constrain
the mind. The idea could be that, since even the profound meditative experiences of nothing-
ness and neither-perception-nor-non-perception had not led to liberating the mind from the
bonds of saṃsāra, the alternative approach was to face things head on by trying to force the
mind into becoming liberated from any bondage.
Once this approach had not brought tangible results, a natural extension of the same attitude
was breath retention. In fact, clenching the teeth and pressing the tongue against the roof of the
mouth already come fairly close to such a mode of practice. The sources agree that Gotama
tried out different modalities of breath retention, yet none of these led him to the goal of his
aspiration.
From attempting to reduce his intake of air to a minimum, the next step was an attempt
to reduce his intake of food to an absolute minimum. This resulted in a serious weakening
of his body to the extent of bringing him close to death, yet it did not bring about the desired
liberation.
The failure of all three ascetic practices to result in progress to awakening led Gotama to
reconsider his approach. This made him realize that wholesome types of pleasure need not be
17
Bhikkhu Anālayo
shunned. The resultant shift of understanding motivated him to abandon asceticism and paved
the way for his attainment of awakening, once his body had regained strength.
The decisive shift of perspective at this juncture reflects a crucial undercurrent of early
Buddhist thought and practice, which is the foundational ethical distinction between what
is wholesome or skillful (kusala/kuśala) and what is unwholesome or unskillful (akusala/
akuśala). The new viewpoint requires stepping out of the ordinary pleasure/pain paradigm.
Based on a thoroughgoing emphasis on what happens in the mind, this novel outlook gives
importance to the ethical repercussions of any experience over its affective quality.
The ordinary untrained mind is prone to pursue what is pleasant and avoid what is
painful. In recognition of the bondage of sensual indulgence that ensues from this, ascetic
ideology tends to advocate the pursuit of self-inflicted pain and the concomitant avoidance
of pleasure. Leaving behind both attitudes as two extremes to be avoided, the solution pro-
posed in early Buddhist thought takes the form of a middle path that emphasizes the ethical
quality of any experience. This ethical quality relates to the intention that stands behind that
experience. The resultant perspective reveals that some agreeable experiences are obstacles
to progress to awakening, such as sensual indulgence, but other pleasant experiences can
support such progress, such as the pleasure and joy of deep concentration. The same dis-
tinction holds for painful or neutral experiences, some of which are detrimental whereas
others are beneficial.
In this way, the evaluation of experience is decoupled from the pleasure/pain paradigm and
instead oriented along the distinction between what is wholesome/skillful and what is unwhole-
some/unskillful (MN 70 and MĀ 195; trans. in Ñāṇamoli 1995/2005, 579 and Anālayo 2017d,
75, respectively). What is wholesome or skillful should be pursued, regardless of whether it is
pleasant or unpleasant. Conversely, what is unwholesome or unskillful should be avoided, no
matter how pleasant (or unpleasant) it might be.
Moreover, the emphasis shifts from the adoption of certain modes of external conduct and
behavior to the mental state that informs them. Of crucial importance here is the question of
intention (cetanā). This is where the distinction between wholesome/skillful and unwhole-
some/unskillful really matters. Intention is what sets the course for what is then undertaken by
speech and action.
The same ethical distinction can be seen at work in relation to the three modalities of asceti-
cism described previously. Even though Gotama stopped his ascetic practices and considered
them as not conducive, on their own, to liberation, he nevertheless retained specific aspects of
each mode of conduct in his instructions to his followers.
The idea of overpowering the mind by sheer force has found a place in series of strate-
gies for overcoming unwholesome thoughts, where it comes as a last resort when all else
has failed (MN 20 and MĀ 101; trans. in Ñāṇamoli 1995/2005, 213 and Anālayo 2013, 153,
respectively). In this way, from being the sole approach to gaining control over the mind, the
use of force is relegated to being an emergency measure for preventing the occurrence of
unwholesome actions, rather than being completely discarded. The potential of attending to
the process of breathing takes the form of detailed instructions on the cultivation of mindful-
ness of inhalations and exhalations, which differ from breath retention by eschewing control
and instead giving prominence to non-interfering mindful observation (MN 118 and SĀ 803;
trans. in Ñāṇamoli 1995/2005, 943 and Anālayo 2013, 228, respectively). The undertaking of
fasting has continuity in instructions to monastic disciples on the observance of intermittent
fasting, in the sense of taking only a single meal per day (MN 65, MĀ 194, and EĀ 49.7; the
first trans. in Ñāṇamoli 1995/2005, 542 and the last in Anālayo 2016a, 90).
18
Gotama Buddha
In this way, the shift of perspective evident in the account of Gotama’s progress to awak-
ening spells out in a middle path that is able to accommodate aspects of the three types of
asceticism to the extent that these can facilitate progress to awakening. What counts – here and
elsewhere in early Buddhist thought – are the ethical repercussions, in terms of fostering what
is wholesome/skillful and avoiding its opposite.
Recollection of Past Lives and the Construction of Personal Identity

With strength regained, after having given up the practice of fasting, Gotama is on record for
having cultivated a recollection of his own past lives during the first part of the night of his
awakening (MN 4 and EĀ 31.3; trans. in Ñāṇamoli 1995/2005, 105 and Anālayo 2016a, 21,
respectively). In the narrative context of Gotama’s progress to awakening, such turning to
the past stands in continuity with his earlier review of the previous practices he had under-
taken in this life in order to gain awakening. This retrospection had led him to the realization
that wholesome forms of joy and happiness need not be shunned. The benefits derived from
looking back at his own past experiences might have inspired his decision to take the same
approach further by way of recollecting his previous lives.
The ability to recollect previous lives, a meditative practice based on mindfulness, features
in the early discourses as being accessible to those who have reached mastery of the four
absorptions. It involves being able to recall one’s former name and living circumstances, the
food eaten and the pleasures and pains experienced in a past life, as well as remembering the
transition from one life to another. The inclusion of such knowledge among what, according to
tradition, was one of the three key insights that brought about the Buddha’s awakening is only
one in a range of passages that show rebirth to have been an integral part of early Buddhist
thought (Anālayo 2018a).
The description of such recollection of former lives does not spell out how this type of
knowledge contributed to Gotama’s eventual awakening. In view of the centrality of the teach-
ing on not self in early Buddhist doctrine, however, it seems fair to propose that the direct wit-
nessing of one past life after another could have been a starting point for Gotama to discern the
constructed nature of the sense of identity. In any of these former lives, which he apparently
recollected vividly, a particular name would have been experienced as “me,” a certain family
and living circumstances been taken to be “mine,” and specific food as well as pleasure and
pain been perceived as “my” experiences. Yet, with a change of scene, so to say, from one life
to the next, a different name would have become “me,” a different family “mine,” and so on.
The notion of a continuity from one life to the next provides a needed perspective for a proper
appreciation of the early Buddhist teaching on not self (anattā/anātman). This doctrine does not
imply a denial of the reality of subjective experience. It only denies that there is a permanent
and unchanging entity involved in such subjective experience. Instead, subjective experience is
seen as nothing but a flow, the product of specific conditions and bereft of anything permanent.
Hence, continuity from one life to another does not involve the reincarnation of a permanent
entity or soul. Instead, it is considered to be simply a continuity of the flow of causes and condi-
tions operative during life. A difference is that, at the time of death, these proceed independently
of the physical body. An illustrative example is a flame which, with the support of wind, may
ignite even what is at a distance (Anālayo 2018a, 22). The flame is just a process of combustion
and does not require any permanent entity to remain afire. It continues as a changing process of
burning as long as the required conditions are in place, and it is even able to ignite something
not immediately contiguous to its fuel, as long as there are supportive conditions for that.
19
Bhikkhu Anālayo
The doctrine of not self is perhaps the most substantial departure of early Buddhist thought
from its ancient Indian setting. In one discourse, this finds expression in an analysis of four
types of clinging, which are clinging to sensual indulgence, to views, to rules and obser-
vances, and to the notion of a self (MN 11, MĀ 103, and EĀ 27.2; the first trans. in Ñāṇamoli
1995/2005, 161 and the last in Anālayo 2016a, 42). Whereas the first three modalities of cling-
ing were according to this discourse within the purview of contemporary Indian practitioners
and thinkers, the problem posed by clinging to a sense of self is seen as a unique contribution
of the Buddha. The doctrine of not self is what makes his teaching exceptional for its compre-
hensive approach to the problem of clinging (see also Anālayo 2022).
The Divine Eye and Conditionality

Recollection of one’s own past lives has its counterpart in the ability to witness the passing away
and being reborn of other sentient beings, called the “divine eye.” These two knowledges stand
in close relationship to each other, just as the doctrine of not self stands in a close relationship
to the early Buddhist teaching on dependent arising ( paṭiccasamuppāda/pratītyasamutpāda).
It is only for ease of presentation that I have taken up the notion of not self in relation to rec-
ollection of past lives and now turn to dependent arising. This is not meant to imply that these
two doctrines are separate and can be matched precisely with the first and second knowledge
gained according to the traditional report by the Buddha on the night of his awakening. Instead,
both are best seen as concomitant dimensions of a gradually dawning insight that found its
completion in the third knowledge, the realization of Nirvana as the actual event of awakening.
With this caveat in place, however, it does seem meaningful to relate the divine eye to a
deeper discernment of conditionality, since the textual sources explicitly connect the vision
of other sentient beings passing away and being reborn to insight into the operative causes
behind this process. The account of the Buddha’s own development of this type of knowledge
conforms to this pattern by relating ethical misconduct to bad rebirths after passing away and,
conversely, moral restraint to rebirth in pleasant circumstances (MN 4 and EĀ 31.3; trans. in
Ñāṇamoli 1995/2005, 106 and Anālayo 2016a, 21, respectively).
The basic principle of conditionality, evident in this way, corresponds to a key doctrine of
early Buddhism that can be considered the other side of the coin of not self: dependent arising.
The standard presentation of this principle takes the form of twelve links, which lead from the
root cause of ignorance to old age and death.
Besides the twelve-link formula, the teaching on dependent arising can take a variety
of different forms in shorter formulations, all of which serve as expressions of the same
principle. This principle constitutes another significant departure from the ancient Indian
philosophical setting. In line with the implications of not self, the doctrine of dependent
arising proposes an interplay of a series of causes and conditions without any permanent
entity or agent involved.
Although a novel conception, the frequent formulation of dependent arising in the form of
twelve links appears to be in dialogue with a Vedic myth depicting the creation of the world
(Jurewicz 2000). In contrast to this myth of creation, the Buddhist formulation describes the
conditioned genesis of the human predicament: old age and death. Moreover, it complements
this by pointing to the way out of this predicament, which requires the eradication of ignorance
and hence the interruption of the chain of conditions that lead from ignorance to old age and
death. In this way, a novel conception appears to have been couched in terminology familiar
to the ancient Indian audience, so as to be more easily assimilated.
20
Gotama Buddha
Awakening
Based on recollection of his own past lives (and hence probably a dawning insight into the
constructed nature of the sense of personal identity) and witnessing the passing away and
being reborn of others (and thus witnessing the cause and effect relationship that stands behind
the notion of karma), the third knowledge on the night of awakening marks the breakthrough
by which Gotama became a Buddha (MN 4 and EĀ 31.3; trans. in Ñāṇamoli 1995/2005, 106
and Anālayo 2016a, 22, respectively). The textual sources consider this breakthrough to corre-
spond to an eradication of all unwholesome influences (āsava/āsrava) in the mind. As a result,
the Buddha (and similarly, those of his disciples who subsequently reached the same level of
mental liberation by becoming arahants/arhats) was forever free from mental defilements like
sensual desire, anger, etc. With all craving and attachments removed, fully awakened ones are
considered to have gone beyond the confines of the predicament of saṃsāra. In other words,
they have attained the deathless.
The way to bring about such awakening (bodhi), sometimes translated as “enlightenment”
(Anālayo 2021), involves seven mental qualities known as the “factors of awakening”
(bojjhaṅga/bodhyaṅga). These are considered a specific discovery of the Buddha and hence
a novel contribution, although they were apparently soon adopted by other practitioners in
the ancient Indian setting (Gethin 1992, 177–83). The first of these factors of awakening is
mindfulness (sati/smṛti), a quality that in recent times has found widespread employment for
secular purposes in the clinical setting, education, and so forth. Its cultivation as a means for
progress to liberation takes the form of four establishments of mindfulness (satipaṭṭhāna/
smṛtyupasthāna), which the discourses present as the direct path to the realization of awaken-
ing (Anālayo 2003, 2013, 2018b).
Building on mindfulness as a mental quality that is required at all times, the remaining six
factors of awakening fall into two groups. One set of three serves to energize the mind and
is for this reason commendable when some degree of sluggishness manifests: investigation-
of-states, energy, and joy. The other set of three awakening factors calms the mind and is
therefore appropriate when some degree of excitement manifests: tranquility, concentration,
and equanimity. The cultivation of these seven awakening factors is a central teaching on
meditation in early Buddhist thought.
Teaching and Compassion

With awakening attained, the Buddha is on record for having hesitated to teach (MN 26; trans.
in Ñāṇamoli 1995/2005, 260). The relevant Pāli discourse reports that this hesitation prompted
the Indian creator god Brahmā to intervene and request that the Buddha share his discovery of
the path to awakening. Notably, this whole episode is absent from the relevant Chinese Āgama
parallel (MĀ 204; trans. in Anālayo 2012b, 32). This version proceeds directly from the Bud-
dha’s awakening to his reflection on whom he should teach first. Nevertheless, the Buddha’s
hesitation is a recurrent element in other discourses and later texts (Anālayo 2011, 178–79).
The rather substantial difference that emerges here can be interpreted in two ways: either
the episode was added to the relevant Pāli discourse (and also influenced other texts), or else
it was lost (or even removed intentionally) from the Chinese Āgama version (Anālayo 2012b,
34–37). Whatever may be the final word on these two interpretations, what remains certain is
that in the textual material common to parallel versions of the early discourses, the Buddha
was not considered to have intentionally prepared himself throughout many past lives for the
21
Bhikkhu Anālayo
task of becoming a compassionate teacher. The idea of having made a corresponding deci-
sion upon meeting a previous Buddha by the name of Dīpaṃkara is only attested in later texts
(Anālayo 2010, 87).
From the viewpoint of the early discourses, the Buddha was simply an exceptionally gifted
person who, on realizing the challenge of old age and death, decided to dedicate his life to
finding a solution for this predicament. Hence, the question of whether he would subsequently
share the solution he had found, by way of teaching others, appears to have been, at least at
the outset, not central to his quest. Needless to say, his actual teaching activity remains an
expression of his outstanding compassion. The point is only that his overall aim, in the way
this is recorded in the early textual sources, had been to liberate himself, rather than to become
a compassionate teacher.
The bodhisattva ideal, the aspiration to become a Buddha in a future life, is clearly a
later development. Its beginnings can be traced through comparative study of the early dis-
courses, which show how relevant elements and notions gradually appear in the course of time
(Anālayo 2010, 2017a).
Compassion itself, however, appears to have entered the stage only at a comparatively later
time. In other words, the original development of the bodhisattva ideal seems to have been
stimulated predominantly by the wish to become equal to the Buddha. This holds particularly
for wanting to acquire omniscience, which in the meantime had come to be attributed to the
Buddha, and for desiring to acquire the marks of his bodily perfection. The compassionate
motivation to deliver sentient beings is not evident in what appear to be the first stages in the
development of the bodhisattva path. Once such motivation had come to be combined with the
bodhisattva ideal, however, it swiftly gained traction.
Hence, from the viewpoint of the contribution made by Gotama to thought and practice in
the ancient Indian setting, the compassionate motivation to deliver sentient beings by undergo-
ing a long course of gradual preparation over a series of lifetimes should be accredited to later
generations.
The Four Noble Truths

With or without being prompted by Brahmā, the Buddha is on record for eventually deciding
to teach. The relevant discourse reports that he first wanted to share his discovery with his
former teachers, Āḷāra Kālāma and Uddaka Rāmaputta, only to find that they had recently
passed away (MN 26 and MĀ 204; trans. in Ñāṇamoli 1995/2005, 262 and Anālayo 2012b, 32,
respectively). He then decided to teach instead five companions from the time of his ascetic
practices. These five had left him upon seeing that he had given up asceticism, believing that
he had become lax and was no longer worthy of their support and companionship.
In order to convey to them that he had not just reverted to a life of sensuality, the Buddha
reportedly began teaching them by proposing a middle path apart from the two extremes of
sensual indulgence and ascetic self-mortification (SN 56.11 and EĀ 19.2, as well as other paral-
lels; trans. in Bodhi 2000, 1844 and Anālayo 2016a, 268, respectively; see also Anālayo 2015b,
347–88). The notion of such a middle path is a recurrent motif in early and later Buddhist
traditions and can be considered another significant contribution to the ancient Indian setting.
The same holds for the actual teaching reportedly given on this occasion, which concerns
the four noble truths. This teaching appears to have been modeled on an ancient Indian medi-
cal diagnosis (Anālayo 2015b, 25–40), which involves four aspects: the disease’s diagnosis,
its etiology, the prognosis, and the cure. Presumably, the Buddha found his realization to be
22
Gotama Buddha
so distinct that he decided to avoid taking up any of the philosophical notions known in the
ancient Indian setting. Instead, he preferred to rely on a medical precedent to formulate in a
pragmatic manner the key aspects of his realization.
In line with its medical precedent, the first noble truth identifies the human predicament.
The term employed here is dukkha/duḥkha, whose import is inadequately rendered as “suf-
fering.” Given that dukkha/duḥkha applies to all conditioned phenomena, a translation has to
be found that does not result in making the teaching itself become incongruous, such as when
pleasant experiences are flatly considered to be “suffering” or else when mountains and rivers
are qualified as “suffering.” They obviously are not in themselves suffering, but still they fall
within the range of applicability of the term dukkha/duḥkha.
Pleasant experiences are pleasant, but they fail to give lasting satisfaction. Mountains and
rivers can be perceived in a variety of ways, depending on the perspective taken, but they also
are incapable of providing lasting satisfaction. Hence dukkha/duḥkha, when employed in a
general sense applicable to all conditioned phenomena, is best rendered as “unsatisfactory.”
In contrast, “suffering” is not a quality shared by all conditioned phenomena, but a reaction of
the untrained mind.
The second noble truth identifies the pathogen responsible for the manifestation of dukkha/
duḥkha, which is craving (taṇhā/tṛṣṇā), in the sense of an incessant thirst for the satisfaction
of desires. The same can be stated in more detail by recourse to the whole chain of dependent
arising, starting with ignorance. This chain leads via several interim links to the experience of
feeling tones (vedanā), which can be pleasant, unpleasant, or neutral. Whereas such feelings
tones are a given of experience, the reaction to them by way of craving as the next link in the
chain is not. Hence, the presence of mindfulness at the juncture of feeling tones can enable the
prevention of reactions impelled by craving.
The third noble truth presents freedom from dukkha/duḥkha as the final goal. Such freedom
requires the eradication of both craving and ignorance, two complementary perspectives on
the root cause of the human predicament. Eradicating ignorance corresponds to the cessation
mode of the chain of dependent arising and thereby to the eventual cessation of old age and
death. Needless to say, even fully awakened ones, including the Buddha, still grow old and
pass away. But they no longer suffer because of that, as their minds are free of craving and
ignorance. Nor will they be reborn and experience old age and death in a future life. It is in
these two interrelated senses that the cessation of ignorance has brought about for them the
cessation of old age and death.
The fourth noble truth presents the path that leads to freedom from dukkha/duḥkha, which
serves as the conduit to the deathless (and birthless) realization of Nirvana. This path is quali-
fied as being eightfold, as it comprises eight dimensions of practice. The first of these requires
establishing the appropriate view (diṭṭhi/dṛṣṭi), in the sense of seeing things in the right per-
spective, namely, in such a way as to lead towards awakening. This first path factor is based
on an initial understanding of the four noble truths as a directive for practice, an understanding
that gradually matures until, with the breakthrough to awakening, the four noble truths are
fully understood and have become a matter of personal realization.
Informed by the directive set by the appropriate view, the next path factor concerns inten-
tions. These should involve the willingness to give precedence to renunciation over sensuality,
the absence of anger over irritation, and the presence of harmlessness over cruelty. The inten-
tional disposition established in this way then informs speech, action, and livelihood, which
need to be streamlined in such a way that they accord with these basic principles. Here and
elsewhere, ethics are a central concern of early Buddhist thought.
23
Bhikkhu Anālayo
The remaining three factors of the eightfold path are properly directed effort, mindfulness,
and concentration. These are interrelated aspects of mental culture, reflecting the importance
given to cultivation of the mind in the early Buddhist path to deliverance.
Conclusion
The chief contributions made by the Buddha Gotama to the ancient Indian setting, in the way
these are reflected in the early discourses, appear to involve the following aspects: the notion
of the deathless, bereft of connotations of immortality or permanent existence in a higher
realm, serves to express freedom from the terror of mortality as a result of reaching awaken-
ing. Deep concentration experiences are divested of any metaphysical connotations and seen
as merely products of certain conditions, whose cultivation leads to the corresponding experi-
ence. The final goal of Nirvana comes to signify a complete transcendence of saṃsāra such
that neither continuity of existence nor its annihilation fits the case, a notion that from ancient
India until today has not been easily understood by those who have not experienced it.
Progress toward this goal involves a stepping out of the pleasure/pain paradigm, in that an
evaluation of experience according to its hedonic quality is superseded by the ethical distinc-
tion between what is wholesome/skillful and what is unwholesome/unskillful as the proper
criterion. Moreover, an assessment along these lines focuses on the mind, in that intention
needs to be monitored, first of all, as it in turn influences actual deeds.
The teaching on the absence of a permanent self of any type, a radical departure from the
ancient setting, has its correlate in the doctrine of dependent arising. Both together explain
continuity of life beyond death without positing any form of a permanent entity, relying
entirely on the notion of a process of causes and conditions.
From the viewpoint of mental cultivation, a key innovation is the teaching of seven men-
tal qualities as factors of awakening, in the sense of their propensity to effect a waking up
to the true nature of reality, thereby issuing in liberation of the mind. The first of these is
mindfulness.
The central teaching of four noble truths that forms the framework for early Buddhist sote-
riology adopts a medical diagnostic scheme to pinpoint the causes for the human predicament,
together with identifying a viable solution and the practical path to be undertaken for that end.
In this way, central teachings of early Buddhist thought – which involve distinct contribu-
tions to Indian philosophy and soteriology – can be related to the Buddha’s own progress to –
and eventual gaining of – awakening.
Abbreviations
AN Aṅguttara-nikāya
DĀ Dīrgha-āgama (T 1)
DN Dīgha-nikāya
EĀ Ekottarika-āgama (T 125)
MĀ Madhyama-āgama (T 26)
MN Majjhima-nikāya
SĀ Saṃyukta-āgama (T 99)
SN Saṃyutta-nikāya
T Taishō edition
Vin Vinaya
24
Gotama Buddha
Notes
1 Although Drewes (2017) is right in pointing out that up to now, Buddhist studies have failed to
reconstruct with certainty “the historical Buddha,” he tends to overstate his case. His claim that “it is
not clear that the tradition itself envisioned the Buddha as an actual person” is not correct, as is the
case for his allegation that the “early texts, such as the suttas of the Pali canon, say hardly anything
about the Buddha’s life” (Drewes 2017, 16). The early discourses contain a range of details about the
Buddha’s life; otherwise, it would hardly have been possible for Ñāṇamoli (1972/1992) and myself
(Anālayo 2017d) to write whole books on this topic based just on the early texts. For critical replies
to Drewes (2017), see also Wynne (2019), Levman (2019), and von Hinüber (2019), with corrections
to the latter two in Anālayo (2019, 93f) and (2020, 21–25).
2 Laumakis (2013) is right in principle that it is problematic to employ the “received” biography of the
Buddha for reconstructing his thought, but in the course of making his case, he tends to overstate the
situation – comparable in this respect to the case of Drewes (2017), discussed in the previous note.
Although of course “the Buddha never wrote anything,” (Laumakis 2013, 13), we do have orally
transmitted records that can be studied by comparing parallel versions. Hence, the situation is more
nuanced than the simplistic assumption that “his supposed teachings were compiled anywhere from a
hundred to a few hundred years after his death.” It is not quite the case that it remains doubtful whether
anything at all “can be said with any degree of certainty with respect to what the man who became the
Buddha actually thought or taught.” In fact, a whole range of later developments of Buddhist thought
can with certainty be identified as such, making it possible to come to the definite conclusion that these
do not belong to what “the man who became the Buddha actually thought or taught.” Although I agree
with Drewes (2017) and Laumakis (2013) that it is important to recognize the limitations of the source
material at our disposition, it seems to me that excessive skepticism by overstating the situation, how-
ever fashionable it may be, risks obfuscating the potential of the material to which we do have access.
3 Bhikkhu Bodhi (2012) departs from the standard way of reference by using the numbering of the Pali
Text Society editions. As a result, in the present case, he refers to the relevant discourse as 3.39.
4 The implications of the first of these mental factors, vitakka/vitarka, are debated, making it difficult
to translate them without implicitly taking a position. Briefly stated, the question at issue is whether
it refers to plain “thought” or not. It does in other contexts, but a perusal of relevant material make
it in my view preferable to understand this term to convey the sense of an “application of the mind”
(Anālayo 2017c, 123–28).
References
Adikaram, E. W. 1946/1994. Early History of Buddhism in Ceylon, or ‘State of Buddhism in Ceylon as
Revealed by the Pāli Commentaries of the 5th Century A.D.’ Dehiwala: Buddhist Cultural Centre.
Anālayo, Bhikkhu. 2003. Satipaṭṭhāna, the Direct Path to Realization. Birmingham: Windhorse
Publications.
———. 2010. The Genesis of the Bodhisattva Ideal. Hamburg: Hamburg University Press.
———. 2011. A Comparative Study of the Majjhima-nikāya. Taipei: Dharma Drum Publishing
Corporation.
———. 2012a. “The Historical Value of the Pāli Discourses.” Indo-Iranian Journal 55: 223–53.
———. 2012b. Madhyama-āgama Studies. Taipei: Dharma Drum Publishing Corporation.
———. 2013. Perspectives on Satipaṭṭhāna. Cambridge: Windhorse Publications.
———. 2015a. “Āgama/Nikāya.” In Brill’s Encyclopedia of Buddhism, edited by J. Silk, O. von Hinüber,
and V. Eltschinger, vol. 1, 50–59. Leiden: Brill.
———. 2015b. Saṃyukta-āgama Studies. Taipei: Dharma Drum Publishing Corporation.
———. 2016a. Ekottarika-āgama Studies. Taipei: Dharma Drum Publishing Corporation.
———. 2016b. Mindfully Facing Disease and Death, Compassionate Advice from Early Buddhist Texts.
Cambridge: Windhorse Publications.
———. 2017a. Buddhapada and the Bodhisattva Path. Bochum: Projektverlag.
———. 2017b. Dīrgha-āgama Studies. Taipei: Dharma Drum Publishing Corporation.
———. 2017c. Early Buddhist Meditation Studies. Barre, MA: Barre Center for Buddhist Studies.
———. 2017d. A Meditator’s Life of the Buddha, Based on the Early Discourses. Cambridge: Windhorse
Publications.
25
Bhikkhu Anālayo
———. 2018a. Rebirth in Early Buddhism and Contemporary Research. Somerville, MA: Wisdom
Publications.
———. 2018b. Satipaṭṭhāna Meditation: A Practice Guide. Cambridge: Windhorse Publications.
———. 2019. Mindfully Facing Climate Change. Barre, MA: Barre Center for Buddhist Studies.
———. 2020. “The Tevijjavacchagotta-sutta and the Anupada-sutta in Relation to the Emergence of
Abhidharma Thought.” Journal of Buddhist Studies 17: 21–33.
———. 2021. “Awakening or Enlightenment? On the Significance of Bodhi.” Mindfulness 12.7: 1653–58.
———. 2022. “Situating Mindfulness, Part 2: Early Buddhist Soteriology.” Mindfulness 13.4: 855–62.
Bodhi, Bhikkhu. 2000. The Connected Discourses of the Buddha: A New Translation of the Saṃyutta
Nikāya. Somerville, MA: Wisdom Publications.
———. 2012. The Numerical Discourses of the Buddha: A Translation of the Aṅguttara Nikāya. Somer-
ville, MA: Wisdom Publications.
Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism. Cambridge:
Cambridge University Press.
Drewes, David. 2017. “The Idea of the Historical Buddha.” Journal of the International Association of
Buddhist Studies 40: 1–25.
Gethin, Rupert. 1992. The Buddhist Path to Awakening: A Study of the Bodhi-Pakkhiyā Dhammā. Lei-
den: E. J. Brill.
Jurewicz, J. 2000. “Playing with Fire: The Pratītyasamutpāda from the Perspective of Vedic Thought.”
Journal of the Pali Text Society 26: 77–103.
Laumakis, Stephen J. 2013. “The Philosophical Context of Gotama’s Thought.” In A Companion to Bud-
dhist Philosophy, edited by Steven M. Emmanuel, 13–25. Chichester: Wiley-Blackwell.
Levman, Bryan. 2019. “The Historical Buddha: Response to Drewes.” Canadian Journal of Buddhist
Studies 14: 25–56.
Liu, Zhen. 2010. Dhyānāni tapaś ca, 禅定与苦修. Shanghai: 古籍出版社.
Ñāṇamoli, Bhikkhu. 1972/1992. The Life of the Buddha According to the Pali Canon. Kandy: Buddhist
Publication Society.
———. 1995/2005. The Middle Length Discourses of the Buddha: A Translation of the Majjhima Nikāya.
Edited by Bhikkhu Bodhi. Somerville, MA: Wisdom Publications.
Salomon, Richard. 2018. The Buddhist Literature of Ancient Gandhāra: An Introduction with Transla-
tions. Somerville, MA: Wisdom Publications.
Vetter, Tilmann. 1995. “Bei Lebzeiten das Todlose Erreichen, Zum Begriff Amata im Alten Buddhismus.”
In Im Tod gewinnt der Mensch sein Selbst, Das Phänomen des Todes in asiatischer und abendlän-
discher Religionstradition, edited by Gerhard Oberhammer, 43–74. Wien: Österreichische Akademie
der Wissenschaften.
von Hinüber, Oskar. 2019. “The Buddha as a Historical Person.” Journal of the International Association
of Buddhist Studies 42: 231–64.
Walshe, Maurice. 1987. Thus Have I Heard: The Long Discourses of the Buddha. London: Wisdom.
Wynne, Alexander. 2019. “Did the Buddha Exist?” Journal of the Oxford Center for Buddhist Studies
16: 98–148.
26
2
SIDDHĀRTHA GAUTAMA
Beyond the Historical Figure
Richard F. Nance
The Limits of Provenance

Each chapter in this volume bears a proper name as its title. The name marks an author who
has made a distinctive contribution to Buddhist philosophy in India, and the chapters are meant
to provide readers with a sense of who these thinkers were as authors. What were the topics
they tended to favor? What was their own sense of the work they were doing, as thinkers, and
what should our sense of that work be? In each case, the name of the author serves as a center
of gravity, around which orbit topical, methodological, and stylistic concerns. The name of the
author indexes forms of consistency across texts, enshrining “relationships of homogeneity,
filiation, reciprocal explanation, authentification, or of common utilization” (Foucault 1977
[1969], 123).1 In cases when one finds it difficult to establish such relationships, one tends to
think there must be different authors at work.
Contemporary scholars are not alone in assigning texts to authors and authors to texts; tra-
ditional Buddhists have long undertaken this work, as well. Traditional attributions are some-
times trustworthy and sometimes troubling. When we are troubled by a traditional attribution,
we may have the sense that the text before us is not wholly consistent with what we think its
putative author would be doctrinally, stylistically, or historically disposed to say.2 Our judg-
ments of consistency are shaped by a complex set of assumptions as to what is possible and
probable, and these assumptions do not always match those of Mahāyāna Buddhist traditions.
So, for example, whereas Mahāyāna traditions have historically described buddhas as bear-
ing prophetic capabilities and superhuman attributes, we contemporary scholars have tended
to dismiss the veracity of such claims. Whereas the traditions have insisted that these bud-
dhas are multiple – that they manifest in countless miraculous ways in the past, present, and
future – we have tended to focus on a “founding figure” of the past: the figure of Siddhārtha
Gautama (alt. Śākyamuni Buddha). And whereas Mahāyāna traditions appear to attribute a
vast and internally diverse corpus of texts to this figure – often speaking of these sūtras as
“the word of the Buddha” or “the word of a buddha” (buddhavacana) – we have tended to
dismiss such attributions. The grounds for our dismissal have been various: doctrinal incon-
sistency, stylistic variance, historical anachronism. From a purely philosophical perspective,
the texts diverge wildly, with some appearing to endorse ontological realism; others, nominal-
ism or idealism; still others, skepticism. We take such divergences to warrant our judgments
DOI: 10.4324/9781351030908-5 27
Richard F. Nance
regarding provenance: the texts strike us as simply too different from one another to constitute
the teachings of a single figure.
Mahāyāna Buddhist traditions, for their part, have taken a rather different approach to
speaking of the/a buddha’s word. Rather than denying the obvious heterogeneity of texts
counted as buddhavacana, they have embraced and thematized it, often making reference
to nine, twelve, or 84,000 divisions of Buddhist teaching.3 They have also tended to resist
the conclusion that textual heterogeneity need imply multiple authorial hands (or mouths) at
work – differences between texts could be read as pointing not to a diversity of authors, but
rather to a diversity of audiences. After all, great teachers “read the room” and adjust their
message accordingly – a capacity that the tradition marks by speaking of a buddha’s unfail-
ing “skillful means” (upāyakauśalya) – and Siddhārtha Gautama was nothing if not a great
teacher. Emphasizing the profound malleability of Śākyamuni’s speech in this way, Mahāyāna
traditions manage to transform the very ground for contemporary scholarship’s attributional
denials into testimony for the Buddha’s pedagogical prowess.
Mahāyāna traditions have not, however, spoken with a single voice on this issue. Across
the centuries, questions over the provenance of texts traditionally counted as the word of the/a
buddha have been pressed not only by scholars of Buddhist studies, but also by Buddhists –
and these questions have left traces in the texts themselves. Consider, for example, the follow-
ing passage, drawn from the Questions of Rāṣṭrapāla (Rāṣṭrapālaparipṛcchā), which likely
reflects attributional anxieties. The speaker is purportedly the Buddha Śākyamuni himself, and
what he says is rather surprising:
When they hear this tranquil dharma, these ignoble ones, who are always hostile
to the dharma, who offend against the teaching, and who are devoid of virtues,
declare: “This was not spoken by the Victor.” My teacher was an ocean of knowl-
edge, very learned, the best of expounders of the dharma. And yet this sūtra was
forbidden by him: “It is by no means the word of the Buddha.” Moreover, he too
had an aged teacher, also possessed of an unlimited abundance of virtues, and this
sūtra was also not accepted by him: “Do not apply yourself to it; it is false.” They
charge: . . . “invented, imagined by the evil-minded and by those who think like
heretical teachers are the Mahāyāna teachings. The Victor could never have said
these words, which are a rebuke against monks.”4
Śākyamuni here situates the teaching of “this sūtra” – the very sūtra he is teaching – as hav-
ing been subjected to repeated dismissal in the past, suggesting that he is somehow at one and
the same time both the source and student of the Questions of Rāṣṭrapāla. This framing is,
at best, awkward; it is very hard to see how a text that records its own rejection by previous
generations of teachers could itself be the very same text rejected by those teachers.5 Problems
of reflexivity can be avoided, however, if we read this passage as presenting the incursion of
a later editorial voice – one concerned to recount moments in the text’s contested history of
transmission. The text may then be read as speaking figuratively of the circumstances of its
own journey – a very turbulent journey – through multiple generations of Buddhist teachers
who question its status as buddhavacana (Boucher 2008, 72). But if what we are encountering
here is a later editorial hand at work, then the conclusion cannot be avoided that Śākyamuni
Buddha did not teach the Mahāyāna sūtras – at least some of them, or some portions of them.
It is tempting to see in this passage a squabble over authorship between non-Mahāyāna
Buddhists and their Mahāyāna brethren. But the sūtra leaves room for the possibility that disa-
greements over provenance also erupted between Mahāyānists themselves. Daniel Boucher
28
Siddhārtha Gautama
has called attention to an irreducible diversity of viewpoints that characterized the nascent
Mahāyāna: what we see in the early Mahāyāna texts, he suggests, is not consensus, but rather
a variety of perspectives, indicating that “Mahāyāna sūtras must have been produced by indi-
viduals situated in vastly different social experiences” (Boucher 2018, 97). Such differences
likely affected attributional claims, as well: it is difficult to imagine that all Mahāyāna practi-
tioners agreed about the provenance of all Mahāyāna sūtras.
Perhaps, therefore, it was not only non-Mahāyāna Buddhists, but also Mahāyāna practi-
tioners (sometimes referred to as “bodhisattvas”), who wished to reject attributional claims
regarding Mahāyāna texts. The following passage, from Asaṅga’s fourth-century Bodhisattva
Stage (Bodhisattvabhūmi), suggests as much:
a bodhisattva who, after hearing the profound subjects contained in the bodhisattva
collection, fails to develop proper devotion toward those supremely profound sub-
jects that pertain to the nature of reality or to the powers of the buddhas and bodhisat-
tvas and denigrates them by saying, “These teachings do not possess any meaning,
they do not possess any genuine dharma, they were not spoken by the Tathāgata,
they do not provide benefit and happiness to sentient beings,” commits an offense,
a transgression; he or she acquires an afflicted offense (kliṣṭam āpattim), regardless
of whether he or she denigrates them because of his or her own improper manner of
thinking or because of his or her desire to comply with the views of others.6
Asaṅga seems to be bothered here by those who find themselves tempted to issue attributional
challenges to texts he favors. Precisely which texts are at issue is not entirely clear. As Ulrich
Pagel has noted (Pagel 1995, 17–36), the term “bodhisattva collection” (bodhisattvapiṭaka) is
used in various ways, some wider than others. It is, however, clear that the term is meant to
pick out one or more Mahāyāna texts; and it is no less clear that the attributional challenge dis-
cussed here is mounted from within the Mahāyāna fold(s), for Asaṅga uses the term bodhisat-
tva to label the person who mounts it. Battles over provenance sometimes pitted Mahāyānists
against the mainstream, but they also seem to have at times pitted them against one another.
It is wrong, Asaṅga suggests, for a Mahāyāna Buddhist to say that Mahāyāna sūtras were not
“spoken by the Tathāgata” (an epithet traditionally used for buddhas); to say this constitutes a
transgression against one’s bodhisattva vows.
The assumptions of provenance endorsed by the Buddhist tradition thus appear to be at
odds with the assumptions of provenance made by scholarship. Clearly, this presents a prob-
lematic situation for a would-be scholar-practitioner. When she takes a historical perspective,
the conclusion seems to her ineluctable: the Mahāyāna sūtras cannot be tied back to the time
of Śākyamuni – and that means, for her, that they cannot be properly classed as buddhava-
cana, much as Mahāyāna traditions might claim otherwise. That seems a responsible scholarly
position to take. But she also is a practitioner, and so she may be brought up short when she
finds statements in śāstric texts such as the Bodhisattvabhūmi that unambiguously suggest
that to challenge the status of the Mahāyāna sūtras as buddhavacana is to commit an offense.
The would-be scholar-practitioner of Mahāyāna traditions today would thus seem to be torn
between the demands of her profession and the demands of her religion. While she might reject
the traditional attribution of the Mahāyāna sūtras to the Buddha and preserve her academic
reputation, she would, in so doing, fail to act as a bodhisattva. Alternatively, she might endorse
the traditional attribution, thereby preserving her religious commitments but at considerable
cost to her reputation as an academic. The conflict between these demands cannot be amelio-
rated by attempting to compartmentalize one’s obligations. When one takes a bodhisattva vow,
29
Richard F. Nance
after all, one does not promise to work tirelessly to heal the sorrows of the world for as long as
the earth and sky shall last except when academic conferences are in session.
Of course, the problem sketched in the preceding paragraph need not trouble a scholar who
is not also a Buddhist. But by insisting that Mahāyāna traditions fail a historical test, she risks
committing a different sort of offense: the scholarly offense of failing to attend adequately to
the complexity of her sources. For Buddhist traditions affirm at least two separable approaches
to settling questions of provenance, only one of which looks to the past and sees teachings as
inheritances that have come down over centuries. A second account, by contrast, would tie
the transmission of Buddhist teachings not back to an awakened figure of the distant past, but
rather across to an awakened figure (or figures, or representatives of such figures) of the pre-
sent. This account manifests in various ways. We are told, for example, that there exist forms
of meditative concentration that afford practitioners the possibility of encountering myriad
“buddhas of the present” ( pratyutpannabuddha); or we are told that the real nature of buddhas
is atemporal and unchanging, and that, for this reason, a buddha can teach at any time via a
mutable “emanation body” (nirmāṇakāya).7
It is easy to dismiss these formulations as forms of magical thinking, but they raise real
interpretative and philosophical issues. To take such points seriously is to imagine our way
into a perspective from which the time of the/a Buddha is not only then, but also now – a per-
spective that is clearly different from the one that tends to inform contemporary scholarship
on Buddhism. This difference affects how one is predisposed to read the whole issue of prov-
enance. For if we grant the possibility of “tying across,” then whether a particular utterance is
properly to be counted buddhavacana is not simply a historical matter.
To see this more clearly, consider the relation between the following two questions.
1. Was p taught by an awakened being who lived a few centuries prior to the Common Era?
2. Should p properly be classed as buddhavacana?
Contemporary scholars tend to collapse these two questions into one. Historians of Buddhist
traditions are, after all, champions of “tying back,” and they are likely to view any attempt
to distinguish between (1) and (2) as ideologically suspect: end-runs around history are not
afforded by appealing to emanation bodies. And while philosophers may find traditional
claims of “tying across” to be metaphysically significant and intriguing, they are, in the end,
also likely to find those claims to be unwarranted.
However we understand the relation between (1) and (2), then – and whether we happen to
think of ourselves as historians or philosophers – we are likely to see traditional attempts to
assign the status of buddhavacana to Mahāyāna Buddhist sūtras as so much wishful thinking.
That might be wrong. In the sections to follow, I will make a case for a different way of read-
ing the relation between (1) and (2). In the reading I will propose, talk of “the word of the/a
buddha” need not be unpacked either in terms of “tying back” or in terms of “tying across”
(or, for that matter, in some uneasy combination of the two). As we will see, such talk need not
concern provenance at all.
This is a possibility that, I will suggest, has been hiding in plain sight in Mahāyāna Bud-
dhist texts. If we are to see it clearly, though, we will need not only to reread those texts, but
also to learn from them how to read them better. Doing this promises to advance us toward a
more historically adequate and – to my mind – more philosophically interesting view of what
Indian Mahāyāna Buddhists of the first millennium CE may have been up to in their talk of
“the word of the/a buddha.” In order to bring this view into focus, we need first to clarify how
30
Siddhārtha Gautama
we ourselves presently understand what it means to describe a text in this way. This is the task
of the next section.
Word: Phrasing and Meaning

What do we take buddhavacana to be? The way in which the Sanskrit compound has been
translated here – as “the word of the/a Buddha” – is meant to point to the fact that the compound
does not completely settle questions of provenance, for the term buddha may or may not index
Siddhārtha Gautama. I want to leave that point aside for a moment to explore another ambiguity
present in the compound: one that besets the term vacana, here translated as “word.”8 In order
not to overwhelm the reader with too many interpretive possibilities at once, I will, for the time
being, assume that the buddha in buddhavacana does index – at least – Siddhārtha Gautama,
and render the compound as “the word of the Buddha” throughout the remainder of this section.
How, then, do we understand the vacana in buddhavacana? What do we mean in speak-
ing of the Buddha’s “word”? Two interpretive paths immediately suggest themselves. On the
one hand, “word” could refer to the specific words and phrases ( pada, vyañjana) uttered by
Śākyamuni on some occasion of teaching. When the phrase is taken this way, then when we
deny that the Mahāyāna sūtras are the word of the Buddha, what we are denying is that the
words and phrases presented in the Mahāyāna sūtras as they exist for us today are the precise
words and phrases spoken by Śākyamuni. On the other hand, “word” could refer in a more
general way to what the Buddha taught: less to specific words and phrases than to the sense
(artha) conveyed thereby. When the phrase is taken this way, then when we deny that the
Mahāyāna sūtras are the word of the Buddha, what we are denying is that the doctrinal content
of the Mahāyāna sūtras corresponds to the doctrinal content of teachings that strike us as more
reasonably attributable to Śākyamuni.
These two ways of interpreting our denials are clearly different. The first interpretation
construes what counts as the word of the Buddha as a collection of words: these words are the
very words spoken by the Buddha, verbatim. I will call this a “locked-to-lexemes” view. In
a locked-to-lexemes view, the word of the Buddha is strictly unparaphrasable and untranslat-
able: although buddhavacana might be memorized or copied out, it cannot be summarized or
glossed without ceasing thereby to be buddhavacana.9
This is an extreme view, with extreme consequences. It calls into question not only the status
of the Mahāyāna sūtras, but also the status of any extant sūtra text, whether it be in Sanskrit or
Pāli. For although we do not know for certain which language(s) were adopted by Siddhārtha
Gautama for the purpose of teaching, it is generally accepted today that whatever linguistic
choices he made are not captured precisely in any version of the teachings that has come down
to us.10 Even so, a locked-to-lexemes view may initially appear to square with what we know
of the reception history of texts counted as buddhavacana within Buddhist traditions. These
texts were clearly memorized (√dhṛ) and collectively recited (saṃ + √gai) within early Bud-
dhist communities. Collective recitation would seem to require agreement on a specific set of
words, recited in a specific order. This can tempt us into thinking that these communities cham-
pioned something resembling a locked-to-lexemes view. But we need to be cautious here, for it
is possible to hold both that one must agree on a specific sequence of words to be memorized
for the purpose of recitation and that certain paraphrases of what one has memorized might
likewise count as buddhavacana. Indeed, there is evidence to suggest a normative traditional
understanding of buddhavacana according to which the meaning of what a buddha teaches is
to be granted equal, or even greater, importance than its lexical form (Lamotte 1988 [1949]).11
31
Richard F. Nance
This brings us to the second view – one that focuses on meaning, conceived as floating free
of specific turns of phrase. I will call this a “mere-meaning” view. When a person who holds a
mere-meaning view denies that the Mahāyāna sūtras comprise the word of the Buddha, she is
raising a worry not about the words of the Mahāyāna sūtras, but rather about their content, here
understood as something that could be paraphrased in more than one way, using more than one
succession of words and phrases. The worry is that this content fails to conform to anything
that Śākyamuni could possibly have meant.
A mere-meaning view may at first seem a more promising route for thinking about issues of
sūtra provenance. For one thing, it seems to hew more closely to what contemporary scholars
likely have in mind when they – we – issue their attributional challenges. What we have in mind
in denying Mahāyāna claims of provenance has to be more a matter of content than phrasing. As
already noted, most contemporary scholars are prepared to accept that the languages in which
the Buddhist teachings have come down to us do not conform in every respect to the language(s)
in which the Buddha is likely to have taught. Even so, many of us would like to be able to say
something about what the Buddha did (or did not) teach – and that suggests that we see Bud-
dhist teaching as amounting to (translatable, paraphrasable) messages, imperfectly recoverable
though those messages may be.12 A mere-meaning view may also look to be more readily com-
patible than its rival with claims about the teaching that are made in Buddhist texts. Consider,
for example, the fact that the Mahāyāna sūtras routinely advise those who endeavor to under-
stand the teaching of buddhas to “rely on meaning (artha, don), not on phrasing (vyañjana, tshig
’bru).”13 A strong reading of this advice would see it as suggesting that attention to phrasing can
be eschewed in favor of attention to meaning – precisely the sort of counsel we would expect
from a Buddhist community that had adopted a version of the mere-meaning view. Those seek-
ing support for this strong reading might look to an analogy voiced in the Sūtra which Incites
Resolve (Adhyāśayasañcodanasūtra), wherein the teaching is likened to sugarcane:
The sugar-cane bark has no juice at all;

The taste-giving delight is all found inside;
The man who chews only on the bark
Cannot find the sugar’s sweet taste.
Talking is like the husk of the cane;
But thinking on meaning is like the taste.
So renounce this joy in talking,
And be ever attentive and think on meaning.14
The practical appeal of the mere-meaning view can, however, lead us to ignore the real
problems it presents. For one thing, it does not square at all well with the traditional empha-
sis – noted previously – on the memorization and recitation of specific words and phrases.
Moreover, a close reading of the sugarcane analogy suggests that we may misunderstand it in
taking it to endorse a mere-meaning view – for the “husks” of talk to be discarded are not the
Buddha’s own words, but rather the words of those who would content themselves with talking
about matters they do not understand. The point of the analogy may thus not be that one need
not attend to the Buddha’s words. It may rather be something like the opposite: in lieu of trum-
peting one’s own superficially acquired knowledge, one should learn to read more carefully.
I would not wish to be read as claiming that no Buddhist communities ever embraced
a mere-meaning view – or, for that matter, a locked-to-lexemes view. Perhaps some did.
I do think, however, that the texts of the tradition suggest that neither of these views ade-
quately captures what is likely to have been a more common traditional understanding of
32
Siddhārtha Gautama
buddhavacana – and that we would do better to take our cue from Buddhist traditions on this
score. Unfortunately, however, we are not very good at bracketing our own assumptions long
enough to get a clear view of the data. When we are asked who is traditionally understood to
be the source of texts counted as buddhavacana, for example, the answer may seem patently
obvious: “the Buddha,” “a buddha,” or “buddhas.” Yet even the most cursory reading of Bud-
dhist sūtra texts – the paradigmatic texts of buddhavacana – suffices to show that this way of
answering the question cannot be correct, for the voices raised in those texts are not always
those of buddhas. This is a point that is at once obvious and routinely overlooked: a sūtra text
may, and typically does, feature multiple voices: the voice of the/a buddha, to be sure, but also
the voice of Ānanda (in the words “thus have I heard” that open many sūtras) and other char-
acters, including kings, brahmins, commoners, monks, nuns, male and female laypersons,
competing teachers – even the Buddha’s nemesis, Māra.15 We will come back to this point.
For now, it is enough to note that Buddhist traditions appear to be quite comfortable using
the term buddhavacana to label texts in which the voices of non-buddhas – and at times, the
voice of an anti-buddha – can intermittently be heard.16 This suggests that we may have gone
awry in thinking that what is at stake in claims to buddhavacana can be adequately captured
in talk of provenance. At the very least, it would seem that, on a traditional view, the word of
a buddha extends well beyond the word of a buddha.
The Well Spoken

What I have just suggested is that to call something buddhavacana may not be to call it the
word of a buddha. That claim, on its face, looks wildly implausible – for a straightforward
translation of the compound just is “the word of the/a buddha” (we will now return to the trans-
lation with which we began).17 How, then, are we to explain the fact that Buddhist traditions
apply the term to texts in which multiple voices can be heard? One explanation would be to
count such cases as instances of synecdoche (what in Sanskrit might be termed upalakṣaṇa).
Buddhist traditions, on this reading, are using a term that properly pertains to some part(s) of
a text – i.e., those parts in which the/a buddha speaks – for the purpose of naming the whole.
Reading the usage in this way enables us to hold on to the idea that buddhavacana must pertain
to provenance, but at the cost of finding the tradition habitually imprecise – even sloppy – in
how it marks texts. That cost strikes me as unreasonably high. Moreover, making this move
raises new questions that are not easily answered, for it is not the case that all texts in which
the voice of a buddha figures are traditionally counted as buddhavacana. Buddhist philosophi-
cal treatises, for example, often contain lengthy passages in which statements of the/a buddha
are quoted – but these treatises are not ipso facto “the word of the/a buddha.” Why would
Buddhist traditions hold synecdochic usage to be acceptable in certain cases, but not others?
There is an alternative explanation: one that would see the practice of labeling sūtra texts
as “the word of the/a Buddha” as amounting to a form of idiomatic usage – what might, in
Sanskrit, be termed lākṣaṇikaśabda. To my knowledge, the compound buddhavacana is never
explicitly marked as an instance of lākṣaṇikaśabda by commentators.18 While this might seem
enough to rule out the possibility of idiomatic usage, it must be remembered that we do not
always mark our idioms as such. The word “idiom” need not be in the vocabulary of an Eng-
lish speaker who uses “beat around the bush” to mean “fail to get to the point.” What idiomatic
usage requires of a speaker is not the knowledge of metalinguistic terms, but knowledge of
certain conventions that bear upon the use of a term or phrase.
What, then, were the operative conventions governing the use of buddhavacana in tradi-
tional contexts? A key to what they may have been could lie in an oft-quoted characterization
33
Richard F. Nance
of Buddhist teaching that echoes across various sūtra texts, according to which the teaching
of a buddha is “excellent in meaning and phrasing” (svarthaṃ suvyañjanam, Tib. don bzang
po tshig ’bru bzang po).19 What if this characterization captured something like a sufficient
condition for counting an utterance or text as the word of the/a buddha? Read in that way, a use
of the term buddhavacana would bear less on provenance than on assigning to a text certain
qualities traditionally held to characterize the speech of buddhas. That is close to assigning
provenance; it requires communal preunderstanding of how a buddha normatively speaks,
so worries over provenance cannot drop out completely. But it is importantly distinct, for it
suggests that “p is buddhavacana” need not imply “a buddha said p.” Rather, “p is buddhava-
cana” may imply that p is good in the beginning, middle, and end; that it comprises excellent
meaning and phrasing; that it is conducive to the attainment of one’s aims; that it has the
capacity to pacify suffering; and so on.20 In short, to call a text buddhavacana may amount to
a Buddhist way of saying that the text is well spoken (subhāṣita).
There is evidence to suggest that this interpretation is less farfetched than it may first appear.
Steven Collins has called attention to the slippage that occurs between an early claim, found
on an Aśokan inscription, that “everything spoken by a buddha is well spoken,” and a claim
that we find voiced in a text that we have already had occasion to cite: the Sūtra which Incites
Resolve, where we read that “whatever is well spoken is spoken by a buddha” (Collins 2013
[1990]).21 These are two very different claims, with very different consequences. According to
the first, the term subhāṣita applies to everything that a buddha happens to say: buddhas say
only what is well said. Note that this claim leaves room for the possibility that the speech of
those who are not (yet) buddhas might also be well spoken. The Sūtra which Incites Resolve,
by contrast, appears to shut down that possibility – for it looks to be committed to the claim
that only buddhas say what is well said. If that is right, then when we identify any particular
utterance as well spoken, we have identified its speaker as a buddha.
Before we assume that we have adequately understood what is going on in the Sūtra which
Incites Resolve, however, we need to explore the relevant passage in context. To do so is to
see a rather different reading emerge. As David Snellgrove has noted, the claim that whatever
is well spoken is spoken by a buddha is invoked multiple times in the text. Fortunately, one
of these invocations is extant in Sanskrit; it is cited in Śāntideva’s Anthology of Training
(Śikṣāsamuccaya) and Prajñākaramati’s Commentary on the Entry into the Practice of Awak-
ening (Bodhicaryāvatārapañjikā). The relevant passage reads:
Furthermore, O Maitreya, inspired speech ( pratibhāna) should be understood to be

spoken by a buddha, for four reasons. Which four? Here, O Maitreya, inspired speech
is meaningful, not meaningless; it accords with the teaching and does not contradict
it; it destroys afflictions and does not exacerbate them; it indicates the virtues and
advantages of nirvāṇa and does not exacerbate the faults of saṃsāra. All inspired
speech endowed with these four qualities should be understood to be spoken by a
buddha. When the inspired speech, Maitreya, of any monk or nun, layman or lay-
woman manifests with these four qualities now or in the future, faithful sons and
daughters will there be induced to conceive an idea of the Buddha (buddhasaṃjñā).
When an idea of the Teacher has been produced, the dharma is bound to be heard
(śāstṛsaṃjñāṃ kṛtvā sa dharmaḥ śrotavyaḥ). Why? Because, O Maitreya, whatever
is well spoken is spoken by a buddha. So, Maitreya, whoever would reject these
four forms of inspired speech and show no respect for them out of personal con-
tempt ( pudgalavidveṣena), saying “These are not words spoken by a buddha,” would
34
Siddhārtha Gautama
thereby reject all of the inspired speech spoken by buddhas. Rejecting the teaching,
he acts in a way that leads to the teaching’s ruin, and so he goes to hell.22
This passage integrates the remark we are working to understand into a discussion of utter-
ances characterized as “inspired speech” ( pratibhāna, Tib. spobs pa).23 Although the remark
itself does not refer to inspired speech, the framing unambiguously suggests that the epithet
“well-spoken” is to be taken as bearing only on speech that is considered inspired. Despite
appearances, therefore, we are not here faced with a universal claim about well-spoken speech.
To see this more clearly, consider a parallel case. Imagine that you are in a faculty meeting,
and a proposal is floated to which you strongly object. You are worried, however, that your
objections will not be preserved accurately in the meeting minutes. A colleague might attempt
to reassure you by saying, “Whatever is said is entered into the record.” That claim is per-
fectly straightforward within the relevant frame. Get rid of the frame, however, and the remark
invites puzzlement. How could the record of a faculty meeting contain everything that has
ever been said? Obviously, it cannot – and, importantly, no one has suggested that it could. The
contextual frame matters, in as much as it conditions what the words one has uttered can be
plausibly construed to say. When we attend to the claim from the Sūtra which Incites Resolve
within the contextual frame provided by the sūtra, we can see more clearly that despite appear-
ances to the contrary, the text is not claiming that anything well spoken is the speech of a bud-
dha. Instead, it is making two rather different claims, both of which bear on inspired speech. It
is claiming, first, that inspired speech is well spoken. And it is claiming, second, that inspired
speech is the speech of a buddha.24 The sūtra does not foreclose the possibility that an utterance
might be well spoken but not inspired – and thus it leaves room for the very same possibility
that one finds in the Aśokan inscription: even non-buddhas might sometimes speak well.
That said, there are still two claims made in this sūtra that look to be on a collision course –
and the two claims look specifically to concern provenance. One is that inspired speech arises
from buddhas; the other is that inspired speech may issue forth from men or women, monastics
or non-monastics. How is one to square these claims?
One way to do so is by reading the passage as claiming that any of these different sorts
of being might attain buddhahood and then go on to teach in inspired ways. That reading
reduces tension, but also goes well beyond what the text explicitly says. We may be helped in
our efforts here, however, by attending to another passage in the same sūtra that invokes the
phrase “whatever is well spoken is spoken by a buddha.” In this second passage, we read that
“bodhisattvas lacking in skillful means” (byang chub sems dpa’ thabs la mi mkhas pa) may
at times accuse others of “teaching the dharma according to their own inspiration” (rang gi
spobs pa chos ston par byed) . . . “in order to please themselves” (rang dgar byas pa), rather
than “teaching the doctrine which has its origin in the sūtra and vinaya.” Such accusations, the
sūtra suggests, can be dismissed. They are mounted by “fools who do not know that whatever
has been well spoken, all is the word of a buddha” (Snellgrove 1958, 623).25
As before, the phrase is invoked in the service of dismissing an objection. The grounds for this
dismissal are, however, left underdetermined. On the one hand, the sūtra could be read as accept-
ing the distinction that the objector presumes – accepting, that is, that a bright line can be drawn
between teachings that derive from one’s own inspiration and teachings that derive from bud-
dhas – but encouraging the objector (and us) to widen the latter category, allowing prima facie
problematic teachings to find a new home on the far side of the line. On the other hand, the sūtra
could be read as challenging the objector’s attempt to draw that line, and as encouraging him (and
us) to re-think the assumptions that look to warrant the attempt – one of which is attachment to
35
Richard F. Nance
the notion that the word of buddhas ought to be constrained to “that which has its origin in the
sūtra and vinaya.” For the sūtra has already suggested the shape of an alternative account: one
that would locate buddhavacana in inspired speech, and see its power to lie less in its perfect
recapitulation of something already said by a figure of the past than in its ability to induce what
the text calls an “idea of the Buddha” (buddhasaṃjñā) or an “idea of the Teacher” (śāstṛsaṃjñā).
What are we to make of these two compounds? As Peter Skilling has noted (Skilling 2009,
76), śāstṛsaṃjñā is most plausibly construed as “perception of the Teacher” or “perception as
the Teacher” (the term “Teacher” is generally held to refer to Śākyamuni – thus the capitali-
zation). Roughly the same might be said about the compound buddhasaṃjñā: “perception of
the Buddha” or “perception as the Buddha.” Beyond these observations, however, we enter
speculative territory, for the Mahāyāna sūtra texts that put the compounds to work tend not to
unpack them. To go further, we must look beyond these texts to see how the terms are handled
in Buddhist philosophical and narrative literature – keeping in mind that whatever picture of
the meaning of these terms we end up developing may not conform to how the terms have been
understood across the various Buddhist communities through which the Mahāyāna sūtras have
historically passed. What is crucial for our purposes is to clarify the term saṃjñā (Pāli saññā,
Tib. ’du shes) – a term that can also be rendered into English as “apperception,” “recognition,”
“idea,” or “conception.” I will adopt the last of these translations in the discussion to follow.
The term saṃjñā, commonly used to index one of the five aggregates invoked in classical
Buddhist discussions of experience, is succinctly defined in Vasubandhu’s Treatise on the Five
Aggregates (Pañcaskandhaka). It is, Vasubandhu tells us, “the grasping of an object’s sign”
(viṣayanimittodgrahaṇam) (Li and Steinkellner 2008, 4).26 Sthiramati’s commentary elaborates:
“Object” (viṣaya) means a mental object (ālambana). “Sign” (nimitta) means that
quality of the object (viṣayaviśeṣa) that serves as a cause for determining that it is
blue or yellow, etc. Thus, to grasp a sign means to make the determination “This is
blue” or “This is yellow,” and the like. . . . While all saṃjñās have the same nature,
they should be understood to be different owing to their basis and mental object.27
Sthiramati goes on to expand on some of the terms that he has used to unpack Vasubandhu’s
definition. He notes that the term “basis” should be understood to refer to the six sense-bases (of
the eye, ear, nose, tongue, body, and mind), and that the term “mental object” ranges over four
types of mental objects, only two of which need concern us here.28 The first of these is an object
that is “associated with signs” (sanimitta), where “sign” picks out “that which is ascribed to
an object’s nature on the basis of words” (śabdadvāreṇa yad artharūpam adhyāropitaṃ). The
second is an object that is “without signs” (animitta), lacking such an ascription (Kramer 2013,
Vol. 1, 30). It is not clear into which of these categories Sthiramati would place an instance
of buddhasaṃjñā or śāstṛsaṃjñā, but whether or not the object of a saṃjñā is associated with
signs, its arising would seem to put one in a position to identify (what we would call) a certain
intentional content as occurrent. On this understanding, a buddhasaṃjñā would then be – at
least – a mental episode in which a buddha-shaped content, identifiable as such, is presented.
When would such a saṃjñā arise? The most obvious occasion would be when one is in the
presence of a buddha. We find just such a circumstance recounted in the voluminous Pāli Leg-
ends of the Buddhist Saints (Apadāna), when a saint named Raṃsisaññaka offers the following
memory report of an encounter that results in a conception of the Buddha (buddhasaññā):
[I saw] the superb Tiger-Bull,

Well-Born-One, on a mountainside,
36
Siddhārtha Gautama
like the rising hundred-rayed [sun],

like the sun [when its] rays are cool.
The majesty of the Buddha
was shining on the mountainside.
[My] heart pleased in the rays; for an
aeon I sported in heaven.
In [all] the subsequent aeons,
goodness was completed by me
because of that mental pleasure,
and remembering the Buddha.
In the thirty thousand aeons
since I obtained that conception then,
I’ve come to know no bad rebirth:
The fruit of conceiving buddhas.29
The conception of the Buddha noted here clearly possesses significant apotropaic properties.
For thirty thousand eons, Raṃsisaññaka tells us, he has experienced no negative rebirth.
Protection from harm, then, is one fruit of conceiving buddhas (buddhasaññāyidaṃ
phalaṃ) – and protection is a quality repeatedly associated with buddhasaññā, even as other
stories from the same collection suggest that the circumstances for its arising may vary. In
one of these stories, a saint named Saddasaññaka reminisces over an occasion on which he
too came into the presence of the Buddha. Saddasaññaka does not see the Buddha but instead
hears his voice – enough, the text suggests, for the arising of a conception of the Buddha that
Saddasaññaka later calls up when it comes time for him to die:
When the time of my death arrived,

I called to mind that conception of the Buddha.
In the ninety-four aeons since
I obtained that conception then,
I’ve come to know no bad rebirth.
Both of these stories suggest that conceptions of the buddha can be generated in the presence
of buddhas, whether that presence is registered visually or aurally. And both stories suggest
that conceptions of the buddha, once generated, can later be remembered, and thus afford
access to that which is not immediately present. We might still wonder, however, about the
circumstances that must be in place for an apotropaic conception of the buddha to be initially
generated. For we might reason that although countless people have brought to mind a concep-
tion of the buddha, the practice of doing so does not seem to have unfailingly protected them.
Is the immediate presence of a buddha required for that?
Perhaps not. A further story from Legends of the Buddhist Saints tells of a master of the
three Vedas named Buddhasaññaka, who hears his students report that a buddha has arisen in
the world. Wishing to see him, Buddhasaññaka sets off from his hermitage. On the cusp of
completing his journey, however, he falls gravely ill:
When a league and a half was left

To go, illness arose in me.
Calling to mind the best Buddha,
I passed away right on the spot.
37
Richard F. Nance
In the ninety-four aeons since

I obtained that conception then,
I’ve come to know no bad rebirth:
The fruit of conceiving buddhas (buddhasaññāyidaṃ phalaṃ).
Note that Buddhasaññaka has heard from his students about the Buddha; he has not yet had the
good fortune to encounter the Buddha directly. Even so, he comes into possession of a concep-
tion of the Buddha that works just a well as a conception prompted by direct encounter. An
effective buddhasaññā can, it seems, arise on the basis of hearsay; thus have I heard.30
Let us return to the question that prompted this digression on the conception of the Buddha:
the question of whether “the word of the/a buddha” (buddhavacana) might be understood as
an idiomatic Buddhist way of branding certain utterances as “well-spoken.” As we have seen,
this interpretation is on the right track, but needs refining: “the word of the/a buddha” is bet-
ter construed as an epithet for inspired speech – which is to say, speech that has the power to
induce a buddhasaṃjñā.31 The story of Buddhasaññaka helps us to see that speech of this kind
need not be issued by a buddha. Paradoxical as it may sound, therefore, the word of the/a bud-
dha need not be a buddha’s word.
It may even be the word of an anti-buddha. In a passage at the close of one of the sūtras of
the Mahāyāna Heap of Jewels (Ratnakūṭa) corpus – the Sūtra of Instruction on the Inconceiv-
able Sphere of Buddhas (Acintyabuddhaviṣayanirdeśasūtra, Tib. sangs rgyas kyi yul bsam gyis
mi khyab pa bstan pa’i mdo) – we find Māra making the following promise (the translations
offered in the following are slightly modified from Chang et al. 1983, 35–36):
I, Pāpīyān the evil one, have always sought opportunities to oppose buddhas and to
create turmoil among sentient beings. Now I vow that, from this day on, I will never
go nearer than one hundred leagues away from the place where this doctrine prevails,
or where people have faith in, understand, cherish, receive, read, recite, and teach it.
Māra immediately goes on to note that his promise cannot, by itself, end all the problems that
the tradition labels “deeds of Māra.” Although Māra is now personally inclined to keep his dis-
tance, he warns the Buddha that “some of my kindred are determined to distract the devotees’
minds so as to destroy the teaching of buddhas.” That worry is, however, swiftly solved – and
by none other than Māra himself, who says:
I will chant the following spell so that devotees can vanquish these demons. If good
men or good women read, write, and recite this incantation, or teach it to others, the
celestial demons will benefit and will, in return, cause the teachers of this teaching to
feel joyful in body and mind, to practice vigorously, to possess unimpeded inspired
speech and spells, and not to lack services, food and drink, clothing, bedding, or
medicine.
The mantra is then provided. The Buddha applauds Māra for his efforts and notes that Mañjuśrī
should be credited for Māra’s eloquence: “ ‘Wonderful, Wonderful! You should know that all
your inspired speech (khyod kyi spobs pa) is a manifestation of Mañjuśrī’s awesome power.’ ”32
We see here something like the same move as we have encountered in the Sūtra which Incites
Resolve: Māra’s inspired speech, though marked as his own (it is, the Buddha notes, “yours”
(khyod kyi), also manifests a power associated with another: the bodhisattva of wisdom. And
while the text does not go so far as to brand Māra’s utterance an instance of buddhavacana,
38
Siddhārtha Gautama
the text from which this passage is derived is unproblematically counted as such by Mahāyāna
Buddhists more generally.33
Beyond Provenance: Untying

We can now return to a worry that we first confronted at the conclusion of the first section
of this chapter: the worry about whether claims regarding provenance that are made by con-
temporary scholars are actually at odds with the ways in which Buddhists have historically
applied the notion of “the word of the/a buddha.” When a contemporary scholar denies that
the specific teachings that Mahāyāna Buddhists brand buddhavacana were taught by a figure
who lived during the fifth–fourth century BCE, the scholar is clearly making a claim about
provenance. When, on the other hand, Asaṅga decries the notion that a particular teaching is
not the speech of the Tathāgata (na tathāgatabhāṣitā), what he is likely objecting to is the idea
that that text does not comprise buddhavacana. But the latter category, as we have seen, may
not carry any implication regarding provenance. To call an utterance (or text) buddhavacana
may amount to no more than branding it an instance of inspired speech ( pratibhāna). Inspired
speech is not restricted to buddhas – much less to a single figure who purportedly lived during
the fifth–fourth century BCE. It may, according to the Sūtra which Incites Resolve, be spoken
by “any monk or nun, layman or laywoman.” What Asaṅga deems an offense may thus not be
the move familiar to us from contemporary scholarship, but something else; what he is worried
over may be less a denial of provenance than a denial of liberating potency. And that sort of
denial is not the sort of denial that the contemporary scholar is generally concerned to make.
This will likely come as a relief to the would-be scholar-practitioner.34 But I hope that it will
also provoke those who are not practitioners to re-think their own assumptions about what
Buddhists of the past may have been up to in speaking of buddhavacana.
What, then, are the more general takeaway points of this, concerning the figure whose
name graces this chapter as its title? If we think of “Siddhārtha Gautama” as naming a histori-
cally placed individual who lived during the fifth–fourth centuries BCE, then much of what
I have been arguing is likely to seem irrelevant – for we may seem to have learned little as to
how to conceptualize the unique bearer of that proper name. To focus on the unique bearer
of the name, however, risks missing what the tradition itself suggests to be worth pursuing –
and I have tried to take my bearings from tradition on this score, downplaying notions of
uniqueness in favor of thinking more broadly about buddhahood, texts, and what is at stake
when questions of provenance are raised. Buddhist traditions, and particularly Mahāyāna tra-
ditions, are not terribly concerned to demarcate the Buddha as a single individual: the name
Siddhārtha Gautama applies not to the, but a, buddha. Buddhas are legion, and the stories
spun about them are stereotyped to a degree that suggests their deep similarity, if not inter-
changeability. “Among buddhas,” the Prophecy for the Magician Bhadra (Bhadramāyākāra-
vyākaraṇasūtra) reminds us, “there is no distinction whatsoever” (D. dkon brtsegs ca 26b6:
sangs rgyas rnams la tha dad bya ba’ang med).
Recall the two questions that were initially raised in the first section of this chapter.
1. Was p taught by an awakened being who lived a few centuries prior to the Common Era?
2. Should p properly be classed as buddhavacana?
Where p = the corpus of Mahāyāna sūtras, a responsible scholarly answer to the first question
must be negative, but a responsible scholarly answer to the second question need not be, if one
is capable of hearing the question in a certain way. The second question may be asking whether
39
Richard F. Nance
a teaching, regardless of its source, should be counted as inspired speech ( pratibhāna) – i.e.,
speech contributing to the production of an idea of the buddha (buddhasaṃjñā).
Such an “idea of the buddha” could, of course, be an idea bearing on provenance (“this
teaching is properly attributable to Siddhārtha Gautama”). But as we have seen, it need not be.
Alternative ideas of the buddha could be, e.g., “The being now before me is a buddha”; “I once
encountered a buddha”; “I would like to come into the presence of a buddha.” These ideas
might arise from the inspired speech of various speakers, none of whom need actually be a
buddha. If we understand buddhavacana to be another name for inspired speech, then the way
is open for us to hold that it is like emptiness and buddha nature: not restricted to buddhas, and
still less to Siddhārtha Gautama alone. This way of interpreting “the word of the/a buddha” is
less a form of “tying back” or “tying across” than it is a form of untying – one meant to free us
from attachment to the thought that buddhas, whether past or present, can only be elsewhere.
In this chapter, I have tried to show that scholars today, whether Buddhist or non-Buddhist,
may have much to learn from Buddhist traditions about how to understand what can at first
appear to be straightforward attributional claims. I do not wish to suggest that the phrase “the
word of the/a buddha” cannot be used to speak of provenance; that is, after all, the most natural
and obvious way to read it. But it is not the only way. We can learn from the tradition how
to read it otherwise. Moreover, the work of doing so is philosophically valuable, in as much
as it encourages us to attend to, and perhaps to reconsider, our own commitments, whether
those commitments be scholarly or existential. It is possible, of course, that the vision of bud-
dhavacana proposed here will turn out, on further reflection, to be no more than “an illusion, a
mirage . . . a dream . . . an apparition . . . a magical creation.”35 If so, then we can at least take
solace in the fact that, in a Buddhist view, that vision will find itself in good company.
Notes
1 The bibliography on concepts of authorship is, of course, immense. For orientation to some central
issues, see Burke (1995); Chartier (1992); Griffin (1999); Haynes (2005); Jaffe (2007); Rosenbaum
(2006). Studies that take up conceptions of authorship as they bear specifically on Indian Buddhist
texts include Cabezón (2001); Cantwell et al. (2013/14) [2015]; Davidson (1990); Gold (2015, Ch.
1); Kapstein (1989, 2018); Kramer (2016).
2 This way of putting the point makes our assessment of a text’s content and aims primary: it is the
assessment of content that puts us in a position to make judgments as to the veracity of traditional
attributions. But it is also true that what we assume about a putative author may itself influence what
we take a text’s content and aims to be. Consider, as a case in point, the Hymn in Praise of Reality
(Dharmadhātustava), a work traditionally attributed to Nāgārjuna. The authorship of this text has
been the subject of scholarly debate (see Liu 2015, ix), and where we stand on this debate may shape
how we read the text. The content and aims of Hymn in Praise of Reality-authored-by-Nāgārjuna
are arguably not the same as the content and aims of Hymn in Praise of Reality-not-authored-by-
Nāgārjuna: We are far more tempted to read the former as deploying a series of similes borrowed
from buddha nature (tathāgatagarbha) discourse in order to affirm the emptiness of the real, and
the latter as invoking ideas associated with the perfection of wisdom literature in order to affirm the
eternality of the real.
3 This move is not restricted to the Mahāyāna. For summary remarks, see Heim (2018, 35–38).
4 Translation slightly modified from Boucher (2008, 137). In keeping with the conventions of this
volume, Boucher’s use of square brackets has been eliminated, and capitalization has been adjusted.
Relevant Sanskrit may be obtained by consulting the Göttingen Register of Electronic Texts in Indian
Languages (GRETIL), at http://gretil.sub.uni-goettingen.de/.
5 We might note, in passing, that attempting to answer worries of the sort identified here by appealing
to a buddha’s unimpeded knowledge across the three times – i.e., by appealing to prophecy – would
come at the cost of introducing new and equally intractable interpretive difficulties. For if I, as a
teacher, were to disparage a sūtra only to discover, upon reading it carefully, that it has precisely
40
Siddhārtha Gautama
anticipated the terms of my disparagement, I would likely no longer disparage it. I would, however,
then need to find a way of explaining how the text fails to anticipate my own change of heart with
regards to it. (Perhaps noticing this failure would itself suffice to convince me once again that the text
is rightly to be disparaged. Should its prophecy then be counted as fulfilled?)
6 Translation slightly modified from Engle (2016, 291). For the Sanskrit, see Dutt (1966, 119).
7 See Harrison (1990, 2003); Griffiths (1994); Makransky (1997).
8 The reader may believe that the discussion to follow is merely tangential to questions of provenance,
in as much as the same string of words can be uttered, or the same content conveyed, by more than
one person. This means that getting clear on what is said is not likely to advance one very far toward
answering the question of who does the saying. In the case of Buddhist sūtras, however, questions of
sourcing cannot be neatly separated from questions of phrasing and meaning, in as much as who does
the saying can come into view for us only on the basis of what is said. How we understand “word”
in “word of the Buddha” thus matters to settling issues of provenance, in as much as it shapes our
judgments of divergence. If we are prone to construe “word” in terms of lexical form, we are liable
to see differences between texts that we might fail to see if we construe “word” in terms of semantic
content – and vice-versa (for just as divergence in form need not imply divergence in content, diver-
gence in content need not imply divergence in form).
9 In a famous passage from the vinaya literature, the Buddha instructs his interlocutors that his words
may be rendered “with one’s own [alt. my own] expression” (sakāya niruttiyā). How this phrase is
properly to be understood has, however, been the subject of considerable scholarly debate in the wake
of Brough (1980). For a useful summary, see Levman (2008–2009).
10 For discussion of the debate, see Levman (2016). For more recent salvos, see Gombrich (2018);
Karpik (2019).
11 Pace Levman, who writes (Levman 2008–2009, 40):
It has long been appreciated that Buddhist monks were expected to memorize and recite bud-
dhavacana as an integral part of their practice . . . [y]et the fact that this contradicts the conven-
tional understanding of Cullavagga v. 3 – that each monk was to learn, and presumably recite the
dharma in his own dialect – has not been noted. For if everyone were reciting something different,
then there is no root text to memorize.
The final claim here is decidedly peculiar. Consider a situation in which different Buddhist communi-
ties memorize and collectively recite slightly different root texts. Within one such community, there
may be collective consensus on what the words of the root text are; within another such community,
there might equally be collective consensus on what the words of the root text are. Under such circum-
stances, the claim that there is “no root text to memorize” would strike the members of any of these
communities as peculiar – for there is a root text in and for that community. When members of the
community have memorized this root text, the practice of collective recitation can proceed smoothly;
the fact of inter-communal divergence is of no consequence. Things might, of course, strike us – those
of us in a position to note inter-communal divergence – rather differently. We might be troubled by
inter-communal divergence in a way that the Buddhist communities we study were not. It should be
noted, though, that Indian Buddhists were not blind to such divergence; it is, for example, explicitly
acknowledged by Vasubandhu in the opening pages of the Logic of Explication (Vyākhyāyukti), when
he notes that reciting “good in meaning, good in phrasing” (svarthaṃ suvyañjanam) is to be preferred
over reciting “possessed of meaning, possessed of phrasing” (sārthaṃ savyañjanam) in a well-known
pericope pertaining to the teaching of buddhas. (We will come back to this pericope later; for a fuller
translation of the context for this passage, see Nance [2012, 131]). Even so, it seems a stretch to think
of the situation sketched here as one in which there is no root text to memorize. If anything, there are
multiple root texts in play.
12 Representative here is Gombrich (2009, 3):
a successful interpretation of the Buddha will make clear not only the ideas he expressed but
also how those ideas lent themselves to the various interpretations which are in fact historically
attested. The Buddha will thus stand as the source for a successful history of Buddha ideas – even
though to compose such a history, even in outline, may be beyond the powers of any single scholar.
13 This advice – one of the so-called “four reliances” (catuḥpratisaraṇa) – is attributed to a buddha in numer-
ous Mahāyāna sūtras, and it is repeated in śāstric texts that span centuries, from the Śrutamayībhūmi
section of the Yogācārabhūmi (c. fourth century) to Ratnākāraśānti’s Ratnāloka (eleventh century).
41
Richard F. Nance
14 Translation slightly modified from Sherburne (2000, 199). This passage is cited in the fifth chapter of
Atiśa’s Commentary on the Lamp for the Bodhisattva Path (Bodhimārgapradīpapañjikā) (Sherburne
2000, 196). The relevant Tibetan (corrected from Sherburne 2000, 196, 198) here reads:
bu ram shing shun snying po ci yang med / dga’ bar bya ba’i ro ni nang na ’dug / shun pa zos pa mis
ni bu ram ro / zhim po rnyed par nus pa ma yin no / shing gi shun pa de bzhin smra ba ste / ro lta bu
ni ’di ni don sems yin / de ltas bas na smra la dga’ spangs te / rtag tu bag yod byos la don sems shig.
Cf. also the *Mahāyānapathasādhanavarṇasaṃgraha (D, dbu ma khi, 299a5–302b6), which offers a
very similar verse.
15 This point regarding the presence of multiple voices could, of course, be expanded by attending to
voices that surface in various ways across the text’s history of transmission, i.e., the voices of interpola-
tors, compilers, redactors, scribes, orators, teachers, etc., responsible for working with texts over time.
16 This multiplicity of voices does not seem to be a late introduction to the scriptural corpus. That is, we
have no reason to suppose that the use of the term buddhavacana is an atavistic holdover from a time
when Buddhist sūtra texts were taken to comprise nothing more than soliloquies of the Buddha (or,
for that matter, Ānanda). As far as we are able to tell, there was no such time.
17 While explicit analyses – vigrahas – of the compound buddhavacana that survive in Indic languages
are hard to come by, the phrase buddhasya vacanaṃ is attested more than once – and the Pāli canon
is riddled with occurrences of the comparable buddhassa sāsanam (“the teaching of the/a buddha”).
The relevant Sanskrit shows up in the Legend of Aśoka (Aśokāvadāna) and elsewhere, and variants of
the Pāli phrase buddhassa sāsanam show up nearly two hundred times across the canonical and post-
canonical materials assembled in the Chaṭṭha Saṅgāyana collection, searchable at https://tipitaka.org/
search. The compound veritably begs to be construed as a straightforward subjective genitive (ṣaṣṭhī
tatpuruṣa) construction: to call something buddhavacana is simply to say that it is a buddha’s speech.
18 A global account of language as constitutively metaphoric (Tzohar 2018) would arguably have little
trouble accommodating this idea. That said, the interpretation of buddhavacana I am forwarding here
does not require assent to such a theory.
19 The Pāli canonical variant is sātthaṃ sabyañjanam. Literally, this means “possessed of meaning and
phrasing,” but commentators in the Pāli tradition take pains to note that what is meant by this is that
both the meaning and the phrasing of a buddha’s teaching are possessed of magnificence or excel-
lence (sampatti) – and it is this idea that is captured more literally in the Sanskrit version, which is
widely attested across Mahāyāna and non-Mahāyāna texts.
20 These characterizations are paraphrased from an oft-cited pericope describing the teaching of buddhas,
discussed at the opening of Vasubandhu’s Vyākhyāyukti. For a translation, see Nance (2012, 130).
21 The first claim occurs in the Bhabrā inscription of Aśoka (Collins 2013 [1990], 7): e keci bhaṃte
bhagavatā budhena bhaṣita save se subhāsite vā. The second claim will be discussed at length in
what follows; a parallel in Pāli occurs at AN 4:164.
22 Translation modified from Snellgrove (1958). Snellgrove offers a Sanskrit text that corrects Bendall’s
edition of the Śikṣāsamuccaya on the basis of the Narthang edition of the Tibetan. As already noted,
much of the relevant Sanskrit shows up also in Prajñākaramati’s commentary to verse 9.42 of the
Entry into the Practice of Awakening (not consulted by Snellgrove), but the Sanskrit that survives
in Prajñākaramati’s text does not entirely match that which survives in Śāntideva’s text – and both
diverge in certain respects from the Tibetan translation of the sūtra (which nevertheless presupposes
an underlying text closer to Prajñākaramati than to Śāntideva).
23 See MacQueen (1981, 1982); Braarvig (1985); Nance (2012, Chapter 2).
24 Again, a parallel case might be instructive. I tell you that all ravens are black; I tell you that all ravens
are birds. I have not licensed you to conclude that I hold everything black to be a bird.
25 Cf. a parallel Pāli passage at AN 4.164, where the remark that whatever is well-spoken is spoken by a
buddha ( yaṃ kiñci subhāsitaṃ sabbaṃ taṃ tassa bhagavato vacanaṃ arahato sammāsambuddhassa)
is made in response to much the same worry – i.e., a worry over whether a teaching that has not
been heard directly from the mouth of the Buddha should be thought to “derive from [the speaker’s]
own inspiration” (sakapaṭibhānaṃ upādāya) or whether it instead should be thought to present “the
word of the Blessed One, the Arahant, the Perfectly Awakened One” (bhagavato vacanaṃ arahato
sammāsambuddhassā). In the Pāli text, this worry is answered with a simile: we are asked to imag-
ine the teaching of the Blessed One as akin to a great heap of grain from which others can draw
sustenance. We are further asked to imagine a situation in which some persons who have retrieved
a portion of this grain are stopped and asked about its source: “Where did you get this grain?” The
correct answer to this question is, the sutta tells us, “We got it from that great heap of grain.” It is in
42
Siddhārtha Gautama
precisely the same way, the sutta goes on to note, that we should understand all that is well-spoken
to be the word of the Buddha, and that “I myself and others derive our good words from him” (tato
upādāyupādāya mayaṃ caññe ca bhaṇāmā ti). One might hear in the analogy an echo of the Vedic
likening of human knowledge to handfuls of dust (Taittirīya Brāhmaṇa 3.10.11.3–5). Precisely how
we are to trace out its implications is not clear; what is clear, though, is that the analogy does not sug-
gest that one who speaks well must for that reason alone be understood to be a buddha.
26 For more on the shifting conception of saṃjñā in Buddhist thought, see Kumagai (2019).
27 Translation slightly modified from Engle (2009, 271). For relevant Sanskrit, see Kramer (2013, Vol.
1, 29). As Engle notes, the opening portion of the passage cited here is quite similar to a discussion of
saṃjñā found in the Triṃśikabhāṣya (Buescher 2007, 56).
28 The remaining two are “those that are immeasurable” (tshad med pa, apramāṇa) and “those that are
associated with the Sphere of Nothingness” (ci yang med pa’i skye mched, ākiṃcanyāyatanam). For
our purposes here, these may be set aside, as they occur only in advanced meditative states.
29 All translations from Legends of the Buddhist Saints have been slightly modified from Walters (2018).
30 The well-known tale of an encounter between Upagupta and Māra that is related in the Legend of
Aśoka (Aśokāvadāna) is also relevant here. In this story, which has been translated and discussed at
some length by John Strong, Māra appears in the guise of a buddha (Strong 1992, 105–9, 192–96).
At the sight of Māra in this form, Upagupta’s mind is overwhelmed by a buddhasaṃjñā, and he
falls prostrate, “like a tree cut off at the root.” Upagupta’s buddhasaṃjñā is clearly prompted by an
encounter not with a buddha, but rather with Māra. His falling prostrate thus looks to be something of
a gaffe, requiring him to explain that his reverential act should be understood as directed not toward
Māra himself, but to the being whose form Māra had so skillfully assumed. If Buddhasaññaka’s
conception of the buddha is prompted by hearsay, Upagupta’s is prompted by something like “see-
say” – that is, he sees the form of a buddha, though no buddha is in fact before him. And whereas the
outcome of Buddhasaññaka’s conceiving of a buddha is unambiguously positive, the outcome (or,
at least, one outcome) of Upagupta’s conceiving of a buddha is a moment of awkwardness. Despite
these differences, however, both stories suggest that a conception of a buddha can be generated even
when no encounter with a buddha has occurred. Moreover, neither story occurs in a Mahāyāna source,
suggesting that the view of buddhasaṃjñā the stories imply may not be a Mahāyāna innovation.
31 An etic way of putting this point, using terms borrowed from Austin (and quite alien to Buddhist
tradition), is to say that buddhavacana can be understood to label a species of perlocution. See Austin
(1975 [1962], 101–4).
32 The relevant Tibetan may be found at D. dkon brtsegs ca 275b3–276a5.
33 The text’s Tibetan translation is catalogued among the texts of the Bka’ ’gyur (“Translation of the
word”) – a collection traditionally held to comprise the discourses of buddhas. Moreover, the most
well-known English translation of texts drawn from the Heap of Jewels corpus (Chang et al. 1983)
bears the title A Treasury of Mahāyāna Sūtras (and, as previously noted, sūtras are paradigm cases of
buddhavacana).
34 Alas, the scholar-practitioner is not out of the woods yet. She will need to think carefully about the
category of inspired speech – and more specifically, what that category excludes. The standing threat
here is that once we begin to think of buddhavacana as speech that prompts buddhasamjñā, the lat-
ter category will expand willy-nilly, and take the former category with it. Consider, for example, the
claim that “all buddhas are manifestations of Māra.” By traditional lights, that is a mistaken claim.
But it is also a claim about buddhas, and so could be understood as prompting an idea of the Buddha.
Why it should not be so understood is something that would take some showing. But the example
suggests that if we go wrong in seeing traditional invocations of buddhavacana as uniformly gestur-
ing toward provenance, we go wrong also in seeing them as uniformly gesturing toward a particular
content. The point here is not to shift the way we translate the compound buddhavacana from “speech
of a Buddha” to “speech about a Buddha.” What is needed for a buddhasaṃjñā to arise clearly must
be more than mere mention of Buddhas.
35 The reference here is to a famous pericope from the Diamond Cutting Transcendent Wisdom
(Vajracchedikā Prajñāpāramitā) describing the status of all conditioned phenomena.
Bibliography
Austin, John Langshaw. 1975 [1962]. How to Do Things with Words: The William James Lectures Delivered
at Harvard University in 1955, edited by J.O. Urmson and Marina Sbisà. Oxford: Oxford University Press.
43
Richard F. Nance
Beyer, Stephan. 1978. “Notes on the Vision Quest in Early Mahāyāna.” In Prajñāpāramitā and Related
Systems: Studies in Honor of Edward Conze, edited by Lewis Lancaster and Luis O. Gomez, 329–40.
Berkeley, CA: Berkeley Buddhist Studies Series.
Boucher, Daniel. 2008. Bodhisattvas of the Forest and the Formation of the Mahāyāna. Honolulu: Uni-
versity of Hawai’i Press.
———. 2018. “Recruitment and Retention in Early Bodhisattva Sodalities.” In Setting Out on the Great
Way: Essays on Early Mahāyāna Buddhism, edited by Paul Harrison, 95–118. Sheffield: Equinox
Publications.
Braarvig, Jens. 1985. “Dhāraṇī and Pratibhāna: Memory and Eloquence of the Bodhisattvas.” Journal of
the International Association of Buddhist Studies 8 (1): 17–29.
Brough, John. 1980. “Sakāya Niruttiyā: Cauld kale het.” In The Language of the Earliest Buddhist Tradi-
tion, edited by Heinz Bechert, 35–42. Göttingen: Vandenhoeck & Ruprecht.
Buescher, Hartmut. 2007. Sthiramati’s Triṃśikāvijñaptibhāṣya: critical editions of the Sanskrit text and
its Tibetan translation. Vienna: Austrian Academy of Sciences Press.
Burke, Seán. 1995. Authorship: From Plato to Postmodernism. Edinburgh: Edinburgh University Press.
Cabezón, José. 2001. “Authorship and Literary Production in Classical Buddhist Tibet.” In Changing
Minds: Contributions to the Study of Buddhism and Tibet in Honor of Jeffrey Hopkins, edited by Guy
Newland, 233–63. Ithaca, NY: Snow Lion Publications.
Cantwell, Cathy, Jowita Kramer, Robert Mayer, and Stefano Zachetti, eds. 2013/14 [2015]. “Authors and
Editors in the Literary Traditions of Asian Buddhism.” Journal of the International Association of
Buddhist Studies 36/37: 195–561.
Chang, Garma et al., trans. 1983. A Treasury of Mahāyāna Sūtras: Selections from the Mahāratnakūṭa
Sūtra. University Park, PA: Pennsylvania State University Press.
Chartier, Roger. 1992. The Order of Books. Stanford, CA: Stanford University Press.
Collins, Steven. 2013 [1990]. “On the Very Idea of the Pali Canon.” In Self and Society: Essays in Pali
Literature and Social Theory, 1988–2010, 3–30. Chiang Mai: Silkworm Books.
Davidson, Ronald. 1990. “An Introduction to the Standards of Scriptural Authenticity in Indian Bud-
dhism.” In Chinese Buddhist Apocrypha, edited by Robert E. Buswell, 291–325. Honolulu: Univer-
sity of Hawai’i Press.
Dutt, Nalinaksha, ed. 1966. Bodhisattvabhumi. Tibetan Sanskrit Works Series, vol. 7. Patna: K. P. Jayas-
wal Research Institute.
Engle, Artemus, trans. 2009. The Inner Science of Buddhist Practice: Vasubandhu’s Summary of the Five
Heaps with Commentary by Sthiramati. Ithaca, NY: Snow Lion Publications.
———. 2016. The Bodhisattva Path to Unsurpassed Enlightenment: A Complete Translation of the
Bodhisattvabhūmi. Ithaca, NY: Snow Lion Publications.
Foucault, Michel. 1977. “What Is an Author?” In Language, Counter-Memory, Practice: Selected Essays
and Interviews, 113–38. Ithaca, NY: Cornell University Press.
Gold, Jonathan. 2015. Paving the Great Way: Vasubandhu’s Unifying Buddhist Philosophy. New York:
Gombrich, Richard. 2009. What the Buddha Thought. London, UK: Equinox.
———. 2018. Buddhism and Pāli. Oxford: Mud Pie Books.
Griffin, Robert. 1999. “Anonymity and Authorship.” New Literary History 30 (4): 877–95.
Griffiths, Paul J. 1994. On being Buddha: The Classical Doctrine of Buddhahood. Albany, NY: State
University of New York Press.
Gyatso, Janet. 2017. Being Human in a Buddhist World: An Intellectual History of Medicine in Early
Modern Tibet. New York: Columbia University Press.
Harrison, Paul. 1990. The Samādhi of Direct Encounter with the Buddhas of the Present. Studia Philo-
logica Buddhica 5. Tokyo: International Institute for Buddhist Studies.
———. 2003. “Mediums and Messages: Reflections on the Production of Mahāyāna Sūtras.” Eastern
Buddhist 35 (1–2): 115–51.
Haynes, Christine. 2005. “Reassessing ‘Genius’ in Studies of Authorship: The State of the Discipline.”
Book History 8: 287–320.
Heim, Maria. 2018. Voice of the Buddha: Buddhaghosa on the Immeasurable Words. Oxford: Oxford
University Press.
Jaffee, Martin S. 2007. “Rabbinic Authorship as a Collective Enterprise.” In The Cambridge Companion
to the Talmud and Rabbinic Literature, edited by Charlotte E. Fonrobert and Martin S. Jaffee, 17–37.
Cambridge, UK: Cambridge University Press.
44
Siddhārtha Gautama
Kapstein, Matthew. 1989. “The Purificatory Gem and Its Cleansing: A Late Tibetan Polemical Discussion
of Apocryphal Texts.” History of Religions 28 (3): 217–44.
———. 2018. “Who Wrote the Trisvabhāvanirdeśa? Reflections on an Enigmatic Text and Its Place in
the History of Buddhist Philosophy.” Journal of Indian Philosophy 46 (1): 1–30.
Karpik, Stefan. 2019. “The Buddha Taught in Pali: A Working Hypothesis.” Journal of the Oxford Centre
for Buddhist Studies 16: 10–86.
Kramer, Jowita. 2012. “Descriptions of ‘Feeling’ (vedanā), ‘Ideation’ (saṃjñā), and ‘the Unconditioned’
(asaṃskṛta) in Vasubandhu’s Pañcaskandhaka and Sthiramati’s Pañcaskandhakavibhāṣā.” Rocznik
Orientalistyczny 65 (1): 120–39.
———. 2013. Sthiramati’s Pañcaskandhakavibhāṣā Part I: Critical Edition; Part II: Diplomatic edition.
Vienna: Austrian Academy of Sciences Press.
———. 2016. “Some Remarks on Sthiramati and His Putative Authorship of the Madhyāntavibhāgaṭīkā,
the *Sūtrālamkāravṛttibhāṣya and the Trimśikāvijnaptibhāṣya.” Buddhist Studies Review 33 (1–2):
47–63.
Kumagai, S. 2019. “Classifications of Perception (saṃjñā) in Buddhist Āgama and Ābhidharma Trea-
tises.” In Reasons and Lives in Buddhist Traditions: Studies in Honor of Matthew Kapstein, edited by
Dan Arnold, Cécile Ducher, and Pierre-Julien Harter, 257–64. Somerville, MA: Wisdom Publications.
Lamotte, Étienne. 1988 [1949]. “Assessment of Textual Interpretation in Buddhism.” Translated by Sara
Boin-Webb. In Buddhist Hermeneutics, edited by Donald S. Lopez, 11–27. Honolulu: University of
Hawai’i Press.
Levman, Bryan. 2008–2009. “Sakāya Niruttiyā revisited.” Bulletin d’Études Indiennes 26–27: 33–51.
———. 2016. “The Language of Early Buddhism.” Journal of South Asian Languages and Linguistics
3 (1): 1–41.
Li, Xuezhu, and Ernst Steinkellner, eds. 2008. Vasubandhu’s Pañcaskandhaka. Vienna: China Tibetology
Publishing House, Austrian Academy of Sciences Press.
Liu, Zhen. 2015. The Dharmadhātustava: A Critical Edition of the Sanskrit with the Tibetan and Chinese
Translations, a Diplomatic Transliteration of the Manuscript and Notes. Vienna: Austrian Academy
of Sciences.
Lopez, Donald S. 1988. “On the Interpretation of the Mahāyāna Sūtras.” In Buddhist Hermeneutics,
edited by Donald S. Lopez, 47–70. Honolulu: University of Hawai’i Press.
MacQueen, Graeme. 1981. “Inspired Speech in Early Mahāyāna I.” Religion 11: 303–19.
———. 1982. “Inspired Speech in Early Mahāyāna II.” Religion 12: 49–65.
Makransky, John. 1997. Buddhahood Embodied: Sources of Controversy in India and Tibet. Albany, NY:
State University of New York Press.
Nance, Richard. 2012. Speaking for Buddhas: Scriptural Commentary in Indian Buddhism. New York:
———. 2013. “The Voice of Another: Speech, Responsiveness, and Buddhist Philosophy.” In A Com-
panion to Buddhist Philosophy, edited by S. Emmanuel, 366–76. Hoboken, NJ: Wiley-Blackwell.
Pagel, Ulrich. 1995. The Bodhisattvapiṭaka (Buddhica Britannica, Series continua V). Tring: The Insti-
tute of Buddhist Studies.
Rosenbaum, Ron. 2006. The Shakespeare Wars: Clashing Scholars, Public Fiascoes, Palace Coups.
New York: Random House.
Sherburne, Richard. 2000. The Complete Works of Atīśa Śrī Dīpaṃkara Jñāna, Jo-bo-rje. New Delhi:
Aditya Prakashan.
Skilling, Peter. 2009. “Seeing the Preacher as the Teacher: A Note on Śāstṛsaṃjñā.” Annual Report of the
International Research Institute for Advanced Buddhology at Soka University XII: 73–100.
Snellgrove, David. 1958. “Note on the Adhyāśayasaṃcodanasūtra.” Bulletin of the School of Oriental
and African Studies, University of London 21 (1): 620–23.
Strong, John. 1992. The Legend and Cult of Upagupta: Sanskrit Buddhism in North India and Southeast
Asia. Princeton, NJ: Princeton University Press.
Tatz, Mark. 1994. The Skill in Means (Upāyakauśalya) Sūtra. New Delhi: Motilal Banarsidass.
Tzohar, Roy. 2018. A Yogācāra Buddhist Theory of Metaphor. Oxford: Oxford University Press.
Vaidya, Parashuram Lakshman. 1961. Mahāyāna-Sūtra-Saṃgraha. Darbhanga: The Mithila Institute of
Post-Graduate Studies in Sanskrit Learning.
Walters, Jonathan, trans. 2018. Legends of the Buddhist Saints (Apadānapāli). http://apadanatranslation.org/.
45
3
THE TANTRIC BUDDHA
Primordial Buddhas as Philosophical Authors
Vesna A. Wallace
A Buddha as an Author in Buddhist Tantric Sources

In Indian Buddhist tantras, the buddha, as a teacher of Buddhist tantric tenets, speaks on a variety
of topics to a variety of audiences in various realms for different purposes, establishing himself
as the ultimate reality appearing in multifaceted forms to lead others to the highest bliss of awak-
ening. Thus, establishing himself as a compassionate and omniscient authority, he teaches tan-
tras in his various manifestations of enjoyment bodies (saṃbhogakāya) and sometimes even in
his ultimate Vajradhara form. According to the Stainless Light Commentary (Vimalaprabhāṭīkā)
on the Wheel of Time Tantra (Kālacakratantra), the buddha teaches by means of limitless terms,
and the redactors wrote down such terms employing the terminology of scholastic treatises
(śāstras), vernacular terms, mantric terms, and single-syllabic terms. A tantric practitioner is
expected to understand all of these terms in both their provisional (neyārtha) and definitive
meanings (nītārtha) (Stainless Light Commentary, vol. 1, 1986, 31–34, 47).
While in some tantras, the buddha is primarily concerned with teaching tantric practices, both
mundane and those deemed indispensable for awakening; in other tantras, he equally delves into
the doctrinal and practical aspects of tantric systems. In the higher classes of tantras, such as the
yoga and unexcelled yoga tantras (niruttarayoga-tantra), his authoritative teachings seem to be
most focused on those aspects of tantric discourses that are systematically pointing out the non-
dual, blissful, and empty nature of ultimate reality present in all sentient beings and attainable
within a single lifetime only by means of a non-conceptual mind of bliss. Although many of his
expositions on emptiness and on the innate, luminous nature of sentient beings’ minds correspond
to those expounded in some Mahāyāna sūtras, it is this persistent focus on bliss and preservation of
the body in tantric teachings that differentiates the tantric buddha from the buddha as an authorita-
tive figure in the sūtras. Regarded as omniscient, he stands as the highest authority on all systems
of thought, Buddhist and non-Buddhist, and as a single, all-pervading mind of the world, he has
the entire world as his body. Thus, there is nothing that stands outside of him, including the deities
worshipped by various non-Buddhist groups. Everything that can be experienced by the senses
and is cognizable by the ordinary mind is inseparable from his provisional forms that manifest
to minds still veiled by mental obscurations. Thus, he is established in Buddhist tantras as a sole
authority. As the ultimate, perfectly enlightened author of tantras, he invariably manifests as male,
although, as we are told in the Wheel of Time Tantra, by nature he is neither male nor female.
46 DOI: 10.4324/9781351030908-6
The Tantric Buddha
To understand a buddha as an author of a tantra, it may be useful to turn to a common definition

of tantra given in Buddhist tantric literature as a “connected discourse” ( prabandha) that consists
of the cause, result, and method of achieving buddhahood. Fundamental to these three aspects of a
tantra is a buddha’s blissful gnosis, present in all sentient beings and expressing itself by means of
the provisional and definitive statements under the names of Kālacakra, Hevajra, Heruka, and so
on. Thus, in Indian Buddhist tantras, the buddha appears as a teacher in diverse peaceful and fierce
forms including Hevajra, Cakrasaṃvara, and Kālacakra, and is referred to by a variety of appel-
lations including the Omniscient One (sarvajña, sarvavid), the Exalted One (bhagavān), Vajrī,
Vajrasattva, and Vajradhara.1 He is presented as superior to all Hindu gods such as Brahmā, Viṣṇu,
and Śiva, all of whom take refuge in him and worship at his feet. In the Hevajratantra, he is called
Bhagavān because he is endowed with six excellent qualities in their entirety, which are mentioned
in the Supreme Primordial Buddha Tantra (Paramādibuddhatantra) as supremacy, completeness,
gnosis, splendor, glory, and beauty.2 In the Hevajratantra and in the Abridged Wheel of Time Tan-
tra (Laghukālacakratantra), he is given this name because he crushes the inner negative forces,
the so-called māras embedded in an ordinary human mind-body complex. He is also referred to as
“ultimate reality” (tattva) because he is true bliss. In the Hevajratantra, we read that
[T]he Buddha has neither existence nor is he of the nature of nonexistence as a

Dharma body; he has a form with face, arms, and legs in his enjoyment and emana-
tion bodies, and yet he is without form due to his supreme bliss ( paramasaukhya).
(Tripathi and Negi 2001, 147)
While in many tantras, the buddha appears as a single teacher, in others, such as the Secret
Assembly Tantra (Guhyasamājatantra), multiple buddhas make their individual declaration
with regard to specific doctrinal tenets, and yet at other times, an instruction is collectively given
by the entire assembly of the buddhas. In tantras in which the opening line, “Thus I have heard”
(which marks texts within the tradition as the authentic words of the buddha), is absent, the
Bhagavān as the teacher (śāstṛ) engages in a dialogue with either bodhisattvas or other buddhas,
making references to his own various enlightened manifestations in the third person. He does
so as if saying, “do not identify me with these forms”; and yet, these manifestations are often
spoken of in tantras as the heart, essence, or secret of the body, speech, and mind of all buddhas.
In Buddhist tantras, the buddha as ultimate reality ( paramārtha, tattva) is also known as a
gnosis ( jñāna) of the sublime bliss of all the buddhas, as the progenitor of the best of victors
( jina). Likewise, as stated in the Guhyasamājatantra, he is known as “the mind of awakening
(bodhicitta), which is without beginning and end, peaceful, beyond existence and nonexist-
ence, and omnipresent due to the equality (samatā) of phenomena and as nonduality of empti-
ness and compassion” (Guhyasamājatantra; Matsunaga 1978, ch. 2, 10).
As taught by the buddha Hevajra in the Hevajratantra:
He alone is a life ( prāṇa) of living beings, and he alone is the supreme indestructible
( paramākṣara). He pervades everything and dwells in the body of all beings.
The Buddha as Philosopher and Doxographer

in the Wheel of Time Tantra
According to the Wheel of Time tantric tradition, the original author of the extensive and now
lost Primordial Buddha Tantra (Ādibuddhatantra) was the historical buddha Śākyamuni, who
47
Vesna A. Wallace
prophesized that the author of its abridged version, the Abridged Wheel of Time Tantra, will be
the Bhagavān Mañjuśrī Yaśas, who will redact it for the Brahmanic sages residing in Kalāpa, the
capital of the mythical land of Śambhala. In the Stainless Light Commentary, the most authorita-
tive commentary on this tantra, which appeared in India in the early eleventh century at the same
time as the tantra itself, we are told that the author of the tantra and its subject matter are one and
the same. The author, who is nothing other than the ultimate truth, is revealing himself in the
conventional and at times symbolic terms that constitute a tantric text. Seated on the lion throne
in the Kālacakra maṇḍala, having been worshipped by the best of gods, he teaches the tantra
at the request of Sucandra, the king of Śambhala, who wants to know the arrangement, dimen-
sions, and constitution of the cosmos, the ways in which the cosmos is present in a human body,
and a tantric method that leads to enlightenment. Prompted by Sucandra’s inquiry, the buddha
expounds these topics in the five chapters of the tantra, dealing with the cosmos, the individual,
tantric initiation, meditational, visualization practice of mentally generating oneself as the cen-
tral deity and its multilayered maṇḍala, and the stage of completion, involving a tantric yoga by
means of which the practitioner actualizes his buddhahood in the form of the deity that one pre-
viously visualized and with whom one imaginatively identified himself. Throughout the tantra,
it becomes increasingly clear that the buddha, being both the author of the tantra and its subject
matter, speaks of himself as being indivisible from the cosmos, a person’s mind-body complex,
and the path and result of a tantric practice. All these are his multifarious appearances perceived
differently by different sentient beings, according to their karmic obscurations.
In the section called “The Great Exposition of Our Own and Others’ Philosophical Views,
Logic, and Analysis,” contained in the second chapter of the Abridged Wheel of Time Tantra,
the buddha Śākyamuni, in his manifestation as Kālacakra, expounds the views of different Bud-
dhist and non-Buddhist systems of thoughts in nineteen verses in order to refute them, making
this tantra the only unexcelled yoga tantra in which a doxography is expounded. He begins his
doxographical exposition by first listing some of the tenets of the Kālacakra tantric system,
namely: the identitylessness of persons and phenomena, maturation of karma, interdependent
origination, the four bodies of the buddha, conscious emptiness (ajaḍaśūnyatā), the eighteen
unique qualities of a buddha, the Four Noble Truths, and so on (Śrīkālacakratantrarāja, ch.
2, 161; Upadhyaya 1986, 255). His counterarguments to the claims made by the adherents of
non-Buddhist and Buddhist systems of thought – including Vaibhāṣikas, Sautrāntikas, and
Yogācārins – are consistent with those given in the treatises of Mādhyamikas. Thus, the bud-
dha appears to give primacy to Madhyamaka thought.
After introducing his own views, he proceeds to point out some of the tenets of other sys-
tems in order to later rebut them. He begins first with a refutation of the Brahmanical asser-
tion of the Vedas as self-arisen and uncreated (akṛtaka) like space and of the word “Veda” as
inherently true:
That Veda is not self-arisen. In the abode of the three worlds, the word Veda signi-
fies the referent (artha). By the word Veda, Brahmā with his four mouths disclosed
the referent long ago. But if a referent is not different from a word, then why does
the uttered word “fire” not burn the mouth? Hence, indeed, there is a teacher (i.e.,
Brahmā) with regard to the unknown topic and with regard to the past and future
meaning.
The Veda is not similar to space because it is created, uttered with the mouths
from different places. In terms of its use, it is local, recited by the mouths of the
twice-born (brāhmaṇas); it is present everywhere, and others read it. Since the low
48
The Tantric Buddha
castes such as śūdras and others read and write it, this Veda is present everywhere.
Therefore, that Veda is not a standard of knowledgeable and learned men
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 260–61, vv. 166–167).
Puṇḍarīka, the author of the Clear Light Commentary on the tantra, explicates these two verses
of the critique of the Brahmanical tradition’s claim that the Veda is self-arisen, eternal, and
inherent in its referent by first asserting that since the word “Veda” signifies a referent (artha)
and is produced by means of a verbal articulation, it cannot be logically established as self-
arisen. Furthermore, the word Veda cannot be logically asserted as eternal or as invariably
inherent and nondifferentiated from its referent, because if a word and its referent are one and
the same, then each time when the word “fire” is uttered, one’s mouth would burn from fire.
Thus, it has been established that the word and referent are two separate things. Likewise,
since god Brahmā has been established as the initial teacher of Vedic text, Veda is logically
established as created and not as self-arisen. Moreover, since it is local in usage, recited not
only by learned brāhmaṇas but also by lower castes, it is not a standard of a learned men
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 160–62).
The next set of rebuttals in the tantra pertains to both Śaivas’ view of God as the independ-
ent creator and Vaiṣṇavas’ position on Īśvara, who is said to incite others to action, though
he himself is not considered as the agent of action. The buddha refutes these tenets from the
viewpoint of interdependent origination in order to demonstrate that nothing in the world has
a single cause or autonomous existence: Īśvara cannot be a creator or an independent instiga-
tor of action if he depends on others to perform actions, and if he were the one who instigates
others to action, he would ultimately be the one to experience the results of karma and not
someone else. Therefore, apart from the karma of living beings, there can be no creator who
bestows upon a person the results of good or bad actions.
In order to explain that everything, including Vaiṣṇavas’ creator Īśvara, arises due to con-
tact with something else and is not from itself but due to the agglomeration of atomic particles,
the buddha states:
Here, if the atoms of the earth, water, fire, and wind in the sky originally do not
belong to the creator, then in the absence of material substance (dravya), and devoid
of the sense objects, the creator of all does not create anything. There is neither a per-
ceptual ( pratyakṣa) nor an inferential ( parokṣa) means of valid cognition ( pramāṇa)
for that creator who is devoid of the sense objects. Thus, everything comes into exist-
ence due to the agglomeration of atoms, O king; and the nature of karma is not due
to the creator’s will.
Water appears even from a moonstone and the reflection of an object in a mirror
due to contact. Due to acidity, there is saliva on the tongue. Due to one’s own voice
there is another voice (i.e., echo). A sprout is due to a pure seed. Due to a lodestone,
there is the rolling of an iron stick. None of these occurs by the will of the creator.
That efficacy of phenomena is not caused by anyone within the abode of the three
worlds.
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 262–63, vv. 169–170)
With those two verses, Īśvara is negated as an independent creator on the grounds that in the
absence of material substances, like atoms, Īśvara is unable to create anything since Vaiṣṇavas
themselves claim that atoms originally do not belong to Īśvara. This refutation is grounded
49
Vesna A. Wallace
in a Buddhist view that not a single phenomenon arises from itself or by the will of a creator
as a single cause, but instead arises in dependence on the sets of causes and conditions. The
efficacy of a cause, or the causality of the cause, is conceived only with regard to the effect and
the effect with regard to the cause. Since cause and effect arise in dependence on each other,
all causes – including Īśvara – are nondetermined.
In the following verses, continuing this line of argumentation from the standpoint of
interdependent origination, or personal and phenomenal identitylessness, the buddha further
refutes the notion of the omnipresent and permanent Self (ātman), asserting that if dependently
arisen phenomena arise after they cease, then they cannot have the permanent, unchanging,
autonomous Self because the ceased phenomenon no longer exists. Hence, one can speak only
of the collection of phenomena in cessation but not of someone who departs to liberation or is
in karmic bondage. Thus, he argues:
If the Self is omnipresent, why does it experience suffering due to separation from
relatives? And if that Self is permanent, why does it fall in love when hit by the arrow
of love? If that Self was active in the waking state, why does it become torpid at the
time of sleep? Thus indeed, that Self is neither omnipresent, nor the Lord, nor was it
formerly active.
There is no Self. There is an origination of interdependent phenomena, and there
are pleasant and unpleasant results, which are devoid of a single agent. There is no
one who departs, but there is a complete departure for liberation. There is bondage
but there is not a bound one.
There is existence, and there is non-existence of interdependently arisen phenom-
ena. The world is devoid of inherent existence and ultimately free from momentari-
ness. These true words of mine are not destroyed by the words of gods and nāgas,
which are with demons of conceptualizations.
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 262–65, vv. 171–172)
According to the Stainless Light Commentary on these verses, the buddha is arguing that if
the Self were omnipresent, then the suffering of all beings would be a suffering of one sentient
being. But if one asserts the existence of many Selves, then one cannot assert the omnipresence
of many Selves. The Self cannot be permanent because it is subject to change; therefore, the
Self is neither omnipresent, not permanent, nor some eternal Lord. The Self of phenomena in
cessation is non-existent because of the non-existence of that to which the Self would belong.
Thus, there is only a collection of phenomena that are in cessation and not someone who departs.
The practical relevance of this refutation of the permanent Self as taught in Hindu systems
of thought is a way in which the buddha expounds the personal identitylessness ( pudgala-
nairātmya) to his original Brahmanic audience and to those who will study and practice the
tantra in the future. He intentionally discusses the identitylessness in this second chapter of
tantra, which precedes the chapters dealing with a tantric practice. This is because without
understanding the identitylessness, or the absence of the permanent, autonomous Self, one can-
not effectively engage in the practice of self-generation as a deity (deity-yoga). After rebutting
the aforementioned views of various Hindu schools, he argues against the views of some Bud-
dhist schools, such as Vaibhāṣikas’ view of the reality of the person ( pudgala), certain tenets of
Sautrāntikas or Arthavādins, as well the Cittamātrins’ view of the reality of consciousness, and
so on. As shown in what follows, in his arguments against these Buddhist schools, he does not
take Nāgārjuṇa’s exposition of the four extremes (catuṣkoṭi) as a standard Madhyamaka exposi-
tion on freedom from any dogmatic position ( pakṣa), as one might expect. Instead, he focuses
50
The Tantric Buddha
on the assertion of the nonduality of emptiness and compassion. It is the author of the Stainless
Light Commentary who resorts here to Madhyamaka’s exposition of the four extremes in his
commentary on this verse.
One (Vaibhāṣika) who asserts the reality called pudgala to be present in the body has
failed due to implying the inherent nature of that pudgala. One (Sautrāntika) who
asserts the object by means of conventional truth has failed, because he considers the
unknown ultimate truth to be nonexistent. Likewise, one (Cittamātrin) who asserts
the reality of consciousness has failed, because he considers everything within the
three worlds to be consciousness. One who asserts the non-duality of emptiness and
compassion is free from dogmatic positions and has not failed.
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 265–66, v. 173)
The author of the Stainless Light Commentary explains why the one who asserts the nondual-
ity of emptiness and compassion is free from dogmatic positions, indirectly pointing to the
buddha himself and asserting that such a one is endowed with compassion without an object
and with emptiness that is free from conceptualization, as well as with the best of all aspects
(sarvākāra) (Śrīkālacakratantrarājasya; Upadhyaya, 1986, 267).
The next target of the buddha’s refutation is a view of the accumulation of the present
karma as a result of the previously performed karma, which he claims is upheld by Tājiks
(Tāyins), descendants of Iranian people and followers of a particular Islamic group whose
members believe that a deceased person experiences either suffering in hell or pleasure in
heaven through another human form due to pleasing Rahman, or Bismilla. It is worth mention-
ing here that this is the only tantra in which the author speaks explicitly about this particular
Islamic group and its theology. He refutes their tenets in this manner:
Tāyins hold that a person experiences his previously performed deeds in this life
through another birth. If it is so, then a removal of people’s karma from one birth to
another could not take place and there would be no escape from the cycle of exist-
ence, nor would there be the entrance into liberation over the course of immeasurable
lives. This, indeed, is the doctrine of Tāyins. But it is discarded by other lineages.
(Śrīkālacakratantrarājasya; Upadhyaya. 1986, 268, v. 174)
By means of the arguments given in the following verse, he refutes the doctrine of Materialists
(Cārvākas, Lokāyata), the followers of Amaraguru and Bṛhaspati who denied the efficacy of
karma, a god as the creator, and the notion of an afterlife, instead propagating the efficacy of
the agglomerated elements of the earth and so on in the production of living beings. This view
was rejected by Buddhists on the basis that it does not explain why inanimate objects, like
trees and the like, do not have consciousness.
If the witnessing mind (sākṣicitta), like the power of intoxicating drink produced
through the combination of different ingredients, arises due to the elements that are
integrated into one, then would not trees have consciousness due to the agglomera-
tion of the earth, water, fire, air, and space? But if those trees do not have the power of
living beings, then the power of the agglomeration of the elements is utterly useless.
That assertion of the Materialists does not bring agreeable results and has destroyed
the path of liberation for people.
(Śrīkālacakratantrarājasya; Upadhyaya, 1986, 268, v. 175)
51
Vesna A. Wallace
His last refutation relates to doctrinal tenets held by Jains, who are committed to the existence
of the eternal, unchanging soul ( jīva), whose size corresponds to the size of the body, and to
the permanence of atoms, and so on. The buddha argues against the Jains in this way:
If the soul has the body as its measure, does it perish after the arms and legs are cut
off? If an atom is permanent through the body, then why does it attain a gross state
after its subtle state? Jains hold that the one who is free from karma departs from
the cycle of existence to the state of bliss, which is located on the top of the world.
The counterargument is: The three worlds, which are made of atoms, never remain
permanent due to the efficacy of time.
The buddha concludes these discussed refutations of the tenets of different systems of thought
with the assertion that his refutation of the aforementioned tenets of different Hindu and Bud-
dhist systems of thought is given in accordance with the peoples’ dull, sharp, and other – mean-
ing low – mental faculties, which are due to the habitual propensities (vāsanā) of people’s
karma. His reason for refuting those tenets is to impart knowledge to those who hold them.
At the same time, he warns his audience not to criticize any Dharma, whether it is one’s own
or of another lineage on the grounds that the mind, like a crystal, becomes colored by colors
of the object in proximity. It is not clear whether this statement expresses his concern that by
studying the tenets of others in order to demolish them, one’s own beliefs may be influenced
by them, or his concern that the spirit of criticism (dūṣaṇa) may negatively influence one’s
mind. The Stainless Light Commentary is silent on this point, and does not give us clues.
Just when one expects the buddha to end his rebuttals, he continues with his warning
against the sacrificial, ritual violence authorized by the Vedas, as he advises the king Sucandra:
The Dharma free from mundane concerns is a service to sentient beings, and the
non-Dharma is harm to them. Violence, which has the Vedas as authority, does not
yield pleasant rewards and always gives suffering. In the words of a fool, a true love,
attached toward all sentient beings, produces supreme bliss. Therefore, for the sake of
sentient beings cultivate with your mind only the lack of inherent existence, O king.
Finally, in the conclusion of this doxographical exposition, the buddha affirms his universal
spiritual authority, stating:
I am Indra, the guru of thirteen men in heaven, a cakravartī on earth, the king of
nāgas in the underworld, revered by serpents. I am omnipresent and foremost. I am
gnosis, a Buddha, the Lord of Sages, the indestructible and supreme sovereign, the
yogīs’ indestructible union, the Veda, self-awareness, and the purifier ( pavitra). King,
take refuge in me with all your being.
The criticism of the Brahmanic interpretation of the Vedas and the sentence in which the
Buddha is identified with the Veda may perplex the reader of this tantra. It may be even more
perplexing in lieu of his statement given in another section of the Wheel of Time Tantra, where
he presents himself to the Brahmanic sages of Kalāpa, whom he seeks to convert to the Wheel
of Time Tantra, as the ultimate author of the four Vedas. The buddha allows himself to make
52
The Tantric Buddha
such claims on the basis of his self-description as a single, unitary, and all-pervasive ultimate
reality, as a source of all phenomena and all teachings, understood and interpreted by different
sentient beings in accordance with their own mental dispositions.
Tantric Buddhas’ Teaching on Emptiness

Just as in the Wheel of Time Tantra, in other tantras, including those already mentioned, all
buddhas teach the doctrine of emptiness (śūnyatā), or identitylessness, each in his own presen-
tational way in accordance with the contexts of the individual tantric systems. A reader familiar
with Mahāyāna canonical sources can easily notice that some of the most foundational doctri-
nal tenets taught by a buddha echo those taught in Mahāyāna Sūtras. It is for this reason that
many Indian commentators on Buddhist tantras saw Vajrayāna as inseparable from Mahāyāna,
calling it a “Profound (gambhīra) Mahāyāna,” “Mantra-Mahāyāna,” “Great Mahāyāna”
(mahāmahāyāna), and so on, differentiating it from the system of perfections ( pāramitānaya)
only in its method and in the amount of time it takes the practitioner to achieve awakening.3
In Buddhist tantras, where the system of perfections and the system of mantras (man-
tranaya) is synthesized, the buddha propounds the doctrine of emptiness (śūnyatā) in ways
similar to those expressed in the Perfection of Wisdom literature and in the works of Indian
Mādhyamikas. Emptiness is one of the most central doctrinal tenets espoused in all Bud-
dhist tantras. In the Vajra Crest Tantra (Vajroṣṇīṣatantra), the transcendence of wisdom
( prajñāpāramitā) that sees emptiness is defined as awakening (bodhi), foundation (ālaya),
liberation (mukti), and a fulfiller of all hopes (cited in Dwivedī and Lāla 1990, 62).
In the Hevajratantra, Hevajra Vajrī speaks of emptiness, or the absence of inherent exist-
ence, of one’s sense faculties and their corresponding sense objects as arising interdependently:
Neither form nor that which sees (eye) it; neither sound nor that which hears (ear) it,
neither aroma nor that which smells it, neither taste nor a taster, neither touch nor that
which touches, neither the mind nor a thinker exists by its own nature (svarūpeṇa).
In another section of the same tantra, he illustrates emptiness in terms of the interdependent
origination of all phenomena with these examples:
For instance, a fire arises in dependence on sticks, on rubbing, and on the effort of a
person’s hands. That fire dwells neither in sticks, in the rubbing, nor in the effort of
a person’s hands. Although being searched for everywhere, it is not found anywhere.
But that fire is neither a true thing nor is it false. Yoginīs should think of all phenom-
ena in that way.
(Hevajratantra 3.37; Tripathi and Negi 2001, 51)
In the same tantra, Hevajra declares that the purity of all things is thusness (tathatā), which,
due to being non-localized, is all-pervading.
In the Secret Assembly Tantra, the five buddhas – who represent here the imperishable real-
ity (tattvākṣara) of the body, speech, mind, secret, and Dharma of all buddhas – assert that the
mind of tathāgatas, known as the mind of awakening (bodhicitta), is the ultimate nature of all
phenomena; that is, emptiness, or identitylessness. A realization of each of the five buddhas
is expressed in their individual states of meditative concentration (samādhi) that correspond
to their particular realizations. Thus, among them, buddha Vairocanavajra, having entered the
53
Vesna A. Wallace
meditative concentration called “the vajra of the realization (abhisamaya) of all tathāgatas,”
describes the mind of awakening in this way:
One’s own mind, which has abandoned all phenomena and is devoid of the psycho-
physical aggregates (skandha) and sense bases (āyatana), or devoid of the apprehender
(grāhaka) and the apprehended (grāhya), is primordially non-arisen and is of the
nature of emptiness due to the equality (samatā) of the identitylessness of phenomena.
(Guhyasamājatantra 2.3; Matsunaga 1978, 10)
He thus asserts that the mind of all buddhas, which is empty by nature, when devoid of
all psychophysical components forming the ordinary mind-body complex, which are
the basis of dual perception, is all-pervasive, just like emptiness. Here also the buddha
Akṣobhyavajra abiding in the meditative concentration called “the vajra of the indestruct-
ibility of all tathāgatas,” teaches that phenomena are ultimately non-arisen; hence, neither
their characteristics nor their reality (dharmatā) can be found. On that basis, he likens iden-
titylessness to space. The buddha Ratnaketuvajra, abiding in the meditative concentration
called “the vajra of the identitylessness of all tathāgatas,” teaches that all phenomena are
without becoming, they are devoid of the characteristics of phenomena and are arisen from
the identitylessness of phenomena. The buddha Amitāyusvajra, while abiding in the medita-
tive concentration called “the vajra of the lamp of the flames of gnosis of all tathāgatas,”
declares that if phenomena are non-arisen, then there is neither becoming nor is there a
meditation (bhāvanā) on phenomena. Finally, the buddha Amoghasiddhivajra declares from
his meditative concentration called “the overpowering vajra of all tathāgatas” that phenom-
ena are luminous ( prabhāsvara) by nature, utterly pure, and similar to space, that from this
perspective, ultimately “there is neither awakening (bodhi) nor is there realization (abhisa-
maya)” (Guhyasamājatantra 2.4–7; Matsunaga 1978, 10).
As in the Eight Thousand-Line Perfection of Wisdom Sūtra, in the Wheel of Time Tantra,
the Vajrī maintains that ultimately there is neither a buddha, nor awakening, nor compassion
that benefits others (Śrīkālacakratantrarāja Vol. 2, 4.6; Dwivedī and Bahulkar 1994a, 153).
Moreover, like in the Eight Thousand-Line Perfection of Wisdom Sūtra, when speaking of
the emptiness of all phenomena, he compares it to space and teaches meditation on space
as a meditation on emptiness. We are told in the Secret Assembly Tantra that space is not
conjoined with any phenomena, nor is it not non-conjoined. All phenomena are dreamlike
appearances, as if arisen in a dreaming state. Space is formless, invisible, and inexplicable,
and this is how all phenomena should be understood. The elements of body, speech, and mind
and the space element are nondual. All phenomena abide in the space element, but the realm
of space does not abide anywhere – neither in the desire realm, nor the form realm, nor in
the formless realm. The realm of phenomena (dharmadhātu) does not abide within the triple
world and is without origination. That which is without origination is not produced by any
phenomenon. Therefore, all phenomena are devoid of inherent existence (niḥsvabhāva). The
mind of awakening, which gives rise to the state of a Vajrī or to the one who has attained the
vajra-like indestructible Perfect Awakening, which, in turns, brings forth the body, speech,
mind, and gnosis of all tathāgatas, does not dwell in the body, speech, or mind. Whatever
phenomenon does not dwell within the triple world, it does not have the origination. It does
not occur to a sleep: “I should manifest the dream within the triple world.” Nor does it occur
to the person: “I should see the sleeping state.” The activity of the triple world is similar to
a dream and is arisen from sleep. As many buddhas and bodhisattvas there are and as many
54
The Tantric Buddha
sentient beings there are in all world systems (lokadhātu) within the ten directions, they all
should be understood to be identityless and dreamlike.
The non-origination and non-abiding of all phenomena are further illustrated with an
example of the way in which one obtains gold or a gem, requesting it from the wish-fulfilling
jewel that grants all jewels. Those who ask for gold or a jewel from the wish-fulfilling jewel
are said to obtain it merely by mind, but the acquired jewel or gold does not dwell in the
person’s mind, nor is it present in the wish-fulfilling jewel. It is in this way that phenomena
should be understood (Guhyasamājatantra, ch. 15; Matsunaga 1978, 84). Similarly, all man-
tra-siddhis, which are ultimately non-arisen, are, nevertheless, present in one’s own body,
speech, and mind – but those body, speech, and mind are not located in the desire realm, in
the form realm, or in the formless realm. We are further instructed that the mind is not located
in the body, and the body is not located in the mind, and that speech is not located in the
mind, and the mind is not located in the speech, because, like space, they are pure by nature.
One’s own body, speech, and mind abide in space, and space does not abide anywhere. Fur-
thermore, in the same tantra, in the context of meditation on the recollection of non-origina-
tion of phenomena, the presiding lord (adhipati) of the vajras of the body, speech, and mind
of all tathāgatas proclaims that everything is luminous by nature, signless (nirmitta), neither
dual nor nondual, peaceful, similar to space, and entirely stainless (Guhyasamājatantra, ch.
7, vv. 34–35: Matsunaga 1978, 23).
A Buddha as a Teacher of the Nonduality of Sam

 sāra and Nirvāna
In the higher classes of tantras, the yoga and unexcelled yoga tantras, the ultimate nonduality
of saṃsāra and nirvāṇa is expressed in different contexts and in a variety of ways. The bud-
dhas’ teachings in these tantras on the nonduality of saṃsāra and nirvāṇa is actually a teaching
on the very nondual, awakened mind of a buddha. In the Hevajratantra, for instance,
The Vajrī asserts that apart from saṃsāra, there is not some other nirvāṇa. Saṃsāra
is the form, sound, and so on; saṃsāra is the aggregate of feeling and so on; saṃsāra
is the sense faculties, and saṃsāra is aversion and other mental afflictions.
These phenomena are actually nirvāṇa but have the form of saṃsāra due to delu-
sion. The non-deluded one, while transmigrating in saṃsāra, brings saṃsāra to ces-
sation through purification.
Nirvāṇa is the mind of awakening (bodhicitta), which has the ultimate and con-
ventional forms.
(Hevajratantra 9.2–3, Tripathi and Negi 2001, 104–5)
In this tantric system, purity (śuddhi) is of the nature of self-awareness (svasaṃvedya);

a self-aware bliss arises due to the pure nature of the sense objects. Although that innate
purity is released by means of various purifications, it is not released by means of some other
purity. On the basis that the sense objects and the sense faculties are pure, self-awareness is
defined as the highest bliss. In the understanding of the eleventh-century Indian Buddhist
scholastic and an abbot of Vikramaśīla monastery, Ratnākāraśānti, this statement implies
that innate purity is not liberated through the purification of mental obscurations (āvaraṇa),
which are accompanied by habitual propensities (vāsanā), because purification is attainable
only by the path meditated upon for a long time. However, everything is pure by nature
(Tripathi and Negi 2001, 105).
55
Vesna A. Wallace
In the Compendium of Tattvas King of Tantras (Tattvasaṃgrahatantrarāja), belonging to the

class of yoga tantras, the buddha, the tathagata, speaks of the way in which the innate luminos-
ity of the mind comes into manifestation through the systematic removal of mental defilements:
O son of the noble lineage, this mind is luminous by nature, like the moon disc. Just
as the moon disc is luminous by nature, so is gnosis. Just as the moon disc gradually
becomes full, so also the jewel of the mind, which is luminous by nature, becomes
full. Just as the moon disc is sequentially seen with incidental lunar digits due to the
departure of the rays of the sun disc, so also the jewel of the mind, which is pure by
nature, is seen full of the qualities of a Buddha due to the gradual departure of the
spots of stains of mental afflictions.
(cited in Dwivedī and Lāla 1990, 22)
In response to various questions posed by Great Bodhisattvas, in the Secret Assembly Tan-
tra, the assembly of tathāgatas declares that all phenomena are nondual (advaya) but char-
acterized, or perceived, with a dual nature (Guhyasamājatantra 18.126a; Matsunaga 1978,
122). We find a similar statement in the Saṃpuṭatantra, the explanatory tantra belonging to
the Hevajra cycle, where:
The vajrī said that saṃsāra is a mind smeared by unbearable (durvāra) stains of
attachment and the like.
The luminous ( prabhāsvara) mind, free from ideation (kalpanā), has the smear of
the stains of attachment and other mental afflictions removed. Due to the absence of
grasping there is neither the apprehending subject (grāhaka) nor is there the appre-
hended object (grāhya). He said, this is the most excellent nirvāṇa.
(Śrībuddhakapālatantrarāja cited in Śrībuddhakapālamahātantrarājaṭīkā
of Abhayākāragupta 2009, ed. Dorje, 14, vv. 2b – 3)
This twofold appearance of the mind is explained in the Vajra Ornament Tantra (Vajra
maṇḍālaṃkāratantra) in terms of its ultimate and phenomenal aspects:
The mind (cintā) that is free of defilements (anāśrava) and the like is stainless (nir-
mala), free from “mine” and “I.” It is empty, similar to space, and devoid of existence
and non-existence.
It is without beginning and end, peaceful, the master of all phenomena; it bears all
forms and rests in the two truths.
In terms of the ultimate truth, this supreme mind is known as the Dharma Body.
In terms of conventional truth, it is consciousness (vijñāna), the mind (citta) that has
all sense faculties at all times and everywhere.
It is called “mental awareness” (manojñāna), wisdom, ( prajñā), intellect (bud-
dhi), thought (mati), mindfulness (smṛti), and insight (vipaśyanā) because of the
purification of a wherever-arisen gnosis.
It is free from eternalism and nihilism and is devoid of the duality of the beginning
and end. There are two conceptions ( prakalpanā) of it in terms of the conventional
and ultimate truths.
Just as a source of fire is established in dependence on kindling, or just as sesame
oil is conceived to be in sesame seeds and sap in a sugar cane, fragrance in flowers,
56
The Tantric Buddha
and fresh butter in milk, so also the mahāmudrā, present in both the form and
formlessness, dwells everywhere, in all beings: stationary and mobile.
(Cited in Dwivedī and Lāla 1990, 58–59)
In the Vajra Crest Tantra, the buddha replies to Vajrapāṇi’s question as to how is the ultimate
reality ( paramārthasatya) is an abode (ālaya) of all phenomena by stating:
Lord of Secrets, all phenomena that are of the nature of psychophysical aggregates
and the sense-bases, and the inanimate and animate worlds are the support of the
ultimate reality due to their impermanence. The ultimate reality is without the Self
(anātmaka); it is unsurpassed and entirely empty; it neither increases nor dimin-
ishes; it is entirely pure, similar to space, and is of the stainless nature, imperish-
able, inexpressible, and devoid of the body, speech, and mind; it does not exist nor
does it non-exist; it does not go anywhere nor does it not go; it is not far, nor it is
near; it is neither empty nor not empty, nor it is middle in between the two. This
is the ultimate reality. Therefore, this supreme, profound, and difficult to compre-
hend non-origination and non-cessation of all phenomena has conventional real-
ity (saṃvṛtisatya) as its co-operating cause ( pratyaya), which itself is arisen from
causes and conditions (hetupratyaya).
(Cited in Lāla 2001, 65–66)
Nirvāṇa is defined in the Wheel of Time Tantra as not having the beginning, middle, and end,
as that which does not have the sense objects, the mind (citta), the material nature ( prakṛti), a
person, a bondage, liberation, the agent, the seed, the time of manifestation, and as that which
is not of the nature of suffering and happiness in the entire world. It is signless (nirnimitta), free
of causes, and without qualities (Śrīkālacakratantrarāja Vol. 3, 1994b, Dwivedī and Bahulkar,
151, v. 244)
Tantric Buddhas’ Teachings on Gnosis

In addition to emptiness, the nondual, self-aware gnosis (svasamvedya-jñāna) is a central doc-
trine taught by buddhas in the unexcelled yoga tantras, where the tantric paths to awakening
are constructed around this concept. This gnosis is the supreme, immutable, and nondual mind
of all the buddhas, connate to all sentient beings but not yet actualized or directly perceived by
them. It is from that gnosis, which is indivisible from wisdom and compassion, that all tantras
emerge in response to the requests posed by various, ten-stage bodhisattvas. In the Hevajrata-
tantra, the Vajrī speaks of gnosis in these words:
Gnosis is free from the notion of oneself and another due to self-awareness.
It is similar to space, free from stains, and is empty. It is of the nature of existence
and non-existence, and is supreme. It is merged with wisdom and compassion and
with passion (rāga) and the absence of passion (virāga).
(Hevajratantra 10.8b – 9, Tripathi and Negi 2001, 113–14)
In this tantric system, gnosis is understood to be similar to space because it perceives itself and
others as being of the same essence (samarasa), and it is merged with passion and the absence
57
Vesna A. Wallace
of passion because the nondual, innate gnosis is passion, or a sublime bliss (mahāsukha), and
it is also the absence of ordinary passion (virāga) (Tripathi and Negi 2001, 114).
The buddha also identifies gnosis as a self-knowledge (ātmavidyā) in the Wheel of Time
Tantra, which is imperishable (akṣara), and from which the entire three worlds arise due to the
experience of bliss and into which they dissolve at the time of death. That kind of gnosis, he
claims, is not associated with scriptures of heretical groups or with the Vedas, Aṅgas, Smṛtis,
Siddhāntas, and so forth. Owing to gnosis, one becomes a buddha, who is a unique teacher,
possessing supernatural powers. A collection of the elements in one’s body, which is due to
the consumption of the six tastes present in the food and drink, arises from space. Therefore,
when one’s gnosis merges into space by means of a yogic, meditative practice, it becomes like
space, eternal, indestructible, and of the same taste. The realized peace, or gnosis, is present
in the elements within the triple world and should be known to be within one’s own body
(Śrīkālacakratantrarāja, vol. 1, Upadhyaya, 1986, v. 3, 159, Ch. 2, 221, v. 96, 221; Dwivedi
and Bahulkar, 1994a, Vol. 2, Ch. 3, v. 86, 89).
In the Supreme Primordial Buddha Tantra, cited in the Stainless Light Commentary, the
Primordial Buddha (Ādibuddha) describes the buddha’s gnosis with its five aspects in his
discussion of meditation on the five psychophysical aggregates as ultimately being of the
nature of the five types of gnosis. The mirror-like gnosis, which is the purified aggregate of
form, knows that all phenomena in space are devoid of the nature of ideation (kalpanā) and
are seen like an image in a virgin’s prognostic mirror. The gnosis of equality, or the aggregate
of feeling, knows that after achieving union with all phenomena, there is only a single, inde-
structible phenomenon that abides; and arisen from the imperishable gnosis, that single phe-
nomenon is without cessation and without eternity. The discriminating gnosis, or the purified
aggregate of discernment, knows that in non-arisen phenomena, which are free from mental
formations (saṃskāra), there is neither awakening nor buddhahood, nor a sentient being,
nor living beings. The accomplishing gnosis, or the purified aggregate of mental formations,
knows the phenomena that transcend the reality of consciousness (vijñānadharmatā) and
that are purified in gnosis, untainted, and luminous by nature as the realm of phenomena
(dharmadhātu). These five types of the gnosis characterizing the mind of the buddha are
represented by five paradigmatic awakened figures, or buddhas: Akṣobhya, Ratnasaṃbhava,
Amitābha, Amoghasiddhi, and Vairocana, who also represent the five types of awakened
families. Thus, one could say that although the awakened author of the tantra is ultimately
a single author, he speaks from the perspective of his five, mutually interconnected types of
gnosis manifesting as the five types of buddhas.
Tantric Buddhas’ Teachings on the Naturally Luminous Mind

and Its Obscurations
Due to the all-pervasiveness of a buddha’s gnosis, the mind of sentient beings is ascertained
in various tantras as inseparable from the nondual enlightened awareness. In the Wheel of
Time Tantra, it is pointed out that just as space does not disappear from a jar when water is
poured into the jar, “so also the sky-vajrī, who pervades the universe and is devoid of sense
objects, is within the body” (Śrīkālacakratantrarājasya; Upadhyaya, 1986, Vol. 1, Ch. 2, 219,
v. 91c–d).
In the Hevajratantra, we are reminded that although the divine pervader (vyāpaka) dwells
in the body of all things (vastu), he is not born from the body. When asked by Vajragarbha
in the Hevajratantra why the body that is pure by nature and primordially devoid of inherent
existence becomes the body composed of the elements, the Vajrī does not offer a fresh answer
58
The Tantric Buddha
to the question. Instead, to some degree, he reiterates what he already taught elsewhere in this
tantra, pointing out that stillness (nistaraṅga), interpreted by the commentator Kāṇha to mean
“the lack of conceptualizations,” is its nature that dwells in every body (Hevajratantram; Tri-
pathi and Negi 2001, Part 1, Ch. 1, v. 12b, Ch. 10, vv. 35–36, 14, 123).
As shown in what follows, we are further told in the Hevajratantra that although the six
sense faculties, the five psychophysical aggregates, the six sense objects, and the five elements
are pure by nature, they are obscured by ignorance and mental afflictions and thereby manifest
as cyclic existence. The sense faculties remain unpurified for as long as the apprehending
subject (grāhaka) and the apprehended object (grāhya) arise. However, by whatever internal
or external sense object and the like one is bound, by that same sense object the bondage is
released. The one who has become purified from mental obscurations sees all sense objects as
pure by nature. For such a one there is no longer a scent, a sound, a form, and so on, because
to that one the world is made of buddhas (Hevajratantram; Tripathi and Negi 2001, Part 1, Ch.
9, vv. 4–5, 19–20, 106).
Kālacakra gives a similar teaching in the Wheel of Time Tantra, wherein the person, like a
silkworm, binds himself with conceptualizations (vikalpa) and liberates himself with his mind
when conceptualizations are no longer present. The mind has four different powers (śakti): the
power of desire; the power of activity (kriyā); and the power of cognition – which respectively
give rise to the perception of phenomena, the penetration into phenomena, and the understand-
ing of the meaning of phenomena – and the nondual power, which does not grasp or relinquish
anything. The person’s karma and the sense faculties are most intimately connected because
karma is present in the sense objects, referred to in this tantra as the qualities (guṇas) of one’s
own materiality ( prakṛti), which is the psychophysical aggregates, elements, and sense facul-
ties. In saṃsāra, suffering and happiness are contained in those material constituents and are
produced by karma, which can be of three types: gross, subtle, and peaceful. When the thought
“I am an agent” arises, this should be known as a “wrong karma”; when the thought “Someone
else or Paśupati is an agent” arises, it is called “karma.” The one who is separated from the con-
stituents of the aforementioned materiality realizes, “Neither I nor someone else is an agent”;
this is considered not to be karma. Therefore, neither the Self nor some creator gives or takes
away the happiness and suffering of beings. It is the previously performed action that brings
about its corresponding result. The mind of deluded persons alone is the agent of creation and
destruction; it gives and takes away suffering and happiness. The rebirth that one enters bound
by karmic fetters is due to one’s own mind. This world, then, is constructed by the karma of
sentient beings (Śrīkālacakratantrarājasya; Upadhyaya 1986, Vol. 1, 85–91, 118–21).
Tantric Buddhas’ Teachings on the Body, Passion, and Bliss

In the yoga and unexcelled yoga tantra classes of Indian Buddhist tantras, the generation of
bliss is deemed indispensable for the actualization of gnosis ( jñāna) and for the attainment of
the ultimate siddhi of awakening. These tantras repeatedly emphasize a soteriological signifi-
cance of the body. In the Mundane and Supramundane Vajra Tantra (Laukikalokottaravajra-
tantra), it is declared:
The siddhi of awakening is obtained by means of bliss; the siddhi is not obtained by
means of bodily mortifications because the buddhahood, arisen in meditative concen-
tration is due to all bliss.
Due to bodily mortification through severe fasting, thirsting, and the like a mental
distraction (vikṣepa) arises; and due to distraction, there is no siddhi. Inferior beings
59
Vesna A. Wallace
(hīnasattva) who practice mortification do not succeed. Therefore, they are declared
to be the ones who have a difficult task.
(Cited in the Lupta Bauddha Vacana Saṃgraha, Dwivedī and Lāla 1990, Vol. 1 55)
Similarly, in the Wheel of Time Tantra, one is advised to preserve the body for the sake of
attaining the siddhi because in the absence of the body, one cannot utilize the body in tantric
practices to generate sensual bliss and thus can attain neither a siddhi nor supreme bliss in this
life (Śrīkālacakratantrarāja; Upadhyaya 1986, Upadhyaya, Vol. 1, Ch. 2, v. 107, 228). It is for
this reason that only embodied human beings can practice tantra.
The idea that a saṃsāric mind, which is invariably embodied in either a subtle or a gross
material form, can achieve the supreme bliss only through the repeated generation of bodily
bliss by means of tantric yogic techniques is unique to tantras. In the Hevajratantra, the Vajrī
Hevajra makes a similar point by stating:
How could there be bliss (saukhya) in the absence of the body? One cannot speak of
bliss in the absence of the body. The world is permeated by bliss that is of the nature
of the pervader (vyāpaka) and the pervaded (vyāpya).
Just as a scent depends on a flower and is unobtainable in the absence of the
flower, in the same way, due to the absence of form and other psychophysical aggre-
gates, bliss is not obtained.
(Hevajratantram; Tripathi and Negi 2001,
Part 2, Ch. 2, vv. 35b–c and 36, 144)
On the basis that awakening can be achieved in this very life, similar statements spoken by the
buddha are scattered throughout many tantras. In the Hevajratantra, the Vajrī Hevajra also states:
There can be no siddhi of awakening by means of all the Vedas and Siddhāntas nor
by means of the multitude of rituals because there is still a rebirth in another world
due to purification.
Without that bliss there is no siddhi in this life or in the other world.
Part 1, Ch. 8, vv. 54–55a, 102)
According to the Ḍākinī Vajra Enclosure (Ḍākinīvajrapañjara), the world arises due to
passion and reaches its destruction due to discarding passion. It is due to the complete
knowledge of the vajra-passion that the mind becomes Vajrasattva (cited in the Lupta
Bauddha Vacana Saṃgraha, Dwivedī and Lāla 1990, Vol. 1, 22). In the Hevajratantra,
the Vajrī emphasizes that by passion, meaning, mundane passion, the world is bound, and
it is liberated by passion (of innate bliss), like a person who knows the nature of poison
and can destroy the poison from which all people die. The sublime, connate bliss is eternal
and supreme, and it manifests as the five types of gnosis. Ultimately, the connate bliss is
the entire world and its nature is nirvāṇa (Hevajratantram; Tripathi and Negi 2001, Part
2, Ch. 2, vv. 44, 46, 51, 59a, 148–50, 152).
According to the Wheel of Time Tantra, to achieve buddhahood, a meditation on emptiness
alone is insufficient. Just as grape does not come from a nimba tree, or nectar from poison,
or lotus from a brahma tree, in the same way, the buddhahood and connate bliss are mil-
lions of eons away for the person who, devoid of immutable bliss, meditates on emptiness
(Śrīkālacakratantrarājasya; Dwivedi and Bahulkar 1994a, Vol. 2, Ch. 5, 38, v. 72).
60
The Tantric Buddha
Tantric Buddhas on the Buddhahood of Sentient Beings

The view of all sentient beings as being ultimately buddhas, expounded in the unexcelled
yoga tantras, has already been either implicitly or explicitly pointed out in various statements
cited in this chapter. This identification is laid out in accordance with the individual models of
tantric yogic practices developed in different tantras. In the Discourse in One Hundred Thou-
sand Stanzas (Lakṣābhidhāna), and the Saṃvara Root Tantra, Vajrasattva is also referred to
as a sublime bliss, which dwells in the mind of all sentient beings (cited in the Stainless Light
Commentary on the Kālacakratantra,; Upadhyaya 1986, Vol. 1, Ch. 1, 33).
As indicated previously, in the Hevajratantra, Hevajra explicitly states that all sentient
beings are buddhas, but their minds are obscured by incidental defilements (āgantukāsrava).
The fact that these defilements are not essential to sentient beings and can be removed means
that sentient beings are essentially buddhas but ignorant of their true nature. From the Vajrī’s
perspective of sentient beings,
The denizens of hell, pretas, animals, gods, humans, and asuras, even the worms on
the excrement and such are always blissful by nature.
No Buddha is found elsewhere, in some place within the world systems (lokadhātu).
Since the mind (citta) alone is a perfect Buddha there is no Buddha elsewhere.
Part 2, Ch. 4, vv. 64–65a, 73–75, 188–90)
In the Wheel of Time Tantra, the buddha gives nearly the identical statement, saying:
Sentient beings are buddhas, and there is not some other great Buddha in this world
system. By serving them, the unlimited cyclic existence is cut off due to the absence
of conceptualizations.
(Śrīkālacakratantrarāja, vol. 3, Dwivedi and Bahulkar
1994b, 37, vv. 66a – b)
In different unexcelled yoga tantras, a buddha gives different explanatory models of the associ-
ation of the Four Bodies of the Buddha4 with living beings’ four bodily cakras and with the four
states of the mind. Thus, in the Hevajratantra, the four bodies of Hevajra – the fourth being the
body of sublime bliss (mahāsukhakāya) – are said to be below the navel, below the heart, in the
throat, and on the head, respectively. An emanation body, considered here to be a material body
(sthāvara), is in the womb cakra ( yoni-cakra) because this is where a sentient being arises and
is fashioned (nirmīyate). On the basis of the view that the Dharma is of the nature of the mind,
it is associated with the heart cakra. An enjoyment body, which is said to correspond to one’s
enjoyments of the six tastes, is placed in the throat; and the cakra of sublime bliss is on the head
(Hevajratantram; Tripathi and Negi 2001, Part 2, Ch. 4, vv. 51–55, 183–84).
In the Wheel of Time Tantra, the locations of the buddha Kālacakra’s four bodies in the
person’s bodily cakras differ from those given in the Hevajratantra. In this tantra, the fourth
body, here called the Innate (sahaja), or Pure (śuddha) body, is associated with three cakras –
the secret cakra, the crown cakra, and the navel cakra. While the locations of the Dharma body
and enjoyment body correspond to those taught in the Hevajratantra, an emanation body is
here associated with the navel cakra. Moreover, the four bodies of the buddha Kālacakra are
identified with the waking state, dreaming, deep sleep, and the state of bliss, respectively. In
the waking state, the mind experiences the world by means of the body and speech; in the
61
Vesna A. Wallace
dreaming state, due to the power of the habitual propensities of the sense objects, the mind
reflects on objects invisible to others; in the deep sleep state, it is devoid of the sense faculties;
and in the fourth state, it attains bliss (Śrīkālacakratantrarāja; Upadhyaya 1986, Vol. 1, Ch. 2,
v. 27, 170; and Dwivedi and Bahulkar, 1994a, Vol. 2, Ch. 4, v. 107, 202).
Conclusion
As author of a Buddhist tantra, a buddha thus takes a normative stance when teaching the
doctrinal tenets that express his ontological and epistemological views regarding his ultimate
nature and the nature of the world that is pervaded by him and ultimately indivisible from him.
He repeatedly directs the attention of his audience to the fact that he is their true nature, and he
does so by resorting to the principles and conventions of a tantric tradition. A tantra as a text,
whether oral or written, is understood by tradition as an authoritative discourse; it is a linguis-
tic expression of a buddha’s mind, given with the intent to bring others to understand his truths
and to convert them, as seen in the Wheel of Time Tantra, whereby he engages in the work of a
philosopher and a doxographer. Thus, the act of listening or reading a tantra is meant to be an
event of interacting with the buddha’s mind, through which one can begin to know awakened
mind and to transform one’s own mind.
Notes
1 Some of his other names mentioned in Buddhist tantras are: pledge being (samayasattva); sublime
vajra (mahāvajra); reality (tattva), the holder of the body, speech, and mind vajras of all tathāgatas;
sublime bliss (mahāsukha); imperishable bliss (akṣarasukha); sublime vajra (mahāvajra); All-Good
(samantabhadra); the Self of all beings (sarvātma); the innate body (sahajakāya); natural body
(svabhāvikakāya); gnostic body ( jñānakāya); pure body (śuddhakāya); vajra body (vajrakāya); gno-
sis-vajra ( jñānavajra); gnosis indivisible from wisdom and compassion; sublime prāṇa (mahāprāṇa);
the vajrayoga; Primordial Buddha (ādibuddha); and so on.
2 A complete list of the six qualities is given in the Supreme Primordial Buddha Tantra, cited in the
Sekoddeśaṭīkā 1941, 3. See Ratnākāraśānti’s Muktāvalī Pañjikā for a slightly different reading (2001, 57).
3 For different interpretations, see Wallace (2011, 95–111).
4 In Vajrayāna, the four bodies of the buddha are the four aspects of Awakened mind. The Gnosis body
( jñānakāya) is the blissful aspect of the gnosis of emptiness, and the Dharma body (Dharmakāya) is an
empty aspect of the awakened mind. These two bodies are accessible only to buddhas. The enjoyment
body (Sambhogakāya) manifests as luminous, divine form in which the awakened mind appears to
those whose mental obscurations are diminished, teaches them, and blesses them; and the emanation
body (nirmāṇakāya) is a form in which the awakened mind appears to ordinary sentient beings due
to their mental obscurations. It is believed that a single buddha can simultaneously manifest in innu-
merable emanation bodies in countless world systems. Thus, buddhas manifest differently to different
beings in accordance with their mental obscurations and propensities.
Bibliography
Abhayākāragupta. 2009. Śrībuddhakapālamahātantrarājaṭīkā. Edited by Chog Dorje. Bibliotheca Indo-
Tibetica Series, vol. 68. Sarnath, Varanasi: Central Institute of Higher Tibetan Studies.
Bagchi, Sitansusekhar, ed. 1965. Guhyasamāja Tantra or Tathāgataguhyaka. Darbhanga: The Mithila Institute.
Dwivedī, Vrajavallabha, and Banārasī Lāla, eds. 1990. Lupta Bauddha Vacana Saṃgraha, vol. 1. Rare
Buddhist Texts Series, vol. 6. Sarnath, Varanasi: Central Institute of Higher Tibetan Studies.
Dwivedī, Vrajavallabha, and S. S. Bahulkar, eds. 1994a. Śrīkālacakratantrarājasya Kalkinā
Śrīpuṇḍarikeṇa Viracitā Ṭīkā Vimalaprabhā, vol. 2. Rare Buddhist Texts Series, vol. 12. Sarnath,
Varanasi: Central Institute of Higher Tibetan Studies.
———. 1994b. Śrīkālacakratantrarājasya Kalkinā Śrīpuṇḍarikeṇa Viracitā Ṭīkā Vimalaprabhā, vol. 3.
Rare Buddhist Texts Series, vol. 13. Sarnath, Varanasi: Central Institute of Higher Tibetan Studies.
62
The Tantric Buddha
Lāla, Banārasī, ed. 2001. Luptabauddhavacanasaṃgraha, vol. 2. Rare Buddhist Text Series, vol. 25.
Sarnath, Varanasi: Central Institute of Higher Tibetan Studies.
Matsunaga, Yukei, ed. 1978. The Guhyasamāja Tantra. Osaka: Toho Shuppen, In Indo-Tibetica Series,
vol. 48. Sarnath, Varanasi: Central Institute of Higher Tibetan Studies.
Newman, John R. 1987. “The Outer Wheel of Time: Vajrayāna Buddhist Cosmology in the Kālacakra
Tantra.” PhD diss., University of Wisconsin.
———. 1992. “Buddhist Siddhānta in the Kālacakra Tantra.” Wiener Zeitschrift für die Künde Südasiens
36: 226–34.
Sekoddeśaṭīkā of Naḍapāda (Nāropā): Being a Commentary of the Sekoddeaśa Section of the Kālacakra
Tantra. 1941. Sanskrit text edited by Mario E. Carelli. Gaekwad’s Oriental Series. Baroda: Oriental
Institute.
Shashni, Thinlay Ram, ed. 1997. Bauddha Tantra Kośa, vol. 2. Rare Buddhist Text Series, vol. 15. Sar-
nath, Varanasi: Central Institute of Higher Tibetan Studies.
Snellgrove, D. L. 1959. The Hevajra Tantra: A Critical Study. Parts 1–2. London Oriental Series, vol. 6.
London: Oxford University Press.
Tripathi, Ram Shankar, and Thakus Sain Negi, eds. 2001. Hevajratantram Mahāpaṇḍitācāryaratnākā
raśāntiviracita Hevajrapañjikā-Muktāvalī Saṃvalitam. Bibliotheca Indo-Buddhica Series. Sarnath,
Varanasi: Central Institute of Higher Tibetan Studies.
Upadhyaya, Jagannatha, ed. 1986. Śrīkālacakratantrarājasya Kalkinā Śrīpuṇḍarikeṇa Viracitā Ṭīkā
Vimalaprabhā, vol. 1. Bibliotheca Indo-Tibetica Series, vol. 11. Sarnath, Varanasi: Central Institute
of Higher Tibetan Studies.
Wallace, Vesna A. 2001. The Inner Kālacakratantra: A Buddhist Tantric View of the Individual. New
York: Oxford University Press.
———. 2004. The Kālacakratantra: The Chapter on the Individual together with the Vimalaprabhā.
Tanjur Translation Initiative, Treasury of Buddhist Sciences Series. New York: American Institute of
Buddhist Studies, co-published with Columbia University’s Center for Buddhist Studies and Tibet
House (Annotated translation from Sanskrit, Tibetan, and Mongolian, together with a critical edition
of the Mongolian text).
———. 2009a. “The Body as a Text and the Text as the Body: A View from the Kālacakratantra’s Per-
spective.” In As Long as Space Endures: Essays on the Kālacakra Tantra in Honor of H. H. The Dalai
Lama, edited by Edward A. Arnold, 179–92. Ithaca, NY: Snow Lion Publications.
———. 2009b. “Medicine and Astrology in the Healing Arts of the Kālacakratantra.” In As Long as
Space Endures: Essays on the Kālacakra Tantra in Honor of H. H. The Dalai Lama, edited by Edward
A. Arnold, 277–300. Ithaca, NY: Snow Lion Publications.
———. 2010. The Kālacakratantra: The Chapter on Sādhana together with the Vimalaprabhā. Tanjur
Translation Initiative, Treasury of Buddhist Sciences Series. New York: American Institute of Bud-
dhist Studies at Columbia University, co-published with Columbia University’s Center for Buddhist
Studies and Tibet House.
———. 2011. “A Brief Exploration of Late Indian Buddhist Exegeses of the ‘Mantrayāna and Mantra-
naya’.” Pacific World: Journal of the Institute of Buddhist Studies, Third Series, No. 13, Special Sec-
tion: Recent Research on Esoteric Buddhism, Fall: 95–111.
———. 2013. “Practical Applications of the Perfection of Wisdom Sūtra and Madhyamaka in the
Kālacakra Tantric Tradition.” In Companion to Buddhist Philosophy, edited by Steven Emmanuel,
164–79. Oxford: Blackwell Publishing.
———. 2017. “Linguistic Arguments and Exegesis in Indian Tantric Buddhism: Intention and Inter-
pretation.” In Buddhism and Linguistics: Theory and Philosophy, edited by Manel Herat, 101–22.
Basingstoke, UK: Palgrave MacMillan.
63
4
MAITREYA
The Future Buddha as an Author
Klaus-Dieter Mathes
In most Buddhist traditions, Maitreya is a bodhisattva who is prophesized to be the fifth, and
coming, Buddha of our present auspicious eon (bhadrakalpa).1 He is held to currently reside
as the future Buddha in Tuṣita heaven, from whence, like his predecessors, he will descend
into the world after the teachings of the previous Buddha, in this case the Buddha Śākyamuni,
are no longer practiced or known. In the Kashmir of the first centuries of our common era,
Maitreya’s cult was at the center of the shift from mainstream to Mahāyāna Buddhism, when
Yogācāra (Yoga Conduct) masters such as Saṅgharakṣa (second century CE) and Asaṅga
(c. 315–390 CE) were believed to have sought and received inspiration from Maitreya (Demiéville
1954, 341). Followers of the Maitreya cult aspired to be reborn in Maitreya’s time and place.
Maitreya, whose name means the “Embodiment of Love,” also came to be considered,
together with Asaṅga and Vasubandhu, as the founder of Yogācāra, one of the two main branches
of the Great Vehicle (Mahāyāna). His status as an author, however, is a bit controversial given
his celestial nature. His works are generally said to have been first taught to Asaṅga, who
then promulgated them in the world. The accounts of how Asaṅga received the teachings from
Maitreya vary. According to Paramārtha’s biography of Vasubandhu (c. fourth or fifth century
CE),2 Maitreya came down to earth from Tuṣita to teach Asaṅga the Stages of Yogic Practice
(Yogācārabhūmi). Most sources specify, however, that Asaṅga was instructed by Maitreya in
Tuṣita heaven (Demiéville 1954, 381–82). Since Tuṣita is considered to be part of the desire
realm, it is within reach of human beings, even those who are still below the bodhisattva levels.
During the transition from mainstream to Mahāyāna Buddhism, it was already consid-
ered possible for practitioners to become bodhisattvas and eventually full-fledged buddhas,
but the Sarvāstivāda idea that there can be only one buddha at a time still remained wide-
spread. Although the Mahāsaṅghikas and Sautrāntikas admitted the simultaneous existence of
multiple buddhas, they nevertheless held that such buddhas must exist in different universes
(Demiéville 1954, 371). Thus, bodhisattvas on the verge of buddhahood had to wait in Tuṣita
(the realm of desire gods), engaging in celestial pleasure, study, and meditation, for their turn
to descend on earth as the buddha of their respective era. In the meantime, however, travels to
and from Tuṣita were imagined as occurring frequently, as, for example, when Śākyamuni, the
buddha of our era, traveled there to give Abhidharma (metaphysics) teachings to his deceased
mother, who had been reborn as a deity there. In as much as it was part of this world, human
beings could also reach it without taking rebirth there, and in China it had been argued that
64 DOI: 10.4324/9781351030908-7
Maitreya
Tuṣita is easier to reach then the better-known “Blissful Paradise” (Sukhāvatī), for the latter is
a pure realm accessible only for those on the bodhisattva levels.3
The question of Maitreya’s role as an author is thus vexed even for traditional scholars,
since he is seen as the source of the teachings but not as their promulgator in the human realm.
In his commentary on An Analysis of the Middle and Extremes (Madhyāntavibhāga),4 one of
the texts Maitreya is said to have revealed to Asaṅga, Vasubandhu praises the author of the
root text as “born from the Buddha himself ” (sugatātmaja),5 indicating an exalted status for
the author, while Sthiramati (sixth century CE) in his subcommentary specifies Maitreya as the
author of the root text.6 Sthiramati describes Maitreya as follows:
Being only one birth away from buddhahood, he has reached the highest perfec-
tion in terms of all bodhisattvas’ supernatural knowledge, memory, analytical knowl-
edge, meditative absorption, power, patience, and liberation, his hindrances on all
bodhisattva levels having been completely left behind.7
As Demiéville (1954, 386) notes, Asaṅga is said to have been on the path of preparation,
an initial stage of the Buddhist path, when he first received teachings from the bodhisattva
Maitreya. Having received the transmission of An Analysis of the Middle and Extremes, he
then proceeded to the first bodhisattva level.
The Five Maitreya Works

Tibetan tradition maintains that Asaṅga traveled to Tuṣita, where he received from the future
Buddha a set of texts that came to be known as the Five Maitreya Works (Mathes 1996,
11–17). In partial support of this legend, namely that Asaṅga is not the origin of these works,
Frauwallner (1994, 296) observes that the doctrines expounded in the Maitreya Works show
a clearly pronounced originality and distinguish themselves noticeably from Asaṅga’s own
works. The Five Maitreya Works include a summary and analysis of the Discourses on the Per-
fection of Insight (Prajñāpāramitāsūtras), an analysis of “buddha nature” (tathāgatagarbha),
and teachings on the Great Vehicle in general. Back in the human world, Asaṅga is said to
have written down these teachings. On this account, Maitreya and Asaṅga, or rather Maitreya
through Asaṅga, can be credited with the following five works, given in the order in which
they are traditionally said to have been taught to Asaṅga. As we shall see, these five texts may
indeed all be by a single hand:
The Ornament of Clear Realization (Abhisamayālaṃkāra)

The Ornament of Great Vehicle Discourses (Mahāyānasūtrālaṃkāra)
An Analysis of the Precious Potential (Ratnagotravibhāga)
An Analysis of the Middle and Extreme (Madhyāntavibhāga)
An Analysis of Phenomena and their True Nature (Dharmadharmatāvibhāga)
The first text, The Ornament of Clear Realization,8 summarizes the Discourses on the Per-
fection of Insight in a way compatible with Yogācāra and Tathāgatagarbha9 thought. This is
followed by The Ornament of Great Vehicle Discourses,10 which groups mainly Yogācāra
topics into twenty-one chapters. The remaining three texts are so-called analyses (vibhāgas);
namely, An Analysis of the Precious Potential,11 An Analysis of the Middle and Extreme, and
An Analysis of Phenomena and their True Nature.12 The first analysis is a standard Indian trea-
tise on buddha nature and buddhahood, the second defines a “Yogācāra middle way” based on
65
Klaus-Dieter Mathes
the Discourses on the Perfection of Insight, and the last is a Yogācāra work that focuses on the
fundamental transformation of the ordinary states of mind, and the phenomena (dharmas) they
display, into the true nature of these phenomena (dharmatā), which amounts to buddhahood.
Some Tibetan traditions place An Analysis of the Precious Potential at the end of the list.13 To
present the two ornaments and three analyses as Maitreya’s revelation invests the Yogācāra
summary and interpretation of Great Vehicle Discourses an extremely high level of authority.
This must be seen in the context of newly forming Yogācāra school’s competition with the
rival Madhyamaka (Centrist Way) interpretation of the Great Vehicle.14
The Chinese tradition knows neither The Ornament of Clear Realization nor An Analysis of
Phenomena and their True Nature (Seyfort Ruegg 1969, 43), and attributes An Analysis of the
Precious Potential to Sāramati (Frauwallner 1994, 255). Instead of these three texts, the Chinese
corpus of the Five Maitreya Works includes the Stages of Yogic Practice (the same one attributed
to Asaṅga via Maitreya and mentioned earlier); An Analysis of Yoga (Yogavibhāga), which is
now lost; and A Commentary on the Diamond Cutter (Vajracchedikāvyākhyā) (Nakamura 1989,
256). The Stages of Yogic Practice differs in general tenor from the Five Maitreya Works of the
Tibetan tradition, as it contains numerous mainstream Buddhism strands and is dominated by
detailed Abhidharma-based descriptions of psycho-physical aggregates and mental factors of
existence. The Tibetan Maitreya Works are more centered on a positively described ultimate
“sphere of qualities” (dharmadhātu), natural luminosity, or buddha nature (Frauwallner 1994,
33; Mathes 1996, 14–16). They thus seem more likely to be by a single author than does the list
of five works as given in the Chinese tradition. The remainder of this chapter provisionally con-
siders “Maitreya” as the author of the Five Maitreya Works according to the Tibetan tradition.
Our work will be to see whether these five texts might reasonably be attributed to a single hand.
The Ornament of Clear Realization

In his commentary on the first among the Five Maitreya Works, The Ornament of Clear Reali-
zation, the eighth-century Indian author Haribhadra writes that while Asaṅga had understood
the meaning of the Buddha’s words and had gained realization, he still was not able to deter-
mine the meaning of the Discourses on the Perfection of Insight due to their profundity and
numerous repetitions. When he saw that this depressed Asaṅga, Maitreya expounded on these
Discourses on the Perfection of Insight, composing the verses of The Ornament of Clear Real-
ization so that Asaṅga might more easily understand the essence of the scriptures (Brunnhölzl
2010, 47). In his initial description of the foundation of the path and practice, Maitreya makes
it clear that an all-pervading, indivisible sphere of qualities does not allow for a distinction
into definite, unchangeable potentials or dispositions (gotra), such as to restrict a particular
practitioner to developing on one particular vehicle ( yāna) only. In The Ornament of Clear
Realization, we thus find:
Given the indivisibility of the sphere of qualities,

A distinction between potentials is not acceptable.15
Needless to say, this excludes the possibility of an absolute cut-off potential (agotra, lit. “those
without a family”). Such a harsh category, which is occasionally found in Yogācāra, contra-
dicts the central Great Vehicle aspiration to lead all sentient beings to the ultimate level of
buddhahood. The softening of any rigid typology of potentials in evidence here reflects a
thread running throughout the Five Maitreya Works in which we find a heavy emphasis on the
inherent capacity of all beings to attain buddhahood. Maitreya’s teaching thus is best described
66
Maitreya
as a synthesis of the Yogācāra and Tathāgatagarbha models of reality, a synthesis that seeks to
avoid the potential flaws of both systems.
Even though several passages in The Ornament of Clear Realization can be interpreted
along the lines of Madhyamaka16 – and numerous Tibetan masters such as Rin chen bzang po
(958–1055) have taken the text as a commentary on the second Dharma Wheel in keeping with
the Madhyamaka School of Autonomous Reasoning (Svātantrika-Madhyamaka, Seyfort Ruegg
1969, 65) – the text’s proximity to Yogācāra concepts cannot be denied (Makransky 1997). Emp-
tiness is thus thought of in terms of the absence of a perceived object and perceiving subject rather
than in terms of the absence of an own nature (svabhāva). This becomes clear in the following
two verses in The Ornament of Clear Realization, which distinguish how ordinary beings and
the noble ones apprehend sentient beings as substantially and nominally existent, respectively:
The two concepts about substantially and nominally existent sentient beings
Are the two perceiving subjects.
Divided into ordinary beings and noble ones,
Each of them is ninefold in terms of their perceived objects.
If the two types of perceived objects do not exist the way they appear,
Of what are these two types of subjects then perceivers?
The emptiness of being a perceiving subject
Is thus the defining characteristic of both of them.17
In the same vein, Brunnhölzl (2010, 84–85) observes that the relinquishment of the perceived
and perceiver, which results in a wisdom beyond these two, is as classical a Yogācāra paradigm
as can be imagined. Moreover, the list of twenty-two examples of enlightened mind (bodhicitta)
in The Ornament of Clear Realization (I.19–20) is not found in the Discourses on the Perfection
of Insight but is almost identical with a list found in The Ornament of Great Vehicle Discourses
(IV.15–20). Likewise, the description of potential in The Ornament of Clear Realization (I.38–
39) is not sheer emptiness, but rather the sphere of qualities serving as a basis for various accom-
plishments on the path. The Yogācāra and Tathāgatagarbha leanings of this text are thus evident.
When it comes to the bodies of the buddha, we find the presentation of the “body of quali-
ties” (dharmakāya) in terms of the “embodiment of true nature” (svābhāvikakāya), “the body of
enjoyment” (sāmbhogikakāya), and the “emanation body” (nairmāṇikakāya). This presentation
matches similar passages in The Ornament of Great Vehicle Discourses and An Analysis of the
Precious Potential. According to The Ornament of Clear Realization, the embodiment of true
nature, which is itself the true nature of the body of qualities, thus contains all buddha qualities:
The Sage’s embodiment of true nature

Has the natural defining characteristic of all uncontaminated qualities,
Which have attained to purity
In every respect.18
In other words, the embodiment of true nature is not conceptualized as the mere emptiness of
the body of qualities, as we find in the four-body system of Haribhadra. In that four-body sys-
tem, the embodiment of true nature is an independent body and not simply another way to talk
about the body of qualities (Mathes 2008, 168; Makransky 1997, 289ff.). But here, the presen-
tation of the buddha bodies recalls the Tathāgatagarbha theory of An Analysis of the Precious
Potential, in which the buddha qualities are inseparably connected with a buddha element that
is equated with the body of qualities and described in terms of the embodiment of true nature
67
Klaus-Dieter Mathes
(Mathes 2008, 8–9). It should also be noted that in The Ornament of Clear Realization, bud-
dhahood is described as all-pervading and permanent, as well:
Given its vast activity,

Buddhahood is termed “all-pervading”;
And being exempt from decay,
It is called permanent.19
This brings to mind the initial verse of praise of An Analysis of the Precious Potential, where
buddhahood is described as being discovered, in the same sense as the historical Buddha with
the given name Siddhārtha opened up to, or found, buddhahood:
Homage to you who opened up to buddhahood,

which is without beginning, middle, or end, and is peaceful,
And who, after awakening, taught the fearless, eternal path
for the sake of the awakening of those who are not yet awake;
To you, who holds the supreme sword and thunderbolt of wisdom and love,
who cuts the tumor of suffering into pieces,
And destroys the rampart of doubts
enclosed by a thicket of various views.20
According to a subsequent commentary verse, being “without beginning, middle, or end”

means that buddhahood is unconditioned.21
The Ornament of Great Vehicle Discourses

The next work attributed to Maitreya is The Ornament of Great Vehicle Discourses. Even
though accepted as a Maitreya work by both the Tibetan and Chinese traditions, the colophon
of the only Sanskrit manuscript mentions the unidentified great bodhisattva Vyavadātasamaya
(Lévi 1911, 7), adding further mystery to the question of authorship. The Ornament of Great
Vehicle Discourses treats various traditional Yogācāra topics and notably includes a theory
that accepts the notion of a cut-off potential. Within his description of different types of
cut-off potentials, Maitreya distinguishes a group that has no “cause” or chance of liberation
from four groups of sentient beings, which are only temporarily cut off.
Some are solely bent on bad conduct.

Then there are those whose positive qualities have been destroyed,
Or those who lack the virtue conducive to liberation.
And some have few positive qualities.
But there are also those for whom there is no cause at all.22
Vasubandhu explains:
What is meant here regarding those who are without the capacity to attain perfect
nirvāṇa is the cut-off potential. In short, there are two types, those who are cut off
temporarily and those who are cut off completely. Of those who are cut off tempo-
rarily, there are four types: those who are solely bent on bad conduct, those whose
roots of virtue are cut off, those who are without the roots conducive to liberation,
68
Maitreya
and those who have few roots of virtue. They are those whose accumulations are
incomplete. Those who are entirely (atyanta-) without the capacity to attain perfect
nirvāṇa – without any cause, so to speak – lack any potential to attain perfect nirvāṇa
at all (eva).23
An absolute (atyanta-) cut-off potential, however, would contradict the statement in The Orna-
ment of Great Vehicle Discourses that all sentient beings have buddha nature:
Even though suchness is undifferentiated in all sentient beings,

In its purified form it is the state of the Tathāgata.
Therefore all living beings
Have the seed/nature (garbha) of him (i.e., the Tathāgata).24
In other words, we have here a strict potential-based system with the explicitly permanent
exclusion of one group of sentient beings from liberation over against an element from An Anal-
ysis of the Precious Potential – buddha nature being defined there as well as in The Ornament
of Great Vehicle Discourses as “suchness accompanied by stains” (samalā tathatā). However,
if that which constitutes a specific potential – i.e., the dependent nature ( paratantrasvabhāva)
within the Yogācāra three-natures model of reality – does not exist ultimately, the ultimate
goodness of an all-pervading sphere of qualities could still inhere as the ultimate nature of all
sentient beings, even those with a cut-off potential (D’Amato 2003, 126ff.).
The gotra issue thus is related to the problem the traditional three-nature model has when
the dependent nature is taken to have substantial, ultimate existence. The problem with this
model is that it posits that something dependently arising really exists, a position which has
been rightfully refuted in Madhyamaka philosophy. The three-nature theory lies at the center of
the Yogācāra model of reality (even though Maitreya himself does not make much use of it25)
and reflects the old Abhidharma distinction of relative and ultimate truths/realities in terms of
nominal ( prajñaptisat) and substantial (dravyasat) existence. In a modified form – there being
no longer material, but only “mental” substance in fully developed Yogācāra – a dependently
arising stream of false imagining (abhūtaparikalpa), i.e., the substantially existing depend-
ent nature, projects the imagined nature ( parikalpitasvabhāva) of a perceived object and per-
ceiving subject, which exist only nominally. Sthiramati follows along these lines in relating
the imagined and dependent natures to the relative and ultimate truths/realities (Salvini 2015,
44–50). The third – the perfect – nature ( pariniṣpannasvabhāva) is the dependent empty of the
imagined. In the Maitreya works, this model is made compatible with Tathāgatagarbha thought
by reducing the substantial existence of the dependent nature to the level of relative truth.
In his commentary on the first verse of the True Reality (tattva) Chapter in The Ornament
of Great Vehicle Discourses, Vasubandhu excludes both existence and nonexistence from true
reality, and in this context takes the ultimate to exist as neither the imagined nor the dependent.
The first line of the first verse in the True Reality Chapter is as follows:
Neither existent nor nonexistent . . .26
Vasubandhu comments:
The ultimate has the meaning of nonduality. It is taught in five points. Neither
existent – in terms of what is called the imagined and dependent – nor nonexistent –
in terms of what is called the perfect characteristic.27
69
Klaus-Dieter Mathes
In other words, the exclusion of what exists nominally and substantially – i.e., the imagined and
dependent natures – from the ultimate and the equation of the latter with buddha nature work
together to smooth out the differences between The Ornament of Great Vehicle Discourses and
An Analysis of the Precious Potential. The problem with the traditional potential (gotra) model
and cut-off potential could be interpreted, as already indicated, along similar lines. One pos-
sible solution would thus be to regard it as a teaching of provisional meaning (neyārtha). But
the problem with that approach is that in The Ornament of Great Vehicle Discourses, it is stated
that the single-vehicle (ekayāna) doctrine – which is a crucial element of the Tathāgatagarbha
teachings – was taught with a purpose, which implies that it has provisional meaning:
The perfect Buddhas have taught

The unity of the vehicle (ekayānatā)
For the sake of those who are unsettled,
To attract some and to hold others.28
In his commentary, Vasubandhu introduces this verse with the following remark:
Buddhahood is the single-vehicle. Thus the unity of the vehicle must be understood
with this or that intent in this or that discourse. But it is not that the three vehicles
do not exist. Why, again, did the buddhas teach the unity of the vehicle with this or
that intent?29
In his subcommentary, Sthiramati concludes:
As for the explanation of “single-vehicle” here, it must be taken to have provisional

meaning,30 because he (i.e., the Buddha) taught it with an intent, namely for the sake
of sentient beings. The teaching of three vehicles has definitive meaning.31
It is not necessary to accept Sthiramati’s conclusion, though, for if one goes by the Yogācāra
hermeneutics of Vasubandhu’s Reasoning of Interpretation (Vyākhyāyukti), every discourse has
an aim or intent, even a discourse of definitive meaning. In other words, the status of provisional
meaning cannot be concluded from the fact that an intent is mentioned (see Mathes 2008, 14–15).
An Analysis of the Precious Potential

In An Analysis of the Precious Potential, the next of the Five Maitreya Works,32 Takasaki
(1966) and Schmithausen (1971) identified older textual layers, which reach back to the begin-
ning of the fourth century CE, the time when the doctrinally close ninth chapter of The Orna-
ment of Great Vehicle Discourses was composed. This is nothing unusual, though, as a teacher
typically draws on earlier memorized passages. The final versions of both texts display a
remarkable synthesis of Yogācāra and Tathāgatagarbha thought, which is indicative of a single
author. In An Analysis of the Precious Potential, the cut-off potential (and thus the strict poten-
tial distinctions) is said to be taught with a hidden intent (and therefore to be of provisional
meaning), since in fact everybody has the naturally pure potential. In his commentary on An
Analysis of the Precious Potential, Asaṅga says:
Again, the saying: “Those with great desire (icchantikas) do not have at all the
capacity to attain perfect nirvāṇa” is taught with the hidden intent of another time to
70
Maitreya
remove hatred towards the Great Vehicle doctrine, this hatred being the reason why
they themselves are those with great desire. Indeed, since everybody has the natu-
rally pure potential, it cannot be that there should be anybody whose nature is the
exact opposite of purity.33
This, in combination with the restriction of the existence of false imagining (i.e., the depend-
ent nature) to the level of relative truth, allows for the notion that all sentient beings possess
buddha nature, the dependent nature being included within the adventitious stains that cover
that nature. The latter is then, as we have seen already in this chapter, identified with mind’s
luminous nature or suchness (i.e., the perfect nature).
The main thesis of An Analysis of the Precious Potential is that all sentient beings possess
buddha nature, which in turn is inseparably endowed with innumerable buddha qualities. This
comes out most clearly in the following verses:
There is nothing to be removed from it

And nothing to be added.
The real should be seen as real,
And seeing the real, one becomes liberated.
The buddha-element is empty of adventitious stains,
Which have the defining characteristic of being separable;
But it is not empty of unsurpassable qualities,
Which have the defining characteristic of not being separable.34
This position is mainly based on the nine similes of buddha nature in The Discourse on Bud-
dha Nature (Tathāgatagarbhasūtra,35 Zimmermann 2002, 64) and other doctrinally related dis-
courses such as The Discourse for Queen Śrīmālā (Śrīmālādevīsūtra)36 and The Chapter on the
Absence of Increase and Decrease (Anūnatvāpūrṇatvanirdeśaparivarta).37 It is not clear, how-
ever, whether all buddha qualities are inseparably contained in one’s nature or only a subset of
them, such as the largely space-like qualities of the embodiment of true nature (svābhāvikakāya)
presented and discussed in An Analysis of the Precious Potential (II.29–47). In other words, the
buddha nature of sentient beings would then not be identical with the body of qualities of a fully
enlightened mind. That this was the intent of An Analysis of the Precious Potential is clear from
its explanation of the fifth simile (a treasure buried under the floor of a poor man’s house) of The
Discourse on Buddha Nature, where buddha nature is fully equated with the thirty-two qualities
of the body of qualities (Zimmermann 2002, 121). An Analysis of the Precious Potential, which
otherwise faithfully renders the nine similes of The Discourse on Buddha Nature, speaks only
of the “treasure of qualities” (dharmanidhi). That this is not simply an unintentional inaccuracy
is clear from An Analysis of the Precious Potential I.149–52, where the treasure illustrates the
naturally present potential, from which the thirty-two qualities of the body of qualities emerge.
In other words, the treasure of buddha nature no longer stands for these thirty-two qualities, but
rather for their cause. It thus is very likely that Maitreya’s An Analysis of the Precious Potential
is a Yogācāra interpretation of older strands of the text that had more closely followed the intent
of the Discourses on Buddha Nature (Tathāgatagarbhasūtras).38 It should be noted, however,
that already two similes from The Discourse on Buddha Nature – namely those of a tree grown
from a seed and the future monarch (cakravartin) in a womb – indicate a growth of the buddha
qualities in terms of strength (Mathes 2008, 8–11).
Whether buddha nature is fully equated with the body of qualities or not, its analysis in
terms of three natures39 results in a luminous perfect nature that is empty of adventitious stains.
71
Klaus-Dieter Mathes
These stains include the imagined nature and dependent nature of false imagining (Mathes
2012, 194). This would not be, then, the original three-nature model of An Analysis of the Mid-
dle and Extreme, the next in line of our Five Maitreya Works, but that of the Great Commen-
tary (Bṛhaṭṭīkā), a large Perfection of Insight (Prajñāpāramitā) commentary usually ascribed
to Vasubandhu and occasionally to Daṃṣṭrasena (Brunnhölzl 2011b, 9–14). If Vasubandhu
truly was the author, the Great Commentary would lend further support to limit the validity of
the three-nature model of An Analysis of the Middle and Extreme and its ontological distinc-
tion between the imagined and dependent natures to the level of relative truth.
An Analysis of the Middle and Extreme

The presentation of false imagining (i.e., the dependent nature) in the beginning of the first chapter
of the An Analysis of the Middle and Extreme is structured around the original Yogācāra model with
its dominant Abhidharma distinction between nominal and substantial existence. As already men-
tioned, this translates into a merely nominal existence of the imagined and a substantially existing
dependent. Salvini (2015, 44–50) shows that substantial existence is equivalent, at least for Stir
amati, to the Abhidharma ultimate. It has been argued that the first part of this chapter also negates
the real existence of nondual mind because in the formulas defining the fourfold Yogācāra practice,
which leads to the realization of a state free from perceived and perceiver, “mind-only” (cittamātra)
and “cognition-only” (vijñaptimātra) are also left behind. False imagining (i.e., “mind-only” as
the dependent) is said to exist, however, and in the formulas describing the fourfold practice,
vijñaptimātra is not the technical term for the Yogācāra tenet of everything existing as cognition-
only but rather simply expresses the logical impossibility of cognition without any object. This is
clear from Vasubandhu’s commentary on An Analysis of the Middle and Extreme I.6cd:
Based on the non-recognition of a perceived object, the non-recognition of mere

cognition (vijñaptimātra) arises.40
Still, the dependent is said to require abandoning in An Analysis of the Middle and Extreme
III.9c,41 and so does not qualify as the ultimate. Line III.10d of the same text goes on to say that
only the perfect nature is accepted as ultimate truth,42 while false imagining (i.e., the dependent)
needs to be abandoned. The obvious solution to the seeming contradiction between an ultimately
and not ultimately existing dependent is that in the description of false imagining in the first part
of the first chapter, the qualifier “ultimately” only refers to an ultimate existence in Abhidharma
terms (where momentary factors of existence constitute the ultimate), while the ultimate in the
third chapter of An Analysis of the Middle and Extreme, as in the sixth chapter of the Ornament
of Great Vehicle Discourses, must be taken as the ultimate in the Tathāgatagarbha system, i.e.,
buddha nature itself. What I propose here, in other words, is that Maitreya attempted a synthesis
of Yogācāra and Tathāgatagarbha thought, a synthesis that can be also found in An Analysis of the
Precious Potential and An Analysis of Phenomena and their True Nature. This requires a shift
from the original Yogācāra model of an ultimate dependent to a Tathāgatagarbha ultimate beyond
the imagined and dependent natures. Understood in this way, the initial verse of An Analysis of the
Middle and Extreme would then present the thesis of such a Yogācāra-Tathāgatagarbha synthesis:
False imagining exists.

Duality is not found in it.
Emptiness is found there (i.e., in false imagining)
And the latter is found in the former.43
72
Maitreya
The first two lines would thus only distinguish nominal and substantial existence on the level of
a Tathāgatagarbha-based relative truth, while the third and fourth lines would be an explanation
of a Tathāgatagarbha-based understanding of the relative and ultimate truths. In other words,
emptiness is not only the absence of duality, but is also understood as luminosity (see An Analy-
sis of the Middle and Extremes I.16 and I.22). Luminous emptiness pervades the dependent
nature’s bearers of properties (dharmin)44 as their true nature (dharmatā), while false imagining
exists in emptiness as bearers of properties. This is as explained in Sthiramati’s commentary:
The existence of emptiness in false imagining must be understood in terms of the

latter’s true nature. False imagining, in turn, exists in emptiness in the form of the
bearers of properties.45
An Analysis of Phenomena and their True Nature

The same relation between phenomena and their true nature is also at the center of the An
Analysis of Phenomena and their True Nature and its commentary by Vasubandhu.46 Their
definitions are as follows:
As to the defining characteristics of phenomena, they are dualistic appearances and

that which appears in accord with expressions; all of them are false imagining.47
As to the defining characteristic of the true nature of phenomena, it is suchness, in which there
is no differentiation between a perceived object and the perceiving subject, an expressed object
and the expression.48
The relation between the two is explained as follows:
The two (i.e., phenomena and their true nature) are neither identical nor separate,
because there is, and also is not, a difference between the existent true nature and
nonexistent phenomena.49
Even though the technical terms neither for relative and ultimate truths/realities nor for the
three natures are found here (and are absent in the commentary, as well), one is left with hav-
ing to include false imagining and the duality it manifests within relative truth, false imag-
ining being the defining characteristic of non-existing phenomena. Still, false imagining is
not completely negated, since it exists as mere delusion, i.e., that which generates dualistic
appearances (see Mathes 1996, 255). Just as in An Analysis of the Middle and Extreme, it must
be eventually abandoned:
Even though false imagining always appears to consciousness, its phenomenal

content is simply nonexistent. Moreover, “when it appears, the true nature does
not appear” means that suchness does not appear. When [false imagining] does not
appear, the true nature appears.50
This happens upon the completion of one’s fundamental transformation (āśrayaparivṛtti):
One comprehends the nature of fundamental transformation known as the stainless-

ness of suchness so that adventitious stains do not appear anymore, and only such-
ness appears.51
73
Klaus-Dieter Mathes
To be sure, the term buddha nature (tathāgatagarbha) is not found in the whole of An Analysis
of Phenomena and their True Nature or its commentary. Still, the proximity of this Yogācāra
text to An Analysis of the Precious Potential is evident, as can be gathered from Vasubandhu’s
final summary, where a permanent true nature (dharmatā) and natural luminosity are disclosed
in just the same way as is the purity of gold, water, and space:
Likewise, in the case of fundamental transformation, it is not that natural luminosity

did not exist before. It only did not appear due to the manifestation of adventitious
hindrances. . . .52
Since that change does not exist, the true nature and the fundamental transforma-
tion (āśrayaparivṛtti), which is characterized by it, are permanent. Here, with the
examples of gold and water, only the quality of these objects of comparison, not their
substance, was taught as being analogous to the transformation. With the example of
space, the transformation was taught completely.53
An Analysis of Phenomena and their True Nature and its commentary display the same syn-
thesis of Yogācāra and Tathāgatagarbha thought as in the first Maitreya works. This synthesis
can be identified as a continuous thread that remedies the flaws Yogācāra has in the eyes of
the followers of the Madhyamaka, namely that a considerable group of sentient beings is com-
pletely cut off from liberation or that a dependently arising mind exists on the level of ultimate
truth. The influence of Yogācāra on the original Tathāgatagarbha doctrine, on the other hand,
prevents a too-substantialist notion of buddha nature.
Whether revealed by a celestial bodhisattva (as tradition would have it) or taught by a Maitreya-
inspired Asaṅga (as David Seyfort Ruegg maintains) or by Maitreyanātha and Sāramati, humans
of flesh and bone (as argued by Erich Frauwallner), there is enough ground for the hypothesis
that the two ornaments and the three analyses, the Five Maitreya Works of the Tibetan tradition,
are from one hand. It is difficult to see how Asaṅga could have played an active role in their
composition, as his own works are largely influenced by the concepts of Śrāvakayāna scholasti-
cism. Instead of Maitreya’s main topic of an inconceivable ultimate empty of duality, Asaṅga
prefers to chart a complex psychological map of mental factors of existence; different types
of consciousness, especially the ground consciousness (ālayavijñāna); and mind’s associated
factors. Following the Chinese tradition, though, Frauwallner argues that the Analysis of the
Precious Potential is by Sāramati while the remaining two analyses and the two ornaments are
by Maitreyanātha. David Seyfort Ruegg replies that the Ornament of Clear Realization and the
group of three Yogācāra works (i.e., The Ornament of Great Vehicle Discourses, The Analysis of
the Middle and the Extreme, and An Analysis of Phenomena and Their True Nature) differ to a
greater extent than the latter three from the An Analysis of the Precious Potential and that if Frau-
wallner tolerates the difference between the three Yogācāra works and The Ornament of Clear
Realization, he should also accept the differences with The Analysis of the Precious Potential. In
line with Seyfort Ruegg, I argue that the Yogācāra influence on the latter and the Precious Poten-
tial’s Tathāgatagarbha influence on the three Yogācāra works can be best explained by assuming
the same author with the agenda of a systematic interpretation of Great Vehicle Buddhism.
Notes
1 That is, following the influential Discourse on the Auspicious Eon (Bhadrakalpikasūtra). See
Tournier (2019, 101).
2 For the dates of Vasubandhu, see Mathes (2008, 45).
74
Maitreya
3 See Vibhāṣā T. 1545, xxxix, 152c (quoted after Demiéville 1954, 376, fn. 4).
4 For a translation of this text, see D’Amato (2012).
5 MAVBh 173: “Having venerated the author of this treatise, who was born from the Sugata himself,”
et passim: The translation of all quoted material is my own.
6 MAVṬ 24: “ Maitreya is the author of this treatise in verse form.”
7 MAVṬ 24–7.
8 For an English translation, see Brunnhölzl (2011a, 475–505).
9 In this contribution, the term Tathāgatagarbha indicates the stream of thought, whereas the term “bud-
dha nature” indicates the idea that is at the heart of that stream of thought.
10 For an English translation, see The Dharmacakra Translation Committee (2014).
11 For an English translation, see Brunnhölzl (2014, 331–460).
12 For an English translation, see Brunnhölzl (2012, 157–71).
13 For the sequence of the Five Maitreya Works, see Mathes (1996, 16–17).
14 It should be noted that the related Yogācāra hermeneutics of Vasubandhu’s “Reasoning of Interpreta-
tion” (Vyākhyāyukti) is at times also taken to stem from Maitreya (see Mathes 2008, 427 fn. 73).
15 AA I.40ab (AA 88).
16 See, for example, AA I.31ab (AA 79): “The fact that matter and the rest lack an own nature refers to
the state of having nonexistence as their nature.”
17 AA V.6–7 (AA 311–4).
18 AA VIII.1 (AA 383–4).
19 AA VIII.11 (AA 397–8).
20 RGVV 79–12.
21 RGV I.6ab (RGVV 81).
22 MSA III.11 (MSABh 1221–24).
23 MSABh 1225–132.
24 MSA IX.37 (MSABh 4013–14).
25 Maitreya invokes the three-nature model of reality only once, in An Analysis of the Middle and
Extremes I.5, where he equates the imagined ( parikalpita) with the cognitive object (artha), the
dependent ( paratantra) with false imagining (abhūtaparikalpa), and the perfect ( pariniṣpanna) with
the nonexistence of duality (dvayābhāva). Generally, though, Maitreya prefers to work with the cat-
egories duality, false imagining, and emptiness. See Mathes (2000, 200ff).
26 MSA VI.1 (MSABh 2212).
27 MSABh 2214–1.
28 MSA XI.54 (MSABh 693–4).
29 MSABh 6827– 692.
30 Tib. bkri ba’i don being an alternative translation to drang ba’i don, Skt. neyārtha.
31 MSAVBh 196a6–7.
32 For an English translation see Brunnhölzl (2014, 337–460).
33 RGVV 371–4.
34 RGV I.154–55 (RGVV 761–4).
35 For an English translation of the Tathāgatagarbhasūtra, see Zimmermann (2002, 93–161).
36 For an English translation, see Wayman and Wayman (1974).
37 For an English translation, see Silk (2015).
38 Tathāgatagarbhasūtras in the plural is a generic term for a group of discourses that teach buddha
nature. The Tathāgatagarbhasūtra is one of them.
39 It should be noted that RGV does not make direct use of three-nature terminology.
40 MAVBh 203–4.
41 MAVBh 4021: “With regard to its comprehension and abandonment,” ( parijñāyāṃ prahāṇe ca).
42 MAV III.10d (MAVBh 4114): “But the ultimate is to be viewed in terms of only one.”
Vasubandhu (MAVBh 4115–16) comments: “Ultimate truth should be understood in terms of the perfect
nature alone.”
Vasubandhu comments (MAVBh 412–3): “With regard to the comprehension and abandonment of the
dependent” ( paratantrasya parijñāne prahāṇe ca |).
43 MAVBh 1716–17.
44 Lit., the bearers of properties.
45 MAVṬ 1517–18.
46 For an English translation, see Brunnhölzl (2012), 173–98.
75
Klaus-Dieter Mathes
47 DhDhVK 19–22.
48 DhDhVK 26–29.
49 DhDhVK 38–41. The additions in brackets are in accordance with Vasubandhu’s commentary (see
Mathes 1996, 122).
50 DhDhVV 571–76.
51 DhDhVVS 12–13.
52 DhDhVV 701–702.
53 DhDhVV 706–708.
Bibliography
Primary Sources
AA Tripathi, Ramshankar, ed. 1993. Abhisamayālaṃkāra. Bibliotheca Indo-Tibet-
ica Series 2. Sarnath: Central Institute of Higher Tibetan Studies.
DhDhVK Dharmadharmatāvibhāgakārikā (Tibetan translation)
Ed. by Klaus-Dieter Mathes. See Mathes 1996, 104–14.
DhDhVV Dharmadharmatāvibhāgavṛtti (Tibetan translation)
DhDhVVS Dharmadharmatāvibhāgavṛtti (Sanskrit fragment)
MAV Madhyāntavibhāga
See MAVBh
MAVBh Nagao, Gadjin M., ed. 1964. Madhyāntavibhāgabhāṣya. Tokyo: Suzuki
Research Foundation.
MAVṬ Yamaguchi, Susumu, ed. 1934. Madhyāntavibhāgaṭīkā. Nagoya: Librairie
Hajinkaku.
MSA Lévi, Sylvain, ed. 1907. Mahāyānasūtrālaṃkāra. Bibliothèque de l’École des Hautes
Études. Sciences historiques et philologiques 159. Paris: Librairie Honoré Champion.
MSABh Mahāyānasūtrālaṃkārabhāṣya
See MSA
MSAVBh Mahāyānasūtrālaṃkāravṛttibhāṣya (Tibetan translation)
Derge bsTan ’gyur 4034, sems tsam, vol. mi, 1b1–283a7
RGV Ratnagotravibhāga
See RGVV
RGVV Johnston, Edward H. 1950. Ratnagotravibhāgavyākhyā. Patna: The Bihar Research
Society, 1950. [The manuscripts A and B on which Johnston’s edition is based are
described in Johnston 1950, vi–vii. See also Bandurski et al. 1994, 12–13.]
References
Bandurski, Frank, Bhikkhu Pāsādika, Michael Schmidt, and Bangwei Wang. 1994. Untersuchungen zur
buddhistischen Literatur. Sanskrit-Wörterbuch der buddhistischen Texte aus den Turfan-Funden.
Beiheft 5. Göttingen: Vandenhoeck & Ruprecht in Göttingen.
Brunnhölzl, Karl. 2010. Gone Beyond, vol. 1. Ithaca, NY: Snow Lion Publications.
———. 2011a. Gone Beyond, vol. 2. Ithaca, NY: Snow Lion Publications.
———. 2011b. Prajñāpāramitā, Indian “gzhan stong pas”, and the Beginning of Tibetan gzhan stong.
Wiener Studien zur Tibetologie und Buddhismuskunde 74. Vienna: Arbeitskreis für Tibetische und
Buddhistische Studien.
———. 2012. Mining for Wisdom Within Delusion. Boston and London: Snow Lion Publications.
———. 2014. When the Clouds Part. Boston and London: Snow Lion Publications.
76
Maitreya
D’Amato, Mario. 2003. “Can All Beings Potentially Attain Awakening? Gotra-Theory in the
Mahāyānasūtrālaṃkāra.” Journal of the International Association of Buddhist Studies 26 (1): 115–38.
———. 2012. Maitreya’s Distinguishing the Middle from the Extremes (Madhyāntavibhāga). Along with
Vasubandhu’s Commentary (Madhyāntavibhāgabhāṣya). A Study and Translation. New York: The
American Institute of Buddhist Studies.
Demiéville, Paul. 1954. “La Yogācārabhūmi de Saṅgharakṣa.” Bulletin de l’École Française d’Extrême
Orient 44 (2): 339–436.
The Dharmacakra Translation Committee. 2014. Ornament of the Great Vehicle Sutras: Maitreya’s
Mahāyānasūtralaṃkāra with Commentaries by Khenpo Shenga and Ju Mipham. Maitreya Texts 3.
Boston and London: Snow Lion Publications.
Frauwallner, Erich. 1994. Die Philosophie des Buddhismus. Berlin: Akademie Verlag.
Johnston, Edward H. 1950. The Ratnagotravibhāga Mahāyānottaratantraśāstra. Patna: The Bihar
Research Society.
Lévi, Sylvain. 1911. Mahāyānasūtrālaṃkāra, vol. 2: Traduction-Introduction-Index. Bibliothèque
de l’École des Hautes Études. Sciences historiques et philologiques 190. Paris: Librairie Honoré
Champion.
Makransky, John J. 1997. Buddhahood Embodied: Sources of Controversy in India and Tibet. SUNY
Series in Buddhist Studies. Albany, NY: State University of New York Press.
Mathes, Klaus-Dieter. 1996. Unterscheidung der Gegebenheiten von ihrem wahren Wesen
(Dharmadharmatāvibhāga). Indica et Tibetica 26. Swisttal-Odendorf: Indica et Tibetica Verlag.
———. 2000. “Tāranātha’s Presentation of trisvabhāva in the gŹan stoṅ sñiṅ po.” Journal of the Inter-
national Association of Buddhist Studies 23 (2): 195–223.
———. 2008. A Direct Path to the Buddha Within: Go Lotsāwa’s Mahāmudrā Interpretation of the
Ratnagotravibhāga. Studies in Indian and Tibetan Buddhism. Boston: Wisdom Publications.
———. 2012. “The Gzhan Stong Model of Reality – Some More Material on Its Origin, Transmission,
and Interpretation.” Journal of the International Association of Buddhist Studies 34 (1–2): 187–226.
Nakamura, Hajime. 1989. Indian Buddhism: A Survey with Bibliographical Notes. New Delhi: Motilal
Banarsidass.
Salvini, Mattia. 2015. “Language and Existence in Madhyamaka and Yogācāra: Preliminary Reflec-
tions.” In Madhyamaka and Yogācāra: Allies or Rivals? edited by Jay L. Garfield and Jan Westerhof,
29–71. Oxford: Oxford University Press.
Schmithausen, Lambert. 1971. “Philologische Bemerkungen zum Ratnagotravibhāga.” Wiener Zeitschrift
für die Kunde Südasiens 15: 123–77.
Seyfort Ruegg, David. 1969. La Théorie du Tathāgatagarbha et du Gotra: Études sur la Sotériologie et
la Gnoséologie du Bouddhism. Paris: École française d’Extrême Orient.
Silk, Jonathan A. 2015. Buddhist Cosmic Unity: An Edition, Translation and Study of the Anūnatvāpūrṇa-
tvanirdeśaparivarta. Hamburg Buddhist Studies 4. Hamburg: Hamburg University Press.
Takasaki, Jikido. 1966. A Study on the Ratnagotravibhāga (Uttaratantra) Being a Treatise on the
Tathāgatagarbha Theory of Mahāyāna Buddhism. Rome Oriental Series 33. Rome: Istituto Italiano
per il Medio ed Estremo Oriente.
Tournier, Vincent. 2019. “Buddhas of the Past: South Asia.” In Brill’s Encyclopedia of Buddhism. Vol. 2:
Lives, edited by Jonathan A. Silk, 95–108. Leiden and Boston: Brill.
Wayman, Alex, and Hideko Wayman. 1974. The Lion’s Roar of Queen Śrīmālā. New York: Columbia
Press.
Zimmermann, Michael. 2002. A Buddha Within: The Tathāgatagarbhasūtra. The Earliest Exposition of
the Buddha-Nature Teaching in India. Bibliotheca Philologica et Philosophica Buddhica 6. Tokyo:
The International Institute for Advanced Buddhology.
77
PART 2
Poet Philosophers
POET PHILOSOPHERS
Texts taught in a typical, contemporary academic philosophy curriculum include a wide vari-
ety of genres. While philosophers today are primarily taught to write academic articles and
monographs, philosophers have historically experimented with a much broader range of liter-
ary forms, including dialogues, aphorisms, personal essays, confessions, epistles, meditations,
autobiographies, and commentaries, in addition to treatises. Some Western philosophers, per-
haps most famously Lucretius and Boethius, expressed their philosophical thinking in poetic
form. For many philosophers, their chosen literary form is not merely one among several
possible ways of clothing their thinking, but is essential to their thought. Consider Plato’s
dialogues, Montaigne’s essays, Kant’s treatises, or Nietzsche’s aphorisms: in each case, the
philosophy is inherently connected to the literary form and would be different if embodied in
another genre. A similar diversity of literary form appears in Indian philosophical traditions,
which have also included treatises, commentaries, dialogues, narratives, and poems. Much
Indian Buddhist philosophy was written according to models of Sanskrit verse. Nāgārjuna,
Vasubandhu, and Śāntideva, for example, and many other prominent Indian Buddhist phi-
losophers, wrote some of their most important works following Sanskrit poetic forms and
integrated elements of dialogue. Thus, in some sense, many Indian Buddhist authors can be
considered “poet philosophers.” The poet philosophers we consider in this part, however, are
poetic in a stronger sense: their philosophy is not merely articulated in verse but embodied in
song and narrative, and their poetic and literary forms do important philosophical work.
In his chapter on Aśvaghoṣa, Vincent Eltschinger notes that this poet from the first cen-
tury of the Common Era – known for Buddhist dramatic works and refined court poetry
(mahākāvya) – was the first Buddhist philosopher to articulate systematic and sophisticated
critiques of other religious-philosophical systems. In Life of the Buddha (Buddhacarita), an
epico-lyrical poem, Aśvaghoṣa defends the rationality of the Buddhist path – in contrast to
other formulations of the path to liberation, especially Sāṃkhya and Vaiśeṣika – and therefore
its effectiveness. In this account, the Buddha is a philosopher who justifies Buddhist doc-
trine and practice on the basis of reason, and reason is necessary for morality and liberation.
Aśvaghoṣa’s views anticipate ideas that are more fully developed by later thinkers associated
with Sautrāntika and Yogācāra. Moreover, Eltschinger argues, Aśvaghoṣa may be the first
Buddhist philosopher we can identify who pursues a distinctive intellectual agenda along with
an argumentative and literary style to achieve it.
DOI: 10.4324/9781351030908-9 81
Poet Philosophers
The following chapter, by Sonam Kachru, is a literary and philosophical reading of the
Questions of Milinda (Milindapañha), a text that presents a dialogue between the Buddhist
monk Nāgasena and the Bactrian King Menander (“Milinda” in Pāli). Kachru is interested
in reading this text as a work of literature that can help us understand what it might be like
to think and feel as a Buddhist. He wants to explore how the Questions of Milinda might be
encountered in ways that can change us as readers. He therefore focuses not only on the argu-
ments, but also on aesthetic features such as the use of examples and the drama of the dialogue.
Kachru reads the drama as an exploration of the role of thought in our lives, and the tragedy of
living a life for which we can give no justification. Kachru also suggests that the text itself can
teach us how to read well and how to reason wisely about our own lives.
In the next chapter, Anne E. Monius addresses philosophy in the work of the Tamil author
Cāttaṉār (ca. sixth–seventh century). Cāttaṉār is said to be the author of The Renunciation
of Maṇimēkalai (Maṇimēkalai Tuṟavu), an extended poetic narrative that traces the path of
Maṇimēkalai from a youthful courtesan to her ordination as a Buddhist nun. It is the first extant
Tamil text of formal philosophy and is representative of the ways in which many Tamil texts
integrated philosophical reasoning into poetic narratives. Three of the chapters are explicitly
philosophical: one critiques non-Buddhist views; the final two present a Buddhist account of
inferential reasoning and interdependent origination. Maṇimēkalai’s understanding of inferen-
tial reasoning and especially interdependent origination as presented by her teacher, Aṟavaṇaṉ,
is what eventually leads her to commit to a life of renunciation. Like Maṇimēkalai, the reader
can look back and see how everything in her life was interdependently originated and what
kinds of conditions can lead to freedom and renunciation of worldly life. As Monius argues,
the beauty and rhetorical power of the narrative help the reader cultivate both an understanding
and an emotional response intended to turn us toward renunciation and freedom.
In the final chapter of this part, Roger R. Jackson presents the late–first millennium Indian
tantric poet, Saraha, as a philosopher. Saraha is informed by the metaphysical and ontologi-
cal thought of Madhyamaka, Yogācāra, and Vajrayāna. Still, as Jackson argues, Saraha is a
kind of philosophical anti-philosopher, as he is critical of philosophical study as a method to
achieve awakening ( just as he is also critical of meditation, pilgrimage, and tantric ritual and
practice – all of which different Buddhists have held to be necessary for liberation). It is only
when we let go of our attachment to conventional forms of practice, Saraha insists, that we can
open ourselves to the ultimate reality beyond language and thought. In his songs, while argu-
ing that the ultimate is beyond the grasp of words, he still seeks to somehow gesture toward it.
This gesturing is accomplished with many terms, including: the innate, stainless mind, inmost
nature, the real, the great seal, nondual mind, emptiness, awakening, Buddha. According to
Saraha’s nondual philosophy, everything in our lives – even what we take to be polluting
and polluted – is ultimately somehow part of the radiant, pure, primordial reality. Instead of
employing sustained rational arguments, Saraha seeks a direct understanding of reality and
poetically presents the human potential for awakening.
82
5
AŚVAGHOS․A
The Dawn of Indian Buddhist Philosophy
Vincent Eltschinger
Aśvaghoṣa (100 CE?) is widely acknowledged as one of the earliest Indian and Buddhist
representatives of ornate court poetry (kāvya).1 But the celebrated poet can with equal right
be regarded as the first Buddhist philosopher – the first, at least, with a clearly identifiable
literary personality, argumentative style, and intellectual agenda. To be sure, the Buddhist
canon on which Aśvaghoṣa so heavily relied contained many interesting arguments against
non-Buddhist – mostly Brahmanical – ideas and practices such as the caste-classes and ritual
violence. But Aśvaghoṣa is the first known Indian Buddhist intellectual to have developed
systematic and often quite sophisticated arguments against concurrent salvation systems,
predominantly Sāṃkhya, likely Buddhism’s most dangerous rival in his time and milieu.
Whereas the poet provides an eloquent outline of the Buddhist path to salvation in his lyri-
cal poem Handsome Nanda (Saundarananda, SNa) – which, like his Drama of Śāriputra
(Śāriputraprakaraṇa), a play dedicated to the conversion of the Buddha’s great disciples
Śāriputra and Maudgalyāyana, contains only very little philosophical discussion – he shows,
in his epico-lyrical Life of the Buddha (Buddhacarita, BC or Life), that the path identified and
experienced by the Bodhisattva is the only rational – hence, legitimate – method for obtain-
ing salvation. In doing so, Aśvaghoṣa portrays the future Buddha as a philosopher assessing
concurrent systems and establishing his own method by means of reasoned arguments alone.
Background
According to the fairly unanimous colophons of his works, Aśvaghoṣa was a native of Sāketa (or
Ayodhyā), the son of lady Suvarṇākṣī (hence probably a brahmin), an eloquent Buddhist monk,
teacher, and poet – perhaps even a great dialectician, depending on the proper interpretation of
mahāvādin.2 Chinese legends, some of which likely go back to Indian prototypes, portray him
as a (likely Śaivite) brahmin who converted to Buddhism after being defeated in debate by the
Sarvāstivādin hierarch Pārśva.3 Be that as it may, Aśvaghoṣa likely was ordained as a monk
in the Sarvāstivāda monastic order. In spite of legends associating him with King Kaniṣka
(crowned in 127 CE) and the so-called Kashmir council, where he would have been in charge of
writing down the Mahāvibhāṣā (a commentary on one of the six Sarvāstivāda canonical Abhi‑
dharma works, the Method for Knowledge (Jñānaprasthāna)),4 there are good reasons to
believe that Aśvaghoṣa was close to a milieu of Sarvāstivādins whose doctrines at times
DOI: 10.4324/9781351030908-10 83
Vincent Eltschinger
conflicted with “orthodox” Sarvāstivādin-Vaibhāṣika tenets and stood much closer to what
later came to be labeled Sautrāntika and/or Yogācāra.5 This is, in fact, hardly surprising, for
Aśvaghoṣa’s Handsome Nanda has been rightly regarded as an early witness to a milieu of
Sarvāstivādins that laid strong emphasis on meditation and visualization practices, the matura-
tion of which eventually resulted in the Yogācārabhūmi (stages of yoga practice/practitioners)
treatises.6 Whatever the case may be, Aśvaghoṣa’s exclusive reliance on sūtras while discuss-
ing doctrinal issues makes him literally – though anachronistically – a Sautrāntika.7
Aśvaghoṣa perhaps would have recognized himself in the description of the Sarvāstivādins
made in a later scripture, the Mahāvaitulya-Mahāsannipāta,8 where the members of this order
appear as experts in Buddhist dogmatics and textual exegesis, but also, and this is perhaps more
surprising, as specialists in the controversy against the non-Buddhists, an area almost com-
pletely absent in the Sarvāstivāda Abhidharma. That the controversy against the non-Buddhists
formed an essential part of the self-representation of the (Mūla)sarvāstivāda order is made per-
fectly clear in a remarkable passage from the Vinaya of the Mūlasarvāstivādins accounting for
the last existence of the Bodhisattva.9 About to leave the Tuṣita heaven where he had spent
his penultimate existence, the future Buddha is held back by the gods, who inform him about
the poisonous conditions obtaining on earth due to the presence of six allodox “reasoners”
(tārkika), six brahmin “traditionalists” (ānuśrāvika), and six “meditators” (samāpattṛ). Far from
being frightened, the Bodhisattva resolutely descends to earth to defeat these perturbators, and
only then to reach nirvāṇa. The Mūlasarvāstivāda Vinaya thus features an agonistic Bodhisattva
coming to earth in order to vanquish non-Buddhist teachers, and there is every reason to suspect
that the (Mūla)sarvāstivādin intellectual elite was thus encouraged to view philosophical con-
troversy against the non-Buddhists as a welcome imitation of the founder. If there is no compel-
ling evidence that Aśvaghoṣa was aware of this passage, there are good reasons to believe that
he was fully sympathetic with its message. This can be seen from his summary of the Buddha’s
career in BC 25.9 (P103a6–7), according to which the Buddha, “after refuting the allodox erro-
neous paths, proceeded on such a path that he could teach the right path,” or in the claim that
by refuting the other systems and by argument he caused men to understand the
meaning which is hard to grasp. By teaching everything to be impermanent and with-
out self and by denying the presence of the slightest happiness in the spheres of exist-
ence, he raised aloft the banner of his fame and overturned the lofty pillars of pride.
(BC 27.30–32, P115b7–116a2)
Aśvaghoṣa’s martial repertoire is thus perfectly in tune with the preceding passage from the
Mūlasarvāstivāda Vinaya and shows that, for him, too, defeating the non-Buddhists was an
essential feature of the Bodhisattva’s spiritual career and a prerequisite for salvation.
Philosophy and Apologetics

Aśvaghoṣa did not build a system of philosophy. Nor did he develop any particular method.
However, if there is nothing like “Aśvaghoṣian” philosophy, his works exhibit a great many
rational arguments concerning issues that eventually became the object of heated philosophi-
cal controversies, as well as a philosophical agenda mirroring distinctly apologetic concerns.
As shown by Patrick Olivelle and others, the highly polysemic notion of Dharma is one of the
Life’s main sites of contention.10 In Chapters 9–10 of the Life, Aśvaghoṣa dismisses orthodox
Brahmanical attempts to prove the Bodhisattva’s abandonment of princely and domestic life to
be morally wrong. For all this, Aśvaghoṣa has nothing to object to the normativity of Dharma
84
Aśvaghos∙a
in the socio-political realm11 as long as – contrary, for instance, to caste regulations – its pre-
scriptions are rationally justified and therefore morally satisfactory. But a significant part of
the Life of the Buddha is intended to demonstrate that no law outside the one identified, expe-
rienced, and preached by the (future) Buddha – i.e., Buddhism – is relevant as far as religion as
a system of salvation is concerned. In other words, Aśvaghoṣa’s Life of the Buddha is destined
to demonstrate Buddhism’s monopoly over salvation, and it is in this apologetic perspective
that Aśvaghoṣa’s activity as a philosopher is best understood.
Let us briefly go back to the canonical accounts of the main events that took place between
the Bodhisattva’s Great Departure and the attainment of Awakening. Gautama had first been
initiated by the teachers Arāḍa Kālāma and Udraka Rāmaputra to higher meditative attain-
ments – the stage of nothingness and the stage of neither-consciousness-nor-unconsciousness.
Judging these practices unsatisfactory, i.e., not conducive to liberation, the Bodhisattva left
these teachers and turned to asceticism, which he subsequently abandoned on the grounds that
it made salvation impossible. However, the canonical accounts provide no argument for the
Bodhisattva’s rejection of these practices. Aśvaghoṣa’s original contribution was to supply rea-
sons (hetu, yukti) in order to account for the future Buddha’s dissatisfaction, thus ascribing him
a systematic assessment of concurrent systems for salvation – in Sanskrit, a parīkṣā, “critical
examination,” or vicāra, “critical analysis.” In Handsome Nanda, the Bodhisattva is described
as a path expert in search of certainty concerning the (un)reliability of the available methods
for salvation and critically analyzing “which of the various scriptural traditions (āgama) in the
world is the highest” (SNa 3.1–5). Critical examination of concurrent truth claims proved to be
of decisive importance for the development of philosophy in the Buddhist environment (at least)
– this is what the Levels of Yoga Practice (Yogācārabhūmi) subjects the “allodoxies” to;12 what
logic and dialectics are designed for according to our earliest extant Buddhist witnesses; the
title of numerous Madhyamaka and epistemological works and chapters; the very way in which
Dignāga (480–540) and Dharmakīrti (around 600?) understand their own philosophical endeav-
ors. Buddhist philosophy essentially consists in a “critical examination,” i.e., in the use of rea-
soning ( yukti) to evaluate concurrent religio-philosophical systems and to establish Buddhism,
or a particular interpretation thereof, as the only rationally founded soteriological method.
If critical examination ( parīkṣā) is the closest one gets to a designation of philosophy in the
early Buddhist environment, can the parīkṣaka – the person who carries out a critical examina-
tion – be regarded as the Indian Buddhist counterpart of the philosopher? Interesting in this
connection is what the formerly Sāṃkhya-oriented ascetic Subhadra says to the Buddha in the
Life (BC 26.7ab, P107a8–b19): “It is said that you have gained a path of salvation other than
that of parīkṣakas like myself.” Significantly, Johnston translates parīkṣaka with “philoso-
phers,” quite rightly so, considering Subhadra’s acknowledgment that
previously he had held with respect to that which is manifested that the self is other
than the body and is not subject to change, and that now that he had listened to the
sage’s words he knew the world to be without self and not to be the effect of self.
(BC 26.17, P108a3–4)13
Subhadra is thus a parīkṣaka in that he holds speculative (and, in his case, ultimately wrong)
views about the world and human nature. Aśvaghoṣa’s parīkṣakas are also particularly inter-
ested in psychology. Accordingly, parīkṣakas who know the soul are said to be doctors for the
mind, and the Buddha is regarded as a parīkṣaka for analyzing the “dispositions, tendencies,
and feelings of all beings” (SNa 8.5 and 9.52cd). It is due to not being a parīkṣaka, i.e., due to
not analyzing things, that the Buddha’s half-brother Nanda is said to have formerly believed
85
Vincent Eltschinger
that heaven was eternal (SNa 12.4ab). Analyzing reality (tattva) is what characterizes parīkṣā
as an intellectual activity and the parīkṣaka as one who devotes themself to it. Whereas Hand-
some Nanda 14.48 shows that being a parīkṣaka does not necessarily presuppose a direct
intellectual intuition of the true nature of things, other passages in the same work clearly link
the parīkṣaka with a soteriologically relevant analysis of reality.14 This was typically the case
of the newly converted Nanda as he finally “examined things according to their prerequisites,
their causation, their nature, the sensation with which they are experienced and their individual
defects” (SNa 17.15). Extrapolating a little, one could perhaps say that parīkṣā as a discursive
operation refers both to philosophy as the critical examination of concurrent speculative views
and to philosophy as a soteriologically oriented analysis of reality. The parīkṣaka’s analysis of
reality causes them both to criticize alternative accounts of the world and to set out on the path
to mental tranquility, insight, and salvation.
A Critique of Rituals and Asceticism

Shortly after departing from his father’s palace, the Bodhisattva visits a forest hermitage (āśrama)
with various sorts of brahmin ascetics and their wives, all “desirous of Paradise and working to
accumulate merit” (BC 7.10). According to Aśvaghoṣa’s lively description, this penance grove
was in full activity, a workshop as it were of dharma, with the transference elsewhere
of the blazing sacrificial fires, with its throngs of seers who had completed their ablu-
tions and with the shrines of the gods humming with the din of prayers.
(BC 7.33)
To achieve their religious ends, the hermits dedicate themselves to austerities (tapas) and prac-
tice various observances and rituals – what the Bodhisattva calls “methods of dharma” (dhar-
mavidhi). “With such austerities accumulated for the due time,” one of the forest-dwellers says,
“they win by the higher to Paradise, by the lower to the world of men. For bliss is obtained
by the path of suffering; for bliss, they say, is the ultimate end of dharma” (BC 7.18). Diverse
as they may be, however, all these ascetic and ritual practices have in common to be aimed at
better forms of rebirth – ideally heaven – but not at getting rid of existence itself, which is the
Bodhisattva’s purpose, he who has “made a vow for the annihilation of existence” (BC 7.44).
Since, at this early stage of his career, the future Buddha has no direct intuition of the
truth, he subjects the hermits’ implicit assumptions to a rational examination. In Life 7.20–31,
the Bodhisattva’s main line of argument against these Brahmanical practices is that they are
absurd as long as they are not aimed at complete liberation from existence (the heavens they
are conducive to are – admittedly ethereal and pleasurable – forms of existence). According
to him, it is absurd to expect heaven, which is itself painful in as much as it is transient (see
what follows), from painful ascetic practices; absurd, to practice restraint in order to achieve
heaven, which, as a “karmically” conditioned state, is nothing but a form of bondage; absurd,
to long for the continuance of existence, with all that is known of the miseries of saṃsāra;
absurd, finally, to expect to rid oneself of the fear of death by ascetic practices that bring about
new births and deaths. As a consequence, “living beings, making themselves miserable in their
hopes of bliss, miss their goal and fall into calamity” (BC 7.24). Moreover, if the mortification
of the body is dharma, then the corporeal bliss that one longs for in heaven is adharma, and
therefore “dharma in this world bears as its fruit what is contrary to dharma” (BC 7.26; and
“in as much as it is under the direction of the mind that the body acts and ceases to act, there-
fore it is the taming of the mind only that is required,” BC 7.27). Does religious merit accrue
86
Aśvaghos∙a
from the purity of food? Then deer and poor people will have much merit. Moreover, water
is just water and will not make a sinner pure. As an ascetic says at the end of the chapter, the
Bodhisattva, “having properly analyzed heaven and release, has opted for release,” and this on
the basis of various arguments (bahuyuktiyukta; BC 7.52 and 32).
The Case of Arāda

Before leaving the forest hermitage, one of the anchorites – likely a Śaiva professing Sāṃkhya
views15 – invites the Bodhisattva to pay a visit to Arāḍa Kālāma, a sage teaching the “path
of the tattvas” (BC 7.55). According to canonical literature, Arāḍa and Udraka Rāmaputra –
Gautama’s first instructors in meditation – were teaching stages of ecstatic trance called the
“incorporeal realms.” These practices likely belonged to what Johannes Bronkhorst has called
“mainstream meditation,” i.e., Jaina and “Hindu” ascetic traditions that – contrary to Bud-
dhism – understood meditation as a painful and “forceful effort to restrain the mind and bring
it to a standstill” and as a medium “by which the practitioner gradually puts an end to all idea-
tions” (Bronkhorst 1993, 22, 81). Aśvaghoṣa thus unsurprisingly presents Arāḍa as an exponent
of liberation (BC 11.69 and SNa 3.3) who has “gained insight into final beatitude” (BC 7.54).
But contrary to the canonical narratives, which remain entirely silent on Arāḍa’s and Udraka’s
theoretical commitments, Aśvaghoṣa introduces Arāḍa’s doctrine as a philosophical system
(darśana, śāstra) entailing both a theoretical and a practical aspect. Thus, far from being just
the meditation specialist depicted in the canonical sources, Aśvaghoṣa’s Arāḍa is, first and fore-
most, a philosopher – more precisely, a representative of the type of “early” Sāṃkhya associ-
ated with the teachers Vārṣagaṇya and Pañcaśikha.16 To the best of my knowledge, however, no
prior Buddhist account of Arāḍa associates him with Sāṃkhya. Moreover, Arāḍa is recorded in
none of the traditional lists of Sāṃkhya and/or Yoga teachers. Arāḍa’s system, as presented by
Aśvaghoṣa, differs markedly from “classical” Sāṃkhya, yet bears close affinities with Sāṃkhya
as it is expounded in the Carakasaṃhitā and in different passages of the Mahābhārata.17
What is striking about Aśvaghoṣa’s reworking of the story is that, far from rejecting
Arāḍa’s and Udraka’s soteriologies after practicing them, as the traditional accounts have it,
the Bodhisattva dismisses them on purely theoretical grounds after submitting them to rational
examination. His argument starts as follows (BC 12.69): “I have listened to this doctrine of
yours, which grows more subtle and auspicious in its successive stages, but I consider it not
to lead to final beatitude, since the field-knower [i.e., the self] is not abandoned.” Arāḍa’s
Sāṃkhyayoga is a soteriological system aiming at the self’s complete isolation from whatever
is not the self, such as passion, corporeality, and mental processes. Once the culmination of
the path – the stage of nothingness – has been reached, the self escapes from the body as a
bird from its cage. In Life 12.69–82 and especially 69–76, the Bodhisattva attempts to dem-
onstrate that this very conception makes salvation altogether impossible.18 Although most of
the subsequent argument operates, I believe, on the ontological level, it can also be interpreted
as suggesting that, since the knower of the field is not abandoned, the very notion of a self –
which, according to most Buddhists, is what enslaves beings to existence and suffering – cannot
be eliminated, either. In his play Drama of Śāriputra, the poet, after “quoting” Life 12.75 (see
what follows), declares that “cessation is not definitive if there still is a belief in a self, and is
due to the perception of selflessness alone.”19 And in Life 17.12–14, Śāriputra – who previously
had been an adherent of Sāṃkhya – criticizes his former creed in the following way:
When one embraces the idea of the soul (ātman) as the origin, there is no abandon-
ment of the ego-principle, and therefore no abandonment of the ego. When a lamp
87
Vincent Eltschinger
and the sun are both present, what is to be known as a cause of the destruction of
light? Just as if one cuts off the roots of a lotus, the fine fibers remain entangled with
one another, so he deemed the [Sāṃkhya] method of salvation not to lead to finality,
while the Buddha’s method was like cutting a stone.
(BC 17.12–14)
Positing a self amounts to reinforcing an already deeply ingrained sense of ego. The percep-
tion of selflessness is the only way to counteract the belief in a self; it is only when the latter
has been eradicated that salvation can be considered truly final. The non-abandonment of the
self makes this impossible.
In Life 12.70, the Bodhisattva gives the reason why in his opinion the self (hence the belief
in a self) cannot be entirely eliminated (BC 12.70): “For I am of opinion that the field-knower,
although liberated from the primary and secondary constituents, still possesses the quality of
giving birth and also of being a seed.” As pointed out by Johnston (1984, 178, n. 70), giving
birth and being a seed are well-known expressions in Sāṃkhya sources, some of which regard
the unmanifest (avyakta) – which, according to them, is none other than the self – as produc-
tive. They then explain how the intellect, the ego-principle, the five great elements, etc., are
successively evolved from the unmanifest, the self. Life 12.70 seems to presuppose similar
ideas on the self: far from being the unproductive witness of classical dualistic Sāṃkhya, the
self is the dynamic source of everything else. But if the qualities of giving birth and being a
seed belong to the very essence of the self/field-knower, they will so remain even beyond lib-
eration. In other words, the self will continue to produce that which it is supposed to rid itself
of, at least provided the causal conditions are present.
This is apparently what Aśvaghoṣa suggests in the next stanzas:
For although the soul by reason of its purity is conceived as being liberated, it will
again become bound from the continued existence of the causal conditions. Just as a
seed does not grow for want of the proper season, soil or water, but does grow when
these causal conditions are present, such I deem to be the case of the soul.
(BC 12.71–72)
To better understand this, let us turn to Life of the Buddha 26.11–14. In BC 26.11, which
presents a preclassical Sāṃkhya path to liberation, the release of the self is claimed to come
about by the increase of sattva (“goodness”) and the destruction of the result of past action
(karman) through the elimination of rajas (“passion”) and tamas (“darkness”).20 Against these
tenets of his former Sāṃkhya persuasion, according to which prakṛti (“Nature”) is permanent,
Subhadra argues as follows:
For in the world they attribute darkness and passion, which delude the mind, to
Nature. Since Nature is acknowledged to be permanent, those two equally do not
cease to exist, being necessarily also permanent. Even if by uniting oneself with
goodness those two cease to exist, they will come into being again under the compul-
sion of time, just as water, which gradually becomes ice at night, returns to its natural
state in the course of time. Since goodness is permanent by nature, therefore learning,
wisdom and effort have no power to increase it; and since it does not increase, the
other two are not destroyed, and since they are not destroyed, there is no final peace.
(BC 26.12–14, P107b3–108a6)
88
Aśvaghos∙a
The parallelism between the arguments of BC 12.71–72 and 26.12–14 is striking.

In addition,
as for the statement that liberation is deemed to come by severance from the power
of the act, from ignorance and from desire, there is no complete severance from them
so long as the soul persists. It is true that advance is obtained by the progressive
abandonment of these three, but where the soul still remains, there these three remain
in a subtle state.
(BC 12.73–74)
To be sure, progress on the path and liberation presuppose the gradual/total elimination
of ignorance, craving, and (bad) action, but the belief in a self makes such an elimination
impossible, for this belief is the very content of ignorance and craving, its consequence.
In other words, ignorance, craving, and action can perhaps be reduced, but certainly not
uprooted.
Besides, the liberation claimed by Arāḍa to take place upon reaching the stage of nothing-
ness is purely illusory, as the Bodhisattva makes clear in BC 12.75: “But such liberation is
a creation of the imagination based on the subtlety of the faults, the inactivity of the mind
and the length of life in that state.” The inactivity of the mind and the length of life alluded
to here are to be understood in the context of the fourth dhyāna, which plays a major role
both in the Buddhist and in Arāḍa’s analysis. Whereas the inactivity of the mind likely refers
to the suspension of intellectual and affective events arrived at by the yogin, the length of
life is that of the deities inhabiting those ethereal realms. And indeed, life expectancy in the
incorporeal realms is well codified in Abhidharma: 20,000 cosmic periods in the stage of
the infinity of space; 40,000 in the stage of the infinity of cognition; 60,000 in the stage of
nothingness; and 80,000 in the highest state of existence.21 According to the Bodhisattva,
then, those rarefied cognitive and affective events occurring in next to endless existences
can easily be mistaken for liberation. But from a Buddhist perspective, there can be no ques-
tion of liberation at that stage. However subtle, defilements are still polluting those living in
these realms. From a Buddhist perspective, the elimination of defilements by non-Buddhist
(“mundane”) paths is neither sufficient nor capable to lead to salvation; the elimination of
false views – first and foremost, the false view of a self – is required, and this can only occur
by perceiving the Buddhist truths.
BC 12.77 opens a new section in the Bodhisattva’s argument, with Vaiśeṣika now under
attack, as suggested by Bronkhorst: “And as the [soul] is not released from number, etc., it is not
devoid of qualities; therefore, as it is not devoid of qualities, it is not admitted to be liberated.”22
In Vaiśeṣika, numbers are qualities (guṇa), and “even a liberated soul will, from the Vaiśeṣika
perspective, possess the quality ‘number’ by virtue of the fact that it has a number: each liber-
ated soul by itself is one in number” (Bronkhorst 2005, 595). And among the many qualities that
can reside in the soul, some (including number) are explicitly said to remain in a liberated soul.
As remarked further by Bronkhorst, the next stanza, BC 12.78, is also to be interpreted against a
Vaiśeṣika background: “For no distinction exists between the qualities and the possessor of the
qualities; for instance, fire is not perceived, when devoid of outward appearance (rūpa) and heat
(uṣṇa).” This argument would make little sense against Sāṃkhya, according to which every-
thing merely consists of a collection of qualities with no distinction between the qualities (guṇa)
and their possessor (guṇin). According to the Bodhisattva, not only does the soul possess quali-
ties such as number, but it also must possess them since nothing can be perceived independently
89
Vincent Eltschinger
of its qualities. With BC 12.79, Aśvaghoṣa seems to move one step further, claiming that no
guṇin can be said to exist without qualities, which corresponds to the Vaiśeṣika understanding
of substances (dravya):
Before a conglomerate mass exists, there cannot be a possessor of the mass; so,
before qualities exist, there cannot be a possessor of the qualities. Therefore the soul,
as possessor of the body, being first released, is subsequently bound to it again.
The Bodhisattva’s last argument against Arāḍa’s position is in the form of a dilemma (BC
12.80–81):
And the knower of the field, when without a body, must be either knowing or
unknowing. If it is knowing, there is something for it to know, and if there is some-
thing for it to know, it is not liberated. Or if your teaching is that it is unknowing,
what then is the use of inventing the existence of a soul? For even without a soul the
existence of the quality of not-knowing is well established as in the case of a log or
a wall.
As far as I can see, the Sāṃkhya, both “early” and “classical,” is immune to the first branch
of the dilemma. The position of classical Sāṃkhya on the issue of the self’s liberated state
can be summarized as follows: although it is conscious by its very nature, the self does not
cognize anything due to its dissociation from prakṛti and its evolutes – the intellect, the mind,
the sense organs, their objects, etc. The liberated self is thus characterized by a kind of non-
intentional, contentless, or non-referential consciousness. This position exposes the Sāṃkhya
to the second branch of the dilemma, following which an unknowing liberated self would
be similar to a log or a wall and thus both unattractive and useless. Such a liberated self is
frequently compared to a stone by the opponents of the Vaiśeṣika and the Naiyāyika view
of liberation.23 And indeed, that the liberated self is insentient (thus incapable of experienc-
ing pain, but also pleasure) and unknowing is a characteristic position of the Vaiśeṣika and
Nyāya schools.24 As objections found in Pakṣilasvāmin’s and Uddyotakara’s commentaries
on Nyāyasūtra 1.1.22 clearly testify, this view has been criticized from very early times on
the grounds that such a liberated state cannot be striven for. Aśvaghoṣa certainly belongs
to the first philosophers to criticize these Sāṃkhya – and, most probably, Vaiśeṣika – ideas
on the liberated state.
The Bodhisattva concludes his critique of Arāḍa’s position on the following stanza (BC
12.82): “But since this successive abandonment is declared to be meritorious, therefore I deem
complete success in reaching the goal to derive from the abandonment of everything.” In
accepting that a self is liberated, Arāḍa’s system fails to abandon everything, and thus makes
the elimination of the belief in a self, hence liberation, impossible. Liberation presupposes
riddance not only of what is falsely believed to belong to the self – the Sāṃkhya prakṛti-cum-
vikāra, the Vaiśeṣika guṇas, and so on – but also of the self.
Against the Self

In his critique of Arāḍa’s Sāṃkhya system, however, Aśvaghoṣa does not refute the self. His
arguments appear somewhat later in the poem, in canto 16, as the newly awakened Buddha
preaches selflessness to King Bimbasāra and his subjects while arriving at Rājagṛha.25
90
Aśvaghos∙a
According to the Buddha,
if a self existed, it would be either permanent or impermanent; great problems (doṣa)

follow in both these hypotheses. To begin with, if it were impermanent, there would
be no fruit of the act; and, since there would thus be no rebirth, salvation would come
without effort.
(BC 16.80–81)
An ephemeral agent bears no moral responsibility for his own actions, for there is no con-
tinuum to link up actions and their results by underlying the maturation process. But insofar
as it is past actions that bind to saṃsāra, there would be no bondage, which simply amounts
to liberation. In other words, the final aim of religious life would be achieved without effort.26
Aśvaghoṣa continues (BC 16.82): “But if the self were permanent and all-pervading, there
would be neither absence of birth nor birth; for space, which is all-pervading and permanent,
neither passes away nor is born.” By definition, whatever is permanent and thus undergoes no
change knows neither rise nor fall. However, living beings are seen to be born and to die.27 If
the self is all-pervading but not permanent, the following absurdity ensues (BC 16.83): “If this
self were all-pervading in nature, there would be no place where it is not; and when it passed
away, there would simultaneously be salvation for everyone together.” Suppose the self is
all-pervading: it will, then, be present in all beings engaged in all retributive destinies, and
its passing will entail the same consequences as those outlined in Life 16.81, except for their
collective, universal application: for all living beings engaged in transmigration, at the time of
this universal self’s destruction, bondage will stop; hence, liberation be achieved, for want of
karmic results and rebirth.
In addition, “as being all-pervading by nature, the self would be inactive and there would
be no doing of the act; and without the doing of acts, how could there be the union with the
fruit of them?” (BC 16.84). Action involves an agent, an instrument, something that is acted
upon, and a specific location where it takes place. An all-pervading entity would be either all
this at the same time, or rule out the possibility of anything other than itself. But the absence
of both action and agent makes karmic retribution and an experiencer impossible. Suppose,
however, that the self – which both Buddhist and Brahmanical sources take to be autonomous
(svatantra) – is an agent. Then “if this self did perform deeds, it would cause no suffering to
itself; for who, that is absolutely autonomous, would cause suffering to himself?” (BC 16.85).28
Let it be recalled here that the prototype of Aśvaghoṣa’s argument is the so-called second ser-
mon of the Buddha in Benares, in which the Buddha speaks as follows (SBhV I.138,10–20):
Corporeality, O monks, is not the self. For if corporeality were the self, O monks,
corporeality would tend neither to harm nor to suffering, and with regard to corpo-
reality wishes such as: “I wish my body were so and so,” or “may it not be so and
so” would be fulfilled. But since corporeality is not the self, therefore corporeality
tends to harm and suffering, and with regard to corporeality wishes such as: “I wish
my body were so and so,” or “may it not be so and so” are not fulfilled. Similarly,
affective sensation, ideation, the conditioning factors and awareness are not the self.29
Aśvaghoṣa continues (BC 16.86): “Due to its being permanent, the self undergoes no
change; but since it experiences pleasure and suffering, we see that it does in fact undergo
change.” The idea expressed in this stanza underlies the entire Buddhist critique of permanent
91
Vincent Eltschinger
entities such as God and the self: either such an entity is permanent – and hence, it cannot
undergo any change – or it undergoes change – and hence, it cannot be regarded as permanent.30
The permanence of the self is not only challenged by the many changes this self is bound to
undergo as an agent, an experiencer, and the very substratum of psychophysical life. Another
threat consists in the fact that salvation presupposes change in the form of a process of gradual
improvement and purification. Aśvaghoṣa spells this out as follows (BC 16.87): “Salvation
comes from the winning of knowledge and the abandonment of defilements; but since the self
is inactive and all-pervading, there is no salvation for it.” As we have seen previously, a per-
manent and all-pervading self neither changes nor acts. However, transformation and action
are necessary conditions for achieving liberation. Indeed, soteriology – and especially Bud-
dhist soteriology – entails both a via illuminativa and a via purgativa: the gradual acquisition
of gnosis through scriptural instruction, philosophical investigation, and meditative practice
parallels (and more often than not conditions) the progressive neutralization and elimination of
gross and subtle defilements. Now, according to Aśvaghoṣa, the self’s all-pervasiveness makes
liberation impossible in that it precludes agency and action.31
On Some Metaphysical Assumptions

In SNa 16.17, Aśvaghoṣa claims that
the cause of this suffering from active being in the world is to be found in the cat-
egory of the vices such as desire and the rest, not in a Creator or Primordial Matter or
Time or the Nature of Things or Fate or Chance.
The same topos underlies BC 18.18–56, a fascinating but difficult passage32 in which Aśvaghoṣa
(alias Anāthapiṇḍada/Sudatta, the rich merchant who gave the Buddhist community its first
monastery, the Jetavana in Śrāvastī) criticizes all available metaphysical hypotheses concern-
ing the origin and/or the nature of the universe: God, Nature, Man, Time, Chance/Causeless-
ness. In these two passages, the poet resorts to a fairly widespread motif whose locus classicus
(and origin?) is often believed to be Śvetāśvatara Upaniṣad 1.1–2ab:
What is the cause of brahman? Why were we born? By what do we live? On what
are we established? Governed by whom, O you who know brahman, do we live in
pleasure and pain, each in our respective situation? Should we regard it as time, as
inherent nature, as necessity, as chance, as the elements, as the source of birth, or as
the Person? Or is it a combination of these?
This ubiquitous motif likely was intended to map the various theoretical possibilities in meta-
physical matters and, perhaps as a sort of scholarly exercise, to allow one to find one’s place
in the picture and to criticize the other hypotheses. Be that as it may, the topos can be seen
to structure much of philosophical controversy until at least Dharmakīrti (PV 2.179ff.) and
Śāntarakṣita’s Tattvasaṅgraha (8th century CE).
Thanks to a sermon delivered by the Buddha (BC 18.3–14), Sudatta has reached the first fruit
of religious life. “With the correct view born in him, he shed the various wrong views, like an
autumnal cloud shedding a shower of stones, and he did not hold that the world proceeded from
a wrong cause, such as a Creator and the like, or that it was uncaused” (BC 18.18, P79a6–7).
The wrong causes (viṣamahetu) criticized in canto 18 consist in a creator God (īśvara, verses
19–28), Nature (either svabhāva or prakṛti, verses 29–41),33 Time (kāla, verse 42), and Man
92
Aśvaghos∙a
( puruṣa, verses 47–51); as for the absence of cause, it is refuted in verses 52–54. According to
Aśvaghoṣa, if God were the creator, the world would know neither process nor change, and the
living beings no transmigration; men would have no doubt about him and love him like a father;
they would have no free will and bear no responsibility. If Nature is one (ekātman), all-pervad-
ing (vyāpin, hence the universal cause of everything), without attribute (nirguṇa), permanent
(nitya, hence not liable to change), productive in essence ( janaka), as its advocates believe,
then its effects should be one, unlimited, without attribute, undifferentiated, and indestructible.
However, one observes that, contrary to this, they are manifold, limited, endowed with specific
qualities, differentiated, and ephemeral. “If Time is postulated as the creator of the world, then
there is no liberation for seekers. For the cause of the world would be perpetually productive, so
that men would have no end” (BC 18.42, P80b3–4). Is man, then, in control of his own destiny?
Certainly not, for, as already argued in the case of the self, he would certainly do only what is desir-
able to himself, obtain whatever he desires, and not let himself be reborn as an animal. Finally, that
all things and events obey causes and conditions can be ascertained by mere empirical observation
(BC 18.52, P81a6–7): does not “corn grow from the seed with the support of soil and water and
by union with the right season”? Verse 56 (P81b2–3) spells out Sudatta’s conclusion: “The various
beings too, moving and stationary, come into existence in dependence on various causes; there is
nothing in the world without a cause, yet the world does not know the universal cause.”
Conclusion
As we can see, Aśvaghoṣa was very well acquainted with contemporary forms of Sāṃkhya and
a witness to the rise of Vaiśeṣika, two philosophical systems that remained the main targets of
the Buddhist intellectual elites until the sixth century. In addition, Aśvaghoṣa likely took an
active part in the formation of Sautrāntika and Yogācāra ideas. An uncompromising defender
of Buddhism, he fought against all non-Buddhist religio-philosophical systems to impose Bud-
dhism as a legitimate form of life and as the only source of salvific truth. Like his near con-
temporary Mātṛceṭa, another celebrated poet, Aśvaghoṣa deeply and permanently impregnated
the Buddhists’ imaginaire, aesthetic sensibility, and philosophical repertoire. Even if pramāṇas
play virtually no role in his philosophy and the Madhyamaka was apparently not known to
him, Aśvaghoṣa surely was one of those who most decisively paved the way for later, “mature”
expressions of Buddhist philosophy. Throughout his works, Aśvaghoṣa insisted that philosoph-
ical analysis is a precondition of religious commitment, and reason, a sine qua non of libera-
tion. This is a lesson many among our contemporaries would be well advised to meditate upon.
Abbreviations
BC (Buddhacarita, by Aśvaghoṣa). See Johnston 1984. Unless otherwise mentioned, all
translations of the BC are Johnston’s.
P (Peking Edition of the Tibetan Buddhist Canon). See Suzuki 1957.
PV (Pramāṇavārttika, by Dharmakīrti). See Miyasaka 1971–1972.
SBhV (Saṅghabhedavastu). See Gnoli 1977.
SNa (Saundarananda, by Aśvaghoṣa). See Johnston 1928 for the Sanskrit and Johnston
1932 for the English translation. Unless otherwise mentioned, all translations of the
SNa are Johnston’s.
ŚvUp (Śvetāśvatara Upaniṣad). See Olivelle 1998, 413–33.
T. (Taisho Edition of the Chinese Buddhist Canon). See Takakusu and Watanabe
1924–1932.
93
Vincent Eltschinger
Notes
1 On Aśvaghoṣa, see Salomon (2015).
2 On these colophons, see Eltschinger (2012, 172, n. 3).
3 On these legends, see Young (2015).
4 See Takakusu (1904, 276–79; Li 2002, 9–13); concerning Kaniṣka, see Lévi (1896, 448–49, 472–75).
5 See especially Yamabe (2003); Eltschinger (2020).
6 See Demiéville’s classic exposition in Demiéville (1951); Yamabe and Sueki (2009, xiii–xviii).
7 See Eltschinger (2020).
8 T. 397 (大方等大集經), XIII.159a16–21. For a translation, see Karashima (2015, 138–39).
9 For a more detailed discussion of this passage, see Eltschinger (2019).
10 See Olivelle (2009, xvii–xlix, especially xliii–xlix); Brocquet (2015).
11 See Eltschinger (2018).
12 On the paravāda section of the Yogācārabhūmi, see Eltschinger (2013–2014).
13 For similar statements, see BC 17.4 (P75b7; Johnston 1984, III.24) and 17.10 (P76a5–6; Johnston
1984, III.25).
14 SNa 16.3, tattvaparīkṣaṇa; 17.15, dharmeṣu cakre . . . parīkṣām; 18.42, vijñātatattvasya parīkṣakasya;
16.47, dhātūn. . . parīkṣamāṇaḥ. On the soteriological aspects of parīkṣā/parīkṣaka, see SNa 9.52,
14.7, 16.3, and 14.7, as well as 18.42 and 18.60.
15 See BC 7.51 and 53, and Johnston (1937, 35–36).
16 Johnston (1937, 9–10). For a description of these teachings, see Johnston (1937, 82–83); on Aśvaghoṣa
and Vārṣagaṇya, see Johnston (1937, 66, 71, n. 1, 78, 88).
17 On the Sāṃkhya known to Aśvaghoṣa, see Ramakrishna Rao (1964); Kent (1982). In my opinion,
these two scholars forcefully interpret Aśvaghoṣa’s description of Arāḍa’s Sāṃkhya according to the
Sāṃkhyakārikā’s dualistic scheme. For another interpretation, see Eltschinger (forthcoming).
18 Dharmakīrti’s critique of the self as the basis of Brahmanical soteriologies, though much more sys-
tematic and sophisticated, is based on the same principle. See Eltschinger and Ratié (2013, 187–283).
19 See Lüders (1911, 194).
20 See previously in chapter, and Johnston (1937, 35–36).
21 See La Vallée Poussin (1923–1931, II.173–174).
22 Translation by Bronkhorst (2005, 594).
23 See Eltschinger and Ratié (2013, 235, n. 188).
24 See Eltschinger and Ratié (2013, 231–36).
25 For a contextualized discussion of these arguments, see Eltschinger (2013).
26 For a similar argument, see Lamotte (1949, 742–43).
27 For a similar argument, see Lamotte (1949, 742). To the self’s permanence, Aśvaghoṣa adds all-
pervasiveness. In other words, there is no point in either time or space that the self would not occupy,
so that this self neither changes nor moves.
28 For a similar argument, see Lamotte (1949, 743); Eltschinger (2010, 318).
29 On this sermon, see Wynne (2009).
30 For a similar argument, see Lamotte (1949, 742).
31 Note, however, that most of the later Buddhist arguments do not argue for the impossibility of libera-
tion due to the self’s omnipresence, but due to its permanence. According to Dharmakīrti (PV 2.202),
the self’s permanence makes both bondage and liberation impossible:
[There can be neither bondage nor liberation for something permanent.] Because the cause of the aris-
ing of suffering is [what we call] ‘bondage’: [since what is permanent is devoid of the capacity to bring
about anything,] how [could this belong] to [something] permanent? [And] being the cause of the non-
arising of suffering is [what we call] ‘liberation’: how [could this belong] to [something] permanent?
(Translation by Eltschinger and Ratié 2010, 204–5)
See also Eltschinger (2010, 321); Lamotte (1949, 742).
32 In spite of Johnston’s heroic and truly remarkable rendering, Aśvaghoṣa’s critique of these five
hypotheses remains very terse due to the Tibetan translation in which it has to be read. Only a short
overview will be provided here. Note that, perhaps due to an interpolation, a form of early Sāṃkhya
is criticized as a sixth view in verses 43–46 (44–46 are missing in Chinese).
33 Johnston suggested that Tibetan rang bzhin renders Sanskrit svabhāva; I believe that it rather renders
Skt. prakṛti. A Sanskrit fragment or quotation would settle the matter.
94
Aśvaghos∙a
References
Brocquet, Sylvain. 2015. “Le mot dharma dans le Buddhacarita. Polysémie et rhétorique de la con-
version.” In Poïkiloï Karpoï. Exégèses païennes, juives et chrétiennes. Études réunies en hommage
à Gilles Dorival, edited by Mireille Loubet and Didier Pralon, 283–311. Aix-en-Provence: Presses
Universitaires de Provence.
Bronkhorst, Johannes. 1993. The Two Traditions of Meditation in Ancient India. New Delhi: Motilal
Banarsidass.
———. 2005. “Aśvaghoṣa and Vaiśeṣika.” In Buddhism and Jainism, Essays in Honour of Dr. Hojun
Nagasaki on His Seventieth Birthday, edited by “Editorial committee,” 596–90. Kyoto: Heirakuji
Shoten.
Demiéville, Paul. 1951. “La Yogācārabhūmi de Saṅgharakṣa.” Bulletin de l’École française d’Extrême-
Orient 44 (2): 339–436.
Eltschinger, Vincent. 2010. “On a Hitherto Neglected Text against Buddhist Personalism:
Mahāyānasūtrālaṅkāra 18.92–103 and Its Bhāṣya.” Études Asiatiques/Asiatische Studien 64 (2):
291–340.
———. 2012. “Aśvaghoṣa and His Canonical Sources: II. Yaśas, the Kāśyapa Brothers and the Buddha’s
Arrival in Rājagṛha (Buddhacarita 16.3–71).” Journal of the International Association of Buddhist
Studies 35 (1–2): 171–224.
———. 2013. “Aśvaghoṣa and His Canonical Sources: I. Preaching Selflessness to King Bimbisāra and
the Magadhans (Buddhacarita 16.73–93).” Journal of Indian Philosophy 41 (2): 167–94.
———. 2013–2014. “The Yogācārabhūmi Against Allodoxies ( paravāda): 1. Introduction and Doxogra-
phy.” Wiener Zeitschrift für die Kunde Südasiens 55: 191–234.
———. 2018. “Aśvaghoṣa on Kings and Kingship.” Indo-Iranian Journal 61: 311–52.
———. 2019. “Why Did the Buddha Come to Earth? Monastic Lineage, Interreligious Polemics, and
Philosophy.” In Reasons and Lives in Buddhist Traditions: Studies in Honor of Matthew Kapstein,
edited by Dan Arnold, Cécile Ducher, and Pierre-Julien Harter, 293–306. Somerville, MA: Wisdom
Publications.
———. 2020. “The ‘dhyāna-Master’ Aśvaghoṣa on the Path, Mindfulness, and Concentration.” In
Mārga. Paths to Liberation in South Asian Buddhist Traditions, edited by Cristina Pecchia and Vin-
cent Eltschinger, 99–176. Vienna: Austrian Academy of Sciences Press.
———. Forthcoming. “Aśvaghoṣa on Sāṅkhya.”
Eltschinger, Vincent, and Isabelle Ratié. 2010. “Dharmakīrti Against the pudgala.” Journal of Indian and
Tibetan Studies (Indogaku Chibettogaku Kenkyu) 14: 185–215.
———. 2013. Self, No-Self, and Salvation. Dharmakīrti’s Critique of the Notions of Self and Person.
Vienna: Austrian Academy of Sciences Press.
Gnoli, Raniero. 1977. The Gilgit Manuscript of the Saṅghabhedavastu. Being the 17th and Last Sec-
tion of the Vinaya of the Mūlasarvāstivādin. Part I. Rome: Istituto Italiano per il Medio ed Estremo
Oriente.
Johnston, E. H. 1928. The Saundarananda of Aśvaghoṣa. Critically Edited with Notes. London: Hum-
phrey Milford, Oxford University Press.
———. 1932. The Saundarananda or Nanda the Fair. Translated from the Original Sanskrit of
Aśvaghoṣa. London: Humphrey Milford, Oxford University Press.
———. 1937. Early Sāṃkhya. An Essay on Its Historical Development According to the Texts. London:
The Royal Asiatic Society.
———. 1984 (1936). Aśvaghoṣa’s Buddhacarita or Acts of the Buddha in Three Parts: Sanskrit Text of
Cantos I – XIV with English Translation of Cantos I – XXVIII, Cantos I to XIV Translated from the
Original Sanskrit Supplemented by the Tibetan Version and Cantos XV to XXVIII from the Tibetan and
Chinese Versions. New Delhi: Motilal Banarsidass.
Karashima, Seishi. 2015. “Who Composed the Mahāyāna Scriptures? – The Mahāsāṃghikas and Vai
tulya Scriptures.” Annual Report of the International Institute for Advanced Buddhology at Soka
University (for the Academic Year 2014) 18: 113–62.
Kent, Stephen. 1982. “Early Sāṃkhya in the Buddhacarita.” Philosophy East and West 32 (3): 259–78.
La Vallée Poussin, Louis de. 1923–1931. L’Abhidharmakośa de Vasubandhu, 6 vols. Paris and Louvain:
Paul Geuthner, J.-B. Istas.
Lamotte, Étienne. 1949. Le Traité de la Grande Vertu de Sagesse de Nāgārjuna (Mahāprajñāpāramitāśāstra),
vol. II. Louvain-la-Neuve: Université de Louvain, Institut Orientaliste.
95
Vincent Eltschinger
Lévi, Sylvain. 1896. “Notes sur les Indo-Scythes.” Journal Asiatique [9e Série] 8: 444–84.
Li, Rongxi. 2002. “The Life of Aśvaghoṣa Bodhisattva. Translated from the Chinese of Kumārajīva
(Taishō Volume 50, Number 2046).” In Lives of Great Monks and Nuns, 3–13. Berkeley: Numata
Center for Buddhist Translation and Research.
Lüders, Heinrich. 1911. “Das Śāriputraprakaraṇa, ein Drama des Aśvaghoṣa.” Sitzungsberichte der
königlich preussischen Akademie der Wissenschaften zu Berlin: 388–411.
Miyasaka, Yusho. 1971–1972. “Pramāṇavārttika-kārikā (Sanskrit and Tibetan).” Acta Indologica 2:
1–206.
Olivelle, Patrick. 1998. The Early Upaniṣads. Annotated Text and Translation. New Delhi: Munshiram
Manoharlal.
———. 2009. Life of the Buddha by Aśvaghoṣa. New York: New York University Press, JJC Foundation.
Ramakrishna Rao, K. B. 1964. “The Buddhacarita and the Sāṃkhya of Arāḍa Kālāma.” Adyar Library
Bulletin 28 (3–4): 231–41.
Salomon, Richard. 2015. “Narratives and Long Poetry: Aśvaghoṣa.” In Brill’s Encyclopedia of Buddhism,
edited by Jonathan Silk, Vincent Eltschinger, and Oskar von Hinüber, vol. I, 507–14. Leiden: Brill.
Suzuki, Daisetz T. 1957. The Tibetan Tripitaka, Peking Edition, Kept in the Library of the Otani Univer-
sity, Kyoto. Tokyo and Kyoto: Tibetan Tripitaka Research Institute.
Takakusu, Junjiro. 1904. “The Life of Vasubandhu by Paramārtha.” T’oung Pao [Second Series] 5:
269–96.
Takakusu, Junjirō 高楠順次郎, and Watanabe Kaikyoku 渡邊海旭. 1924–1932. Taishō shinshū daizōkyō
大正新脩大蔵経. Tokyo: Taishō Issaikyō Kankōkai 大正一切經刊行.
Wynne, Alexander. 2009. “Miraculous Transformation and Personal Identity: A Note on the First anātman
Teaching of the Second Sermon.” Thai International Journal for Buddhist Studies 1: 85–113.
Yamabe, Nobuyoshi. 2003. “On the School Affiliation of Aśvaghoṣa: ‘Sautrāntika’ or ‘Yogācāra’?” Jour-
nal of the International Association of Buddhist Studies 26 (2): 225–54.
Yamabe, Nobuyoshi, and Fumihiko Sueki. 2009. The Sutra on the Concentration of Sitting Meditation
(Taishō Volume 15, Number 614). Translated from the Chinese of Kumārajīva. Berkeley: Numata
Young, Stuart H. 2015. Conceiving the Indian Buddhist Patriarchs in China. Honolulu: University of
Hawai’i Press, Kuroda Institute.
96
6
THE MILINDAPAÑHA
How to Use a Philosophical Resource and
Find a Literary Gem
Sonam Kachru
– “Reverend Nāgasena, will you converse with me?”

– “I will . . . if you converse in the speech of the wise and learned,
but not if you converse in the speech of kings.”
The Questions of Milinda (after Horner 1963, 38–39)
The Questions of Milinda – or Milindapañha in Pāli1 – is the most widely known version
today of The Questions of Menander, the name I will initially use for what scholars believe to
have been a genre of works from ancient South Asia. The Questions of Menander ostensibly
records for us conversations held at Sāgala (today, Sialkot) over a few days, featuring the
Bactrian King Menander and a Buddhist monk named Nāgasena. (The name “Milinda” is how
“Menander” survives in Pāli.) In addition to the Pāli work, we also possess in Chinese The
Sūtra of the Monk Nāgasena (a fourth-century CE translation prepared from an earlier third-
century CE translation), possibly reflecting materials from Gandhāra (Demiéville 1924; Fuss-
man 1993, 63–66; Baums 2017, 33). There is evidence that there were other representatives of
this genre, now – except for fragmentary traces – lost (Skilling 1998, 91–92).
It is important to recall that Menander enjoyed a life off the page, as it were. He reigned
over the northwest of the subcontinent and the Punjab in 155–130 BCE, and Sāgala may
indeed have been his capital (Bopearachchi 1990, 1992). If Menander is remembered in Greek
sources primarily for the military victories which extended his power over northern India
(Tarn 1951, 225, 266), for Buddhists he was “one of theirs” (Lamotte 1988, 425). I am unsure
as to whether Richard Stoneman is right to say that “there is little doubt that the King himself
was a Buddhist” (Stoneman 2019, 365); the numismatic evidence is silent on the issue of
his personal rather than official convictions (Bopearachchi 1990, 48; Fussman 1993, 85–90).
Indic sources, however, do emphasize his fostering of Buddhist institutions (Neelis 2011, 105),
and in the Pāli tradition, as the ending of the Questions of Milinda would have it, Menander
converts to Buddhism (Neelis 2011, 105). Though we are not compelled to believe that, it is
hard to shake off an intriguing fact noticed very early by James Prinsep (see Tarn 1951, 266):
Menander’s funeral is described by Plutarch in a way that cannot but now be reminiscent of
the distribution of the relics of the Buddha.2
DOI: 10.4324/9781351030908-11 97
Sonam Kachru
Unlike Menander, we have no independent record of the Nāgasena of this dialogue; nor
need we believe that he ever existed as an individual, despite the philosopher Vasubandhu in
the late fourth and early fifth century CE treating him as a sthavira, an authoritative Buddhist
elder and forebear to several traditions (Kapstein 2001, 363; see also Rhys Davids 1891). For
our purposes, we can take it that Nāgasena lives, where he lives, as a dramatic character in
The Questions of Menander.
What of The Questions of Menander? Some might wish to treat the dialogues as invested in
documenting rather than imagining a world. Others may hope to read it like detectives, sleuth-
ing for inadvertent (if not deliberate) traces of a multicultural and multi-ethnic past worthy of
remembrance.3 Here, however, I ask the reader to practice a variety of forgetting. I want the
reader to treat both Menander and Nāgasena as dramatic characters, no more and no less. For-
getting can be productive. I hope that it will allow us to attend to the work as a literary achieve-
ment exploring, among other things, what it can be like to learn to think and feel as a Buddhist.
I will confine my attention to the Pāli work The Questions of Milinda, to which I will make
reference (for the most part) using or adapting the translations in the two volumes of Rhys
Davids (1890, 1894, hereafter RD1 or RD2, respectively) or the two volumes by I. B. Horner
(1963, 1964, hereafter H1 and H2, respectively).4
The Questions of Milinda and This Guide

Not being an expert in this work, I begin like Menander: as an enthusiastic outsider. And like
him, I find myself increasingly pulled into the textual universe this work holds inside of itself.
My goal is practical: how shall a student make use of this remarkable though forbidding work?
The Pāli work can seem heterogenous. Though it is presented as a single work consisting
in seven books, one may also treat it as a library of five works redacted possibly over centuries
(from the second–fifth centuries CE). On one view (following Von Hinüber 2000, 84), the first
three books of the work we now have were once a complete unit. In Book I, we get backstory,
featuring the past lives of the protagonists and social and cultural contexts for the conversations.
The dialogue proper takes place in Books II–III, which are arguably older still (cf. Norman
1983, 110–12), ending on the evening of the second day. Book III then concludes with a meeting
on the morning of the third day as the protagonists offer a retrospective appraisal of the conver-
sations, thus concluding this larger unit.
With Book IV, On Difficult Questions, the conversation and text resume on the morning
of the fourth day. From this point on, each book introduces a new theme and constitutes for
philologists a new historical addition. Book V introduces another day and a new topic, the
uniqueness of the Buddha. Book VI discusses laypeople, asceticism, and the way to nibbāna,
while Book VII takes a new route, which I call Thinking With Similes. There is then another
formal conclusion.
The following guide is divided into two parts, each part emphasizing distinct ways of
approaching and reading (sometimes overlapping parts of) the work. Part One deals with
Books II–VI, and Part Two with the literary unity of Books II–III, which I will refer to as *The
Questions of Menander to highlight a difference.5 Reading the work through from Books I–VII
can be a very different experience than reading Books II–III alone. The former is triumphal;
the latter tragic. Furthermore, reading The Questions of Milinda as a whole, from Books I–VII,
does not obviously present itself as a dramatic unity. Nevertheless, one may discern the fol-
lowing trajectory: Milinda travels a pathway from ignorance and skepticism into greater famil-
iarity with the world as Buddhists conceive it to be. Such a trajectory involves the “thickening”
98
The Milindapañha
of theory (discussed in what follows), one feature of which is a growing familiarity with the
imagery and values of a Buddhist textual universe.
Part One: The Triumph of Wisdom (Books II-VI)

As one goes farther and farther into the work, Milinda seems to gain knowledge not only about
Buddhist concepts but also of Buddhist texts in a wide variety of genres, exhibiting thereby con-
siderable skill as a reader. To begin with, however, the focus is on the acquisition of initially unfa-
miliar conceptual vocabularies and the exercise of skill in reasoning. In a meta-poetic moment
in the opening verses of the prologue (Trenckner 1880, 1), the discourse (kathā) of Nāgasena
is described as being aesthetically captivating (citra) through the use of illustrative examples
(opamma) and arguments – or, more generally, epistemic stances and strategies (naya). This is a
characterization with which the Buddha, later in the prologue, will agree (RD1, 6).
The Questions of Milinda is not merely a dictionary, conceptual lexicon, or encyclopedic
FAQ.6 To be sure, historians may use the work to elicit patterns of intellectual concern. For
example, why was it important to ask whether we may ascribe a variety of attention that is
appraisable according to ethical and epistemological norms to animals like sheep and goats?
Or other complex epistemic states, like wisdom? (II.1.6 in RD1, 51). Such use will not, how-
ever, help us address the distinctive analytic method enshrined in conversations employing
illustrative cases. Though much more work needs to be done to get anything approaching an
inventory of the diverse ways in which illustrations are used, I offer two ways of using illustra-
tions that strike me as being important.
Intelligibility and Acknowledgment

Milinda does not only ask questions of the form “What is [the definition of] X?”; he also asks
questions like “How could anyone not already committed to Buddhism believe that P?” Here
is an example (III.7.2; cf. RD1, 124): “Your people, Nāgasena, say that though a man should
live an evil life for a hundred years, at the moment of death, were a thought of the Buddha to
enter his mind, he will be reborn among the gods. This I don’t believe.’ ” Sometimes his ques-
tions take the form, “I cannot believe P.” Milinda’s questions can challenge the intelligibility
and not alone the factuality of certain positions. To not find a proposition intelligible might
mean not knowing what someone is saying, not knowing how they could consistently say it,
or not knowing why anyone would wish for it to be true.
In such cases, conversation often serves to reveal that there is a substantive principle, belief
in which blocks the intelligibility of the proposition in question. For example, the King once
elicits from Nāgasena that a monk ought to act so that “this dissatisfaction should cease, and
no dissatisfaction arise” (III.7.3; cf. RD1, 124–25). The King takes this to entail that some-
thing nonexistent, like what is to come, can figure as a reason, and so as a cause, for action. But
the King takes it on principle that nothing nonexistent can play that role.
Nāgasena uses an illustration to bring the King to see how he himself already acts as if he
does not always believe the principle. For example, the King’s prudential reliance on precau-
tionary measures – exemplified in fortification, defense, or the digging of wells – commit him
to endorsing sentences of the same logical shape as Nāgasena’s. A more famous example of
this occurs in the first conversation. Milinda finds incredible Nāgasena’s saying that his name
does not refer to any single entity – no single entity, as it turns out, from among the parts into
which he might be resolved, nor any combination of these parts. Milinda finds this incredible
99
Sonam Kachru
because he appears to believe that words must correspond to referents in the world if they are
to be meaningful and that metaphysical analysis can always supply us with the referents. In
reply, Nāgasena elicits from this, in combination with the King’s use of the word “chariot” and
his beliefs about the parts of a chariot, that the King will say about chariots what Nāgasena
has said about selves – it is not selflessness that is the source of unintelligibility, but the King’s
untested, inconsistently held, and therefore unwise reliance on a dubious principle of meaning
(RD1, 43–45).
Such illustrations provide conceptual and psychological gains. By proposing counterexam-
ples to a general principle, they suggest that principle’s possible falsity; and by getting someone
to acknowledge that their reliance on a principal has been superficial – and, in fact, inconsistent
with commitments tacitly held – an illustration may redirect the use of reason. Instead of being
used to challenge an interlocutor, reasons may be used to realize that what seems unfamiliar in
an interlocutor may have to do with an imperfect understanding that one has of oneself.
Reading the arguments in sequence, it is helpful to imagine them finding traction in the
mind of a dramatic character. By having his own (hitherto tacit) commitments rendered per-
spicuous with the help of the illustrations Nāgasena has drawn from ways of life familiar to
the King, Nāgasena’s reasons help Milinda become less hostile and more open: more open to
the intelligibility of vocabularies and beliefs hitherto alien to him, more open to formerly tacit
propositions to which he is really committed and of which he was hitherto ignorant, and thus,
more open to the claims of another’s beliefs on him as being possibilities for his own life.
Conceptual Innovation
Illustrations may also be used more narrowly: to modulate one’s concepts, for example. At one
point, for instance, the King effectively asks if it is worse to do something bad knowingly or
unknowingly (III.7.8; RD1, 129). Forensic practices of ascription of blame and punishment –
to which the King makes reference in the discussion – as well as received Buddhist concepts
of action underscoring the intent with which something is done7 might lead one to reason as
follows: if one punishes what is done with some requisite kind of awareness – and typically,
we do not punish what is done without some kind of awareness or intent behind it – it must be
the case that this is because we account it worse when someone does bad knowingly.
Look, however, at Nāgasena’s example: “But what do you think, O King: Take a heated
iron ball glowing with flame and heat – one person takes hold of it unknowingly; another takes
hold of it knowingly. Who is more severely burnt?” (cf. RD1, 129) What “worse” means is
here fixed by the example of “being burnt” – a relatively direct, deleterious and causal con-
sequence of the action for the agent. This is not quite the moral notion defined by reliance
on social practices of blame or punishment. Nāgasena has used a clever example to shift the
concept of action from a normative context to a (somewhat) more naturalized one. This latter
context also modifies what “knowingly” might mean: it need not mean prospective intent, but
rather something like being attentive with awareness of what one is doing, here defined by
knowledge of the nature of what one is acting with and what it can do.8
Nāgasena’s point, I take it, is not to answer an old question (and so yielding an answer
conflicting with orthodoxy). The point, instead, through the proliferation of cases, might be the
creation of more context-specific variegated concepts of action.9 Earlier, Nāgasena identifies
the following as the hallmark of reasoning as a paṇḍit (II.1.3; RD1, 46):
When paṇḍits talk a matter over with one another, there is a winding up, an unwind-
ing . . . distinctions are drawn, and contra-distinctions. . . .
100
The Milindapañha
Winding up and unwinding appear to be phrases in a process of getting an interlocutor to

clarify their position by virtue of its presuppositions and entailments (cf. Ganeri 2012, 20–25).
I take this to correspond to what is involved in clarifying the intelligibility of a position.
The careful drawing of distinctions and corresponding conceptual refinement (and innovation,
even) is, on my reading of The Questions of Milinda, essential to the wise exercise of reason.
The word “paṇḍita” – a word translated in my epigraph as someone “wise and learned” – com-
bines the virtues of long study, analytic intelligence, and action guiding wisdom (see Collins
2020, 5, 75–76): a paṇḍita/ā knows his or her way around difficult terrain, so to speak, by
exercising the virtues which the conversations of Nāgasena and Milinda exemplify, and which
a reader, perhaps, is intended to enact.
Logical Topography of the Doctrine

In Books IV–VI, the questions acquire more logical structure. The title of Book IV, “meṇḍaka-
pañha,” literally means “A question about a ram.” In the Ummagga Jātaka (Cowell and Rouse
1907, 177–78), the future Buddha solves a puzzle involving a strange fact, the friendship
between a dog and a ram observed by a King and encoded in an enigmatic question. By exten-
sion, the phrase “ram-question” can mean a riddle, or anything difficult to interpret without
wide experience or knowledge. Milinda does not ask riddles, exactly – he poses dilemmas
(ubhato-koṭika). For example: Did the Buddha not answer a certain kind of question because
he did not know the answer or because he wished to keep back certain truths? Either way,
inconsistency looms for a Buddhist (RD1, 204).
There is some precedent in the Pāli canon for the use of dilemmas, though interestingly,
they are not principally associated with either the Buddha or his students. The leader of the
Jains, Nigaṇṭha Nātaputta, is shown to be keen on having his followers use them in debate and
is portrayed as holding that being put a dilemma is like having something stuck in the throat
that one can neither swallow nor spit out (see Abhayarājakumāra Sutta, Majjhima Nikāya,
i.392; Ñāṇamoli and Bodhi 1995, 498). Unlike what this story suggests of the role of dilem-
mas, Milinda does not use them as a coercive tool in debate – or if he does, Nāgasena soon
repurposes them as tools of clarification.
A dilemma exposes a person to prima facie incompatible commitments. In previous books,
Nāgasena helps Milinda to put his own commitments into rational order; here, the dilemmas
force Nāgasena to contour doctrinal statements into a space of reasons and exhibit their con-
ceptual structure. We begin to see statements of the form: “The Buddha said P because of x, y,
z”; “His saying Q* is an instance of believing (the more general principle) Q”; “He said P in
the sense of P*, which can be made consistent with ~P”; and so on.
Take the following case (IV.5.16; RD2, 16–18). Milinda asks a question that reveals
him to be a meticulous reader of Buddhist texts in that he picks up on a contradiction in
the Buddha having said the following (which I offer in paraphrase) at different times and
in different texts:
1. In former births when human, I acquired the habit of nonharming.

2. When I was Lomasa Kassapa, I had hundreds of creatures killed while performing the
Vājapeyya ritual.
The Vājapeyya (in Sanskrit), an important and complex Vedic ritual, is sometimes (as here)
said to have involved the sacrifice of animals. Milinda is right to note that it is inconsistent
to say, “Considering all actions in all lives, I have never Ø-ed,” while also maintaining that
101
Sonam Kachru
“I have Ø-ed.” But there is no inconsistency, says Nāgasena, if the second sentence is read
as saying:
2.* When I was Lomasa Kassapa, I Ø-ed, being out of my mind because of lust.
Lomassa performed the ritual for a King as a way of securing the latter’s daughter in return.
The interesting thing here is not only that Nāgasena produces a sentence which avoids the
appearance of flat-out inconsistency; Nāgasena is also using the dilemma to reconstruct a
principle with far-reaching implications: What does it mean to have acted, but not as oneself?
The King does not understand this. And this may not be his fault. He goes on to provide a
reason for his incomprehension (RD2, 17). He argues that it is only possible to Ø out of some
strong motivating pro-attitude or condition – involving states like lust, anger, pride, avarice,
poverty, and so on – when these express the kind of person one is. On his view, it is a cruel
person, for example, who is said to act out of some such state as anger, as that is part of what
it means, on this account, to be a cruel person. And if that is true, the clause “by way of strong
emotion” or the like can never be exculpatory.
Nāgasena disagrees. One can Ø through a motivating force when it neither befits one nor
expresses who one is. Even if the motivating state is internal to a person, for example, it can
be a state from which the person is alienated. One needs a more refined criterion for internal
source-hood of an action, one which recognizes that desires which lie within the boundaries
of a person’s psychology may nonetheless be external to who that person is (for a contempo-
rary account, see Sripada 2016). Nāgasena does not only see the need for such an account;
he shows us the kind of questions to which such an account will lead and the implications of
such a view.
How can one know that alienation has taken place? One must learn to listen to another as
a literary critic might listen to a poem. Nāgasena cites a verse ascribed to Lomasa and says of
it that the sentiment it expresses is inconsistent with any desire on the future Buddha’s part
to engage in violence. It shows, instead, that as Lomasa, the future Buddha wished to act one
way, but that at the sight of his beloved “lust deranged his mind” (cf. RD2, 18). An unwilling
madman and addict through lust, as it were, he was constrained to act harmfully.
Nāgasena then goes on to explain the implicit analogy through an extended discussion on
the loss of prudential self-interest and moral compass in madness. If to act when alienated
from the motivational sources for the actions one commits is analogous to derangement, there
is reason to withhold ascribing responsibility to the agent. But then, readers might well ask: is
talk of responsibility ultimately tenable? According to Nāgasena, and other Buddhist texts, as
long as we do not acquire meta-cognitive control of our action-motivating states – such control
being a variety of achievement of regimens of training – we all, in some sense, act as if mad
(Collins 2015, 197–202). Recently, there has been growing interest among contemporary phi-
losophers in such issues. The consequences of this “as if ” for ethics, after all, are far-reaching.
Thickening Theory
The exploration of the logical topography can be an involved affair. But there are ways in
which things can become more involved that do not have to do with logical structure.
Consider the following questions. Do wondrous events take place at the shrines memorial-
izing the relics of the worthy dead of the Buddhists (IV.8.51; RD2, 174)? Can the dead derive
any benefit from offerings made in their name (IV.8.29; RD2, 151)? Did the Buddha think that
trees speak, or not (IV.3.20; RD2, 241)?
102
The Milindapañha
Such questions exhibit the “thickening” of theory. By that I mean to say that the concepts
they use and the facts they seek to bring into view have become more “world” involving, with
the content increasingly coming to reflect the meanings and values of participants in a par-
ticular social world. Such normatively inflected content evinces a particular emphasis on such
features as are capable of rendering intelligible or motivating action in the social world they
are suited to capture in thought (cf. Scanlon 2003). The thickening of theory in The Questions
of Milinda, beginning with Book IV, is reflected in the topics (increasingly having to do with
ritual as well as cosmology), as well as in the greater variety of sources to which Milinda and
Nāgasena appeal, spanning verses, customs, and prose narratives. For a good example of why
one needs to pay attention to thinking “thickly,” we can turn to Milinda’s questions about when
the Buddha was Vessantara, which we might term a “master class” in thinking with narrative
(IV.8.1, RD2, 114).
Vessantara’s birth story is a popular story of the Buddha’s penultimate life as a human in
which, as Vessantara, he practiced the ethical excellence of generosity to the extreme of giving
away his wife and children. Steven Collins describes it as a story not only extolling certain
virtues but thinking critically about them (Collins 2016, 1–2). As Milinda sees it, the ethical
difficulty presented by the narrative is magnified by the way the future Buddha gives away
his children in that Vessantara (a) gave them without their consent while (b) seeing the suffer-
ing this entailed for them and (c) remaining impassive in the face of the suffering his actions
manifestly brought about. How can goodness consist in this?
I will quote Milinda’s way of making this point at some length with only a few ellipses.
The appeal to narrative detail is not incidental to the argument. In accord with what I meant
previously by “thickening,” it serves as a paradigmatic way of bringing a moral fact into view
(cf. RD2, 115; H2, 96):
Nāgasena, the future Buddha did something hard to do when he gave away his own
beloved children to the Brahmin as slaves. And the second thing he did was harder
still: He bound his own children . . . young and tender though they were, with creep-
ing vines, and seeing them being dragged along by the Brahmin, the Brahmin flog-
ging them with vines, he could bear to look on. And third . . . When his boy came
back to him having struggled to free himself from his bonds, the future Buddha again
bound him with vine and again gave him away. Fourth . . . When the children, weep-
ing, cried: “Daddy, this ghoul is leading us away to eat us!” he did not comfort them
by saying “Don’t be afraid.” Fifth . . . The prince Jāli, weeping, fell at his feet and
begged him: “Enough, Daddy, be content with this. Get back Kaṇhājinā. I’ll go with
the ghoul. Let him eat me!” The future Buddha did not agree even to this.
Sixth . . . He showed no compassion as the boy Jāli lamented: “Have you a heart
of stone, Daddy, that you can look upon us, suffering, being led away by the ghoul
into the vast, wild forest and not hold us back?”
And the seventh thing he did was even harder still – when his children were led
away out of sight to something frightening and terrifying his heart did not split into a
hundred pieces, a thousand pieces!
Milinda’s appeal to the narrative is masterly. It places us in the narrative by inviting us to inhabit
alternative points of view. Reading this, we begin with a narrator’s third-personal point of view
on an event thinly described. We are then invited to contrast the perspective of Vessantara’s boy
with that of the future Buddha. There is also the invoked, but not explored, mute and terrified
perspective of Jāli’s little sister, Kaṇhājinā. We look on the future Buddha as his children see
103
Sonam Kachru
him, and with them, perhaps, interrogate his perspective on this horror. Unlike the boy, he is
silent. Unlike the girl, he appears initially to be silent and indifferent, without compassion or
moral concern. As we watch, at last, the children disappear from his point of view, what might
we have felt were we in his place? What do we feel? What ought he (or we) to feel?
I. B. Horner has Milinda grant, at the end, that Vessantara was “in great distress and terror”
seeing his children being led out of sight (H2, 97). I am not convinced. The text is unclear here,
perhaps deliberately so. The frightfulness and terribleness evoked by the sentence describing,
from his point of view, the disappearance of his children – “When his children were led away
out of sight” – might more naturally be read as qualifying, instead, the murky destiny to which
the children are being led, or, though this is strained, the ghoulish Brahmin leading them away;
perhaps the ambiguity reflects the limit of the future Buddha’s sight and his growing horror
as they vanish (RD2, 115).10 We do know that Vessantara said and did nothing overtly and,
perhaps, that he ought to be feeling something; else, Milinda’s narrative argument implicitly
asks what distinguishes the future Buddha’s silence from that of the ghoulish Brahmin who
leads the children away. Doesn’t Vessantara also stand to gain from this travesty? “What,”
Milinda asks, “has a man desiring merit to do with bringing anguish to others? Should he not
have given himself away?” (cf. H2, 97).
After long discussion, including an argument we shall briefly consider in what follows,
Nāgasena, too, appeals to narrative. He does this by invoking Vessantara’s experience after the
terrible deed has happened:
Vessantara, having given away his wife and children, entered the leaf hut and lay
down. Powerful grief rose in him, anguished as he was because of his exceeding love
for them. His heart grew hot, and his breath, hot, could not pass through his nose.
He breathed in and out of his mouth. The tears which fell were drops of blood that
poured from his eyes.
(cf. H2, 105; RD2, 126; cf. Cone and Gombrich 2010, 55)
For Nāgasena, the grief is caused by his love for his children, but it may not be directed at the
suffering of his children and his wife alone. It may signify a revolt at what the supererogatory
ideal and excellence of giving requires of him; it may also signify embodiment of the cost of
excellence (RD2, 126; see also Heim 2003, 540, 545). Narrative, as appealed to by Milinda
and Nāgasena, provokes but does not settle interpretation.
Before Nāgasena turns to narrative, he offers some striking ways of conceptualizing the
moral character of the life of a being directed at awakening. He suggests thinking of such a
being’s life as a public shareable good – a public feast, for example. He also goes on to offer
some thoughts on awakening as a goal, readjusting the scope and weight of traditional moral
goods. Each of these arguments will require careful unpacking, as Nāgasena himself antici-
pates. He imagines the moral contestability of Vessantara’s actions and their ambiguous and
transgenerational emotional impact as being bound up together (Heim 2003, 546). In fact,
when Nāgasena speaks to Milinda of how successive tradition gradually makes it possible for
such an event to reach “our meeting here today so that, defaming and disparaging that gift,
we question whether it were well given or ill given” (H2, 97), I like to think that he also has
in mind future readers of the dialogue, identifying moral deliberation rather than definitive
opinion as the vital part of tradition (Heim 2003, 546). While there is much to be said about
this, I wish to note just this. Nāgasena’s decision to contest Milinda’s arguments by appeal-
ing to the point of view and the reality of psychological cost is telling, for Nāgasena thereby
directs us to the following questions. Why would one wish to do something which wrenches
104
The Milindapañha
one’s entire being this way? What is the world (and its moral structure) like that realizing our
ideals must exact such a price?
The King assumes a kind of indifference on the part of the future Buddha and conceives
of such indifference as part of the pursuit of this excellence. Perhaps this is because he can-
not imagine a world in which doing good and feeling good do not align. Is he thinking of our
world? Here, Nāgasena is effectively telling us that the story of Vessantara presents us with a
world wherein, to adapt a phrase from John Rawls (and Isaiah Berlin), there can be no moral
life or world without loss (cf. Pogge 2007, 187). Nāgasena is instructing Milinda in tragic
choice. We cannot pursue or maintain all that matters to us, and doing what one has overrid-
ing reason to do is consistent with ruinous psychological and physical cost. Were one to push
back on either of these points, one would inhabit a different moral universe from that in which
Vessantara and Nāgasena live, a point that such appeals to narrative, with their richly felt per-
spectives, help make apparent.
Part Two: The Tragedy of Wisdom (Books II – III)

The Questions of Milinda ends in Book VII with the King requesting to be ordained as a monk,
though not before building a monastery for Nāgasena. This secures the success of Buddhism in
the marketplace of ideas, an image invoked at the very beginning of the prologue, where we are
told how Sāgala, rich in markets, features innumerable varieties of well-displayed goods, along
with the competing cries of welcome from teachers of every philosophical tradition and many
ascetic practices (H1, 2). Discourse is for “the stability of true Dhamma” (cf. H2, 305) – so says
the first of the last verses of the text. On this view, it makes sense that the conversations recorded
in this text be seen as a means to shore up the faith against such times as when “the tradition of
the Buddha is in danger of crumbling away” (to use, once more, the language of the prologue)
(RD1, 14).
But that is not the only story we might tell. At the end of *The Questions of Menander
– or at the end of what is now Book III of the Questions of Milinda11 – the camera does not,
so to speak, pan back far enough to take in the historical vicissitudes of the Buddhist tradi-
tion. Instead, we find two men who have grown close during their exchange of questions and
answers, their virtues and excellences beginning to mirror one another’s. They are said to spend
the night examining themselves, rehearsing their conversation. The next morning, the King is
keen to convey that the felicity of mind (somanassa) with which he passed the night did not
derive from the fact of having questioned the monk; rather, there are virtues beyond victory
and defeat: “I was debating with myself as to whether I had asked felicitously, and whether
I had been rightly answered” (cf. RD1, 135). Nāgasena concurs. To speak well together, and to
examine oneself after, is what matters. On this ending, Milinda does not convert to Buddhism.
If he changes at all, he is changed by coming into a cooperative relationship with Nāgasena
based on the felicity of mind philosophy offers when exercised as a way of life.12
A More Intimate Drama

Milinda’s conversion to the art of wise, cooperative reasoning is the more intimate drama and
the text’s minor masterpiece. More specifically, Books II–III offer an intriguing response to the
following narrative problem: how shall thinking be represented as a dramatic activity?
*The Questions of Menander creates dramatic movement by invoking transitions in what
I shall call the context, the mode, and the character of thought. The work leads its readers from
an initially public and intrinsically confrontational mode of thought concerning the beliefs of
105
Sonam Kachru
individuals to a mode of thought that is private, dialogical and cooperative, and increasingly
impersonal. Here I mean two things by “impersonal”: thinking, as the text has it, involves
being able to master a vocabulary and way of framing our experience on which (a) there is no
reference to selves or what belongs to a self; and (b) what is offered can be used to bring any
individual under a description and not only biographically individuated persons.
At the end of Book III, Milinda wishes to lavish gifts on Nāgasena, but Nāgasena refuses.
The King reminds Nāgasena that they are constrained by public opinion. Milinda and Nāgasena
have been speaking oriented by their desire to arrive at truth. They have done so in the King’s
chambers, beyond the reach of the public. But the public still exists as a horizon of possible
relevance, and as an available way of reframing the conversation – as an exchange of goods,
for example, or as earlier, at the very beginning of the text, when the conversation began in
public and was framed as a public contest and trial.
It is only after Nāgasena successfully defeats Milinda in a public debate that the conversa-
tion changes (see Vasil’kov 1993). The next day, they choose to speak in private and not public;
Nāgasena insists that they continue to speak only as pandits and not as Kings.13 Among other
things, this means the following: by holding different opinions on a topic one may cooperate
with and not only challenge another; one may reason not only to win an argument but to get at
truth; an argument need not entail contestation; and a new model of arguments exchanged in
conversation may involve a process of trying to expose error in one another’s positions, with
such exposure being productive rather than threatening and to be met with acknowledgment
and not anger (RD1, 46). Outside the constraints of unreasonable power, conversation, analy-
sis, and the mutual exploration of possible positions may lead to truth.
Such cooperation need not eschew hierarchy. As Milinda puts it at the end, they have rea-
soned together as master and student to get at the truth (RD1, 88). Where may such intimate
conversations of truth unconstrained by power and fear be held? At the beginning, Menander
suggests speaking as one might with a brother, a novice, a lay disciple, or a servant (RD1, 88). It
is, perhaps, into some such intimate and edifying theater that the text invites the reader, as well.
The mode of conversation shifts, as does the topic. We move from discursive frames, like
our conventional ways of speaking, which presuppose selves, to frames which do not. More
generally, the conversation involves Milinda making himself more at home with the existential
context for our lives provided by Buddhism in its understanding of rebirth, saṃsāra, and time
and making himself at home with the virtues emphasized by Buddhists as a means to freedom
within this context. If the going gets tough here, and the questions too numerous and too dis-
parate, one may hold to the following dramatic thread: Milinda’s continuing trouble with both
aspects of impersonalism.
Consider his question about the self as a principle of cognizance or awareness, for which he
uses the Pāli word vedagu (RD1, 86).14 Apparently unconvinced by the impersonal vocabulary
for experience Nāgasena has been laying out and in an attempt to motivate an idea of a self
as perceiver, Milinda helps himself to a striking illustration: Imagine a person in a house with
many windows. The self might be like such a person: as one might see using the windows, the
self sees, or hears, and the like, with the relevant sense capacity.
One might think that Milinda has learned something from Nāgasena about the use of illus-
trations. But perhaps he has not learned the lesson perfectly. Nāgasena uses Milinda’s illus-
tration to highlight an important disanalogy with the case of sensory experience. Experience
might be analogous to a person in a house with many windows were something sensible an
object like the King’s manservant, for example, something capable of coming in and out of
view: something, that is, that might be (a) reidentified over time (b) through any window
equally and (c) available to us to be aware of wherever they (and we) are located. But this,
106
The Milindapañha
Nāgasena believes, is not so. Sensible things are too tightly individuated to their senses to
motivate the sense of Milinda’s autonomous perceiver and the content-indifferent windows.
This argument needs more careful study. It is strikingly reminiscent of Socrates’ themati-
cally continuous discussion with Theaetetus on the difference it makes to think of the soul as
seeing with in contrast to seeing through the senses (Burnyeat 1976).15 Whether in Greek or
Pāli, impersonalism can impose changes to the way we speak and think about ourselves. These
changes are difficult for Milinda to fully accept. As he effectively admits in his next question
(RD1, 89–92), what he is having difficulty with may be put this way. He is accustomed to
speaking and thinking about experience in terms of sentences structured by appeal to actions
and their agents. Even when he withholds speaking of a self, he still thinks of events as actions
brought about by agents, albeit insentient ones. Nāgasena, in attempting to teach him Buddhist
thought (Abhidhamma), is trying to teach him an entirely new conceptual grammar, featuring
causal and structural explanations rather than agential ones.
There is another source for the difficulty of impersonalism based on a more familiar sense
of “personal.” As he is learning to think in this new impersonal idiom for exploring experi-
ence, Milinda is not content to learn what may hold true of anyone’s experience; he wants to
know about Nāgasena in particular. Why did Nāgasena, rather than anyone else, leave home?
(II.1.5, RD1, 49). It is a pattern of concern that does not leave him. Later, Milinda happens to
raise again a question which he had asked sometime earlier in the conversation. Tellingly, it
is a highly personal question: “Will you, Nāgasena, be born again?” Milinda waits to ask his
question again just after a lengthy exposition of an impersonal (in the sense of “without self ”)
account of rebirth and continuity. Nāgasena has to rebuke him into dropping the matter for
good (II.2.7; see I, 6; RD1, 76–77; 50).
The incident is, to me, suggestive. Milinda asks some fifteen questions before raising again
the one that so obviously matters to him. As with music, the silences between questions, and
what is not said in them, can be as significant as what is said.
The Lion in a Gilded Cage

Shifts in the context, mode, and character of conversation, I have suggested, are dramatically
significant. They allow us to track what it is like to begin thinking with – and as – a Buddhist
about the world as Buddhism describes it. A later literary critic on the subcontinent might ask:
what aesthetic savor does such drama sustain? A variety of tragic bittersweetness, I would ven-
ture to say. Unlike The Questions of Milinda, *The Questions of Menander is not triumphal.
On the night before the formal conclusion of their private dialogue, when Nāgasena and
Milinda agree in their assessment of the virtues in play in philosophical conversation and
acknowledge the need to take into account the wider social world to which they must eventu-
ally return, the narrative has Milinda say this (3.7.18; cf. RD1, 135; see also Sick 2007, 276):
Just as . . . a lion, the king of beasts, trapped in a golden cage is ever staring outward,
I . . . too, though I dwell on the inside, ever stare outward.
Philosophy need not always result in freedom. For some, it can result in a sense of imprisonment.
Let me elucidate the poetic argument that is suggested by the image of the lion (also found in the
Chinese work). An inscription in Kharoṣṭhi found in Tillya Tepe, Afghanistan, refers to “the lion who
has driven away fear,” perhaps meaning the Buddha and his first sermon (see DeCaroli 2015, 20).
One might also expect it to be a symbol of royalty, as it was for the Mauryan Aśoka, as well as for
Menander’s descendants: the seated lion is found on the Indian-standard square coins of the apparently
107
Sonam Kachru
Buddhist Menander II (90–85 BCE).16 In Pāli, however, the King’s first glimpse of Nāgasena has the
latter, quixotically enough for a bald monk, appear “like a shaggy lion who knows no fear” (RD1, 39).
I recommend reading Milinda’s image as an allusion to a popular Buddhist text which
exhorts the celibate ascetic to “not tremble, as a lion does not tremble at sounds, not caught up
with others, as the wind is not caught in a net” (verse 71, Suttanipāta, in Norman 2001, 9; see
also Salomon 2000, 108; 38). The text continues: “Living victorious, having overcome like
a strong-toothed lion, the king of beasts, one should resort to secluded dwellings” (verse 72,
Suttanipāta, in Norman 2001, 9). Given this background, Milinda’s use of the image is surely
ironic: the lion, unlike the wind, now finds himself caged “inside” his own quotidian life.
Redescribing oneself with such commitments one has acquired through reasoning in conver-
sation, but without a corresponding change in one’s way of life, is not straightforward. If it is
philosophical reasoning that lets Milinda see through the bars of the cage to a broader vista of
possibilities, the same reasons now led him to experience his life as a cage in the first place.
When they begin their private dialogue, Nāgasena effectively explains to Milinda the dif-
ference between reasons and causes. A monk is said to leave home with nibbāna as the goal
or reason, though individual monks have any number of causes that have led them to leave
home, including, as Nāgasena wryly mentions, the tyranny of kings (RD1, 49–50). At the end
of their conversation, Milinda cites political necessities as holding him back. Is this to cite a
mere cause? Could one ever have normatively weighted reasons not to take up the way of life
Nāgasena has worked so hard to render intelligible? If Milinda finds himself locked into a life
for which he knows he has no justification, that might suggest a quiet tragedy.
Conclusion
“[T]here are . . . passages in the early part of the Questions of King Milinda,” said H. H. Price,
once Wykeham Professor of Logic in Oxford, “which have a very modern ring, and might
almost have been written in Cambridge in the 1920s” (Price 1955, 229). For the most part, the
attention of philosophers continues to be directed only at early moments in The Questions of
Milinda, employing the same criterion of relevance: what topics might appear in the leading
journals of the discipline today? This is too restrictive. There is much cutting-edge material to
be found elsewhere in the text. But it is perhaps better to do away with the parochial criterion
of relevance altogether and read more of the text and with greater sensitivity to the way in
which it directs us to read it.
If we incline to Milinda’s point of view, the work might seem to offer us a path from doubt
through reason farther and farther into the imaginative world of Buddhism; from Nāgasena’s
(rather less variable) vantage point, the work may be read as the enactment of faith through
the victory of reason in service of Buddhism. But there is a danger in identifying too much
with either character at the expense of the other. While the answers to difficult questions are
important, so too is the art of questioning, which is why the Buddha in the prologue, predicting
the conversations to come in the future, says that both Milinda and Nāgasena shall explain his
teaching, unraveling and disentangling its difficulties through questions and illustrations
(cf. RD1, 6). Another way to put this is to say that the true protagonist might be wisdom itself,
and that the true stage for this dialogue might lie off the page.
In a work dedicated to the skillful use of similes and images, it is fitting to consider the
image with which the work concludes (cf. H2, 305; see also Drewes 2007, 105–6):
Therefore, let one who is wise,

beholding his own good,
108
The Milindapañha
reverence and honor those who have wisdom

As a memorial ought to be honored.
Notice the invitational quality: a reader is expected to exemplify wisdom in the sense of pruden-
tial reason, enough to keep in view wisdom – now meaning something like analytic skill – as
something worthy of regard as a memorial housing the Buddha’s relics is worthy of regard, and
for the same reason: honoring such a site brings merit. A memorial (cetiya) housing relics of the
Buddha or other venerated teachers, whether in the literary universe of Pāli or Sanskrit, can trig-
ger strong emotions and make available transformative experiences. From about the first century
CE, by analogy, books containing scripture could be such memorials (Drewes 2007); any body
(khandho), in fact, as the conclusion of The Questions of Milinda has it, might serve in this
capacity, as long as it embodies wisdom. Thus, we are to treat the fictional characters who are
wise – pandits embodying wisdom ( paññā) – and, presumably, the text with its wise depiction
of their wisdom-exemplifying conversation, as being effectively memorial shrines. To honor and
reverence a work this way implies a certain way of reading. I leave it to the reader to imagine
ways in which one might read this text worshipfully, another possible accent of the verbal root
pūj-, along with “honoring” and “reverencing.” To me, being asked to read a text as if venerating
a memorial shrine suggests, at the very least, that the reader is being encouraged to enter into
an intimate, affectively charged (and possibly long-term) transformative relationship with it.17
One might here recall the conclusion of *The Questions of Menander and the suggestion
that analytic conversation must be accompanied by self-examination. To echo Nāgasena’s
characterization of the dialogue of pandits, perhaps we are to read forward and backward,
occupying now the role of Milinda/Menander, now that of Nāgasena, alternatively touching
freedom and seeing our own gilded cages through which possible forms of life beckon.
Notes
1 The work is included within the Pāli canon in Burma as the eighteenth book of the Khuddaka Nikāya,
the latter being a part of the Group of Scriptures (Sutta-piṭaka) and rendered either as The Short Group
of [Canonical] Texts, or The Group of Short [Canonical] Texts; neither name, as Steven Collins wryly
remarks, is accurate (Collins 2020, 2). In the introduction to the standard edition of the work, V. Trenckner
records that in Sri Lanka, it is sometimes known as Milindapañho, which uses the masculine ending as
collective singular; manuscripts typically refer to it as Milindapañham (using the neuter collective singu-
lar) or Milindapañhā, which is either plural or a collective singular feminine ending (Trenckner 1880, vi).
2 See Hall (2015, 200), citing Plutarch, Moralia, 21d-2c, who has Menander incorrectly be a King of
Bactria (Stoneman 2019, 365); though see also Fussman (1993, 65) and Lamotte (1988, 421), quoted
in Neelis (2011, 105, n117). On the connection with Hellenistic hero cults and not alone Buddhist
relic cults, see Neelis (2011, 105), citing Burkert (1985, 203).
3 For students interested in such questions, I recommend now beginning with Chapters 12–13 in Stone-
man (2019), supplemented with Baums (2017). This is not, however, to endorse every claim in Stone-
man (2019). For example, I do not think it is right to say that The Questions of Milinda “contains
nothing that would remind one of Greek philosophy” (Stoneman 2019, 366).
4 On translations into other European languages, see de Jong (1996, 382–83). Excitingly, a much-
needed new translation into English is being prepared by Maria Heim for the Murty Classical Library
of India and Harvard University Press.
5 I also wish with this title to signal the likely provenance for the latter: this stretch of text – where the
Pāli and the aforementioned Chinese work, The Sūtra of the Monk Nāgasena, overlap most – may
well reflect materials held in common the second century C.E. (see Baums 2017, 33; possibly earlier,
according to Nolot 1995, 9).
6 It is not merely, that is, a Buddhist iteration of the South Asian narrative penchant – one also found
in late-antique Mediterranean genres (erotapokriseis or problemata; see Volgers and Zamagni 2004)
– for organizing knowledge through questions and answers.
109
Sonam Kachru
7 In standard historiography, Jains and Buddhists present their views of action as offering contrast-
ing perspectives of the nature of action and the part of action which ought to receive emphasis and
weight. It is claimed by Buddhists that Jains emphasize what is done (by overt action), sometimes
even at the expense of awareness or intent, while it is claimed by Jains that Buddhists downplay the
overt action and what is done thereby. For rival perspectives on the same examples, see Jacobi (1895,
414–15); Pruden (1991, 650–51); the issue is raised as early as the Upāli Sutta (Ñāṇamoli and Bodhi
1995, 477–93).
8 Compare the example of the iron ball in the prudential advice offered in verse 371 of the Dhamma-
pada (Norman 2000, 53).
9 This kind of conceptual variegation is only strengthened when one reads over the several cases in
the work. See Rhys Davids (1890, Vol. I., 224; 1894, Vol. II., 78). See also McDermott (1977). My
account is informed by the methodological commitments to context-sensitivity and resistance to gen-
eralization which Maria Heim has reconstructed for Buddhaghosa in Heim (2018, especially pp. 50,
149, 167).
10 The entire description of the seventh difficult thing to do requires careful analysis, well beyond the
scope of our discussion. I should say that I am inclined to follow Rhys Davids and not Horner, and
I am grateful to Maria Heim for helping me think through the possible readings, including those of
the commentary.
11 A terminological reminder: Earlier, I called Books II–III *The Questions of Menander to highlight a
philological and literary point. This stretch of text antedates the Pāli work, as it is found in different
languages, and it is literarily distinctive. As we consider this part of the work, I shall use the name
“Milinda” to say things that are true of the character of the King we find in the Pāli work as well as
in the The Sūtra of the Monk Nāgasena, the work surviving in Chinese which overlaps with Books
II–III.
12 A model of such friendship (and this style of conversation) may be found in the conversation of
Sāriputta and Puṇṇa Mantāṇiputta, who speak employing similes and as “two great beings rejoining
in each other’s good words” (Ñāṇamoli and Bodhi 1995, §14, 245).
13 There is canonical precedent for the implied contrast between truth and power as orienting values in
conversation. In the Upāli Sutta (Majjhima Nikāya i.376), the Buddha asks the householder Upāli,
someone keen on adversarial debate, whether or not he “will debate on the basis of truth,” given
which “[they] might have some conversation on this” (Ñāṇamoli and Bodhi 1995, 481). The sense
that some topics are best treated in private conversation and not public debate is found in the way that
the narrative changes in the Bṛhadāraṇyaka-Upaniṣad 3.2.13, when Yājñavalkya takes Ārtabhāga
aside to discuss karma. On thinking in conversation more generally, see the papers in Black and Pat-
ton (2016).
14 The word has an interesting history. In Brahmanical usage, it meant “one who knows the Vedas”;
in Buddhist usage, it was reinterpreted to mean “One who has achieved knowledge of release from
samsara.” See Norman (2012, 198).
15 Though Menander in Pāli uses the instrumental case to make his point – the self is “the living prin-
ciple which sees hue with the visual sense ( jīvo cakkhunā rūpaṃ passata . . .)” – his appeal to the
example of windows, and his subsequent explanation, suggest instead that he agrees with Socrates’
case for the soul seeing “through” the senses (see Burnyeat 1976, 32–33, 37). Nāgasena, meanwhile,
adopts a position reminiscent of the Heraclitean flux of unowned sensations of transient proper sen-
sibles against which Socrates sets himself.
16 Whitehead (1914, 62). See also the maneless lions of Pantaleon (Whitehead 1914, 16).
17 My thanks to Ishan Chakrabarti for conversation on this. Caley Smith has drawn my attention in
conversation to a similarly metatextual moment in the Bhagavad Gītā (verse 18.70) when study or
reading of certain texts can take the form of an epistemologically transformative variety of (inner)
ritual, a jñāna-yajña, and result in new forms of relations to what is real. See Flood and Martin (2012,
190).
Bibliography
Baums, Stefan. 2017. “Greek or Indian? The Questions of Menander and Onomastic Patterns in Early
Gandhāra.” In Buddhism and Gandhara: An Archaeology of Museum Collections, edited by Himan-
shu Prabha Ray, 33–47. London: Routledge.
110
The Milindapañha
Black, Brian, and Laurie Patton. 2016. Dialogue in Early South Asian Religions. London and New York:
Routledge.
Bopearachchi, Osmund. 1990. “Ménandre Sôter, un roi Indo-grec. Observations chronologiques et
géographiques.” Studia Iranica 19: 39–85.
———. 1992. “Was Sagala Menander’s Capital?” In South Asian Archaeology 1989: Papers from the
Tenth International Conference of South Asian Archaeologists in Western Europe, Musée national
des Arts asiatiques – Guimet, Paris, 3–7 July 1989, edited by Catherine Jarrige, 327–37. Madison,
WI: Prehistory Press.
Burkert, Walter. 1985. Greek Religion. Trans. John Raffan. Cambridge, MA: Harvard University Press.
Burnyeat, Myles. 1976. “Plato on the Grammar of Perceiving.” The Classical Quarterly 26 (1): 29–51.
Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism. Cambridge:
———. 2015. “Madness and Possession in Pāli Texts.” Buddhist Studies Review 31 (2): 195–214.
———, ed. 2016. Readings of the Vessantara Jātaka. New York: Columbia University Press.
———. 2020. Wisdom as a Way of Life: Theravāda Buddhism Reimagined. Edited by Justin McDaniel.
New York: Columbia University Press.
Cone, Margaret, and Richard F. Gombrich. 2010. The Perfect Generosity of Prince Vessantara: A Bud-
dhist Epic. Bristol: Pali Text Society.
Cowell, E. B., and W. H. D. Rouse. 1907. The Jataka, or Stories of the Buddhas Former Births, vol. 4.
Cambridge: Cambridge University Press.
DeCaroli, Robert. 2015. Image Problems: The Origin and Development of the Buddha’s Image in Early
South Asia. Seattle and London: University of Washington Press.
de Jong, J. W. 1996. “Review of Entrietens de Milinda et Nāgasena. Traduit du p‫‏‬âli, présenté et annoté.”
Bulletin of the School of Oriental and African Studies 59 (2): 382–83.
Demiéville, Paul. 1924. “Les versions chinoises du Milindapañha.” Bulletin de l’École Française
d’Extrême Orient 24: 1–26.
Drewes, David. 2007. “Revisiting the phrase ‘sa pṛthvīpradeśaś caityabhūto bhavet’ and the Mahāyāna
cult of the book.” Indo-Iranian Journal 50 (2): 101–43.
Eliot, T. S. 1919. “The Preacher as Artist.” Athenæum 4674: 1252–53.
Flood, Gavin, and Charles Martin. 2012. The Bhagavad Gita. New York and London: W. W. Norton and
Company.
Fussman, Gérard. 1993. “L’Indo-grec Ménandre ou Paul Demiéville revisité.” Journal Asiatique 231:
61–138.
Ganeri, Jonardon. 2012. Identity As Reasoned Choice: A South Asian Perspective on the Reach and
Resources of Public and Practical Reason in Shaping Individual Identities. New York: Continuum
International.
Hall, Edith. 2015. Introducing The Ancient Greeks: From Bronze Age Seafarers to Navigators of the
Western Mind. New York: W. W. Norton and Company.
Heim, Maria. 2003. “The Aesthetics of Excess.” Journal of the American Academy of Religion 71 (3):
531–54.
———. 2018. Voice of the Buddha: Buddhaghosa On The Immeasurable Words. New York: Oxford
University Press.
Horner, I. B. 1963. Milinda’s Questions: Volume I. London: Luzac and Company.
———. 1964. Milinda’s Questions: Volume II. London: Luzac and Company.
Jacobi, Hermann. 1895. Jaina Sūtras: Part II. Oxford: Clarendon Press.
Kapstein, Matthew T. 2001. Reason’s Traces: Identity and Interpretation in Indian and Tibetan Buddhist
Thought. Boston: Wisdom Publications.
Lamotte, Étienne. 1988. History of Indian Buddhism: From the origins to the Śaka era. Translated by
Sara Webb-Boin. Publications de l’Institut orientaliste de Louvain, 36. Louvain-la-Neuve: Université
catholique de Louvain, Institut orientaliste.
McDermott, James P. 1977. “Kamma in the Milindapañha.” Journal of the American Oriental Society
97 (4): 460–68.
Ñāṇamoli, Bhikkhu, and Bhikkhu Bodhi, trans. 1995. Majjhima Nikāya. Boston: Wisdom Publications.
Neelis, Jason. 2011. Early Buddhist Transmission and Trade Networks: Mobility and Exchange Within
and Beyond the Northwestern Borderlands of South Asia. Leiden: Brill.
Nolot, Édith, trans. 1995. Entrietens de Milinda et Nāgasena. Traduit du pâli, présenté et annoté. Paris:
Galimard.
111
Sonam Kachru
Norman, K. R. 1983. Pāli Literature: Including the Canonical Literature in Prakrit and Sanskrit of All
the Hīnayāna Schools of Buddhism. Wiesbaden: Otto Harrassowitz.
———. 2000. The Word of the Doctrine (Dhammapada). Oxford: Pali Text Society.
———. 2001. The Group of Discourses (Sutta-nipāta). Oxford: Pali Text Society.
———. 2012. “Theravāda Buddhism and Brahmanical Hinduism: Brahmanical Terms in a Buddhist
Guise.” The Buddhist Forum 11: 193–200.
Pogge, Thomas. 2007. John Rawls: His Life and Theory of Justice. Translated by Michelle Kosch. New
Price, H. H. 1955. “The Present Relations Between Eastern and Western Philosophy.” The Hibbert Jour-
nal 53: 222–29.
Pruden, Leo M. 1991. Abhidharmakosabhasyam of Vasubandhu. Volume II. Berkeley: Asian Humanities
Press.
Rhys Davids, T. W. 1890. The Questions of King Milinda. Translated from the Pāli. Part I. Oxford: Clar-
endon Press.
———. 1891. “Nāgasena.” The Journal of the Royal Asiatic Society of Great Britain and Ireland (July):
476–78.
———. 1894. The Questions of King Milinda. Translated from the Pāli. Part II. Oxford: Clarendon Press.
Salomon, Richard. 2000. A Gāndhārī Version of the Rhinoceros Sūtra: British Library Kharoṣṭhī Frag-
ment 5B. Seattle and London: University of Washington Press.
———. 2018. The Buddhist Literature of Ancient Gandhāra: An Introduction with Selected Translations.
Somerville, MA: Wisdom Publications.
Scanlon, T. M. 2003. “Thickness and Theory.” Journal of Philosophy 100 (6): 275–87.
Sick, David H. 2007. “When Socrates Met The Buddha: Greek and Indian Dialectic in Hellenistic Bactria
and India.” Journal of the Royal Asiatic Society 17 (3): 253–78.
Skilling, Peter. 1998. “A Note on King Milinda in the Abhidharmakośabhāṣya.” Journal of the Pāli Text
Society 24: 81–101.
Sripada, Chandra. 2016. “Self-Expression: A Deep Self Theory of Moral Responsibility.” Philosophical
Studies 173 (5): 1203–32.
Stoneman, Richard. 2019. The Greek Experience of India. From Alexander to the Indo-Greeks. Princeton
and Oxford: Princeton University Press.
Tarn, W. W. 1951. The Greeks in Bactria and India. Cambridge: Cambridge University Press.
Trenckner, V. 1880. Milindapañha. London: Pali Text Society.
Vasil’kov, Y. 1993. “Did East and West Really Meet in Milinda’s Questions?” Petersburg Journal of
Cultural Studies 1: 62–77.
Volgers, Annelie, and Claudio Zamagni, eds. 2004. Erotapokriseis: Early Christian Question and Answer
Literature in Context. Leuven: Peeters.
Von Hinüber, Oskar. 2000. A Handbook of Pāli Literature. Berlin and New York: Walter de Gruyter.
Whitehead, R. B. 1914. Catalogue of Coins in the Panjab Museum, Lahore. Vol. I: Indo-Greek Coins.
Oxford: Clarendon Press.
112
7
CĀTTAN‒ ĀR
Poet-Philosopher in Tamiḻ
Anne E. Monius1
Who Is Cāttan‒ār?
Since the rediscovery and first publication of his only known work, The Renunciation of
Maṇimēkalai (Maṇimēkalai Tuṟavu),2 in 1898, Cāttaṉār3 has continued to generate intense
debate among scholars of South Indian Tamiḻ literature, history, and religion. Who was he?
When and where exactly did he live? What manner of Buddhist was he, and what texts did he
know and draw upon in composing this long poetic narrative about a beautiful young cour-
tesan, Maṇimēkalai, who becomes a Buddhist nun? Embedded in the story are three densely
worded philosophical chapters in verse: one chapter is devoted to ten non-Buddhist teachings,
another is on Buddhist inferential reasoning, and the final chapter provides a detailed account
of interdependent origination. What purpose do these chapters serve? Do they reveal anything
of the author’s own particularly Buddhist inclinations in a Tamiḻ-speaking region of southern
India where the history of Buddhist thought and practice remains obscure?
In the century or more since The Renunciation of Maṇimēkalai became widely known as
a classic work of Tamiḻ literature and one of the earliest long narratives composed in Tamiḻ,
knowledge of its author, Cāttaṉār, has advanced little. All that can be said of him with any
certainty derives from the sole text that he has left to us, alongside several brief mentions
in another long Tamiḻ narrative, The Story of the Anklet (Cilappatikāram), with which The
Renunciation of Maṇimēkalai has long been aligned. As scholar Kamil V. Zvelebil aptly sum-
marizes, “the only certain thing” that can be said of Cāttaṉār “is that he was a convinced and
well-educated Buddhist and a great poet” (1995, 128).
Cāttaṉār’s name first appears in the preface to The Story of the Anklet, the earliest long-
form narrative in Tamiḻ, that tells the story of a young married couple, Kōvalaṉ and Kaṇṇaki.
Kaṇṇaki weathers adultery, hardship, and exhausting travel, and her husband is unjustly put
to death by a local king as he attempts to sell his wife’s anklet, their sole remaining valuable.
In response, the ever-faithful Kaṇṇaki tears off her breast, hurls it at the royal city of Maturai
(which promptly bursts into flames), and ascends into heaven as a goddess. Dating of the text
continues to be debated, but a scholarly consensus has emerged that places the narrative in the
fourth or fifth century. While the preface names one Iḷaṅkō Aṭikaḷ as the author of the text,
Iḷaṅkō is said to compose the story at the specific request of “Cāttaṉ, a grain merchant from
Maturai” (preface, lines 88–9),4 who has already been described as “Cāttaṉ of graceful Tamiḻ”
DOI: 10.4324/9781351030908-12 113

Anne E. Monius
(line 10). Cāttaṉ tells Iḷaṅkō of the events he has witnessed, and then listens as Iḷaṅkō narrates
the story back to him “in poetry and song” (line 60). Cāttaṉ appears once more in the text, as
he is called again to describe the events he has witnessed; here he is described as “the master
of graceful Tamiḻ” (25.66). Cāttaṉ, as a character in The Story of the Anklet, emerges as an
observant poet who witnesses extraordinary events first-hand, yet humbly yields the telling of
the narrative to his companion, Iḷaṅkō.
The preface to The Renunciation of Maṇimēkalai names Cāttaṉ as the author only once,
again pairing him with Iḷaṅkō but with their respective roles reversed. The dating of The
Renunciation of Maṇimēkalai is no more certain than that of The Story of the Anklet, although
continuing the story of minor characters in the first text perhaps suggests, at the very least, a
somewhat later date for Cāttaṉār’s work; a rough scholarly consensus has tentatively settled on
the sixth or seventh century. This time it is Iḷaṅkō – here called a “king” – who requests a story
from Cāttaṉ, described as “a prosperous grain merchant,” who tells the story “of the renuncia-
tion of Maṇimēkalai in thirty songs of eloquent Tamiḻ” (preface, lines 95–8).5
Beyond these sparse literary references,6 all that can be said of the historical Cāttaṉār must
be gleaned from the beautifully written poetic narrative that he authored, The Renunciation of
Maṇimēkalai. Continuing the story of minor characters in the earlier narrative, The Story of the
Anklet (where Maṇimēkalai is the daughter of Kōvalaṉ and his mistress), the complex narra-
tive is told in 4,758 lines arranged in thirty chapters, with sixteen “branch stories” or subplots
(Richman 1988, 2–3) woven into the primary story. With its technical terminology; references
to a wide range of Sanskrit, Pāli, and Tamiḻ literary sources; and dense philosophical mate-
rial, Cāttaṉār clearly anticipates a sophisticated Buddhist audience for his text, well-versed in
multiple poetic and philosophical traditions (see Monius 2001, 16–18).
The Basic Story of The Renunciation of Man imēkalai7

Cāttaṉār’s narrative opens in the coastal city of Pukār. Mātavi, Kōvalaṉ’s former lover and
mother of Maṇimēkalai, has just heard the news of her ex-lover’s tragic end and has renounced
her former life as a courtesan to follow the Buddhist teacher, Aṟavaṇaṉ; she has vowed that
her young daughter will do the same. Sent out with a friend to gather flowers for a garland,
Maṇimēkalai encounters a young prince of the city, who immediately decides that he must
possess her. The heroine’s namesake goddess, Maṇimēkalā, whisks the girl off to the island
of Maṇipallavam to save her from the prince’s advances; there Maṇimēkalai encounters a
magnificently bejeweled seat of the Buddha, and she begins to realize her previous karmic
trajectory that has brought her to the present-tense narrative. The goddess reappears and grants
the girl three mantras that enable her to fly, to change physical form at will, and to appease
her own hunger whenever necessary. Maṇimēkalai then receives a miraculous begging bowl
from a mysterious lake on the island, one that never empties of food if used to feed the poor.
Maṇimēkalai then returns to the city of Pukār to be reunited with her mother and friends.
They meet the Buddhist teacher, Aṟavaṇaṉ, and he proceeds to tell his young disciple the his-
tory of her miraculous alms bowl. Maṇimēkalai then feeds the hungry masses of Pukār with
the bowl. The prince, of course, has not forgotten the object of his desire, and he resumes his
pursuit. In order to avoid the prince’s amorous advances, Maṇimēkalai deploys the power of
her mantra to change form; the prince is then killed in a case of mistaken identity. Maṇimēkalai
is then imprisoned for her part in the crime. The queen attempts to have her murdered, but the
power of Maṇimēkalai’s mantras keeps her safe from harm. Aṟavaṇaṉ briefly relays the basics
of interdependent origination before Maṇimēkalai meets the previous holder of the miraculous
alms bowl. She then proceeds to the western capital of Vañci, where she visits the temple
114
Cāttan
‒ ār
dedicated to the heroes of The Story of the Anklet, Kōvalaṉ (her father) and his virtuous wife,
Kaṇṇaki. She assumes the form of a male ascetic and listens to ten non-Buddhist teachers
discuss their basic doctrines. Maṇimēkalai then travels east to Kāñci and, at the request of her
grandfather, commands the king to build a replica of Maṇipallavam there in recognition of the
power of the magical alms bowl. In the penultimate chapter, Aṟavaṇaṉ teaches proper inferen-
tial reasoning to Maṇimēkalai, and, in the final chapter, interdependent origination.
Finally, in the last two lines of the text, Maṇimēkalai, “having heard the dharma, took up
ascetic practice, and vowed to eradicate the karmic effects of birth” (30.263–4).
Cāttanār’s Buddhist Philosophy

¯
Coming at the very end of the narrative, the two final chapters of The Renunciation of
Maṇimēkalai – on proper inferential reasoning (chapter twenty-nine) and interdependent orig-
ination (chapter thirty) – are, by position, a culmination of sorts, serving as the immediate
precursors to the heroine’s final vow of renunciation. Both chapters have received significant
amounts of scholarly attention over the past century, largely in an effort to place Cāttaṉār his-
torically and in terms of doctrinal affiliation more precisely.
Buddhist Logic in The Renunciation of Manimēkalai

The penultimate chapter of Cāttaṉār’s The Renunciation of Maṇimēkalai contains the most
extensive discussion of inferential reasoning in Tamiḻ literature. Maṇimēkalai’s teacher,
Aṟavaṇaṉ, begins his discussion of Buddhist logic with the assertion that the Buddha accepts
only two valid means of knowledge: direct perception ( pirattiyam) and inference (karuttaḷavu)
(29.47–48). After a very brief treatment of direct perception, Aṟavaṇaṉ states that “all other
valid means of knowing properly belong to inference” (55–6). He then moves on to propose a
five-limbed syllogism, using the common fire-on-the-mountain example (57–63). The propo-
sition ( pakkam) is that “the mountain possesses fire.” The reason (ētu) is “because it possesses
smoke.” “Just like a kitchen” is the example (tiṭṭāntam). The application (upanayam) is “and
the mountain also possesses smoke.” The conclusion (nigamaṉam) is thus “because the moun-
tain possesses smoke, it also possesses fire.” Aṟavaṇaṉ then proceeds to critique this five-part
syllogism, ultimately arguing that “application and conclusion are connected to the example
and are thus included in it” (109–10).
For the remainder of his discourse on inference, Aṟavaṇaṉ focuses on the proper working
of a three-part inferential reasoning process, addressing himself first to the three valid forms
of proposition (113–21), three valid forms of reason (121–35), and two valid forms of exam-
ple (136–41), then turning to address the nine invalid forms of proposition (147–90), fourteen
types of invalid reason (191–325), and twelve varieties of invalid example (325–468). From
the sheer number of verses devoted to each topic, it would seem that Cāttaṉār particularly
concerns himself with guarding against faulty inferential reasoning, enumerating the varieties
of incorrect proposition, reason, and example, and providing brief examples of each.
A significant amount of scholarship to date has considered the system of Buddhist infer-
ential logic explicated by Aṟavaṇaṉ in its potential relationship to other Buddhist traditions
of logic, initially spurred by the recovery and publication of the Sanskrit Primer on Logic
(Nyāyapraveśa), a text originally attributed to Dignāga (whom many believed hailed from
Kāñcipuram, in what is now the Indian state of Tamil Nadu) in the early twentieth century
(1927–1930), but now largely viewed as the work of his disciple, Śaṅkarasvāmin. Kandaswamy
(1978, 253–312) correctly summarizes the many decades of research into Cāttaṉār’s possible
115
Anne E. Monius
Sanskrit sources for his twenty-ninth chapter on Buddhist inferential reasoning by describing
the author as “more an interpreter than a mere translator of Nyayapravesa” (287). Kandswamy
locates what he rightly terms “minor” discrepancies between the Tamiḻ and Sanskrit texts
(286–90): a slight variation in the order of the fallacies of proposition (286–87), two original
subcategories of fallacious example authored by Cāttaṉār (287), a host of original examples
provided in the Tamiḻ text (287–88, 291–305), and some variation in the adoption of Sanskrit
philosophical terms into Tamiḻ (288–90). Hikosaka (1989, 121–52), like Kandaswamy, devotes
a full chapter to discussion of the Tamiḻ narrative’s presentation of inferential reasoning, notic-
ing major similarities and minor differences between The Renunciation of Maṇimēkalai and the
Primer on Logic. Overall, Hikosaka concludes that Cāttaṉār provides more of an introduction
and less of a conclusion than the Primer on Logic, and that the differences in terminology and
number suggest that Cāttaṉār’s reference to this particular Sanskrit text is only partial, and
“that he should have used some other sources also” (135). Nearly a century of scholarship
on Cāttaṉār’s earliest presentation of formal Buddhist logic in Tamiḻ demonstrates an affinity
to Śaṅkarasvāmin’s Primer, yet what marks the Tamiḻ text’s wholly unique contribution to
Buddhist logic is its form, its embedding of dense philosophical discourse into highly stylized
poetic narrative (see in what follows).
Interdependent Origination in The Renunciation of Man·imēkalai

Cāttaṉār’s final chapter, “The Story of the Young Woman’s Renunciation in Order to Eradicate
the Qualities of Existence,” features Aṟavaṇaṉ’s teachings to his young student on the nature
of human existence, emphasizing the twelve-part chain of interdependent origination, with the
links in the chain (nidāna) here described in Tamiḻ as “appearance of dependent things” (carpiṉ
tōṉṟu) (30.17). Aṟavaṇaṉ emphasizes that “the cause of undying liberation is realizing” that
“sorrow is impermanent, caused by karma and the fruits of that karma” (29–31). The twelve
dependent links in the chain of causation “have no creator, nor have they been created” (39),
but are rather “dependent” entirely upon “karma and its results” (39). Aṟavaṇaṉ then describes
each of the twelve “dependent things” in turn, from ignorance ( pētaimai) (51–81), awareness
(uṇarvu) (82–3), name and form (aruvu uruvu) (84–5), six gateways (vāyil āṟu) (86–7), sense
contact (uṟu) (88), sensation (nukarvu) (89–90), craving (vēṭkai) (91), and attachment ( paṟṟu)
(92), to becoming ( pavam) (93–4), birth ( piṟappu) (95–7), old age (mūppu) (98–101), and
death (cākkāṭu) (102–3). Each arises in a chain of dependence (104–18) – a “circle of exist-
ence without end” (118) – that can only be undone, step by step, through the cessation of each
of the links in turn, beginning with ignorance (119–33). Particular attention (thirty-one verses:
51–81) is paid by Aṟavaṇaṉ to explaining the nature of ignorance, glossed in terms of lack
of awareness of the chain of interdependent origination, of “forgetting what is seen through
natural reason” (53), and lack of wisdom to discern good karma (nal viṉai) from bad karma
(tī viṉai). Having explained the basics of interdependent origination, Aṟavaṇaṉ then relates to
Maṇimēkalai the same teaching in terms of four groupings of the links in the chain of causa-
tion and three junctures among them (134–52); three categories of birth (153–58); the three
temporalities of past, present, and future (159–68); and the working out of karma (169–75).
Cāttaṉār, through his learned Buddhist character, then briefly explains the Four Noble
Truths (176–88), the five aggregates (189–90), six types of designation (191–215), four
methods (nayam) (217–34), and four forms of question and answer (235–49). He concludes
by affirming that “there is no one who fixes either bondage or liberation” (250–51), only
the elements ( poruḷ ) in the chain of causation that he has just outlined. His final call to his
young disciple – to contemplate what he has just taught and thus to eradicate hatred (ceṟṟam),
116
Cāttan
‒ ār
delusion (mayakkam), and darkness of mind (maṉattu iruḷ ) (256–60) – immediately results in
Maṇimēkalai’s final act of renunciation (263–4).
While Cāttaṉār clearly displays a wide knowledge of Buddhist forms of inferential reason-
ing in Sanskrit in his previous chapter, he here in the final chapter follows far more closely
Theravāda-aligned traditions in outlining each of the topics described earlier. As with the
twenty-ninth chapter, many scholars have sought potential sources for Cāttaṉār’s final phil-
osophical discussion. Overall, Cāttaṉār appears to follow most closely extant Pāli sources,
particularly Buddhaghosa’s Path of Purification (Visuddhimagga), and particularly the seven-
teenth chapter entitled “Explanation of the Soil of Wisdom” (Paññābhūminiddesa). Cāttaṉār’s
organization of the twelve links in the cycle of interdependent origination, for example, fol-
lows closely Buddhaghosa’s discussion of the same in the Path of Purification (17.289–90),
as does the Tamiḻ discussion of the junctures among the groupings, temporality (284–87),
the Four Noble Truths (300), and the four methods (309–13). As in the chapter on inferential
reasoning, however, Cāttaṉār’s presentation contains unique definitions and details, and lies
wholly embedded in a long and complex poetic narrative.
Philosophy Embedded in Poetic Narrative

No matter what his specific sources might have been, Cāttaṉār’s particular contribution to the
history and study of Buddhist philosophy lies less in the content of his philosophical chapters
and far more in the form that each takes and the narrative purposes that each serves. Cāttaṉār
organizes his narrative in such a way that the story itself leads naturally to these final two sets
of Buddhist teachings. Long before the final chapter, to cite but one brief example, Aṟavaṇaṉ
gives a short foretaste of his later teaching on interdependent origination as he counsels the
young girl’s mother on the true nature of reality (24.105–40). The gradual transformation of
Maṇimēkalai’s character, to cite another example, culminates naturally in the final two chapters
of the narrative that mark the final removal of, as previously, “hatred, delusion, and darkness of
mind.” Without the final two chapters, the young woman’s story would be incomplete; she would
not be ready to renounce the world. In its complex plot that interweaves the main narrative and
multiple branching stories, The Renunciation of Maṇimēkalai slowly both reveals and embod-
ies its sophisticated concept of karma in the context of interdependent origination, revealing an
intricate pattern of interlocking lives, deeds, and events. Cāttaṉār spins a narrative web filled
with characters, events, and imagery that compels its audience to confront and reflect upon both
the ultimate transience and profound interconnectedness that he describes theoretically only in
the final chapter. In similar fashion, a long story that Aṟavaṇaṉ relays to Maṇimēkalai before he
speaks of formal logic in chapter twenty-nine – of the goddess, Maṇimēkalā, and her dramatic
rescue of one of Kōvalaṉ’s ancestors as the reason for the young girl’s name – gestures toward
the overall logic in the text of ripening karmic conditions and renunciation, of past deeds com-
ing to full fruition in the future: “On that day of your birth, the goddess, Maṇimēkalā, with her
body of soft petals, spoke to your father in a dream, as if he were awake, about your resolve
to renounce” (29.31–3). In such a narrative context, the long and dense discussions of proper
inferential reasoning and interdependent origination are especially significant.
Philosophy as Narrative
Critical to full understanding of the philosophical chapters in Cāttaṉār’s The Renunciation of
Maṇimēkalai is a careful consideration of the role that those chapters play in the larger narra-
tive that each serves. As the story of a beautiful young courtesan’s journey toward her final act
117
Anne E. Monius
of renouncing worldly life, the first twenty-eight chapters of Cāttaṉār’s text narrate the epony-
mous heroine’s gradual awakening to the truth of the Buddha’s dharma and the challenges she
faces (primarily from the infatuated prince and his royal parents) as she turns away from the
lifestyle of a courtesan and embraces the Buddhist path. Two themes dominate the overall nar-
rative: the working out of karma in human lives and the unfolding of those conditions that lead
the primary character, Maṇimēkalai, to renounce in the final lines of the text.
Cāttaṉār’s text repeatedly invokes karma to explain events and circumstances in which vari-
ous characters find themselves. Yet karma in The Renunciation of Maṇimēkalai clearly oper-
ates in a very different way than in Iḷaṅkō’s earlier narrative, The Story of the Anklet. Early
in Cāttaṉār’s text, the young prince, enchanted by a glimpse of the beautiful Maṇimēkalai,
is told that “the body is the result of karma and the cause of future karma” (4.113). Later,
Maṇimēkalai attributes her fear and loneliness as she finds herself alone on the isolated island
of Maṇipallavam to the same bad karma (ve viṉai) that led to her father’s earlier execution
(8.40–3). The violent death of the infatuated prince in hot pursuit of Maṇimēkalai is explained
to his father, the king of Pukār, as a result of bad karma (tī viṉai), and karma plays a dominant
role in the sequence of actions that leads up to the moment of the murder (22.193–203). Karma
in Cāttaṉār’s narrative becomes a powerful force at work in a complex web of conditions gener-
ated by the processes of interdependent origination (that, as previously, are spelled out explic-
itly only in the text’s final chapter). The logic of karma can perhaps be seen most clearly in
Cāttaṉār’s unique treatment of the two main characters in the earlier narrative, The Story of the
Anklet: Kōvalaṉ and Kaṇṇaki. Kaṇṇaki, for example – eventually emerging as a great goddess
in the earlier narrative – in Cāttaṉār’s hands becomes simply another karma-accumulating indi-
vidual who “will return to the ocean of births and be caught up in the cycles of birth and death”
after enjoying the fruits of her good karma in heaven (26.40–1). Kaṇṇaki and Kōvalaṉ will be
reunited on earth to live out their remaining karma until the Buddha comes again, performing
acts of kindness to all for a long time (26.54–61). In The Renunciation of Maṇimēkalai, the logic
of karma across past, present, and future requires a far more complex narrative structure than in
The Story of the Anklet, including flashbacks and episodes that combine past, present, and future
actions and a far greater number of characters and subplots. The earlier narrative’s straightfor-
ward A-in-the-past–yields–B-in-the-present causal sequences are replaced in The Renunciation
of Maṇimēkalai by an intricate tapestry of characters, narratives and subnarratives, and scenes
that suggest: A yields B, provided that C, D, E, and F are present to provide the proper environ-
ment in which both A and B can come to full fruition.
Another dominant theme that links the many plots and subplots of The Renunciation of
Maṇimēkalai builds upon the logic of karma noted previously and ties the flow of the narrative
to the text’s dense concluding chapter on interdependent origination and the heroine’s final act of
renunciation. Signaled by the “causes” or “conditions” (ētu, from Sanskrit or Pāli hetu), in combi-
nation with verbal forms such as “blossoming” (etirtal), “maturing” (murtital), or “manifesting”
(nikaḻtal) – literally a blossoming, maturation, or manifestation of the beneficial root conditions
that lead to enlightenment – The Renunciation of Maṇimēkalai structures its complex web of
narratives around its central character’s karmic ripening, around Maṇimēkalai’s growing ability
to hear and truly comprehend the Buddha’s teachings (Monius 2001, 24–32). Although the Tamiḻ
narrative never defines what the “good” conditions (nal ētu) are exactly, wider Theravāda Bud-
dhist tradition holds the primary beneficial “roots” (Pāli kusalamūla) to be absence of attach-
ment (alobha), enmity (adosa), and delusion (amoha) (Buddhaghosa 1969, 17.66–70). Cāttaṉār
grants the manifestation of beneficial root conditions (ētunikaḻcci) tremendous agency through-
out the story he tells, beginning with the introduction of his heroine. Overhearing her mother,
Mātavi, tell a friend about the violent death of her former lover, Kōvalaṉ, Maṇimēkalai sheds
118
Cāttan
‒ ār
fateful tears that propel her outside her home and spark the main events of the narrative to come:
“Because there had blossomed (etirntu uḷatu) for Maṇimēkalai a manifestation of beneficial con-
ditions (ētunikaḻcci) – like the fragrance emanating from a flower – she felt tremendous sorrow”
(3.2–4). Later in the story, Manimēkalā (the goddess) connects the appearance of these benefi-
cial conditions with the young woman’s future commitment to the Buddhist path: “Because the
conditions (ētu) that lead to the path of dharma of the First Sage have matured (mutirntatu),” she
goes on to explain that she has whisked Maṇimēkalai off to an isolated island to save her from
the prince’s amorous advances (7.19–22). As the young heroine moves toward her final desti-
nation of Kāñcipuram to learn of logic and interdependent origination from her teacher before
renouncing the world, references to the arising of beneficial conditions shift significantly to the
future. Twice, characters predict what will happen in Kāñci to result in Maṇimēkalai’s final act:
“All the manifestations of the beneficial conditions (ētunikaḻcci yāvum) will occur there. They
are many” (12.104–7 and 21.155–58). Immediately prior to her renunciation, Maṇimēkalai’s
Buddhist teacher, Aṟavaṇaṉ – having just explained to her the doctrine of interdependent origi-
nation – commands his student to eradicate those conditions (ētu) that are unwholesome or
non-beneficial: attachment ( paṟṟu), enmity (ceṟṟam), and delusion (mayakkam) (30.254–260).
After this mandate from her teacher, now with all the beneficial conditions in place for her to
advance along the path toward enlightenment, only then can Maṇimēkalai make her final vow
of renunciation “to eradicate the karmic effects of birth” (30.264). The manifestation of benefi-
cial conditions – and the final eradication of the non-beneficial – explains the transformation of
Maṇimēkalai’s karmic state as witnessed throughout the narrative.
If the themes of karma and the manifestation of the beneficial conditions leading to the
heroine’s renunciation and enlightenment tie the larger narrative to the philosophical discus-
sions of Cāttaṉār’s final two chapters, then his twenty-seventh chapter speaks directly to the
topics that those discussions address. The narrative presentation of ten distinct schools of
non-Buddhist philosophical and religious thought, presented to Maṇimēkalai shortly after she
arrives in Vañci, might, on the one hand, simply be read as a catalog of philosophical thinking
present in sixth- or seventh-century southern India, a narrative cousin of sorts to Mādhava’s
fourteenth-century Collection of All Philosophies (Sarvadarśanasaṃgraha). Yet a careful
reading of the presentation of each non-Buddhist school of thought reveals the full extent to
which Cāttaṉār carefully shapes what each teacher says – primarily focusing on logic, the role
of divine intervention in human life, and the constituent elements of the universe – to pave the
way for the “correct” Buddhist teachings on similar topics in the final two chapters of the text.
For example, Maṇimēkalai first encounters in Vañci a Vedic logician (aḷavai vāti) (27.3).
Citing Vedavyāsa (Vētaviyātaṉ), Kirutakōṭi, and Jaimini (Caimiṉi) (5–6), the logician enumer-
ates ten valid means of knowledge and proceeds to explain each in some detail. Of the second,
inference (karuttaḷavu), he lists three types; each is defined and an example provided. He con-
cludes his discussion of inference by noting that inferential reasoning works validly across the
three times of past, present, and future (38). The remaining modes of valid knowledge are then
discussed in turn, with brief examples provided, followed by a short discussion of eight falla-
cies ( piramāṇāpācaṅkaḷ ) (57–77). The Vedic logician then concludes his presentation by nam-
ing six distinct religious-philosophical traditions (Lokāyata, Buddhist, Sāṃkhya, Naiyāyika,
Vaiśeṣikha, and Mīmāṃsā) in terms of the modes of knowing that each holds to be valid
(78–85). Maṇimēkalai simply hears all of these and departs, immediately encountering a wor-
shipper of Śiva who boldly declares, “Īcaṉ is lord!” (86). No comment is offered on the valid-
ity – or lack thereof – of the presentation of any non-Buddhist teacher until the young woman
enters Kāñcipuram to sit at the feet of her Buddhist teacher, Aṟavaṇaṉ. To him, she describes
all the doctrines of other communities learned in Vañci as being “the negation of what is
119
Anne E. Monius
good” (29.44), and she pleads with the respected teacher to impart to her “the true doctrine”
(45). What follows is the lengthy discussion of proper and improper inferential reasoning,
culminating in Aṟavanaṉ’s confident assertion that “through these proper methods” of infer-
ential reasoning just described to her, she will “know what to accept as truth and as falsehood
without any doubt” (471–72). The Śaiva’s confident assertion of his lord that “his appearance
removes sorrow” (27.93) and the Vaiṣṇava’s equally confident claim that “Nāraṇaṉ is our pro-
tection” (99) similarly provide the salient point of departure for the final chapter’s discussion
of interdependent origination in which “no lord assigns bondage and liberation” (30.250–1).
The presentations of five teachers – the Ājīvaka, the Jain, the Sāṃkhyan, the Vaiśeṣikan, and
the materialist – all focus on the nature of substance ( poruḷ ), providing the perfect foil for the
final chapter’s discussion of all things ( poruḷ ) as “dependent” (cārpu). To all the Vañci teach-
ers, Maṇimēkalai responds simply by laughing (27.280), as the proper Buddhist understanding
of all the doctrines taught is yet to come.
Philosophy as Poetry
Not only does Cāttaṉār embed his philosophical chapters in a complex narrative of which
they form an integral part, as described previously, but he also composes both chapters – as he
does the entire text – in sophisticated poetic form. The text describes itself in terms of “songs”
( pāṭṭu) (preface, line 98), and the oldest among the surviving manuscripts refers to each chap-
ter of the narrative with the same term. Composed primarily in one of Tamiḻ literature’s oldest
meters (akavaḷ ), Cāttaṉār’s text resounds with all the stylistic and rhetorical embellishment
possible in early medieval Tamiḻ poetry, rendering even the most basic statement of a proper
syllogism elegantly beautiful. Consider, for example, the following straightforward rendering
of a syllogism:
The proposition is: “This mountain possesses fire.”

“Because it possesses smoke” states the proper reason.
“Just like a kitchen” states the example. (29.59–61)
In Cāttaṉār’s poetic hands, however, this becomes an elegant three lines filled with playful
alliteration, rhyme, and tightly structured internal rhythm:
pakkam immalai neruppuṭaittu eṉṟal

pukaiyuṭaittu ātalāl eṉal poruntu ētu
vakaiyamai aṭukkaḷai pōl tiṭṭāntam
Although a full discussion of Tamiḻ poetic form lies far beyond the scope of this chapter, the
preceding lines lend the syllogism a coherence and beauty that the English translation (or a
similar statement in prose) simply cannot capture. The initial rhyme of pukai and vakai ties
together the second and third lines, as does the alliteration of pakkam and pukai in the first and
second lines. The internal coherence of each poetic line is bound together by the repetition of
“pa” sounds in the first line in pakkam and neruppuṭaittu; the long vowels of ātalāl and ētu in
the second line – like the repetition of the -ai in vakaiyamai and aṭukkaḷai in the third line –
similarly grant each line an elegant coherence.
Cāttaṉār’s discussion of inferential reasoning, in other words – like his final discussion
of interdependent origination – is meant to be not only philosophically meaningful, but also
120
Cāttan
‒ ār
poetically beautiful. The final two chapters of the text not only perform significant work within
the overall narrative, as previously described, but each also contributes to the poetic impact
of The Renunciation of Maṇimēkalai as a whole. The precise nature of that poetic impact on
Cāttaṉār’s learned audience cannot be known for certain, but literary theory in both Tamiḻ
and Sanskrit suggests that a poetic text such as this aims to cultivate certain emotional experi-
ences to be savored among connoisseurs. The many scenes of hunger and profound suffering
throughout the narrative – no less Maṇimēkalai’s final vow to eradicate the karmic effects
of birth (30.264) – taken together suggest an overall ethos of pathos and compassion, given
philosophical framing in the final two chapters (see Monius 2001, 33–57).
Conclusion
The seemingly odd combination of narrative poetry and philosophy found in Cāttaṉār’s sole
extant work suggests that the complexities of both the text’s vision of karma and the gradual
unfolding of the beneficial conditions leading to the heroine’s renunciation require the power
of story-telling to convey the full impact of such forces at work in the world. The preceding
narrative does more than merely illustrate the points of doctrine noted in the final two chapters;
rather, it conveys more fully how proper reasoning and interdependent origination operate in
human lives. Logic and the twelve links in the chain of interdependent origination cease to be
merely abstract ideas; they have been shown throughout the entire narrative to be the forces
that govern everyday human existence.
The dense philosophical discussions of Cāttaṉār’s final two chapters, placed at the end of
an equally complex narrative of primary plot suffused with multiple branch stories, offer a the-
orization of sorts for the narrative that precedes it; one without the other would be insufficient.
The Renunciation of Maṇimēkalai provides a rich resource for considering Tamiḻ-language
forms of narrative philosophy, as Cāttaṉār seamlessly blends extensive philosophical discus-
sion into his core narrative. Cāttaṉār’s text presents the earliest extant example of formal
philosophical discourse in Tamiḻ, and the embedding of philosophy in poetic narrative appears
to be a unique feature of Tamiḻ literature of this general period. Composed perhaps in the tenth
century by an unknown Jain author, for example, the Nīlakēci features its eponymous heroine
traveling throughout India to debate the celebrated teachers of other religious schools and
convert them to the Jain path. The fourteenth-century commentary on the text, attributed to
Camaṉa Tivākara Vāmaṉa Muṉivar, preserves a verse from a long narrative no longer extant,
the presumably Buddhist Kuṇṭalakēci,8 to which the anonymous Jain text is taken to be a
direct response (Nīlakēci viḷakka urai 1973, 138). Unlike The Renunciation of Maṇimēkalai,
in which more than twenty-five chapters of beautifully wrought narrative prose precede the
later chapters on the philosophies of other religious schools – the treatment of inference and
the final chapter on interdependent origination – the Jain text devotes far more time to debate
and considerably less space to building a narrative. Other poetic philosophical texts are known
only through stray references, such as a handbook of logic in metrical form called The Book
of Logic (Aḷavainūl) (Cāminātaiyar 1978, 137), perhaps pointing to a larger Tamiḻ tradition of
composing philosophical argument in verse.
Cāttaṉār’s The Renunciation of Maṇimēkalai provides a brief but rich resource for consider-
ing the knowledge of – and significance of – Buddhist philosophy in Tamiḻ intellectual culture
in the early medieval period. Composed in verse, rather than the prose of philosophical com-
mentary, and embedded in sophisticated poetic narrative in which both play a critical role, each
chapter’s presentation of philosophical argument has few parallels in Sanskrit or the regional
121
Anne E. Monius
languages of pre-colonial South Asia. Cāttaṉār’s sole work offers significant opportunities for
fresh consideration of the relations among literature and philosophy, narrative, and logic.
Notes
1 [Editors’ note:] We feel extremely fortunate to have received this unique contribution from Anne E.
Monius, who inspired so many scholars with her sensitivity to literary culture in the history of reli-
gions in South Asia, not long before her unexpected passing on August 3, 2019. Anne was renowned
for her quiet yet profound generosity toward her students, as well as all others who came into her orbit.
As she translates elsewhere, quoting the protagonist in The Renunciation of Maṇimēkalai, “Those who
(never) cease loving all human beings are those who realize the final truth that eradicates suffering”
(Anne E. Monius, “Literary Theory and Moral Vision in Tamil Buddhist Literature,” Journal of Indian
Philosophy 28 [2000], 211). By this logic, surely Anne is one who has realized that final truth and thus,
we must believe, suffers no more.
2 The best available translation of The Renunciation of Maṇimēkalai into English is by Daniélou
(Cāttaṉār 1989).
3 Although the author of The Renunciation of Maṇimēkalai is named Cāttaṉ in the preface to the
text, scholarly convention adds the Tamiḻ honorific -ār to his name. This practice will be followed
throughout.
4 All quotations from The Story of the Anklet are drawn from Iḷaṅkō Aṭikaḷ (1978). All translations from
the Tamiḻ throughout the chapter are by the author.
5 All quotations from The Renunciation of Maṇimēkalai are drawn from Cāttaṉār (1981).
6 While many scholars have sought to align the author of The Renunciation of Maṇimēkalai with earlier
classical Tamiḻ poetry (known collectively as “Caṅkam” literature and dated to the early centuries
of the Common Era), no clear connection can be found thematically or stylistically with such earlier
work attributed to a poet known as Cīttalai Cāttaṉār. Named as the author of roughly eleven poems
from these early anthologies of themes of “inner” (akam) love and “outer” ( puṟam) war, nothing aside
from the name, Cāttaṉ, seemingly ties this poet to the author of The Renunciation of Maṇimēkalai.
7 See Richman (1988, 19–23) and Monius (2001, 18–24) for more detailed plot summaries.
8 Like the Nīlakēci, presumably known by the name of its heroine.
Bibliography
Buddhaghosa. 1969. Visuddhimaggo with Paramatthamañjūsaṭīkā. Edited by Revatadhamma. 4 vols.
Varanasi: Varanaseya Sanskrit Vishvavidyalaya.
———. 1991. The Path of Purification (Visuddhimagga). Translated by Bhikkhu Ñāṇamoli. Kandy:
Buddhist Publication Society.
Cāminātaiyar, U. Vē. 1978. Caṅka tamiḻum piṟkāla tamiḻum. Ceṉṉai: Ṭākṭar U. Vē. Cāminātaiyar Nūl
Nilaiyam.
Cāttaṉār. 1981. Maṇimēkalai. Edited by U. Vē. Cāminātaiyar. Ceṉṉai: Ṭākṭar U. Vē. Cāminātaiyar Nūl
Nilaiyam.
———. 1989. Manimekhalaï: The Dancer with the Magic Bowl. Translated by Alain Daniélou. New
York: New Directions.
Dignāga. 1927–1930. Nyāyapraveśa. Edited by Anandshankar B. Dhruva. Gaekwad’s Oriental Series,
38–39. Baroda: Oriental Institute.
Hikosaka, Shu. 1989. Buddhism in Tamiḻnadu: A New Perspective. Madras: Institute of Asian Studies.
Iḷaṅkō, Aṭikaḷ. 1978. Cilappatikāram mūlamum arumpata uraiyum aṭiyārkkunallār uraiyum. Edited by
U. Vē. Cāminātaiyar. Ceṉṉai: Ṭākṭar U. Vē. Cāminātaiyar Nūl Nilaiyam.
———. 1993. The Cilappatikāram of Iḷaṅkō Aṭikaḷ: An Epic of South India. Translated by R. Parthasar-
athy. New York: Columbia University Press.
Kandaswamy, S. N. 1978. Buddhism as Expounded in Manimekalai. Annamalainagar: Annamalai
University.
Krishna Raja, S. 1998. “Theory of Anumāna in The Renunciation of Maṇimēkalai.” In Buddhism in
Tamiḻnadu: Collected Papers, edited by R. S. Murthy and M. S. Nagarajan, 279–300. Chennai: Insti-
tute of Asian Studies.
122
Cāttan
‒ ār
Mādhva. 1997. Sarvadarśanasaṃgraha of Mādhavācārya. Edited by K. L. Joshi. Translated by E. B.

Cowell and A. E. Gough. New Delhi: Parimal Publications.
Monius, Anne E. 2001. Imagining a Place for Buddhism: Literary Culture and Religious Community in
Tamiḻ-Speaking South India. New York: Oxford University Press.
Nīlakēci viḷakka urai. 1973. Edited by Po. Vē. Cōmacuntaraṉār. Tirunelvēli: Teṉṉintiya Caivacittānta
Nūṟpatippu Kaḻakam.
Richman, Paula. 1988. Women, Branch Stories, and Religious Rhetoric in a Tamil Buddhist Text. Syra-
cuse, NY: Maxwell School of Citizenship and Public Affairs, Syracuse University.
Zvelebil, Kamil V. 1995. Lexicon of Tamiḻ Literature. Handbuch der Orientalistik, Zweite Abteilung:
Indien, 9. New York: E. J. Brill.
123
8
SARAHA
The Anti-Philosopher as Philosopher
Roger R. Jackson
Introduction
The late-first-millennium tantric poet Saraha (“the Archer”)1 seems an unpromising candidate
for inclusion in a volume governed by an author-centered approach to Indian Buddhist phi-
losophy. He is, to begin with – a bit like Friedrich Nietzsche in the Western tradition – more an
anti-philosopher than a philosopher, at least if we conceive of philosophy as a set of rational
operations and arguments utilized in an allegedly dispassionate search for objective truth. He
scorns his Indian contemporaries for their arrogant posturing and unwarranted confidence, and
he insists, as a gnostic and mystic, that if we are to be free, we must transcend conceptuality,
with its attendant dichotomies, and directly realize the primordial purity, emptiness, luminos-
ity, bliss, and freedom of our own mind – and then express our realization not in the desiccated
prose of the academician, but in the evocative and resonant language of the poet or singer of
songs. Furthermore, if Saraha is an author at all, he is a most elusive one, almost impossible to
locate in place and time. Indeed, some scholars have questioned his historicity and suggested
that he may be a literary invention, a name and a set of colorful stories slapped onto a collec-
tion of works that may not all have been composed by the same person. Yet for South Asian
and, especially, Tibetan Buddhists of the past millennium, Saraha has been a supreme exem-
plar of a radical way of being human that points beyond reason to the “joyful wisdom” pos-
sessed by those courageous enough to reject social, religious, and philosophical conventions;
see themselves as they truly are; and live their lives, accordingly, with spontaneity, wisdom,
and compassion. Following tradition, we will conjure “Saraha” here as a single individual, do
our best to locate him within the Indian and Buddhist landscape from which he emerged, and
explain why this arch-critic of rationality may deserve the title of “philosopher.”
Saraha in His Setting

In the pantheons of late Indian Buddhism and Tibetan Buddhism, Saraha belongs to a class of
saints known as great adepts – mahāsiddhas – who flourished in South Asia in the late first and
early second millennia CE. Often said to number eighty-four, they are known primarily through
the colorful stories of their lives and deeds composed in India, Nepal, and Tibet, and through
spiritual songs – aphoristic couplets (dohā), adamantine songs (vajragīti), and performance
124 DOI: 10.4324/9781351030908-13

Saraha
songs (caryāgīti) – handed down under their names in oral and written traditions. The adepts
are notable for subverting social and religious norms, glorifying the body and senses, celebrat-
ing spiritually potent female wisdom figures called yoginīs or ḍākinīs, instructing ordinary
people through vernacular songs, displaying supernormal powers, and unabashedly exalting
the freedom and joy entailed by direct knowledge of ultimate reality – variously referred to as
the real (tattva), gnosis ( jñāna), the connate nature (sahaja), the natural mind (nijacitta), great
bliss (mahāsukha), and the great seal (mahāmudrā) – or simply “that” (tat) or “mind” (citta).
Most great adepts were followers of the esoteric Mahāyāna traditions known as Vajrayāna, or
Secret Mantra, that dominated Indian Buddhism from the seventh to thirteenth centuries. Based
on teachings retroactively attributed to the Buddha called tantras, these restricted – yet, para-
doxically, widely popular – traditions insisted that their followers could become fully awakened
in this very lifetime through receiving initiation from a qualified guru and undertaking practices
in which they “made the goal into the path” by seeing themselves, other beings, and their envi-
ronment as divine. In the late and highly esoteric systems known as the Mahāyoga and Yoginī
tantras, practitioners would transform themselves into buddhas through elaborate and some-
times transgressive yogic techniques that involve manipulating mental and physical energies
within the subtle body (sukṣmaśarīra) that interpenetrates our coarse physical body, sublimating
such “negativities” as desire, anger, ignorance, and even death into their awakened equivalents.
No great adept is more important than Saraha. He was influential in South Asia, but it is
really in the Tibetan tradition that he gained prominence, especially in his capacity as the human
source of the widespread Buddhist philosophical and practical teaching known as the great seal
(mahāmudrā). Originally tantric, the term came to denote many things in Buddhist circles,
including a ritual hand-gesture; clear visualization of oneself as a buddha–deity; a consort for
sexual yoga practice; the primordial purity of the mind; a meditation technique whereby mind
settles into its own empty, luminous, and blissful nature; and the final fruition of the Mahāyāna
path, perfect buddhahood. In Tibetan Buddhism, especially but not solely in the Kagyü (bka’
brgyud) order, the great seal became the basis for much contemplative experimentation, creative
mythmaking, scholastic analysis, poetic expression, and philosophical debate – and amidst all
this, Saraha was a central figure: not only were his songs constantly read, memorized, quoted,
and discussed, but long after his death, Tibetan masters from across the religious spectrum
encountered him in visions and claimed to have received teachings directly from him.
Some of Saraha’s dohā collections and performance songs have been preserved in two
medieval Indic languages, Apabhramśa and Old Bengali,2 which were common in northeast
India late in the first and early in the second millennium. The largest trove of Saraha texts –
twenty-five or so – is found in various editions of the Tengyur (bstan ’gyur), the collection
of works by Indian followers of the Buddha translated into Tibetan between the eighth and
fourteenth centuries; the Saraha Tengyur texts were all translated in the eleventh century or
later. Many are devoted to tantric deities and rituals, but most are tantric songs. The best
known of these, the Treasury of Dohā Songs (Dohākoṣagīti) – often referred to by Tibetans as
the People Dohās (PD) to reflect its original audience – is typically included in an “essential
trilogy” (T. snying po skor gsum) with two other collections, the Queen Dohās (QD) and King
Dohās (KD),3 also named for their audience. Other Saraha songs in the Tengyur include a
trilogy of adamantine songs (vajragīti) – the Body Treasury (BT), Speech Treasury (ST), and
Mind Treasury (MT)4 – and numerous instruction songs on such topics as the great seal, non-
conceptual awareness, and the Buddhist path. Saraha’s songs typically involve a rich mix of
sardonic social criticism, ecstatic affirmation of the purity and potency of the mind, and direct
instructions to his listeners, all brought indelibly to life through his skillful deployment of
simile, metaphor, paradox, humor, word-play, and poetic rhythm.
125
Roger R. Jackson
The surviving biographical material on Saraha was written beginning in the eleventh cen-
tury and is available only in Tibetan, though some works may have been translated from Indic
originals. As noted by Kurtis Schaeffer (2005, 15–25), there are two major Saraha narratives.
In the “radish-girl narrative” found in Abhayadatta’s Lives of the Eighty-Four Great Adepts,
Saraha is an observant east Indian brahmin by day and a practitioner of Buddhist tantra by
night. Accused by his fellow brahmins of drinking alcohol, he denies the charge, then proves
his veracity by surviving various physical ordeals. Asked by his repentant accusers to teach the
path to liberation, he sings dohās to the country’s king, queens, and people, respectively. He
then retreats to another country with a fifteen-year-old servant girl, who assists him in his prac-
tice. One evening, as she cooks him a radish stew, he enters a contemplation that lasts twelve
years. When he emerges from it, he asks the girl for his radish stew. She mocks him for his
attachment to radishes and his enslavement to conceptual thought. At this, he attains liberation,
and departs with the girl for the celestial realm of the ḍākinīs. In the “fletcheress narrative,”
first mentioned by the eleventh-century Nepali commentator Balpo Asu and filled out by later
Tibetan writers, Saraha is a south Indian brahmin who one day encounters a low-caste female
arrow maker in the marketplace of his town. When he asks her about her profession, she
replies, “Noble son, the intent of the Buddha can be understood through symbols and actions,
not through words and texts,” and proceeds to give him an explanation of the way the arrow
and archery symbolize the realization of one’s own natural mind. At this, Saraha is instantly
liberated. He moves to a cremation ground with the fletcheress, where they engage in tantric
rituals that soon draw the condemnation of the townspeople. In response, Saraha teaches his
dohā collections to the people, queens, and king, respectively, then departs to the southern
mountain of Śrī Parvata, where he resides to this day.
Saraha as Critical Philosopher

We saw in the biographical traditions that by living with a low-caste woman Saraha courted
the opprobrium of people obsessed with caste purity. His rejection of this cornerstone of
the Indian social system is quite evident in a song attributed to him by the Tibetan historian
Tāranātha (1575–1634):
Oh ho! I am a brahmin and I live with a girl who works with bamboo.
I see neither caste nor no caste.
(Tāranātha 1983, 35)
He provides further evidence of his unconcern for caste in his Queen Dohās, where he
describes females suitable to serve as a yogī’s great seal (mahāmudrā), his partner for sexual
yoga practice, as including “low-caste women, laundresses, whores, or pickers of rags” (QD
436; Jackson 2012, 179). Later, he speaks of how a yogī who has practiced sexual yoga with
his consort “fades into conceptless space” so that even when he lives with her in a cremation
ground, “he makes friends with outcastes, and draws the corpse cart along” (QD 47; Jackson
2012, 179). Similarly, he asserts in his People Dohās that “enjoying yourself in an outcaste’s
home, you don’t get covered with dirt” (PD 56b, Jackson 2004, 857). And, on the most basic
level, “in the connate nature where things taste the same, you’ll find neither śūdra nor brah-
min” (PD 46, Jackson 2004, 80).
Saraha is even more pointed in his critique of religion. His most famous attack on the prac-
tices and practitioners of his day comes at the outset of his People Dohās, where he denounces
members of his own brahmin caste for their vain incantation of the Vedas, pointless oblations,
126
Saraha
and spiritual charlatanism (PD 1–5; Jackson 2004, 53–55) and goes on to question the utility
of such Hindu rituals as lamp lighting, food offering, mantra practice, pilgrimage, forest pen-
ance, and immersion in holy rivers (PD 14–15; Jackson 2004, 60). He also decries the extreme
asceticism of Jains, whom he describes as “mocking the path” by going naked and plucking
out their hairs (PD 5; Jackson 2004, 55). Nor does Saraha spare his fellow Buddhists. He
complains about “so-called novices, monks, and elders” who “seek to dry up the intellect,” as
well as proponents of the Mahāyāna, “where scripture turns to sophistry and word-play” (PD
10–11a, Jackson 2004, 57–58). To those Buddhists who assert that the heart of their practice
is meditation, he remarks,
The whole world is deluded by meditation –

and no one perceives their inmost nature.
(PD 35b; Jackson 2004, 70)
His scorn is also directed toward proponents of the Vajrayāna. In the Queen Dohās, he sings
of how the adamantine gnosis, “lovely in its natural state,” is
devoid of maṇḍalas or fire offerings,

quite without mantras, mudrās, and consecrations,
and can’t come about through any tantra or treatise.
(QD 31, Jackson 2012, 177)
Elsewhere, he dismisses advanced subtle-body practices, noting that “drawing energies up and
down/in the turning cakras, you will not find . . . truth” (MT 25, Braitstein 2014, 165–66), and
asks enthusiasts of tantric sexual yoga,
If you don’t grasp everything as it is,

how in the midst of sex will you perfect great bliss?
( PD 91, Jackson 2004, 105)
Saraha’s critique of religious practice extends to nearly every form of human behavior, so that
his advice sometimes seems to amount to “do nothing”:
Mind at its core is free from virtues and faults.

Truth itself requires no deeds at all,
while the mind that’s done with deeds is great bliss sublime.
(QD verse 77, Jackson 2012, 183)
In short, so much for religion.

Saraha also harshly criticizes philosophy and philosophers. In his Body Treasury, he attacks
the four major Indian Buddhist philosophical schools:
Vaibhāṣikas and Sautrāntikas,

Yogācārins and Mādhyamikas and so forth,
criticize each other and argue;
ignorant of the real . . .,
they turn their backs on the connate nature.
(BT 2, Braitstein 2014, 125)
127
Roger R. Jackson
He consistently mocks the pretensions of thinkers who “intellectualize based on signs/but . . .

can’t see the real,” and “die of thirst in the desert of variant texts and meanings” (PD verses
78a, 56, Jackson 2004, 98, 84). In the Queen Dohās, he notes the falsehood of “every term and
symbol you impose” and the importance of “clearing away the husk of logical complication”
so as to comprehend what is true, real, and essential (QD 75, 79, Jackson 2012, 183, 184).
Indeed, not just intellect but mind itself is problematic: “the world is bound by mind,/and no
one at all can grasp the non-mind” (PD 78, Jackson 2004, 97) because “the mind . . ./can’t be
considered anything by anyone, there is no mind” (QD verse 24, Jackson 2012, 176). Further-
more, that which is to be realized – the real, the connate nature, the great seal – is
without syllable or color or qualities . . .;

it can’t be spoken or known.
( PD 58, Jackson 2004, 87)
One important consequence of the rejection of language and thought is that the dichotomies
through which we understand and operate in the world are effaced. Thus, the distinctions
between self and other, saṃsāra and nirvāṇa, pure and impure, suffering and pleasure, and
even good and bad, simply do not apply. Saraha describes reality as a state in which “there is
no self or other” and we are “free from existence and nirvāṇa” (PD 96, 110, Jackson 2004,
108, 115). Regarding ethics, he observes:
Even the label ‘good’ is a sickness plunging you into saṃsāra,

and when you label an act as ‘evil,’ you can’t prevent its fruition.
(QD 75, Jackson 2012, 183)
The consequences of this stance for a yogī’s conduct are dramatic:
Behaving without regard for dos and don’ts, like an elephant

plunging impulsively into a pond, his mind perpetually crazed,
he performs the basest of deeds yet is free.
(QD 50, Jackson 2012, 180)
Saraha rejects language and thought – and by extension social, religious, and philosophical
norms – because the ultimate nature of all phenomena is emptiness (śūnyatā): their lack of
self-existence, their inconceivability, their unfindability apart from the mind. As he sings in
the King Dohās,
There’s nothing to be negated, nothing to be

affirmed or grasped; for it never can be conceived.
(KD 35, Guenther 1969, 70)
As deluded beings, however, we believe – either instinctively or on the basis of reflection –

that there must exist a true self – either “in here” or “out there” – that is the real, metaphysical
foundation of ourselves and the world. We are reinforced in this belief by the immediacy of
our senses, the persistence of our body and mind, and the apparent precision and coherence
of our language and concepts. But all these deceive us: nothing has a real foundation, no such
self will ever be found without or within, and all our fine-spun theories are vain exercises in
imputing complexity where there is only simplicity, externality where there is only mind, and
128
Saraha
something where there is nothing. Indeed, to be undeceived, we must simply recognize, with
Jack Kerouac, that the Buddha is “philosophy’s dreadful murderer” (Kerouac 1959, 179).
Saraha as Mystical Philosopher

Our discussion thus far might make Saraha seem like a nihilist, as Nietzsche and Nāgārjuna
appeared to critics in their own eras. However, just as Nietzsche and Nāgārjuna turn out on
closer inspection not to be nihilists but simply radical thinkers who use their critiques to clear
the field for a fresh way of seeing and being in the world, so Saraha’s rhetoric of critique and
negation reflects his attempt to cut through conventional thinking and practice so as to lay bare
the fundamental transrational reality that must be realized directly if we are to be free. In this
respect, Saraha is like the mystics of many traditions, who insist that the divine only can be
reached by silencing the deluded, dichotomous chatter of words and ideas and letting it shine
on its own. And, like other mystics, he must paradoxically assert that the divine is utterly
beyond conception or description, but because it is the most important thing we can possibly
know, it must somehow be indicated – or perhaps sung – so that we can experience it, not
through a miasma of concepts, but face-to-face.
Mind and Reality in Saraha

As we have seen, the single thing Saraha says we must understand goes under many names,
but most often, perhaps, it is simply “mind” (citta). Mind here refers not to a set of capricious
emotions, rational operations, fanciful or deliberate imaginings, or “experiences” of any sort,
but mind as it really is – pure, empty, luminous, aware, ecstatic, and free. As Saraha puts it in a
famous verse, “the single seed of everything is mind, where existence and nirvāṇa both arise”
(PD 41, Jackson 2004, 73). His assertion can be read several ways. The most obvious would
be that all things are the result of the karma of beings, and karma, being mental “intention”
(cetanā), is really “mind.” By another interpretation, Saraha is claiming – in the phenomenalist
manner of Asaṅga, Vasubandhu, and other figures in the Yogācāra school – that everything in
some sense is mind, or at least cannot be known outside mind. He sings, for instance, of how
All Buddhas, sentient beings, and phenomena

arise simultaneously with mind itself.
(BT 48, Braitstein 2014, 135)
More specifically,
Mind, intellect, and the formed contents of mind are it,

So too are the world and all that seems from it to differ,
All things that can be sensed and the perceiver,
Also dullness, aversion, desire, and awakening.
(KD 33, Guenther 1969, 69)
Saraha’s assertion that everything in saṃsāra and nirvāṇa – even our delusions – partakes
of the primordial purity, emptiness, luminosity, awareness, ecstasy, and freedom of the ulti-
mate raises several questions. First, how do we stray from our own original nature? How
could something essentially pure ever be tainted by impurity? On one level, the answer should
already be clear. Our senses, body, and intellect present reality to us in deeply misleading ways:
129
Roger R. Jackson
as multiple rather than single, defiled rather than pure, substantially existent rather than empty.
To return to our original nature, then, seems to require transcending the body, senses, and intel-
lect – for in doing so, we eliminate the main basis for fixation on self, and what remains is just
the primordially pure and blissful mind itself. On a different and deeper level, however, we
never have departed from our original nature since, as Saraha repeatedly assures us, all things –
including the body, senses, and intellect, and even delusion and suffering – are “it,” or “that,” or
“the real.” A further question therefore arises: what is the point of trying to regain our original
nature when we have never lost it to begin with? If we all already are buddhas, why practice the
path? And if everything already partakes of the ultimate, why restrain the body and senses, why
seek to transcend intellect, why bother to behave ethically? A further question inevitably arises:
if all things are empty, is the primordial nature insisted upon by Saraha – the pure, luminous,
blissful mind – also empty or, as his language sometimes suggests, is it an exception, the one
thing that truly exists? Saraha is equivocal on such matters, delighting in paradox and apparent
contradiction, but we will briefly attempt to tease out some possible answers.
In attempting to understand Saraha’s paradoxes, the key, as is often the case in Mahāyāna
Buddhism, is an appreciation for the interplay between the two truths: the ultimate ( parmārtha)
and the conventional or phenomenal (vyāvahārika, saṃvṛti). Although these are variously defined
and distinguished by different thinkers, conventional truth typically corresponds to the more or
less reliable everyday deliverances of the senses and intellect, while the ultimate denotes the
real and abiding nature of persons and phenomena. Mahāyāna thinkers generally concurred that
the ultimate nature of things is emptiness but disagreed over what emptiness is empty of. Phi-
losophers associated with the Mādhyamika school, such as Nāgārjuna and Candrakīrti, typically
described emptiness as a “non-affirming negation” ( prasajya-pratiṣedha): the lack – in any entity
or concept – of intrinsic, self-sufficient existence, with emptiness itself being the ontological con-
dition for the possibility of the functioning of the environments, things, persons, and minds that
constitute the conventional world. A different view was taken by proponents of Yogācāra and the
buddha–nature doctrine, who typically regarded the emptiness of worldly things to be their lack of
difference from the mind perceiving them and the emptiness of buddha mind to be an “affirming
negation” ( paryudāsa-pratiṣedha): its lack of anything other than buddha mind, which itself is
replete with eternally perfect qualities.8 Philosophers may attempt to draw clear lines between the
two truths, but many Mahāyāna masters – from the authors of the Perfection of Wisdom sūtras, to
Nāgārjuna, to great tantric adepts like Saraha – switch between ultimate and conventional levels
of discourse without notice, leaving readers to fend for themselves amidst the apparent contradic-
tions in such statements as Rāhulabhadra’s celebration of the goddess Prajñāpāramitā:
Seeing you, one is bound to suffering,

not seeing you, one is also bound.
Seeing you, one is freed from suffering,
not seeing you, one is also freed.
(verse 15, Strong 2002, 142)
Returning to the previously mentioned questions posed to Saraha, we might note first that he
gives no real explanation of how or why our original purity has been tainted by delusion: it is
simply the way things are – a mystery, perhaps. As to whether we ought to bother practicing
the path if we all already are buddhas, a possible answer is supplied through the concept of two
truths: yes, we all are eternally in possession of perfect realization, but – for whatever reason –
we have fallen asleep spiritually and do not recognize our original, undimmed enlightenment.
Thus, while we may not really be deluded, we are conventionally in deep distress – and amidst
130
Saraha
our distress, conventional notions of bondage and liberation, pure and impure, and good and
evil, must be applied, and we must be induced to recognize our original nature as if we had
strayed from it. The concept of two truths also helps us understand what it means to say, as
Saraha does, that ultimate reality is at once empty and real. If, as emphasized in Yogācāra and
buddha–nature circles, the emptiness of the primordial mind is an “affirming negation,” then
the mind’s emptiness need not contradict its eternal perfection, since emptiness here simply
means the mind’s emptiness of anything other than its own perfection – a view that would
be referred to in Tibetan Buddhism as “other-emptiness” (gzhan stong). Although Saraha
seems naturally aligned with this perspective, proponents of the opposing “self-emptiness”
(rang stong) position could and did read his verses as suggesting that the primordial mind is
empty of self-existence, just as everything else is, and that his apparent assertion of the mind’s
true existence must, like similar claims in buddha–nature literature, be taken as inspiring and
important conventional claims, but conventional claims nevertheless.
The Path to Realization

Given that, conventionally speaking, we must return to our original nature, how do we do so?
From early on, Buddhists have discussed whether awakening is achieved suddenly (the “subitist”
position) or in progressive stages (the “gradualist” position). When the discussion has become a
debate – as in late first-millennium India, China, and Tibet – subitists have insisted that awaken-
ing depends solely or primarily on seeing reality directly, and that no amount of conventional
philosophizing or gradual practice will bring it about, so we must simply drop concepts and
realize our original nature at a single blow. Gradualists have countered that sudden awakening,
while not impossible, requires sharp intelligence and prior training in this or a previous life, and
that prematurely dropping such concepts as self and other, saṃsāra and nirvāṇa, and good and
evil, is extremely dangerous for those who are unprepared – the vast majority of practitioners.
The Subitist Approach

Saraha himself acknowledges that “a yogi desiring great gnosis/may proceed by stages or
at once” (QD 80, Jackson 2012, 184), but at first glance, he seems partial to subitism. He
declares, for instance, that “the great seal is instantaneous full awakening” (BT 71, Braitstein
2014, 139), and in a verse that is virtually a charter for meditation on the great seal, he sings,
If mind points to mind, concepts are held in suspension;

just as salt dissolves in water, so mind dissolves into its nature.
( PD 78b, Jackson 2004, 98)
A key way for mind to see mind is to “make the mind like space” (PD 42a, Jackson 2004, 74)
by dropping concepts and resting within it: “this mind so tightly bound – relax it, and you’re
free, no doubt” (PD 42h–43, Jackson 2004, 78). Similarly,
the free mind settles into its nature

just as it’s always been: without need to label or examine.
(QD 74, Jackson 2012, 183)
For the most part, Saraha’s subitism is straightforward: do not think or do anything, he says,
and do not meditate in the usual sense; just see your own mind – and everything else – as the
131
Roger R. Jackson
pure, empty, blissful reality they always have been. In several texts, however, he introduces
a quasi-systematic element to his subitist instruction through a discussion of four symbolic
terms: recollection, non-recollection, the unborn, and beyond-intellect. The first two are the
most crucial, for they describe the way we misconstrue reality and the way we can overcome
our error. Recollection, or mindfulness (smṛti), is a venerable term in accounts of Buddhist
meditation: typically, it is re-cognition of a focal object when one has lost concentration. Here,
it connotes any deliberate application of the intellect, which, as we have seen, is anathema for
Saraha. The way we eliminate it is by “non-recollection” (asmṛti): simply dropping thought
and letting our original nature shine forth. A more or less synonymous term used by Saraha is
cognitive disengagement (amanasikāra), which negates the classical Buddhist mental factor
of “cognitive engagement” (manasikāra), whereby we intentionally bring some object within
the purview of cognition. As with non-recollection, cognitive disengagement points to our
need to immediately and completely abandon intellect and enter directly into a state whereby
we see reality as it is. The third and fourth symbolic terms are simply elaborations on the
characteristics of this reality: “the unborn” indicates that our original nature is beyond either
origination or cessation, while “beyond-intellect” indicates that this unarisen reality is beyond
all analytical comprehension or verbal predication.
Gradual Paths
The subitist approach seems simple, but it is deceptively so, and Saraha concedes as much
by presenting a variety of more gradual practices that might lead to the same realization. The
major context for these gradual practices is the “completion stage” (saṃpannakrama) tech-
niques required by the esoteric and sometimes transgressive Mahāyoga and Yoginī tantras.
Although Saraha mocked such practices because they could be misconstrued and misapplied
by the unwary, they are in fact the key to realization for that majority of practitioners unsuited
to subitism. Because completion-stage practices are carried out within the body, we may
read Saraha’s body-positive verses – e.g., “what’s bodiless is hidden in the body;/know this,
and you’re freed there” (PD 89, Jackson 2004, 104) – as implicitly referring to advanced
yogic techniques involving the 72,000 channels (nāḍī) that interpenetrate our coarse physical
body, the “hormonal” drops (bindu) and breath-related energies ( prāṇa) that move within the
channels, and the energy centers (cakras) found where the outer channels intersect with (and
typically constrict) the all-important central channel. The way we are freed within the (subtle)
body is through practices that involve moving wind-related energies from the outer channels
into the central channel, and then gaining proficiency in moving energies and drops up and
down there, opening up the cakras, and inducing various states of gnostic bliss. Typically, the
sequence culminates in the dissolution of the energies, drops, and awareness itself into a drop
at the heart cakra, where we purify the subtlest mental and physical aspect of ourselves and
create the buddha bodies that complete the Mahāyāna path. There are various ways of describ-
ing these culminating practices, and while Saraha does not lay out any particular system or
sequence, he alludes to certain operations, such as making the mind like space and dissolving
the winds and other elements within the subtle body until we attain “utmost great bliss” (PD
42b–c, Jackson 2004, 75).
Elsewhere, he refers to techniques of sexual yoga that help induce the movement of ener-
gies into and within the central channel. In the People Dohās (verses 84–86, Jackson 2004,
101), he celebrates a mysterious yoginī who may be a goddess he worships or his flesh-and-
blood partner – or both. In his Queen Dohās, he details the qualities possessed by a proper con-
sort (QD 41–46, Jackson 2012, 179) and specifies that the tantric yogī “who’s . . ./aware that
132
Saraha
all things are suchness. . . ./ . . . sees great bliss during sex” (QD 39–40, Jackson 2012, 178).
Elsewhere, he refers to the consort as an “action seal” (karmamudrā), one of four seals – or
confirmations – of spiritual practice described in the later tantric systems (BT 79–96, Braitstein
2014, 140–44).9 While Saraha accepts the necessity and efficacy of the action and other seals,
it is clear that for him the great seal (mahāmudrā) – the unmediated, complete, and final reali-
zation of the nature of mind and reality – is supreme, and perhaps sufficient on its own (BT 81,
ST 31, 34–35, Braitstein 2014, 141, 157, 158).
In later tantric formulations, the completion stage can only be approached when one has
mastered the “generation stage” (utpattikrama), a complex system of practices – a sādhana
– in which we change how we see ourselves and the world, visualizing ourselves and others as
divine buddha–deities, regarding all words and sounds as mantras, and imagining our environ-
ment as a pure land or maṇḍala. Saraha asserts that “the stages of generation . . ./are to the stage
of completion like little stars to the moon and sun” (ST 32, Braitstein 2014, 157) and repeatedly
dismisses the “yoga with signs” involved in the generation stage (e.g., ST 45, Braitstein 2014,
160), yet he often refers to mystic syllables, maṇḍala circles, and other elements of sādhana
(e.g., PD 90, 90a, 98, Jackson 2004, 104, 105, 109), making it likely that his subtle-body yoga
practice was indeed preceded by the re-envisioning required by the generation stage. To prac-
tice the generation stage, in turn, requires an esoteric initiation, or empowerment (abhiṣeka).
Saraha rarely recommends, and occasionally denigrates, this linchpin of tantric practice, but
because his main frame of reference remains the completion stage, which typically presup-
poses both initiation and sādhana practice, he assumes it in his discourse. And in fact,
A person deprived of the precious initiations

is like a lowly servant aspiring to be king.
(QD 16, Jackson 2012, 178)
The source of initiation – the sine qua non for receiving it – is the guru, whose importance as
initiator, instructor, and object of devotion cannot be overstated. As Saraha asserts, “in your
reverence for the guru, the discipline of the secret vehicle is complete” (BT 34, Braitstein
2014, 132). More fully,
Through the guru, teachings, transmissions, commentaries, and special instructions are
shown.
You will attain your own intrinsic identity . . .
through pure special instructions from the guru;
if you worship the guru, you will obtain supreme connate bliss.
(ST 5, Braitstein 2014, 194; cf. trans. 151)
In the Queen Dohās, Saraha describes the guru as “the king of physicians,” “the sublime
boat that delivers us/from saṃsāra’s deep, vast sea,” “the mighty boatman who’s gained great
bliss/ . . . possessed of sun-like gnosis,” “the wheel-turning king who’s wise in the ways/of
changing all dharmas to bliss as elixirs turn metals to gold,” and the one whose “mind, like a
river, submerges dichotomous views” (QD 53–55, Jackson 2012, 180–81). In short, “the root
of every attainment is the adamantine master” (QD 62, Jackson 2012, 181). Elsewhere, Saraha
insists repeatedly that the only reliable basis for realization of our original nature is the guru’s
“ambrosial” instruction. Without it, he says, we are lost; with it, we have access to “that,” the
connate nature, the rootless real, our inmost nature, our inmost thought, our purified awareness
(PD 17, 29, 37, 38, 39, 69, respectively).
133
Roger R. Jackson
In principle, a disciple should only accept a guru who is pure, insightful, compassionate,
and experienced on the path. Conversely, a guru is only supposed to initiate or instruct a dis-
ciple who already possesses the qualities emphasized by non-tantric Buddhism: restraint of
the senses, detachment from saṃsāra, a desire for spiritual liberation, an intellectual and con-
templative appreciation of the nature of reality, and profound compassion. Saraha instructs his
audience not to be bound by sensuous things, noting that the connate nature is found “where
senses subside and self-nature cannot stand” (PD 71, 29, Jackson 2004, 94, 68); and in one of
his most oft-quoted passages, he sings,
Rejecting compassion, stuck in emptiness,

you won’t gain the utmost path.
Nurturing compassion all by itself,
you’re stuck in rebirth and won’t win freedom.
But if you’re able to join emptiness with compassion,
you won’t remain in existence – or nirvāṇa.
(PD 15a–b, Jackson 2004, 61)
Thus, a disciple’s suitability for initiation is – at least in principle – dependent upon previous
mastery of basic Buddhist virtues. The tantric path is difficult and potentially dangerous and,
transgressive rhetoric aside, actually requires an even greater degree of renunciation, under-
standing, and compassion than non-tantric practices, lest the disciple be overwhelmed by the
mental and physical forces evoked in the Vajrayāna. For Saraha, the guru is important above
all because they point out our original nature, which we recognize either through receiving ini-
tiation and undertaking various tantric yogas or perhaps by the guru’s blessing and instruction
alone. In this sense, the guru may open the way either to the gradual practices of the Vajrayāna
path or to a sudden approach to awakening that short-circuits the slow accretion of virtues in
favor of attaining the qualities of a buddha all at once.
The Fruits of the Path

Whether approached suddenly or gradually, the final fruit of the path is buddhahood, which
was subdivided by Mahāyānists into variously-enumerated “bodies” (kāya). Saraha admits
four bodies: (1) a truth body (dharmakāya) consisting of a buddha’s pure and perfect mental
qualities; (2) a celestial enjoyment body (sambhogakāya) that instructs high-level bodhisat-
tvas in Mahāyāna theory and practice; (3) an emanation body (nirmāṇakāya) consisting of
various manifestations for the sake of ordinary sentient beings; and (4) an essential body
(svabhāvikakāya), which may signify either the unity of the first three bodies or the empty
nature of the truth body. Saraha refers to these buddha bodies in the Queen Dohās:
Non-dichotomy devoid of “is” and “isn’t” is the truth body,

its essential bliss is the great-enjoyment body,
its manifold guises for wandering beings are the emanation body,
and the very gnosis that knows they’re indivisible is the essential self of them all.
(QD 63–64, Jackson 2012, 182–83)
Saraha does not mention the buddha bodies explicitly in the People Dohās, but expresses their
nature and function metaphorically in a series of verses that bring together his poetic gifts and
134
Saraha
delight in paradox so as to emphasize the unity, emptiness, serviceability, and ineffability of

the awakened state:
The precious tree of nondual mind spreads through the triple world;
it bears compassion flower and fruit, though there is no other or doing good.
The precious tree of emptiness blooms compassion, various and many hued;
another’s pleasure is its final fruit: mind intent on others’ joy.
The precious tree of emptiness lacks compassion; it has no root or branches,
so if you prattle about it, your troubles will only grow.
It’s just one, just one tree – that’s why the fruit is one.
Knowing it can’t be split, you’re free from existence – and from nirvāṇa.
(PD 108–110, Jackson 2004, 114–15)
Conclusion: Saraha, Buddhism, and Philosophy

The anti-philosophical strain in Saraha’s rhetoric might lend the impression that he has noth-
ing to do with traditional Buddhist thought, but it should be evident by now that he is deeply
indebted to the Buddhist thinkers preceding him. His critical spirit is actually quite akin to that
of many Indian Buddhists, including Śākyamuni himself, Nāgārjuna, and many of his fellow
great adepts. From a positive standpoint, he accepts conventionally such basic Buddhist notions
as saṃsāra and nirvāṇa, the four noble truths, and the necessity to transcend self-grasping for
the sake of liberation. He is equally beholden to Mahāyāna thought, sharing with Great Vehicle
traditions such ideas as emptiness, non-duality, mind-only, buddha nature, and the bodhisat-
tva’s need to master wisdom and compassion. More specifically, Saraha reflects the ideas and
images developed by three major Mahāyāna philosophical traditions. Madhyamaka influence
is evident from his repeated references to emptiness – however construed – as the nature of
things and his use of paradox to undermine our certainty about the categories through which
we encounter and attempt to control reality. Yogācāra influence is evident in his strong empha-
sis on the metaphysical primacy of mind and the importance of attaining a non-dichotomous
understanding of its true nature. The influence of buddha-nature discourse is clear from his
consistent assertion that all beings not only can become buddhas but in a basic sense already
are buddhas because their minds are naturally pure, luminous, and empty of all afflictions. To
these ideas derived from early and Mahāyāna Buddhism, Saraha adds distinctively Vajrayāna
elements, including reliance upon the guru, celebration of the body and senses, an emphasis
on idealized female figures, and techniques for attaining buddhahood within the subtle body.
So, is Saraha rightly considered a great Indian Buddhist philosopher? It depends, of course,
on what we mean by philosophy. If using conventional logical analysis and rational argu-
mentation are the sole criteria, then the answer is no – his approach is a far cry from that of
Nāgārjuna or Dharmakīrti. As Herbert Guenther notes, however, the tantric ethos forces us
to reconceive philosophy as the attempt “to know reality directly, not by rumor or descrip-
tion . . . [but as] a way of life and quest for meaning [involving] the whole of man and not
merely his brain” (Guenther 1963, 112, 113). This description easily could be applied to all
of Buddhism – and indeed, to wisdom traditions everywhere. Such an approach – in which
theory and practice are united in pursuit of human freedom – in turn suggests that figures like
Nietzsche in the West and Saraha in India may help us to think about philosophy not as the
allegedly dispassionate application of rational methods for the sake of arriving at some sup-
posedly objective truth, but rather as the poetic articulation of a unified vision of reality and the
135
Roger R. Jackson
potential humans have for attaining ultimate wisdom, joy, and goodness within that reality. As
we have seen, the reality Saraha sings is the pure, empty, luminous awareness that is our origi-
nal nature, and in some sense the original nature of everything, with the consequence that the
purpose of life is to realize – to reacquaint ourselves with – the basic nature we never have lost.
This sense of philosophy may seem distant from the meaning of the term in recent Anglo-
American thought, but it finds intriguing echoes in the work of twentieth-century Continental
European philosophers. Martin Heidegger (1889–1976), for instance, expressed an ongoing
concern with the question of being and emphasized in his later works the importance of replac-
ing “calculative thinking” with a “meditative thinking” that “contemplates the meaning in eve-
rything that is” through “releasement toward things and openness to the mystery” (Heidegger
1966, 46, 65). Pierre Hadot (1922–2010) undertook a radical reconsideration of the ancient
Epicureans, Skeptics, and Stoics, seeing them not as mere logicians but as thinkers concerned
with how to live a meaningful life through the practice of “spiritual exercises” aimed at self-
transformation. He also argued that these thinkers’ use of rational argumentation was just one
of many techniques conducive to humans’ attunement to themselves and the cosmos (Sharpe
n.d.). And, if philosophy as a rational method is merely one of multiple ways to approach and
identify reality, then perhaps poetry like that of Saraha (or Rumi, or Rilke, or Stevens) may be
just as powerful – or even more powerful – a means to that end as reasoning, and just as “philo-
sophical” as the syllogisms of learned scholars. Heidegger sets them on the same lofty plane
in proclaiming that “the thinker utters being; the poet names what is holy” (Heidegger 1968,
391) – and in Saraha, we see the two roles coalesce so completely that we are forced to recon-
sider both poetry and philosophy, and ask whether, at the highest level, they really differ at all.
Notes
1 For modern scholarship on Saraha, see especially Guenther (1969, 1993); Jackson (2004, 2012),
Schaeffer (2005); Shahidullah (2007); Braitstein (2014).
2 For editions, see Jackson (2004, 53–115) and Kvaerne (1977, 168, 222, 226), respectively.
3 These are the Performance Song Dohā Treasury (Dohākoṣa-nāma-caryāgīti) and Instruction Song
Dohā Treasury (Dohākoṣopadeśagīti), translated in Jackson (2012, 173–84) and Guenther (1969,
63–71), respectively.
4 The full titles are A Body Treasury Called the Immortal Adamantine Song (Kāyakośāmṛtavajragīti), A
Speech Treasury Called the Gentled Voiced Adamantine Song (Vākkośarucirasvaravajragīti), and the
Unborn Treasury of Mind Adamantine Song (Cittakośājavajragīti); for editions and translations, see
Braitstein (2014, 125–208).
5 In this and most other cases, I have slightly modified existing translations for the sake of terminologi-
cal and stylistic consistency.
6 A parenthetical abbreviation followed by a number indicates the verse number in a given collection.
7 In citations from my translations of the Apabhraṃśa (but not the Tibetan) of Saraha’s People Dohās,
I have restored the line length found in the original.
8 The terms for the two types of negation were not coined until later in the Indian tradition, but may
reasonably be read back into the writings of early figures.
9 The four are the action seal (karmamudrā), pledge seal (samayamudrā), dharma seal (dharmamudrā),
and great seal (mahāmudrā).
References
Braitstein, Lara. 2014. The Adamantine Songs (Vajragīti) by Saraha. New York: American Institute of
Buddhist Studies.
Guenther, Herbert. 1963. The Life and Teaching of Nāropa. London: Oxford University Press.
———, trans. and annot. 1969. The Royal Song of Saraha: A Study in the History of Buddhist Thought.
Seattle: University of Washington Press.
136
Saraha
———. 1993. Ecstatic Spontaneity: Saraha’s Three Cycles of Dohās. Berkeley: Asian Humanities Press.
Heidegger, Martin. 1966. Discourse on Thinking. Translated by John M. Anderson and E. Hans Freund.
New York: Harper and Row.
———. 1968. “What Is Metaphysics?” Translated by R. F. C. Hull and Alan Crick. In Existence and
Being, edited by Werner Brock, 353–92. London: Vision.
Jackson, Roger R. 2004. Tantric Treasures: Three Collections of Mystical Verse from Buddhist India.
New York: Oxford University Press.
———. 2012. “Saraha’s Queen Dohās.” In Yoga in Practice, edited by David Gordon White, 162–84.
Princeton: Princeton University Press.
Kerouac, Jack. 1959. Mexico City Blues. New York: Grove Press.
Kvaerne, Per. 1977. An Anthology of Buddhist Tantric Songs: A Study of the Caryāgīti. Oslo:
Universitetsforlaget.
Schaeffer, Kurtis R. 2005. Dreaming the Great Brahmin: Tibetan Traditions of the Buddhist Poet-Saint
Saraha. New York: Oxford University Press.
Shahidullah, Muhammad. 2007 [1928]. The Mystic Songs of Kāṅha and Saraha: The Dohā-koṣa and the
Caryā. Translated by Pranabesh Siṇha Ray. Biblioteca India no. 329. Kolkata: Asiatic Society.
Sharpe, Matthew. n.d. “Pierre Hadot.” Internet Encyclopedia of Philosophy. Accessed March 19, 2022.
https://iep.utm.edu/hadot/.
Strong, John S. 2002. The Experience of Buddhism: Sources and Interpretations. 2nd ed. Belmont, CA:
Wadsworth.
Tāranātha, Jonang. 1983. The Seven Instruction Lineages. Translated and edited by David Templeman.
Dharamsala: Library of Tibetan Works and Archives.
137
PART 3
Abhidharma Philosophers
ABHIDHARMA PHILOSOPHERS
The Sanskrit term “Abhidharma” (“Abhidhamma” in Pāli) has multiple meanings. It often
refers to a corpus of texts, one of the “three baskets” (tripiṭaka; tipiṭaka) of the Buddhist
canon. In contrast to the discourses (Sūtra, Sutta) and the monastic code (Vinaya) – the other
two “baskets” of the canon – Abhidharma literature systematically analyzes forms of con-
scious experience and their objects in an explicitly Buddhist philosophical framework. One of
the distinguishing features of the Abhidharma texts is the comprehensive account of mental
and physical factors authors try to achieve, often in the form of lists. There are two complete
canonical collections of Abhidharma texts extant today – one preserved in Pāli claimed by the
Theravāda tradition, and one preserved in Chinese ascribed to the Sarvāstivāda tradition – but
previously there were additional Abhidharma texts associated with other schools, some of
which exist in fragments while others are completely lost. The term “Abhidharma” can also
refer to commentaries on these canonical texts, as well as to other texts that are intended to
provide a systematic analysis, metaphysics, phenomenology, and evaluation of the varieties of
consciousness and its objects using the earlier Abhidharma categories. Abhidharma can thus
be a way of characterizing a style and approach to doing philosophy in Buddhist traditions,
one of the first manifestations of Buddhist scholasticism. What these diverse meanings of
the term share is that Abhidharma takes up the teachings from Buddhist discourses and gives
a more formal and methodical account of Buddhist doctrine. While Abhidharma developed
in the early centuries of Buddhism in India, later thinkers, including those associated with
the Mahāyāna traditions of Madhyamaka and Yogācāra, were working within an Abhidharma
philosophical context, and many of them wrote Abhidharma texts. This is why Abhidharma
can be seen as the backbone of so much Buddhist philosophy in India and remains a constant
reference for Buddhist philosophers.
In the first chapter of this part on Abhidharma philosophers, Maria Heim describes the
Enumeration of Phenomena (Dhammasaṅgaṇī) and Analysis (Vibhaṅga), which constitute the
first two books of the canonical Pāli Abhidhamma. Most traditional Theravāda commenta-
tors, as well as modern scholars, have interpreted the Pāli Abhidhamma as an ontology of
ultimate reals. Heim defends an alternative interpretation for which she finds evidence in the
Abhidhamma as well as in some postcanonical commentary, particularly that of Buddhaghosa.
According to this interpretation, the Abhidhamma is a method of unending phenomenological
analysis, a practice of continually exploring and then transforming experience. Heim uses the
DOI: 10.4324/9781351030908-15 141

Abhidharma Philosophers
canonical Abhidhamma texts to demonstrate how they invite practitioners into a therapeutic
and soteriological process of transformation through the endless exploration of experience that
demonstrates that there is no final ontology, no final account of ultimate reality.
The next chapter, by Rupert Gethin, focuses on Points of Discussion (Kathāvatthu), the
fifth of the seven Abhidhamma texts collected in the Pāli Canon. The text is therefore regarded
as “the word of the Buddha” (buddhavacana), even as it is also attributed to a particular figure,
Moggaliputta Tissa, who is said to have lived in the third century BCE, because he is thought
to be correctly explicating the Buddha’s teachings in the work. Through dialogues between
one voice advocating the views of the text and another voice presenting the opponent’s views,
Points of Discussion addresses more than 200 points of contention in early Buddhist doctrine.
The dialogues follow a structure of reductio ad absurdum arguments, in which the opponent
comes to understand that their beliefs are wrong because they are inconsistent. We are thus
able to gain an understanding of both what some Buddhists were arguing about and the meth-
ods of argumentation they were employing. Gethin discusses these methods and then focuses
on several particular points of doctrine debated in the text.
As Buddhism entered new geographical locations in the Indian subcontinent in its early
centuries, different groups formulated distinctive interpretations of the Buddha’s teachings.
Abhidharma texts were composed in part to articulate and distinguish the doctrinal positions of
diverse monastic communities. In his chapter, Bart Dessein discusses Kātyāyanīputra (ca. first
century CE) and his Abhidharma Treatise in Eight Sections (*Abhidharmāṣṭagranthaśāstra or
*Abhidharmāṣṭaskandhaśāstra), also known as the Source of Knowledge (Jn͂ ānaprasthāna).
Kātyāyanīputra is often regarded as the founder of the Sarvāstivāda school of Buddhism, and
his text played a central role in the formation of the Sarvāstivāda identity with its own philoso-
phy and an accompanying understanding of the Buddhist path. In the Abhidharma Treatise, he
developed a comprehensive analysis of the material world and defended the role of reason and
insight on the Buddhist path. Dessein presents some of Kātyāyanīputra’s original contributions
to Buddhist philosophy and traces two distinct developments that emerged from his work: the
pedagogical digests that articulate the essence of Sarvāstivāda views compiled in Bactria and
Gandhāra, and the commentaries on the Abhidharma Treatise composed by the Vaibhāṣikas
of Kaśmīra.
In the final chapter in this part, Peter Skilling discusses what he terms the “Vātsīputrīya/
Saṃmitīya” complex of schools – including the Bhadrayānīya, Dharmottarīya, Ṣaṇṇagarika,
and Kaurukulla-Sāṃmitīya – which are often referred to by their opponents as “Pudgalavādins.”
One of the many views these traditions held was that the person ( pudgala) was neither dif-
ferent from nor the same as the constituent elements of sentient beings, a position that many
other Buddhists critiqued as an abandonment of the traditional Buddhist understanding of
“non-self.” While the Vātsīputrīyas/Saṃmitīyas probably did not think of this one idea as
defining their tradition – the Saṃmitīya metaphysical texts surviving in the original language
do not address the question of the pudgala – their opponents did. The Vātsīputrīya/Saṃmitīya
were absorbed by other traditions by the sixteenth century, and for many years they have
been regarded as not genuinely Buddhist. In contrast, drawing on a variety of sources, Skill-
ing argues that for well over a millennium the Vātsīputrīya/Saṃmitīya schools were in fact
a dominant form of mainstream Buddhism in South Asia. With this in mind, he makes the
broader point that we should avoid thinking of one or another Buddhist school as somehow
original, orthodox, or heterodox, and instead recognize the dynamic changes that have always
marked the development of Buddhist thought.
142
9
THE DHAMMASAṄGANĪ AND
VIBHAṄGA ˙
The Perfectly Awakened Buddha and the First
Abhidhammikas
Maria Heim
The Enumeration of Phenomena (Dhammasaṅgaṇī) and Analysis (Vibhaṅga) are the first two
books of the Pāli Abhidhamma. They are attributed by the tradition to the Buddha (in an
account that I will consider briefly in what follows) and, along with the other five books of
the Abhidhamma, are considered to be canonical; all seven books comprise the third collec-
tion of early Buddhist texts in the Theravāda called the Abhidhamma Piṭaka (the other two
piṭakas, or genres of Buddhist knowledge, are the Suttanta, the discourses, and the Vinaya, the
monastic rules). There are several good summaries of all seven Abhidhamma books, and I will
make no attempt to recap their lengthy contents here (see Nyanatiloka 2008). While traditional
exegetes did not consider historical questions of development, holding as they did that all of
this material came from the Buddha (even if mediated by disciples and other scholars), mod-
ern scholars have tended to see the Enumeration of Phenomena and Analysis as the earliest
to emerge. Both texts also have extensive layers of exegesis on them, at both the commentary
(aṭṭhakathā) and subcommentary (ṭīkā) layers, as well as medieval compendia (saṅgaha) that
attempt to synthesize and summarize the Abhidhamma tradition as a whole. The Abhidhamma
has been influential throughout Theravāda history, particularly in Burma/Myanmar, where a
lively scholarly tradition on it continues to the present day.
In this chapter, I focus on these first two books in order to concentrate on this earliest itera-
tion of what became a long and rich philosophical tradition. My concerns are philosophical
rather than historical, and I will read these texts to ask, first, what they themselves tell us
they are doing, and second, how their philosophical implications have been understood by
traditional and modern scholars. There have been widely divergent views on what this early
philosophical tradition is about, ranging from those who see it as an ontological system posit-
ing ultimate reals, to those who read it as a system of phenomenological analysis to investigate
and transform experience. While I will show the nature of the scholarly arguments for both
interpretations, my own view tends to the latter reading, and I demonstrate why. I also offer
additional philosophical possibilities that take special notice of the potential contributions sug-
gested by the early Abhidhamma as a modal analysis of experience.
I should note that “Abhidhamma” means the “higher” or “further” Dhamma, or teaching,
in that it offers additional development of the basic doctrines of Pāli Buddhism as they are
articulated in the suttas, such as the Four Noble Truths, the five aggregates, the twelve-fold
DOI: 10.4324/9781351030908-16 143

Maria Heim
dependent origination, the four foundations of mindfulness, etc. Readers new to Indian Bud-
dhism should become familiar with these core doctrines before attempting to understand
Abhidhamma and the interpretation of it that I advance here.
Some Preliminaries: Matrices and the Fecundity of Lists

The Pāli tradition as represented by Buddhaghosa (the ascribed commentator of the Abhi‑
dhamma texts) holds that the Abhidhamma Piṭaka was discovered and taught by the “Per-
fectly Awakened Buddha” (Sammāsambuddha). It was discovered during the fourth week of his
awakening as he attained omniscience and sat with unblinking eyes contemplating “endless and
immeasurable” methods for interpreting experience.1 These endless methods (naya) became the
Abhidhamma texts. But how could he teach endless methods? The teaching of them has various
accounts, one of which is that the Abhidhamma, being coextensive with the Dhamma itself,
was taught when the Wheel of the Dhamma was turned (the suttas contain the same teachings,
but in them the teachings are given in the contexts of particular narratives when they are taught
to particular interlocutors). Another account, one that suggests that the Abhidhamma can also
be conceived as a distinct body of material, has it that the Buddha taught the Abhidhamma to
his mother in heaven during a three-month period (where he began to teach her all the lists of
phenomena as we see them in the Dhammasaṅgaṇī), while also coming down to earth peri-
odically and teaching it to his disciple Venerable Sāriputta. When teaching to Sāriputta, “the
foremost disciple renowned for analysis,” he gave the method “much like one gesturing to the
ocean that is seen by stretching out one’s hand while standing at the shore.”2 That is to say,
the Buddha initiated through a gesture the oceanic methods of the Abhidhamma to a disciple
skilled in analysis who could expand on them. In this reading, the Abhidhamma is the begin-
ning of a series of methods and listings of formulas for analyzing experience that is inherently
generative, indeed “endless and immeasurable.” These considerations, while perhaps fanciful
from a modern point of view, may be instructive as we begin to interpret the prolixity of lists in
this genre of canonical teachings and how we might conceive of them as methods.
Indeed, the Abhidhamma texts consist mostly of lists. A key term for understanding Abhi‑
dhamma lists is mātikā, matrix, though the use of mātikā was by no means exclusive to this
genre. A mātikā is a table of contents, outline, or listing that can lead to further expository
development of the items on it. In both Pāli and English, the word is derived from “mother” and
retains the sense of being that from which something further issues (Gethin 1992b, 160–61).
Much like a professor jotting down a brief outline of lecture notes which she will then expand in
class, a Buddhist teacher would have ready (though without the jotting down since for centuries
the tradition was transmitted orally) a list which she could expand via commentary. Of course,
this can work conversely: mātikās also function as brief (saṅkitta) summaries or contractions
of teachings that are elsewhere expansive (vitthāra), and thus offer the Dhamma in nutshells.
Rupert Gethin has shown that mātikās in the Buddhist sources were used not only to flesh
out the details of items on the original list, but also to generate further lists. Lists generate other
lists until one begins to get a sense of the overall teaching in a highly systematic way. To show
this, he starts with the familiar list of the Four Noble Truths. We begin with a list of four items
(Suffering, Origin, Ceasing, and the Path), where three of these mention other lists: the Truth
of Suffering mentions the five aggregates (form, feeling, perceiving, volitional constructions,
and consciousness); the Truth of Origin mentions three types of craving; and the Truth of the
Path is, of course, eightfold. Pressing further, each of the aggregates itself subsumes more lists
(four kinds of form, three [or six] kinds of feeling, six kinds of perceiving, et cetera) and then
each of the constituents of the Eightfold Path is parsed by further lists (right view contains
144
The Dhammasaṅganī and Vibhaṅga
˙
the Four Noble Truths [here the parts contain the whole], right thought is the three wholesome
thoughts, right speech is of four kinds, et cetera). One could go on. Gethin pursues many of the
possible lists (but not all) to which the Four Truths can lead to arrive eventually at all thirty-
seven factors of awakening (and even that is an arbitrary stopping place), and an overall sense
of core Buddhist teachings.3
The utility of lists in an oral culture relying on memory is obvious and often noted. But
further, Gethin argues, this capacity of lists to subsume and generate other lists was invaluable
for teaching and understanding in a way that goes beyond rote learning because “the lists help
one learn the Dhamma with a view to its inner structure and dynamic” and can “act as a kind
of flowchart for the composition of a discourse” (Gethin 1992b, 156). The matrices become
guidelines or methods for recitation and composition within the oral performative culture in
which teachings were taught and known. If any list can lead potentially to all other lists, the
Abhidhamma specialist can speak to any feature of the Dhamma and indeed the overall pat-
terns and connections of the entire system. These features lead Buddhaghosa to insist that only
Abhidhamma specialists really know how to preach the Dhamma, because they do not get
muddled.4
Another generative listing practice useful for both memorization and teaching is to order
teachings according to the lists of “twos,” “threes,” “fours,” and so on. Lists by numbers,
often in a question-and-answer format, is a very common practice in the suttas, as we see in
the “Chanting Together” (Saṅgiti) and “Expanding Tens” (Dasuttara) suttas, and of course in
the entire structure of the Aṅguttara Nikāya.5 This provided not only an orderly and systematic
teaching, a memory aid, and a collective and performative style of preaching (how many twos
can we remember and recite together?), but it also allows one to see new identifications and
relationships and to improvise within the terms of the larger system. The “Great Questions”
sutta, for example, is the improvisational expansion of the Buddha’s teaching by a pundit nun
from Kajaṅgalā. She is asked by lay people to preach on a list of questions about ones, twos,
threes, et cetera. Claiming that she had heard neither the Buddha nor the monks expand on
these particular points, she offers to explain, in detail, how she understands these questions.
She expands the lists in a way that uses the numerical schema to discuss core teachings. The
lay people honor her for this and she invites them to check her answers with the Buddha, who
confirms that he would have answered exactly as she did.6
Finally, if we are right in finding the kernel of Abhidhamma practices in some of the sutta
literature, another purpose of lists was for the Buddha to teach his disciples the skills to ana-
lyze experience from multiple angles and in multiple ways. Such aspectual and modal inter-
rogations of experience are important, both in the contemplative practices he advocates and in
teaching. The latter can be shown in “The Many Types of Elements Sutta,” where the Buddha
urges his disciples to train to become “learned” ( paṇḍita) and “inquisitive” (vīmaṃsaka). They
do so by becoming “skilled” (kosala) in the elements, bases, and dependent origination (these
fundamental teachings, some discussed in what follows, are not necessary to elaborate here for
the purpose at hand). Ānanda takes this to heart so that as the Buddha begins to give numerical
listings of the elements, Ānanda asks him, repeatedly, “but venerable sir, is there another mode
of teaching ( pariyāya) whereby one can become skilled with the elements?” In each case, the
Buddha says yes, and gives a different numerical listing. It is thus through Ānanda’s persistent
questions that we get first eighteen elements, then a different listing of how the elements can
be described by way of six, and then a different list of six, then yet another list of six, then a
breakdown of them into three, then into two.7 This exercise suggests two main points to bear
in mind as we enter this literature. First, lists of phenomena can be variable: there is no single
list of the elements because they can be divided up differently according to different methods
145
Maria Heim
of looking at them. Second, there is a valued skill in questioning to arrive at different modes of
breaking down and teaching about experience. A skillful inquisitor does not stop with one list.
The Enumeration of Phenomena

This first text of the Abhidhamma Piṭaka, the Enumeration of Phenomena (hereafter, Enu-
meration) is structured into three books. Book I begins with a matrix of twenty-two triplets
(that is, lists of threes) and one hundred pairs (lists of twos). It then takes up the first list of
triplets which are three questions: “which are the good (kusala) dhammas? Which are the
bad (akusala) dhammas? Which are the indeterminate dhammas?”8 The text itself does not
define dhamma or kusala, but at this stage in the literature, dhamma seems to mean “object of
experience” that one can observe in contemplative introspection.9 “Phenomenon” will serve
as a translation. For its part, the commentary adds the nuance that these are not essentialist or
living entities.10 As for kusala, the commentary defines it as “salutary, blameless, competent,
and producing happy results,”11 a range which I attempt to capture quite generally as “good.”
Akusala, “bad,” is the opposite of this, and indeterminate are undetermined or neutral.
Answering these three questions of this first matrix takes up half of the entire Enumera-
tion, as it proceeds to ask about specific types of moments of awareness (citta). For example,
it begins by taking up a type of good awareness that is associated with sensual desire and
accompanied by joy and knowledge.
What are the good phenomena? On whatever occasion there is the arising of a good
awareness of the sensory realm accompanied by joy and knowledge, having as its
object a (visual) form, sound, smell, taste, touch, or mental phenomenon, then at that
occasion there are: contact, feeling, perceiving, intention, awareness, initial think-
ing, sustained thinking, joy, pleasure, oneness of mind, faculty of faith, faculty of
energy, faculty of mindfulness, faculty of concentration, faculty of wisdom, mental
faculty, faculty of happiness, faculty of vitality, right view, right thought, right effort,
right mindfulness, right concentration, power of faith, power of energy, power of
mindfulness, power of concentration, power of wisdom, power of shame, power of
apprehension, non-greed, non-hatred, non-delusion, non-covetousness, non-malice,
right view, shame, apprehension, tranquility of body, tranquility of mind, lightness
of body, lightness of mind, softness of body, softness of mind, workableness of body,
workableness of mind, proficiency of body, proficiency of mind, uprightness of body,
uprightness of mind, mindfulness, meta-attention, calmness, insight, exertion, bal-
ance, and whatever other dependently-arisen formless dhammas occur on that occa-
sion – these are good phenomena.12
This particular occasion of experience can include at least these fifty-six phenomena (dham-
mas), though it need not include all of them; elsewhere, these types of dhammas are referred
to as cetasikas, psychological phenomena. The rest of Book I lists the cetasika dhammas
occurring in various types of good moments of awareness, then in the various occasions of bad
moments of awareness, and then in occasions of the indeterminate, and in every case working
these classifications of experience through the distinctions that occur in the other triplets and
pairs, as well as many others.
There are several things to notice about this first list of fifty-six dhammas. First, most of its
first few items have precedent in the “One-by-One Sutta,” which describes Sāriputta listing the
phenomena he experienced and identified “one by one” during his meditation experiences. His
146
˙
introspection revealed that, among other phenomena, contact, feeling, perceiving, intention,
awareness, initiative, resolve, energy, mindfulness, equanimity, and attention arose, endured,
and faded away in his jhānic meditation.13 While initiative (chando), decision (adhimokkho),
and attention (manasikāro) enter into these Abhidhamma lists only at the postcanonical level,
Sāriputta’s observations might be the seed of this method of cataloging the phenomena present
through introspective practice.
Second, one should notice that some of the items repeat under different modes. For exam-
ple, “mindfulness” (sati) occurs no fewer than four times: as a faculty, as a power, as “right
mindfulness” (that is, one of the Eightfold Path factors), and listed on its own. This might be
seen as unnecessarily redundant: should not a single mention of mindfulness (sati) be enough?
Buddhaghosa resists forcefully the idea that the list is baggy or ill-considered because each of
the different groupings and modalities (mindfulness as a power or a faculty or a path factor)
speaks to different aspects or workings of these phenomena. They are like artisans who work
in different guilds in which their functions and roles vary: the “same” artisan can work as a
carpenter, plumber, et cetera, for the king.14 We might say that mindfulness might be present in
a particularly strong way in a particular occasion of this type of experience, whereby it oper-
ates as a faculty governing the experience as a whole. Or in a different instance of a similar
occasion of experience, mindfulness might not function as a ruling faculty but rather as a mild
trace of a path factor. These differences suggest that these various classifications of how a
phenomenon can present were intended to be captured in the list itself. The modern interpreter
Nyanaponika Thera argues similarly that these repetitions under different aspects are not dis-
pensable (as the later summarizations of the tradition suggest) but rather indicate “the different
functions and ways of application of a single quality,” and are essentially practical in orienta-
tion for the psychological and transformative purposes at the heart of the whole business. He
cites a modern psychologist to indicate the importance of this – “in psychology a difference in
aspects is a difference in things” – and Nyanaponika explores with great nuance the implica-
tions of this aspectual or modal method of analysis.15
Another key feature of the list of fifty-six phenomena in this occasion of good awareness
is that it ends with an “et cetera.”16 The list is not described as final or complete. Indeed, all
lists of phenomena occurring in particular occasions of awareness (bad and indeterminate,
too) in this text (and there are a good many) end with “et cetera” in this way. Finally, note
that by the end of this list, these phenomena are described as “the good phenomena” (ime
dhammā kusalā). This is important because many of the same items (contact, feeling, perceiv-
ing, intention, and so on) will appear again when the text begins to list the varieties of bad and
indeterminate moments of awareness;17 there they will change normative valence and become
bad and indeterminate phenomena. Again, this feature has been noted by Nyanaponika, who
argues for the importance of these groupings for indicating how items in them are “open” and
changeable according to what else occurs on the occasion in which they arise. This builds into
the system a resistance to seeing dhammas as single, discrete, self-contained units because
their functions, intensities, roles, and valences change according to whatever else occurs in the
occasion of experience being considered (Nyanaponika 1998, 40–41).
Book II takes up matrices for analyzing form (rūpa). Here it is important to note that form
is not “matter” in the sense of the physical stuff of the world, rendered in a reductionist attempt
to get at atoms, as it is sometimes understood. The text’s many matrices defining form labor in
much the same vein of enumerations that classify and elaborate it according to single defini-
tions, pairs, triplets, and so on up to lists of eleven. Throughout, rūpa consists of dhammas
occurring always and only with other dhammas (of the other four aggregates, often grouped
together as that part of our phenomenality associated with naming [nāma] experience), and
147
Maria Heim
is analyzed as it is experienced by our six senses, karmic results, and other psychological
factors and features of experience. For his part, Buddhaghosa interprets rūpa as that part of
human phenomenality that is impacted (literally “molested,” ruppana) by sensory contact. For
example, when we begin to analyze our experience of forms, we find that the “earth element”
is experienced in terms of its hardness, how it acts as a foundation, and how it receives other
things; and the “water element” is analyzed by its flowing, spreading, accumulating, and so
on.18 This is not a physics of matter and processes that occur “out there” independent of an
observer, but rather an analysis of the human experience of them – how they are for us.
Book III is a summary, but instead of reducing or tidying up the numerous analytical matri-
ces and practices in the first two-thirds of the text, it offers yet further matrices and classifica-
tions that consider dhammas in terms of yet further modalities and aspects. These groupings
consider various kinds of phenomena in terms of their association with the causes (hetu), con-
ditions ( paccaya), hindrances (nīvaraṇa), fetters (saṃyojana), oozings (āsava), views (diṭṭhi),
forms of clinging (upādāna), defilements (kilesa), path factors (magga), and so on. In other
words, it appears that every core teaching, classification, and distinction given in the Suttanta
finds a place in the Enumeration as a mode of analysis of dhammas.
Analysis
Our second text, Analysis (Vibhaṅga), is structured rather differently than Enumeration, but
it interacts with it in interesting ways. It consists of eighteen chapters that treat, with great
analytical depth, core teachings found in the suttas. These are: aggregates, bases, elements, the
Four Truths, the faculties, dependent origination, the four foundations of mindfulness, right
striving, the bases for magical power, awakening factors, path factors, the jhānas, the immeas-
urables (that is, the divine abidings), the five moral precepts, the four kinds of discriminating
analysis, understanding, miscellaneous items, and an analysis of the “heart of the Dhamma.”19
These lists function much as we have described previously with the help of Gethin as lists
leading to further lists in a highly generative way that can develop one’s understanding of
both the intricacies of the details and the overarching structure of the teachings. The lists also
develop a modal treatment of experience as it is interpreted by those teachings.
The text is very systematic in that each chapter has three parts: Suttanta Analysis, Abhi‑
dhamma Analysis, and Lines of Questioning. The first is an analysis that picks up on classifica-
tions and distinctions mentioned in the Suttanta, but according to Buddhaghosa, this is only a
partial analysis. The fuller analysis is given in the Abhidhamma Analysis which is more elabo-
rate, looking at the category under analysis from “every side,”20 by offering matrices of single
definitions, pairs, triplets, and so on. The Lines of Questioning will be familiar from the ques-
tions of the Enumeration beginning with which items are good, which are bad, and which are
indeterminate, and going on from there to offer additional matrices of the twenty-two triplets
and one hundred pairs used in that text. So in this generative way, Analysis puts to creative
use for its groups and items the same method that the Enumeration uses to treat dhammas. Of
course, many of the items in Analysis’s teachings are themselves dhammas.
There is space only for a brief example of the first chapter, the analysis of the aggregates.
This offers first a relatively short Suttanta Analysis of each of the five aggregates: form (rūpa),
feeling (vedanā), perceiving (saññā), volitional constructions (saṅkhāra), and consciousness
(viññāṇa). The five aggregates doctrine is of paramount importance because it breaks down
and groups the entire experience or phenomenality of a human being into these five clusters
of phenomena. Beginning with form, the Suttanta Analysis breaks it down further into what is
experienced as past, present, and future – according to whether it is experienced subjectively
148
˙
or by others, whether it is gross or subtle, inferior or superior, and distant or proximate. (Notice
that these analyses of rūpa, like those in Enumerations, do not itemize the physical stuff in the
world, but rather list modalities of human experiences with this cluster of phenomena; again,
this is because rūpa refers not to matter but to one side of our phenomenality.) The Abhi
dhamma Analysis piles on more analyses according to single ways of describing form, and
matrices of pairs, triplets, and up to elevens; the Lines of Questioning runs rūpa through the
twenty-two triplets and one hundred pairs. Then the chapter takes up feeling, the second of the
aggregates, and analyzes it by the same modes in all three parts that it did with the treatment
of form. And so on for the other three aggregates.
And this is only the first chapter of Analysis – there are seventeen more. Each chapter
deploys somewhat different modalities as are appropriate to it, and anyone with an interest in
any of the eighteen teachings would be well served by exploring these chapters that analyze
it exhaustively (or exhaustingly, as the case may be). The final chapter, the “Heart of the
Dhamma,” circles back to many of the categories the whole text has covered, to ask about fur-
ther types and classifications of the aggregates, bases, elements, and so on. As in the Enumera-
tion’s final book, something that looks like it might summarize winds up elaborating further.
Philosophical Interpretations
Abhidhamma as an Ontology of Ultimate Reals

To what end, all this analysis? There is a substantial body of scholarship that treats the Pāli
Abhidhamma as, at bottom, a metaphysics or ontology of ultimate reality, and in fact we can
say that this is the dominant position in both traditional Theravāda scholarship and modern
Buddhist studies. To begin with representatives of the latter, contemporary scholars of Buddhist
philosophy have not always made much of a distinction between the Pāli Abhidhamma and the
Sarvāstivāda Abhidharma on this point, and some have assumed that what holds for the latter
also describes the former. For example, Jan Westerhoff asserts that while there emerged differ-
ent Abhidharma canonical traditions “all were united by a common core of philosophical prin-
ciples,” which include a notion of “primary existent objects” that are “ultimately real.” Because
these ultimately real objects, dharmas, are said to have an “intrinsic nature” (svabhāva), they
“exist no matter what, without depending on the existence of any other dharma or on any
conceptualizing mind” (Westerhoff 2013, 130). Other texts make the same move in conflating
the Abhidharma/Abhidhamma systems and characterizing them as offering, as The Princeton
Dictionary of Buddhism puts it in its single entry on both, “an objective, impersonal, and highly
technical description of the specific characteristics of reality” (my italics) (Buswell and Lopez
2014, 4); when this dictionary turns to the Enumeration it treats its purpose in the same vein as
“a systematic analysis of all the elements of reality” (241). While Paul Williams distinguishes
between the Sarvāstivādin tradition and the Theravāda, and suggests that the latter might be
characterized as an “event ontology” rather than a “substance ontology,” he is insistent that
Abhidhamma analysis involves “seeing things as they are, and that is a matter of ontology”
(Williams and Tribe 2000, 92). For many scholars, the matter is settled.
As we turn to scholars trained within the Pāli tradition, we find a long history of an onto-
logical reading of the Abhidhamma dating at least to the medieval compendium literature, and
likely initiated in some of the layers of exegetical tradition that precede it. Far and away the
most significant development after the first commentaries is the Compendium on the Meaning of
the Abhidhamma (Abhidhammatthasaṅgaha), attributed to the eleventh-century monk, Ācariya
Anuruddha. This text is the most widely known Abhidhamma text and has almost entirely
149
Maria Heim
supplanted the canonical literature in both monastic curricula in Theravāda countries and mod-
ern Western scholarship. Indeed, it is this text’s system of Abhidhamma that is regularly cited
and discussed as the Pāli Abhidhamma in textbooks and scholarly works, despite the fact that
it differs in several highly significant respects from the canonical literature (a point that is itself
seldom noted). For example, textbooks that cite a standard enumeration of Abhidhamma catego-
ries of eighty-nine classes of cittas and fifty-two cetasikas are drawing from the Compendium
rather than the Enumeration or any other canonical listing, where the lists do not correspond
exactly with this system.21
The Compendium owes its reputation in part to its concise and highly systematic presentation
of what had grown to be a huge and unruly body of material when we consider not just the seven
books of the Abhidhamma but the many layers of commentary and subcommentary that followed
in their wake. The Compendium is only about fifty pages, making it a handy and useful summary
and primer. While this is not the place to compare systematically this text with the Enumeration,
I can mention several features of it that differ from its canonical predecessor, and these features
suggest a far more ontological orientation. As we have seen, the Enumeration lists its phenom-
ena always in reference to questions about what occurs in particular occasions of momentary
awareness; for example, the list we began with previously describes a certain specific type of
momentary awareness, an “occasion” in which some combination of fifty-six or more dhammas
could occur. Other lists describe bad or indeterminate moments of awareness, different occasions
of experience. In other words, the text does not attempt a single generic list of items of experience
abstracted from the occasions in which they occur. But the Compendium tidies all of this up to
offer a single abstract list in an arrangement of eighty-nine cittas and fifty-two cetasikas.
The Compendium also adds distinctions not present in the canon, such as a category of “beau-
tiful cittas,” and a distinction between universally present and occasional cetasikas. Further,
when we considered the first kind of awareness in the Enumeration, we noted that some fifty-six
cetasikas (and counting) could be present in that particular type of good awareness and that some
of the items could be said to be repeating under different aspects (mindfulness, for example,
appears four times). In the Compendium’s generic list of fifty-two cetasikas, all such apparent
repetitions have been pruned away, and the list is presented without the possibility that it could be
supplemented; it appears to be closed and total. Moreover, even a cursory glance at the two lists
reveals that a not insignificant number of the cetasikas differ from the lists of cetasikas occurring
in the Enumeration, as for example: the Compendium does not list citta itself (it may have seemed
redundant to treat citta as both a cetasika and a moment of awareness comprised of cetasikas),
calm meditation (samatha), and vipassanā. Faith, mindfulness, energy, and others are listed once
as such, but not in their aspects as powers and faculties. Among akusala dhammas, missing in the
Compendium but present in the Enumeration are wrong thought, wrong effort, wrong concentra-
tion, and some of the dhammas acting as powers and faculties (such as the power of energy). In
addition, the Compendium includes items that are not in any of the Enumeration’s listings, and
came into the Abhidhamma tradition only with the aṭṭhakathā, including attention, resolve, ini-
tiative, and among bad cetasikas, conceit, envy, avarice, rigidity, and sluggishness.22 Doubtless,
the field would benefit from a more sustained comparative treatment of these two texts and the
implications of their differences than I can offer here, but I have tried to suggest enough of these
differences to suggest that we are quite possibly dealing with different systems.
Of course, some may view these differences as simply the Compendium’s more rational,
simplified, and elegant rendering of the same basic text. To be sure, it lists many of the Enu-
meration’s distinctions and classifications that may capture its complexity when put into
action ramifying these classifications with its basic list of dhammas. And we have no standing
to require that the “tradition” should adhere to the canonical formulations rather than what
150
˙
it may consider to be a more concise maturation of it. But the differences in both content
and form are significant enough that they suggest that modern scholars may wish to avoid
speaking of a single Abhidhamma theory, represented by this medieval summary of it. In
terms of content, the Compendium presents new and different dhammas and new distinctions
and categories. Perhaps even more significantly, in terms of formal presentation it presents a
complete, abstract, “view-from-nowhere” list, and many of the features of the modal qual-
ity of presentation (such as how a particular phenomenon occurs under different aspects and
operating in different modes, such as powers and faculties) are no longer evident. These formal
features suggest not just differences in the details, but a substantial development of the tradi-
tion over time away from what appeared as open-ended sets of modal analytical practices to a
more reductive, closed, and complete system. The differences and tensions between a total and
totalizing system and an open-ended one deserve further scholarly scrutiny.
Perhaps the Compendium’s most significant ontological move is the use of language that
was not present in the Enumeration but that came to be used and then invested with onto-
logical significance in later exegetical layers. Two critical terms in this regard are paramat-
tha (ultimate meaning) and sabhāva (the particular nature of a thing). Neither is present in
the Enumeration or Analysis.23 We can begin with paramattha. The Compendium frames its
project in terms of listing the four main types of phenomena from an “ultimate” standpoint:
citta, cetasika, rūpa, and nibbāna, which then receive further elaboration as they come to
structure the whole system. In earlier commentarial literature, paramattha just meant a kind
of ultimate, analytic, “furthest-sense” language, but there is reason to believe that by this
medieval text (and likely much before in the ṭīkā literature), it had come to refer to ultimate
reality – the things or events that really exist. We have traveled considerably from the early
usage of dhamma as merely an object of experience. The Compendium says that “in this way
the Tathāgatas reveal what is ultimate as four: awareness, psychological phenomena, form,
and nibbāna.”24 This assertion suggests that Tathāgathas are not just speaking in furthest-sense
language, but rather are revealing ultimate reality. The text also speaks of paramattha in terms
of what exists when it contrasts dhammas, such as form and feeling, with what is not found to
exist in this sense, such as earth and mountains.25 This suggests a distinction about how our
concepts capture what really exists, and the terms that capture what really exists are dhammas
like rūpa and vedanā. It now becomes possible to refer to “ultimate dhammas” ( paramat-
thadhamma) as the commentaries on the Compendium go on to do.26 In his commentary on the
Compendium, modern interpreter Ledi Sayadaw takes himself to be offering “an explanation
of the ultimate” (Sayadaw 1913–1914, 129; see Braun 2015, 20–21). The ontological inter-
pretation is of course a very powerful religious and philosophical claim: the Buddha, in this
highest form of his teaching, declared and revealed in the most direct terms what really exists.
Another crucial term in the ontological shift is sabhāva, which is not mentioned by any of
the canonical Abhidhamma texts. It enters the Pāli tradition in the aṭṭhakathā as referring to the
particularity or specific characteristic a thing has that makes it different from other things, and
it can be used very broadly in this way. But after Buddhaghosa, the term sabhāva came to be
freighted with maximalist ontological significance to mean the “intrinsic nature” of something
(much as the Sanskrit term svabhāva came to be understood by the Madhyamaka critics of the
Abhidharma traditions). Further, dhamma comes to be identified with sabhāva, so dhammas
come to be that which have their own, intrinsic nature.27 While the Compendium is terse enough
to not always be entirely clear about what it means by such terms, it does indicate that all of the
fifty-two cetasikas, for example, are established “by way of their particular nature.”28 The project
has come to involve listing phenomena according to their intrinsic natures, and so we come to
find modern accounts that speak of dhammas as “things in themselves, ultimates.”29 Whereas
151
Maria Heim
the Enumeration began to list phenomena that can occur in particular occasions of awareness
and that must be analyzed in a modal way, the Compendium simply lists the bare phenomena
themselves, recruiting the notion of paramattha to indicate ultimacy and the notion of sabhāva
to indicate an intrinsic nature.
Y. Karunadasa characterizes the “dhamma theory” that has at this point emerged in strongly
ontological terms, although he acknowledges that this theory was “not precisely articulated”
in the canonical texts.30 Combined with the ontological sense of paramattha, the theory is now
seen to be positing “ultimate existents” that can be said to have “objective existence,” and the
ṭīkās begin to speak of “ultimate intrinsic realities” ( paramatthasabhāva).31 A dhamma now
represents “a fact having an objective counterpart” and “an actual datum of objective experi-
ence” (Karunadasa 1996, 18). For Karunadasa, dhammas are the “ultimate irreducible data” of
cognition and objective existence, and the “ultimate elements of existence”; they are “not ame-
nable to further analysis” (Karunadasa 1996, 5). He further suggests that the tradition can be
considered a kind of “critical realism” in that it “recognizes the distinctness of the world from
the experiencing subject” and that “the dhammas do not exist in dependence on the operation
of the mind” (Karunadasa 1996, 28). While I think that much of this language considerably
overstates what we actually see in the literature (especially the idea that the dhammas are
somehow independent of experience and not amenable to further analysis), I think he is right
to say that the ṭīkā literature, as represented here primarily by the Compendium, has generated
an ontology of ultimate reals as the highest teaching that is the Abhidhamma.
Abhidhamma as Endless Phenomenological Analysis

While the preponderance of tradition and scholarly weight falls on the side of the ontological
reading, we can identify an entirely different set of choices going as far back as the aṭṭhakathā
(the earliest postcanonical exegetical layer that we have). This material – codified in the fifth
century, likely by a team of scholars headed by Buddhaghosa – emphasizes not ontology but
analysis in its reading of the Enumeration and Analysis. Before turning to Buddhaghosa’s reading
of the Abhidhamma, it may be useful to recap several features of the canonical texts themselves
that suggest that assuming them to be fundamentally ontological in nature and purpose may be
misleading.
As we have seen, Abhidhamma matrices, which generate other lists to expound the teach-
ing and its structures, continue ever more finely grained analyses and questioning “from all
sides” in what Analysis calls the Abhidhamma Analysis. I hope that even in my brief synopses
of both books, it is clear that analysis begets analysis without ever landing on a final, single,
ultimate list of phenomena as the point or outcome of all the many modes of analysis these
books deploy. The point rather seems to be to generate analyses that consider and reconsider
teachings about experience in a modal and aspectual way. I have also mentioned the practical
nature of these methods of analyzing in contemplative practice experience. They were used
not to canvass all of psychological life and the material world, but rather to examine further
the particular teachings the Buddha gave in the suttas to observe and change one’s experience
for therapeutic and soteriological purposes.
The distinction between conventional teachings and ultimate teachings occurs first in the
Kathāvatthu (and is not mentioned in any of the other six canonical books) and then is picked
up at the aṭṭhakathā layer, where it refers not to different kinds of truth or reality, but to differ-
ent registers of the Buddha’s teachings. In the Kathāvatthu the distinction can be read not as
describing ultimate reality per se, but as indicating how one arrives at different terms: “does
one arrive at ‘person’ by means of the furthest-sense, by what is realized?”32 The answer is no,
152
˙
because “person” does not survive the analytic dismantling of it into parts, such as the aggre-
gates, so it is not a paramattha term. By contrast, the technical language of the Abhidhamma,
such as aggregates, dhammas, form, feeling, et cetera, are arrived at via furthest‑sense ( para-
mattha) analysis. (This does not entail, however, that any mention of “person” ( puggala) is
inappropriate,33 at least not when we consider that an entire book of the Abhidhamma, the
Puggalapaññatti, explores the nature of the person, and not just reductively, but also socially,
morally, in terms of progress, and so on).
For Buddhaghosa, paramattha teachings are given in a register that is technical and analyti-
cal – language in the “furthest sense.” This is contrasted with the Buddha’s more ordinary lan-
guage that is given in conventional (sammuti) or customary (vohara) teachings. He is insistent
that the distinction refers to different types of the Buddha’s language (bhāsā) and teachings
(desanā), and not to different kinds of truth (sacca).34 One would not say that ultimate lan-
guage is true and conventional language is false, because both refer to the Buddha’s teachings
and are thus (for him) both incontrovertibly and unqualifiedly true. It is not an ontological
distinction, and the mention of paramattha is not an occasion to describe reality. It may seem
like a short step to say that something that cannot be analyzed further is what actually exists,
but this is precisely what the canonical materials, and Buddhaghosa, nowhere say (and indeed,
they never grant that these categories cannot be analyzed further).
Buddhaghosa mentions sabhāva but does not imbue it with ontological significance,
and he uses the term quite broadly. For example, he describes the different sabhāvas or
particularities of each of the decaying corpses in the “meditations on the disgusting” in the
Visuddhimagga. When he does so, he is not mounting an argument that these different types
of decaying corpses are final units of analysis and thus ultimate reals; rather, this is part of
a contemplative practice identifying particular features of a meditation object. Dhammas,
too, can be defined in terms of their particularities because it can be useful to define a thing
by its specific characteristic, as we see here: “dhammas have particularities (sabhāva);
alternatively, dhammas are brought about by conditions, or dhammas are brought about
by their particularities.”35 The word dhamma can be understood in terms of sabhāva, but
for him, it does not mean “essence” or “intrinsic nature,” but only a thing’s particularity
whereby it can be defined as different than something else. The claim that dhammas can be
“brought about” by either their own particularities or by conditions itself rules out that they
are self-existent, sui generis phenomena. It says that they can be understood or grasped by
either their particularities or by how they are conditioned by other phenomena. Further,
Buddhaghosa emphasizes that dhammas are “empty.” Discussing the “Section on Empti-
ness” (suññatavaro) in the Enumeration, he says this:
“These are just dhammas” is mentioned to show that, due to their emptiness, “they
are only dhammas, without essence, without a leader.” Therefore, the meaning is to
be understood in this way: on an occasion that the first main good awareness arises in
the realm of sense desires, on that occasion fifty-six dhammas arise by virtue of being
constituents of that awareness, in the sense of their particularities.36
Dhammas have particularities but are without essences due to their emptiness. Of course,
emptiness as a mode of analysis had long been part of the canonical Abhidhamma and the
Paṭisambhidāmagga (which explicitly discusses the “emptiness of sabhāva” in the case of
each of the five aggregates), and Buddhaghosa follows suit.37
While Buddhaghosa’s interpretation of these two terms, paramattha and sabhāva, begins to
show that he reads the tradition differently than do his successors, there are other features of his
153
Maria Heim
interpretation relevant to the point. As we have seen, he reads the Abhidhamma texts as an “end-
less and immeasurable” series of methods and practices that begin to give room for the Buddha’s
active and dynamic omniscient practices of knowing things without obstacle. He identifies his
school of commentary as the Vibhajjavāda, which among other things entails that one steers clear
from launching one’s own view and instead “takes up the meaning of a text and then returns
again to that meaning by explaining it with different methods ( pariyāya).”38 We have already had
occasion to note the “Many Types of Elements Sutta,” in which training in being inquisitive about
multiple modes of teaching and numerous lists is a fundamental skill for a teacher; the “Many
Types of Feeling Sutta” does similar work with feeling (vedanā), showing that this dhamma may
be sliced and diced variously and is hardly irreducible (see Heim 2021). This is the spirit and style
in which Buddhaghosa read the canonical texts and carried out his own analyses in the Visud-
dhimagga. I have elsewhere argued that when we take the time to work through Buddhaghosa’s
theory of scripture and the interpretative practices he engages in, we see him articulating a view
of the Abhidhamma Piṭaka as a kind of “phenomenological analysis,” a modal examination of
experience through methods designed to explore experience and change it (Heim 2018; Heim
and Ram-Prasad 2018). He is not engaged in metaphysical or epistemological arguments.
Among modern scholars, as may be apparent, Nyanaponika Thera shares much of this view
in an elegant and insightful discussion focusing on the Enumeration. He describes the canonical
Abhidhamma as a whole as a phenomenology in that it deals with “phenomena, that is, with the
world of internal and external experience,” which he contrasts with “ontology, or metaphysics,
that inquires into the existence and nature of an essence, or ultimate principle, underlying the
phenomenal world” (Nyanaponika 1998, 190). He modulates this to some extent, however, by
suggesting that the practices of analysis in the Abhidhamma do in fact aim at contributing to
“ontological problems,” namely that it shows that in reality there no abiding essence to be found
in the world of experience (Nyanaponika 1998, 21); and he does slip back into talking about “real-
ity” from time to time (though he is careful to say that it deals with “actuality from an exclusively
ethical and psychological viewpoint and with a definite practical purpose”) (Nyanaponika 1998,
2, 46). Notably, Bhikkhu Bodhi, in his introduction to Nyanaponika’s book, modulates the phe-
nomenological reading further, and works it back to ontology, asserting that “the Abhidhamma
draws up a list of ontological actualities” and does in fact discriminate between what is real and
what is only apparently real (Nyanaponika 1998, xvi – xviii). It is important to many that the Abhi
dhamma be seen to be the Buddha offering an account of ultimate reality, the way things really are.
My own reading is to take seriously the possibility that the intellectual purpose of the
canonical Abhidhamma is to offer a complex set of analytic practices and methods that allow
the practitioner to explore the many – perhaps infinite – facets of experience without ever land-
ing on a final, single, essential list of the contents or aspects of it. The lists aim not to arrive at a
single, irreducible account of the nature of our experience, but rather to engage in the methods
that would resist such a final account (and in this sense, the later tradition contravenes this
spirit). The lists do not seek to provide an account of reality, even the reality of experience, but
rather to provide methods and practices of analysis as part of the contemplative and analytical
purpose to inquire into and transform experience. This is philosophy of a different sort than
metaphysical assertion and argument. It is a practice of exploring and transforming experience
within the therapeutic and soteriological aims of Buddhist dogma.
Part of how I have arrived at this view is that I find much of the modern ontological account
unconvincing, either as a reasonable account of the canonical texts or in terms of how it could
work philosophically. We may begin with some of the textual considerations. If the Enumera-
tion was attempting to arrive at a final list of irreducible dhammas, why did it not simply and
clearly state, at some point, what these are? Given that it emerged in the ancient Indian context
154
˙
in which sophisticated ontological systems were ubiquitous, including other traditions of Bud-
dhist thought, why did it not avail itself of such ontological terminology? For that matter, why
not define dhamma in terms that would make clear that it is an element of ultimate reality,
instead of assiduously avoiding language that could be (and later was) construed in ontological
ways (sabhāva, paramattha)? It was either innocent of such ideas, or it deliberately avoided
them. If we are to be persuaded that the later medieval tradition seamlessly represents the
canonical tradition, we need to see how exactly this works in a treatment that first reckons with
the obvious differences in the texts.
Philosophically, I find some of the readings we have explored about dhammas either unde-
fended or incoherent. We can begin with widespread slippage in the secondary scholarship that
begins by noting that the dhammas are objects of experience as they are conveyed by core teach-
ings, but then slides into asserting that they refer to elements of reality, without ever explaining
why this must be. It is worth recalling that the original matrices provided lists of doctrines, and
doctrines are always practices – the Four Noble Truths is a practice of diagnosing experience
and understanding its conditions; the five aggregates is a practice of dismantling the idea of
person; dependent origination is a practice of understanding mutual causes and conditions in
analyzing experience. The items that come to be treated as dhammas are taken from these lists
of practices: things like feeling, perceiving, mindfulness, path factors, and so on. These come to
be objects of experience in introspective methods designed to recognize them “one by one,” to
see their conditionality, and to transform the practitioner’s phenomenality.
This makes it most unclear what it would mean to say that dhammas exist, as we saw pre-
viously, “objectively” or that they exist “no matter what, without depending on the existence
of any other dharma or on any conceptualizing mind” as “impersonal,” untethered “from any
experiencing subject.” Nor is it clear how dhammas can be “irreducible” or “not amenable to
further analysis” when the canonical texts can always be counted on to provide another list
that analyzes, in another mode, each of the phenomena it takes up – elements can be eighteen,
or six, or a different six, or a still different six, or three, or two; feelings can be carved up vari-
ously and reconsidered by lists of two, three, five, six, eighteen, thirty-six, one hundred and
eight, and so on (Heim 2021). Rather, the question is: where, exactly, do these analyses end?
How do we find a stopping point that is not arbitrary? I think Gethin is closer to the mark when
he notes the danger that comes from breaking things up into parts in that “we might then take
the parts as real and begin to reify the world again, if in a different way.”
It seems to me that the early Abhidhamma authors sought to avoid precisely this
same danger through the elaboration of the various mātikās. Try to grasp the world of
the Dhammasaṅgaṇi, of the Paṭṭhāna, and it runs through one’s fingers.39
He goes on to emphasize the use of the mātikās in “practical psychology,” in a system “primar-
ily concerned to distinguish states and processes of mind on the basis of actual observation.”
And he is right, I think, to note that one often encounters details that are beautiful and profound
(though other tasks at hand have precluded space for me to begin to do justice to these).
Finally, I remain deeply intrigued by the philosophical potential of a modal and aspectual
exploration of experience. I am persuaded by Buddhaghosa and Nyanaponika Thera that prolif-
erating groupings and classifications of phenomena can help resist essentialist accounts of what
we experience. Our feelings and sensations and moral dispositions do seem, phenomenologi-
cally, to be open and changeable in their workings, valences, and intensities to whatever else
occurs with them in any given occasion of awareness. Is “pleasure” (to take at random just one
dhamma explored in the matrices) really the same thing in every instance in which it occurs?
155
Maria Heim
Do “right view” and “right thought” (to take two more) have single uninflected unchanging
natures across all moments of experience in which they might occur? The abstractions that we
use to get at our experience – whether we are talking about the solidity we experience in the
things we touch or the hedonic impact we encounter when we feel – are always going to be at
some remove from the singular particularity of our actual moments of experience. But it could
be that those abstractions do a more useful job of describing experience when they are subjected
to numerous, perhaps even endless, modes of inquisitive analysis that classify and reclassify,
divide and redivide, consider and reconsider, than when they aim to land on a single, final,
highly abstract, essentialist, objective, and ultimate “own nature” that they must be.
The early Abhidhamma texts teach us that the philosophy of experience can be an “end-
less and immeasurable” application of methods useful for taking up concrete moments of
experience. This suggests more of a process of analytic work than a final description of reality.
Scholars of the early Western phenomenological tradition might see affinities in their approach
and this repeated application of phenomenological questions that remain open to further analy-
sis. In the Buddhist tradition, however, this philosophy as practice is, of course, ultimately
therapeutic as its methods provide programmatic ways to examine and then fundamentally
change the patterns of experience.
Notes
1 See Atthasālinī (As) 14–17 (Tin and Rhys Davids 1976, 16) for the details in this paragraph; see also
Heim (2018, 40–44, 154–63) on the authorship and teaching of the Abhidhamma as Buddhaghosa
understood these. “Endless and immeasurable” is ananto aparimāṇo. While all translations are my
own unless otherwise noted, I also cite the available English translation. All abbreviations follow the
Pāli Text Society.
2 As 16 (Tin and Rhys Davids 1976, 20): Evaṃ sammāsambuddhe nayaṃ dente paṭisambhidāppattassa
aggasāvakassa velante ṭhatvā hatthaṃ pasāretvā dassitasamuddasadisaṃ nayadānaṃ hoti.
3 Gethin 1992b, 150–56. My translations are slightly different than Gethin’s, most notably, “perceiv-
ing” for saññā instead of “recognition” (though he is of course right that it involves identifying and
naming what one perceives).
4 Atthasālinī 29 (Tin and Rhys Davids 1976, 37).
5 There is substantial scholarship discussing the origins of the Abhidhamma in terms of suttas that
are similar to, or precursors for, some of what we see in the Abhidhamma Piṭaka; for some of it see
Anālayo (2014) and Cousins (2015). Both are also very helpful on the oral and performative aspects
of these practices.
6 A v.54–59 (Bodhi 2012, 1376–79). This the second of two suttas called Mahāpañhā. At the conclu-
sion of Kijaṅgalā’s teaching, the Buddha calls her “paṇḍitā,” a learned pundit.
7 Bahudhātuka Sutta (M iii.61–67; Ñāṇamoli and Bodhi 1995, 925–30). Briefly, the eighteen elements
(dhātu) are each of the six senses (mind is a sense) parsed into three components of sensing; the first six
elements are: earth, water, fire, air, space, and consciousness; the next six are: pleasure, pain, joy, distress,
equanimity, and ignorance; the next six are: sensual desire, renunciation, ill will, non-ill will, cruelty,
and non-cruelty; the three are sense-sphere, form-sphere, and formless-sphere elements; the two are:
conditioned and unconditioned. Another sutta that refuses to settle on a single list of phenomena (in its
case, feeling) so we get lists of two, three, five, sixteen, thirty-six, and one hundred and eight (“The Many
Types of Feeling Sutta,” Bahuvedanīya Sutta at M i.396–400; cf. S iv.223–28, 231–32 see Heim 2021).
8 Dhs 1: [katame] kusalā dhammā akusalā dhammā abyākatā dhammā (Rhys Davids 1975, 1–2).
9 In the suttas, dhammas in this technical sense are phenomena observed in meditation, as in the
Mahāsātipaṭṭhāna Sutta (D ii.290) and the Anupada Sutta (discussed in what follows).
10 As 38: nissattanijjīvatā (Tin and Rhys Davids 1976, 49).
11 As 38: ārogyaanavajjachekasukhavipākesu (Tin and Rhys Davids 1976, 48).
12 Dhs 8–9 (Rhys Davids 1975, 1–5).
13 Anupada Sutta, M iii.25–29 (Ñāṇamoli and Bodhi 1995, 899–902).
156
˙
14 As 135–136 (Tin and Rhys Davids 1976, 178–80). Gethin discusses the implications of listing the
“same” phenomena as powers and faculties (1992a, 143).
15 Nyanaponika (1998, 88–89; 37–41), citing James Ward. These are points he emphasizes throughout
ch. IV. “Modal” and “aspectual” are my terms, however.
16 Ye vā pana tasmiṃ samaye aññepi atth paṭiccasamuppannā arūpino dhammā: that is, literally, “and
whatever other dependently-arisen formless dhammas occur on that occasion” (Dhs 8, 75, 87, et cetera).
17 As, for example, Dhs 75, 87.
18 Visuddhimagga XI.93, 96. Gethin makes the same point about rūpa: the analyses of rūpa “focus on
the physical world as experienced by a sentient being” (1986, 36). See also Heim and Ram-Prasad
(2018) for a systematic demonstration of what Buddhaghosa (and I think the Abhidhamma on which
he is relying) means by nāma and rūpa.
19 The Vibhaṅga is translated by U. Thiṭṭila (1969), and his introduction provides a useful overview.
20 As 2–3 (Tin and Rhys Davids 1976, 4–5).
21 It is standard to cite the Compendium’s system, as for example, Williams and Tribe (2000, 90–91),
Gethin (1998, 210–15), Lamotte (1988, 594–97), and Nyanatiloka (2008, Appendix II); following
Nyanatiloka, the system is given in charts in appendices in Ñāṇamoli’s translation of the Visuddhim-
agga (Ñāṇamoli 1999), even though the Visuddhimagga itself nowhere articulates it as such.
22 Compare Dhs 8 and Dhs 75 to the Compendium’s chapter II. The latter’s main commentary (the
Abhidhammatthavibhāvinī) occasionally notes a discrepancy between the Enumeration and the Com-
pendium, as for example, the way that the “heart base” (hadayavatthu) is not mentioned in the former
but warrants listing among rūpa dhammas by the Compendium since it is “in accordance with the
scriptural tradition and reasonable” (Wijeratne and Gethin 2007, 221).
23 It is surprising the extent to which scholars read these terms back into the canonical texts, however
(as, for example, Nyanatiloka 2008, 16).
24 Saṅgaha VI.65; Wijeratne and Gethin 2007, 255. My translation.
25 Saṅgaha VIII.42; Wijeratne and Gethin 2007, 322.
26 For example, Abhidhammatthavibhāvinīṭīkā §2 (see Wijeratne and Gethin 2007, 7) and Ledi Say-
adaw’s Paramatthadīpanī 332.
27 Karunadasa (1996, 15 and 42n37), citing the Abhidhamma Mūlaṭīkā (121): dhammo ti sabhāvo.
28 Saṅgaha III.1; Wijeratne and Gethin (2007, 88).
29 U. Thiṭṭila (1969, xxv). This kind of reading comes to be read back into the canonical texts, as in this
instance.
30 Karunadasa (1996, 2). Noa Ronkin, who frames the Pāli Abhidhamma in strongly metaphysical and
ontological terms, also acknowledges, drawing on Gethin, that the canonical tradition did not posit
irreducible elements and is more open-ended than is usually recognized (2005, 109).
31 Karunadasa (1996, 19 citing the Abhidhammārthasaṁgraha-sannaya); the use of paramatthasabhāva
is also found in the Vibhaṅga-Mūlaṭīkā and the Mahāṭīkā on the Visuddhimagga.
32 Kvu 1: puggalo upalabbhati saccikaṭṭhaparamatthenāti.
33 Many scholars assume that this distinction about language is really about reality: Karunadasa says
that the passage denies “the reality of a person” (1996, 11), and that this concerns “two levels of
reality” (1996, 12). Ledi Sayadaw asserts that from the standpoint of ultimate reality, the notion of a
personal entity is “just an erroneous view” (1913–1914, 129).
34 Papañcasūdanī i.137 and Manorathapūraṇi i.94. See Heim (2018, 85–94) for translations and dis-
cussions of the relevant passages from Buddhaghosa.
35 As 39: Attano pana sabhāvaṃ dhārentīti dhammā. Dhāriyanti vā paccayehi, dhārīyanti vā
yathāsabhāvatoti dhammā. It is not easy to translate dhārenti and dhārīyanti – to wear, have, possess,
bear, bring.
36 As 155: Dhammāva ete dhammamattā asārā apariṇāyakāti imissā suññatāya dīpanatthaṃ vuttā.
Tasmā evamettha attho veditabbo – yasmiṃ samaye kāmāvacaraṃ paṭhamaṃ mahākusalacittaṃ
uppajjati, tasmiṃ samaye cittaṅgavasena uppannā atirekapaṇṇāsadhammā sabhāvaṭṭhena dhammā
eva honti; this is on Dhs 121.
37 Paṭisambhidāmagga II.178 (sabhāvena suññaṃ). For a translation of its chapter on emptiness see
Ñāṇamoli (1982, 356–61). The Paṭisambhidāmagga is an Abhidhamma-like text that was placed in
the Suttanta Piṭaka; Buddhaghosa relies on it heavily. For another example of Buddhaghosa noting
the emptiness of dhammas, note how he sees the twelve parts of dependent origination as empty in
Vism XVII.273 and 283.
157
Maria Heim
38 Vism XVII.25 and Sammohavinodaṇī 130: atthaṃ saṅgāhentena tamevatthaṃ punarāvattetvā apare-
hipi pariyāyantarehi niddisantena.
39 Gethin (1992b, 165). It should be noted that Gethin from time to time slides into a metaphysical treat-
ment of dhammas and the whole system; as for example, dhammas are physical and mental events
that are “the way things are” and very much like “atoms” (1998, 209).
Bibliography
Anālayo. 2014. The Dawn of Abhidharma. Hamburg: Hamburg University Press.
Bodhi, Bhikkhu, ed. 2007. A Comprehensive Manual of Abhidhamma: The Abhidhammattha Saṅgaha of
Ācariya Anuruddha. Kandy: Buddhist Publication Society.
Bodhi, Bhikkhu, trans. 2012. The Numerical Discourses of the Buddha: A Translation of the Aṅguttara
Nikāya. Boston: Wisdom Publications.
Brahmavamso, Ajahn. 1995. Chaṭṭhasaṅgāyana Saṅgāyana Tipiṭaka. Igatpuri: Vipassana Research
Insitute.
Braun, Erik. 2015. “The Great War of the Commentaries: The Abhidhamma and Social Change in Colo-
nial Burma.” History of Religions 55 (1): 1–40.
Bronkhorst, Johannes. 1985. “Dharma and Abhidharma.” Bulletin of the School of Oriental and African
Studies 48: 305–20.
Buswell, Robert E., and Donald S. Lopez, eds. 2014. The Princeton Dictionary of Buddhism. Princeton:
Princeton University Press.
Cousins, Lance S. 1981. “The Paṭṭhāna and the Development of the Theravādin Abhidhamma.” Journal
of the Pāli Text Society 9: 22–46.
———. 2014. “The Case of the Abhidhamma Commentary.” Journal of the International Association of
Buddhist Studies 36–37: 389–422.
———. 2015. “Abhidhamma Studies III: Origins of the Canonical Abhidha(r)mma Literature.” Journal
of the Oxford Centre for Buddhist Studies 8: 96–145.
Gethin, Rupert. 1986. “The Five Khandhas: Their Treatment in the Nikāyas and Early Abhidhamma.”
Journal of Indian Philosophy 14 (1): 35–53.
———. 1992a. The Buddhist Path to Awakening: A Study of the Bodhi-Pakkhiyā Dhammā. Leiden: E.
J. Brill.
———. 1992b. “The Mātikās: Memorization, Mindfulness and the List.” In In the Mirror of Memory:
Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, edited by Janet
Gyatso, 149–72. Albany: State University of New York.
———. 1998. The Foundations of Buddhism. New York: Oxford University Press.
———. 2009. “He Who Sees Dhamma Sees Dhammas: Dhamma in Early Buddhism.” In Dharma: Stud-
ies in its Semantic, Cultural, and Religious History, edited by Patrick Olivelle, 91–120. New Delhi:
Motilal Banarsidass.
Heim, Maria. 2018. Voice of the Buddha: Buddhaghosa on the Immeasurable Words. New York: Oxford
University Press.
Heim, Maria, and Chakravarthi, Ram-Prasad. 2018. “In a Double Way: Nāma-Rūpa in Buddhaghosa’s
Phenomenology.” Philosophy East and West 68 (4): 1085–115.
Heim, Maria. 2021. “Some Analyses of Feeling.” In The Bloomsbury Research Handbook of Emotions
in Classical Indian Philosophy, edited by Maria Heim, Ram-Prasad Chakravarthi, and Roy Tzohar,
87–105. London: Bloomsbury.
Karunadasa, Y. 1996. The Dhamma Theory: Philosophical Cornerstone of the Abhidhamma. Kandy:
Buddhist Publication Society.
Lamotte, Étienne. 1988. History of Indian Buddhism From the Origins to the Śaka Era. Translated by
Sarah Webb-Boin. Louvain: Institut Orientaliste.
Ñāṇamoli, trans. 1982. The Path of Discrimination (Paṭisambhidāmagga). London: The Pāli Text Society.
———. 1999. The Path of Purification: Visuddhimagga By Bhadantācariya Buddhaghosa. Kandy, Sri
Lanka: Buddhist Publication Society.
Ñāṇamoli, and Bhikkhu Bodhi, trans. 1995. The Middle Length Discourses of the Buddha: A Translation
of the Majjhima Nikāya. Somerville, MA: Wisdom Publications.
Nyanaponika Thera. 1998. Abhidhamma Studies: Buddhist Explorations of Consciousness and Time.
Edited by Bhikkhu Bodhi. Boston: Wisdom.
158
˙
Nyanatiloka. 2008. Guide Through the Abhidhamma Piṭaka: A Synopsis of the Philosophical Collection
Belonging to the Buddhist Pāli Canon. Kandy: Buddhist Publication Society.
Rhys Davids, C. A. F., trans. 1975. A Buddhist Manual of Psychological Ethics (Dhammasaṅgaṇī). Lon-
don: Pāli Text Society.
Ronkin, Noa. 2005. Early Buddhist Metaphysics: The Making of a Philosophical Tradition. New York:
RoutledgeCurzon.
Sayadaw, Ledi. 1913–1914. “Some Points in Buddhist Doctrine.” Journal of the Pāli Text Society 7:
115–69.
Thiṭṭila, U., trans. 1969. The Book of Analysis (Vibhaṅga). Bristol: Pāli Text Society.
Tin, Pe Maung, and C. A. F. Rhys Davids. 1976. The Expositor (Atthasālinī). London: Pāli Text Society.
Westerhoff, Jan. 2013. “Metaphysical Issues in Indian Buddhist Thought.” In A Companion to Buddhist
Philosophy, edited by Steven Emmanuel. Oxford: Wiley-Blackwell.
Wijeratne, R. P., and Rupert Gethin, trans. 2007. Summary of the Topics of Abhidhamma
(Abhidhammatthasaṅgaha) by Anuruddha and Exposition of the Topics of Abhidhamma
(Abhidhammatthavibhāvinī) by Sumaṅgala. Oxford: Pāli Text Society.
Williams, Paul, and Anthony Tribe. 2000. Buddhist Thought: A Complete Introduction to the Indian
Tradition. London: Routledge.
159
10
MOGGALIPUTTA TISSA’S
POINTS OF DISCUSSION
(KATHĀVATTHU)
Reasoning and Debate in Early
Buddhist Thought1
Rupert Gethin
Introduction
The Points of Discussion (Kathāvatthu) is a collection of early Buddhist debates about dis-
puted points of Buddhist thought. It thus provides insight into the philosophical issues that
preoccupied early Buddhists and more especially their methods of argument and reasoning. It
is counted the fifth of seven works that constitute the Abhidhamma-piṭaka, the third division
of the Pali canon of the Theravāda school of Buddhism. As such, it is awarded by followers
of the school the status of ‘the word of the Buddha’ (buddhavacana). Yet like two other works
of the Pali canon, it is traditionally attributed to a named individual. The Path of Discrimina-
tion (Paṭisambhidāmagga) and Explanation (Niddesa) of the Khuddaka-nikāya are attributed
to one of the Buddha’s immediate disciples, Śāriputra. The Points of Discussion, however, is
associated with an individual who is acknowledged to have lived several generations after the
death of the Buddha (ca. 400 BCE), during the reign of the emperor Aśoka in the middle of the
third century BCE, the elder Moggaliputta Tissa.
The Theravāda exegetical tradition acknowledges that some might object to regarding a
work composed a century or more after his death as ‘the word of the Buddha.’ However, it
justifies doing so by using a strategy derived from the discourses of the Sutta-piṭaka. In a num-
ber of discourses, we find the Buddha giving a brief outline that is subsequently explained in
full by a disciple; questioned about this, the Buddha confirms that he would have given a full
explanation in exactly the same terms as the disciple has done. In a similar vein, Theravāda tra-
dition claims that the Buddha, having foreseen Moggaliputta’s situation, established the basic
headings of the Kathāvatthu, and on the basis of these Moggaliputta later expounded the full
text not by his own knowledge but according to a method and scheme received from the Bud-
dha (As 4–5; Kv-a 1–2). In this matter, it seems we must side with the ancient skeptics: that the
contents and method of the Points of Discussion derive directly from the Buddha is hardly pos-
sible. The Points of Discussion assumes the development of attempts to set out the Buddha’s
teaching systematically and in detail on the basis of an established collection of discourses of
the Buddha; it assumes the development of different schools of Buddhist systematic thought
160 DOI: 10.4324/9781351030908-17

Moggaliputta Tissa’s Points of Discussion (Kathāvatthu)
(Abhidharma) with their own positions on a variety of issues. In short, the Points of Discussion
reflects the Buddhist intellectual world of the third–first centuries BCE (Cousins 1991, 34).
The Points of Discussion itself provides no explicit indication of its authorship or of the cir-
cumstances of its composition. The text is a collection of discussions (kathā) of various points
(vatthu) of Buddhist doctrine. Each discussion takes the form of a dialogue between represent-
atives who take opposing positions on the points in question. The oldest extant commentary
(Kv-a, composed probably at the beginning of the fifth century by an unnamed monk at the
invitation of Buddhaghosa) tells us that in each case, these two opposing positions are those of
the advocate (sakavādin), who presents the position of the author’s own school, the Theravāda
school of Lanka, and an opponent ( paravādin), who presents the position of another school.
The text provides no details of which schools of Buddhism these opponents might represent.
According to the details provided by the commentary, the opponents are representatives of
various Buddhist schools. This seems to be corroborated by what we know of the positions on
points of doctrine taken by different Buddhist schools from other sources (see Bareau 2013).
The discussions are arranged in twenty-three chapters (vagga) each comprising between five
and twelve discussions. The chapters are simply numbered sequentially and not given titles,
and while discussions of related points are sometimes grouped, there appears to be no overall
topical method to the arrangement. In all, the Points of Discussion consists of between 216 and
226 topics of discussion.2 The text is difficult to date as a whole since its structure clearly lends
itself to the addition of further points of discussion as they might arise. The lack of an overall
method of ordering suggests that this is, in fact, how the text developed.
Tradition tells us that Moggaliputta Tissa was born into a brahman family and educated in
Vedic tradition. He was induced to join the Buddhist order by a monk supported by his family. He
subsequently mastered the Buddhist textual tradition, achieved awakening, and became a monk of
great renown. As the story has it, the ascendancy of the Buddhist monastic order during the reign
of Aśoka attracted spurious monks. Ascetics belonging to non-Buddhist schools infiltrated the
order and continued with their own practices while neglecting the rules of the Buddhist order. The
king himself became involved and sought the help of a respected monk, namely Moggaliputta
Tissa. With his help, the spurious monks were expelled and the order declared pure and unified.
Moggaliputta Tissa then convened a Buddhist council at which he recited the Points of Discus-
sion, defeating any opposing views (Jayawickrama 1962, 33–55). This account of the origin of
the text seems ill fitted to its nature: a text that is concerned with the nuances and finer points of
Buddhist doctrine seems largely irrelevant to these circumstances, and it is likely that different
issues – keeping monastic discipline, guarding against schismatic tendencies, disagreement on
points of doctrine, establishing an authoritative Buddhist textual corpus – have been conflated.
The method of argument

The Points of Discussion is a difficult text. There is no critical edition; the only available trans-
lation into a modern European language is a paraphrase (Aung and Rhys Davids 1915, li–lii).
A full study remains a scholarly desideratum; Warder’s 1963 article remains the most reliable
guide to the overall method of argument. In order to begin to navigate the Points of Discus-
sion, we must rely in good measure on the oldest commentary. While there exists an English
translation of this commentary (Law 1940), it too must be regarded as provisional in nature.
The literary form of the Points of Discussion is that of a series of dialogues between an
advocate of the text’s own positions and an opponent. The text begins abruptly with the first
dialogue and continues without ever providing any indication of who is speaking which lines.
Although who (Theravādin or opponent) must be saying what is – for the most part – apparent,
161
Rupert Gethin
it is not always so (Aung and Rhys Davids 1915, lii–liii), which makes following an argu-
ment’s twists and turns sometimes difficult.
Each dialogue begins with the Theravādin advocate putting the disputed point to the oppo-
nent in the form of a proposition and asking if this is the position the opponent holds. The
opponent assents to the proposition. The advocate then puts further propositions to the oppo-
nent, which – in nearly all cases – the opponent denies. The Points of Discussion’s advocate
then suggests that the first proposition to which the opponent has assented implies these further
propositions, so in denying these, the opponent should acknowledge that his initial position
must be false. As we shall see, the method in part adumbrates the prasaṅga method of reductio
developed by Nāgārjuna (second century CE) and his Mādhyamika followers.
In principle, a complete dialogue for a single disputed point follows the same pattern (see
Table 10.1). First, there is what the commentary calls ‘the way of setting out the thesis’ (vāda-
yutti),3 a set of eight exchanges (A) termed ‘refutations’ (niggaha) in which the opponent’s
position is shown to be incoherent. This is followed by a series of further discussions (B)
addressing various propositions supplementary to the main proposition and aimed at further
demonstrating the unwanted implications of the opponent’s position; often, these supplemen-
tary discussions end with both sides citing canonical sūtra sources (suttāharaṇa) in support of
their respective positions. In the case of the first disputed point, the series of supplementary
discussions takes up some sixty pages (Kv 11–69, §§ 17–245); as the text progresses, the num-
ber of supplementary points tends to become fewer.
The initial set of eight refutations (A) follows a formalized pattern that provides a template
for demonstrating the contradiction involved in the opponent’s assenting to an initial proposi-
tion while denying a second proposition. This template is intended to be adapted and applied
as appropriate to the supplementary propositions (B) and also to all other disputes in the Points
of Discussion.
Each of the eight refutations consists of five sections: (i) an initial demonstration in five
steps of the contradiction involved in affirming a thesis; (ii) a rejoinder in four steps; (iii) a
refutation in four steps based on the rejoinder; (iv) the application of the refutation in four
steps; and (v) a conclusion in four steps asserting the success of the rejoinder.4
In the first section of the first of the eight refutations (initial demonstration), the Theravādin
advocate sets out the refutation of the opponent’s position; the opponent then responds in the
second section by fallaciously turning the Theravādin’s method of reduction back on him and
apparently winning the first round. In the first section of the second refutation, the opponent
repeats his counter to the Theravādin’s position, who then responds in the second section
by repeating his refutation of the opponent’s position and so winning the debate. As Warder
points out, these first two refutations are considered basic; the subsequent six refutations can
be applied as appropriate. In effect, they form three pairs that parallel the initial pair. Thus, in
the third and sixth refutations, the question is whether the opponent’s position applies in all
circumstances; in the fourth and seventh, whether it applies always; in the fifth and eighth,
whether it applies to everything. But even in the first two refutations, it is apparent that much –
from the perspective of the logic of the argument – is redundant.
This is in practice acknowledged by the text itself insofar as the eight refutations are set
out in full only in the case of the first disputed point (Kv 2–11, §§ 1–16); in all subsequent
debates, only the initial part of the first exchange is given, with the rest indicated by ‘and so on’
( peyyāla). This is possible since the first section (i) – the straightforward direct statement of
the argument in five steps – of the first refutation contains the essential reasoning that informs
the argument in the next seven refutations. The remaining sections and refutations can in prin-
ciple be constructed for each disputed point on its basis.
162
Table 10.1 The structure of the points of discussion.
First point of discussion
A. The way of setting out the thesis (vāda-yutti): 8 ‘refutations’ (niggaha)

1. Refutation in 5 sections
i. Direct (anuloma): exchange initiated by Theravādin advocate (5 steps)
ii. Rejoinder ( paṭikamma): exchange initiated by opponent (4 steps)
iii. Refutation (niggaha): statement by opponent (4 steps)
iv. Application (upanayana): statement by opponent (4 steps)
v. Conclusion (niggamana): statement by opponent (4 steps)
2. Refutation in 5 sections
i. Counter ( paccanīka): exchange initiated by opponent (5 steps) = 1.ii
ii. Rejoinder: exchange led by Theravādin (4 steps) = 1.i
iii. Refutation: statement by Theravādin (4 steps)
iv. Application: statement by Theravādin (4 steps)
v. Conclusion: statement by Theravādin (4 steps)
[1–2 = 10 sections; 42 steps]
3. Refutation in 5 sections: everywhere (sabbattha) – paired with 6
4. Refutation in 5 sections: always (sabbadā) – paired with 7
5. Refutation in 5 sections: in everything (sabbesu) – paired with 8
6. Refutation in 5 sections: everywhere – paired with 3
7. Refutation in 5 sections: always – paired with 4
8. Refutation in 5 sections: in everything – paired with 5
B. Supplementary applications (saṃsandana)
1. Additional discussions using the template (abbreviated) set out in A
2. Citing authoritative sūtra sources (suttāharaṇa)
Remaining 216/226 points of discussion

A. The way of setting out the thesis: 8 ‘refutations’
1. Refutation in 5 sections (only section i. stated)
i. Direct exchange initiated by Theravādin (5 steps)
‘And so on. . .’ ( peyyāla)
B. Supplementary applications
1. Additional discussions
2. Citing authoritative sūtra sources
Given the significance of this first section, it is worth examining its steps in detail. The com-
mentary points out that the Theravādin demonstrates his refutation in two ways: direct (anu-
loma) and reverse ( paṭiloma). The first disputed point concerns the existence of the person: the
opponent is a Buddhist who asserts the reality, in some sense, of the person, a pudgalavādin.5
Its first section is as follows (with names of the speakers, the distinction between the direct and
reverse ways of refutation, and the five steps as indicated by the commentary added):
Theravādin: The person exists in a real and ultimate sense?

Opponent: Correct.
Theravādin: The person exists in a real and ultimate sense in the same way as something
which is real and ultimate?
Opponent: No, this cannot be said.
Theravādin: Understand this refutation:
163
Rupert Gethin
[Direct (anuloma)]
– If the person exists in a real and ultimate sense, (1) then clearly it should be
said that the person exists in a real and ultimate sense in the same way as some-
thing which is real and ultimate.
– What you say is false, namely that (2) it can be said that the person exists in a
real and ultimate sense, but it cannot be said that the person exists in a real and
ultimate sense in the same way as something which is real and ultimate.
[Reverse ( paṭiloma)]
– But (3) if it cannot be said that the person exists in a real and ultimate sense in
the same way as something which is real and ultimate, (4) then clearly it cannot
be said that the person exists in a real and ultimate sense.
– What you say is false, namely, (5) that it can be said that the person exists in a
real and ultimate sense, but it cannot be said that the person exists in a real and
ultimate sense in the same way as something which is real and ultimate.
To use the terminology of the commentary, it is apparent that the first proposition (the
opponent’s position) takes the form of the protasis or antecedent (ṭhapanā) of a condi-
tional statement, and the second proposition its apodosis or consequent ( pāpanā). The
logical steps involved in this initial argument have been analyzed by Aung (Aung and
Rhys Davids 1915, xlviii–l), Schayer (1933, 91–92; 2001, 28), Bocheński (1961, 421–23),
Warder (1963, 64, 67), Watanabe (1983, 157–59), and Matilal (1998, 35–37), though with-
out complete agreement.
In the first place, there is disagreement as to whether this argument is better expressed
by way of the logic of terms or the logic of propositions. Schayer (1933, 90–91; 2001, 28)
is clear that we should use the latter; Bocheński (1961, 422–23) disagrees; Warder (1963,
64, 67), without reference to the issue – or to Schayer and Bocheński – uses both; Matilal
(1998, 37) leaves the issue open. Expressing the debate by way of the logic of terms gives
the following:
Theravādin: Is A B?
Opponent: Yes.
Theravādin: Is A C?
Opponent: No.
Theravādin:
[Direct (anuloma)]
– If A is B, then A is C.
– Therefore not: (A is B) and not (A is C).
– If A is not C, then A is not B.
– Therefore not: (A is B) and not (A is C).
Expressing it by way of the logic of propositions gives the following:
Theravādin: p?
Opponent: Yes. [⊦ p]
Theravādin: q?
Opponent: No. [∼q]
Theravādin:
164
[Direct (anuloma)]
– If p then q. [p ⊃ q]
– That p and not q, is false. [∼ ( p. ∼q)]
– If not q, then not p. [∼q ⊃ ∼p]
– That p and not q, is false. [∼ ( p. ∼q)]
As Warder (1963, 63–64) points out, not all disputed points are of the form: Is A B? Is A C?
He distinguishes seven basic forms on the basis of an analysis of the first twenty-one debates.
It would seem that at least some disputes (e.g. II.11, ‘there are two cessations’) could only be
expressed as propositions. There is also some disagreement as to whether the argument is bet-
ter reduced to the logical form of modus tollens – if it is true that p, then it is true that q; it is
not true that q, therefore it is not true that p (Schayer 1933, 91, 2001, 28) – or modus ponens
– if it is not true that q, then it is not true that p; it is not true that q, therefore it is not true that
p (Matilal 1998, 37). But a preoccupation with the logical form of the argument may in part
miss the point. The Points of Discussion is concerned with more than mere syllogism.
I have set out in full just the first section of the first refutation, which, as noted, contains the
essentials of the argument. Warder (1963, 64–65, 67–68) and Watanabe (1983, 163–67) analyze
and set out the logical steps involved in each of the five sections of the first and second refuta-
tions; Watanabe (1983, 167–74) continues by analyzing the initial sections of the third–eighth
refutations. But even if we take just the first and second refutations as constituting a full debate,
we have ten sections with a total of forty-two steps. As Warder (1963, 66) notes, this seems
‘unnecessarily long-winded’ and clearly can never have been intended as a neat syllogism.
The reasons for the redundancy of much of the argument from a logical point of view remain
unclear. Repetitions that appear redundant are a feature of early Buddhist literature and, besides their
mnemonic purpose in a literature that was composed and transmitted orally, they seem also to have
served a meditative purpose. Certainly to construct and recite, on the basis of the template of the
first point of discussion, the sections of the first two refutations for subsequent points would demand
sustained attention and considerable mental agility in order to follow the inversion of statements and
swapping of terms. Moreover, the proliferation of layers of repetition appears to be a deliberate liter-
ary device employed in the canonical Abhidhamma works, which, as the commentaries point out,
are supposed to be infinite in extent when expanded in full and thus beyond our grasp (Gethin 2020).
There is also perhaps a rhetorical purpose in the redundant arguments: in each debate, the opponent
is shown to resort to fallacious sophistry (chala) and so is thoroughly defeated. And yet this does not
mean that the opponent is not allowed some good arguments, for a basic purpose of the Points of Dis-
cussion is also to probe the implications and coherence of the teaching of the Buddha as set out in the
received discourses of the Buddha, which both the Theravādin and opponents accept as authoritative.
Nonetheless, despite the logical redundancy of the series of refutations, this formalized
system of debate and stereotyped analysis appear to mark a significant step in the development
of Indian logic.6 Schayer thus suggests that the basic form of argument reveals that the author
of the Points of Discussion understood the ‘definition of implication’ (namely, the equivalence
of ‘if p then q’ and ‘that p and not q, is false’) and the law of transposition (the equivalence of
‘if p then q’ and ‘if not q, then not p’) (1933, 92, 2001, 28).
The application of the method: two examples

Having considered the basic method of argument of the Points of Discussion, I shall briefly
consider its application to two different disputed points: one purely conceptual and abstract
165
Rupert Gethin
(that past, future, and present realities exist), the other with more immediate and practical
implications since it concerns something that might be adduced as evidence of the attainment
of awakening or not (that arhats may have seminal emissions).
Past, present, and future realities exist

The sixth disputed point of the first chapter is concerned with the position that fundamental and
irreducible realities (dharmas) – the mental and physical ‘events’ that make up the world – of the
present, past, and future exist (Kv 115–143). This is, of course, the position held by the Sarvāstivāda
school. This point of discussion is of particular interest in the present context since we possess
an account of the same dispute in a text belonging to a different school that must be more or less
contemporary with the Points of Discussion. The Consciousness Group (Vijñānakāya) – a text of
the Sarvāstivāda Abhidharma-piṭaka that survives only in Chinese translation (T. 1539) – opens
with an account of a debate between an advocate of the Sarvāstivāda position and an opponent
named Maudgalyāyana (Muganlian). It seems likely, though it cannot be proved, that this oppo-
nent should be identified with Moggaliputta Tissa (La Vallée Poussin 1925, 344).
In the Points of Discussion, the debate begins with a simple statement of the Sarvāstivāda
position: ‘Everything exists?’ The opponent assents. In trying to refute such a view, the advo-
cate of Theravāda interrogates the opponent by placing a series of propositions and questions,
which should be understood as debated according to the template set out previously. This
method is such that it does not allow for an initial account of why some Buddhist thinkers
might want to hold such a position, though at the close of the debate, the opponent is allowed
to quote some sūtra sources that give some grounds for the position. The assumption must
be that those who come across this discussion already know what it is about. I will return
to the grounds for such a position in what follows after first considering the attempt by the
Theravādin advocate to demonstrate its incoherence.
The initial sections seem to aim at confirming that the opponent does indeed wish to main-
tain the counterintuitive position that things – the five bundles (skandha) of material form,
feeling, perception, volitional forces, and consciousness – that have ceased (the definition
for what is past) or have not yet arisen (the definition for what is future) nonetheless exist
(Kv 115–119, §§ 1–4). The argument is that this is self-evidently incoherent: ‘ceased’ means
ceased to exist, and ‘not yet arisen’ means not yet existing.
The Theravādin advocate’s next move is to argue that the opponent’s position must surely
compromise the doctrine that all irreducible realities, all dharmas (apart from nirvāṇa), are
impermanent, which is a core principle of Buddhist doctrine. Indeed, according to the oppo-
nent’s position, material realities must retain their intrinsic material quality in the future, in
the present, and in the past, given that they are said to ‘exist in all times’ (the meaning of
the school’s name). Further, there seems to be an internal incoherence since, while they lose
the qualities of being future and present, dharmas never lose their quality of being past (Kv
120–125, §§ 7–20).
Following this, the Theravādin suggests that the opponent’s position has absurd conse-
quences: it leads to a position where we must say that the eye that exists in the past sees a
visible form that is past with a cognition that is past; similarly, a future eye must see a future
visible form and a present eye a present visible form (Kv 126–129, §§ 23–28). The argument
then shifts to the situation of the noble ones, such as arhats, who, by definition, have eradicated
greed, hatred, and delusion: if their past greed, hatred, and delusion exist, then why do they not
affect them? For the opponent accepts that an ordinary person’s past greed, hatred, and delusion
still affect them, since he accepts them as being existent in all times (Kv 131–136, §§ 35–46).
166
The Points of Discussion continues by spelling out further absurd consequences of the
opponent’s position: that if past hands, and so on, exist, then all the actions performed by those
hands must exist (Kv 136–137, §§ 47–49); that there must exist three instances of everything,
one in the future, one in the present, one in the past (Kv 137–138, §§50–54). Then once more
it returns to what seems to be regarded as the fundamental point that when we ordinarily say of
things that they are ‘past’ or they are ‘future,’ we must mean that they do not exist. Yet asked if
what is past exists, the opponent replies yes; asked if what exists is past, the opponent replies
that what exists may be past or it may not be past, since he holds the position that things that
are present and future also exist. But this, the Theravādin suggests, is to say that what is past
is not past (since it exists), and what is not past is past (since both are alike in existing). This
applies also to the future and the present (Kv 138–140, §§ 55–58). The debate ends with the
Theravādin and opponent each citing sūtra sources that apparently lend support to their respec-
tive positions (Kv 140–143, §§ 59–64).
It is possible to distinguish three lines of argument in the overall discussion. The first con-
cerns linguistic usage (to say of things they are past or future can only mean they do not exist).
The second concerns the implications of the disputed position for Buddhist doctrine (e.g., it
compromises the fundamental doctrine that things are impermanent). The third attempts to
demonstrate that the position has absurd consequences (there must be three of everything).
Turning to The Consciousness Group, which presents Sarvāstivādin arguments from their
own perspective, the basic method of argument is similar. The Sarvāstivādin repeatedly puts his
opponent Maudgalyāyana’s position to him: the past and future do not exist; only the present
and what is beyond conditions (nirvāṇa) exist. When Maudgalyāyana assents, the Sarvāstivādin
confronts him with a statement to which he must say yes or no: initially, in this debate, whether
it is possible to know that greed, hatred, and delusion are ethically unwholesome (akuśala).
Maudgalyāyana agrees that it is. He must then answer whether the greed, hatred, and delusion
that are the object of such a cognition are past, future, or present. If he says past or future, he is
defeated, because they must nonetheless exist as the objects of the cognition. If he says present,
he must admit that a person can have two cognitions at once: (i) an unwholesome cognition
associated with greed and, at the same time, (ii) a wholesome cognition that takes the first cog-
nition as its object and understands the greed to be ethically unwholesome. But, according to an
accepted Abhidharma principle, two simultaneous cognitions are not possible. The Conscious-
ness Group proceeds by adducing other circumstances that seem to demand that something
past or future must exist in some sense (La Vallée Poussin 1925, 346–49; T. 1539, 531a27–c2).
Bronkhorst (1993) has suggested that a comparison of this disputed point in the Points of
Discussion and The Consciousness Group reveals that the author of the Points of Discussion
was unaware of the Sarvāstivādin arguments for their position. I am not so sure. Bronkhorst
claims (on the basis of Kv 127–29, §§ 1.6.23–28) that the position of the Points of Discussion
is that only present cognitions of present objects with a present sense organ are possible; this
shows that the author of the Points of Discussion was unaware of the Consciousness Group’s
arguments based on the cognition of past objects. But I think Bronkhorst misreads the Points
of Discussion’s attempts at reductio of the Sarvāstivādin position. First, it is quite clear that
the Theravāda Abhidhamma did not deny that past material and mental dharmas might be
the objects of present cognition (Dhs 187, §§ 1041–3; 241, §§ 1417–8); they simply denied
that such objects should be said to exist. Whether their position is coherent is not the issue
in the present context. The point is that neither the Points of Discussion nor The Conscious-
ness Group shows explicit awareness of their opponent’s arguments for their positions. But
this does not mean they were unaware of them. On the contrary, the authors of these treatises
assume a certain familiarity with the background to these disputed points. The strategy of both
167
Rupert Gethin
is to bypass the opponent’s specific arguments and to identify and attack his position’s weak
points directly by drawing out its implications with reference to shared axioms of Buddhist
doctrine – and thereby demonstrate its incoherence.
Thus, in the first disputed point of the Points of Discussion, the opponent does not attempt
to set out the positive philosophical arguments for the position that the person exists, nor does
the Theravādin set out the positive philosophical arguments for not-self. Rather, the debate is
an intra-Buddhist affair that hinges on a single point: if you want to say the person exists, then
you can only meaningfully make this claim if you mean that the person exists in precisely the
same way that systematic Buddhist thought, in the form of the Abhidharma, states that other
things (namely, dharmas) exist. The argument is only effective because the opponent, as a
Buddhist, must avoid saying the self exists in these terms.
The arhat has impure emissions of semen

The second chapter of the Points of Discussion opens with a discussion of the proposition
that an arhat (one who is awakened and has eradicated all greed, hatred, and delusion) may
have impure seminal emissions. The commentary suggests that such a position is maintained
by those who have noticed that some whom they believe to be arhats have seminal emissions;
according to the Theravādins, though, these are not genuine arhats – they claim arhatship
through self-delusion, conceit, or deceit (Kv-a 55). That an arhat might have seminal emis-
sions is one of five points (all found in the Points of Discussion, though not grouped as five
points) referred to in some later Buddhist sources as ‘the five points of Mahādeva,’ which all
focus on what it means to be an arhat (Cousins 1991).
The opponent assents to the proposition that an arhat has impure emissions of semen. The
Theravādin suggests that this is incompatible with other propositions that the opponent also
accepts, namely (i) that an arhat has no desire; and (ii) that it is ordinary men with desire who
have seminal emissions. The opponent continues to maintain his initial thesis according to
which the arhat, despite having no desire, has seminal emissions (Kv 163–164, §§ 1–2).
The Theravādin then switches to a different line of questioning: if not through desire, how
precisely do arhats have seminal emissions? The opponent claims there are mischievous gods
belonging to the retinue of Māra (the god who, in Buddhist texts, continually tempts with the
objects of the five senses, attacking the Buddha on the night of his awakening and offering
him his daughters): these gods provide the arhat’s seminal emission. The Theravādin then asks
if these gods bring their own seminal emissions, that of others, or the arhat’s, or if the emis-
sion exudes from the pores or the skin. The opponent denies all these possibilities, and the
Theravāda suggests a contradiction (Kv 164–165, §§ 3–4).
Again, the Theravāda switches to a different line of questioning: why do the gods pro-
vide seminal emissions? The opponent suggests that it is to induce doubt in the arhat. The
Theravādin asks if arhats have doubt. If the opponent answers no, he must acknowledge the
contradiction; if he answers yes, then the Theravādin asks whether the doubt is in the Buddha,
the Dharma, the Sangha, and so on. The opponent denies that it is doubt of this sort. Again,
there is an implied contradiction (Kv 165–166, §§ 5–6).
The Theravādin now switches to physiological matters. First, where does the seminal emis-
sion come from? The opponent replies that it issues from food and drink. The Theravāda asks
whether all those who eat and drink, including women and eunuchs, have seminal emissions.
The opponent admits that this is not so. Further, if semen is produced by food and drink, there
should be somewhere in the body where it is stored, as there is for feces and urine. The oppo-
nent again denies this (Kv 166–167, §§ 7–8).
168
The Theravādin now returns to matters related to the initial line of questioning. If the arhat
has seminal emissions, is he interested in such things as sex, having a family, perfumes, and
money, as ordinary men who have seminal emissions are? The opponent denies this. The
Theravādin asks whether the arhat has abandoned greed, hatred, delusion, and other afflic-
tions; whether he has developed the constituents of awakening and the path and understood the
four truths. The opponent agrees he has. The Theravādin suggests this is not compatible with
having seminal emissions (Kv 167–170, §§ 9–13).
At this point, the opponent changes tack. When asked if an arhat has seminal emissions,
instead of responding simply yes, he replies that an arhat ‘skilled in his own practice’ (sadham-
makusala) does, but an arhat ‘skilled in superior practice’ ( paradhammakusala) does not. This
terminology only occurs in the Points of Discussion. We can do little better than follow the
commentarial explanation of the distinction: the former is an arhat who has reached awaken-
ing without developing full concentration (samādhi) based on the four absorptions ( jhāna) and
four formless meditation attainments (samāpatti), while the latter is one who has developed
all eight. The significant point is that the opponent has introduced a distinction between arhats
that the Theravādin now argues is beside the point. He does so by revisiting previous lines of
question: Are the two types of arhat alike in having abandoned greed, hatred, and delusion and
other afflictions? Are they alike in having developed the constituents of awakening and the
path? The opponent agrees they are. The Theravāda once again suggests the opponent’s posi-
tion is incoherent. (Kv 170–172, §§ 14–20).
In response to the opponent’s restating his position, the Theravādin now cites a canonical
source (Vin I 295) that states explicitly that there is no occasion on which an arhat’s semen
might be released (Kv 172, § 21). The opponent acknowledges the source; the Theravādin sug-
gests this is a final contradiction. But this is not quite the end of the discussion. The opponent
takes the initiative and asks if the Theravādin denies that an arhat can have seminal emissions.
The Theravādin denies this. The opponent returns to the question of providing things for an
arhat: can people provide robes and the other monastic requisites for arhats? The Theravādin
agrees they can. The opponent now asks whether the Theravādin agrees that, since others can
provide arhats with robes, there is provision of things by others. The Theravādin agrees (Kv 172,
§§ 22–23). The discussion thus ends with the opponent claiming that since provision of things
by others is admitted, the Theravādin should not deny that an arhat can have seminal emis-
sions. The point is presumably that if gods can provide the material requisites for arhats, those
of Māra’s ilk might theoretically provide unwanted things, such as semen (Cousins 1991, 42).
The overall argument in the case of this disputed point revolves around three principal
issues: (i) the connection between seminal emissions and sexual desire; (ii) the physiology
of the production and storage of semen; and (iii) whether a relevant distinction can be made
between types of arhat.
Conclusions
The Points of Discussion exemplifies a particular style of debate and method of argumentation
focused on disputed points of Buddhist doctrine. The method of argumentation works by dem-
onstrating the incompatibility of two propositions. The first is a particular position taken on the
disputed point; the second may be an accepted point of Buddhist doctrine, something gener-
ally accepted to be true by ordinary human beings, or some unwanted implication of the first.
The incompatibility is demonstrated using what is essentially a logical formula, although it is
elaborated rhetorically. The opponent is in this way cornered, having either to deny essential
points of Buddhist doctrine or admit absurd consequences in order to maintain the position.
169
Rupert Gethin
To what extent we might view the style of debate, the method of argumentation, and the
compilation of a set of disputed points as the work of a particular author (call him Moggali-
putta Tissa) is unclear. What is clear is that the set of disputed points is easily added to; the
same is true of the supplementary points that are introduced in the course of debating the
principal point. The parallels to some parts of the Points of Discussion that are found in The
Consciousness Group suggest a situation in which Buddhist thinkers attempting to articulate
the finer points of Buddhist doctrine were engaged in debate with each other. Such debates
would have been a context for the exchange and development of methods of argumentation.
That Moggali (of Theravāda sources)/Maudgalyāyana (of Sarvāstivāda sources) was a sig-
nificant figure at some point in this process seems quite possible. The Middle Indian diction
of certain of the repeated phrases of the Points of Discussion is distinctly eastern, which sug-
gests that something of its core has relatively early origins in Magadha close to the ancient
Mauryan capital of Pāṭaliputra.
The question of the valid means for establishing knowledge ( pramāṇa), which became
central to Buddhist philosophy from the fifth century CE, is not explicit in the Points of Dis-
cussion, but two valid means of knowledge are implicitly accepted in the text: the authority of
scripture (sūtra) and reason ( yukti). These are put to use in a particular way. Denied the oppor-
tunity to articulate their positions at length, opponents generally must answer yes or no, and
are thereby continually confronted with the unacceptable and absurd consequences of their
positions. It seems not unreasonable to suggest that such a method of argumentation paves the
way for someone like Nāgārjuna, who exploits such a method to its full: whatever position you
adopt categorically, there will be unwanted consequences and incoherence.
Abbreviations
As Atthasālinī. Müller, Edward, ed. 1897. London: Pali Text Society.
Dhs Dhammasaṅgaṇi. Müller, Edward, ed. 1885. London: Pali Text Society.
Dhs-mṭ Dhammasaṅgaṇi-mūlaṭīkā. 1960. Rangoon: Buddhasāsana-samiti.
Kv Kathāvatthu. Taylor, Arnold Charles, ed. (1894, 1897) 1979. London: Pali Text
Society.
Kv-a Kathāvatthuppakaraṇa-aṭṭhakathā. Jayawickrama, N. A., ed. 1979. London: Pali
Text Society.
T Taishō Shinshū Daizōkyō. Junjirō Takakusu and Kaigyoku Watanabe, ed. 1924–
1932. Tokyo: Taishō Issaikyō Kankōkai
Vin Vinayapiṭaka. Vols. I–V Oldenberg, Hermann, ed. 1879–1883. London: Pali Text
Society.
Notes
1 I am grateful to Pierre-Julien Harter for his careful reading and his comments and suggestions.
2 There is some variation in the editions on what counts as a separate point of discussion, especially in
chapters nine and twenty-three. The PTS edition counts 216 points, the Burmese edition 226.
3 Kv-a 2.2–3; explained at Dhs-mṭ 15 as yuttī ti upāyo, vādassa yutti vādayutti, vādappavattanassa
upāyo ti attho.
4 The number of steps in each section is stated in the text itself, but how to count those steps is taken
from the commentary. As Warder (1963, 66) notes, it is not clear why the first section is said to consist
of five steps and the second four, since they both have the same form.
5 The commentary suggests that those affirming the existence of the person include Buddhists and
non-Buddhists: ‘Who affirms the person? Within the teaching [of the Buddha], the Vajjiputtakas and
170
Sammitiyas, and outside it, many followers of other systems of thought.’ (Kv-a 9.16–18: ke pana
puggalavādino ti. sāsane Vajjiputtakā c’ eva Sammitiyā ca bahiddhā ca bahū aññatitthiyā.)
6 Bocheński (1961, 422) criticizes Randle’s’ assessment (1930, 13–14) of the Kathāvatthu as exhibiting
disregard for logic.
Bibliography
Aung, Shwe Zan, and Caroline A. F. Rhys Davids. 1915. Points of Controversy or Subjects of Discourse:
Being a Translation of the Kathā-Vatthu. London: Pali Text Society.
Bareau, André. (1955) 2013. The Buddhist Schools of the Small Vehicle. Honolulu: University of Hawai’i
Press.
Bocheński, Józef Maria. 1961. A History of Formal Logic. Notre Dame: University of Notre Dame Press.
Bronkhorst, Johannes. 1993. “Kathāvatthu and Vijñānakāya.” In Premier Colloque Étienne Lamotte:
Bruxelles et Liège, 24–27 Septembre 1989, 57–61. Louvain-la-Neuve: Université catholique de Lou-
vain, Institut orientaliste.
Cousins, Lance Selwyn. 1991. “The ‘Five Points’ and the Origins of the Buddhist Schools.” The Buddhist
Forum 2: 27–60.
Frauwallner, Erich. 1995. Studies in Abhidharma Literature and the Origins of Buddhist Philosophical
Systems. Albany, NY: State University of New York Press.
Gethin, Rupert. 2020. “Reading Repetitions in the Saṃyutta-nikāya and Early Abhidhamma: From the
Mahā-vagga to the Dhammasaṅgaṇi.” In Research on the Saṃyukta-āgama, edited by Dhammadinnā,
109–169. Taipei: Dharma Drum Corporation.
Jayawickrama, N. A. 1962. The Inception of Discipline and the Vinaya Nidāna: Being a Translation and
Edition of the Bāhiranidāna of Buddhaghosa’s Samantapāsādikā, the Vinaya Commentary. London:
Luzac.
La Vallée Poussin, Louis de. 1925. “La controverse du temps et du pudgala dans le Vijñānakāya.” In
Études Asiatiques publiées à l’occasion du vingt-cinquième anniversaire de l’École française
d’Extrême-Orient, 343–76. Paris: G. Van Oest.
Law, Bimala Churn. 1940. The Debates Commentary (Kathāvatthuppakaraṇa-aṭṭhakathā). London: Pali
Text Society.
Matilal, Bimal. 1998. The Character of Logic in India. Albany, NY: State University of New York Press.
Randle, H. N. 1930. Indian Logic in the Early Schools. London: Humphrey Milford.
Schayer, Stanislaw. 1933. “Studien zur indischen Logik. II. Altindische Antizipationen der Aussagen-
logik.” Bulletin International de l’Académie Polonaise des Sciences et des Lettres, Classe de Phi-
lologie (1–6): 90–96.
———. 2001. “Studies on Indian Logic. II. Ancient Indian Anticipations of Sentential Logic.” Translated
by Piotr Balcerowicz. In Materials of the International Seminar: Argument and Reason in Indian
Logic, edited by Piotr Balcerowicz and Marek Mejor, 27–33. Warsaw: Instytut Orientalistyczny, Uni-
wersytet Warszawsk. www.balcerowicz.eu/indology/schayer2001.pdf.
Warder, Anthony Kennedy. 1963. “The Earliest Indian Logic.” In Trudy Dvadtsat’ Piatogo Mezhdun-
arodnogo Kongressa Vostokovedov, Moskva 9–16 Avgusta 1960 [Proceedings of the Twenty-Fifth
International Congress of Orientalists], vol. IV, edited by B. G. Gafurova, 56–68. Moscow: Izd-vo
vostochnoi lit-ry.
Watanabe, Fumiro. 1983. Philosophy and its Development in the Nikāyas and Abhidhamma. New Delhi:
171
11
KĀTYĀYANĪPUTRA AND
THE LARGE COMMENTARY
(MAHĀVIBHĀS․Ā)
The Development of Abhidharma Literature
and of a Sarvāstivāda Self-identity
Bart Dessein
As Buddhist doctrines spread across the Indian subcontinent, during and immediately pre-
ceding the Mauryan period, local monastic communities formed in the early stages of Bud-
dhist history adapted to the local circumstances of the regions where they settled. Distinctive
monastic schools (nikāya) began to develop in the fourth–second centuries BCE, each based
upon the acceptance of particular interpretations of monastic discipline (Bechert 1973, 8). This
period is often referred to as Nikāya Buddhism, and is characterized by textual reorganization:
whereas the sacred texts of Buddhism were, at first, grouped together according to their style,
this initial classification gradually gave way to the well-known “three baskets” (tripiṭaka)
schema that groups texts according to content. Texts that describe the life of the historical
Buddha and his sermons were collected in the Sūtra literature. Texts relating to monastic life
were grouped in the Vinaya literature. Philosophical treatises were grouped in the Abhidharma
collection. Buddhist historiography traces the origin of the Sūtra and the Vinaya literature to
two direct disciples of the Śākyamuni Buddha: Ānanda and Upāli, respectively. Abhidharma
texts postdate Sūtra and Vinaya texts. They typically cull lists of doctrinal items (dharma) and
matrices (mātṛkās) from the Sūtra and Vinaya texts and bring them into a more or less organ-
ized philosophical system. It is very likely that the term “Abhidharma” refers to such lists of
doctrinal items. One possible interpretation of the term “Abhidharma” is thus “concerning
the components (dharma) of existence,” a term that also reflects the ontologies developed in
Buddhist philosophy.1 The use of numerical lists in Sūtra, Vinaya, and Abhidharma literature
points to the overall oral/aural Indian literary context in which Buddhism originated. These
numerical lists hereby served as mnemotechnical devices in the oral recitation of texts. The
monastic code of the Sarvāstivādins refers to both the oral origin of the Abhidharma and to the
numerical lists as the backbone of these texts (in this case, a series of numerical lists compris-
ing five elements each) with the following words that are attributed to Ānanda, a direct disciple
of the Buddha (T.23.1435, 449a19 ff.):
Ānanda was further asked: “Where did the Buddha start to preach the Abhidharma?”
Ānanda answered: “Thus have I heard: At one time, the Buddha was staying in
172 DOI: 10.4324/9781351030908-18

Kātyāyanīputra and Mahāvibhās∙ā
Śrāvastī. At that moment, the Buddha told all bhikṣus (monks): ‘When talking about
the five . . ., five . . ., five . . ., five. . . .’ ”
As the historical Buddha and his early disciples faded into the past, discussions between dif-
ferent monastic groups and subgroups about the correct interpretation of the often enigmatic
numerical lists around which the Buddha had formulated his sermons increased. This con-
tributed to the development of doctrinal sects within existing monastic communities. The
Abhidharma literature is therefore also the literature of so-called sectarian Buddhism. Bud-
dhist historiography refers to this phase of sectarian Buddhism as the eighteen schools of the
Buddhism of the Disciple (Śrāvakayāna). A more common, albeit denigrating, name to refer
to these eighteen schools is Lesser Vehicle Buddhism (Hīnayāna). This name contrasts early
Buddhism with Mahāyāna or Greater Vehicle Buddhism that gradually developed from within
the different schools and sects of early Buddhism.
While the Abhidharma tradition may have had an oral origin, the texts as we possess them
now are the product of an expanding written tradition. This is the logical outcome of both the
increasing complexity of philosophical debate and the need for sectarian self-identification
and concomitant recognition of specific philosophical texts as peculiar to one’s own group.
It is to the self-identification of the Sarvāstivādins as a distinct group – a process that post-
dates the reign of the famous king Aśoka the Maurya (r. 268–232 BCE) and the committing
to writing of philosophical texts – that the historical figure of Kātyāyanīputra, a Buddhist
scholar born to a Brahman family around the beginning of the common era, is connected.2 The
commentator Paramārtha (499–569 CE) acknowledges this distinctive role of Kātyāyanīputra
in his Compilation of the Investigation into the Secrets of the Subtle Meaning of the Three
Treatises (San lun xuanyi jian you ji 三論玄義檢幽集) when stating that:
In the beginning of the three hundred years after the demise of the Buddha,
Kātyāyanīputra left this world, having invoked the creation of two schools. . . . Hav-
ing come to the middle of the three hundred years after the demise of the Buddha,
two schools were formed because of minor reasons: the first are the Sarvāstivādins
who are also called the Hetuvādins, and the second are the Haimavatas who are also
called the Sthaviravādins.
(T.70.2300, 363a8–12)
Tradition has it that it was on the occasion of a council that was convened by the Kuṣāṇa king
Kaniṣka (c. 127–150 CE) that the Sarvāstivādins committed their canon to writing. Tarānātha’s
History of Buddhism (Chos ’byung) more precisely informs us that at this council, the Sūtras and
the Abhidharma that had not yet been committed to writing were written down, and those texts
that had already been written down were purified. The reason given for committing these texts to
writing was that ordinary people recited the scriptures incorrectly, making omissions and interpo-
lations (Obermiller 1931–1932, vol. 2, 101). This statement implies that the oral medium of trans-
mitting texts must – at least for some period – have coexisted with a written transmission. This also
renders plausible the possibility that some texts were transmitted in both ways simultaneously.
The importance of this for Kātyāyanīputra’s text is that it is very likely that two versions existed
simultaneously: the Abhidharma Treatise in Eight Sections (*Abhidharmāṣṭagranthaśāstra or
*Abhidharmāṣṭaskandhaśāstra) and the Source of Knowledge (Jñānaprasthāna).
The self-identification of the Sarvāstivādins as a distinct philosophical group hinges on
their standpoint that all entities (dharma) exist in the past, present, and future. That is to say,
through the dynamics of causality, it is impossible for any entity to exist in the present without
173
Bart Dessein
the causes that shaped it but that, at present, already belong to the past. All present entities
cast forth their future existence through the continuous operation of causality. In this sense,
all present entities will actually also exist in the past, when they will have become the cause
of that as of yet still future entity that will, at that moment, become a present entity. It is this
doctrine that has given its name to this school: sarvam asti, meaning “all exists.” This highly
ontological trait of Sarvāstivāda thinking is also representative of a more general development
within Buddhism from a stage that focused on an individual’s liberation from their psychologi-
cal suffering that was prominent in early Sūtra texts toward a doctrine that engaged with issues
like the world in a temporal and a spatial sense and an individual’s position within this world.3
Within the Sarvāstivāda monastic community, there may have been agreement on the fact that
“all exists.” However, ongoing philosophical discussions within this particular monastic group
gradually led to the formation of two prominent doctrinal subgroups: the Sarvāstivādins of Bac-
tria and Gandhāra, and the Vaibhāṣikas of Kaśmīra. Though each subgroup took Kātyāyanīputra’s
text as a foundation, they engaged with different aspects of his works and formulated their ideas
in different types of texts: while the more conservative group of Bactria and Gandhāra compiled
pedagogical digests that organized and systematized the materials Kātyāyanīputra had selected
in his Abhidharma Treatise in Eight Sections, the Sarvāstivāda subgroup that moved to Kaśmīra
declared the Source of Knowledge to be the primary body (śarīra) of Sarvāstivāda thinking,
identifying six Abhidharma texts that predated the self-identification of the Sarvāstivādins as a
separate doctrinal group as the support or feet ( pāda).4 They compiled extensive and polemical
commentaries on this Source of Knowledge version. One of these commentaries is the Large
Commentary (Mahāvibhāṣā), a text that also contains Mahāyāna elements.
Kātyāyanῑputra and the Abhidharma Treatise in Eight Sections/Source

of Knowledge
Authorship and Extant Versions of the Text

While the Sanskrit and the Tibetan traditions attribute the Abhidharma work that is enti-
tled Source of Knowledge (Jñānaprasthāna) to Kātyāyanīputra,5 the Chinese tradition takes
Kātyāyanῑputra as the creator (Chinese zao 造) of an Abhidharma work that is either called Api-
damo fazhi lun 阿毘達磨發智論 (Source of Knowledge) or (Jñānaprasthāna) Apitan ba jiandu
lun阿毘曇八犍度論 (Abhidharma Treatise in Eight Sections) (*Abhidharmāṣṭagranthaśāstra
or *Abhidharmāṣṭaskandhaśāstra) (T.26.1543, 771b16; T.26.1544, 918a5). Kātyāyanῑputra is
characterized as the compiler rather than the composer of the Source of Knowledge in the
Large Commentary (on the Abhidharma) (*[Abhidharma]mahāvibhāṣā[śāstra], Apidamo da
piposha lun 阿毘達磨大毘婆沙論), where we read that Kātyāyanῑputra “received the lec-
tures” (shouchi yanshuo受持演說) the Buddha gave in a question-and-answer format and
compiled them “in order to transmit them widely” (guang ling liubu廣令流布) (T.27.1545,
1a11–19). This suggests that the concept of authorship for Abhidharma works is not sim-
ply attribution of the text to an original figure who composed it; rather, Abhidharma texts
used doctrinal materials that were transmitted over time. It also suggests that these materials
were shaped into philosophical constructs that conformed to the developing interpretations
of the words of the historical Buddha. Furthermore, it makes it clear that the argument of
authority was an important issue for Abhidharma texts, as the compilers presented the Abhi‑
dharma texts to be the result of lectures received from the Buddha.6 The history of the transmis-
sion of the Abhidharma Treatise in Eight Sections reveals that there were several translations
174
in Chinese from presumably different “original” texts written in different languages (Middle
Indic and Sanskrit), which makes it difficult to speak of one original text identifiable as the
work of Kātyāyanīputra (Demiéville 1961, 472 n. 5; Takakusu 1904–1905, 84).
Giyū Nishi (1934) analyzed the doctrinal positions associated with different doctrinal
groups in the extant vibhāṣā compendia and concluded that the Source of Knowledge and the
Abhidharma Treatise in Eight Sections represent different lineages of transmission within the
larger Sarvāstivāda group. According to Ryōgen Fukuhara (1965, 218–19), the translitera-
tions in the Abhidharma Treatise in Eight Sections indicate that the work may have been the
Gandhāran version of the text, while the Source of Knowledge represents the recension of the
work by the Kāśmῑri Vaibhāṣikas. It therefore is not unlikely that Kātyāyanῑputra compiled his
Abhidharma Treatise in Eight Sections in central India, before the Sarvāstivāda school began
to flourish in Kaśmῑra, and that the Source of Knowledge is a reworked version of that text.
This corroborates the preceding hypothesis that the Abhidharma Treatise in Eight Sections,
which Kātyāyanῑputra compiled by placing into the mouth of the historical Buddha doctrinal
materials that had been transmitted over time, was later referred to as the Source of Knowledge.
This hypothesis is relevant to the importance of each of these texts – the Abhidharma Treatise
in Eight Sections and the Source of Knowledge – in the development of two new types of texts,
each connected to a different doctrinal subgroup that developed within the larger Sarvāstivāda
monastic community. The subgroup of the Sarvāstivādins of Bactria and Gandhāra, later called
“Sautrāntika” (literally, “those following the Sūtras”), formulated their interpretations of Buddhist
doctrine in summary pedagogical digests in which the essential elements of the doctrine were
given in stanzas (kārikā), followed by an auto-commentary. This helps us understand why the titles
of their texts often use the term hṛdaya (“heart,” “essence”; Chinese xin 心). The Sarvāstivāda
subgroup of Kaśmīra, on the other hand, formulated their interpretations of the words of the Bud-
dha in the form of extensive and polemical prose commentaries on the Source of Knowledge. It
is these prose commentaries, called vibhāṣā (lit. extensive commentary) in Sanskrit, from which
the Sarvāstivāda subgroup takes its name. The Bactrian and Gandhāran subgroup objected to
the Kāśmīri Sarvāstivādin developments and claimed to adhere to the word of the Buddha more
strictly; thus, they were called Sautrāntika to indicate that their doctrinal position was closer to the
original words of the Buddha (as preserved in the Sūtras) than that of the Kāśmīri Sarvāstivādins.
Doctrinal Innovations of the Text: The Path Structure, the Material

World, and the Liberating Power of Understanding
Like the Abhidharma authors before him, Kātyāyanīputra selected numerical lists of doctrinal
items (dharma) from Sūtra and Vinaya texts. The name Abhidharma Treatise in Eight Sections
indicates that this material was organized into eight main sections. These eight are further sub-
divided into forty-four subsections. The Source of Knowledge takes over this general division
into eight sections and forty-four subsections as outlined in Table 11.1.
Kātyāyanīputra states that the “highest mundane factor” (laukikāgradharma) – i.e., the
transition point from an ordinary being ( pṛthagjana) to being a noble (ārya) person – is “aban-
doning the affairs of an ordinary being and obtaining the doctrine of the Buddha; abandon-
ing heterodox affairs and obtaining the true doctrine” (T.26.1543, 771c15–16). Moreover, he
argues, all heterodox views are to be abandoned through one of the four noble truths that
encapsulate Buddhist doctrine: through insight into the truth of suffering, the truth of the origin
of suffering, the truth of the extinction of the origin of suffering, or the truth of the path lead-
ing to the extinction of the origin of suffering (T.26.1543, 772c9–20). Thus, Kātyāyanīputra
175
Bart Dessein
Table 11.1 Organization of Abhidharma materials in the Abhidharma Treatise in Eight Sections and the
Source of Knowledge.
Main Section Subsection Main Topics
1. Miscellany (*saṃkῑrṇa) 1–8 Structure of the path and

T.26.1543, 771b21–784b25 foundational topics in Buddhist
T.26.1544, 918a7–929b2 philosophy
2. Fetters (*saṃyojana) 9–12 Three fetters responsible for
T.26.1543, 784c7–812a12 the arising of ninety-eight
T.26.1544, 929b10-951a6 contaminants
3. Knowledge ( jñāna) 13–17 Knowledge, wisdom, insight, and
T.26.1543, 812a21–841a27 the types of persons who possess
T.26.1544, 951a6–972a8 them
4. Action (karman) 18–22 Virtuous and unvirtuous actions,
T.26.1543, 841b7–854a1 their causes, and their results
T.26.1544, 972a16–981c1
5. The four fundamental material 23–26 The four material elements
elements (mahābhūta) and their derivatives, and the
T.26.1543, 854a21–867a5 relationship of the fetters to
T.26.1544, 981c9–991b17 materiality
6. Faculties (indriya) 27–33 Relationship of the faculties to the
T.26.1543, 867a16–887a16 material world and to the fetters
T.26.1544, 991b18–1007c23
7. Concentration (samādhi) 34–38 Forms of concentration obtained in
T.26.1543, 887b7–905a16 various cosmic realms
T.26.1544, 1008a6–1027b11
8. Views (dṛṣṭi) 39–44 Range of Buddhist and non-
T.26.1543, 905a27–917b8 Buddhist views
T.26.1544, 1027b12–1031c28
introduces the innovative element of a “path structure” – i.e., the path of vision (darśanamārga)
that is the instrument to annihilate obstructions to attaining nirvāṇa – as a foundational topic
of Sarvāstivāda philosophy. The name “path of vision” refers to the idea that it is through
attaining vision and realization of the four noble truths that defilements (kleśa) are gradually
eliminated and eventually will no longer arise, thus leading to the attainment of nirvāṇa.
It is noteworthy, however, that the concept of the “highest mundane factor” and the “path of
vision” are the first topics discussed in Kātyāyanīputra’s work. It thus appears that this text was
not meant for ordinary beings who would first need to be instructed on the nature of “ordinari-
ness” and who, once convinced of the loathsomeness of ordinary life, would start to follow the
Buddhist path, but for adepts (śrāvaka) who are already convinced of the necessity to develop
the path of vision (Armelin 1978, 15). This observation is corroborated by the fact that the
remaining seven subsections of this first main section, “Miscellany,” discuss highly complex
doctrinal topics such as dependent origination ( pratῑtyasamutpāda) and the characteristic marks
(lakṣaṇa) of conditioned factors (saṃskṛta dharma). The issue of dependent origination con-
cerns the list of the so-called twelve members of dependent origination that explains that karmic
activity is the force that continues the cycle of transmigration (saṃsāra). This is a list of predom-
inantly psychic concepts such as ignorance regarding the Buddhist doctrine, the primary per-
ceptual consciousness of a newborn baby, contact (sparśa) of the senses with the outside world,
the feelings that result from these contacts, the craving for continuation or discontinuation of
176
these feelings, and concepts such as existence, birth, and decay and dying that directly relate to
a human being’s physical existence.7 This list is related to the following famous teaching of the
Buddha that denies the existence of a “self ” and that is to be read in the Vinaya:
Then the Lord addressed the group of the five monks, saying: “Body, monks, is not
self. . . . Feeling is not self. . . . Perception is not self. . . . Consciousness is not
self. . . . What do you think about this, monks? Is body permanent or impermanent?”
“Impermanent, Lord.” “But is that which is impermanent painful or pleasurable?”
“Painful, Lord.” . . . “Seeing in this way, monks, the instructed disciple of the ari-
yans disregards body and he disregards feeling and he disregards perception and he
disregards the habitual tendencies and he disregards consciousness; disregarding he
is dispassionate; through dispassion he is freed; in freedom the knowledge comes to
be: ‘I am freed,’ and he knows: ‘Destroyed is birth, lived is the Brahma-faring, done
is what was to be done, there is no more of being such or such.’ ”8
The twelve links in the chain of dependent origination may explain how transmigration is caused
by karmic activity (for example, craving results from sense contact). However, it does not
explain how karmic activity itself functions. For this reason, a list of four conditions ( pratyaya)
and a redundant list of six causes (hetu) from the Sūtra literature were brought into the philo-
sophical discussion.9 As the concept of “causes and conditions” is less directly linked to the
concept of a human being’s individual life than the concept of the twelve members of depend-
ent origination is, it gradually took over the role of the twelve members to explain the general
principle of causality. Meanwhile, the list of twelve members was restricted to explaining the
rebirth of a human being as one feature of the general process of causality. It is important to note
here that Kātyāyanīputra explains the concept of dependent origination in the subsection on
human beings of the first main section of his text, and tradition has it that when the Abhidharma
Treatise in Eight Sections was translated into Chinese, a section on “causes and conditions”
(hetupratyaya) was at first forgotten and added later. This may suggest that: (1) Kātyāyanīputra
had intended his text for the adept who envisaged his personal liberation, not for a larger audi-
ence of ordinary human beings; and (2) that it was only after the compilation of the Abhidharma
Treatise in Eight Sections that the concept of “causes and conditions” gained prominence.
Intrinsically linked to the concept of “dependent origination” are the concepts “birth”
(utpāda), change in continuance (sthityanyathātva), and passing away (vyava) as character-
istic marks of the conditioned. In the subsection that deals with these characteristic marks
of conditioned factors, the question of whether entities that transgress through time because
of the dynamics of causality exist as actual entities in the three time periods of past, present,
and future is addressed. While all present entities are the karmic result of a cause that, in the
present moment of time, already belongs to the past, they cannot exist without this past cause.
Given the dynamics of causality, all present entities continuously form a future result. The
momentariness of time implies that all present entities instantaneously become entities of the
past and will live on as causes for the future entities that are now being formed. Causality and
momentariness thus make it impossible that any conditioned factor would possess an enduring
self-nature (svabhāva). Commenting on the Source of Knowledge, the Large Commentary (on
the Abhidharma) states:
Furthermore, another reason why the Source of Knowledge presents this topic is that
there are some who are deluded with regard to the self-nature of the entities of the
three time periods, denying the existence of the past and future entities, and who
177
Bart Dessein
maintain that the present entities are unconditioned. To repudiate their position, it
is shown that the essence and the characteristic marks of the past and future entities
truly exist, and that the present entities are conditioned.
(T.27.1545, 393a18–20)
It is because the Buddhist adept knows that all conditioned factors (including human beings
as conditioned factors) are without self-nature that he turns away from worldly life, seeking to
annihilate all volitional actions that would form fetters that would bind him to this world. The
introduction of concepts such as a path of vision, dependent origination, and factors that arise
in dependency upon causes (hetu) and conditions ( pratyaya) thus all relate to the ultimate aim
of Buddhist practice: the removal of defilements that is liberation.
In the four subsections of the second main section, “Fetters,” Kātyāyanīputra analyzes
defilements in more detail. This section explains how three fetters (saṃyojana) are respon-
sible for the arising of ninety-eight contaminants (anuśaya), i.e., proclivities that result in
the karmic activities of human beings. The text states (T.26.1543: 785c19–20): “How many
of these three fetters have a retribution (savipāka) and how many do not have a retribution?
Answer: All that is unvirtuous (akuśala) has a retribution.” These three fetters are: (1) the
view that there is a self (satkāyadṛṣṭi); (2) doubt (vicikitsā) regarding whether defilements
can be annihilated through vision of the four noble truths; and (3) the view that merely adher-
ing to precepts and vows will lead to liberation (śῑlavrataparāmarśa). The view that there is
a self can be regarded as the fundamentally erroneous conviction that causes a human being
to perform volitional actions for the benefit of maintaining human life and making it pleasur-
able.10 Each of these volitional actions relates to craving (tṛṣṇā) and is incited by a proclivity.
All volitional actions have a karmic result and lead to rebirth and suffering in a future life.
Knowledge of the non-self of all conditioned factors and the conviction that insight in the four
Buddhist truths can annihilate these proclivities, both of which are addressed in the first main
section of the text, will bring the cycle of rebirths to a halt. Conversely, doubt regarding the
efficacy of insight into the noble truths in eliminating defilements may put the working of the
proclivities back in operation. As Kātyāyanīputra writes (T.26.1543, 785a23), “In the sense
realm (kāmadhātu) of ordinary existence, doubt is unvirtuous.”
While these first two fetters are related to an adept’s knowledge and understanding, the
third fetter is more related to his sincerity: it is not the keeping of the monastic precepts and
vows as such that will determine whether or not an adept will attain liberation, but rather the
degree of his conviction. This final fetter suggests that Kātyāyanīputra indeed intended a read-
ership of monastics.
This section also explains the annihilation of the contaminants in its relation to the three
realms of the Buddhist cosmology: the sense realm (kāmadhātu) of ordinary existence; the
material realm (rūpadhātu) where, through meditation, sensual desire for the material is brought
to a halt; and the etheric realm of formlessness (ārūpyadhātu) that pertains to the domain of
higher forms of meditation. This content is presented in a more systematic way in the Heart of
the Abhidharma (*Abhidharmahṛdaya), a text that will be discussed in the following section.
The third main section, “Knowledge,” deals with the types of knowledge ( jñāna) the
adept gains in the different stages that lead to becoming an arhat, i.e., a liberated person.
Numerical lists of types of knowledge are hereby discussed according to their mutual relations
(T.26.1543, 821a17–20). Although it is clear from this discussion that the types of knowledge
are related to the four noble truths and to the different cosmic realms – conventional knowl-
edge relating to the sense realm, and subsequent knowledge relating to the material realm
178
and the realm of formlessness – the precise nature of this relationship is not systematically
explained.
The fourth main section, “Action,” discusses the difference between virtuous, or skillful
(kuśala), and unvirtuous, or unskillful (akuśala), actions, their causes, and their karmic result
in a particular form of rebirth. As with the treatment of the types of knowledge, a multi-
tude of doctrinal lists are mutually interlinked. The question of which type of wrong behavior
is the gravest offense is answered in a way that again points to the monastic readership of
Kātyāyanīputra’s text (T.27.1543, 843a26–27): “Dividing the Saṃgha (Buddhist community)
through lies (mṛṣāvāda) is the gravest offence. Because of this behavior, one will be retributed
with a kalpa-long life in the Avīci hell (the deepest of all hells).”
While the first four main sections of the text thus share their concern with the issues of
defilement, karmic consequences, and the method to reach liberation, the fifth main section,
“The four fundamental material elements,” stands out as the first occurrence in Sarvāstivāda
Buddhism of a discussion of the material aspect of the cosmos. Indeed, while the second main
section of the text brought in the concept of the three cosmic realms, materiality as such was
not yet discussed. This fifth main section explores the four fundamental material elements
(mahābhūta): earth ( pṛthivī), water (ap), fire (tejas), and wind (vāyu), as well as the different
forms of material elements derived therefrom. This materiality is distinguished from the men-
tal domain of thoughts (citta), from thought-concomitants (caitasika) such as mindfulness,
and from those formations that are dissociated from thoughts but cannot be considered mat-
ter either (cittaviprayuktasaṃskāra).11 Toward the end of the section, these material elements
are brought into relation with the “fetters” discussed in the second main section, as it is more
precisely explained how a human being is bound to the material world through fetters that are
the result of sensual activity.
From here, it is a logical step to the sixth main section, “Faculties.” The faculties are also
discussed in their relation to the items of other doctrinal lists. One of the relations addressed
is between the faculties and the material world, leading to defilements and their karmic
consequences.
The seventh main section, “Concentration,” goes on to discuss the different forms of con-
centration that are attained in various cosmic realms. Special attention is paid to the states of
the nonreturner (anāgāmin) and the once-returner (sakṛdāgāmin), two stages preceding the
attainment of arhat-ship.
Finally, the last main section, “Views,” mentions some erroneous views (mithyādṛṣṭi): the
view that the present and the future do not exist (T.26.1543, 913a17); the view that there is
permanence (T.26.1543, 913a26); and the view that human beings have defilements that have
no causes and conditions (T.26.1543, 913b5). These critiques mark the self-identification of
Sarvāstivādins as a distinct doctrinal group within the larger Buddhist community.
In these eight sections, Kātyāyanīputra combined concepts enumerated in doctrinal lists,
from time to time introducing innovative interpretations. He did not, however, construct a philo-
sophical system. This gives the text the impression of a scholastic patchwork. Indumati Armelin
thus characterizes Kātyātanīputra’s text as merely “a testimony of an effort to systematization”
(1978, 6). The lack of system is justified in the Large Commentary (on the Abhidharma) as
follows: “The true nature (bhāva) and characteristic marks (lakṣaṇa) of the entities (dharma)
within the Abhidharma should be sought, not their order of presentation” (T.27.1545, 1c23–24).
As with earlier Abhidharma texts, then, Kātyāyanīputra’s work is not a fully elaborated
philosophical system. Nevertheless, his presentation of doctrinal concepts such as the prin-
ciple of causality; the three time periods of past, present, and future; and the possibility of
179
Bart Dessein
annihilating defilements through applying insight into the four noble truths and thereby bring-
ing an end to the dynamics of causality all became fundamental for later Sarvāstivāda phi-
losophy. This explains why Kātyāyanīputra is often seen as the founder of the Sarvāstivāda
school. Moreover, while Kātyāyanīputra’s text may focus on the adept’s individual liberation,
by bringing in concepts such as “causes and conditions” and the “three time periods,” it also
testifies to an ontological development within Buddhism. This general ontological approach
began to replace Buddhist doctrine as a method for personal, psychological healing, and came
to characterize the Sarvāstivādins as a self-identified doctrinal group.
The Abhidharma Treatise in Eight Sections: The Source for New

Doctrinal and Textual Traditions
The Heart Treatises: Sarvāstivāda Pedagogical Compendia

The doctrinal elements that were presented in an often unsystematized manner in the Abhi‑
dharma Treatise in Eight Sections/Source of Knowledge were for the first time brought into
a well-structured philosophical treatise in the Tocharian *Dharmaśreṣṭhin’s Heart of the
Abhidharma (*Abhidharmahṛdayaśāstra). The work is preserved in a Chinese translation
(T.28.1550) completed in 391 CE and seems to be intended to be a widely accessible peda-
gogical digest aimed at a non-Buddhist readership. The Heart of the Abhidharma offers a
systematic account of what must be understood and practiced in order to follow the path that
leads from the condition of an ordinary human being to that of an arhat. This path is outlined
in Table 11.2.
The first chapter, “Elements,” sets out from an analysis of the world in which ordi-
nary human beings live. This chapter presents reality according to five general categories
( pañcavastuka): aggregates (skandha), sense(field)s (āyatana), elements (dhātu), thoughts
(citta), and formations dissociated from thoughts (cittaviprayukta saṃskāra). The aggre-
gates are a list of five concepts that – according to the Buddhist doctrine – constitute a
human being: matter (rūpa), feelings (vedanā), conceptualizations (saṃjñā), conditioning
factors (saṃskāra), and consciousness (vijñāna). The sense(fields)s are the faculties of sight,
hearing, smelling, tasting, touching, and thinking, and their respective object-fields (matter,
sound, smell, taste, the tangible, and ideas and concepts). The elements are the same sense
faculties and their respective object-fields, with the addition of the six types of conscious-
ness that are the result of the contact of faculties and objects. *Dharmaśreṣṭhin thus starts
his doctrinal outline with an explanation of the nature of a human being (the aggregates),
the interaction of a human being with the outside world (the sense[field]s that constitute a
human being as a psychological being [the elements]), and his life as a conscious being in the
sensual world (thoughts and formations dissociated from thoughts). *Dharmaśreṣṭhin also
connects these different categories through the concept of causes and conditions by explain-
ing that because of the dynamics of causality, everything in the mundane world is devoid of a
self-nature (svabhāva). After a description of the material world in which human beings live,
the way a human being, through her faculties, interacts with this material world is described
in the second chapter, “Conditioning Factors.” This chapter addresses all the psychic forma-
tions of a human being. Hence, these first two chapters can be regarded as corresponding
to the first of the Buddhist four noble truths, the truth of suffering. From a pedagogical per-
spective, it is necessary to understand the true nature of human life before one can become
convinced of the necessity of following the Buddhist path.
180
Table 11.2 The path of vision and the spiritual path.
Spiritual Path
Arhat
Non-Returner
(has annihilated all nine types of proclivities of the sense realm)
Once-Returner
(has annihilated eight types of proclivities of the sense realm)
Stream-Enterer
(has annihilated five types of proclivities of the sense realm)
Path of Vision
Sense Realm Realm of Form

Realm of Formlessness
(14) Knowledge regarding the doctrine (16) Subsequent knowledge regarding
of the path: path of liberation the doctrine of the path: path of liberation
(13) Forbearance regarding the doctrine (15) Subsequent forbearance regarding
of the path: immediate path the doctrine of the path: immediate path
of cessation: path of liberation the doctrine of cessation: path of liberation
of cessation: immediate path the doctrine of cessation: immediate path
of the origin: path of liberation the doctrine of the origin: path of liberation
of the origin: immediate path the doctrine of the origin: immediate path
of suffering: path of liberation the doctrine of suffering: path of liberation
of suffering: immediate path the doctrine of suffering: immediate path
Four Elements That Are Conducive to Insight
– Highest mundane factor
– Forbearance
– Summit
– Warmth
Four Applications of Mindfulness
– Application of mindfulness on entities
– Application of mindfulness on thoughts
– Application of mindfulness on feelings
– Application of mindfulness on the body
Three Preparatory Exercises
– Development of the analysis of the entities
– Mindfulness of breathing in and breathing out
– Contemplation of the repulsive
181
Bart Dessein
Knowledge of the true nature of the mundane world reveals that the cause of rebirth and suf-
fering are volitional actions. This is explained in the third chapter, “Actions,” in which volitional
actions are differentiated into unwholesome corporeal, verbal, and mental actions. This is then
followed in the fourth chapter by an analysis of the basic moods, i.e., the proclivities (anuśaya)
that are the fundamental reason why human beings perform these actions. There is a basic set of
ten proclivities: five false views, attachment, repugnance, pride, doubt, and ignorance. Actions
and proclivities correspond to the second Buddhist truth, the truth of the origin of suffering.
The analysis of actions and proclivities enables the adept to eliminate the proclivities
through progressing on the path to liberation, which is the subject of the fifth chapter, “Nobil-
ity.” Subsequently, the types of knowledge and the types of concentration a Buddhist adept
attains while progressing on the path to liberation are discussed in the sixth and seventh chap-
ters, “Knowledge” and “Concentration.” These chapters thus correspond to the third and fourth
Buddhist truths, the truths of the extinction of the origin of suffering and the path leading to
the extinction of the origin of suffering. As noted by Frauwallner (1971, 123), the fourth and
fifth chapters of the Heart of the Abhidharma should be seen as the real core of the doctrine:
“Proclivities” are the dynamic force that lead a human being to perform actions and, hence, are
the ultimate cause of rebirth, and “Nobility” is the description of the different stages one goes
through as a result of eliminating the proclivities on the path of vision.
The central importance of the fifth chapter explains why it is in this chapter that the text tran-
sitions from theoretical approach to practice. This chapter explains with more precision how
three preparatory exercises that were briefly mentioned in the first chapter lead to the convic-
tion that life is only characterized by suffering and that the Buddhist path is the only path that
can bring this suffering to a halt: contemplation of the repulsive (aśubhabhāvanā), mindfulness
on breathing in and breathing out (ānāpānasmṛti), and the development of the analysis of the
entities (dhātuvyavasthā). Contemplation of the repulsive consists of nine subsequent stages in
which the repulsiveness of the human body is contemplated. Mindfulness on breathing in and
breathing out consists of concentration on the stream of air when breathing, and on how this
stream of air manifests itself in the body. Analysis of the entities is the analysis of six elements:
earth ( pṛthivī), water (ap), fire (tejas), wind (vāyu), space (ākāśa), and consciousness (vijñāna).
Following these three preparatory exercises, there are four applications of mindful-
ness: on the body (kāyasmṛtyupasthāna), on feelings (vedanāsmṛtyupasthāna), on thoughts
(cittasmṛtyupasthāna), and on entities (dharmasmṛtyupasthāna). These four are a logical con-
tinuation of the realization that the body is subject to suffering. Contemplation on the body is
the remedy against the erroneous view that what is unclean (aśubha), i.e., the body, is seen as
clean (śubha); contemplation on feelings serves as remedy against the wrong conceptualiza-
tion (mithyāsaṃjñā) of suffering (duḥkha) as happiness (sukha); contemplation on thoughts
is a remedy for the conceptualization of what is impermanent (anitya) as permanent (nitya);
and contemplation on the entities serves as remedy against the conceptualization that what is
without self-nature (anātmaka) has a self-nature (ātman). Four elements that are conducive to
insight (nirvedhabhāgīya) follow. In the first stage, “warmth” (ūṣman), the four noble truths
are understood. Because of warmth, correct effort (samyagvyāyāma) and correct deliberation
(samyaksaṃkalpa) of the adept increase, and these two attitudes are accompanied by virtuous
actions. This brings the adept to the second mental state conducive to insight, called “summit”
(mūrdhan), in which the four noble truths are fully understood. In the third stage, “forbear-
ance” (kṣānti), the adept enjoys the four noble truths. The most extreme form of forbearance
is the condition that is a direct antecedent of the highest mundane factor (laukikāgradharma),
which is the highest quality an ordinary being can attain before she becomes an arhat. In this
stage, the adept is able to contemplate the unsatisfactoriness pertaining to the mundane realm
182
in one moment only. The fact that “the highest mundane factor” appears only at this stage
of the text again indicates that it was oriented toward a different readership than the text by
Kātyāyanīputra.
Once the adept has developed the four elements that are conducive to insight, he enters the
actual noble path (āryamārga) that has two components: the path of vision (darśanamārga)
and the spiritual path (bhāvanāmārga). The path of vision has sixteen moments: eight forms of
forbearance and eight forms of knowledge. The first moment of the path of vision is forbear-
ance regarding the doctrine of suffering (duḥkhe dharmakṣānti), which is the moment when the
adept gains insight in the doctrine of suffering. Forbearance regarding the doctrine of suffering
is called an immediate path (ānantaryamārga), because as soon as the adept has this insight,
proclivities bound to the sense realm are immediately annihilated. The second moment of the
path of vision produces a form of knowledge ( jñāna) that has the same object, the doctrine of
suffering (duḥkhe dharmajñāna). It annihilates the proclivities of the sense realm through insight
into the truth of suffering, ensuring that the adept will remain free of this type of defilement.
While the first two moments were concerned with the sense realm, the third and fourth moments
of the path of vision are, respectively, the subsequent forbearance regarding the doctrine of
suffering (duḥkhe’nvayakṣānti) and the subsequent knowledge regarding the doctrine of suffer-
ing (duḥkhe’nvayajñāna), which are directed toward the suffering of the higher two realms of
existence.
These four types of forbearance and knowledge that regard suffering ensure the complete
annihilation of that part of the proclivities that is to be annihilated through insight in the truth of
suffering. The same structure of four moments applies successively to the insights of the origin
of suffering, the extinction of the origin of suffering, and the path that leads to the extinction
of the origin of suffering. But even after the path of vision has been completed, proclivities are
not fully eliminated – a fifth method of elimination, belonging to the spiritual path, is necessary.
The sixteenth moment of the path of vision is thus considered to be the first moment of the spir-
itual path. Through spiritual practice, one also has to annihilate those proclivities that have not
been completely annihilated on the path of vision. As soon as the adept enters on the spiritual
path, he is on his way to attaining four successive fruits of the seeker (śrāmanyaphala), culmi-
nating in arhat-ship. The arhat is no longer reborn because he has stopped all karmic actions.
*Dharmaśreṣṭhin’s text became the fundamental work of the Sarvāstivāda subgroup that resided
in Bactria and Gandhāra and that formulated its interpretation of the Buddhist doctrine in a series
of pedagogical digests, i.e., texts explaining the essence or heart (hṛdaya) of the Abhidharma. As
mentioned earlier, this subgroup of the Sarvāstivāda monastic community claimed that its inter-
pretation of the doctrine was closer to the original sermons of the Buddha. *Dharmaśreṣṭhin’s text
informed later Sarvāstivāda developments in Gandhāra. These developments in turn informed
Vasubandhu’s Treasury of the Abhidharma (Abhidharmokośa). That Vasubandhu, when requested
by the masters of Kaśmῑra to write commentaries on the stanzas (kārikā) of his Treasury of the
Abhidharma, criticized the Vaibhāṣika viewpoints attests to the different doctrinal interpretations
that existed within the larger Sarvāstivāda monastic group and to the knowledge of Vaibhāṣika
viewpoints in Gandhāra. The Treasury of the Abhidharma is regarded as the final pedagogical
digest of Sarvāstivāda thought, and also as anticipating Yogācāra thought.12
The Vibhāsā Commentaries

As mentioned, with the vibhāṣā commentarial works, we touch upon the distinctive develop-
ment of Kaśmīra in contrast to Gandhāran Buddhism. This contrast can be traced to the second
council of Pāṭaliputra (contemporary Patna in northeastern India), which, according to tradition,
183
Bart Dessein
took place at the beginning of the second century after the historical Buddha’s demise. This
synod was supposed to settle a dispute between the Sthaviravādins, presided over by Tissa
Moggaliputta (Maudgaliputra), and the Sarvāstivādins, who were headed by Kātyāyanīputra.13
When the debate was judged in favor of Tissa Moggaliputta, some of the Sarvāstivādins are
reported to have moved to Kaśmῑra. They considered the Source of Knowledge to be the defini-
tive text of Sarvāstivāda philosophy and they referred to this text as the “body” (śarῑra) of
Sarvāstivādin doctrine that is sustained by six earlier (central Indian) Sarvāstivāda works to
which they referred as “feet” ( pāda). They also composed commentarial works (vibhāṣās)
on Kātyāyanῑputra’s work. These vibhāṣās gave their name to the Vaibhāṣika subgroup of
Sarvāstivādins. Those Sarvāstivādins who stayed behind emphasized *Dharmaśreṣṭhin’s Heart
of the Abhidharma, a work that gave rise to a series of works culminating in Vasubandhu’s
Treasury of the Abhidharma.
That the Sarvāstivādins of Kāśmῑra valued the Source of Knowledge over the Abhidharma
Treatise in Eight Sections can be traced back to a further council that was held in Kaśmῑra.
In his biography of Vasubandhu, Paramārtha informs us that 500 years after the Buddha’s
parinirvāṇa, Kātyāyanῑputra, who had gone forth in the school of the Sarvāstivādins, convened
a synod in Kaśmῑra, which was attended by five hundred arhats and five hundred bodhisattvas.
It is at this synod that all sayings concerning the Sarvāstivāda Abhidharma were collected
and compiled (zhuan ji sapoduobu apidamo 撰集薩婆多部阿毘達磨) and made (zhi 製) into
a work that consists of eight sections: the Source of Knowledge (T.50.2049, 189a1–26). We
may also reiterate here that Nāgārjuna, in the Instruction on the Great Perfection of Wisdom,
claimed that the text that Kātyāyanῑpitra composed in eight sections serves to “initiate knowl-
edge.” The Life of Vasubandhu also claims that the Kuṣāṇa king Kaniṣka, who appears in the
account of the synod, prohibited the vibhāṣā that was written on the Source of Knowledge
from being taken out of the country (T.50.2049, 189a26–b24). This suggests that the synod in
Kaśmῑra was meant not to settle a dispute with rival schools but rather to establish Kaśmῑra,
i.e., Vaibhāṣika, self-identity.14
There is good reason to date the extant Chinese version of the Large Commentary (on the
Abhidharma) later than Kaniṣka and, more precisely, in a period in which the philosophy of
the Mahāyāna was already well developed in Kaśmῑra. The vibhāṣā compendia were not only
commentaries on earlier Abhidharma texts, but some of the categories that figure prominently
in the Large Commentary (on the Abhidharma) – the analysis of materiality (rūpa), the con-
cept of the three vehicles (the vehicle of the disciples, the vehicle of the solitary buddhas,
and the vehicle of the bodhisattva), and the bodhisattva’s spiritual development in ten stages
(bhūmi), for example – would also figure prominently in Yogācāra Buddhism. As Robert E.
Buswell and Padmanabh S. Jaini suggest, the Large Commentary (on the Abhidharma) is
therefore also “a major source of material on early Mahāyāna developments” (Buswell and
Jaini 1996, 119).
Abbreviations
Abhidh-k-vy Sphuṭārthā Abhidharmakośavyākhyā by Yaśomitra
Chos ’byung Chos ’byung by Bu ston
T. Taishō shinshū daizōkyō大正新修大藏經
Tāranāthae Tāranāthae de Doctrinae Buddhicae in India Propagatione Narratio contextum Tibeticum
184
Notes
1 For a discussion of the development of Abhidharma texts from mātṛkās, see Dessein (2013); for a
discussion of the term “Abhidharma,” see Willemen, Dessein, and Cox (1998, 13–14).
2 A date of the second century BCE for the self-identification of the Sarvāstivādins has been suggested
by Hirakawa (1974, vol. 1, 143), and a date of the first century BCE by Shizutani (1978, 48 ff).
3 In early Buddhist texts, it is sometimes stated that practicing the items of the list of thirty-seven
“limbs of enlightenment” is the way to make an end to defilement (āsrava). See Dessein (2013, 33).
For the transition from what has been referred to as the period of “psychological Buddhism” as it was
succinctly expressed in the list of thirty-seven “limbs of enlightenment” to the period of “ontological
Buddhism,” see Bronkhorst (1985).
4 See Abhidh-k-vy 9.11–12. These six are the T.26.1536, T.26.1537, T.26.1538, T.26.1539, T.26.1540,
and T.26.1541 (also translated in T.26.1542).
5 For the Sanskrit tradition, see Abhidh-k-vy 9.12–14 and 12.4 ff.; 11.24–29. For the Tibetan tradition,
see Chos ’byung, vol. 1, 49.
6 For an in-depth study of these elements, see Dessein (2012).
7 The standard list of twelve mutually interlinked members of dependent origination appears to be a
compilation of smaller series, the two most prominent of which are a seven-membered (starting with
ignorance and leading to feeling) and a five-membered formula (starting with craving and leading to
decay-and-dying). See Wayman (1970–1971, 186).
8 Mahāvagga of the Pali Vinaya. Edition Oldenberg (1880, 13–14); translation Horner (1962, 20–21).
9 For a discussion of the four conditions and six causes, and the distinction between conditions and
causes, see Stcherbatsky (1958, vol. 2, 138).
10 Note that according to the Buddhist doctrine, only volitional actions have a karmic result.
11 The formations dissociated from thoughts are a mixed list of concepts such as types of meditative
attainment (samāpatti) and the linguistic concepts of syllables, words, and sentences.
12 There are two translations into Chinese of Vasubandhu’s Treasury of the Abhidharma: one done by
Paramārtha in 565 CE, entitled Apidamo jushe shi lun阿毘達磨俱舍釋論 (T.29.1559), and one done
by Xuanzang in 653 CE, entitled Apidamo jushe lun 阿毘達磨俱舍論 (T.29.1558). The version by
Xuanzang was translated into French by Louis de La Vallée Poussin ([1923–1931] 1971) and into
English by Lodrö Sangpo (2012). The French translation by Louis de La Vallée Poussin was trans-
lated into English by Leo Pruden (1991). For more on Vasubandhu, see Chapter 16 in this volume.
13 See Bareau (1955, 168–69).
14 For debates on the purpose of this synod, see Roerich (1949, 25); Obermiller (1931–1932, 96–97);
Tāranāthae 47, 7 ff; Schiefner (1869, 58–61); Frauwallner (1952, 252); Bareau (1955, 132).
Bibliography
Primary Sources
Chos ’byung. Obermiller, Evgeniy Evgenevitsj, trans. 1931–1932. History of Buddhism by Bu-Ston, Part
1: The Jewelry of Scripture; Part 2: The History of Buddhism in India and Tibet. Materialen zur
Kunde des Buddhismus 18–19. Heidelberg: In Kommission bei O. Harrassowitz, Leipzig.
Sphuṭārthā Abhidharmakośavyākhyā. Wogihara, Unrai, ed. 1932. Sphuṭārthā Abhidharmakośavyākhyā:
The Work of Yaśomitra. Tokyo: The Publishing Association of the Abhidharmakośavyākhyā.
Taishō shinshū daizōkyō大正新脩大藏經. Takakusu, Junjirō 高楠順次朗 and Watanabe, Kaigyoku 渡邊
海旭, ed. 1924–1934. Taishō shinshū daizōkyō 大正新脩大藏經. Tokyo: Taishō issaikyō kankōkai.
———. T.23.1435: Daśādhyāyavinaya, Shi song lü 十誦律, Puṇyatara, Dharmaruci, Kumārajīva.
———. T.26.1536: Mahākauṣṭhila (Skt.), Śāriputra (Ch.), [Abhidharma]saṃgῑtiparyāya[pādaśāstra],
Apidamo jiyimen zulun 阿毘達磨集異門足論, Xuanzang玄奘.
———. T.26.1537: Śāriputra (Skt.), Mahāmaudgalyāyana (Ch.), [Abhidharma]dharmaskandha[pāda
śāstra], Apidamo fayun zulun 阿毘達磨法蘊足論, Xuanzang玄奘.
———. T.26.1538: Maudgalyāyana (Skt.), Mahākātyāyana (Ch.), Prajñaptiśāstra, Shishe lun 施設論,
Dharmapāla, Weijing 惟淨.
185
Bart Dessein
———. T.26.1539: Devaśarman, [Abhidharma]vijñānakāya[pādaśāstra], Apidamo shishen zulun阿毘

達磨識身足論, Xuanzang玄奘.
———. T.26.1540: Pūrṇa (Skt.), Vasumitra (Ch.), [Abhidharma]dhātukāya[pādaśāstra], Apidamo
jieshen zulun阿毘達磨界身足論, Xuanzang玄奘.
———. T.26.1541: Vasumitra, [Abhidharma]prakaranapāda[śāstra], Zhongshifen apitan lun 眾事分阿
毘曇論, Guṇabhadra and Bodhiyaśas
———. T.26.1542: Vasumitra, [Abhidharma]prakaranapāda[śāstra], Apidamo pinlei zulun阿毘達磨品
類足論, Xuanzang玄奘.
———. T.26.1543: Kātyāyanῑputra, *Abhidharmāṣṭagrantha[śāstra]/*Abhidharmāṣṭaskandha[śāstra],
Apitan ba jiandu lun 阿毘曇八犍度論, Saṃghadeva and Zhu Fonian 竺佛念.
———. T.26.1544: Kātyāyanῑputra, Jñānaprasthāna, Apidamo fazhi lun 阿毘達磨發智論, Xuanzang
玄奘.
———. T.27.1545: 500 arhats, *[Abhidharma]mahāvibhāṣā[śāstra], Apidamo da piposha lun 阿毘達磨
大毘婆沙論, Xuanzang 玄奘.
———. T.28.1550: Dharmaśreṣṭhin, *Abhidharmahṛdaya[śāstra], Apitan xin lun 阿毘曇心論,
Saṃghadeva, Huiyuan 慧遠.
———. T.29.1558: Vasubandhu, Abhidharmakośaśāstra, Apidamo jushe lun 阿毘達磨俱舍論. Xuan-
zang 玄奘.
———. T.29.1559: Vasubandhu, Abhidharmakośaśāstra, Apidamo jushe shi lun阿毘達磨俱舍釋論.
Paramārtha.
______. T.50.2049: Paramārtha, Posoupandou fashi zhuan 婆藪槃豆法師傳.
———. T.70.2300: Paramārtha, San lun xuanyi jian you ji 三論玄義檢幽集.
Tāranāthae de Doctrinae Buddhicae in India Propagatione Narratio contextum Tibeticum. Schiefner,
Anton, ed. 1868. Petropoli: E. Codicibus Petropolitanis, Eggers.
Secondary Sources
Armelin, Indumati. 1978. Le Coeur de la Loi Suprême. Traité de Fa-cheng. Abhidharmahṛdayaśāstra de
Dharmaśrῑ. Paris: Librairie Orientaliste Paul Geuthner.
Bareau, André. 1955. Les sectes bouddhiques du petit véhicule. Paris: École française d’Extrême-orient.
Bechert, Heinz. 1973. “Notes on the Formation of Buddhist Sects and the Origins of Mahāyāna.” In
German Scholars on India. Contributions to Indian Studies, edited by Cultural Department of the
Embassy of the Federal Republic of Germany, 6–18. Varanasi: Chowkhamba Sanskrit Series Office.
Bronkhorst, Johannes. 1985. “Dharma and Abhidharma.” Bulletin of the School of Oriental and African
Studies 47 (2): 305–20.
Buswell, Robert E., and Padmanabh S. Jaini. 1996. “The Development of Abhidharma Philosophy.” In
Encyclopedia of Indian Philosophies. Volume VII. Abhidharma Buddhism to 150 A.D., edited by Karl
H. Potter, 73–119. New Delhi: Motilal Banarsidass.
Demiéville, Paul. 1961. “Un fragment Sanskrit de l’Abhidharma des Sarvāstivādin.” Journal Asiatique
249: 461–65.
Dessein, Bart. 2012. “ ‘Thus Have I Heard’ and Other Claims to Authenticity: Development of Rhetorical
Devices in the Sarvāstivāda Ṣaṭpādābhidharma Texts.” In Zen Buddhist Rhetoric in China, Korea,
and Japan, edited by Christoph Anderl, 121–62. Leiden and Boston: Brill.
———. 2013. “Lists as Dynamic Devices in Early Buddhist Doctrine and Textual Tradition.” Antiquo-
rum Philosophia 7: 29–48.
Frauwallner, Erich. 1952. “Die buddhistischen Konzile.” Zeitschrift der Deutschen Morgenländischen
Gesellschaft 102: 240–61.
———. 1971. Die Entstehung der buddhistischen Systeme. Göttingen: Vandenhoeck & Ruprecht.
Fukuhara, Ryōgen福原亮厳. 1965. Ubu Abidatsumaronjo no hattatsu 有部阿毘達磨論書の発達.
Kyoto: Nagata Bunshōdō.
Hirakawa, Akira 平川彰. 1974. Indo Bukkō shi インド佛教史, 2 vols. Tokyo: Shunjūsha.
Horner, Isaline Blew. 1962. The Book of the Discipline (Vinaya-Pitaka). Volume IV. Mahāvagga. London:
Luzac.
La Vallée Poussin, Louis de. (1923–1931) 1971. L’Abhidharmakośa de Vasubandhu, 6 vols. Bruxelles:
Institut belge des Hautes Études chinoises.
186
Lodrö Sangpo, Gelong. 2012. Abhidharmakośa-Bhāṣya of Vasubandhu. The Treasury of the Abhidharma
and Its Commentary, 4 vols. New Delhi: Motilal Banarsidass.
Nishi, Giyū 西義雄. 1934. “Ubushūnai ni okeru hotchi hi-hotchi kei nado no shoshu no gakusetsu
oyobi gakutō no kenkyū 有部宗內に於ける發智系‧非常發智系等の諸種の學說及學統の研究.”
Shūkyō kenkyū 宗教研究11 (4): 564–79; 11 (5): 768–89.
Oldenberg, Hermann. 1880. The Vinaya Piṭakaṃ: One of the principal Buddhist holy scriptures in the
Pāli language. Vol. II: The Cullavagga. Pali Text Society. London: Williams & Norgate.
Pruden, Leo. 1991. Abhidharmakośa of Ācārya Vasubandhu, 4 vols. Berkeley: Asian Humanities Press.
Roerich, G. N. 1949. The Blue Annals, 2 vols. Royal Society of Bengal, Monograph Series, vol. VII.
Calcutta: The Royal Asiatic Society of Bengal.
Schiefner, Anton, trans. 1869. Tāranātha’s Geschichte des Buddhismus in Indien – aus dem Tibetischen
übersetzt. Eggers, St. Petersburg: Kaiserliche Akademie der Wissenschaften.
Shizutani, Masao 静谷正雄. 1978. Shōjō Bukkyōshi no kenkyū 小乘仏教史の研究. Kyoto: Hyakkaen.
Stcherbatsky, Theodore. 1958. Buddhist Logic, 2 vols. The Hague: Mouton.
Takakusu, Junjirō. 1904–1905. “On the Abhidharma Literature of the Sarvāstivādins.” Journal of the Pali
Text Society: 67–146.
Wayman, Alex. 1970–1971. “Buddhist Dependent Origination.” History of Religions 10: 185–203.
Willemen, Charles, Bart Dessein, and Collett Cox. 1998. Sarvāstivāda Buddhist Scholasticism. Hand-
book of Oriental Studies 2/11. Leiden, New York and Köln: Brill.
187
12
THE VĀTSĪPUTRĪYA/
SĀM. MITĪYA
Buddhist Personalism as a Mainstream School
of Thought
Peter Skilling
Introduction1
The self is life’s perpetual puzzle, and the nature of the self is one of the fundamental ques-
tions of philosophy. Its scope ranges from the psychological, intellectual, and social makeup
of the individual – the human being – and his or her relations to other human beings and to the
world, to metaphysical reflections on origins and destiny of the soul and its relation to divinity
and time. In many systems of thought, these ruminations respond to and address core ethical
questions of agency and moral responsibility. “Longue durée” systems in which the self rein-
carnates over long and even astronomical periods have to explain how this happens. In Bud-
dhist moral philosophy, it is axiomatic that human deeds or acts entail reactions: they are like
seeds that in the course of time will bear fruit for the doer. The rich narrative literature of the
Avadānas (tales of the past-life exploits of people connected with Buddhism) that developed to
illustrate the workings of karma frequently calls up a verse that makes this point:
Human deeds do not just disappear

Even after hundreds of millions of eons.
When their time comes and conditions ripen
Deeds bear fruit for breathing beings.
Most of the well-known religious systems of India (Hinduism, Jainism, Buddhism) are such
systems. Buddhism is unique in that the landmark teaching of Gautama Buddha is “non-self.”
Writing perhaps in the second century CE, a poet praised the Buddha:
When the mind is filled with egotism, the succession of rebirths does not cease.
Egotism does not leave the heart when the view of self is present.
No other teacher in the world proclaims non-self:
Therefore, outside of your system of thought there is no other pathway to peace.
(Wogihara 1932–1936, 697)
Gautama Buddha did not posit the existence of an enduring self, but – on the contrary – decon-
structed the individual into components and constituent parts. These he formulated in several
188 DOI: 10.4324/9781351030908-19

The Vātsīputrīya/Sām.mitīya
ways. An early summary was the five aggregates, the twelve sense doors, and the eighteen ele-
ments (skandha-āyatana-dhātu), a framework that developed into what modern thinkers have
described as a “dharma theory” – an analysis of experience in terms of complex interactions
between clusters of material and mental conditions (dharma). For example, Buddhaghosa’s
commentary on the Enumeration of Phenomena (Dhammasaṅgaṇi) of the Pali Abhidhamma
explains how certain types of mentality arise:
When the first of the great skillful mentalities (cittas) of the sense-sphere arises, at
that time the fifty and more constituents (dhammas) that have arisen by way of fac-
tors of mentality (citta) are indeed ultimate constituents in the sense of self-existents
(sabhāva). There is nothing else, whether a being, or an entity, or a man or a person.
(Ronkin 2005, 117)
The question that Buddhist philosophy had to face is: how does moral responsibility work
when there is no soul?
In the Sūtra on the Ultimate Meaning of Emptiness (Paramārthaśunyatāsūtra), the Buddha said:
It is like this, O monks: moral action (karma) exists and its fruition ( phala, vipāka)
exists, but there is no doer or agent (kāraka) who casts aside the aggregates of this
present life and takes up the aggregates of the next – except as a conventional desig-
nation (dharmasaṃketa).
There was, however, one group of schools within the traditionally enumerated eighteen early
monastic orders that asserted the existence of a person, an individual ( pudgala) who was
neither the same as nor different from the constituent elements of being. To maintain the exist-
ence of the pudgala seems, from the contemporary understanding of Buddhism, to be unprec-
edented and unique. But this notion had a long evolution, starting from the time of the Buddha
himself. It is not possible, however, to trace the evolution in any chronological detail.
The group of schools that asserted the existence of the person was the Vātsīputrīya/Sāṃmitīya
complex of schools. In this chapter, I attempt to delineate the contours of the Vātsīputrīya/Sāṃmitīya
presence in India. The topic raises questions about the place of the Vātsīputrīya/Sāṃmitīyas in
Buddhist history and in Indian and world philosophical thought – how this has been assessed and
how to assess it. This entails examining the history of ideas in the aggregation of Buddhist sanghas,
or monastic orders, both synchronically and diachronically – a task that is far beyond the scope of
this contribution. Put simply, we need to avoid taking one or the other Buddhist school as original,
orthodox, or mainstream, and instead regard the development of Buddhist thought as an organic
process of exchange between the sanghas within the broader dynamics of Indian philosophy. The
“eighteen Buddhist schools” did not develop unilaterally as discrete and self-contained units. Clas-
sical texts preserve several competing narratives of how the schools emerged. The problems raised
in this contribution concern the overall development of Buddhism, both within and outside of the
framework of the traditional monastic orders, schools of thought, and vehicles.
Buddhist Debates About Self and Person

Many canonical passages discuss aspects of non-self and state that the view of a self is a wrong
or mistaken view.2 Centuries of discussion and debate on the nature and status of the “person”
are intricately intertextual: the different sides cite the Buddha’s words to support their theses.
It is a debate among different interpretations of the import of the Buddha’s speech.
189
Peter Skilling
In the Sūtra on the nun Śailā’s encounter with Māra, the latter approaches the arhat nun
Śailā and quizzes her about the origin of the “being” (satva). She retorts:
Why do you think there’s a sentient being?

Māra, this is only your mistaken idea.
Empty is this mass of formations:
No being is there here.
Just as when the parts are assembled
We recognize what we call a “chariot”:
Like this, dependent on the aggregates,
One speaks by convention of a “being.”
(La Vallée Poussin 1912, 257–58)
Śailā’s riposte was often cited in polemics about the “being”; for instance, in the Topics of
Debate (Kathāvatthu, Taylor 1894, 66) and Questions of King Milinda (Milindapañha, Trenck-
ner 1880, 27), and in the works of Kumārajīva (Lamotte 1944, 28), Vasubandhu, Candrakīrti,
and other philosophers.
We see here a shift from the ātmā to the being, which represents a major reformulation
of the philosophical stakes from the critique of the self to a broader spectrum of modes of
individuality. It appears that the early discourses accepted and used the notions of the conven-
tional individual, the conscious being or person (satva, pudgala), without much reservation.
The Vātsīputrīyas took what might seem to be a natural step and recognized the “person” as a
valid constituent of human existence and adopted it into their dharma system. Other schools,
including the Theravaṃsa (the lineage of schools claiming to descend from the early theras,
the direct followers of Śākyamuni), saw this as a concession to the idea of an enduring ātmā
– soul, self, or substantial ego – of mainstream Indian thought and emphasized in their writings
that the being or person does not exist and cannot be epistemologically proven. Belief in an
enduring soul was considered a serious obstacle to liberation. The shift is evident in the Topics
of Debate. The Sarvāstivāda school went further and introduced a long list of synonyms of
the being and the person into their Sūtra and Abhidharma literature, denying that any of these
iterations of individuality existed. This line of thought strongly influenced the Perfection of
Wisdom and Great Vehicle literatures, where similar lists occur regularly.
The debate seems in part to have been fueled by the ambiguities of language. The word
“person” occurs throughout the Buddhist scriptures, and one Sūtra in particular is known as
“discourse on the bearer of the burden” (Bhārāhārasūtra). The Buddha teaches the monks that
the burden is the five aggregates and the bearer of the burden is the person. But, if the person
who bears the burden is different from the burden, then isn’t the person something more than
just a name for the five aggregates? There were many attempts to explain the anomaly. For
some, the Sūtra proved that the pudgala is an autonomous reality, but this gave rise to further
questions. Their answer was that the person was neither the same as the aggregates nor differ-
ent from the aggregates: it did not fit the usual abhidharmic categories. It was inexpressible
(avācya), an “indeterminate self.”
Vātsīputrīya, Sām
· mitīya, Pudgalavāda: Quandaries of Nomenclature
Modern academic writing most commonly calls the school pudgalavāda – “those who assert the
factual existence of the person or individual” – in a word: “personalists.” But the Vātsīputrīya/
190
Sāṃmitīya family of schools itself did not use this term. Pudgalavāda is not a self-descriptor but
rather a label coined by other schools to indicate what they saw as the school’s distinguishing,
and most pernicious, tenet: the pudgala. Pudgalavāda has been the term of choice in Buddhist
studies for over a century (Lusthaus 2009, 275),3 but I generally prefer to avoid the label –
not only because it is polemical and pejorative, but even more because it obfuscates rather
than clarifies the historical record. The label exaggerates, from the perspective of philosophical
adversaries, the role of the pudgala. I use instead the somewhat awkward term “Vātsīputrīya/
Sāṃmitīya” or, for short and for the later period, simply “Sāṃmitīya.” I use the term to include
the affiliated Bhadrayānīya, Dharmottarīya, Ṣaṇṇagarika, and Kaurukulla-Sāṃmitīya schools.
Several modern scholars stand out in the study of the pudgalavāda in the English language:
Theodor Stcherbatsky (1866–1942), Louis de La Vallée Poussin (1869–1938), André Bareau
(1921–1993), Thích Thiện Châu (1931–1998), and Leonard Priestley.4 We might say that spe-
cific attention to the notion of the pudgala starts with Stcherbatsky’s provocatively titled “The
Soul Theory of the Buddhists,” written in English and published in 1919 (Stcherbatsky 1919a,
823–54; 1919b, 937–58): that is, modern pudgalavāda studies are now a hundred years old.
In 1920, La Vallée Poussin published an entry on the “Sammitiyas” in the Encyclopaedia of
Religion and Ethics (La Vallée Poussin 1920, 168–69), but he also published many articles
and books relevant to the theme before and after this (La Vallée Poussin 1925a, 34–38). Bar-
eau wrote extensively on Buddhist sects (Bareau 1954, 1955a, 1956). By the end of the last
century, Vietnamese monk Thích Thiện Châu ([1996] 1999) and Leonard Priestley (1999) pro-
duced balanced studies of the concept of the pudgala, demonstrating that the pudgalavādins
had something to say and that this was worth listening to. New sources for the study of the
Vātsīputrīya/Sāṃmitīyas were identified, and in some cases were edited and translated.
Vātsīputrīya/Sām
· mitīya and the Doctrine of the Person
What is the relation between the “person” or “individual” and the “soul” or the “self ” –
between pudgala and the ātman? Are they the same or different? Both are everyday terms of
speech and reference and both have metaphysical usages. In Jainism, pudgala, material and
non-sentient, is one of six fundamental substances. Ātman, in grammar a noun as well as an
emphatic and reflexive pronoun, carries significant metaphysical baggage in Hindu schools of
thought. The Vātsīputrīya’s term of choice is pudgala, not ātman, but the criticisms of other
schools react (or perhaps overreact) to it as a kind of ātman. This is why some authors singled
them out as “Buddhist tīrthikas.”5
When we treat the “person” as the signature tenet of a belief system or vāda, we need to
consider the relationship between received doctrine and quotidian monastic life. The belief
in or assertion of the pudgala – the person or the individual – looms large in the doxographi-
cal literature compiled by Buddhist authors from other, non-Vātsīputrīya/Sāṃmitīya schools
or traditions, but the word pudgala is noticeably absent in most of the available texts of the
school itself. Only two metaphysical treatises of the Sāṃmitīya school survive in their original
Indian language – The Great Discourse on the Evolution and the Devolution of the Universe
(Mahāsaṃvartanīkathā, Okano 1998) and the Stanzas of the Compendium of Abhidharma
(Abhidharmasamuccayakārikā). The two works survive in more or less complete form, but
neither of them discusses or even mentions the doctrine of the pudgala.
This leads to the question: how did the Buddhist orders manage their identities? The cut and
style of monastic costume maintained a physical, visual identity for the different orders, as it
does today for the subdivisions within the Theravādin, Tibetan, and East Asian sanghas. Robes
191
Peter Skilling
announce hierarchical and gender identities. But barba non facit philosophum, l’habit ne fait
pas le moine: we are not concerned here with matters of hairstyle and dress, but of ideas and
philosophy. Was every Vātsīputrīya/Sāṃmitīya monastic expected to study and adhere to the
idea of the pudgala? Did the pudgala mean anything to rank and file monastics? Did the pud-
gala insinuate itself into the monastic codes of the Vinaya, the teaching of the Buddha in the
Sūtras, or the daily liturgies and spiritual exercises? What, if anything, did the pudgala mean to
lay followers and supporters of the school? In the absence of any Vātsīputrīya/Sāṃmitīya can-
ons, these questions remain open, but on the evidence of currently available texts, I suspect that
the answer is that no, the pudgala does not seem to feature in the canonical texts of the school.
Methodological Considerations in the Study

of the Vātsīputrīya/Sām
 mitīya
The existing paradigms for the study of school development are conditioned by the resources
available today and the choices that we make as researchers (further dependent on our lan-
guage skills or lack thereof). Since records of most other schools are no longer extant, the
resource base is largely limited to the textual traditions of two schools, the Theravāda and the
Sarvāstivāda/Vaibhāṣika. Most modern research has sided with these two schools and pre-
sumed that pudgalavāda is an aberration – that it is unorthodox and even non-Buddhist. The
South Asian pudgalavāda group of monastic orders died out nearly five centuries ago and
today has no voice whatsoever in the forums of philosophy (despite the fact that the debate
on the pudgala continues to be part of Tibetan Buddhist curricula). Today, whether in Tibetan
monasteries or in modern Buddhism and Buddhist studies, there is no one to speak on behalf of
the Vātsīputrīya/Sāṃmitīyas or to argue their case. Nonetheless, we ought not to ignore them.
Buddhism is not a monolith, and the Vātsīputrīya/Sāṃmitīyas are one of the several cogent
streams of thought that constitute the complex of ideas that we recognize as “Buddhism.” Like
other sanghas, the Vātsīputrīya/Sāṃmitīyas trace their lineage directly back to Śākyamuni
and the first two recitation/redactions of his teachings. Like other sanghas, the Vātsīputrīya/
Sāṃmitīyas were fully fledged corporate bodies which interacted with state and society, and
responded to the continual flux of philosophical thought around them. Like other sanghas,
the Vātsīputrīya/Sāṃmitīyas were historical and cultural currents of Indian Buddhism with
distinct identities. The Sāṃmitīya self-consciousness is nurtured by a distinctive account of
the how their scriptures were redacted (see in what follows). They claim a distinct pedigree, a
distinct history, through five recitation convocations.
Balanced studies of the historical and philosophical evolution of the Vātsīputrīya/Sāṃmitīyas
have been impeded by methodological issues. One is the persistence of Victorian frameworks
that were developed at the inception of Buddhist studies. Early European scholars interpreted
Indian religions, including Buddhism, on their own terms, importing inappropriate models
from contemporary historiographies of the Christian churches and the fledgling discipline of
“world religion.” They saw the evolution of Buddhist schools in terms of “orthodoxy” and
“heresy,” as a history of schism, dissension, and departure from the original pure and untram-
meled teaching of Gautama Buddha. They found this untrammeled teaching in the Theravāda
that they encountered in Sri Lanka – or at least in its Pali scriptures, since they did not take long
to decide that the latter-day understanding and practice of the religion was degraded and cor-
rupt. The Theravāda was among the Buddhist schools that rejected the Sāṃmitīya theory, and
modern Buddhist historiography generally echoes their polemics. The very idea of the person
was seen as wrong – and there were no pudgalavādins around to argue for it.
192
Geography and Localization of the Vātsīputrīya/Sām

· mitīya
From the very start, Buddhist monastics were mobile. Perennially peripatetic, they carried
ideas, practices, ceremonies, and rituals with them wherever they went. Theories and doctrines
were nurtured in shifting landscapes of exchange and debate. There is no systematic history
or geography of the Vātsīputrīya/Sāṃmitīyas – or, indeed, of any other school. We can only
attempt to sketch the geography and demography of the Buddhist Vinaya schools by consult-
ing Chinese sources like the fabled accounts of Xuanzang and Yijing, which enable us to imag-
ine the geographical distribution of the schools at the time of their visits: that is, in the seventh
century. Later sources in Tibetan are piecemeal but often helpful. Indic literary references are
few and far between, but script and inscriptions help to locate the early Buddhist schools.
Xuanzang’s reports show that the Sāṃmitīyas were well represented at the major sites asso-
ciated with the Buddha’s life – Kapilavastu, Varanasi, Sarnath, Sravasti, and Vaisali. At the
great monasteries of Nālandā and Vikramaśila, the Sāṃmitīyas were represented side by side
with the four main schools of the age. They were an active presence throughout the Ganga-
Yamuna basin from Mathura down to Bengal, and Xuanzang reports that the Vātsīputrīya/
Sāṃmitīyas were widespread in the middle and lower Indus valley (present-day Pakistan) and
in the coastal regions of Pakistan and western India.
Yijiing writes that “in Lāṭa and Sindhu . . . the Sāṃmitīya has a large number of adherents”
(Li Rongxi 2000, 11). This idea is perpetuated by the rubric “Saindhava Śrāvaka” encountered
in Tibetan sources like Tāranātha’s history describing adherents of the school in north India as
“Sindhi,” “residents of Sindh.” It appears that the Vātsīputrīya/Sāṃmitīyas were the predomi-
nant school in Sindh up to the Arab invasions of the eighth century. Little remains of early
Buddhist or pre-Muslim antiquities of the region, but what does remain suggests that there
were numerous stūpas and elegant decorative styles in brick (Cousens [1929] 1975; van Lohu-
izen 1981, 43–57; Pal 2008). In Gujarat, there are a number of rock-cut monasteries. They are
much abraded and lack any inscriptions. Which sanghas inhabited them? We do not know, but
from the blanket statements about the Buddhism of the region in the Chinese travel accounts,
they could well have been Sāṃmitīyas. At an early period, the Vātsīputrīya/Sāṃmitīyas were
present along the western coast in Konkan and the adjacent western Ghats, as well as inland
in the Deccan plateau, where at Nasik they enjoyed royal patronage of prestigious projects
like monumental caitya halls. Xuanzang associates many of the Sāṃmitīya centers with the
visits of Śākyamuni or previous Buddhas and with the celebrated emperor, Aśoka (note, how-
ever, that during his journeys across the subcontinent, the redoubtable Chinese pilgrim-scholar
reported on many sites that claimed such illustrious associations).
Buddhist monuments once dotted the north Indian landscape. Across north India, from
Sindh in the west to Bengal in the east, they were interactive nodes in an extensive and extraor-
dinary network of knowledge. By the time the British arrived and began – haphazardly, at least
at first – their colonial research in archaeology and history, nothing was left of thousands of
monasteries and stūpa sites beyond their foundations, and as economic activities expanded,
even these were looted for bricks to build railway lines and new buildings – the nineteenth cen-
tury was an age of both exciting archaeological discoveries and of thoughtless despoliation.
Some, if not many, of these ruins would have once been Vātsīputrīya/Sāṃmitīya institutions,
given that Xuanzang’s records mention at least 1,351 monasteries of the school with more than
66,500 monks (he rarely distinguishes between monasteries and nunneries) (Lamotte 1958,
tables, 596–601). His figures are approximations – but even so, in terms of monastic land
holdings and population, the Sāṃmitīya were by far the largest school: more than half of the
monks of India belonged to the Sāṃmitīya order. The monasteries would have been centers
193
Peter Skilling
of training and education where monks and nuns practiced the Vinaya and studied the scrip-
tures and the Vātsīputrīya/Sāṃmitīya dharma system. The only structures associated with the
school to survive are early rock-cut monuments in western India.
Vātsīputrīya/Sām
 mitīya Scriptures and Canons
The historical geography of the Vātsīputrīya/Sāṃmitīyas – as far as we can reconstruct it –
shows a widespread and dynamic monastic order that was well established in the lower
Indus and Ganga-Yamuna valleys. Like other schools, the Vātsīputrīya/Sāṃmitīyas must
have had rich textual collections centered on Tripiṭakas containing their redactions of the
Vinaya, Sūtras, and Abhidharma. The school’s scriptural language was a form of Middle
Indo-Aryan, specified by Tibetan sources as Apabhraṃśa (Thích Thiện Châu [1996] 1999,
31–32). It was written in at least the Nagari and the Bhaikṣukī scripts; the latter, apparently
unique to the school, is also called Saindhavī (Dimitrov 2020). But the textual products of
Vātsīputrīya/Sāṃmitīya thought do not survive. The seventh-century Chinese scholar Yijing,
who had first-hand experience in India, describes the Sāṃmitīya Tripiṭaka as 300,000 stan-
zas in length. Xuanzang carried fifteen texts and treatises of the school back to China, but
the manuscripts do not survive (Li Rongxi 1996, 394). Many of the short passages cited
in extant Sāṃmitīya works like the Treatise of the Sāṃmitiya School (*Sāṃmitīya-nikāya-
śāstra, Sānmídibù lùn, T. 1649) have parallels in the canons of other schools like the Pali
canon. While a few fragments of their canonical literature survive, nothing is left of their
ritual, liturgical, and mythological literature. The result is that the Vātsīputrīya/Sāṃmitīyas
are known to us through the writings of outsiders to the tradition and its textual sources.
This includes great masters like Vasubandhu and Bhāviveka, who at least had some direct
access to the textual tradition. We modern scholars do not even have access to the school’s
texts and write using the reports of opponent schools. We may claim the benefits of historical
consciousness, but we have access only to fragmentary records. Reasoned debate or criticism
presupposes an understanding of the intellectual milieux of the historical audiences of the
Vātsīputrīya/Sāṃmitīyas, but, in the absence of texts or social or literary records, such an
understanding is lacking – even impossible – at present. This is hardly satisfactory.
Some details of the Vinaya rules for monks and nuns are given in Buddhatrāta’s Exposition
of Twenty-two Aspects of the Monastic Rules (*Vinaya-dvāviṃśativyakti, Lü èrshiêr míngliăo
lùn, translation by Paramārtha, T. 1461). We do not even know what the canonical Abhidharma
of the Vātsīputrīya/Sāṃmitīyas looked like – how many books it had, their titles, whether any
of them had introductions to tell whether they were spoken by the Buddha, by disciples, or by
others. The Vātsīputrīya/Sāṃmitīyas certainly had a fully developed dharma theory of their
own. This is further evidence that their pudgala did not contradict or replace the Buddhist
deconstruction of the self; whatever the “indeterminate person” was, it depended on a complex
and sophisticated theory of interactive dharmas. We see this in the Stanzas of the Compendium
of Abhidharma (Abhidharmasamuccayakārikās) and in Daśabalaśrīmitra’s long excerpts from
scholastic texts; if not from their Abhidharma, then from a manual or manuals.
The Abhidharma Treatise, Description of the Universe (*Lokaprajñapty-abhidharma-
śāstra, Li shi a pi tan lun 立世阿毘曇論, T. 1644), a cosmological work translated into Chi-
nese in 559 by Paramārtha, has been identified as belonging to the Sāṃmitīyas (Mahāsaṃvar
tanīkathā, Okano 1998, 22 foll.) This text is one of the sources for the Description of the Uni-
verse (Lokapaññatti, Denis 1977), a cosmological compendium transmitted in Pali in South-
east Asia. The Sāṃmitīya affiliation is contested by others.
194
Recitation Convocations (Saṃgīti) of the Vātsīputrīya/Sām

· mitīyas
Accounts of the Buddhist convocations or recitation/redactions consolidate the lineage
and identity of the different schools. They establish a sense of continuity – a direct link to
Śākyamuni and the great arhats. Not many accounts survive, and modern scholarship and
piety have largely taken the Theravādin account as foundational for all schools.6 Remark-
ably, two roughly contemporary accounts of the recitation convocations of the Vātsīputrīya/
Sāṃmitīyas are preserved. One is “from the Āgama of the Ārya-Sāṃmitīya-school” (Okano
1998, 397, § 1) in Tibetan translation in Daśabalaśrīmitra’s Examination of the Compounded
and Uncompounded; the other is in Sanskrit verse in Sarvarakṣita’s Great Discourse on Evolu-
tion and Devolution (Okano 1998, § 4.2.15). The account runs as follows:
Long after the final nirvāṇa of the great sage Kāśyapa, in the age of strife, the last age,
our teacher, the Tathāgata Śākyamuni arose in the world . . . the Tathāgata, the lamp,
after establishing the continuity of the lamp of the holy Dharma, himself passed away.
In the second month after the Tathāgata’s final nirvāṇa, . . . five hundred monks free
of desire recited and compiled the teaching of the Buddha in the Saptaparṇa cave.
And again, one hundred years after the Tathāgata’s final nirvāṇa, seven hundred
monks free of desire recited and compiled the Dharma.
And again, four hundred years after the Tathāgata’s final nirvāṇa, when the Ascet-
ic’s community had become divided into different groups, each adhering to its own
school, Vātsīputra recited and compiled the Dharma of one of these schools. From
that time on, that school has been known as the Vātsīputrīyas, the expounders of the
Dharma.
And again, seven hundred years after the Tathāgata’s final nirvāṇa, a senior monk,
the sage *Sammita, recited and compiled the traditions of that school. From that time
on, that school has been known as the Sāṃmitīya school.
And again, eight hundred years after the Tathāgata’s final nirvāṇa, the senior
monks Bhūtika and Buddhamitra recited and compiled the traditions of that school.
These are known as “the five recitation convocations of the Dharma of the
Sāṃmitīya school.”
The Sanskrit Great Discourse and Daśabala’s Tibetan excerpt name four prominent teachers
(we might call them the patriarchs) of the school: Vātsīputra, *Saṃmita, and, as a pair, Buddhi
and Buddhamitra. We do not know anything about their dates or their lives and intellectual
careers, only their names and their association with the third, fourth, and fifth Tripiṭaka recita-
tion/redactions. Their textual legacies do not survive – not a single extant work. The result is
that we cannot trace their roles in the philosophical development of pudgalavāda.
Sām
 mitīya Authors
Not many authored works of the Sāṃmitīyas are preserved. Vasubhadra composed the Sutra of
the Three Categories of Phenomena (*Tridharmakhaṇḍaka-sūtra), which is lost in the original
Indic version but preserved in two different Chinese translations: the Sì ēhánmù chāojiĕ (T.
1505) and the Sānfădù lùn (T. 1506). Both translations include a commentary by Saṅghasena,
the Book on the Three Categories of Phenomena (*Tridharmakhaṇḍaprakaraṇa).
The author and translator of the Treatise of the Sāmmitīya School, or, from its colophon,
Treatise on the Description of the Basis (*Āśrayaprajñaptiśāstra), are not known. It deals
195
Peter Skilling
with controversial philosophical topics, including the pudgala and the intermediate state, and
quotes many Sūtras (Thích Thiện Châu [1996] 1999, 99–117).
A short, independent commentary on the ye dharmā stanza entitled Commentary on the
stanza of dependent arising (Pratītyasamutpādagāthā-vyākhyāna) is preserved in Tibetan trans-
lation. According to the colophon, it was written by Ācārya Vinayavarman of the Sāṃmitīya
school. It awaits study and translation to see whether it throws light on any specific doctrines
of the school (La Vallée Poussin 1962, 50, § 127.2; Yamaguchi et al. [1977] 1986, 51–52).7
Indic versions of the works of only two authors, Saṃghatrāta and Sarvarakṣita, are pre-
served. We know nothing about Saṃghatrāta, author of the Stanzas of the Compendium
of Abhidharma. The author who stands out is the poet Sarvarakṣita; the colophon of the
Mahāsaṃvartanī-kathā describes him as “the master, the excellent venerable one Sarvarakṣita,
the great poet.” He is also the author of a Sanskrit grammar and the Prakrit Birth Story of
Prince Maṇicūdā (Maṇicūḍā-jātaka), which is composed in verse (Hanisch 2006, 109–61;
2008, 195–342) – a glance at only a few verses immediately confirms that he is a “great poet.”
Another Indian author is known only from his works and translations in Chinese. This is
Paramārtha (499–569 CE), who is said to have been ordained as a Sāṃmitīya and to have stud-
ied at Valabhī, in present-day Gujarat, which in Xuanzang’s time was a flourishing Sāṃmitīya
center (Okano 1998, 58–59). Paramārtha brought 240 bundles (about 20,000 fascicles) of San-
skrit books by sea to China and was a prolific translator who rendered important treatises by
Asaṅga and Vasubandhu into Chinese. He was a leading proponent of the Vijñānavāda school
in China, and through his exegesis, he was an influential philosopher in his own right. This
does not preclude a Sāṃmitīya backgound: regardless of their monastic affiliation, monastics
were free to follow their philosophical consciences.
The association of teachers and philosophers to the Vinaya schools is not straightforward.
Monks ordained within the Sāṃmitīya tradition could master and write on non-Sāṃmitīya phi-
losophy. An outstanding example is Vimuktisena, who composed a commentary on the Orna-
ment of Realization (Abhisamayālaṃkāra), a verse summary of the Perfection of Wisdom that
tradition ascribes to Maitreya bodhisatva,8 who transmitted it to Asaṅga in the Tuṣita heaven.
The colophon of the Sanskrit manuscript and the Tibetan translation describe Vimuktisena as
a monk of the Kaurukulla tradtion of the Ārya Sāṃmitīyas. His Sāṃmitīya background did
not inhibit him from mastering an abstruse and decidedly non–pudgala-oriented work like the
Ornament of Realization.
Toward a Chronology of Vātsīputrīya/Sām

 mitīya Ideas
The Vātsīputrīya/Sāṃmitīyas did not invent the word pudgala, which is frequent in the early
Sūtras. It is embedded in the dialogues and discourses that record the Buddha’s teachings
over his forty-five–year career; the literature on the topic is immense and I am unable even to
attempt to summarize it here.
It is hard to see how terms for “individual” or “person” can be avoided in any language. In
standard Buddhist usage, the Buddha is a great man (mahā-puruṣa) and the eight types of real-
ized person are called the “eight individual persons” (Pali, aṭṭha-purisa-puggala), a term that
is part of the homage to the three jewels in the hymnology of the recollection of the Buddhist
order.9 In some passages, the word pudgala is prominent; the Vātsīputrīya/Sāṃmitīya draw
attention to these in the formalized debates in the treatises. The Designation of Persons (Pug-
galapaññatti), one of the seven books of the Theravādin Abhidhamma, even uses the word
puggala in its title. It deals with types or categories of individuals. Whenever the Designation
of Persons was compiled, the term puggala was still quite respectable. Similar lists of pudgala
196
were drawn up by other schools, although none of them devote a whole work to the categories.
The lists and definitions define the grades of attainment and map the paths and goals. The dif-
ference seems to be that the Vātsīputrīya/Sāṃmitīyas chose to integrate the pudgala into their
dharma system and relate it directly to the aggregates, sense-bases, and elements, and to the
processes of rebirth and the realization of nirvāṇa. This was novel, if not radical.
Critiques and Doxographies

Some of the great Buddhist metaphysicians incorporated critiques of the concept of the
pudgala into their works: for example, Harivarman (c. 250–350), Asaṅga (fourth century),
Bhāviveka (c. 500–570), Candrakīrti (c. 600–650), and Śāntarakṣita (eighth century, com-
mented upon by his disciple Kamalaśīla) (Priestley 1999, 48–52; Duerlinger 2009, 80–85).
Inventories of Sāṃmitīya tenets are given in Vasumitra’s Treatise of the Wheel of the Different
Divisions of the Tenets (Samayabhedacakra) and related doxographies preserved in Tibetan
and Chinese, and in Bhāviveka’s Blaze of Reason (Tarkajvālā). These doxographies are promi-
nent in Tibetan scholarship up to the present day, and major Tibetan doxographical works
deal with the subject in some detail. That the schools gave so much space to the Vātsīputrīya/
Sāṃmitīyas reflects the historical weight of their ideas.
Not much information is available after the travel accounts of the itinerant seventh-century
Chinese scholar-monks Xuanzang and Yijing. By the eleventh century, we get miscellaneous
mentions in written sources from India and Tibet. A Sanskrit history of the Vajrayoginī cult
from Nepal notes that the (future siddha) Maitrīgupta ordained as a Sāṃmitīya at Vikramapura
in the first half of the eleventh century and “listened to the Sūtra, Abhidharma, and Vinaya”
(Lévi 1931, 423).10 The Tibetan historian Tāranātha refers to a Sāṃmitīya master named
Jñānamitra, a śūdra (a man of lowest caste) of Tripura in eastern India, who was ordained
at Jagaddala monastery in northern Bengal and was learned in the Vinaya and Abhidharma
Piṭakas according to the Sāṃmitīya system (Tāranātha in Templeman 1983, 53). These ref-
erences suggest that Sāṃmitīya ordination rites could be performed at Vikramapura and
Jagaddala monasteries, where instruction in their Tripiṭaka was also available – if not at these
monasteries themselves, then elsewhere in northern India.
This was the time when Daśabalaśrīmitra, a twelfth-century paṇḍita, compiled his Exami-
nation of the Compounded and Uncompounded, a copious compendium of philosophy
according to the Sarvāstivāda, Sthāvira, and Sāṃmitīya systems, as well as the Yogācāra and
Madhyamaka. The presentation of the Sāṃmitīya doctrines, integral to his work, shows that
Daśabalaśrīmitra had direct access to their texts. Sarvarakṣita’s Great Discourse on Evolution
and Devolution continued to be copied in the Kathmandu valley. The Patna Dhammapada and
the Stanzas of the Compendium of Abhidharma were recovered in Tibet in the first half of the
twentieth century by Sankrityayan and Tucci, respectively. In addition to this, an unknown
number of palm-leaf manuscripts preserved in Tibet are written in bhaikṣukī script, in what we
might call a “Sāṃmitīya Prakrit,” showing that the school’s manuscripts continued to circulate.
At a certain point – we cannot say when – ordinands of the school began to study, practice,
and contribute to Mahāyāna thought. This may be equally true for members of other schools.
Modern historians, influenced by the model of Christian history, have tended to overemphasize
exclusivity, division, and sectarianism, schismatic thought that might be alien to the ecology of
Buddhist and Indian ideas. An outstanding example is Vimuktisena, previously mentioned. This
raises the question of the significance and boundaries of school affiliation. Monks like Vimuk-
tisena seem to have been Sāṃmitīya by ordination only – not by adherence to its philosophical
system.
197
Peter Skilling
The last report of Sāṃmitīya activity comes from the 1460s from central Tibet, where a monas-
tic document records the visit of an itinerant monk named Lokottara who hailed from eastern India
and was traveling from Kashmir to central Tibet. He stated that he belonged to the Sāṃmitīya
school and was a student of Madhyamaka, Pramāṇa, and Vajrayāna (Van der Kuijp 1995).
Recovery: Sām
 mitīya Texts in Tibetan and in Indian Languages
Daśabalaśrīmitra
A significant advance in Vātsīputrīya/Sāṃmitīya studies has been the study of Daśaba
laśrīmitra’s Analysis of the Compounded and the Uncompounded, the only Indian work pre-
served in Tibetan translation that gives direct access to the thought of the Saṃmitīya school.
The Paṇḍita, who most probably wrote in the twelfth century, devotes six complete chapters to
their dharma system and gives other doctrines elsewhere in his work. This encyclopedic work
is almost entirely lost in the original Sanskrit (but for three folios preserved in Cambridge Uni-
versity Library), but it is preserved in an excellent Tibetan translation (Szántó 2020, 129–45).
Each chapter is a direct citation of (a) Sāṃmitīya text(s) chosen by Daśabalaśrīmitra. In this
regard, the Sāṃmitīya chapters parallel Daśabalaśrīmitra’s citations of Sthāvira material
which are taken directly from a single source, the *Vimuttimagga.
Sāṃmatīya Buddhology
Daśabalaśrīmitra provides invaluable information on the concept of the Buddha (Skilling
2006). The Great Discourse on Evolution and Devolution gives a short verse summary of the
events of Śākyamuni’s life.
Sāṃmatīya Cosmology
Cosmology and cosmogony deal with the shape, structure, and formation and dissolution of
the world system in which we humans live, including systems of time and eons. Two texts
deal with this subject. One is Sarvarakṣita’s Sanskrit verse, Great Discourse on Evolution and
Devolution, and the other is a prose section of the Tibetan Analysis of the Compounded and the
Uncompounded. These show that the school had a fully articulated cosmology that – in broad
strokes – resembles the cosmology of other schools.
Sāṃmatīya Scholasticism
The Vātsīputrīya/Sāṃmitīyas developed a unique interpretation of the stages of the path
(mārga) to release from suffering. The different schools disagreed on the number of kar-
mic destinies or realms (gati). Some maintained that there are five, others that there are six.
A major point of disagreement was whether or not there is an intermediate state (antarābhava)
between death and rebirth. Daśabalaśrīmitra confirms other available texts that the Sāṃmitīyas
accepted six destinies and the intermediate state.
Most of the material cited by Daśabalaśrīmitra is abhidharmic in subject matter and style.
Since the school’s Abhidharmapiṭaka does not survive, we cannot trace their source. The
Stanzas of the Compendium of Abhidharma is an Abhidharma treatise composed in concise
and tightly packed stanzas (kārikās); it resembles and at times even parallels Vasubandhu’s
Treasury of Abhidharma (Abhidharmakośa). Because the verses are compact and laconic,
198
they are difficult to understand without a commentary. Some of the categories outlined in
Daśabalaśrīmitra’s citations and the verses of the Stanzas of the Compendium of Abhidharma
are the same as those of the Sarvāstivādin Abhidharma. This suggests that the dharma theories
of the two schools developed in tandem and that, despite developing into distinct schools,
they share core categories and doctrines. A text recently recognized as a Sāṃmatīya work is
a “preacher’s manual,” dating perhaps to the fifth century. Composed in Sanskrit verse and
prose, grounded in canonical citations in “Sāṃmitīya Prakrit,” this unique text deals with
many of the practical undertakings of Buddhist practitioners in early South Asia (Szántó 2019).
Were the Sām

 mitīyas and Vātsīputrīyas Buddhists?
One question that has rankled Buddhist intellectuals is this: were the Vātsīputrīya/Sāṃmitīyas
Buddhists? (La Vallée Poussin 1925b, 227–28). The doubts began in India, from, most prob-
ably, the very beginnings of the great pudgala debate. The question continued to be discussed
in Tibet, and it is still debated in modern Buddhist studies. The question depends on context
and the definition of “Buddhist”; there is no easy answer and no single answer.
Buddhism had several canons, not a single canon. It had several Tripiṭakas – perhaps eight-
een or more – not a single Tripiṭaka. Tripiṭakas were dynamic repositories of ideas, not static
and passive collections; even when they were fixed and the texts did not change, the uses and
interpretations did. Buddhism did not have a single track or path, but developed several –
even multiple – systems of practice and thought. One of these systems was the Vātsīputrīya/
Sāṃmitīya, who transmitted their own canon. A canon is a collection and accumulation of the
ideas and practices of a school, transmitted orally for centuries before being written down; each
school needed such a corpus for the study and practice of Śākyamuni’s Dharma. Debate on the
canonicity of texts continued among the reciters and was never fully resolved. For example, the
Miscellaneous Collection (Khuddakanikāya) of the Theravādin canon contains – in addition to
early works shared with the other schools – a number of oddities, commentaries, Abhidhamma,
and advanced buddhologies that were and became canonical. But they were canonical for the
Theravādins alone: other schools would not consider them canonical or even know them.
The members of all eighteen Buddhist schools regarded Śākyamuni as their teacher and they
were all deeply and sincerely devoted to his person and teaching. All of the eighteen schools
traced their scriptural collections, their Piṭakas and Tripiṭakas, back to the Buddha himself
through the recitation-redaction of the scriptures held at Rājagṛha after his decease. All of them
preserved and nurtured Sūtra and Vinaya collections, and developed Abhidharma systems.
Yaśomitra comments on whether or not the Sāṃmitīyas’ notions of the pudgala disqualify
them as Buddhists, saying, “It is not that there is no liberation for the Vātsīputrīyas: there is,
because they are Buddhists” (Wogihara 1932–1936, 699.4). In his commentary on the Intro-
duction to the Bodhisatva Path (Bodhisatvacaryāvatāra), however, Prajñākaramati refers to
the pudgalavādins as “tīrthikas within the Buddhist fold” (La Vallée Poussin 1898, 297.8). In
his General Presentation of the Classes of Tantra, Mkhas grub rje (1385–1438), a close dis-
ciple of Tsongkhapa, presents another angle: “pudgalavādins are not Buddhist because they
believe in the pudgala, but they are Buddhist because they take refuge and practice ethics,
etc.” Some say that the question rests with whether or not the schools upholds the “four seals”:
1. The saṃskāras are all impermanent.

2. Everything with flux (sāsrava) is suffering.
3. All natures are devoid of self.
4. Nirvāṇa is tranquil and solitary.
199
Peter Skilling
Some say they do, and some say they do not. Once again, there are no Sāṃmitīyas to speak up
and clarify their position.
New Perspectives: Fragments From a Sām

 mitīya World
Inscriptions on Sāṃmitīya statues or their pedestals bear a canonical citation that appears to
be a Prakrit version of the summary of the Buddha’s teaching given by the monk Aśvajit to
Śāriputra – a celebrated event that is recounted in the scriptures of all known schools, in San-
skrit, Pali, hybrid Sanskrit, and in a variety of translations.
The Fortunate One, dear sir, teaches the nature of dependent arising and the cessation
of dependently arisen dhammas:
The Tathāgata teaches the cause

Of the dhammas that arise from causes
As well as their extinction.
Such is the teaching of the Great Ascetic.
The ye dharmā verse or stanza of dependent arising was, and is, widely disseminated in lit-
urgies, inscriptions, and manuscripts across the Buddhist world. That the Sāṃmitīyas chose
to inscribe this particular text shows that like the other schools they regarded, or revered,
conditionality as the very heart or core of the Buddha’s teaching. The school’s public face is
dependent arising and the stanza of causation – not the pudgala.
The Vātsīputrīya/Sāṃmitīya was not a marginal group: it was a fully fledged Buddhist
school. Priestley notes that “this was not a small sect on the fringes of Buddhism, but rather an
integral part of the Indian Buddhist tradition” (Priestley 1999, vii). It was one among equals.
We now have access to some of their works in Indian languages and translation, and further
studies will allow us to appreciate more clearly their role in Buddhist thought and culture. These
fragments from a Sāṃmitīya world give a new perspective on the school, suggesting that it was
a vital and creative presence in north India up till the final age, the decline and disappearance
of Buddhism. There seems to be a particular nexus of activity in the eleventh–twelfth centu-
ries in north India: Sarvarakṣita’s literary productions, the copying of Bhaikṣukī manuscripts,
the inscribing of images, and the ordinations as Sāṃmitīya monks, for example. This activity,
and the accessibility of Sāṃmitīya materials, is reflected in Daśabalaśrīmitra’s decision to give
significant space to Sāṃmitīya doctrines. Marginality and centrality are subjective categories,
dependent on the writer’s choice of a center. I am ready to say that throughout the history of
Indian Buddhism, the Vātsīputrīya/Sāṃmitīyas were anything but marginal, either as a monas-
tic order or as an intellectual current. They were a significant presence in the material and meta-
physical landscape of northern and western India from the beginnings of post-Nirvāṇa Buddhist
philosophy, over a period of 1,500 years. It is time to recognize this in Buddhist studies.
In a work published in 1999, Leonard Priestley wrote:
It is hardly necessary to point out the importance, both philosophically and histori-
cally, of a form of Buddhism which differs strikingly in its interpretation of the Bud-
dha’s teaching from what we have come to regard as orthodox, and yet was for some
time, at least, the dominant form of Śrāvakayāna Buddhism in India. But the difficul-
ties facing us in investigating the Pudgalavāda are considerable.
200
Orthodoxy and unorthodoxy are troublesome notions, and what role, if any, they have in
the study of a fluid and dynamic Indian Buddhism – or, if we like, Indian Buddhisms – that
evolved for more than 2,000 years is open to question.
In this short chapter, I have attempted to point out the significance of the Vātsīputrīya/
Sāṃmitīya tradition. Many difficulties still face us in the study of this school of practice and
thought, but the new materials outlined in this chapter should advance our understanding con-
siderably. The title of an earlier work referred to “rehabilitating the pudgalavādins.” By this,
I meant bringing them in from the cold by showing that they were one of the Buddhist schools
that significantly contributed to the development of Buddhism. I hope to encourage the non-sec-
tarian study of Buddhist monastic culture in the broadest sense, embracing literature, ritual, art,
and archaeology, as far as resources allow. By non-sectarian, I mean without taking any tradition
as original or authentic, but rather examining Buddhist intellectual culture with open minds.
Notes
1 I am grateful to Eng Jin Ooi for help with the final stages of this chapter.
2 I use the word “canonical” to mean belonging to the speech of the Buddha as preserved in the scrip-
tural collections of the different Buddhist schools without prejudice for any particular school.
3 See the remarks on “the term pudgalavāda, which scholars continue to use,” in Lusthaus (2009, 275),
and his n. 1, “I leave pudgalavāda in the lower case . . . in order to indicate that it is not the proper
name of a school or sect but a label attached to the Vātsīputrīya, Sāṃmitīya, etc. schools by their
opponents.” With this politics of capitalization I concur.
4 I do not refer here to an extensive literature in Japanese, which includes translations of some of the
key works from Chinese and Tibetan.
5 A tīrthika, “one who leads or guides to and across the ford to liberation,” is a general Buddhist term
for other spiritual traditions. It is often inappropriately translated as “heretic.”
6 Exceptions are the work of early generations of European scholars such as Schiefner, Wassilief, Przy-
luski (1927), and La Vallée Poussin, and those of Bareau mentioned previously, as well as his Les
premiers conciles bouddhiques (Bareau 1955b).
7 It is found only in the London Dunhuang manuscript collection; present-day Tanjurs – Tibetan canons
of Buddhist scriptures – do not include the work, or any others by any author of this name.
8 In this essay, bodhisatva is spelled with a single “t” as it is spelled in the overwhelming majority of
manuscripts and inscriptions, as also in loan words in other languages like Tibetan or Thai. Bodhisat-
tva with a double “tt” is a modern normalization.
9 Arthaviniścayadharmaparyāya: text Samtani (1971, 47–48; trans. Samtani 2002, 33). The Commen-
tary (Nibandhana) gives an etymological explanation (Samtani 1971, text 253.2; 2002 trans. 182).
10 For the date, see Tatz (1987, 695–711). I suppose that Vikramapura = Vikramaśilā.
Bibliography
Bareau, André. 1954. “Trois traités sur les sectes bouddhiques attribuès à Vasumitra, Bhavya et Vinītadeva
1.” Journal Asiatique 242: 228–66.
———. 1955a. Les sectes bouddhiques du Petit Véhicule. Saigon: Ecole française d’Extrême-Orient.
———. 1955b. Les premiers conciles bouddhiques. Paris: Presses Universitaires de France.
———. 1956. “Trois traités sur les sectes bouddhiques attribuès à Vasumitra, Bhavya et Vinītadeva 2.”
Journal Asiatique 244: 167–200.
Cousens, Henry. (1929) 1975. The Antiquities of Sind with Historical Outline. Archaeological Survey of
India, vol. XLVI, Imperial Series. Calcutta, reprt. Varanasi: Bhartiya Publishing House.
Denis, Eugène. 1977. La Lokapaññatti et les idées cosmologiques du bouddhisme ancien, 3 vols. Lille:
Université de Lille.
Dimitrov, Dragomir. 2020. The Buddhist Indus Script and Scriptures: On the So-called Bhaikṣukī or Saindhavī
script of the Sāṃmmitīyas and their Canon. Wiesbaden: Harrassowitz Verlag (Veröffentlichungen der
Fächergruppenkommission für Außereuropäische Sprachen und Kulturen: Studien zur Indologie).
201
Peter Skilling
Duerlinger, James. 2009. “The Theory of Persons of the Pudgalavādins.” The Indian International Jour-
nal of Buddhist Studies 10: 75–90.
Hanisch, Albrecht. 2006. “Progress in Deciphering the So-called ‘Arrow-head’ Script Allowing Access
to Sarvarakṣita’s Maṇicūḍajātaka, a Text of the Buddhist Sāṃmitīya School.” Journal of the Centre
for Buddhist Studies Sri Lanka 4: 109–61.
———. 2008. “Sarvarakṣita’s Maṇicūḍajātaka. Reproduction of the Codex Unicus with Diplomatic
Transcript and Palaeographic Introduction to the Bhikṣukī Script.” In Sanskrit Texts from Giuseppe
Tucci’s Collection, edited by Francesco Sferra. Manuscripta Buddhica, 1; Serie orientale Roma.
Rome: Istituto italiano per l’Africa e l’Oriente.
La Vallée Poussin, Louis de. 1898. Bouddhisme, études et matériaux: Ādikarmapradīpa,
Bodhicaryāvatāraṭīkā. London: Luzac & Co.
———, ed. 1912. Madhyamakāvatāra par Candrakīrti – Trad. tibétaine. St.-Pétersbourg: Imprimerie de
l’Académie Impériale des sciences.
———. 1920. “Sammitiyas.” In Encyclopaedia of Religion and Ethics XI, edited by James Hastings,
168–69. Edinburgh: T. & T. Clark; New York: Charles Scribner’s Sons.
———. 1925a. Nirvâṇa (Études sur l’Histoire des Religions 5). Paris: Gabriel Beauchesne.
———. 1925b [1971]. L’Abhidharmakośa de Vasubandhu V. Repr. Brussels: Institut Belge des Hautes
Études Chinoises.
———. 1962. Catalogue of the Tibetan Manuscripts from Tun-huang in the India Office Library. Lon-
don: Oxford University Press for the Commonwealth Relations Office.
Lamotte, Étienne. 1944. Le traité de la grande vertu de sagesse de Nāgārjuna (Mahāprajñāpāramitāśāstra),
vol. I. Louvain: Bureaux du Muséon.
———. 1958. Histoire du bouddhisme indien: Des origines à I’ère Śaka. Louvain: Institut Orientaliste.
Lévi, Sylvain. 1931. “Un nouveau document sur le bouddhisme de basse époque dans l’Inde.” Bulletin of
the School of Oriental Studies, University of London 6 (2): 417–29.
Li Rongxi, trans. 1996. The Great Tang Dynasty Record of the Western Regions. Berkeley: Numata
———, trans. 2000. Buddhist Monastic Traditions of Southern Asia: A Record of the Inner Law Sent
Home from South Seas. Berkeley: Numata Center for Buddhist Translation and Research.
Lusthaus, Dan. 2009. “Pudgalavāda Doctrines of the Person.” In Buddhist Philosophy: Essential Read-
ings, edited by William Edelglass and Jay L. Garfield, 275–85. New York: Oxford University Press.
Okano, Kiyoshi. 1998. Sarvarakṣitas Mahāsaṃvartanīkathā: ein Sanskrit-Kāvya der Sāṃmitīya-Schule
des Hīnayāna-Buddhismus. Sendai: Seminar of Indology, Tohoku University.
Pal, Pratapaditya, ed. 2008. Sindh: Past Glory, Present Nostalgia. Mumbai: The Marg Foundation.
Priestley, Leonard. 1999. Pudgalavāda Buddhisṃ: The Reality of the Indeterminate Self. South Asian
Studies Papers, no. 12, Monograph no. 1. Toronto: University of Toronto.
———. n.d. “Pudgalavada Buddhist Philosophy.” Internet Encylopedia of Philosophy. Accessed June 25,
2020. www.iep.utm.edu/pudgalav/.
Przyluski, Jean. 1927. Le concile de Rājagṛiha: introduction à l’histoire des canons et des sectes boud-
dhiques. Paris: Paul Geuthner.
Ronkin, Noa. 2005. Early Buddhist Metaphysics: The Making of a Philosophical Tradition. London:
Routledge.
Samtani, Narayan Hemandas, ed. 1971. The Arthaviniścaya-Sūtra & Its Commentary (Nibandhana)
(Written by Bhikṣu Vīryaśridatta of Śri-Nālandāvihāra), Critically Edited and Annotated for the First
Time with Introduction and Several Indices. Patna: K. P. Jayaswal Research Institute.
———, trans. 2002. Gathering the Meanings: The Compendium of Categories: The Arthaviniścaya
Sūtra and Its Commentary Nibandhana. Berkeley, CA: Dharma Publishing.
Skilling, Peter. 2006. “Daśabalaśrīmitra on the Buddhology of the Sāṃmitīyas.” In Nagoya Studies in
Indian Culture and Buddhism – Sambhasa 25, 98–123. Nagoya: Department of Indian Studies, Grad-
uate School of Letters, Nagoya University.
Stcherbatsky, Theodar. 1919a. “The Soul Theory of the Buddhists.” Bulletin de l’Académie des Sciences
de Russie VI série 13 (12–15): 823–54.
———. 1919b. “The Soul Theory of the Buddhists.” Bulletin de l’Académie des Sciences de Russie VI
série 13 (16–18): 937–58.
Szántó, Péter-Dániel. 2019. “The *Saddharmaparikathā, a Buddhist Preachers’ Manual from Late Antiq-
uity.” Presented at the University of Vienna, 22 March; Unpublished Draft Courtesy the Author.
202
———. 2020. “A Sanskrit Fragment of Daśabalaśrīmitra’s Saṃskṛtāsaṃskṛtaviniścaya (Ch. 29 & 30).”

Annual Report of the International Research Institute for Advanced Buddhology 23: 129–45.
Tatz, Mark. 1987. “The Life of the Siddha-Philosopher Maitrīgupta.” Journal of the American Oriental
Society 107 (4): 695–711.
Taylor, C. Arnold, ed. 1894. Kathāvatthu. London: The Pali Text Society.
Templeman, David, ed. and trans. 1983. The Seven Instruction Lineages (Jo Nang Tāranātha). Dhar-
masala: Library of Tibetan Works & Achieves.
Thích Thiện Châu, Bhikshu. (1996) 1999. The Literature of the Personalists of Early Buddhism. Trans-
lated by Sara Boin-Webb. New Delhi: Motilal Banarsidass Publishers.
Trenckner, V., ed. 1880. The Milindapañho: Being Dialogues Between King Milinda and the Buddhist
Sage Nāgasena. London: Williams and Norgate.
Van der Kuijp, Leonard W. J. 1995. “Some Indian and Sri Lankan Peregrinators in Central Tibet and
Glo Bo Smon Thang During the Fifteenth Century.” Presented at the 7th Seminar of the International
Association for Tibetan Studies, Graz.
van Lohuizen, J. E. 1981. “The Pre-Muslim Antiquities of Sind.” In Sind Through the Centuries: Pro-
ceedings of an International Seminar Held in Karachi in Spring 1975 by the Department of Culture,
Government of Sind, edited by Hamida Khuhro, 43–57. Karachi: Oxford University Press.
Wogihara, Unrai, ed. 1932–1936. Sphuṭārthā Abhidharmakośavyākhyā by Yaśomitra, Part II. Tokyo: The
Publishing Association of Abhidharmakośavyākhyā.
Yamaguchi, Zuihō et al. (1977) 1986. スタイン蒐集チベット語文献解題目錄 (A Catalogue of the
Tibetan Manuscripts Collected by Sir Aurel Stein), Part Two. Tokyo: Toyo Bunko.
203
PART 4
Philosophical Founders
PHILOSOPHICAL FOUNDERS
The great flowering of Buddhist philosophy that takes place in India has no founder. Even the
Buddha himself is said to be just the most recent of those who have turned the wheel of the
Dharma for the sake of the world; according to Buddhist traditions, other buddhas preceded
him. But conventionally, it is nevertheless possible to identify developments that can, in retro-
spect, be recognized as new modalities in Indian Buddhist philosophical life. Of course, such
beginnings are happening constantly, and this volume aims to show how a great number of
individual philosophers leave their marks on how Buddhist philosophy gets done. Still, as in
the West, the philosophical work of some figures affects subsequent thinkers in the tradition;
in India, such figures have traditionally been identified as constituting new lineages of thought
and practice.
We identify the thinkers in this part as “philosophical founders,” taking our inspiration from
the popular Tibetan schema of the “six ornaments and two supreme ones” (rgyan drug mchog
gnyis). This group of eight luminaries, a popular subject for monastic scroll paintings, includes
the six philosophers treated in this part – Nāgārjuna, Āryadeva, Asaṅga, Vasubandhu, Dignāga,
and Dharmakīrti – along with two masters of monastic discipline (vinaya), Guṇaprabha and
Śākyaprabha. The six founders broadly also represent three significant streams in Mahāyāna
philosophical thought, that of the Madhyamaka, Yogācāra, and Pramāṇa traditions, with two
masters associated with each stream. We begin with Nāgārjuna, the founder of Madhyamaka,
along with his chief immediate disciple, Āryadeva. Following this, we have Asaṅga, the founder
(with Maitreya) of the Yogācāra, along with his chief disciple and brother, Vasubandhu. And
we finish with Dignāga, the founder of the Buddhist Pramāṇa stream, along with his most
influential disciple, Dharmakīrti. Each of these six figures is treated in this part by authors who
take distinct approaches. As they are all iconic for so many of the Buddhist thinkers who fol-
low them, readers will find their names referenced repeatedly throughout the rest of the book.
Nāgārjuna (second–third century CE) is widely regarded as one of the most influential philos-
ophers in any Buddhist tradition. His Root Verses of the Middle Way (Mūlamadhyamakakārikā),
along with the other analytic treatises and devotional hymns that articulate his account of depend-
ent origination and emptiness, became the basis of the Madhyamaka tradition of Mahāyāna
Buddhism. Nāgārjuna’s notoriously terse verses led to a great diversity of interpretations by
Buddhist intellectuals in India and in Central and East Asia, as well as among contemporary
scholars. In her chapter, Amber D. Carpenter engages the most prominent interpretations of
DOI: 10.4324/9781351030908-21 207

the Root Verses – as metaphysical, semantic, epistemological, and anti-rationalist – and offers
a reading that emphasizes Nāgārjuna’s return to the Buddha’s teachings on dependent arising
( pratītyasamutpāda). Thus, according to Carpenter, Nāgārjuna is a fundamentalist reformer,
responding to the intellectual world of second-century Indian Buddhism by returning to what he
believes is the true teaching of dependent arising and its rightful centrality in Buddhist thought.
In contrast to the view of dependent arising as the way in which things arise or are in relation
with each other, Nāgārjuna’s understanding of dependent arising functions to deconstruct meta-
physical conceptions of individuation as such; according to Carpenter, for Nāgārjuna, depend-
ent arising means there are no things arising. By linking the original teaching of dependent
arising with the Mahāyāna understanding of emptiness and the distinction between ultimate
and conventional truth, Nāgārjuna defends the Mahāyāna as the authentic embodiment of the
Buddha’s teachings. Carpenter sees in Nāgārjuna’s Precious Garland (Ratnāvalī) the implica-
tions for ethics of Nāgārjuna’s philosophy of dependent arising, emptiness, and the two truths.
Reading Nāgārjuna as embedded in communities of practice, she suggests that his ontology,
epistemology, and philosophy of language do indeed inform the Buddhist practical-soteriolog-
ical project.
Āryadeva (second–third century) is traditionally said to have been Nāgārjuna’s student and
is credited as a cofounder of the Madhyamaka. While many Indian philosophers argue that
possessing an essence or intrinsic nature is a condition for something to be real, Āryadeva
believes that there is no entity that meets this condition. Like Nāgārjuna, he employs a dia-
lectical method to deconstruct any claim of intrinsic nature; he argues that there is no essence
or identity that is somehow independent of causes and conditions. Phenomena are then pre-
cisely empty of the intrinsic natures that are attributed to them. Thus, we must abandon any
philosophical views that presuppose real entities. Focusing on Āryadeva’s most significant
work, the Four Hundred Verses (Catuḥśataka), Tom J. F. Tillemans gives an account of both
Āryadeva’s anti-realist metaphysics and ontology and his ethics, as well as how they may – or
may not – relate to each other. Āryadeva’s ethics are, as Tillemans argues, radically revisionist.
That is, Āryadeva proposes a supererogatory morality that calls for a kind of moral sainthood.
Tillemans suggests that there is a tension between Āryadeva’s descriptive ontology and revi-
sionary ethics, as the former undermines any possible philosophical foundation for the later.
Tillemans also explores Dharmapāla’s Yogācāra interpretation of the Four Hundred Verses
and what it might tell us about early Yogācāra-Madhyamaka syntheses.
The writings attributed to Asaṅga (fourth–fifth century) are foundational to the develop-
ment of the Yogācāra tradition of Mahāyāna Buddhism. Indeed, Asaṅga’s texts are some of
the earliest systematic presentations of core Yogācāra ideas. Jowita Kramer, in her chapter,
focuses on Asaṅga’s two most important texts: the Compendium of the Abhidharma (Abhi
dharmasamuccaya) and the Summary of the Mahāyāna (Mahāyānasaṃgraha). The Compen-
dium of the Abhidharma is a systematic work of what some scholars call Mainstream (i.e., not
Mahāyāna) Abhidharma that integrates characteristically Yogācāra ideas. In the Summary of
the Mahāyāna, Asaṅga articulates the core ideas that distinguish Yogācāra from other Bud-
dhist schools. These include the idealist ontology of the three natures, and that there is nothing
external to the mind – that there is nothing other than presentation to the mind (vijñaptimātra).
He also presents the doctrine of the eight forms of mind, including the storehouse mind
(ālayavijñāna), a form of mind beneath the other forms that is the store of karma. Kramer sug-
gests that while it may not be the case that Asaṅga himself first articulated these ideas, he was
likely the first author to develop them into one coherent philosophical system.
Perhaps no Indian Buddhist philosopher justifies a wariness about categorizing authors accord-
ing to schools as much as Vasubandhu (fourth–fifth century), traditionally said to be Asaṅga’s
208
younger brother. Vasubandhu is widely regarded as one of the founding fathers of the Yogācāra
school. Yet he wrote a comprehensive account of the Sarvāstivāda Abhidharma, the Treasury of
Abhidharma Verses (Abhidharmakośakārikā). At the same time, in his own detailed and very
influential Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya) – the touch-
stone reference for Abhidharma in Tibetan scholasticism – he subjects the root verses to critique
from multiple perspectives, especially from a view associated with the Sautrāntika school. Some
of Vasubandhu’s other texts defending an idealist ontology and associated Yogācāra ideas were
also highly influential. Still, as Jonathan C. Gold notes, while Yogācāra and Madhyamaka came
to be regarded as rival schools, Vasubandhu never actually uses the term “Yogācāra,” and he
does employ characteristically Nāgārjunian arguments. According to Gold, then, Vasubandhu
should be read not as sequentially representing one school after another, but rather as a phi-
losopher who works within multiple inherited traditions. Focusing on questions of causality and
language, Gold traces a common theme across the diversity of Vasubandhu’s texts. In particular,
Gold unpacks Vasubandhu’s view that while concepts and words cannot actually capture real-
ity, even false concepts, when used skillfully, can point to an ineffable reality beyond language.
Dignāga (late fifth–early sixth century) is credited with founding the influential Buddhist
tradition of logic and epistemology known as Pramāṇa. Of course, Dignāga inherited and was
informed by a rich tradition of philosophical inquiry into perception, cognition, and reasoning.
He was especially influenced by Vasubandhu’s writings on argumentation. Still, Dignāga’s
work, culminating in the Compendium of the Means of Valid Cognition (Pramāṇasamuccaya)
together with his autocommentary, constitutes a significant development in Indian logic and
epistemology. In contrast to some other Indian philosophers who accept verbal testimony
(śabda) as a distinct valid means of cognition ( pramāṇa), Dignāga understands it as a form
of inference (anumāna), one of only two valid means of cognition, along with perception
( pratyakṣa). He subdivides inference into inference-for-oneself and inference-for-others,
which makes it possible to treat argumentation within the broader epistemological category of
inference. As Kei Kataoka notes, this was part of Dignāga’s project of bringing epistemology
(theories of the means of valid cognition), dialectics (theories of argumentation), and semantics
(theories of word-referents) into one unified field of study. Relevant to all three domains, and
one of Dignāga’s most original and important contributions, is the idea of pervasion (vyāpti), a
form of invariable concomitance that can guarantee inferences. In addition to giving an over-
view of Dignāga’s project in the Compendium, Kataoka focuses on some of Dignāga’s most
crucial concepts, including the exclusion of what is other (anyāpoha), the three conditions
of a proper reason (trairūpyaliṅga), the self-awareness of cognitions (svasaṃvitti), and non-
perception (anupalabdhi). Kataoka carefully traces what Dignāga drew from his predeces-
sors, what constitutes his original contributions, and where his successors – the Mīmāṃsaka
Kumārila and the Buddhist Dharmakīrti – recognized the limitations of his system.
Developing the work of his predecessors, especially Vasubandhu and Dignāga, Dharmakīrti
(ca. sixth–seventh century) is regarded as a cofounder of the Buddhist school of logic and
epistemology, and was for centuries seen as the main representative of Buddhist philosophy
for non-Buddhist philosophers in India. His most important work, the Commentary on the
Means of Knowledge (Pramāṇavārttika), explicates and refines the Compendium of the Means
of Knowledge (Pramāṇasamuccaya), synthesizing Dignāga’s epistemology, dialectics, and
semantics with traditional Buddhist doctrines to give them a whole new significance. While
Dharmakīrti is generally associated with the fields of logic and epistemology, as John Taber
points out, he is primarily interested in defending the rationality of core Buddhist doctrines
as he understands them. Thus, in addition to addressing his original contributions to theories
of perception, inference, universals, and exclusions, Taber devotes much of his chapter to
209
Dharmakīrti’s apologetic concerns. These include Dharmakīrti’s defense of the Buddha as

tantamount to a means of knowledge ( pramāṇa) and the authority of Buddhist scripture, his
critique of physicalism and defense of antirealism and momentariness, and his disagreements
with Brahmanical philosophers. Throughout, Taber makes clear, Dharmakīrti is primarily
interested in logic and epistemology not for their own sake, but insofar as they can provide a
systematic and rational defense of the Buddhist soteriological project.
210
13
NĀGĀRJUNA
Dependent Arising Without
Any Thing Arising1
Amber D. Carpenter
With Nāgārjuna (c. second–third century CE), we come to the first named philosopher in
the Buddhist tradition, and certainly one of the most significant. As with Plato, half the fun
of Nāgārjuna (when it does not lead to blood on the carpet) is that there are nearly as many
Nāgārjunas as there are serious students of his work – with the added layer of complication that
it is even more radically contested what counts as Nāgārjuna’s work. Yet, so little is securely
known about Nāgārjuna the person that the best way to anchor a referent for the name is via his
works, traditionally including ‘analytical’ treatises, several practical works (for instance, the
Letter to a Good Friend [Suhṛllekha]), and a number of hymns. While scholars have disputed
whether the same person is indeed responsible for all these works, for philosophical purposes,
it is not necessary to adjudicate the matter. For within philosophical discourse, the analytic text
Root Verses of the Middle Way (Mūlamadhyamakakārikā; abbreviated as MMK) may be taken
as criterial: Nāgārjuna is, by definition, “the author of the Root Verses of the Middle Way,” the
founding text of what became over time a distinctive understanding of Mahāyāna Buddhist
thought known as Madhyamaka.2
About Nāgārjuna-the-author-of-the-Root Verses of the Middle Way as a person, we know
very little with certainty. It is said that he was born into a Brahmin family and converted to
Buddhism, and possibly even worked as advisor at the Sātavāhana court on the Deccan Pla-
teau.3 He wrote in Sanskrit, not in one of the various vernaculars as was common practice
among Buddhists at the time, thus situating his thought within the wider pan-Indic intellectual
discourse, both Buddhist and non-Buddhist.
It is broadly accepted that Nāgārjuna so defined also wrote the Dispeller of Dis-
putes (Vigrahavyāvartanī)4 and other treatises grouped together in Tibetan doxography as
Nāgārjuna’s ‘analytic’ works, namely: the Seventy Verses on Emptiness (Śūnyatāsaptati),
with accompanying commentary by Nāgārjuna and later commentaries by Candrakīrti and
Parahitabhadra;5 the Sixty Verses on Reasoning (Yuktiṣāṣṭika), widely-cited in Madhyamaka
debates (quotes appear in, for instance, Bhāviveka, Śāntarakṣita, and Candrakīrti, who also
wrote a commentary on the whole);6 Crushing the Categories (Vaidalyaprakaraṇa), which
critiques non-Buddhist metaphysical categories;7 and, possibly in this group, and at any rate
mostly agreed among scholars to be by the author of the Root Verses, the Precious Garland
(Ratnāvalī; abbreviated as RĀ), a letter of advice to a prince containing within it a healthy
measure of articulation of Nāgārjuna’s Madhyamaka.8
DOI: 10.4324/9781351030908-22 211

Amber D. Carpenter
Naturally, disagreements continue.9 But in some cases, at least, settling these disagreements
may not be vital for philosophical purposes: if the ideas in the Letter to a Good Friend, for
instance, are so generic that any Mahāyāna Buddhist might have composed them, then we
may allow the attribution to stand,10 so long as any interpretation of it is guided by an under-
standing of the Root Verses and not the other way round. The question acquires philosophical
interest only if taken as evidence for some broader claim – for instance, that the only ethi-
cal implications of Madhyamaka are generically Mahāyāna. Similarly, regarding the hymns
attributed to Nāgārjuna but possibly retrospectively ascribed to him: the Buddhist tradition
of associating these with Nāgārjuna is long and deep, and from the point of view of what the
figure of Nāgārjuna stands for in the development of Mahāyāna Buddhism, this association is
significant. For philosophical purposes, however, the Root Verses will be used to interpret the
hymns, and not the other way round.11
There is, of course, a great deal this leaves open. In particular, one might wonder whether
the largely metaphysical12 – or semantic,13 or epistemological,14 or anti-rationalist15 – Root
Verses have any bearing at all on ethical conduct. Here the beloved but disputed text, the Pre-
cious Garland, comes to the fore. If this text is by Nāgārjuna, in our sense of that name, then
it could give, first of all, some confirmation that Nāgārjuna took the views articulated in the
Root Verses to be consistent with continued engagement with conventional reality and with
hortatory reform within that discourse; and second, it would give some indication of what such
a Madhyamaka engagement might look like – or it might alternatively confirm that Madhy-
amaka differs from other Buddhist views only in its conception of the ultimate goal, without
that bearing on everyday conduct.
The Buddhism Nāgārjuna adopted upon entering the Buddhist fold was of the Mahāyāna
sort – although precisely what Mahāyāna meant, apart from the growing clarity that it was
different from the other Buddhist views and practices dominant at the time, was still under
construction. Indeed, Nāgārjuna’s Root Verses constitutes an important move in the formation
and definition of the Mahāyāna, as it gradually acquired that distinctiveness from other Bud-
dhist views which retrospectively appears so categorical.
In the early part of the first millennium of the Common Era, a subset of views within
the eclectic Abhidharma tradition became increasingly correlated and distinguished from all
other Abhidharma strands of thought. Usually associated with the Mahāsāṃghikas, these ideas
included the supermundane nature of the Buddhas, and bodhisattvas who make vows to remain
in saṃsāra for the sake of helping suffering beings. This subset of associated views began to
come together with the new injection of ideas – such as Buddha nature (tathāgatagarbha) and
an emphasis on emptiness – from the Perfection of Wisdom Sutras (Prajñāpāramitā Sūtras),
texts ‘discovered’ from perhaps the turn of the millennium and recognised by some – but not
all – Buddhists as authentic buddha-vacana, words of the Buddha.16 While Mahāyāna was,
by Nāgārjuna’s time, a distinct label for those working out an interpretation of the Buddha-
dharma centred on these notions, such an affiliation was not yet so distinct as to warrant
distinct communities of practice, and Mahāyānists of Nāgārjuna’s time and place would have
been practicing in mixed monasteries alongside other assorted Abhidharma Buddhists.17
It is within this dialectical context that Nāgārjuna offers his interpretation of the Buddha-
dharma. He is a Mahāyānist, and in that, he can be expected to be critical of Abhidharma
views where they differ from the Mahāyāna. At the same time, his aim is to legitimate the
Mahāyāna, and his strategy for doing so will be two-fold: first, to offer the most defensible
interpretation of key notions any Buddhist must recognise (and this will include the criti-
cal enterprise of demonstrating alternatives to be incoherent); second, to demonstrate that
this reading more accurately captures the original thought of the Buddha, from which his
212
Nāgārjuna
Abhidharma competitors offer unwarranted departures. Nāgārjuna is thus typical of a funda-

mentalist reformer, legitimating what is innovative by presenting it as a reclamation of the
original thought from which more established interpretations have in fact departed.
Nāgārjuna seems not to have had much immediate impact on Buddhist thought, whether
because his view was simply unexceptionably Mahāyānist18 or because his argument was to
Sanskrit speakers so obviously fallacious.19 Although there were apparently others, only one
early commentary on the Root Verses, entitled Fearless (Akutobhayā), survives at all. And
there was an immediate successor, Āryadeva, who took up the Madhyamaka banner ener-
getically enough to be considered nearly a co-founder of this distinctive interpretation of the
Buddha’s teachings. A flurry of interest some centuries later, in the commentaries and debates
of Buddhapālita and Bhāviveka in the late fifth century, and Candrakīrti in the sixth, revived
Madhyamaka as a (to varying degrees distinct) Mahāyāna alternative to Yogācāra.20 Since the
Root Verses is an enormously opaque text, most attempts to come to grips with Nāgārjuna’s
thought therefore do so by reaching back through these layers of interpretation, particularly
through the Tibetan reception of Candrakīrti’s interpretation of Madhyamaka.21
In what follows, however, I will use a reading of the Root Verses to anchor an articulation
of Nāgārjuna’s thought from the other direction – viz., as it arose in the intellectual context of
second-century CE Buddhist thought. Against the background of his own times, Nāgārjuna’s
fundamentalist reformation hinges specifically on a recovery of what he takes to be the origi-
nal and correct teaching of dependent arising ( pratītyasamutpāda), and the centrality this
ought to hold in relation to all other teachings. Associating the distinctively Mahāyāna notion
of emptiness with dependent arising should legitimate the Mahāyāna as more authentically
capturing the Buddha’s original thinking than subsequent Abhidharma thought had managed.
On Nāgārjuna’s view, dependent arising should replace – and not supplement or qualify –
conceptions of being, thus undermining all metaphysical individuation tout court. Only this,
he argues, will free us from the mesmeric power of individuating misconceived as appre-
hension instead of creation. The epistemic implications of abandoning individuation (without
individuation, there is no assertion) explain both the style of the Root Verses and the elusive,
apparently sceptical manoeuvres of the Dispeller of Disputes. And finally, so understanding
the nature and centrality of dependent arising indeed revises the shape of our practical-soteri-
ological task, from one of seeing ultimate reality as it is (instead of as it is constructed) to one
of seeing experienced reality as dependent, or even seeing the ultimate reality of dependence
in the fabrication of ordinary thought and life. The implications of this can be seen in the Pre-
cious Garland, which bridges everyday practical and ultimate concerns.
Fundamentalist Reformer: Centering Dependent Arising

To see which fundamentals of the Buddha’s teaching Nāgārjuna wants to reclaim, there is
nowhere better to look than the verse of praise of the Buddha that frames the Root Verses.
These apparently formulaic verses which open Buddhist Sanskrit philosophical texts indi-
cate the author’s priorities and the lens through which we are to understand the text which
follows.22
The Root Verses are framed with:
I salute the Fully Enlightened One, the best of orators, who taught the doctrine of
dependent arising, according to which there is neither cessation nor arising, neither
annihilation nor the eternal, neither singularity nor plurality, no coming nor going,
for the auspicious cessation of prapañca.23
213
Amber D. Carpenter
Exactly what this prapañca is which ceases will have to emerge from an interpretation of the Root
Verses. According to Siderits and Katsura (2013), “pra + √pañc literally means to be prolix or
excessively wordy” (NMW, 126); it is often glossed in Buddhist-English discourse as ‘conceptual
proliferation,’ and that will do as a working approximation. The relevant point for present pur-
poses is that it is the teaching of dependent arising for which the Buddha is praiseworthy; and this
teaching is the key to realising the supremely desirable end, characterised here as the stilling of
prapañca. Nāgārjuna does not pick out the Buddha as a knower nor as supremely compassionate,
and he entirely neglects familiar themes distinctive of the Buddha’s teaching – no self, non-cling-
ing, impermanence, suffering and the roots of suffering.24 The opening verse does not so much as
hint at themes more closely associated with the Mahāyāna, such as emptiness (śūnyatā).25 Instead,
Nāgārjuna selects dependent arising as the sole basis for the Buddha’s praiseworthiness.
Half the verse,26 in fact, is given over entirely not to praise of the Buddha but to characteri-
sation of dependent arising – and the manner of its characterisation is striking. First, depend-
ent arising is characterised entirely through matched pairs of negations, which is even more
evident in the Sanskrit where each word of the first two lines begins with an alpha-privative.
This is very far indeed from earlier sutta formulations of dependent arising as “from this, that”
– as, for instance, “with ignorance as condition, formations [come to be]; with formations as
condition, consciousness; with consciousness as condition, materiality-mentality.”27 Second,
one of those negations is actually a negation of arising itself – unarising (anutpādam) seem-
ing to negate the very thing that is asserted in dependent arising ( pratītysamutpādam). This
is an acute form of the tension pervasive in the whole verse, which holds dependent arising
responsible for an auspicious stillness (upaśamaṃ, translated ‘cessation’ above). This ten-
sion in honouring the Buddha as the teacher of unarising dependent arising signals that not
only will Nāgārjuna hang his exposition of the Buddha-dharma on pratītyasamutpāda but his
understanding of this central term will be challenging to the received orthodoxy.28
Nāgārjuna therefore opens the Root Verses proper by squarely situating his project with
respect to the incontestable received teachings, the Discourses of the Buddha:
Not from itself, not from another, not from both, nor without cause; Never in any way
is there any existing thing [bhāvāḥ] that has arisen.
(MMK 1.1)
In its deployment of the tetralemma – presenting and rejecting four ways beings (bhāvāḥ)
might arise – this verse, and the chapter it announces, epitomises what the Naked Ascetic Kas-
sapa Sutta (Acelakassapasutta, SN 12.17/ii.18–21) expresses more expansively, also deploy-
ing the tetralemma.29 In the Naked Ascetic Kassapa Sutta, the insistent ascetic Kassapa asks
the Buddha whether suffering is created by oneself, by another, or both, or whether it arises
spontaneously, and in each case, the Buddha replies negatively. He also rejects the suggestion
that perhaps there is no suffering or he does not know that there is suffering. The baffled Kas-
sapa asks the Buddha to teach him about suffering. “Kassapa,” the Buddha replies,
[if one thinks,] “The one who acts is the same as the one who experiences [the
result],” . . . this amounts to eternalism. But, Kassapa, [if one thinks,] “The one
who acts is one, the one who experiences [the result] is another,” . . . this amounts
to annihilationism. Without veering towards either of these extremes, the Tathagata
teaches the Dhamma by the middle: “With ignorance as condition, volitional forma-
tions [come to be]; with volitional formations as condition, consciousness. . . . Such
is the origin of this whole mass of suffering.”30
214
Nāgārjuna
While this discussion clearly has a pragmatic import – it is the origin of suffering that is under
discussion – the Buddha’s answer addresses the point from the most general angle, namely,
how one should think of the origins of anything that arises. The false view of eternalism arises
if one thinks, quite generally, “the one who acts is the same as the one who experiences” the
result, while nihilism arises from supposing quite generally “the one who acts is different from
the one who experiences.” Nāgārjuna’s first verse, seen in this connection, only makes explicit
the generality implicit in the sutta’s tetralemma and, picking up on the ‘created by’ of the sutta,
places the discussion squarely in the context of causes or conditions of arising.
With such a sutta passage in the background, Nāgārjuna’s decision to present a discus-
sion of conditions in the form of a tetralemma reinforces the message of the praise verse,
that dependent arising is the focus of attention, since that is what is positively asserted as the
Buddha’s teaching at the end of the sutta passage. Nāgārjuna’s audience will more specifically
be primed already to anticipate the untenable implications associated with each limb of the
tetralemma: the first option implying eternalism, the second nihilism, the third presumably
inheriting the faults of both – since the “Dhamma by the middle” avoids both extremes, rather
than mixing them – and the last is simply rejected.
The explicit reiteration in Root Verses 1 of so much of the Naked Ascetic Kassapa Sutta
draws attention to what it leaves tacit, namely “the Tathāgata teaches the Dhamma by the
Middle.” If there is any work for Nāgārjuna to do, it will be in interpreting this, the proposed
alternative to the four rejected options – “with ignorance as condition, volitional formations
[come to be]; with volitional formations as condition, consciousness” and so on, up to “this
whole mass of suffering.”31 Here is where Nāgārjuna will challenge prevailing interpretations
and offer his own version of dependent arising as an alternative.
What are these prevailing interpretations that Nāgārjuna challenges? While the centuries
of interpretation since the Buddha’s parinirvāṇa took his teachings in various directions, and
while it is difficult to determine with precision how exactly the various interpretations were
developed at the time of the composition of the Root Verses, the evidence of a canonical
Abhidharma text such as the Conditional Relations (Paṭṭhana) suggests that “the Dhamma by
the middle: ‘With ignorance as condition, volitional formations [come to be]’ ” was predomi-
nantly interpreted as a claim about conditions and causes. Conditions for arising were many
and various – and identifying, distinguishing, and describing the conditions by which things
arise was tantamount to grasping dependent arising.32 For anything arising, there are multiple
conditions for its arising and multiple further events or items for which it is a condition. This
is dependent arising as a sort of hyper-distributed, massively ramified causation.33
This view is most naturally put in an ontological mode, but it can also characterise the phe-
nomenological practices ascribed to the Theravādins of identifying and describing mental phe-
nomena as conditions under which further mental phenomena arise.34 Whether ontological or
phenomenological, appeal to dependent arising so conceived works as a part of Buddhist doc-
trine and practice by deflecting away from the tendency to pick out and privilege one thing as
the cause, without asserting the sheer non-existence of causes (thus avoiding being stricken
by nihilism). Instead of one cause, there are always many causes, none of them autonomous
(self-causing, or uncaused). Meditative practices of distinguishing the many contributing factors
from one another enable us to recognise this massively interconnected network of dependencies.
Any apparent object of desire or aversion, fear, or pride thus dissolves into countless strands of
endlessly ramifying causal conditions, undermining the coherence of these afflictive emotions.
This general metaphysical point translates immediately to the personal context: there are
no autonomous acts, and therefore no agents as we might feel ourselves to be or aspire to be.
The self – both as agent and subject – dissolves into countless dependencies on both personal
215
Amber D. Carpenter
and impersonal factors. This eliminates the coherence of the notion of a real self that might be
successfully distinguished from the rest. The result is not the discovery of the true self, but the
acceptance that there is no self to seek, and no empty hole where a self could be.
Nāgārjuna’s Critique: Its Target and Motivation

If this aptly, if impressionistically, characterises Abhidharma thinking about dependent aris-
ing, what, we might ask, could Nāgārjuna possibly find objectionable?
We might begin at the broad exegetical level. In the tetralemma of SN 12.17, all four limbs
are rejected, and Root Verses I.1 explicitly picks up on this. The “Dhamma by the Middle” –
that is, dependent arising – is offered instead of any of those four options. But any analysis of
dependent arising that figures it as causes and conditions, however multiple, must necessar-
ily present these multiple conditions and their results as either the same as or different from
each other – that is, as either the first or second limbs of the tetralemma. But these options (as
well as the option of combining or rejecting both), were rejected by the Buddha, as leading to
eternalism and nihilism, respectively. So, any apt understanding of the Doctrine by the Middle
must construe it as an alternative to causes and conditions, not as a special version of it.
This would be a textual basis for Nāgārjuna’s objection, and for supposing this is what he
is objecting to. But his objection, seen in this light, is not to any particular theory of causation,
and it would hold even when all parties acknowledge that causation, as a relation, is merely
nominal and not ontologically robust. Indeed, Nāgārjuna’s objection is only incidentally about
causation at all: in the first place, because distributed causation is how Ābhidharmikas mistak-
enly construe dependent arising (the topic it is essential to get right); and in the second place,
because causal relations share problematic features of any relations and model the relevant
features of any asymmetrical or asynchronous relation. The Buddha offered the Dhamma of
the Middle as an alternative to either identifying or distinguishing cause and effect. What his
Dhamma of the Middle rejects, then, is the identity or difference of any cause(s) and its/their
effect(s) – and by extension, says Nāgārjuna, of any relata. The objection, it seems, is to the
individuation of relata and, in Root Verses 1, of cause and effect in particular.
Thus, Root Verses 1 aims to show that if the relata ‘cause’ and ‘effect’ can be individuated
from each other, then they cannot arise dependently. If the relata exist prior to and indepen-
dently of their relation,35 then each must have a distinct identity: “Nowhere does there exist
any such thing as an existent [bhāvaḥ] without defining characteristic”;36 and “There is no
existing thing [bhāvaḥ] that is without svabhāva.”37 But, Nāgārjuna reasons, if each relatum
has its distinct identity independently of any relation to any other, then those distinct individu-
ating natures themselves cannot be dependently arising.38 They were either there all along and
so did not arise at all (the first limb of Nāgārjuna’s tetralemma) or else they pop into existence
apropos of nothing (the fourth limb).39 To make that which distinguishes an existing thing
from everything else – that which makes it this individual – dependent on another, as the
second limb of the tetralemma suggests (and as perhaps the Ābhidharmikas thought they were
doing) is to acknowledge that there is no individual there to be related prior to the relation,
and thus the whole notion of a web of causal relations between discrete individuals falls apart.
Massively distributed causation is not merely not dependent arising; it is antithetical to it,
because it presumes the individuation of the relata, and such individuation requires independ-
ence. Thus, it is significant that the primary assertion of Root Verses I.1 is that “Never in any
way is there any existing thing [bhāvāḥ] that has arisen.”40 No bhāvaḥ arises in any of the four
ways because a ‘being,’ were there any, would be something distinct from other beings, some-
thing independently identifiable as ‘this.’ But there are no such things.
216
Nāgārjuna
The practical objection to the Abhidharma conception of dependent arising as distributed

causation is that it leaves individuation intact, not just as an apt characterisation of reality
but also as a mode of thought, with its attendant implicit values.41 There may not be well-
individuated personal selves on the Abhidharma view; but this is explained – as in the much-
discussed chariot analogy in The Questions of Milinda (Milindapañha), which itself is rooted
in SN 5.10 – by appeal to the constituent elements on which such apparent wholes depend.
These constituent dharmas enjoy the independent identity which persons and chariots lack,
and enjoy greater claim to reality in virtue of that fact. The Abhidharma conception of depend-
ent arising thus dissolves the personal self into countless “selves,” retaining the notion that to
be real is to have well-individuated identity independently of others – so that instead of point-
ing away from the self and the search for the self, the Abhidharma understanding of dependent
arising as a way dharmas associate risks simply attaching us more to the notion that a clear
delineation of independently specifiable individuals-in-relation gets at how things really are.
Nāgārjuna’s Alternative
Nāgārjuna rejects distributed causation as the correct construal of dependent arising, and there-
with the fundamental picture of reality as beings-in-relation. If well-individuated relata cannot
be dependently arising, then conversely what dependently arises cannot be well-individuated.
Thus, in the course of the Root Verses, the point from Chapter 1 is generalised across any
relations and relata,42 with asymmetrical or asynchronous relations modelled on the examina-
tion of causation in Chapter 1 and symmetrical or synchronic relations modelled on the analy-
sis of traversing-traversed in Chapter 2, which claims that it is both necessary and impossible
to distinguish two conceptually dependent things (in this case, traversing and the space tra-
versed, 2.1–5, or motion and mover43); in Chapter 6 on Desire-Desirer, where there is depend-
ence, one cannot have co-occurrence and distinctness together (6.8); Chapter 10 on Fire-Fuel
concludes, upon consideration of the necessary mutual dependency of fire and fuel (10.10–12,
et passim), that “they are not considered by us to be wise instructors in the teachings of the
Buddha who describe the subject and existents in terms of identity and difference” (MMK
10.16). In each case, it turns out to be both necessary and impossible for what is distinct to
be in relation.44 This is a wholesale critique of beings as individuals, and one which can be
extended to practices of individuation with minimal ontological commitment.45
Just as the Buddha’s anatta teaching was not the assertion of an absence where a ‘self’
could have been, so Nāgārjuna’s critique of individuation does not assert the contingent non-
existence of what might have existed. On the contrary, Nāgārjuna insists that it is rather the
very dichotomy between ‘exists’/‘does not exist’ – set up by thinking of discretely identifiable
individuals in terms of identity and difference – which is the problem:
What entity is prior to arising and the rest, what entity is simultaneous, and what
entity comes after — these do not exist; the concepts of existence and nonexistence
no longer apply there.
(MMK 9.12)46
So this is not the view that there is nothing instead of something; only “those of little intel-
lect . . . take there to be existence and non-existence with respect to things” (MMK 8.13).
Rather, the conclusion is that there are no things – with the additional specification that it did
not take much to be an Abhidharma ‘thing’ in the first place. Space, nirvāṇa, and the Four
Noble Truths are, on one standard Abhidharma analysis, all distinguishable items – existing
217
Amber D. Carpenter
things in the sense of knowable dharmas – without being substances (dravya). And Nāgārjuna
criticises these only tenuously thing-like individuals, too, on the same grounds that he criti-
cises all purported beings (space, MMK 5; the Noble Truths, MMK 24; nirvāṇa, MMK 25).
This is thus not merely the rejection of robust Aristotelian substances as bearers of properties
or as primary existents, but of any kind of individuatable ‘this’ at all.47
In place of things-in-relation, there is dependent arising. “Whatever exists in dependence,
that is free of svabhāva” (MMK 7.16).48 Thus, it is incoherent to demand an account of what
arises dependently, for the very conception of dependent arising is the erasure of the model of
individuals-in-relation.49 In a move to appropriate, and perhaps implicitly critique, Mahāyāna
notions of emptiness, Nāgārjuna goes on in Root Verses Chapter 4 to characterise this as the
appropriate sense of that key Mahāyāna term, emptiness (śūnyatā). Drawing on Chapter 1’s
critique of causal relations, Nāgārjuna says that bodily form or rūpa (and likewise, the other
“aggregates,” or skandhas, 4.7) is neither separate from, nor similar to nor dissimilar to its
cause, nor is it without cause. And he concludes by characterising this as a refutation and
explanation based on emptiness (MMK 4.8–9). An argument based on emptiness, that is to say,
is an argument based on the unreality of discrete individuals, as evidenced in the neither-same-
nor-different character of any purported individuals.50
What arises dependently cannot be said to be identical to or different from that on which it
depends (22.6–8); nor, for the same reason, can it be said to exist by means of svabhāva (22.9).
“Thus,” according to Nāgārjuna, “both that on which he depends and the one who is dependent
are altogether empty” (22.10). Emptiness is thus presented as non-individuatedness, or what
follows from it51 – not just the lack of svabhāva in the sense of substantial (as opposed to insub-
stantial, merely apparent) reality, but lack of individuatedness, of the distinctness in virtue of
which things are thought to exist as the identifiable things they are. Where this lack of individu-
ation was the account of dependent arising proposed in place of any causal model in Chapter 1,
it should be no surprise to find emptiness then used explicitly to characterise dependent arising
in Chapter 24: “Dependent arising we declare to be emptiness” (MMK 24.18).52
Nor, given this assimilation, is it any wonder that prapañca is extinguished in emptiness
(MMK 28.5). Where there is no individuation (the emptiness which is ubiquitous is emptiness
of svabhāva), language and concepts – which function, essentially, to discriminate discrete
items as ‘same’ (as itself) and ‘different’ (from not-this) – necessarily lose their function.53
Recognising that dependent arising is absence of any individuation, there is quite simply noth-
ing left for thought and language to do. The stillness of conceptualising promised in the fram-
ing verse is the natural result.
Thus, the declared assimilation of emptiness and dependent arising is immediately fol-
lowed by the observation that emptiness itself “is a dependent concept” (MMK 24.18). ‘Emp-
tiness’ is not a thing or a substance or a quality that pervades reality – nothing enjoys that
status; nor, moreover, is it independently specifiable in some more attenuated way. Emptiness
is a way of conceptualising dependent arising, meaningful within this specific context as a way
of indicating the ubiquitous absence of independent discriminability, or the pervasive non-
individuatedness due to dependency. In a precursor to Diṅnāga’s apoha theory, any apparently
well-individuated concept, ‘emptiness’ included, derives whatever meaning it has through its
association with other concepts, and not by accurately referring to real individuals.54 Perhaps
‘emptiness’ is a particularly useful way of characterising dependent arising; dissolving the
apparent independent specifiability of anything whatsoever, it may well be “taught by the
conquerors as the expedient to get rid of all views” (MMK 13.8).55 But it remains, for all that,
a concept or a way of grasping something, dependent for both its sense and its usefulness on
the very conceptions of individuation and existence which it seeks to undermine. Without a
218
Nāgārjuna
notion of discrete identities, there is nothing for emptiness to be empty of;56 and were there no
habits of individuating and relating discrete items, there would be no good for understanding
emptiness to do.57 As far as ‘the thing itself’ which the word ‘emptiness’ presumably refers
to, this just is dependent arising; if there was ever a place for necessary self-predication, it is
surely in the dependent arising of dependent arising. Bringing us back to the Discourse with
Kassapa, the verse concludes, “just that is the middle path.”58
Most contemporary philosophical scholarship on Nāgārjuna recognises this as articulating
a kind of anti-foundationalism, whether metaphysical, explanatory, or semantic.59 There is
nothing that explains, grounds, gives being to, or validates anything else, which does not also
depend on that very thing for its very identity. In this way, anything that could be picked out
or individuated in any way is equally non-basic.60 This equalising of everything as depend-
ent seems reinforced by the infamous claim at Root Verses 25.19, that “there is no distinc-
tion whatsoever between saṃsāra and nirvāṇa.” Even in the most basic practical distinction
between that from which we move and that towards which we move, their individuation as
such is mutually dependent, and they are no different from each other in this respect.61 Such
anti-foundationalism, however, may be compatible with considering dependent arising as, in
a certain sense, ultimate: dependent arising explains or grounds the possibility that anything
whatsoever should arise, without ever explaining that this rather than that should arise here.62
As Nāgārjuna understands it, dependency necessarily impugns the well-individuated nature
of any purported existent, rendering impossible the individuation necessarily prior to any relat-
ing. This is the true heart of the Buddha’s teaching of dependent arising – not the fragmenta-
tion of being into countless impersonal beings and a frenetic mental hygiene of chasing down
endlessly unfolding relations between them,63 but rather the utter absence of real individuals,
together with a mental hygiene of ceasing to individuate what is not really individual.
Tetralemma, Scepticism, and Speechlessness

There are, then, significant implications of Nāgārjuna’s construal of dependent arising for
language, thought, and the sorts of philosophical argument for which these are the necessary
medium. For all language and all thought is individuating, and this is so even if we think that
meaning arises from use and not by reference to entities – if words do not delineate, they can-
not mean or be used as words at all.64 Likewise, all argument relies on the explanatory priority
of explanans to explanandum. If individuation itself is incoherent, there will be no way to
state the position which does not rely on the very incoherence it diagnoses. Even to say that
dependent arising is ultimate reality – as we must in English – says too much, for it invokes
the existence and identifying-individuating power of ‘is’ which Nāgārjuna eschews. Similarly
regarding “emptiness is lack of svabhāva,” a claim Nāgārjuna immediately follows up with
“It is empty” is not to be said, nor “it is not-empty,” nor that it is both, nor that it is
neither; [“empty”] is said only for the sake of instruction.
(MMK 22.11)
One might use words as a kind of paedagogical gesture, to facilitate the appreciation of reality
as not carved in the way words necessarily do carve: “The ultimate truth is not taught indepen-
dently of customary ways of talking and thinking” (MMK 24.10). But any such gesture must
be immediately undercut by reminders that “one should assert neither real nor non-real”65 in
order to dispel the false impression that the customary ways of speaking are accurate articula-
tions of how things are, or indeed that things (distinct and related to one another) are. It is an
219
Amber D. Carpenter
unavoidable implication of this construal of dependent arising that the via negativa is the only
way, and even here one must be careful that the negations do not imply an assertion of their
opposites: not being must be supplemented by not non-being.
This makes Nāgārjuna’s way of presenting his project an elegant – perhaps even neces-
sary – fit for the particular aims of his project. Since the aim is to present a construal of
dependent arising which dislodges the individuals-in-relation picture, and its associated iden-
tify-and-relate mode of thinking, the manner in which this is presented matters. That manner
is most conspicuously the destructive tetralemma itself, which structures the first chapter of
the Root Verses, and looser but similar constructions in subsequent chapters (see MMK 3.2, 5;
4.1, 12.1, 18.8, 22.17–18). Much scholarly work has accordingly been done on the logic of the
tetralemma;66 more attention still has been given to Candrakīrti’s codification of Nāgārjuna’s
method as exclusively prasaṅga – exclusively making use of others’ views in order to reveal
an unwanted (to the opponent) implication of those views.67 More immediately relevant than
the logic of the tetralemma to Nāgārjuna’s philosophical goals as outlined is the fact that
Nāgārjuna negates each limb – just as the Buddha did in Saṃyutta Nikāya 12.17, consid-
ered previously. And more salient than potential codifications of method is the regularity with
which (apparently) exclusive and opposite pairs are both negated.68
One might worry, however, that this leaves the Mādhyamika in an awkward position. If eve-
rything sayable is inadequate simply in virtue of being sayable, then it looks as if every claim
is equally inadequate. But if every claim in conventional language is equally inadequate, then
it would seem that, conversely, any claim is as valid as any other: the teachings of Gotama the
Nyaiyāyika would be just as (in)valid as the teachings of Siddhartha Gautama, Sage of the Śakyas.
And there is a deeper problem. The very practice of reason-giving is a practice of asserting
relations of priority and dependency among things. (The ‘cause’ discussed in Root Verses I.1 is
hetu, which means ‘reason’ as much as ‘cause’; and the critique counts as much against expla-
nation as against causation.69) And fundamentally, any knowledge claim implicitly asserts the
independence and priority of the means by which one knows. For any knowledge claim, my
knowledge is only as secure as my basis for believing it. The proximate basis is my evi-
dence – thus, the asymmetry implicit in any explanatory enterprise. But any evidence must
be accessed in some way, whether through direct experience or through reasoning. And thus,
any evidence can only be as strong as my mode of access to it is reliable. That is to say, any
knowledge whatsoever depends upon reliable means of cognition, or pramāṇas (if my vision
is faulty, no visual evidence in the world can deliver accurate knowledge of visibles to me). In
this way, a reliable means of cognition underlies the very possibility of valid cognition – and
so is prior to any particular valid cognition, and independent of it. Seeing must be determinate
independently of any particular visual cognition, if any particular visual evidence is to count
for anything. But this sort of independence and priority is just what Nāgārjuna has undermined
with his new interpretation of dependent arising.70 He has, therefore, undermined the very pos-
sibility of having a valid cognition of emptiness or dependent arising – and with this, he seems
to have lost the means for asserting dependent arising, or emptiness, itself.
This seems to be the animating concern of Nāgārjuna’s Dispeller of Disputes. Valid means
of cognition, if they are determinate and distinct from their objects and the knowledge they
give rise to, must be either validated by some yet prior means of cognition (perhaps inference
grounds confidence in perception) or must be self-validating. If Nāgārjuna rejects both of
these – as he must, since the former leads to a vicious regress, while the latter reintroduces
independent foundations the Root Verses argued against – then he is left unable to claim any
validity for his own claims about emptiness, lack of individual being (bhāvāḥ) or svabhāva,
and dependent arising.
220
Nāgārjuna
While Nāgārjuna presents this dilemma as an objection to his proposed interpretation of

the Middle Way in terms of emptiness, this is in fact a pretext for an extension of that inter-
pretation into a trenchant critique of the whole epistemic project. For, after quickly disposing
of the trivial form of the objection based on the misconstrual of ‘emptiness’ as non-existence,
Nāgārjuna rebuts the objection by arguing that any opposing view is as much skewered on
this objection as his own – and indeed, more so. No means of cognition is self-validating,
nor is it possible to have infinitely stacked modes of cognition validating each other – this is
so, Nāgārjuna argues, whatever one takes one’s metaphysics to be. The dependency which
was incompatible with individual identities in the Root Verses likewise infects all the tools
by which one might argue for or against it (or anything else). So, trying to launch an argu-
ment against Νāgārjuna’s Madhyamaka by relying on such tools is misguided – the tools
themselves would be revealed as either only conditionally or capriciously authoritative, thus
demonstrating rather than defeating Nāgārjuna’s point. Knowledge as the objector conceives
it, when he complains that Nāgārjuna cannot lay claim to it, is simply impossible in any case.
Nāgārjuna’s strategy and conclusion here is quite similar to the classic Problem of the Cri-
terion, which Sceptics of Graeco-Roman antiquity deployed to such effect.71 But Nāgārjuna’s
aim is not the same as the sceptic’s.72 For the argument leads to sceptical conclusions only for
one who is committed to a certain conception of knowledge – a conception on which causes
are prior to, independent of, and explanatory of their effects; and a conception on which there-
fore causes of valid cognition, the apparatus for cognising itself, must be prior to and inde-
pendent of the object of cognition. Dispense with this prejudice and the non-independence
of means of cognition does not impugn cognitions at all. One recognises merely that their
credibility is neither freestanding nor indefeasible. This makes honest humility and a standing
openness to reconsideration the best epistemic condition to be in, for seeking certainty (even
relative certainty) is a fool’s errand. But only Nāgārjuna’s Madhyamaka is willing to make this
move, and so only the Mādhyamika is free of the very scepticism of which they stand accused.
Of course, dispensing with the epistemic foundationalist prejudice is more easily said than
done, for if one goes about it via vacua as Nāgārjuna demands, it seems to entail eschewing
the claim to have asserted anything at all. Indeed, Nāgārjuna encourages this understanding
of the implications: “If I had any thesis (pratijñā)” he (in)famously says (Dispeller 29), “that
fault would apply to me. But I do not have any thesis, so there is indeed no fault for me”73 –
leaving uncomfortably ample room for speculation about what it means for whatever we read
in the Root Verses to not be tantamount to expressing a thesis, position, or claim (pratijñā
can mean any of these), and likewise what it meant there to say that emptiness rids one of
all views. An extremely capacious understanding would include any assertion or claim what-
soever about anything – indeed, any articulate speech, thought, or view at all, since we saw
previously that thought and language necessarily individuate where there are no correspond-
ing individuals. On such an understanding of ‘view,’ the Root Verses might be understood as a
work of literary fiction, or at any rate a mere device to induce a certain mental state.74 A more
restricted understanding of ‘view’ construes it specifically as ‘speculative’ views, theoretical
or metaphysical claims, or – as Siderits and Katsura have it – even more specifically with
metaphysical views about the ultimate nature of reality.75 Then the point would be not the
cessation of conceptualising altogether, but rather to cease understanding concepts as even
potentially well-defined terms for referring to a well-delineated reality.
An approach to the question which would integrate it into the preceding discussion of
dependent arising and emptiness might begin by associating ‘view’ with emptiness: some-
thing counts as a ‘view’ just in case it purports to be non-empty, which is to say lays claim to
being independently well-defined or referring to some well-defined individual whose identity
221
Amber D. Carpenter
as that individual does not depend upon the very thing it explains, grounds, or refers to. So
understood, there is an obvious sense in which Nāgārjuna has expressed no views, for he has
already acknowledged the dependency of everything he has said – emptiness itself depends for
its intelligibility and coherence on prior and mistaken notions of individual beings. His own
claims are explicitly empty in just the sense that everything is: arising dependently and not
identifiably distinct from the conditions on which they depend, including the tools for cogni-
tion which are themselves dependent on the cognising and object cognised. It is, then, not a
thesis or a claim in the sense that the interlocutors mean it at all. On the one hand, Nāgārjuna
has been careful not to even try to give a positive specification of his position as distinct from
all other positions, preferring instead to give us reason to stop adopting views and mental prac-
tices that even implicitly presume independent identity. On the other hand, in as much as even
the via negativa relies implicitly on distinguishing thoughts, recognising reasons, and relations
between statements, Nāgārjuna can happily accept that any characterisation of or argument for
his view is dependent – and not just incidentally dependent on prior causal factors, but depend-
ent for whatever apparent identity and distinctness it has on precisely those mistaken notions
of individuality and being they aim to undermine.
But this preliminary understanding of the sense in which Nāgārjuna has no views – and
how having no views might enable him to escape the self-refutation objection while redeem-
ing the meaningfulness of the Root Verses – only goes so far. If Nāgārjuna gets the better
of his opponents, it is only in as much as he recognises the inevitable emptiness of his own
claims, their non-absoluteness or lack of ultimate grounding in some self-sufficient mode
of cognition, while his opponents lack awareness that their own views are in the same posi-
tion. But dialectically, this should only put all contending views on all fours: all views arise
dependently, lack coherent and distinct content, and rely for their intelligibility on similarly
dependently arising factors. If this is where we are left, we might think that Madhyamaka
avoids logical self-contradiction only at the drastic cost of a kind of practical self-abnega-
tion: if the Madhyamaka view is correct, that very fact eliminates the point of saying so and
the possibility of showing it.
Two distinct worries may be distinguished here, the first to do with everyday practice and
the second with insight and the ultimate Buddhist goal. If thought and language can never get
it right about how things really are, and if everyday speech is not even trying to make claims
about how things really are, then everyday language can never be criticised for having got it
wrong. Everyday language is an effective instrument for achieving everyday ends, and so long
as it succeeds in that objective, there is nothing going wrong and nothing in need of correcting.
This, however, suggests that there is no space for a reforming Buddhist outlook on practical
everyday affairs – we must simply leave everything in its place.76 This would leave even the
Buddha’s ambiguous talk of the self, so at odds with everyday notions, as itself a bit of idle
chatter without practical relevance. The second concern is that if the means of substantiat-
ing arguments is itself undermined, then there is no adjudicating between views – not even
between the view that we should abandon all views, and the view that we ought to aim, as the
Eightfold Path suggests, for Right View.
In short, it seems the Mādhyamika is saved from speechlessness only by destroying the
inherently teleological character of the Buddhist path. For if ordinary speech generally is fine
just as it is, so long as it is intended and taken pragmatically – as indeed it ordinarily is –
and without implying some conviction about the really real basis underwriting it, then what
needs reforming? If “there is no distinction whatsoever between saṃsāra and nirvāṇa” (MMK
25.19), where is there to go? And if all language that does try to say anything about ‘how
things really are’ is equally falsifying, this seems to render Buddhist insight utterly mute.77
222
Nāgārjuna
Practical Implications, Proximate and Ultimate

We might see our way out of this by turning to the text in which Nāgārjuna seems most explic-
itly concerned with just this issue.78 The Precious Garland is presented as ‘advice for a prince’
– so it is explicitly concerned with everyday life and how that might be lived better. What there
is to be said on this score, however, does not come neatly disentangled from ultimate concerns
of a specifically Madhyamaka kind, nor is it disconnected from the ultimate aim of escaping
worldly life altogether.
The text sets up two distinct ends – happiness and the highest good — with virtuous prac-
tices and wisdom as their respective means (RĀ 3–4).79 This may seem at first to reinforce
the idea that the Root Verses’ view of dependent arising as lack of individuation leaves us
with two disconnected alternatives: either we fully embrace its implications for language and
thought, and thus remain mute and disengaged (literal quietism); or we accept the catego-
ries of the world as they stand and aim for worldly goods as these are understood (practical
quietism).80 Even if one could alternate between the two, one could never use the former to
reform the latter, nor the latter to approach and articulate the former. This disjunctive quiet-
ism would seem to leave the practical commitments of the committed bodhisattva impotent.
So it is essential to see how the text gives us more than that and, in doing so, addresses how
it could be that “the ultimate truth is not taught independently of customary ways of talking
and thinking” (MMK 24.10) – indeed, how it could be that Nāgārjuna follows his extended
critique and re-conception of dependent arising in the Root Verses with a presentation of the
twelve-fold chain of dependent arising (MMK 26) that is, as Katsura and Siderits (307) put it,
“the perfect model of orthodoxy.”81
The Precious Garland articulates two crucial ways in which the two ends (worldly well-
being and ultimate attainment) and their respective paths converge. First, on a practical level,
we are assured that aiming for happiness will indeed eventually issue in attaining the highest
good.82 The text then bears this out in its extended discussion of sense-pleasures, which takes
the main constituent of happiness as ordinarily conceived and uses the prince’s interest in it to
motivate a concern not only for virtue but for wisdom and liberation. Note, for instance, how
what starts as a familiar Abhidharma exercise on the foulness of the body as a means to induc-
ing disenchantment with sensual pleasures (RĀ 148–70) – essential for virtuous practice –
leads to an orientation towards “unparalleled enlightenment” rooted in “wisdom not relying on
duality” (RĀ 174); and later how the analysis of pleasure at RĀ 348–63 relies on specifically
Madhyamaka positions and techniques concerning emptiness and insubstantiality. Indeed, it
turns out that one of the four main virtuous practices for a happy life is none other than wisdom
(RĀ 139), so that to aim at happiness turns out, as promised, to lead one to enlightenment. The
higher goal is actually built into the path to the lesser goal.
This convergence of higher and lower goals grounds a second convergence, which oper-
ates in the opposite direction. For now, leading to wisdom operates as a constraint on what
could count as virtuous practice. The fact that a certain ‘ordinary outlook’ or way of concep-
tualising leads – or does not lead – to wisdom as the Mādhyamika understands it is a reason
for considering such an outlook as either apt (134–35) or in need of reform. Thus, we have
a basis for critique of ordinary practice, after all, and indeed a basis in ultimate truth and not
just in the several prescriptions found in scripture – though not in the way the Abhdidharma
might have suggested. Common ways of speaking and thinking are indeed riddled with
error, not because they get at the wrong objects, but because thinking in that way causes suf-
fering. Given dependent arising/non-individuation/emptiness, no distinctions and ways of
individuating are correct characterisations of how things really are, and wisdom will consist
223
Amber D. Carpenter
in recognising this. Ordinary modes of thought can be criticised if and when they block this
recognition.
Thus, those conventional uses of language are apt which lead one towards recognition of
ubiquitous dependency and its concomitant insubstantiality – and this is indeed what the Pre-
cious Garland does with the prince, a paradigm worldly figure, in verses 125–75, and elsewhere.
A perfectly orthodox exposition of the twelve links of dependent arising may do the same.
Uncritical participation in the language of individual persons (‘I,’ ‘mine,’ and associated
desires: ‘I want this to be mine’) probably will not. A thorough examination of the relation-
ship between individuation, causation, existence, and reality – such as the Root Verses itself –
would be a use of language which facilitates this, as would thinking of emptiness as the lack of
substantial individuation necessarily implied by dependency. Going along with whatever the
world happens to say about individuation and causation may or may not be fruitful, depending
on the circumstances. Perhaps there is no immediate harm in using the language of persons for
everyday transactions, but it can be difficult to do so while simultaneously maintaining the per-
spective that recognises the purposes of these transactions themselves are most often riddled
through with confusions about self and individuation that cause suffering.83 And so it may well
be appropriate for even ordinary folk with worldly commitments – the prince, for instance – to
engage in active critique of ordinary language and concepts (RĀ 37–73). The Mādhyamika is
not restricted to using all ordinary language with only its ordinary meanings without critique,
for critique may arise from the conduciveness of practice to happiness and final liberation. Nor
then, conversely, is the Mādhyamika restricted to using ordinary language for only its ordinary
non-philosophical purposes; reasons can be given and reasoning efficaciously engaged in,
without indulging in the self-refuting presumption that one’s reasons and reasoning enjoy an
implausible epistemic or ontological independence.
None of this requires that philosophical reasoning be a necessary and exclusive means
to ultimate realisation of the Buddha’s teaching of dependent arising. There is no reason to
restrict the effect of language on the mind in that way, and every reason to suppose that non-
argumentative use of language may also have salutary effects on one’s phenomenology, state,
or outlook – indeed, nothing speaks against reading the Root Verses itself in such a literary
way, if that accomplishes the same end.84 It is noteworthy in this regard that the four hymns
most centrally associated with Nāgārjuna engage explicitly in reasoning to achieve their aim,85
as well as in rehearsing terms of approbation in a more poetic way.86 We need not suppose
some strict distinction between philosophical engagement on classic philosophical topics such
as cause, relation, individuation, essence, and existence, on the one hand, and ‘merely’ skillful
teaching on the other.87 On the contrary, these might often be one and the same, as is demon-
strated not just in the Root Verses, but also in the Precious Garland.
But if one does engage philosophically with the Root Verses (and other analytical texts,
such as the Dispeller of Disputes), whether negatively or constructively,88 what one can no
longer do is naïvely suppose that such engagement is a search for some lynchpin, whether that
be one thing or many dharmas. Nor may one engage in practices of individuating, identifying,
relating, and classifying – making Nāgārjuna’s work into a worldview – as if such a mental
exercise were essentially different from the more modest individuating and relating we do in
everyday life. The Root Verses, however, has this advantage over ordinary thought: it carries
its own antidote within it. By enabling us to recognise the individuating we do while we are
doing it – making, indeed, such recognition inescapable – the Root Verses ensures that we are
no longer implicitly convinced of the superficial unimpeachability of our thinking. To engage
in thinking through the Root Verses is to become aware of the dependency of that very activ-
ity – and it may possibly even suggest an alternative.89
224
Nāgārjuna
Conclusion
To sum up: Nāgārjuna is a fundamentalist reformer, taking us back to what he asserts is the
core of the Buddha’s teaching – viz., dependent arising – and the correct understanding of that.
His critique of the Abhidharma aims to transform how we think of dependency and therewith
the distinction between ultimate and conventional, and of the Buddhist goal and the path to it.
As Nāgārjuna tells it, the Ābhidharmikas are working with a view according to which exist-
ing things dependently arise. Dependently arising characterises the behaviour or relations of
things with respect to one another; but it can only do so on the presupposition of individuation;
that is, of a prior distinction between two things (elements of experience, existing things) sub-
sequently related. When seen in this light, it becomes clear that this Abhidharma understand-
ing of dependent arising does not, after all, do away with a picture of reality that fundamentally
sees individuals as acting on and being affected by each other; it only complicates that picture.
It depersonalises the picture, to be sure, but the basic elements into which we analyse experi-
ence or existent things retain the characteristics of persons – having a distinct identifying
nature which makes them what they are by distinguishing them from everything else.
Such a model, Nāgārjuna aims to show, positively precludes dependency and eo ipso depend-
ent arising. For dependency is incompatible with having a discrete identity and therefore incom-
patible not only with causation, but with any relating that presumes (as it commonly does) the
distinctness and identifiability of the relata. If causation is especially relevant, it is because (i) it
is the terms in which his Buddhist opponents understand dependent arising, so it is the view of
dependent arising he has particularly to displace in his project to recover the Buddha’s original
teaching by the middle; and (ii) causation can be used as a model for any asymmetrical relation,
and in particular for epistemological dependency relations inherent in explanation.
Nāgārjuna’s larger project in the Root Verses is to defend a replacement conception of depend-
ent arising ( pratītyasamutpāda instead of existence/non-existence) against the Ābhidharmika
supplementary conception of dependent arising ( pratītyasamutpāda as a way in which things
arise). Instead of conditioned relations between discrete individuals, instead of the being/non-
being, same/different dichotomy this implies: dependent arising. Eschewing individuation itself
(and not just an ontological commitment, since it is as much the mental habit of individuating that
generates suffering as any additional existence claim on top of that); recognising the imperfect
individuation of things, their imperfect sameness and difference from others and themselves; and
recognising this incomplete differentiation as due to dependency – just this is the Middle Way.
This leaves much still up for debate. For instance: does relating – even the conventional,
epistemological, and non-ontological activity of relating – indeed presume the prior individu-
ation of its relata? Are existence and essence so mutually bound up that rejecting the latter
implies losing also the former? Is Nāgārjuna’s alternative proposal even coherent – or has it
not, perhaps even wittingly, removed its own conditions for the possibility of being coherent?
And in spite of the suggestions offered in this chapter, there is still ample room for debating the
practical implications of this: is the implication a kind of recommendation to extreme quietism,
a re-conception of the ultimate prajñā to be obtained as simply not thinking anything at all, and
a rejection of critical engagement with everyday notions? Is Nāgārjuna a sceptic in the classical
mould, using critical thinking to defeat any ambitions of critical thought, for the purposes of a
disengaged equanimity? Or if we should stay engaged, how does the Madhyamaka view sup-
port or revise that project? How will correct understanding of pratītyasamutpāda – a distinc-
tively Madhyamaka, non-Abhidharma understanding – make one a better leader in the world?
It is indeed fitting that such questions should remain open, for the power and inescapability
of Nāgārjuna in Buddhist philosophy consist in how much new terrain his thought opened up
225
Amber D. Carpenter
and the standard he set for the degree of penetration, subtlety, and systematicity to be brought
to the task.
Notes
1 I am very grateful to Paul Livingston for insightful and thoughtful comments on an earlier draft, and
to Jay L. Garfield and William Edelglass for their comments on a later draft. Graham Priest’s com-
ments on the penultimate version helped me to clarify important points. Enormous gratitude goes also
to Nicholas Lua for exceptional research assistance. And finally, I thank the editors of the volume –
particularly William Edelglass, by whom to be edited is a joy – for their invitation, for their patience,
and for constructive support in bringing this to fruition.
2 Ruegg (1981) is still unsurpassed for a scholarly philosophical overview of the history and literature
of Madhyamaka thought in India.
3 Joseph Walser (2005, 61–88), argues for the dates and place, for which there is cautious scholarly
support; Walser argues more specifically that Nāgārjuna “wrote the Ratnāvalī within a thirty-year
period at the end of the second century in the Andhra region around Dhānyakaṭaka (modern-day
Amaravati)” (Walser 2005, 61).
4 Westerhoff (2010) offers a lucid translation of, cogent commentary on, and introduction to the Dis-
peller of Disputes.
5 Komito (1987) translates just the verses from Tibetan into English, as does Tola and Dragonetti
(1987), which includes the Tibetan, as well. Lindtner (1982, 34–65) translates verses with auto-com-
mentary from Tibetan into English. Erb (1997) offers a partial translation of Candrakīrti’s commen-
tary into German.
6 Lindtner (1982, 102–19) offers a translation of the Sixty Verses, as does Tola and Dragonetti (1995);
Loizzo (2007) includes a translation of Candrakīrti’s commentary along with the verses.
7 Westerhoff (2018a) offers a translation of Crushing the Categories, tackling in the introduction the
question of attribution. Tola and Dragonetti (1995) also offers a translation into English.
8 Translation of the Precious Garland from the Sanskrit can be found in somewhat old-fashioned Eng-
lish in Tucci (1934; 1936). Hopkins (1998) provides a more modern English translation, from the
Tibetan.
9 Pind (2001) and Tola and Dragonetti (1995) have challenged the attribution of Crushing the Catego-
ries, for instance, while in the opposite direction, Lindtner (1982) has a more expansive (and there-
fore more contentious) list of authentically Nāgārjuna texts, including in addition to the previous:
the Letter to a Good Friend, Proof of Convention (Vyavahārasiddhi), Four Hymns (Catuḥstava), and
Verses on the Heart of Dependent Arising (Pratītyasamutpādahṝdayakārika), among others.
10 Nāgārjuna need not think that his Middle Way is distinct from other Mahāyāna interpretations of the
Buddha-dharma in all respects in order to think there are important enough differences.
11 We are, incidentally, similarly situated regarding Plato’s dialogues and the letters traditionally attrib-
uted to him. See Burnyeat and Frede (2015) for recent discussion.
12 See Arnold (2005, 6, 117; 2010; 2012); also see Westerhoff (2009a) – an excellent resource for getting
to grips with Nāgārjuna’s Madhyamaka philosophically – and Shulman (2011). Thomas Wood (1994)
proposes a metaphysical nihilist interpretation; see also Westerhoff (2016). Oetke (2007) proposes
metaphysical illusionism as a viable alternative (though Oetke [2011] rejects metaphysical readings).
Ferraro (2013; 2014) proposes ‘realist antimetaphysics,’ which is as much metaphysical as Kant’s
critical philosophy may be. Priest and Garfield (2002) offer a logical reading of the Root Verses, more
metaphysical than epistemological in its claim that Nāgārjuna endorses a paraconsistent logic that
sanctions true contradictions, though it may also be given a semantic construal, as Garfield endorses
in Siderits and Garfield (2013).
13 Ruegg (1977; 1983); Garfield (1996); Siderits (2003b; 2007) (semantic non-dualist).
14 Arnold (2010) focuses on explanation, while Ferraro (2017) is epistemological in a rather different, anti-
metaphysical way. Matilal (1986, Ch. 2) and Mills (2018) present Nāgārjuna as a skeptic, as does Gar-
field (1994) (explored further in relation to later Madhyamaka in Dreyfus and Garfield [2010; 2021]).
15 Huntington (2018, 156) argues that
to insist that Nāgārjuna’s sole or overriding concern is to make rational sense is to forget that
looking for meaning in these terms is – at least in principle – only one possible methodological
approach to interpretation of his writing.
226
Nāgārjuna
Huntington characterizes the Root Verses as a literary work of art aimed at inducing “a recognition
of the limits of reason and a willingness to surrender to the groundless space below or beyond those
limits, which are as well the limits of memory and imagination, of desire and fear” (Huntington 2007,
125). See also de la Vallée Poussin (1933, 59), and more recently Stepien (2019) for an anti-argument
reading of Nāgārjuna and of the Root Verses. Such an approach has less difficulty explaining the prac-
tical implications of the Root Verses, but while it may arise from and lead to provocative philosophi-
cal reflections, it does entail removing the content of the text from the sort of philosophical discourse
in which Bhāviveka and Candrakīrti, for instance, placed it – a discourse in which, as Tillemans
(2017, 111) puts it, “the particularity of the Madhyamaka is that it emphasizes philosophical analysis
as a method of leading to the quietening of thought.”
16 The Perfection of Wisdom in Eight-Thousand Lines (Aṣṭasāhasrikā Prajñāpāramitā Sūtra), which
by some accounts was composed by – or at least prevalent among – the Mahāsaṃghikas of the
Andhra region (where Nāgārjuna may also have hailed from), may be especially relevant; accord-
ing to P. L. Vaidya (1960, xvii) it was the exclusive prajñapāramitā text to which Indian Buddhist
philosophers attended.
17 This is a cause for some of the difficulty in determining just how distinctive and exclusive Mahāyāna
and non-Mahāyāna Buddhist ideas were. As Walser (2005) observes, under these conditions of com-
munal living, Mahayanists had to present themselves as not so very different from their brethren if
they hoped to have resources devoted to the copying and preservation of their texts.
18 As Westerhoff claims, the Root Verses is “not primarily conceived as a treatise with a specific sectar-
ian orientation, but as a fundamental Mahāyāna text” (Westerhoff 2018b, 121).
19 Hayes (1994) argues that transparent equivocation explains the delayed effect of Nāgārjuna on the
development of Buddhist philosophical thought.
20 Garfield and Westerhoff (2015) largely explores the rivalry between these two Mahāyāna views, but
Westerhoff’s and Shulman’s contributions suggest that the hostility may not be there in Nāgārjuna.
There were also important Mahāyāna syncretists.
21 This is so for at least the bulk of the Anglophone scholarship. A notable exception to this is Kalupa-
hana (1986); Shulman (2008) is a more recent instance, while Shulman (2010) identifies the tendency
to assimilate Nāgārjuna and Candrakīrti in the contemporary Anglophone scholarship on Madhy-
amaka, and gives specific reason why we should be cautious about taking Candrakīrti as our authority
in understanding Nāgārjuna.
22 Diṅnāga’s Compendium of Pramāṇas (Pramāṇasamuccaya), for instance, opens by praising the Bud-
dha as “the personification of the pramāṇas.” Vasubandhu’s Treasury of Abhidharma, with Com-
mentary (Abhidharmakośabhāṣya), largely concerned with prajñā as the “discernment of dharmas”
(AKBh. I.2a), begins by honouring the Buddha as the “teacher of the truth” who “has in an absolute
manner destroyed all blindness,” by which, the commentary notes, “the Blessed One is sufficiently
designated” (AKBh. I.1). In each case, the Buddha is praised for having those features which are most
valued by and essential to the ideas that follow.
23 The translation is that of Siderits and Katsura, leaving out material they supply in square brackets,
using ‘(dependent) arising’ for their ‘(dependent) origination,’ and leaving untranslated ‘prapañca,’
which they render as ‘hypostatisation.’
24 Notice that both Vasubandhu and Diṅnāga, by contrast, connect their selected qualities of the Buddha
to his beneficence: “he has drawn the world out of the mire of transmigration” (AKBh. I.1); he “seeks
the benefit of [all] living beings” (PS Ι.1.1). They praise other qualities of the Buddha, as well, in
particular his beneficence, while Nāgārjuna selects exclusively the Buddha’s teaching of dependent
arising for praise.
25 Nor is there reference to svabhāva (translated, variously, as ‘own-nature,’ or ‘self-nature,’ ‘intrinsic
nature,’ ‘essence,’ and even ‘own-being’), as that which is lacked, despite the centrality of this to dis-
cussions of Madhyamaka; indeed, svabhāva is somewhat less central to the Root Verses generally than
its prevalence in Madhyamaka discourse would suggest. Even Chapter 15, which Candrakīrti presents
as an analysis of svabhāva, may rather be, as Bhāviveka’s title suggests, “An Analysis of the Existent
and Nonexistent.” See Shulman (2007, 148–499) for a sustained argument that this term of art indeed
holds less central significance for Nāgārjuna than most contemporary interpretations suggest.
26 In the Sanskrit, this is the first half and not, as in the English translation, the middle portion.
27 MN 38; cf. SN 12.1 (PTS ii.1)
28 See Arnold (2005, 169) for a similar understanding that in the Root Verses, “pratītyasamutpāda
must be associated with a novel, not commonly accepted import.” Oetke (2007) concludes that “the
227
Amber D. Carpenter
linguistic meaning” of pratītyasamutpāda “and its notional significance is preserved whereas its con-
ceptual import or its intension is completely novel” (Oetke 2007, 27).
29 Alex Wayman argues that “Nāgārjuna, in the matter of the catuṣkoti, is heir to and continuator of
teachings in the early Buddhist canon,” the catuṣkoti (tetralemma) in its distinct uses being “found
in early Buddhism and later in the Mādhyamika school,” and “well represented in passages of early
Buddhism, as preserved in the Pāli canon” (Wayman 1977, 10–11).
30 Translated by Bhikkhu Bodhi in Connected Discourses of the Buddha (Wisdom Publications).
31 That the focus is on interpreting dependent arising is underscored by the MMK’s only explicit refer-
ence to a named discourse: “The Instructing of Katyāyana” (MMK 15.7a) refers to SN 12.15 (the
Kaccāyanagotta Sutta), the core of which again has the Buddha not “veering towards either of these
extremes,” here between existence and non-existence, and again teaching “the Dhamma by the mid-
dle: ‘With ignorance as condition, volitional formations [come to be]; with volitional formations as
condition, consciousness’ ” (trans. Bodhi 2000). Kalupahana (1992, 161) calls the Root Verses as a
whole “a grand commentary on the Buddha’s discourse to Kaccāyana” (viz., SN 12.15); but the Naked
Ascetic Kassapa Sutta (SN 12.17), which includes the “Doctrine of the Middle” passage of SN 12.15,
prefigures also the tetralemmic approach to expressing it.
32 Ronkin (2005) does some of the careful work necessary for articulating the interpretive moves
made from Nikāya to Abhidharma Buddhism. Chapter 5 in particular details the transition from
the Nikāya view of dependent arising to the early Abhidharma interpretation of it in terms of
causation.
33 This need not imply a well-worked-out theory of causation as such; contrast Kalupahana (1975) with
Ronkin (2005) on whether there is or need be a theory of causation implicit in the analysis of condi-
tions. Shulman (2008) argues for the priority of mental conditioning in the early Buddhist notion
of dependent arising (rejecting pratītyasamutpāda as a “general principle of causality,” 315), while
Meyers (2018) argues that ‘dependent arising’ and ‘causation’ are false friends with only superficial
and misleading similarity, but is concerned specifically with what modern notions of causation infe-
licitously import.
34 Gethin 1992 and 2004 both focus on the practical aspect of Abhidharma classifications. Heim (this
volume) offers an anti-metaphysical phenomenological reading of texts from the Pāli Abhidhamma.
The advantage of the understanding of Nāgārjuna’s critique which I unfold below is that its target
might include not only such an ontologically reticent version of the Abhidharma, but also Abhi
dharma views at any point along the trajectory from dhamma as classificatory category to dhamma
as ultimately existing entity (see Cox 2004; Ronkin 2005, Ch. 2.). Whether it would so readily tell
against the sort of ‘philosophical perception’ and meditative phenomenon that Shulman 2014, Ch. 1,
attributes to early, sutta Buddhism is less evident. Indeed, one reading of Nāgārjuna may be as argu-
ing that just this is the original Buddha-dharma that the Mahāyāna returns us to, after Abhidharma
deviations and codifications.
35 Bastow (1994, 496) calls this “a principle which in general seems to be noncontroversial: that the
existence of a relation demands the existence of its terms” and cites Saṃghabhadra as explicitly
recognising it.
36 MMK 5.2 uses lakṣaṇa for defining characteristic, accentuating that it is distinct identity which is at
stake.
37 MMK 13.3. See note 25 on translating svabhāva. This is, as Siderits and Katsura render it, said in
the opponent’s voice, but it thereby confirms that this is the conception at issue (see Shulman 2007,
150n29 for discussion of this point). This translation drops bracketed texts added by Siderits and Kat-
sura, and retains “existing thing” for bhāvaḥ, to underline the connection between this and MMK I.1:
The very same thing which “does not arise in any way” is something which necessarily has a proper,
individuating nature.
38 Ronkin (2005, Ch. 4) presents Abhidharma dharma theory as concerned centrally with individu-
ation; and this includes reading the associated notion of svabhāva as this arises from the early
Abhidharma (Ronkin 2005, 112–22) as that which identifies a discrete individual as what it is, and
thus individuates it.
39 All of this can be read in causal terms, or in more broadly explanatory terms.
40 Compare MMK 12.1, “Some say that suffering is self-made, some that it is made by another, some
that it is made by both, and some that it is without cause; but it is not correct to think of suffering as
an effect” (emphasis added).
41 See Carpenter (2014).
228
Nāgārjuna
42 The focus on relata and relations is even clearer in the titles given these chapters by Buddhapālita
and Bhāviveka than those more commonly used in English translations, given by Candrakīrti. For
instance, where Candrakīrti has, “What is Prior” (Chapter 10), Bhāviveka offers “Appropriator and
Appropriated”; Chapter 18 on “Self ” is, for Bhāviveka, “On Self and Things”; for Chapter 20, instead
of Candrakīrti’s “Assemblage,” Bhāviveka offers “Assemblage and Causal Factors.”
43 MMK 2.18–20 reads:
It is not right to say that the goer is identical with the act of going; nor again is it right to say that
goer and act of going are distinct. If act of going and the goer were identical; then it would also
follow that agent and action are one. If, on the other hand, the goer were thought to be distinct
from the act of going, then there would be the act of going without a goer, and a goer without an
act of going.
The analysis of sense-fields (āyatanas, Ch. 3), for instance, is modelled on this. See also the “Hymn to
Him Who Has Gone Beyond,” 11; “Hymn to the Unthinkable One,” 11. Shulman (2010) argues that
even the first half of this chapter concerns the relation between agent and activity, rather than between
locus and activity as Candrakīrti’s interpretation, and the Siderits and Katsura translation, suggest.
44 If Bastow (1994, 494) is correct and the Sarvāstivādins already defined their dharmas by their rela-
tions to each other (“the very being of a dharma is interpenetrated by those other dharmas which are
causally responsible for its arising”), then Nāgārjuna is here either bringing to the surface an incoher-
ence with their explicit commitments, or else pushing those commitments to their logical conclusions.
45 This is important if Heim is right in interpreting Pāli Abhidhamma, of which Nāgārjuna should cer-
tainly have been aware, as primarily doing phenomenology, without making claims about ontology
or the nature of existing things (see Heim, Chapter 9 in this volume).
46 Chapter 8, on object and agent, particularly stresses the inadequacy of the real–unreal dichotomy;
“intrinsic nature and extrinsic nature, existent and nonexistent – who see these do not see the truth of
the Buddha’s teachings” (MMK 15.6). See also 4.4, 5.7 (and 5.2 quoted in what follows). MMK 15.6
is immediately followed by a reference to SN 2.17.
47 Hence, the sense to be found in calling Nāgārjuna a nihilist, in a specific metaphysical sense. It is
not that he asserts there is nothing, but he does conclude there are no things. See Westerhoff (2016)
for a careful and incisive discussion of the various senses of nihilism, those that Nāgārjuna would
be keen to avoid, those with which his opponents (then and now) might reasonably charge him, and
the specific sort of nihilism which avoids those objections, and is distinct from the nihilism decried
ubiquitously in Buddhism as an ‘extreme view.’ This latter sense, which Westerhoff attributes to
Nāgārjuna, is rather similar to the view set out here, but Westerhoff underscores the ‘substantiality’
supposedly implied by svabhāva, while this account focuses on its individuating character.
48 Compare RĀ 38–62, which associates correct understanding of dependent arising with rejecting the
existent/non-existent dichotomy (v. 38, 42cd, 48–49, 56).
49 Similarly, to suppose that after the critique of Root Verses I Nāgārjuna owes some alternative account
of causation, some explanation of how things do in fact arise, is to miss the main point of the critique.
Causation is critiqued to clear space for an alternative understanding of dependent arising. If the
result of the critique is an understanding that dependency impugns individuation, then to even sup-
pose there might be an account of causation is to be in the grip of an illusion.
50 Emptiness as an alternative to existence (and non-existence), understood as having a distinct identity,
is worked out in Chapter 13 (on suffering) and Chapter 15 (which Candrakīrti considers an analysis
of essence (svabhāva), but which Bhāviveka considers “Analysis of the Existent and Nonexistent”);
from here, as talk of svabhāva diminishes somewhat, references to emptiness (śūnyatā) – previously
sparse – accelerate. Chapters 13 and 15 particularly look like they may be Nāgārjuna’s attempt to
correct mistaken notions of emptiness arising among the Mahāyānists.
51 “Dependent arising is emptiness” ’ in 24.18 is immediately followed up in 24.19 with, “There being
no dharma whatsoever that is not dependently originated, it follows that there is also no dharma what-
soever that is non-empty” (MMK 24.19). That is, because dependently arising, therefore emptiness.
This might be understood as the reason why dharmas are empty (they are empty because depend-
ently arising); and it might equally be articulating how we are to understand that pervasive emptiness
(dharmas are empty in the sense that they depend and are not independent in any respect whatsoever,
including in respect of their distinct, individuating identity).
52 Compare, from the “Hymn to Him Who Has Gone Beyond the World,” 22: “Dependent Origination
has been considered by you to be just emptiness. ‘There is not an independent being’: (this is) your
229
Amber D. Carpenter
incomparable lion’s roar” (in Tola and Dragonetti 1985). All quotations from the hymns are taken
from this translation.
53 From the “Hymn to the Unthinkable One,” 40 (minor adaptations from Tola and Dragonetti’s
translation):
What is neither one nor not one, neither both nor not both, without base and not manifest, unthink-
able; what cannot be pointed out, what neither arises nor disappears, is neither liable to destruction
nor eternal – that, similar to space, is not within the range of words and knowledge.
54 If one asks about Emptiness, the thing itself, the answer is that there is no thing, Emptiness, just as
there are generally no things, as discrete individuals with their own definitive identity. But one can
use the circle of related concepts through which it gets its meaning to better grasp the alternative to
things-in-relation, viz. dependent arising.
55 There is considerable controversy over how exactly to understand the dṛṣṭi (views) which emptiness
should get rid of. See discussion following.
56 And likewise, wherever one wishes to indicate the specific emptiness of any particular thing, this nec-
essarily depends on first individuating it: the emptiness of the chariot or of the rūpa dharma derives
any specificity it has dependently upon the reference to the chariot or the rūpa dharma.
57 This is a rather deflationary take on the ‘emptiness of emptiness,’ which has been the subject of a great
deal of careful and ingenious debate in the contemporary Anglophone literature. Garfield (1994) argues
that MMK 24.18 “provide(s) the fulcrum for Candrakīrti’s more explicit characterization of the empti-
ness of emptiness as an interpretation of Nāgārjuna’s philosophical system – the interpretation that is
definitive of the Prāsaṅgika-Mādhyamika school” and offers an interpretation of MMK I that agrees
in making “the doctrine of the emptiness of emptiness” the heart of Nāgārjuna’s view (Garfield 1994,
221–22). Siderits (2003b)’s semantic reading also focuses on the emptiness of emptiness, arguing
that the upshot is “that the ultimate truth is that there is no ultimate truth – there is only conventional
truth” (Siderits 2003b, 11). While Siderits’ view avoids paradox through disambiguation, Priest and
Garfield (2002) and Garfield and Priest (forthcoming) embrace the paradox. Such a reading, however,
risks being lopsided – recoiling from the fact that words do not correspond to reality to conclude that
“there is only conventional truth” (Siderits 2003b, 11) looks very much like the semantic equivalent of
discovering nothing exists as an identifiable individual and concluding non-existence. As Nāgārjuna
reminds us, the Buddha taught two truths/realities (MMK 24.8), worldly (lokasaṃvṛtisatyaṃ) and
ultimate ( paramārthataḥ) truth/reality, the distinction between which must be understood (and fail-
ure to grasp the latter prevents one attaining nirvāṇa, 24.10). If worldly truth (lokasaṃvṛtisatyaṃ) is
saṃsāra, then this reading of the emptiness of emptiness adopts the extreme position that the claim
“there is no distinction whatsoever between saṃsāra and nirvāṇa” (MMK 25.19) is tantamount to the
positive assertion that there is only saṃsāra. We should be wary of following Candrakīrti’s lead here
and taking the emptiness of emptiness as the core of the Root Verses, particularly when alternatives
exist. Nāgārjuna might rather be understood, for instance, as reining in the enthusiasm of his Mahāyāna
brethren, by advocating that ubiquitous emptiness be understood merely as a way of designating the
essenceless non-individuatedness of reality. See note 51 on 24.19. The opening thirteen verses of the
Seventy Verses on Emptiness could be read as arguing in the same way, namely “Because dependently
arising, therefore without distinct nature, and that just is to say ‘empty.’ ”
58 Oetke (2007) offers patient analysis of subtly distinguished interpretations of MMK 24.18, including
the reading, not pursued here, which takes the verse to identify ‘conceptual designation,’ and not just
dependent arising, with emptiness.
59 Presumably that is the force of the ‘non-dualism’ in Siderits 2007’s ‘semantic non-dualism.’
60 It is not to be taken for granted here that this ‘equally’ is warranted; everything could be dependent,
yet some things might be more basic or fundamental than others. Ubiquitous dependency does not
itself entail equality of all dependent entities.
61 This difficult verse seems to claim further that nirvāṇa is not a distinct state or mode of being from
saṃsāra, and this is hard to make good sense of. Nirvāṇa is frequently understood as cessation, and
cessation looks very much like the ‘stillness’ that should arise upon the correct grasp of dependent
arising, in the frame verse, but will not otherwise arise. If nirvāṇa is the stilling of prapañca and
saṃsāra the proliferation of concepts (non-stilling of prapañca), then there does seem to be a great
difference between the two, and the purpose of the Root Verses is to actualise the change from one to
the other. If, however, nirvāṇa is not the stillness recommended in the frame verse, then it is also no
longer the goal of the Buddhist path, but something quite different he is talking about. It is tempting
230
Nāgārjuna
to consider here whether Nāgārjuna might be overstating his case for a certain literary-soteriological
effect at 25.19 – and in general, this text deserves much closer consideration of how the literary ele-
ments function argumentatively, as seen for instance in Plato’s dialogues or the Introduction to the
Bodhisattva Way (Bodhicaryāvatāra).
62 In this respect, it is rather like Plato’s chora in the Timaeus. In just the way Plato would take this to
be an incomplete and inadequate explanation of reality and what exists, pratītyasamutpāda could be
understood as non-foundational, for it can never adequately explain or ground any particular being.
(My thanks to Alexis Pinchard, whose work on Nāgārjuna and Plato suggested the association.)
63 Gethin (1992) brings out the open-ended fecundity of the Abhidharma matrices as a salutary practice.
See also Heim (2014, 86): “The Abhidhamma is engaged simultaneously in open-ended possibility
and reductive analysis. The Abhidhamma’s work with the Dhamma takes us deeply into ultimate mat-
ters ( paramattha) and into the irreducible factors of our experience that cannot be analysed further,
even while the relations between them can extend and vary almost infinitely.”
64 We cannot even use language – thinking of Wittgenstein’s example at the beginning of his Philo-
sophical Investigations – if ‘block,’ ‘slab,’ and ‘bring’ are indistinguishable from each other. Diṅnāga
recognises this when he argues that any naming or categorising involves conceptualising distinctions
(ad PS I.1.3–4), and his comment on PS I.1.3d explicitly indicates that whether language refers to
a real individual or not has no bearing on the matter. Even his apoha theory of meaning necessarily
individuates – which is why, although the theory may be a useful tool for getting us to cease con-
ceptualising or to navigate everyday reality more skilfully, apoha theory cannot connect concepts
and percepts (nor does it pretend to: “a general term does not express particulars,” PS V.2). Even if
a word means by excluding other words (PS V.1, 11), these exclusions are just an alternative way of
individuating. Plato’s worry in the Theaetetus that the extreme Heraclitean flux-theorists fall into a
similar predicament (Tht. 156a–157c) is operating in similar terrain.
65 For instance, “it is correct to call nirvāṇa neither existent nor an absence” (MMK 25.10cd).
66 Nāgārjuna’s deployment of the tetralemma has been the topic of significant scholarly attention, par-
ticularly as it is later associated by Candrakīrti with a distinctive interpretation of Nāgārjuna’s thought.
Oetke (1996) discusses the use of the tetralemma in Indian philosophy generally; Ruegg (1977) is a
classic discussion of the tetralemma in Mahāyāna. Tillemans (1999, 200), after considering a “logically
trivial interpretation of the tetralemma” that could take it as “a uniquely ‘therapeutic’ use of language”
(Tillemans 1999, 190), argues instead that it is an example of non-classical logic that does not deviate
considerably from classical logic. See also Ganeri (2001, Ch. 2) and Priest and Garfield (2002).
67 While Candrakīrti’s dispute with Bhāviveka on this point is the origin of it, this became an especially
important marker of a distinct interpretation of Mādhyamaka as Indian Buddhism travelled into Tibet;
see Dreyfus and McClintock (2003).
68 Examples include: “An action does not possess conditions; nor is it devoid of conditions” (MMK I.4);
“vision does not see; nor does nonvision see” (MMK 4.5); “an agent that is both real and unreal does
not bring about an object that is real or one that is unreal” (MMK 8.11); “Dissolution does not at all
exist either with or without arising; arising does not at all exist with or without dissolution” (MMK
21.1); and so on. Cf. “ ‘Hymn to the Unthinkable One,” 12–15.
69 Arnold 2010 and 2012 lean heavily on the explanatory priority aspect.
70 Nāgārjuna recognises this explicitly in Root Verses 9 and in the Dispeller.
71 Garfield (1996) notes the similarity to Pyrrhonists in strategy and aims.
72 For a contrary view, see Matilal (1986, Ch. 2), and more recently Mills (2018).
73 Translation, Westerhoff (2010); the choice of ‘thesis’ for pratijñā reflects a narrow rather than capa-
cious interpretation of what is disavowed, and what emptiness eliminates. Westerhoff (2009b) offers
detailed consideration of three ways in which one might understand, and the tradition in fact has
understood, this verse. A similar claim is found in the Sixty Verses, v. 50.
74 Huntington (1995) and more recently Stepien (2019).
75 The capacious reading would seem to connect back to the ‘auspicious stilling of prapañca’ associated
with emptiness in the opening verse, but this itself depends on whether prapañca is any conceptuali-
sation at all, or merely (as Siderits and Katsura translate it) hypostasisation – that is, conceiving of
things as ‘having intrinsic nature.’ The restricted view is closely related to understanding Nāgārjuna
to be concerned specifically to reject realist semantics: e.g. “Nāgārjuna makes no claims that would
be true or false in virtue of real entities with intrinsic natures” (Tillemans 2017, 117). See Ruegg
(1983, 2003, Part I) for articulation of this semantic interpretation of Madhyamaka thesislessness.
76 This concern about quietism is raised by Tillemans (2011) and in this volume.
231
Amber D. Carpenter
77 Tillemans (2016, Ch. 5) characterises Candrakīrti (but not Bhāviveka) as such a quietist. Tillemans
(2011; 2019) identifies this quietist reading of Candrakīrti as the “typical Prāsaṅgika” perspective
(2011, 156; 2019, 640), with Tillemans (2011) exploring an alternative, more philosophically promis-
ing version of Prāsaṅgika. Tillemans (2017) presents ‘Nāgārjuna and Early Madhyamaka’ as quietist
of a sort, though perhaps not a pernicious sort. It is worth noting that classical Mediterranean skepti-
cism faced this same threat of two-fold quietism: a default conservatism regarding what is commonly
accepted (since all grounds for critique have been undermined), and a simultaneous ideal mental calm
(ataraxia) consisting in holding and asserting no convictions whatsoever (not even that everything is
unknowable).
78 Many other and ingenious ways out have been tried – see for instance Priest, Siderits and Tillemans
(2011).
79 These two run throughout the discussion, returning for instance at vv. 44–57, 212–13, 221–23, 230
and 381. Carpenter (2015) discusses in detail the correct understanding of these two goals, and the
structure of the ethics in the Ratnāvalī which avoids presenting them as parallel alternative ends.
Shulman (2011) offers insightful close comment on what he calls the ‘philosophical’ portions of the
Precious Garland, although one might (and the reading set out here does) query the aptness of cutting
off certain portions of the text from others as being ‘philosophical.’
80 One way of reading verses 174–75 might support ‘happiness’ and ‘enlightenment’ as non-intersecting
alternatives, though it relies on taking them out of the context of the whole.
81 What follows would be an alternative to the Katsura-Siderits suggestion that Chapter 26 is compatible
with what came before by being pegged to conventional reality, while the foregoing chapters concern
ultimate reality; the worry is that resolving tensions in this way so separates the two truths that there
is no way that one could be used to lead to the other.
82 “In one who first practices high status, definite goodness arises later” (RĀ 3). Translations from the
Precious Garland are those of Hopkins (1998), which translates from the Tibetan translation (‘high
status’ is roughly worldly well-being, while ‘definite goodness’ is the ultimate enlightened state).
Translation from the Sanskrit is available in Tucci (1934; 1936), though the English is quite outdated.
83 Maintaining an ironic engagement, as Siderits (2003a, Ch. 5) calls it, is not easy, but the stuff of
bodhisattvas.
84 See Huntington (2018). However, determining whether the end achieved is indeed the same, and the
correct one, would seem to require engaging in philosophical reflections on the explicit content of the
Root Verses. Only direct philosophical engagement with the Root Verses could suggest that reading it
as one reads fiction would be salutary – it does not come with such reading instructions on its surface,
or present as fiction at first glance.
85 Analogical reasoning, for instance, at “Hymn to the Unthinkable One,” 4–6; a recap of MMK 1 at
“Hymn to the Unthinkable One,” 9–12; “Hymn to Him Who Has Gone Beyond the World,” 4–7, 11,
14 17.
86 The “Hymn According to the Supreme Truth” especially, and much of the “Hymn to the Incompara-
ble One.”
87 Contra Schroeder (2000).
88 Negatively, as in Candrakīrti’s way, doing no more than turning others’ commitments against them, or
constructively, as in Bhāviveka’s way, offering suitably contextualized assertions and reasons which
are then undercut or their own dependency (their emptiness) made explicit.
89 One of the great debates within the Madhyamaka originating from Nāgārjuna is whether there is
‘something else,’ or whether the recognition of the dependency of one’s thought, as well as any
objects of thought, suffices to effect the blissful transformation to stillness.
Bibliography
Arnold, Dan. 2005. Buddhists, Brahmins and Beliefs. New York: Columbia University Press.
———. 2010. “Nāgārjuna’s ‘Middle Way’: A Non-Eliminative Way of Understanding Selflessness.”
Revue Internationale de Philosophie 64 (253): 367–95.
———. 2012. “The Deceptive Simplicity of Nāgārjuna’s Arguments Against Motion.” Journal of Indian
Philosophy 40: 553–91.
Bastow, David. 1994. “The Mahā-Vibāṣā Arguments for the Sarvāstivāda.” Philosophy East and West
44 (3): 489–500.
232
Nāgārjuna
Bodhi, Bhikkhu, trans. 2000. The Connected Discourses of the Buddha: A Translation of the Saṃyutta
Nikāya. Somerville, MA: Wisdom Publications.
Burnyeat, Myles, and Michael Frede. 2015. The Pseudo-Platonic Seventh Letter. Edited by Dominic
Scott. Oxford: Oxford University Press.
Carpenter, Amber D. 2014. “Ethics of Substance.” Aristotelian Society Supplementary Volume 88 (1):
145–68.
———. 2015. “Aiming at Happiness, Aiming at Ultimate Truth – In Practice.” In Moonpaths: Ethics and
Emptiness, edited by the Cowherds, 21–42. New York: Oxford University Press.
Cox, Collett. 2004. “From Category to Ontology: The Changing Role of ‘Dharma’ in Sarvāstivāda
Abhidharma.” Journal of Indian Philosophy 32 (5–6): 543–97.
Dreyfus, Georges B. J., and Jay L. Garfield. 2010. “Madhyamaka and Classical Greek Scepticism.” In
Moonshadows: Conventional Truth in Buddhist Philosophy, edited by the Cowherds, 115–30. New
———. 2021. “Madhyamaka and Skepticism.” In Skepticism in the Indian Philosophical Tradition,
edited by Matthew R. Dasti and E. Mills. Leiden: Brill.
Dreyfus, Georges B. J., and Sara L. McClintock, eds. 2003. The Svātantrika-Prāsaṅgika Distinction:
What Difference Does a Difference Make? Boston: Wisdom Publications.
Erb, Felix. 1997. Śūnyatāsaptativṛtti: Candrakīrtis Kommentar zu den “Siebzig Versen über die Leer-
heit” des Nāgārjuna (Kārikas 1–14); Einleitung, Übersetzung, textkritische Ausgabe des Tibetischen
und Indizes. Stuttgart: F. Steiner.
Ferraro, Giuseppe. 2013. “Outlines of a Pedagogical Interpretation of Nāgārjuna’s Two Truths Doctrine.”
———. 2014. “Grasping Snakes and Touching Elephants: A Rejoinder to Garfield and Siderits.” Journal
of Indian Philosophy 42 (4): 451–62.
———. 2017. “Realistic-Antimetaphysical Reading Vs Any Nihilistic Interpretation of Madhyamaka.”
Journal of Indian Philosophy 45: 73–98.
Ganeri, Jonardon. 2001. Philosophy in Classical India: An Introduction and Analysis. London: Routledge.
Garfield, Jay L. 1994. “Dependent Arising and the Emptiness of Emptiness: Why Did Nāgārjuna Start
with Causation?” Philosophy East and West 44 (2): 219–50.
———. 1996. “Emptiness and Positionlessness: Do the Mādhyamika Relinquish All Views?” Journal of
Indian Philosophy and Religion 1: 1–34.
Garfield, Jay L., and Graham Priest. Forthcoming. “Madhyamaka, Ultimate Reality, and Ineffability.” In
Reasons and Empty Persons: Mind, Metaphysics, and Morality: Essays in Honor of Mark Siderits,
edited by Christian Coseru. New York: Springer.
Garfield, Jay L., and Jan Westerhoff, eds. 2015. Madhyamaka and Yogācāra: Allies or Rivals? New York:
Oxford University Press.
Gethin, Rupert. 1992. “The Mātikās: Memorization, Mindfulness and the List.” In In the Mirror of Mem-
ory, edited by Janet Gyatso, 149–72. Albany, NY: SUNY Press.
———. 2004. “He Who Sees Dhamma Sees Dhammas.” Journal of Indian Philosophy 32 (5/6): 513–42.
Hayes, Richard P. 1994. “Nāgārjuna’s Appeal.” Journal of Indian Philosophy 22 (4): 299–378.
Heim, Maria. 2014. The Forerunner of All Things: Buddhaghosa on Mind, Intention, and Agency. New
Hopkins, Jeffrey. 1998. Nāgārjuna’s Precious Garland: Buddhist Advice for Living and Liberation.
Ithaca, NY: Snow Lion Publications.
Huntington, C. W. Jr. 1995. “A Way of Reading.” Journal of the International Association of Buddhist
Studies 18 (2): 279–308.
———. 2007. “The Nature of the Mādhyamika Trick.” Journal of Indian Philosophy 35 (2): 103–31.
———. 2018. “Nāgārjuna’s Fictional World.” Journal of Indian Philosophy 46 (1): 153–77.
Kalupahana, David J. 1975. Causality: The Central Philosophy of Buddhism. Honolulu, HI: University
of Hawai’i Press.
———. 1986. Nāgārjuna: The Philosophy of the Middle Way. Albany, NY: SUNY Press.
———. 1992. A History of Buddhist Philosophy: Continuities and Discontinuities. Honolulu, HI: Uni-
versity of Hawai’i Press.
Komito, David Ross. 1987. Nāgārjuna’s “Seventy Stanzas”: A Buddhist Psychology of Emptiness. Ithaca,
NY: Snow Lion Publications.
Lindtner, Christian. 1982. Nagarjuniana: Studies in the Writings and Philosophy of Nāgārjuna. Copen-
hagen: Akademisk Forlag.
233
Amber D. Carpenter
Loizzo, Joseph. 2007. Nāgārjuna’s Reason Sixty (Yuktiṣaṣṭikā) with Chandrakīrti’s Commentary
(Yuktiṣaṣṭikāvṛtti). New York: American Institute of Buddhist Studies at Columbia University.
Matilal, Bimal K. 1986. Perception: An Essay on Classical Indian Theories of Knowledge. Oxford:
Clarendon Press.
Meyers, Karin. 2018. “False Friends: Dependent Origination and the Perils of Analogy in Cross-Cultural
Philosophy.” Journal of Buddhist Ethics 25: 785–818.
Mills, Ethan. 2018. Three Pillars of Skepticism in Classical India: Nāgārjuna, Jayarāśi, Śrī Harṣa. Lan-
ham, MD: Lexington Books.
Ñāṇamoli, Bhikkhu, and Bhikkhu Bodhi, trans. 2009. The Middle Length Discourses of the Buddha:
A Translation of the Majjhima Nikāya. 4th ed. Somerville, MA: Wisdom Publications.
Oetke, Claus. 1996. “Ancient Indian Logic as a Theory of Non-Monotonic Reasoning.” Journal of Indian
Philosophy 24 (5): 447–539.
———. 2007. “On MMK 24.18.” Journal of Indian Philosophy 35 (1): 1–32.
———. 2011. “Two Investigations on the Madhyamakakārikās and the Vigrahavyāvartanī.” Journal of
Indian Philosophy 39 (3): 245–325.
Pind, Ole Holten. 2001. “Why the Vaidalyaprakaraṇa Cannot Be an Authentic Work of Nāgārjuna.” Wie-
ner Zeitschrift für die Kunde Südasiens/Vienna Journal of South Asian Studies 45: 149–72.
Priest, Graham, and Jay L. Garfield. 2002. “Nāgārjuna and the Limits of Thought.” In Beyond the Limits
of Thought, edited by Graham Priest, 249–70. Oxford: Oxford University Press.
Priest, Graham, Mark Siderits, and Tom Tillemans. 2011. “The Two Truths About Truth.” In Moon-
shadows: Conventional Truth in Buddhist Philosophy, edited by The Cowherds, 131–50. New York:
Ronkin, Noa. 2005. Early Buddhist Metaphysics: The Making of A Philosophical Tradition. Abingdon,
Oxon: RoutledgeCurzon.
Ruegg, David Seyfort. 1977. “The Uses of the Four Positions of the ‘Catuṣkoṭi’ and the Problem of the
Description of Reality in Mahāyāna Buddhism.” Journal of Indian Philosophy 5 (½): 1–71.
———. 1981. The Literature of the Madhyamaka School in India. Wiesbaden: Otto Harrossowitz.
———. 1983. “On the Thesis and Assertion in the Madhyamaka/ dBu ma.” In Contributions on Tibetan
and Buddhist Religion and Philosophy, edited by Ernst Steinkellner and Helmut Tauscher, eds. Wie-
ner Studien zur Tibetologie und Buddhismuskunde 11. Vienna: Arbeitskreise für Tibetische und Bud-
dhistische Studien.
———. 2003. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy. Wiener Stu-
dien zur Tibetologie und Buddhismuskunde. Vienna: Arbeitskreise für Tibetische und Buddhistische
Studien.
Schroeder, John. 2000. “Nāgārjuna and the Doctrine of ‘Skillful Means’.” Philosophy East and West 50
(4): 559–83.
Shulman, Eviatar. 2007. “Creative Ignorance: Nāgārjuna on the Ontological Significance of Conscious-
ness.” Journal of the International Association of Buddhist Studies 30 (1–2): 139–73.
———. 2008. “Early Meanings of Dependent Origination.” Journal of Indian Philosophy 36 (2):
297–317.
———. 2010. “The Commitments of a Madhyamaka Trickster: Innovation in Candrakīrti’s Prasanna-
padā.” Journal of Indian Philosophy 38 (4): 379–417.
———. 2011. “Ratnāvalī: A Precious Garland of Buddhist Philosophical Systems.” Indo-Iranian Jour-
nal 54: 301–29.
———. 2014. Rethinking the Buddha: Early Buddhist Philosophy as Meditative Perception. New York:
Siderits, Mark. 2003a. Personal Identity and Buddhist Philosophy: Empty Persons. Aldershot, England:
Ashgate.
———. 2003b. “On the Soteriological Significance of Emptiness.” Contemporary Buddhism 4 (1): 9–23.
———. 2007. Buddhism as Philosophy: An Introduction. Aldershot, England: Ashgate.
Siderits, Mark, and Jay Garfield. 2013. “Defending the Semantic Interpretation: A Reply to Ferraro.”
Siderits, Mark, and Shōryū Katsura. 2013. Nāgārjuna’s Middle Way: Mūlamadhyamakakārikā. Somer-
Stepien, Rafal K. 2019. “Abandoning All Views: A Buddhist Critique of Belief.” The Journal of Religion
99 (4): 529–66.
234
Nāgārjuna
Tillemans, Tom J. F. 1999. “Is Buddhist Logic Non-classical or Deviant?” In Scripture, Logic, Language:
Essays on Dharmakīrti and His Tibetan Successors, 187–205. Translated by John Dunne. Somerville,
MA: Wisdom Publications.
———. 2011. “How Far Can a Mādhyamika Buddhist Reform Conventional Truth? Dismal Relativism,
Fictionalism, Easy-Easy Truth, and the Alternatives.” In Moonshadows: Conventional Truth in Bud-
dhist Philosophy, edited by The Cowherds, 151–65. Oxford: Oxford University Press.
———. 2016. How Do Mādhyamikas Think? And Other Essays on the Buddhist Philosophy of the Mid-
dle. Somerville, MA: Wisdom Publications.
———. 2017. “Philosophical Quietism in Nāgārjuna and Early Madhyamaka.” In The Oxford Handbook
of Indian Philosophy, edited by Jonardon Ganeri, 110–32. New York: Oxford University Press.
———. 2019. “Mādhyamikas Playing Bad Hands: The Case of Customary Truth.” Journal of Indian
Philosophy 47 (4): 635–44.
Tola, Fernando, and Carmen Dragonetti. 1985. “Nāgārjuna’s Catustava.” Journal of Indian Philosophy
13 (1): 1–54.
———. 1987. “Śūnyatāsaptati: The Seventy Kārikās on Voidness (According To The Svavṛtti) Of
Nāgārjuna.” Journal of Indian Philosophy 15 (1): 1–55.
———. 1995. On Voidness: A Study on Buddhist Nihilism. New Delhi: Motilal Banarsidass.
Tucci, Giuseppe. 1934. “The Ratnāvalī of Nāgārjuna.” Journal of the Royal Asiatic Society of Great
Britain and Ireland 66 (2): 307–25.
———. 1936. “The Ratnavali of Nagarjuna.” Journal of the Royal Asiatic Society of Great Britain and
Ireland 68 (2–3): 237–52, 423–35.
Vaidya, P. L., ed. 1960. Aṣṭasāharśrikā Prajñāpāramitā, with Haribhadra’s Commentary Called Āloka.
Darbhanga: The Mithila Institute.
de la Vallée Poussin, Louis. 1933. “Réflexions sur le Madhyamaka.” Mélanges chinois et bouddhiques
3: 4–59.
Walser, Joseph. 2005. Nāgārjuna in Context: Mahāyāna Buddhism and Early Indian Culture. New York:
Wayman, Alex. 1977. “Who Understands the Four Alternatives of the Buddhist Texts?” Philosophy East
and West 27 (1): 3–21.
Westerhoff, Jan. 2009a. Nāgārjuna’s Madhyamaka: A Philosophical Introduction. Oxford: Oxford Uni-
versity Press.
———. 2009b. “The No-Thesis View: Making Sense of Verse 29 of Nāgārjuna’s Vigrahavyāvartanī.”
In Pointing at the Moon: Buddhism, Logic, Analytic Philosophy, edited by Mario D’Amato, Jay L.
Garfield, and Tom J. F. Tillemans, 26–37. New York: Oxford University Press.
———. 2010. The Dispeller of Disputes: Nagarjuna’s Vigrahavyāvartanī. Oxford: Oxford University
Press.
———. 2016. “On the Nihilist Interpretation of Madhyamaka.” Journal of Indian Philosophy 44 (2):
337–76.
———. 2018a. Crushing the Categories (Vaidalyaprakaraṇa). New York: American Institute of Bud-
dhist Studies.
———. 2018b. The Golden Age of Indian Buddhist Philosophy. Oxford: Oxford University Press.
Wood, Thomas E. 1994. Nāgārjunian Disputations: A Philosophical Journey Through an Indian Look-
ing-Glass. Honolulu, HI: University of Hawai’i Press.
235
14
ĀRYADEVA
Quietism and Buddhist Ethics
Tom J. F . Tillemans
Introduction
Āryadeva was a second–third century CE Indian Buddhist author who – along with his teacher
Nāgārjuna (c. 150–250 CE) and perhaps an obscure, lesser figure Rāhulabhadra – is con-
sidered a co-founder of the “Philosophy of the Middle” (Madhyamaka). This Great Vehicle
(Mahāyāna) philosophy is characterized by its use of dialectics to reject any attribution of
intrinsic natures (svabhāva) to things, an intrinsic nature being a type of aseity – what a thing
is in itself, completely independently of other factors. And while possessing an intrinsic nature
is thought by most Indian thinkers to be a necessary condition for anything to be a fully
real entity (bhāva), Āryadeva, Nāgārjuna, and their followers, the Mādhyamikas, held that
all things are what they are only dependently. The conclusion: things are “empty” (śūnya) of
intrinsic natures; it is impossible for there to be fully real entities and impossible to ascribe
any fully real properties to them. Like Nāgārjuna, Āryadeva is thus not a metaphysical real-
ist (Tib. dngos smra ba = Skt. *bhāvavādin); his thought is in sharp contrast with that of his
Indian opponents and most of his coreligionists. I will take up his philosophical perspectives
on metaphysics and ethics in more detail in what follows.1
Life
Following Candrakīrti (sixth century CE), one of Āryadeva’s two Indian commentators,
Āryadeva was born in Siṃhaladvīpa as a prince.2 He subsequently traveled from Siṃhaladvīpa –
likely Śrī Laṅka – to South India and became a disciple of the great Mādhyamika thinker
Nāgārjuna, the author of “the sixfold corpus of [Madhyamaka] reasoning,” including the Verses
on the Philosophy of the Middle (Madhyamakakārikā). After supposedly founding monasteries
in South India, he moved to Northern India, to Nālandā.3 All accounts concord on Āryadeva’s
skill in defending Buddhism from non-Buddhist thinkers. And indeed, his major work, The
Four Hundred Verses (Catuḥśataka), does show a strong familiarity with non-Buddhist doc-
trines of the time, especially those of Brahmanical schools such as the Sāṃkhya and Vaiśeṣika,
which it subjects to trenchant criticism. In a famous debate, supposedly with a Mātṛceta or
Durdharṣakāla – whom the Tibetan historian Tāranātha (1575–1634) implausibly claims to
become the second century CE Buddhist poet Aśvaghoṣa (Chimpa and Chattopadhyaya 1970,
236 DOI: 10.4324/9781351030908-23

Āryadeva
131–32) – he is said to have defeated his opponent with dialectics, mantras, and supernatural
feats, converting him finally to Buddhism (see Cabezón 2008, 75–78). The Chinese biography
goes a step further: Āryadeva ended his days murdered by the disciples of a non-Buddhist
philosopher whom he had defeated in debate (Taishō 2048, Robinson 1967, 27–28, Lang 1986,
7). As is frequent in hagiographical versions of Indian debates, we learn about the buildup and
the denouement, the magical tricks and the perfidy, but relatively little of philosophical content
(Cabezón 2008, 89–90).
Works
The philosophies one attributes to a second- or third-century thinker will significantly differ
depending upon the texts one regards as authentic. Such questions of textual authenticity are
generally not simple matters. Various works, besides the Four Hundred Verses, have been
attributed to Āryadeva, especially in the Tibetan and Chinese canons (see Tillemans 1990,
6–7). In fact, there can be little doubt that the name “Āryadeva” was widely applied to authors
of works that could not have been written by the second- or third-century author of the Four
Hundred Verses.4 We shall, in what follows, examine aspects of the thought of the author of
the Four Hundred Verses.5 I consider this treatise – and, to a much lesser degree, the Hundred
Verse Treatise – as the best representative of Āryadeva’s thought, although it may well be that
at least some other texts that critique non-Buddhist systems, e.g., the Hundred Letter Treatise
(akṣaraśataka; trans. Gokhale 1930), are by the same historical individual, too.6
Overview of the Four Hundred Verses

Before giving a philosophical analysis of Āryadeva’s ideas, here is a brief overview of the con-
tent of the Four Hundred Verses. Chapters one–eight discuss ethical themes, many of which
are common to most Buddhist thinkers whether of the Great Vehicle or other traditions. The
first four chapters, for example, criticize the four basic illusions (viparyāsa) governing our
psyches, viz., our mistakenly taking transitory life as permanent, what is painful as pleasur-
able, what is dirty as clean, and what is selfless as having a self.7 The next chapters deal,
respectively, with the bodhisattva’s practices leading to enlightenment, the elimination of the
passions (kleśa) that impede such practices, and the elimination of attachment to the objects of
the senses; the eighth chapter deals with the proper practices of Buddhist disciples.
It is in the latter half of the Four Hundred Verses that we find the detailed refutations
of various ontologies, Buddhist and non-Buddhist, as well as instruction on methodologi-
cal principles for Madhyamaka thought. Chapters nine–eleven refute, respectively, the real-
ity of permanent entities (nitya), personal selves (ātman), and time (kāla); Chapter twelve
refutes “heretical views” (dṛṣṭi) – i.e., inter alia, non-Buddhist views on liberation, scripture,
and asceticism – and critiques the importance attributed to high-caste birth; Chapter thirteen
refutes Buddhist and non-Buddhist positions on the reality of the sense faculties and their
objects (indriyārtha); Chapter fourteen refutes our deep-seated hypostatization of dichoto-
mies, or “extremes” (antagrāha); Chapter fifteen refutes the reality of conditioned things
(saṃskṛtārtha); Chapter sixteen, entitled “Cultivating understanding in the master and dis-
ciple,” is on Madhyamaka method and logical issues that arise in executing that method. In
what follows, I will first seek to better understand Āryadeva’s philosophical orientation and
method in the latter half of the Four Hundred Verses, and then take up the first half on ethics
and religion; I close with the question of the systemic coherence of the work.
237
Tom J. F. Tillemans
Quietism about ontology

Āryadeva’s orientation in the last chapters of the Four Hundred Verses and the Hundred
Verse Treatise can, in my opinion, best be described as quietistic. He neither advances nor
defends any theses that are ontologically committed; that is to say, that presuppose accept-
ance of fully real entities (bhāva). This includes positive statements as well as negations
and all conjunctions or disjunctions of the two. In having no theses and no debates about
them he closely follows Nāgārjuna, who had famously said in verse 50 of his Sixty Verses
on Reasoning (Yuktiṣaṣṭikā):
Superior individuals have no theses (Tib. phyogs = Skt. pakṣa) and no debates; how
could there be any opposing theses for those who have no theses [themselves]?8
Now, it might be objected that quietism is, in one way or another, a relatively widespread
theme in Buddhism, Madhyamaka or not. There are, for example, a variety of meditative tech-
niques to arrive at an irenic state beyond conceptual thought, at nirvāṇa, a state of peace, or the
“silence of the noble ones” (āryatūṣṇīṃbhāva). One finds some form of quietism in the Pāli
canon, Tantra, and Zen. True, but the particularity of the Mādhyamikas, including Āryadeva, is
that they emphasize philosophical analysis as a method leading to that quietening of thought,
diagnosing rationally and with dialectics where it goes wrong in making ontologically commit-
ted truth claims and debating about them. That is why it is of special interest to philosophers.9
Two senses of “no thesis,” “no debates”

Understanding what Āryadeva says presupposes clarity on the thought he inherited. This is par-
ticularly so when it comes to Āryadeva’s “no thesis” methodology and metaphilosophy. Unfor-
tunately, it seems inadequately recognized in much modern scholarship that Nāgārjuna and
Āryadeva both promote two markedly different ways in which a Mādhyamika philosopher does
not have theses, that is to say, one way that does not clearly concern the propriety or legitimacy
of philosophical debate and another that certainly does. First, the Mādhyamika does not have
a thesis because there is not anything fully real, including thesis-statements. And second, the
Mādhyamika does not have a thesis because she does not make, or personally endorse, truth
claims presupposing anything fully real, all such claims being refutable. It is the second sense of
“no thesis” which most clearly leads to quietism, as it embodies the normative stance that one
should not be committed to the content of thesis-statements or debate about that content. The sec-
ond also lends itself to a semantic interpretation of the no-thesis stance, viz., that a Mādhyamika
does not accept or debate about any theses whose terms range over fully real entities (bhāva).
Although it seems to have been important in Mādhyamikas’ debates with their Brahmanical
contemporaries, what I am calling the first sense does not turn on the general stance that all
ontologically committing propositions are somehow refutable. Nor is it about the impossibil-
ity of making claims via statements with a realist semantics. It turns on the real existence or
nonexistence of just one statement, viz., the Mādhyamika’s own fundamental principle that
nothing has any intrinsic nature (svabhāva). Let us look at the textual data a bit closer. In
Nāgārjuna’s treatise, the Dismissal of Disputes (Vigrahavyāvartanī), a non-Buddhist meta-
physical realist – a Naiyāyika or Vaiśeṣika – argues in the first verse that Nāgārjuna’s own
statement that nothing has any intrinsic nature is self-refuting, because if true, it would imply
that this very thesis-statement itself does not really exist and therefore could not do anything;
238
Āryadeva
it could not mean anything, prove anything, or refute anything. Nāgārjuna replies in verse 29
of the Dismissal:
If I had any thesis ( pratijñā) at all, then I would, for that precise reason, have faults.
I don’t have any theses and thus I don’t have the fault [of self-refutation of which
you accuse me].10
Here, Nāgārjuna recognizes that if he had a fully real thesis-statement to the effect that nothing
is fully real, then the real existence of that very statement would falsify what it asserts; but,
so his reply goes, he does not have any such real thesis-statement, precisely because nothing
is real, and hence does not fall into self-refutation. Indeed, the Dismissal repeatedly argues
that reasoning, statements, etc., can function all the while being unreal, just as two illusory
or magical projections can do things (or, more accurately, seem to do things) to each other. In
chapter sixteen of the Four Hundred Verses, and in a broadly similar passage in the Hundred
Verse Treatise, chapter ten (Tucci 1929, 82), Āryadeva takes up that charge of self-refutation.
He says in the Four Hundred Verses 16.2:
[Objection:] Since the [Mādhyamika] speaker, what he states [and his own words]
would also be [unreal], then his saying that [all] is empty would be incoherent. [We
reply:] That [i.e., the speaker, assertion and words] which arises in dependence is not
existent in any of the three.
Instead of appealing to reasoning as simply an efficacious play of illusion, he invokes the

basic idea that all things – including speakers, what is stated, and words – are dependent upon
each other. However, for our purposes, the objection and reply are the same as in Nāgārjuna’s
Dismissal. In both cases, the no-thesis stance invoked is not the generalized, quietistic refusal
to make truth claims about the real; it is about the mere existence, or illusoriness, of one
statement and its speaker. In short, this is a somewhat specialized debate that should not be
mistakenly conflated with Mādhyamika’s generalized rejection of truth claims or the rejec-
tion of any semantics that presupposes fully real entities (see Oetke 2003 against Seyfort
Ruegg’s interpretation of Dismissal, verse 29, in such a generalized fashion). The undeniable
philosophical interest and importance of a rejection of realist semantics does not mean that the
specific passages in the Dismissal, like verse 29, are to be read as themselves expressing that
rejection. The same critique would apply to semantic readings of Āryadeva’s 16.2.
Let us now turn to the second sense of “no thesis.” The stance that one should not make
or endorse any truth claims that imply commitment to fully real things, or that have a realist
semantics, is at its clearest in Āryadeva’s systematic refutation of each of the “four alterna-
tives,” or the tetralemma (catuṣkoṭi). According to the tetralemma, an entity either is . . .; is
not . . .; both is . . . and is not . . .; or neither is . . . nor is not. . . . These four supposedly exhaust
the possible formulations of any metaphysical thesis (on the tetralemma, see Seyfort Ruegg
2010, chapter 3). Here is how Āryadeva describes the tetralemma in verse 21 of chapter four-
teen in the Four Hundred Verses:
Being, nonbeing, [both] being and nonbeing, neither being nor nonbeing: such is the
successive method that the wise should always use with regard to identity and the
like [i.e., all other theses].
239
Tom J. F. Tillemans
For Āryadeva, each of these four is typically refuted by reductio ad absurdum arguments,
so-called prasaṅga; i.e., consequences that would be unacceptable to the proponent of the
alternative in question. Āryadeva’s extensive use of absurd consequences is thus taken by
many as evidence that he is a “Consequentialist Mādhyamika” (Tib. dbu ma thal ‘gyur pa;
Skt. *prāsaṅgikamādhyamika) – roughly, a Mādhyamika who only refutes others and does not
positively prove any ontological positions of his own.
Consequentialist Madhyamaka is the way to read Āryadeva’s work via the commentary of
the sixth century Mādhyamika Candrakīrti, the Catuḥśatakaṭīkā, and it does lead to a refusal of
theses, for the would-be intrinsic natures of entities turn out to be so riddled with inconsisten-
cies on analysis that no one can rationally make full-fledged truth claims about them. While
Tibetan thinkers, and indeed most modern scholarship, have embraced some variant of this
Candrakīrtian interpretation of Āryadeva, it is not the only way to understand his thought and
refusal of theses. The other important way is in terms of Yogācāra idealism, by relying upon
the Yogācāra’s trademark stance, viz., the three-nature (trisvabhāva) theory.
Āryadeva as a Yogācāra quietist

While Consequentialist Madhyamaka rejects any and all intrinsic natures (svabhāva) and all truth
claims that presuppose their existence, Āryadeva – if we take a Yogācāra interpretation of his
thought – rejects only one type of nature among three: his quietism is thus correspondingly cir-
cumscribed. In particular, according to this interpretation, Āryadeva refutes the existence of sub-
ject-object dualities and the linguistically conditioned in favor of “mind alone” (cittamātra). More
generally, his Madhyamaka dialectic would serve to reject all intrinsic natures conditioned by
conceptual thought and language; they are shown to be “imagined natures” ( parikalpitasvabhāva)
that are unreal and nonexistent. Āryadeva would leave the ineffable, dependent nature
( paratantrasvabhāva) and the perfect nature ( pariniṣpannasvabhāva) intact as existent.11
It is quite likely that some commentators on Nāgārjuna – i.e., the fifth–sixth-century
Yogācāra commentators, Guṇamati, Gunaśrī or Sthiramati – would have held a three-nature
interpretation of Nāgārjuna, although the surviving textual fragments of the first two authors
are very meager (see Tillemans 1990, 57–58). In the case of Sthiramati (An hui 安慧; c. 510–
570 CE), his commentary (Taishō 1567, Da cheng zhong guan shi lun 大乘中觀釋論) remains
only in a difficult Chinese translation by Wei jing 惟淨. That said, the Yogācāra interpreta-
tion of early Madhyamaka was clearly a strong contender in India. Yogācāra-Madhyamaka
synthesis continues all the way to the eleventh-century thinker Ratnākaraśānti and beyond to
Tibet, its followers being designated under a huge variety of names – “those who profess emp-
tiness of what is other (gzhan stong pa)” and other appellations. Pertinently, Ratnākaraśānti’s
self-definition was as a three-nature Mādhyamika (rang bzhin gsum gyi dbu ma pa =
*trisvabhāvamādhyamika) (see Luo 2018).
The three-nature reading of Madhyamaka is indeed the interpretation promoted by
Dharmapāla (Hu fa 護法; c. 530–561 CE), who commented upon the last half of the Four
Hundred Verses and whose work does not survive in Sanskrit, was never translated into
Tibetan, but only comes to us in Xuanzang’s Chinese translation (Taishō 1571, Guang bai
lun shi lun 廣百論釋論). It was Dharmapāla’s commentary on the last chapter of Āryadeva’s
Four Hundred Verses that provided his responses in a crucial debate with the Mādhyamika
Bhāviveka (c. 500–570 CE), a debate that set Yogācāra and Madhyamaka on significantly dif-
ferent trajectories. (For a translation of key parts of this chapter of Dharmapāla’s commentary,
see Hoornaert [2004]).12 See also Kajiyama (1968/1969) on the historical relationship between
Bhāv(av)iveka, Sthiramati, and Dharmapāla.) Āryadeva was thus the focal point for a major
240
Āryadeva
parting of the ways in India. Interestingly enough, he was also probably a focus for the Tibetan
synthetic movements – his commentator Dharmapāla sometimes figures in their lineages, too.
What textual evidence might have prompted Dharmapāla to see Yogācāra in Āryadeva?
As Hoornaert pointed out (2004, n. 11), it seems that Dharmapāla was working with a San-
skrit manuscript of the Four Hundred Verses that was significantly different from that on
which Candrakīrti commented. Dharmapāla’s Sanskrit source must have contained some key
Yogācāra terminology lacking in that used by Candrakīrti. In particular, Dharmapāla’s three-
nature stance on Āryadeva comes out most clearly in his commentary to Four Hundred Verses
16.23, a verse that can be translated as follows from the Chinese:
[Objection:] If things are empty of intrinsic natures (ben xing 本性 = svabhāva),

what benefit [do we gain] by seeing that emptiness (kong 空 = śūnyatā)? [Response:]
It is the realization of emptiness that eliminates the fetters of false conceptualizations
(xu wang fen bie 虛妄分別 = abhūtaparikalpa).13
The term abhūtaparikalpa is a notoriously important Yogācāra term, one that easily
evokes the three-nature theory. False conceptualizations are themselves dependent natures
( paratantrasvabhāva), as they arise causally and produce effects, but their content consists of
imagined natures ( parikalpitasvabhāva). Indeed, Dharmapāla does go into an exposition of the
Yogācāra three-nature theory immediately after this verse. If, however, we translate the same
verse as it figures in Tibetan in Candrakīrti’s commentary – the actual Sanskrit of the passage
is unfortunately lost – it is apparent that abhūtaparikalpa would not have occurred.14 We sim-
ply have the basic and recurrent Buddhist idea that conceptualization obscures or blocks direct
seeing of the truth.
[Objection:] If entities exist by their intrinsic natures, what benefit do we gain in

seeing emptiness? [Response:] Seeing is fettered by conceptualizations (rtog
pa = kalpanā); they are what is to be refuted here [in the Four Hundred Verses].
It is also clear that the rest of verse 23 is very different in the manuscript used by Candrakīrti
and thus not at all surprising that it led to a very different commentary from that of Dharmapāla.
In Candrakīrti’s explanation, there is no clear reference to Yogācāra ideas.
The crucial point of Dharmapāla’s interpretation of Āryadeva is to avoid what he sees
as a nihilistic notion of ultimate truth ( paramārthasatya).15 While many Mādhyamikas con-
strue the ultimate truth, emptiness (śūnyatā), as a simple negation ( prasajyapratiṣedha) or an
absence of any real intrinsic nature whatsoever, Dharmapāla, in his commentary to the Four
Hundred Verses 16.23, stresses that it is only the linguistically conditioned imagined natures
that are customary truth, while the ineffable dependent and perfect natures are ultimate truths.
Moreover, he makes it clear (f. 247c) that those linguistically conditioned natures are the same
as universals, or common characteristics (sāmānyalakṣaṇa), while the dependent natures are
particulars (svalakṣaṇa). There is thus a deliberate rapprochement with the nominalist philoso-
phy of the Epistemological school of Dignāga (c. 480–540 CE). As Dharmapāla says:
Words designate only common characteristics (共相, sāmānyalakṣaṇa), but the indi-
vidual characteristics (自相, svalakṣaṇa) of things are beyond the scope of words.
Individual characteristics are not nonexistent (非無), and common characteristics are
not existent (非有).
(247c, translation Hoornaert 2004)
241
Tom J. F. Tillemans
The upshot is that ultimate truth for him is certainly not the simple absence of all intrinsic nature,
as it is for Mādhyamikas from Bhāviveka and Candrakīrti on; it is the combination of the inef-
fable dependent nature – i.e., particulars perceived without any linguistic conditioned natures –
and the perfect nature. The two philosophical interpretations of Āryadeva thus differ markedly:
the Yogācāra sees him as accepting a wordless, conceptless reality composed of particulars,
whereas for a Mādhyamika, Āryadeva’s ultimate truth is the denial of any reality whatsoever.
Ethics
What are we to make of Āryadeva’s ethics in the Four Hundred Verses? Importantly, if Āryadeva,
like Nāgārjuna, rejected Buddhist and non-Buddhist claims that involve ontological commit-
ment, he did not think that he also had to reject or remain quietistic about ethical positions and
religious notions of enlightenment, bodhisattva paths, and the like. Indeed, as we saw in our
outline of the chapters of the Four Hundred Verses, half of the work argues for basic Buddhist
ethical views on impermanence, suffering, purity, selflessness, altruistic intentions, freedom
from attachments, and the conduct of bodhisattvas. Āryadeva thus clearly thought that such
argumentation did not fall victim to his metaphysical critiques or his quietism about having
theses. The result is a complex position that reoccurs, in one way or another, in major Madhya
maka writing, from Nāgārjuna’s Jewel Garland (Ratnāvalī) to Candrakīrti’s commentaries
(vṛtti, ṭīkā) on Nāgārjuna and Āryadeva, Śāntideva’s Engaging in the Practices of a Bodhisattva
(Bodhicaryāvatāra), and on to the Path and its Stages (Lam rim) of Tsong kha pa (1357–1419).
In these works, we find a more or less nuanced rejection of ontologies coupled with a robust
advocacy of the canonical Buddhist ethics and religious practices promoted in the Discourses
(Sūtra) traditionally attributed to the Buddha, the Scholastic Teachings (Abhidharma), Code of
Discipline (Vinaya), or the Great Vehicle literature laying out the bodhisattva path.
Descriptive versus revisionary

The Oxford philosopher P.F. Strawson made a now well-known distinction between descrip-
tive and revisionary metaphysics: “Descriptive metaphysics is content to describe the actual
structure of our thought about the world, revisionary metaphysics is concerned to produce
a better structure” (Strawson 1959, i). We could extend Strawson’s descriptive-revisionary
contrast to ethics. An ethics that advocates supererogation, moral sainthood, extreme self-
sacrifice, passionless non-attachment to anything, and in general very substantial betterment
to our present views and behavior will be said to be revisionary, while one based more on how
people do think and act will go in the direction of description. What would the Madhyamaka
of Āryadeva, Nāgārjuna, Śāntideva et al. look like in this light?
These philosophers will end up with a significant tension in their overall stance: they adopt
a descriptive approach to all matters ontological but a strongly revisionary approach to ethics.
Ethics and bodhisattva practices concern customary (saṃvṛti) truths, or customarily real mat-
ters about which one can and should argue, while ontology is about ultimately ( paramārtha)
real things about which one should neither take a position nor argue; instead of bettering the
world’s customary views on what there is, one supposedly limits oneself, faute de mieux, to
description of what the world acknowledges (lokaprasiddha) in keeping with its epistemic
procedures.16 The ethical discussions in the Four Hundred Verses, however, do not leave the
world’s ethical views largely intact, but rather seek to show that the world should very signifi-
cantly better its views and behavior. In effect, Āryadeva legitimizes the revisionism inherited
from canonical non-Madhyamaka Buddhist sources. A problem of systemic coherence then
242
Āryadeva
arises: if Madhyamaka tells us to do little more than describe what the world acknowledges
on customary matters concerning what is and is not – essentially because there is no realist
ontology that would provide leverage for anything better – then why the radical revisionism
in ethics? That philosophical tension between description and revision is not just limited to
Āryadeva: it is what we find in Nāgārjuna, Śāntideva, and many others. Let us look at how it
plays out in more detail in the Four Hundred Verses. The picture, in my opinion, may well need
some significant rethinking if Madhyamaka Buddhism is to be persuasive to a modern reader.
Misogyny
It must be said from the outset that, for many contemporary readers, the ethical chapters in
Āryadeva’s Four Hundred Verses will meet with stiff resistance, in part because arguments
for the Buddhist goal of passionless detachment are often focused on strong deprecation of
women and their sexuality. Misogyny was, no doubt, widespread in the ascetical monastic
milieu in which much of the Great Vehicle (Mahāyāna) historically originated.17 It is also
apparent in the Pāli Discourses (Sutta) of early Buddhism, like the Numerical Discourses
(Aṅguttaranikāya) (see Bhikkhu Bodhi 2012, 683–84), although the most virulent forms seem
more frequent in the Great Vehicle (Sponberg 1992, 21). In any case, Āryadeva manifestly
inherited monastic asceticism and deprecation of women, bolstering it with questionable dia-
lectics in chapter three of the Four Hundred Verses on the uncleanliness of the female body;
Nāgārjuna had done likewise extensively in his Jewel Garland (Ratnāvalī).
The origins of Āryadeva’s deprecation of women are thus partially explicable historically.
Nonetheless, the disturbance it poses to those who think about his ideas philosophically will
not simply go away with basic historical facts about the early Mahāyāna context. There is, in
effect, a question as to what is core and what is peripheral in this ethical thought. Minimizers
of the disturbance might well try to point out that no less an ethicist than Aristotle defended
condemnable institutions such as slavery as just and natural in Politics I.iii–vii, and that his
promotion of slavery does not vitiate the many important ideas in the Nicomachean Ethics, nor
virtue ethics in general, as it is somehow unessential and peripheral to them.
The Four Hundred Verses’ deprecation of women, however, may present a more intractable
problem than Aristotle’s acceptance of slavery, as the deludedness of male feelings for women
fits very closely with a larger Buddhist discourse on the suffering of saṃsāra and the need for all-
pervasive renunciation to be free of it. True, Āryadeva and Candrakīrti, in the first chapters of the
Four Hundred Verses (see Lang 2003), do not argue only against the cleanliness of the female
physique and the value of women generally: they also argue against the illusion of pleasure,
against physical beauty, romantic love, love for family members, grief at their death (1.12–13),
dedication to work and all other worldly exertions (2.18), and many other key features of human
emotional life. Nonetheless, in keeping with much of Buddhist canonical literature, Āryadeva
and both his Indian commentators took men’s love and desire for women as a commanding
object lesson in deep-rooted, saṃsāric delusion. These are treated as paradigm cases of tainted
ignorance and sources of suffering. The uncomfortable truth is that in most Indian Buddhist
traditions, and in non-Buddhist texts (e.g., the Upaniṣads) alike, there was a large consensus
that powerful contempt and repulsion directed toward women should be cultivated as salvific.18
Intuitions, consensus, moral disagreement

Āryadeva’s method in the first eight chapters of the Four Hundred Verses often seems to be
an attempt to plumb the depths of people’s ethical intuitions, showing that if we were not in
243
Tom J. F. Tillemans
the grip of self-grasping, were informed, and were consistent with our fundamental intuitions,
we would accept canonical Buddhism’s ideas and change much of what we now think. Is that
approach, based on an actual or ideal consensus, or perhaps just a convergence of views, likely
to justify Āryadeva’s revisionism? Not easily, for justifying canonical Buddhism’s ethics purely
or predominantly by means of a metaethics of intuitions and convergence is a very tall order (see
Finnigan 2015). While consistency with intuitions can lead to changes of beliefs without appeals
to metaphysics, scripture, omniscient buddhas, divine will, or some other source of argumen-
tational leverage, the changes would usually leave the larger more fundamental framework of
beliefs untouched. Unacceptable, radical conclusions will instead lead people to say that prem-
ises are false or that the reasoning is riddled with non sequiturs. At some point, revisionary ethi-
cists typically need a set of facts – moral and non-moral – to solve disagreements about the truth
of counterintuitive ethical statements. Āryadeva, as we shall see, acknowledges that he does, too.
Some might well object, however, that Āryadeva could not have thought that Buddhist eth-
ics were true and based on facts because Mādhyamikas, in their metaphysics, do not accept any
statements of any customary truths (saṃvṛtisatya) to be true and about facts. Āryadeva, thus,
only used argument and rhetorical techniques as expedients, or skill in means (upāyakauśalya),
to induce changes of mind for religiously beneficial results but himself rejected truth and facts
in all matters customary. This – or some variant upon it – is a quite widespread interpretation
of Nāgārjuna’s and Āryadeva’s stance and is not without textual support. Dharmapāla, notably,
seems to have read Āryadeva’s ethics largely in this fashion as only upāyakauśalya:
For only emptiness is true, all else is not the truth, but is [just] the Tathāgata using
skillful means (shan quan fang bian 善權方便 = upāyakauśalya) to preach and pub-
licize [the Dharma] in accordance with the desires of sentient beings.
(Tillemans 1990, 92)
There was also a recurring, philosophically significant debate in Tibetan Madhyamaka about
whether the customary is just, at most, wrongly thought to exist (Tib. blo ’khrul ba’i ngor yod
pa) and whether there can ever be right knowledge (tshad ma = pramāṇa) of it (Tillemans
2016, 51–52). Indeed, given the Madhyamaka philosophy of emptiness and its idea that people
always mistakenly hypostatize the customary, the question legitimately arises whether any
customary truth is true.19
Scripture and the humanly unknowable

It is thus a live issue as to what Indo-Tibetan Mādhyamika thinkers could have meant when
they spoke of artha (fact, state of affairs) and satya (truth) – whether they only recognized,
in the end, ersatz, semblant “facts” and “truths” or, perhaps, deflationary facts and truths, as
I and others have argued elsewhere (see Tillemans 2019; Cowherds 2011, chapter 8). We shall
henceforth proceed in a neutral fashion staying uncommitted. It would be wrong, however,
to say that Āryadeva’s ethical arguments and claims are immune from rational disagreement
because he simply never relied on any notion of facts and truth. Mādhyamikas, on the contrary,
repeatedly argue that facts and states of affairs have important implications, and, notably, that
facts about moral causality and retribution, or karma, ground what one should and should
not do. Āryadeva, Candrakīrti, and – to a somewhat lesser degree – Dharmapāla, too, are no
exceptions.20
For Āryadeva and successive Mādhyamikas, the issue in grounding ethics on karma is
essentially epistemic and not to be conflated with the metaphysical issue of the full-fledged
244
Āryadeva
reality (or lack of it) of facts, moral or otherwise, or the propriety of truth claims about them.
The problem is one of human knowledge: the relevant facts – or “facts” – that would eliminate
disagreement cannot, in principle, be adequately understood by ordinary rational humans with
their own reason alone. Many of those facts are, crucially, suprasensible (atīndriyārtha).
Āryadeva, in effect, granted that justification of Buddhist revisionary ethical positions
could not be based only on the religiously unschooled intuitions of the common person and
the observable or inferable facts recognized by the world. Instead, ethics required belief in
the special kinds of customary facts described in Buddhist scriptural statements, i.e., “radi-
cally inaccessible” (atyantaparokṣa) facts about the details of moral causality, or karma. For
Āryadeva, as for subsequent Mahāyānists, much of the workings of that subtle, retributive
causality will be unfathomable by any ordinary human beings in that such causality is humanly
unobservable, not inferable from anything observable, and thus only understandable through
trustworthy scriptures authored by extraordinary individuals with understanding of the supra-
sensible (Tillemans 1999, chapters 1–2).
How could such subtle (but nonetheless customary) facts make rational individuals abandon
their opposition and radically revise their beliefs? Mādhyamikas, especially those who follow
Candrakīrti, stress that the customary is that which is, or should be, acknowledged by the world
(lokaprasiddha) in keeping with the world’s own epistemic procedures. The obvious question,
then, is whether a rational individual could believe that Buddhist scriptures are trustworthy on
unfathomable matters if what they say is, in principle, not fathomable by the world itself.
Four Hundred Verses 12.5 attempts a response. Āryadeva tells us that when, in an ethical
deliberation, there is doubt about the veracity of the Buddha’s descriptions of imperceptible
karmic consequences, we should nonetheless believe them to be true because the Buddha has a
proven track record of truth in other significant areas; notably, in his teaching on emptiness. This
response is crucially important in the history of Buddhism and is still current; it is invoked regu-
larly by Tibetans, including the Fourteenth Dalai Lama. Dharmapāla and Candrakīrti develop
an account of scriptural trustworthiness on the basis of 12.5, and, if we accept Tibetan accounts,
that Āryadevan account was the same as what we find in the sixth and seventh centuries in
Dharmakīrti’s discussions of scripture, i.e., the “triple analysis” (Tib. dpyad pa gsum) of the
reliability of scriptures (see Tillemans 1999, chapters 1 and 2; 2000, 78–79). We can go further:
Āryadeva would have played a major role in grounding Mahāyāna ethics on the suprasensible,
because he was, as far as we can see, the first to have explicitly formulated the dominant Bud-
dhist epistemic method later known as “scripturally based inference” (āgamāśritānumāna).
Here is what Āryadeva said in 12.5:
When someone entertains doubt concerning the imperceptible things ( parokṣa)

taught by the Buddha, he should develop a [rationally founded] belief in these very
things on account of emptiness (śūnyatā).21
His point, especially if we follow Candrakīrti’s interpretation, is not that the truth of statements
about emptiness (śūnyatā) logically entails the truth of statements about karmic facts. It is
rather an epistemic point about justified belief: humans supposedly can, with their own critical
acumen, determine that the teachings on emptiness are an example (dṛṣṭānta) of a time when
the Buddha got a very important matter perfectly right; therefore, so it is argued, because of
his reliability on something essential like emptiness, it is also rational for humans to believe
his other statements even when they cannot – on their own – determine the truth-values. There
is, in effect, a transfer of credibility from one set of teachings to another. In his commentary
to 12.5, Candrakīrti will go further and claim that transferring credibility is in keeping with
245
Tom J. F. Tillemans
worldly practices and intuitions – it is, as he puts it, following “your very own principles alone
(svanayenaiva).” In other words, it follows intuitions and epistemic practices that are acknowl-
edged by the world (lokaprasiddha). We regularly believe people who are proven experts.
As I have argued in more detail elsewhere (Tillemans 2016, chapter 8), that argument can
be seen to be badly flawed upon reflection. Having attested competence and credibility in one
subject area does not transfer to another, unless the two are demonstrably closely related. Most
of us are probably familiar with the trap of wrongly according extended credibility to people
simply because they happen to be qualified in one important field. Those would-be universal
experts become unreliable, not because they never correctly understood significant things, but
because they overstep their qualifications: their expertise is made to extend to subjects not
related to the area in which it has been rightly recognized. The flaw in the appeal to the Bud-
dha’s track record is similar. It is not at all clear that trustworthiness concerning important gen-
eral principles like emptiness does reasonably transfer to explanations concerning the details
of karma in all their specificity and complexity, because the relationship is not clear. While
emptiness of intrinsic nature, as a general principle, may be closely linked with the general
feature that phenomena arise dependently due to causes and conditions – as Buddhists from
Nāgārjuna on have stressed – knowing that much would hardly suggest that one somehow
knows the specific details of what is retribution for what.
In the end, I do not think that Āryadeva’s verse 12.5, Candrakīrti’s construal of it, or the
closely related Dharmakīrtian triple analysis will provide a publicly debatable proof of scrip-
tural trustworthiness to convince people outside an already committed Buddhist community.
In other words, it will not eliminate moral disagreement among rational people of very differ-
ent views. Indeed, Dharmakīrti and his commentators seem to concede as much. Dharmakīrti
was quite skeptical about the probative status of “scripturally based inferences” establishing
such things as subtle karmic causality, and he said clearly in his Own Commentary (svavṛtti)
to verse 1.217 that reasoning “in this way is not a bona fide inference at all (na khalv evam
anumānam anapāyam) as there is no necessary connection between words and states of affairs
(anāntarīyakatvād artheṣu śabdānām)” – the mere fact that something is said can never entail
its truth.22 Śākyabuddhi (seventh–eighth centuries) and Karṇakagomin (c. 800) give us another
angle: “it is not an inference because of factual matters” (vastutas tv ananumānam); it is
only one “on account of the thought of people who want to engage [on the spiritual path]
( pravṛttikāmasya puṃso ‘bhiprāyavaśāt)” (see Tillemans 1999, 43 and n. 11). In other terms,
it is a type of faith-based reasoning that will be used among religious aspirants, but not one
that could be invoked in a neutral, public context of unbelievers.23 Dharmakīrti, to his credit,
recognized that limitation. It is unlikely that Āryadeva and Candrakīrti did.
Epistemic humility?
A larger problem then arises. Is basing ethics, in part or in whole, on a retributive, moral
causality that is nonetheless outside the range of human beings’ own knowledge a coherent,
rational standpoint at all? Arguably, for Buddhist metaphysical realists of the non-Madhyam-
aka schools, like Dharmakīrti, it is not an incoherent position to say that actions have humanly
inaccessible dispositions to lead to certain results, dispositions that are as they are irrespective
of all beliefs and dissent that rational people might have. Such realist philosophers could, in
effect, accept a type of epistemic humility about karmic facts – unknowability does not count
against real existence, but only calls for acceptance of our personal cognitive limitations.24
A Mādhyamika who disavows metaphysical realism and professes adherence to what is cus-
tomarily acknowledged by the world (lokaprasiddha) does not, however, have the ready option
246
Āryadeva
of such humility. Candrakīrti’s argument appealing to “your own principles alone” does not work
well, and it is hard to see that any other like it will. For a Mādhyamika, the conclusion, then, is
not epistemic humility, but is one that is potentially more disturbing to traditional Buddhists: if
subtle karmic cause and effect is not acknowledged as a fact by most of us in the world and not
justifiable by the world’s epistemic procedures, much talk of karma-based distinctions in reli-
gious ethics risks being simply false. In the end, what the unresolved tension between descrip-
tion and revision in the two sections of the Four Hundred Verses shows us is that a Madhyamaka
philosophy that emphasizes quietism about ontology and a strong reliance on worldly intuitions
will have to move in a different direction from Buddhist traditions that have heavily relied on the
suprasensible and the faith-based, discussable only within protected confines. The price to pay
for the philosophy of the second eight chapters of the Four Hundred Verses would seem to be
much of the ethics of the first eight. Instead of Āryadeva’s positions, Madhyamaka’s emphasis
on the world’s ideas would seem to require a much more secularly oriented, publicly debatable
ethics that values what people do think with their own autonomous reason.
Abbreviations
CS Catuḥśataka of Āryadeva.
CST Catuḥśatakaṭīkā of Candrakīrti.
GBLSL Guang bai lun shi lun 廣百論釋論. Taishō 1571. Dharmapāla’s commentary on the
last eight chapters of CS.
Taishō Taishō Shinshū Daizōkyō. The Tripiṭaka in Chinese. Ed. by J. Takakusu and K. Wata-
nabe. 100 volumes. Tokyo, 1925–1935.
Notes
1 I am indebted to Finnigan (2015; 2018). In Tillemans (2016, chapter 8), I had said that Mādhyamikas
try to justify their ethical views to non-believers by appeals to intuitions and what is acknowledged
by the world. No doubt they tried, but I doubted they would succeed without major changes to their
ethics. Finnigan (2015) shows just how hard it would be to pull their rabbit out of such a hat.
2 The biographical details found in Candrakīrti are amplified in Tāranātha’s History of Buddhism in
India (Chimpa and Chattopadhyaya 1970, 123 et seq.)
3 Āryadeva’s being a native of Śrī Laṅka may possibly be corroborated by references in the Ceylonese
chronicles Dīpavaṃsa and Mahāvaṃsa to a “Deva” who lived in the second half of the third century
at the time when the Indian Vetullavāda sect of Great Vehicle Buddhism was temporarily implanted
in Śrī Laṅka (Seyfort Ruegg 1981, 50). His Ceylonese origins are to be taken with some caution,
as Chinese sources have him being born into a South Indian brahmin family. It is far from clear
who, if anyone, was a student of Āryadeva, but some Tibetan sources speak of Rāhulabhadra, who
wrote an important Praise of the Perfection of Wisdom (Prajñāpāramitāstotra); others claim that
Rāhulabhadra was the teacher of Nāgārjuna. See Seyfort Ruegg (1981, 54–56) on Rāhulabhadra.
4 A case in point is the Destruction of Errors about Madhyamaka (*Madhyamakabhramaghāta), assigned
to an Āryadeva by Tibetans: this text copiously borrows from the Verses on the Heart of Madhy-
amaka (Madhyamakahṛdayakārikā) and Torch of Dialectics (Tarkajvālā) of Bhāviveka, a celebrated
Mādhyamika who lived in the sixth century (500–570 CE). Another such case is the Compendium on
the Essence of Knowledge (Jñānasārasamuccaya), a text which the Tibetan canon ascribes to Āryadeva,
but which gives the fourfold presentation of Buddhist doctrine typical of the doxographical (siddhānta)
literature, a genre which considerably post-dates the third century (trans. Mimaki 1976). Less clear are
the Hand Treatise (Hastavālaprakaraṇa) and its Commentary (Hastavālaprakaraṇavṛtti), both of which
the Tibetan canon attributes to Āryadeva, although the Chinese canon most likely has it right in attribut-
ing them to the fifth century author Dignāga (trans. Thomas and Ui 1918). Finally, the Chinese canon
has five works attributed to Āryadeva, the longest two of which are known as “Hundred Verse Treatises”:
Taishō 1570, the Extensive Hundred Verse Treatise (Guang bai lun 廣百論), is the last half of the Four
247
Tom J. F. Tillemans
Hundred Verses (Catuḥśataka); Taishō 1569, the Hundred Verse Treatise (Bai lun 百論 = *Śata(ka)
śāstra; trans. Tucci 1929), bears many thematic affinities with the Four Hundred Verses but is nonethe-
less a different work and not simply an extract of the Four Hundred. The Hundred Verse Treatise became
a key text of the Chinese Madhyamaka school known as the “Three Treatise” (San lun 三論) tradition
(May 1979; May and Mimaki 1979). If the identification of the author Qing mu 青目 with Āryadeva
should be correct, then the author of the Four Hundred Verses may have also written a commentary on
Nāgārjuna’s Verses on the Philosophy of the Middle (May 1979, 481; Tillemans 1990, 7, fn. 17).
5 The Four Hundred Verses has been translated into English in Lang (1986), who uses the Sanskrit
fragments and the Tibetan; Sonam (2008) translates the Four Hundred Verses and rGyal tshab rje’s
commentary from the Tibetan; chapters nine–sixteen are translated into Italian in Tucci (1925) based
on the Sanskrit fragments, Tibetan, and Chinese (Taishō 1570). Lang (2003) introduces and translates
the first four chapters along with Candrakīrti’s commentary; May (1980, 1981a, 1981b, 1982, 1984)
provide a French translation of chapter nine of the Four Hundred Verses and Candrakīrti’s commen-
tary; Tillemans (1990) contains a translation of chapters twelve and thirteen, along with Candrakīrti’s
and Dharmapāla’s commentaries; Johnson (2012) translates chapter fifteen and Candrakīrti’s com-
mentary. An edition of the Sanskrit fragments and their corresponding Tibetan translation is found in
Suzuki (1994), following upon the works of Haraprasād Shāstrī (1914) and Bhattacharya (1931).
6 Tibetan tradition identifies major Mādhyamika authors – i.e., Nāgārjuna, Āryadeva, and Candrakīrti – with
later tantrikas having the same name. In particular, the Mādhyamika Nāgārjuna is identified with tantric
Nāgārjunapāda, also known as Ārya Nāgārjuna (c. seventh century CE); Āryadeva is said to be the disci-
ple of Ārya Nāgārjuna and the author of the Lamp that Integrates the Practices (Caryāmelāpakapradīpa;
trans. Wedemeyer 2007), which furthers the Ārya (noble) tradition of exegesis on the Tantra of the Eso-
teric Community (Guhyasamājatantra); he is also said to be the author of the Treatise on the Purity
of Mind (Cittaviśuddhiprakaraṇa) connected with Ārya Nāgārjuna’s Five Stages (Pañcakrama). To be
clear, however, one cannot follow the Tibetan tradition in attributing tantric texts to the same Mādhyamika
Āryadeva; hence, one badly needs a differentiation between at least two individuals, Āryadeva I being the
Mādhyamika and Āryadeva II a much later tantric author (Mimaki 1987; Tillemans 1990, 6–7).
7 For the canonical schema of four illusions (viparyāsa), see Abhidharmakośabhāṣya ad V.8, transla-
tion La Vallée Poussin (1971, tome IV, p. 21).
8 Tibetan text in Lindtner (1990, 114).
9 There is a tension here in that while Āryadeva practices analysis and uses argumentation adroitly, he
insists that the Buddhist’s invoking emptiness (śūnyatā) is somehow not simply a move in debate:
“The tathāgatas [= buddhas] did not state this [idea that things do not have real intrinsic natures] for
the sake of debate (vāda). Nonetheless, it burns up rival assertions, just as fire [burns up] fuel” (Four
Hundred Verses 12, verse 15; see Tillemans 1990, vol. 1, 128, vol. 2, 46–47).
10 Vigrahavyāvartanī 29, cited in Candrakīrti’s Prasannapadā (ed. La Vallée Poussin 1970, 16.7–10).
11 For more on the three natures, see Chapter 16 and Chapter 22 in this volume.
12 There is also a rather significantly flawed translation in Keenan (1997).
13 Taishō 1571, 246a: 若法本性空見空有何德虛妄分別縛證空見能除. Our translation.
14 The Tibetan is: gal te rang bzhin gyis dngos yod//stong mthong yon tan ci zhig yod//rtog pas mthong
ba ‘ching ba ste//de ni ‘di ru dgag par bya//. If the Sanskrit text translated into Tibetan had used
abhūtaparikalpa, one would have expected something more along the lines of the usual equivalent,
yang dag pa ma yin pa’i kun tu rtog pa, instead of simply rtog pa. See Gadjin Nagao, Index to the
Mahāyāna-sūtrālaṃkāra. Part One. Tokyo, 1958, s.v. abhūtaparikalpa.
15 Interestingly, the broad lines of the polemic against a purely negative ultimate truth are also found
in Tibetan Mādhyamika writers like Go rams pa bSod nams seng ge (1429–1489), but without the
allegiance to Yogācāra’s three-nature theory.
16 Such is the position of Candrakīrti in his commentary to Four Hundred Verses, especially in his
arguments in chapter thirteen against the logicians’ misplaced attempts to revise worldly epistemic
standards (see Tillemans 1990, vol. 1, 176–79).
17 Contrary to how Mahāyāna has been typically depicted, it was not a predominantly egalitarian lay
movement (see Boucher 2008, 50–52; Nattier 2003, 96–100).
18 See Cabezón (2017, chapter 4). While there certainly are tantras, as well as a few sūtras (e.g., the
Mahāyānopadeśa or Vimalakīrtinirdeśa), that challenge misogyny, that challenge was not the norm
in Indian Buddhism. It should be noted, in this vein, that Theravāda and some Mahāyāna Buddhists
practiced an extreme contemplation on disfigured female bodies in morgues and cemeteries as a way
of cultivating renunciation (see Cabezón 2017, 227 et seq.; Wilson 1996). The practice is known as
248
Āryadeva
aśubhabhāvanā, “meditation on the hideous,” and is for the edification of men to impress upon them
the actual repulsive nature of the female physique. It continues to this day. Buddhist canonical litera-
ture has no remotely similar place for an aśubhabhāvanā on the male physique.
19 Cowherds (2011, vi, 15, emphasis in original) sets out a book-length treatment of four questions, one
of which is formulated as “What is true about customary truth?” or “What kind of truth is it, and what
sort of truth theory would best fit it?”
20 For example, Candrakīrti uses the phrase “concerning suprasensible states of affairs” (atīndriyeṣv
artheṣu) in connection with Āryadeva’s 12.5. Dharmapāla ad 12.5, in Chinese, seems to gloss only
emptiness (śūnyatā) as yi 義 (real fact, state of affairs, artha), preferring shi 事 (phenomenal facts,
things) for all that is customary, including imperceptible things, i.e., shen shi 深事, parokṣa.
21 Translation and text of 12.5 and Candrakīrti’s commentary in Tillemans (1990, vol. 1, 120 and vol. 2,
17–19). For Dharmapāla, Tillemans (1990, vol. 1, 91–92 and vol. 2, 132–33).
22 Tillemans (1999, 42). Cf. Dharmakīrti’s Svavṛtti to 1.318 (ed. Gnoli 1960, 167): na kvacid askhalita
iti sarvaṃ tathā/vyabhicāradarśanāt, “It is not the case that when one is unmistaken on something,
all the rest is similarly [unmistaken], for we see that this [implication] is deviant.”
23 Cf. Pramāṇavārttika 4.106 (Tillemans 2000, 148–50). Human understanding of subtle karmic facts
and the like invariably depends on “acceptance” (abhyupagama), i.e., on having previously accepted
(iṣṭa) scriptures that talk about them as reliable, because otherwise “no [human] deliberation will
occur at all (na cintaiva pravartate).”
24 Cf. what is known as “Kantian humility,” i.e., human beings’ inevitable incapacity to understand
things in themselves (Ding an sich). See Langton (1998; 2004).
References
Bhattacharya, Vidhushekhara. 1931. The Catuḥśataka of Āryadeva: Sanskrit and Tibetan Texts with Copious
Extracts from the Commentary of Candrakīrti [Chapters 8–16.] Part 2. Calcutta: Visvabharati Bookshop.
Bhikkhu Bodhi. 2012. The Numerical Discourses of the Buddha: A Translation of the Aṅguttara Nikāya.
Boucher, Daniel. 2008. Bodhisattvas of the Forest and the Formation of Mahāyāna: A Study and Transla-
tion of the Rāṣṭrapālaparipṛcchāsūtra. Honolulu: University of Hawaii Press.
Cabezón, José I. 2008. “Buddhist Narratives of the Great Debates.” Argumentation 22: 71–92.
———. 2017. Sexuality in Classical South Asian Buddhism. Studies in Indian and Tibetan Buddhism.
Chimpa, Lama and Alaka Chattopadhyaya. 1970. Tāranātha’s History of Buddhism in India. Simla:
Indian Institute of Advanced Study.
Cowherds (= Georges Dreyfus, Bronwyn Finnigan, Jay L. Garfield, Guy Newland, Graham Priest, Mark
Siderits, Koji Tanaka, Sonam Thakchoe, Tom Tillemans, and Jan Westerhoff). 2011. Moonshadows:
Conventional Truth in Buddhist Philosophy. New York: Oxford University Press.
Finnigan, Bronwyn. 2015. “Madhyamaka Buddhist Meta-ethics: The Justificatory Grounds of Moral
Judgments.” Philosophy East and West 65 (3): 765–85.
———. 2018. “Madhyamaka Ethics.” In The Oxford Handbook of Buddhist Ethics, edited by Daniel
Cozort and James Mark Shields, 162–83. Oxford: Oxford University Press.
Gnoli, Raniero. 1960. The Pramāṇavārttikam of Dharmakīrti: The First Chapter with the Autocommentary:
Text and Critical Notes. Serie Orientale Roma 23. Rome: Istituto Italiano per il Medio ed Estremo Oriente.
Gokhale, Vasudev Vishwanath and Budhiruci. 1930. Akṣara-Çatakam: The Hundred Letters, a Madhy-
amaka Text. After Chinese and Tibetan Materials Translated by Vasudev Gokhale. Materialen zur
Kunde Buddhismus 14. Heidelberg: O. Harrassowitz.
Haraprasād Shāstrī. 1914. “Catuḥśatika of Ārya Deva.” Memoirs of the Asiatic Society of Bengal 3 (8):
449–514 [Sanskrit fragments of CS and CST].
Hoornaert, Paul. 2004. “The Dharmapāla-Bhāvaviveka Debate as Presented in Dharmapāla’s Commentary
to Catuḥśataka XVI.23.” Kanazawadaigaku bungakubu ronshū, Kōdōkagaku tetsugaku hen 24 – Studies
and Essays, Behavioural Sciences and Philosophy, Faculty of Letters. Kanazawa University 24: 119–49.
Johnson, Dennis A. 2012. “Refuting the Conditioned – The Saṃskṛtārthapratiṣedha of Candrakīrti’s
Catuḥśatakaṭīkā: Introduction, Translation and Summary Together with a Critical Edition of the
Tibetan Translation.” MA Thesis, University of Vienna. Accessed January 18, 2019. https://othes.
univie.ac.at/20891/ [Translation of chapter 15 of CS and CST].
249
Tom J. F. Tillemans
Kajiyama, Yūichi. 1968/1969. “Bhāvaviveka, Sthiramati and Dharmapāla.” Wiener Zeitschrift für die
Kunde Süd- und Ostasiens und Archiv für Indische Philosophie 12–13: 193–203.
Keenan, John P. 1997. Dharmapāla’s Yogācāra Critique of Bhāvaviveka’s Mādhyamika Explanation
of Emptiness. The Tenth Chapter of Ta-ch’eng Kuang Pai-lun Shih Commenting on Āryadeva’s
Catuḥśataka Chapter Sixteen. Lewiston: Edwin Mellen.
La Vallée Poussin, Louis de. 1970. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna, avec le
commentaire de Candrakīrti. First published in St. Petersburg, Bibliotheca Buddhica IV, 1903–1913.
Reprinted Osnabrück: Biblio Verlag.
———. 1971. L’Abhidharmakośa de Vasubandhu: traduction et annotations, 6 vols. Paris: Geuthner.
Louvain: J.-B. Istas, 1923–31. New edition prepared by É. Lamotte, Mélanges chinois et bouddhiques
16. Brussels: Institut Belge des Hautes Études Chinoises.
Lang, Karen C. 1986. Āryadeva’s Catuḥśataka: On the Bodhisattva’s Cultivation of Merit and Knowl-
edge. Indiske Studier VII. Copenhagen: Akademisk Forlag [Translation of CS].
———. 2003. Four Illusions:. Candrakīrti’s Advice to Travelers on the Bodhisattva Path. Translation
and Introduction by Karen C. Lang. New York: Oxford University Press [Translation of chapters 1–4
of CS and CST].
Langton, Rae. 1998. Kantian Humility: Our Ignorance of Things in Themselves. New York: Oxford
University Press.
———. 2004. “Elusive Knowledge of Things in Themselves.” Australasian Journal of Philosophy 82
(1): 129–36.
Lindtner, Christian. 1982 [1990]. Nāgārjuniana. Studies in the Writings and Philosophy of Nāgārjuna.
Indiske Studier 4. Copenhagen: Akademisk Forlag. Reprinted Delhi: Motilal Banarsidass.
Luo, Hong. 2018. “Is Ratnākaraśānti a gZhan stong pa?” Journal of Indian Philosophy 46: 577–619.
May, Jacques. 1979. “Chūgan.” In Hōbōgirin, Dictionnaire encyclopédique du bouddhisme d’après les
sources chinoises et japonaises V, edited by Sylvain Lévi, J. Takakusu, Paul Demiéville, Hubert Durt
et al., 470–93. Paris: Librairie d’Amérique et d’Orient. Tokyo: Maison franco-japonaise.
———. 1980. “Āryadeva et Candrakīrti sur la permanence (I).” In Indianisme et Bouddhisme, Mélanges
offerts à Mgr Étienne Lamotte, 214–32. Louvain: Publications de l’Institut orientaliste de Louvain
[French translation of chapter 9 of CS and CST].
———. 1981a. “Āryadeva et Candrakīrti sur la permanence (II).” Bulletin de l’École française
d’Extrême-Orient 69: 75–96.
———. 1981b.“Āryadeva et Candrakīrti sur la permanence (III).” Études Asiatiques 35 (2): 47–76.
———. 1982. “Āryadeva et Candrakīrti sur la permanence (IV).” Études de Lettres, University of Lau
sanne 3: 45–76.
———. 1984. “Āryadeva et Candrakīrti sur la permanence (V).” In Acta Indologica, Studies of Mysti-
cism in Honor of the 1150th Anniversary of Kōbō Daishi’s Nirvāṇam, 115–44. Narita: Naritasan
Shinshoji.
May, Jacques, and Katsumi Mimaki. 1979. “Chūdō.” In Hōbōgirin, Dictionnaire encyclopédique du
bouddhisme d’après les sources chinoises et japonaises V, edited by Sylvain Lévi, J. Takakusu, Paul
Demiéville, Hubert Durt et al., 456–70. Paris: Librairie d’Amérique et d’Orient. Tokyo: Maison
franco-japonaise.
Mimaki, Katsumi. 1976. La Réfutation bouddhique de la permanence des choses (sthirasiddhidūṣaṇa) et
la preuve de la momentanéité des choses (kṣaṇabhaṅgasiddhi). Publications de l’Institut de Civilisa-
tion Indienne, fascicule 41. Paris: Institut de Civilisation Indienne.
———. 1987. “Āryadeva.” In The Encyclopedia of Religion, edited by Mircea Eliade et al., vol. 1,
431a–432a. New York: Macmillan Publishing.
Nattier, Jan. 2003. A Few Good Men. Honolulu: University of Hawaii Press.
Oetke, Claus. 2003. “Some Remarks on Theses and Philosophical Positions in Early Madhyamaka.”
Robinson, Richard. 1967 [1976]. Early Mādhyamika in India and China. Madison: University of Wis-
consin Press. Reprinted by Motilal Banarsidass, New Delhi.
Seyfort Ruegg, David. 1981. The Literature of the Madhyamaka School of Philosophy in India. Wies-
baden: O. Harrassowitz.
———. 2010. The Buddhist Philosophy of the Middle: Essays on Indian and Tibetan Madhyamaka.
Studies in Indian and Tibetan Buddhism. Somerville MA: Wisdom Publications.
250
Āryadeva
Sonam, Ruth. 2008. Āryadeva’s Four Hundred Stanzas on the Middle Way, with Commentary by Gyel-
tsap. Translated by Ruth Sonam with additional commentary by Geshe Sonam Rinchen. Ithaca, NY:
Snow Lion Publications [Translation of CS, CST and rGyal tshab rje’s commentary].
Sponberg, Alan. 1992. “Attitudes toward Women and the Feminine in Early Buddhism.” In Buddhism,
Sexuality, and Gender, edited by José I. Cabezón, 3–36. Albany: State University of New York Press.
Strawson, Peter F. 1959. Individuals: An Essay in Descriptive Metaphysics. London: Methuen.
Suzuki, Kōshin. 1994. Sanskrit Fragments and Tibetan Translation of Candrakīrti’s
Bodhisattvayogācāracatuḥśatakaṭīkā. Tokyo: Sankibo [Sanskrit fragments of CS and CST with their
corresponding Tibetan].
Thomas, Frederick W., and Hakuju Ui. 1918. “The Hand Treatise, A Work of Āryadeva.” Journal of the
Royal Asiatic Society: 267–310.
Tillemans, Tom J. F. 1990. Materials for the Study of Āryadeva, Dharmapāla and Candrakīrti. The
Catuḥśataka of Āryadeva, Chapters XII and XIII, with the Commentaries of Dharmapāla and
Candrakīrti: Introduction, Translation, Sanskrit, Tibetan and Chinese Texts, Notes. 2 vols. Wiener
Studien zur Tibetologie und Buddhismuskunde 24.1 & 24.2. Vienna: Arbeitskreis für Tibetische und
Buddhistische Studien Universität Wien. Reprinted in one volume by Motilal Banarsidass Publishing
Co., New Delhi 2008 [Translation of chapters 12 and 13 of CS, CST and GBLSL].
———. 1999. Scripture, Logic, Language: Essays on Dharmakīrti and His Tibetan Successors. Studies
in Indian and Tibetan Buddhism. Somerville, MA: Wisdom Publications.
———. 2000. Dharmakīrti’s Pramāṇavārttika: An Annotated Translation of the Fourth Chap-
ter ( parārthānumāna), vol. 1 (k. 1–148). Vienna: Verlag der Österreichischen Akademie der
Wissenschaften.
———. 2016. How Do Mādhyamikas Think? And Other Essays on the Buddhist Philosophy of the Mid-
dle. Studies in Indian and Tibetan Buddhism. Somerville, MA: Wisdom Publications.
———. 2019. “Deflating the Two Images and the Two Truths: Bons baisers du Tibet.” In Wilfrid Sellars
and Buddhist Philosophy: Freedom from Foundations, edited by Jay L. Garfield, 80–96. Routledge
Studies in American Philosophy. New York and London: Routledge.
Tucci, Giuseppe. 1925. “Studi Mahāyānici. La versione cinese del Catuḥçataka di Āryadeva confrontata
col testo sanscrito e la traduzione tibetana.” Rivista degli Studi Orientali 10: 521–90 [Italian transla-
tion of chapters 9–16 of CS and some extracts of GBLSL].
———. 1929. Pre-Diṅnāga Buddhist Texts on Logic from Chinese Sources. Gaekwad’s Oriental Series,
no. XLIX. Baroda: The Oriental Institute [Translation of Āryadeva’s Śatakaśāstra (Bai lun), Taishō
1569].
Wedemeyer, Christian K. 2007. Āryadeva’s Lamp that Integrates the Practices (Caryāmelāpakapradīpa):
The Gradual Path of Vajrayāna Buddhism According to the Esoteric Community Noble Tradition.
American Institute of Buddhist Studies. New York: Columbia University Press.
Wilson, Liz. 1996. Charming Cadavers: Horrific Figurations of the Feminine in Indian Buddhist Hagi-
ographic Literature. Chicago: University of Chicago Press.
251
15
ASAṄGA
Great Systematizer of Yogācāra Thought
Jowita Kramer
The Indian scholar-monk Asaṅga (Chinese: Wuzhuo 無著; Tibetan: Thogs med) was one of
the most influential masters of the Yogācāra tradition. Almost all information concerning his
life and works available to us today derives from Chinese and Tibetan sources whose historic-
ity is uncertain. Most modern scholars date Asaṅga to the fourth or fifth century.1
Asaṅga’s Life
The four main traditional sources that provide details of Asaṅga’s life and works are the Biogra-
phy of the Master of the Doctrine Vasubandhu (Posoupandou fashi zhuan 婆藪槃豆法師傳, T
2049) by Paramārtha (Zhendi 眞諦; 499–569), the Records of the Western Regions (Datang xiyuji
大唐西域記, T 2087) by Xuanzang 玄奘 (600[?]–664), the History of the Doctrine (Chos ’byung)
by Bu ston (1290–1364), and the History of the Doctrine in India (rGya gar chos ’byung) by
Tāranātha (1575–1634).2 The oldest of these texts is the Biography of the Master of the Doc-
trine Vasubandhu (sixth century), according to which Asaṅga – who is said by Paramārtha to be
Vasubandhu’s brother – was born in Puruṣapura (located in Gandhāra; present-day Peshawar) as
the eldest son of a brahmin and ordained as a monk in the Sarvāstivāda tradition early in his life.
Unsatisfied with the teachings, he is said to have visited Tuṣita heaven by means of his spiritual
power, where he received instruction in the Mahāyāna sūtras from the future Buddha Maitreya.
Moreover, Paramārtha explains that Maitreya also came down to earth in order to teach the Foun-
dation of the Yoga Practitioners (Yogācārabhūmi) to Asaṅga, who subsequently composed sev-
eral commentaries on the Mahāyāna sūtras Maitreya had explained to him. Paramārtha also states
that Vasubandhu later became a Mahāyāna proponent himself after Asaṅga, under the pretext of
an illness, had Vasubandhu visit him and attributed the cause of his illness to Vasubandhu speak-
ing poorly of the Mahāyāna.3
The second important biographical source, Xuanzang’s travelogue Records of the West-
ern Regions, additionally states that Asaṅga was active in Ayodhyā and that his disciple was
Buddhasiṃha. Furthermore, Asaṅga is said to have received not only the Foundation of the
Yoga Practitioners but also the Ornament of Mahāyāna Sūtras (Mahāyānasūtrālaṃkāra,
Dasheng zhuangyanjing lun 大乘莊嚴經論) and the Distinguishing the Middle from the
Extremes (Madhyāntavibhāga, Bian zhong bian lun song 辯中邊論頌) from Maitreya. Unlike
Paramārtha, Xuanzang assigns Asaṅga to the Mahīśāsaka school, not the Sarvāstivāda.4 Some
252 DOI: 10.4324/9781351030908-24

Asaṅga
modern scholars follow Xuanzang’s assumption and state that Asaṅga’s Abhidharma scriptures
are based on many Mahīśāsaka ideas. However, others argue that there is not enough ground to
determine which school affiliation Asaṅga had before he became a Mahāyāna teacher.5
In contrast to Chinese sources, Asaṅga’s Tibetan biographies do not speak of a con-
version from Mainstream to Mahāyāna Buddhism. Bu ston reports that Asaṅga was the
son of a woman named Gsal ba’i tshul khrims (reconstructed as “Prasannaśīlā” in Ober-
miller [1932, 37]), who wanted to prevent the downfall of the Buddhist doctrine through
the birth of three extraordinary sons. According to Bu ston’s report, Asaṅga was the son
of this woman and a “warrior” (kṣatriya) father, while his (half-)brother Vasubandhu is
said to have been the son of a brahmin. Further, Bu ston reports that Asaṅga, after many
years of unsuccessful practice, which he began at a cave in Ri bya rkang (reconstructed
as “Kukkuṭapādaparvata” in Obermiller [1932, 37]) and through which he intended to
propitiate Maitreya, met a sick dog whose body was partially eaten by worms. In order
to free the dog from the worms, Asaṅga is said to have carved a slice of meat from his
own body. But at the very moment he wanted to use his tongue to remove the worms from
the dog’s body and put them on his amputated flesh, the dog disappeared and Maitreya
appeared in his place. Maitreya is said thereafter to have taken Asaṅga to Tuṣita heaven,
where he instructed him in various Mahāyāna sūtras, the Foundation of the Yoga Prac-
titioners and the “five treatises of Maitreya” (i.e., according to the Tibetan tradition, the
Ornament of Clear Realization [Abhisamāyālaṃkāra], the Ornament of Mahāyāna Sūtras,
the Distinguishing the Middle from the Extremes, the Distinguishing the Factors from
Their True Nature [Dharmadharmatāvibhāga] and the Distinguishing the Germ of the
[Three] Jewels [Ratnagotravibhāga]).6 On another occasion, Bu ston states that Asaṅga
composed, after his return to earth, the Foundation of the Yoga Practitioners, the Com-
pendium of the Abhidharma (Abhidharmasamuccaya), the Summary of the Mahāyāna
(Mahāyānasaṃgraha),7 the Determination of Reality (*Tattvaviniścaya, a commentary on
the Ornament of Clear Realization; Delhey 2019, 75), and commentaries on the Distin-
guishing the Germ of the [Three] Jewels and the Sūtra of the Explanation of the Profound
Meaning (Saṃdhinirmocanasūtra). He moreover is said to have written down the “five
treatises of Maitreya.”8 Bu ston also tells the story of how Asaṅga converted Vasubandhu
to the Mahāyāna, after the latter had sarcastically remarked: “Alas! Though Asaṅga medi-
tated for twelve years in the forest, instead of attaining success in his meditation, he has
composed a work useless in sense but heavy enough to be an elephant’s load.”9
In his History of the Doctrine in India, Tāranātha presents a similar story of the life and
works of Asaṅga, but adorns it with more details, relating, for example, that Asaṅga estab-
lished a residence called “Dharmāvikuravihāra” in “Veluvana,” Magadha. There he is said
to have written down the five works of Maitreya, as well as composed the Foundation of the
Yoga Practitioners, the Summary of the Mahāyāna, the Compendium of the Abhidharma, and a
commentary on the Ornament of Clear Realization. Tāranātha further reports that under King
Gambhīrapakṣa, there was a monastic gathering in the monastery of “Uṣmapuri” in the city of
“Sagari,” “near the west,” where Asaṅga gave teachings. The king is also said to have invited
Asaṅga to his court to test his abilities through philosophical questions about the Sūtra of the
Perfection of Wisdom (Prajñāpāramitāsūtra). The king is said to have been so pleased with
Asaṅga’s answers that he had 25 Mahāyāna monasteries built. During this time, Asaṅga is also
supposed to have converted to the Mahāyāna his brother Vasubandhu, who had previously
commented upon Asaṅga’s Foundation of the Yoga Practitioners with the sarcastic remark
mentioned already by Bu ston.10
253
Jowita Kramer
Furthermore, Tāranātha ascribes to Asaṅga the establishment of monastic centers in

“Kṛṣṇarāja,” “Urgyana,” and Magadha. When Asaṅga taught near Ayodhyā, he and his fol-
lowers are said to have been attacked by an army of fighters. After Asaṅga instructed eve-
ryone to continue with contemplative practice and not to be disturbed by the attackers, they
withdrew without harming anyone. Tāranātha also states that Asaṅga taught not disrespect-
ing any of the Buddhist traditions, for which reason the adherents of mainstream Buddhism
(śrāvaka) respected him and studied the sūtras and the Abhidharma with him. In the later
part of his life, Asaṅga is said to have spent twelve years in Nālandā and to have died in
*Rājagṛha, where his disciples are said to have erected a shrine (caitya) with his relics.11
In other parts of his report, Tāranātha names *Buddhadasa and *Dharmadasa as Asaṅga’s
disciples.12
The questions of whether Maitreya(nātha) (Chinese Cizun 慈尊; Tibetan Byams pa, Byams
pa mngon po) is to be considered a historical person or a deity, and whether the works Asaṅga
received from him should be regarded as his or as the works of a separate author, have sparked
much discussion among contemporary scholars. The traditional Chinese and Tibetan accounts
do not agree on this question, either.13
Asaṅga’s Works
In total, the following nineteen works are traditionally associated with Asaṅga as primary
author:
1. Foundation of the Yoga Practitioners (Yogācārabhūmi)14

2. Summary of the Mahāyāna (Mahāyānasaṃgraha)15
3. Compendium of the Abhidharma (Abhidharmasamuccaya)16
4. Treatise on Propagating the Teaching of the Noble One (Xianyang shengjiao lun 顯揚聖
教論)17
5. Treatise on Six Aspects of Meditation Instruction (Liumen jiaoshou xiding lun 六門教授
習定論)18
6. Treatise on the Perfection of Wisdom That Cuts [Like] a Vajra (Jingang bore lun 金剛般
若論, *Vajracchedikā Prajñāpāramitāśāstra)19
7. Treatise Conforming with the Middle (Shunzhong lun 順中論)20
8. Seventy Stanzas on the Perfection of Wisdom in Three Hundred Stanzas (Triśatikāyāḥ
Prajñāpāramitāyāḥ Kārikāsaptatiḥ)21
9. Hymn on the Qualities Based on the Body [Consisting in the True Reality] of Phenomena
(Chos kyi sku la gnas pa’i yon tan la bstod pa, *Dharmakāyāśrayāsāmānyaguṇastotra),22
10. Meditation Manual of the Perfection of Wisdom (Shes rab kyi pha rol tu phyin ma’i sgrub
thabs, *Prajñāpāramitāsādhana),23
11. Noble Maitreya Meditation Manual (’Phags pa byams pa’i sgrub thabs,
*Āryamaitreyasādhana),24
12. Noble Commentary on [the Sūtra of] the Explanation of the Profound Meaning (’Phags
pa dgongs pa nges par ’grel pa’i rnam par bshad pa, *Āryasaṃdhinirmocanabhāṣya),25
13. Commentary on the Recollection of the Buddha (Sangs rgyas rjes su dran pa’i ’grel pa,
*Buddhānusmṛtivṛtti)
14. Commentary on the Recollection of the Doctrine (Chos rjes su dran pa’i ’grel pa,
*Dharmānusmṛtivṛtti)
15. Commentary on the Recollection of the Monastic Community (Dge ’dun rjes su dran pa’i
bshad pa, *Saṃghānusmṛtivyākhyā)26
254
Asaṅga
16. Commentary on Distinguishing the Germ of the [Three] Jewels (*Ratnagotravibhāgavyākhyā)27

17. Instruction Called Lamp of Absorption (bSam gtan gyi sgron ma zhes bya ba’i man ngag,
*Dhyānadīpanāmopadeśa)28
18. A (lost) commentary on the Ornament of Clear Realization (Abhisamayālaṃkāra)29
19. Commentary on the Ornament of Mahāyāna Sūtras (Mahāyānasūtrālaṃkārabhāṣya)30
In trying to evaluate Asaṅga’s contribution to Buddhist philosophy, we first face the difficulty
that it remains controversial which works can actually be attributed to him as their original
author. Second, with regard to many Yogācāra texts, as with earlier Buddhist texts in general,
there is often no consensus regarding the chronological order. This means that for many texts
and passages, we can know neither exactly who wrote them nor which of the texts’ passages
that contain a particular teaching are to be considered the earlier.
Among the nineteen texts ascribed to Asaṅga, only the Compendium of the Abhidharma,
the Summary of the Mahāyāna, and the Treatise on Propagating the Teaching of the Noble
One are widely accepted by modern scholars to have been composed by a single author called
Asaṅga. In contrast, most modern scholars take the Foundation of the Yoga Practitioners to be
not the work of an individual author but rather a compilation that grew into its present shape
over the course of many decades or even centuries. If Asaṅga is to be associated with the
Foundation of the Yoga Practitioners, then it is only as its compiler, not as its original author.
As Lambert Schmithausen notes, the text appears to be a compilation of different strands and
materials that have been transmitted by an early Yogācāra community, which seems not to
have been entirely homogeneous (Schmithausen 2014, 11).
The Treatise on Propagating the Teaching of the Noble One31 is a summary of the teachings of
the Foundation of the Yoga Practitioners, which is partially reproduced verbatim (see Delhey 2019,
78). The work is only available in Xuanzang’s Chinese translation. Most modern scholars accept
the attribution of this work to Asaṅga, some of them assuming that only the verses were composed
by this author, while regarding the prose as the work of Vasubandhu (Delhey 2019, 74). The text
seems to be closely related to the Compendium of the Abhidharma (von Rospatt 1995, 85).
The authorship of the other fifteen works ascribed to Asaṅga remains doubtful.32 While the short
commentary on the Sūtra of the Explanation of the Profound Meaning might indeed have been
written by him, it can be stated with some certainty that the same Asaṅga who composed the three
Yogācāra works just mentioned is not the author of the two tantric ritualistic meditation manuals
(sādhana) ascribed to Asaṅga (Delhey 2019, 79). It is also not very probable that this Asaṅga is the
composer of the commentaries on the Distinguishing the Germ of the [Three] Jewels and the Root
Verses on the Middle Way (Mūlamadhyamakakārikā) (i.e., the Treatise Conforming with the Mid-
dle), which shows no relation to Yogācāra thought, as well as the two works related to the perfec-
tion of wisdom ( prajñāpāramitā).33 The two meditation manuals, the Instruction Called Lamp of
Absorption and Treatise on Six Aspects of Meditation Instruction, as well as the three works on the
recollection of the Buddha, the doctrine (dharma), and the monastic community (saṃgha) have not
been investigated thoroughly enough in modern scholarship to come to any final conclusions regard-
ing their authorship. The hymn (stotra) on the “body consisting in the true reality of phenomena”
(dharmakāya) was probably extracted from the Ornament of Mahāyāna Sūtras, while the commen-
tary Illumination of the Ornament of Clear Realization (Abhisamayālaṃkārālokā) is not available
to us and therefore nothing can be said about its composer (Delhey 2019, 78). The authorship of the
Commentary on the Ornament of Mahāyāna Sūtras is commonly ascribed to Vasubandhu.
The Yogācāra tradition is characterized by a number of specific concepts. The one that
has probably attracted most attention and is subject to controversial debates until the present
day is the doctrine of “nothing but presentation” (vijñaptimātra). According to this tenet, the
255
Jowita Kramer
whole world is nothing but mind (cittamātra), and external objects exist not as material enti-
ties but only as mental manifestations. The second central theory of the Yogācāras is the doc-
trine of eight forms of mind, including the five sense perceptions – already propagated in the
early Buddhist schools – plus the mental perception (manovijñāna), the defiled notion [of “I”]
(kliṣṭaṃ manas), and the store mind (ālayavijñāna) (Schmithausen 2018, 264). The third doc-
trine for which the Yogācāras are famed is the set of three natures (svabhāva) or characteristics
(lakṣaṇa) describing three aspects of all phenomena.34 The first refers to the assumption that
we superimpose a false, conceptualized ( parikalpita) nature on the phenomena we perceive,
which does not reflect their real existence. The second aspect of all entities is that their nature
is “dependent on causes and conditions” ( paratantra). The final aspect describes the perfect
( pariniṣpanna) nature of phenomena, which means that, in actuality, manifestations are com-
pletely free of the conceptualizations ascribed to them.
Another doctrine that was developed in Yogācāra circles is the concept of the three bodies
(kāya) of a Buddha. The body through which a Buddha makes his appearance in the world
is called the “manifestation body” (nirmāṇakāya), while the one through which he teaches
advanced Bodhisattvas in the pure lands is referred to as his “enjoyment body” (saṃbhogakāya).
The true reality realized by a Buddha is his “essential body” (svābhāvikaḥ kāyaḥ) or his “body
consisting in the true reality of phenomena” (dharmakāya).35 The concept of the three bodies
of the Buddha seems to appear for the first time in the Ornament of Mahāyāna Sūtras.36
The question of whether the three Yogācāra tenets of “nothing but presentation”
(vijñaptimātratā), “store mind” (ālayavijñāna), and the three “natures” (svabhāva) are inextri-
cably connected has given rise to a number of controversial discussions. Some scholars assume
that they were developed interdependently and complemented each other from the very begin-
ning.37 In contrast, Lambert Schmithausen tried to show in several of his publications (1987,
2014, 2018) that such a logical inseparability of the three concepts is neither necessary nor
reflected in early Yogācāra sources. In most of the earlier Yogācāra texts, the three tenets are
treated in separate contexts, and it is often the case that one is even completely absent (Sch
mithausen 2018, 263). According to Schmithausen (2018, 266), the concepts of the store mind
and the three natures originate from two different sections of the Foundation of the Yoga Prac-
titioners. There, the tenet of “nothing but presentation” (vijñaptimātratā) is not mentioned
except for a quotation from the Sūtra of the Explanation of the Profound Meaning, which
was almost certainly added at the very end of the compilation process of the Foundation of
the Yoga Practitioners (Schmithausen 2014, 31) – and may indeed be the text in which the
idea of “nothing but presentation” (vijñaptimātra) occurs for the first time.38 Notably, the store
mind (ālayavijñāna) is not mentioned in what we might consider the “Māhāyanistic” parts of
the Foundation of the Yoga Practitioners, such as the section called Stage of the Bodhisattva
(Bodhisattvabhūmi), but only in the “traditional” parts of the text (Schmithausen 1987, 98ff.).
In the treatises Ornament of Mahāyāna Sūtras, Distinguishing the Middle from the Extremes,
and Distinguishing the Factors from Their True Nature, which are commonly ascribed to
Maitreya(nātha), the store mind is not mentioned at all. The doctrine of “mind only” occurs in
the Ornament of Mahāyāna Sūtras in the form of cittamātra (not vijñaptimātra).
If we assume that Asaṅga is not the original author of the Foundation of the Yoga Practi-
cioners (and, if at all, only its compiler) or of the three works ascribed to Maitreya(nātha), but
that he is the composer of the Compendium of the Abhidharma, Summary of the Mahāyāna,
and the Treatise on Propagating the Teaching of the Noble One, we have to conclude that
Asaṅga is not the originator of the central Yogācāra concepts of the store mind, mind only, the
three natures, or the three bodies of a Buddha. Rather, what we can credit to this scholar is the
combination of these doctrines into a coherent system and a nuanced elaboration of that system.
256
Asaṅga
The Compendium of the Abhidharma and the Summary of the Mahāyāna are considered
Asaṅga’s key works by modern scholars (Delhey 2019, 77). Both these texts summarize and
systematize Yogācāra thought and have had a lasting impact on the development of Mahāyāna
Buddhism in India and beyond. According to Dan Martin (2002, 344), the Compendium of
the Abhidharma played an important role in the early transmission of Abhidharma teachings
in Tibet and was only later superseded by the Commentary on the Treasury of Abhidharma
(Abhidharmakośabhāṣya). Compared to the (three) works ascribed to Maitreya (mentioned
previously in this chapter), both the Compendium of the Abhidharma and Summary of the
Mahāyāna show a much closer affiliation with Abhidharmic thought (Schmithausen 1987, 100).
The Compendium of the Abhidharma

The Compendium of the Abhidharma is extant in some Sanskrit fragments, as well as in Tibetan
and Chinese translation.39 One of the notable aspects of the Compendium of the Abhidharma is
that it is a combination of the Abhidharmic thought of mainstream Buddhism with innovative
Yogācāra ideas. Drawing mainly on old sūtras and elements of earlier Yogācāra texts, such as
the Foundation of the Yoga Practitioners, Asaṅga thus creates a distinctive system of Yogācāra
Abhidharma (Kritzer 1999, 13). While the Compendium of the Abhidharma is closely related
to the Foundation of the Yoga Practitioners, it further systematizes and develops its ideas
(Kritzer 1999, vii).40 The theory of the store mind (ālayavijñāna) and the scheme of three natures
(svabhāva) appear already well established in the Compendium of the Abhidharma. At the same
time, as noted by Bayer (2010, 19), the store mind (ālayavijñāna) is mentioned only six times in
the Compendium of the Abhidharma, and only in its first part called Compendium of Characteris-
tics (Lakṣaṇasamuccaya). Remarkably, the theory of “nothing but presentation” (vijñaptimātratā)
is not mentioned at all (Schmithausen 2018, 269ff.). The doctrine of the three bodies (kāya) of a
Buddha, which is absent from the Foundation of the Yoga Practitioners (Deleanu 2006, 227 n.
171), is mentioned in the Compendium of the Abhidharma (Bayer 2010, 22). Another doctrine
that we do not find in the Foundation of the Yoga Practitioners but in the Compendium of the
Abhidharma is the concept of the “non-abiding nirvāṇa” (apratiṣṭhitanirvāṇa), which refers to
the Buddha’s quality of neither being fixed in nirvāṇa nor in cyclic existence (saṃsara).41
Many parts of the Compendium of the Abhidharma seem to follow the traditional doctrine
of mainstream Buddhism in affirming the nonexistence of a permanent self (ātman) and do
not propagate or presuppose the idea of an ultimate unreality of all factors (dharma) or the
nonexistence of external material objects expressed in other Yogācāra texts (Schmithausen
1987, 190). The concept of emptiness (śūnyatā) is also not explained in terms of the nonexist-
ence of the factors of existence, but instead as their lack of a substantial and unchangeable
self. The true essence (tathatā) of these factors, which is equated with this emptiness, is there-
fore – according to the Compendium of the Abhidharma – not their total unreality but the fact
that they do not have such a self (Schmithausen 1972, 155ff.). Notably, the Compendium of
the Abhidharma also explains the three natures (lakṣaṇa = svabhāva) in accordance with this
understanding. The conceptualized (vikalpita) nature is understood as the wrong ascription
of the idea of a self to the five constituents (skandha), the eighteen elements (dhātu), and the
twelve bases (āyatana). The dependent ( paratantra) nature is explained to be the constituents,
elements, and bases themselves, whereas the true nature (dharmatālakṣaṇa = pariniṣpannaḥ
svabhāvaḥ) is defined as the absence of a self in the constituents, elements, and bases.42
As mentioned by Bayer (2010, 20), the term “mind only” (cittamātra) is only used in the
Compendium of the Abhidharma in connection with mental factors, such as feeling (vedanā),
etc., which are assumed to be associated with only the mind (citta) and not with a self (ātman).
257
Jowita Kramer
At the same time, some scholars have pointed out that there are a few statements to be found in
the Compendium of the Abhidharma that seem to indicate the unreality of matter.43 Schmithausen
(1972, 158) refers to a passage stating that atoms ( paramāṇu) are “without a body” (niḥśarīra);
that is, they do not exist as particles of a specific size. Another passage of the text to suggest an
idealistic point of view appears within an explanation of the meditative practice concerned with
realizing that the previously visualized object is nothing but one’s own mind (Bayer 2010, 21).
Although the doctrine of the store mind is mentioned only a few times in the Compendium
of the Abhidharma, it appears in a well-developed shape in the text. However, it is important
to note that it only has the function of a substratum, which ensures that the mind can continue
uninterruptedly (Bayer 2010, 19), and of a container of all mental seeds or the collection of
these seeds (sarvabījaka). We do not find any statements in the Compendium of the Abhi
dharma, which would define the store mind as having the role of cognizing an object.44
One of the remarkable parts of the Compendium of the Abhidharma is the discussion of
“Vaipulya,” a synonym of the Mahāyāna (Skilling 2013, 95), defined by Asaṅga as the “canon
of the Bodhisattvas” (*bodhisattvapiṭaka).45 In this context, the author gives reasons why
some are afraid of these teachings and lists twenty-eight “wrong views” that result in the fact
that some, even though they follow the Mahāyāna, are not liberated by it. Asaṅga also explains
some central statements of the Mahāyāna, such as “all dharmas are without substance” and all
dharmas are “unarisen, undestroyed, at peace from the beginning” (Skilling 2013, 88).
Another notable section is the Determination of the Talk (Sāṃkathyaviniścaya), which dis-
cusses various aspects of “talk” (sāṃkathya), one of them being the topic of “speech” (vāda).46
In addition to forms of everyday talk, such as “common talk” (sarvalokavacana) and “rumor”
( pravāda), Asaṅga deals with various components of a (formal) debate in this passage. He
lists the possible locations (adhikaraṇa), including the royal court (rājakula); the subjects
(adhiṣṭhāna), including the various forms of proofs (sādhana); and the qualities of the debater,
such as the perfection of eloquence (vākkaraṇasampad) and self-confidence (vaiśāradya).
Another remarkable category mentioned in this context is “defeat” (nigraha), which can
involve ceding (tyāga) a debate in recognition of one’s own faults or because of manifest-
ing anger or pride. The text enumerates nine kinds of defects (doṣa) in a discussion, such as
it being confusing (ākula), unmeasured (amita), meaningless (anarthaka), or disconnected
(aprabaddha). According to Asaṅga, there is also the possibility of withdrawing (niḥsaraṇa)
from the debate – for instance, after having realized the opponent’s inability, the assembly’s
inferiority, or one’s own incompetence. Finally, Asaṅga states that one should only engage in
a debate in order to develop one’s own knowledge, and not only for the purpose of debating
with others. In this context, he quotes the Abhidharmasūtra, which lists twelve reasons why a
Bodhisattva should not get involved in a discussion with others.
It is noteworthy that in contrast to Abhidharmic works such as the Commentary on the
Treasury of Abhidharma, the Compendium of the Abhidharma does not usually deal with topics
in a controversial way. While the Commentary on the Treasury of Abhidharma often presents a
number of different and opposing positions on certain concepts and doctrines, the Compendium
of the Abhidharma is, doctrinally, a rather uniform piece of work (Bayer 2010, 16). However,
at the same time, scholars have also noticed features that indicate its compilatory character.47
The Summary of the Mahāyāna

The Summary of the Mahāyāna, which is lost in its Sanskrit original and only available in one
Tibetan (D 4048) and four Chinese translations (T 1592, 1593, 1594, 1596), can be regarded
as one of the most important works of the Yogācāra school.48 The text consists of ten chapters
258
Asaṅga
devoted to the exposition of all the main Yogācāra teachings. Compared to earlier Yogācāra
treatises, such as the Ornament of Mahāyāna Sūtras, the text is written in a clearer and more
systematic style and appears to represent a more mature stage of Yogācāra philosophy. As
the main sources of the Summary of the Mahāyāna, scholars have identified the Sūtra of the
Explanation of the Profound Meaning, the Abhidharmasūtra, the Sūtra(s) of Perfect Wisdom
(Prajñāpāramitāsūtra), the Foundation of Yoga Practicioners, the Ornament of Mahāyāna
Sūtras, and the Distinguishing the Middle from the Extremes.49 Notably, some arrive at the
conclusion that the Summary of the Mahāyāna was probably composed before the Compen-
dium of the Abhidharma,50 even though the Compendium of the Abhidharma seems closer to
the teachings of traditional Abhidharma than the Summary of the Mahāyāna.
The ten chapters of the Summary of the Mahāyāna deal respectively with the store mind
(ālayavijñāna), the three natures (svabhāva), the theory of “nothing but presentation”
(vijñaptimātratā), the six perfections ( pāramitā), the ten stages (bhūmi), the three kinds of
training (śikṣā) – i.e., in highest discipline (adhiśīla), highest mind (adhicitta) and highest
knowledge (adhiprajñā) – the “transformation of the basis” (āśrayaparāvṛtti), and Buddha-
hood (including the three Buddha bodies [kāya] and four awarenesses [ jñāna]).51
As pointed out by Lambert Schmithausen, the arrangement of the first three chapters of the
Summary of the Mahāyāna follows the structure of chapters 5–8 in the Sūtra of the Explana-
tion of the Profound Meaning. However, in contrast to earlier Yogācāra works, in the Sum-
mary of the Mahāyāna, the three Yogācāra concepts are explicitly integrated into a unified
system (Schmithausen 2014, 27–40). At the same time, the concept of “nothing but presenta-
tion” (vijñaptimātratā) is fully established and becomes the central idea of spiritual practice
and philosophical theory (Schmithausen 2018, 278). While in earlier texts, the store mind is
taught independently from the theory that all matter is only a mental manifestation, the Sum-
mary of the Mahāyāna seems to be the first Yogācāra text that fully combines the store mind
(ālayavijñāna) with the “mind only” system and regards the store mind as the basis of all phe-
nomena of the whole world. The original meaning of the store mind, namely the mind that lies
hidden in the material sense faculties, is modified in the Summary of the Mahāyāna to denote
a mind to which all conditioned factors stick (ālīyante), in the sense that they are bound to it
as their cause (Schmithausen 1987, 65).
The first chapter of the text is one of the most important, providing a detailed explanation
of the store mind that should be considered as the most elaborate of the older Yogācāra texts
(Schmithausen 1985, 140). The store mind, on the one hand, is still ascribed its traditional
function of “biological appropriation” and of keeping the corporeal basis of personal exist-
ence alive. On the other hand, it is also presented as a container of all seeds (bīja) and is
called a “maturation mind” (vipākavijñāna) (Schmithausen 1987, 61ff. and 100ff.). Asaṅga,
moreover, describes the store mind as being constituted by three kinds of imprints (vāsanā),
namely the imprints of (1) expressions (abhilāpa); (2) the (false) view of the self (ātmadṛṣṭi);
and (3) the limbs of existence (bhavāṅga) (Lamotte 1938, 22 [58]). As already noted by Erich
Frauwallner (2010, 333), especially the first two classes of imprints reflect the Yogācāra’s
emphasis on two specific doctrines: the first stresses the relevance of conventional expressions
for the human conceptualization of phenomena; the second is related to the Yogācāra theory
of the defiled notion [of “I”] (kliṣṭaṃ manas), which is also rooted (in the form of its seed) in
the store mind.
The theory of a supramundane (lokottara) insight, arising from the imprint (vāsanā) of
hearing (śruta) the Buddhist teachings, is also proposed in the first chapter of the Summary
of the Mahāyāna.52 According to this teaching, this type of insight cannot arise from seeds
originating in the store mind. Instead, it is assumed that the first moment of the supramundane
259
Jowita Kramer
insight arises from an imprint left in a person’s mental continuum when hearing the Bud-
dhist teachings, which is considered the outflow (niṣyanda) of the true reality (dharmadhātu).
Asaṅga further explains that the imprint of the supramundane insight exists together with the
store mind but nonetheless is not part of it (Schmithausen 2014, 585ff.), a state he compares
to milk mixed with water. It cannot be the store mind itself as it actually forms its antidote.
Frauwallner (2010, 332) points out that Asaṅga’s approach to the state of liberation and Bud-
dhahood is more psychological than what we find in the works attributed to Maitreya(nātha):
while the Maitreya treatises seem more concerned with the description of the highest reality
(and what we might call its ontological status), Asaṅga aims at explaining the psychological
complex underlying the process of liberation.
In chapter three, and partly also chapter two, Asaṅga treats the topic of “nothing but presen-
tation” (vijñaptimātratā) in detail. In this context, we find a list of arguments by means of which
Asaṅga tries to prove that external objects are not existent.53 This list may be considered the
earliest transmitted enumeration of this kind (Schmithausen 2014, 626). According to the first
argument, the assumption that the same object appears in various ways to the different classes
of beings – such as humans, animals or gods – can only be reasonable if the referents of their
perception are not really existent. Second, Asaṅga states that all objects perceived in dreams,
images visualized in meditation, past and future objects (appearing in thoughts), and reflections
in mirrors do not presuppose a real (material) object as their basis. The third argument explains
that if external objects were really existent, everyone would effortlessly see true reality when
perceiving them. Furthermore, nonconceptual awareness would never be experienced as com-
pletely objectless. The following two arguments point out that if external objects really existed,
it would be impossible for spiritually advanced practitioners to transform material things (for
instance, earth into gold) or make material things appear according to their will. Finally, Asaṅga
states that nonconceptual awareness could not be experienced with really existing external
objects (Schmithausen 2014, 627ff.). As Schmithausen points out (2014, 628), most of these
considerations regarding the nonexistence of external matter seem to be associated with (spir-
itual) experience in the first place and less with the solution of theoretical problems.
Another two sets of proofs to be found in the Summary of the Mahāyāna that are of par-
ticular importance are those concerning the store mind (ālayavijñāna) and the defiled notion
[of “I”] (kliṣṭaṃ manas) (Lamotte 1938, 5ff. [7] and 13–22 [29–57]). While the existence
of the store mind has already been proved in (at least) one earlier Yogācāra source (namely
the Foundation of Yoga Practitioners), the list of six proofs provided in the Summary of the
Mahāyāna does not seem to rely directly on the nine proofs of the Foundation of the Yoga
Practitioners and includes a number of innovative aspects.54 At the same time, the Summary
of the Mahāyāna seems to be the earliest transmitted Yogācāra source that explicitly presents
a list of arguments showing that the existence of the defiled notion of “I” is necessary because
its rejection would result in unacceptable consequences (Kramer 2016, 162ff.).
Asaṅga not only treats the topic of “nothing but presentation” (vijñaptimātratā) in detail,
but also defines what constitutes basic categories of conceptualization (vikalpa) and mental
presentation (vijñapti). Most notable in this context are two lists of ten kinds of vikalpa and
of eleven kinds of vijñapti,55 both summarizing various categories of a human being’s experi-
ence of the world. These classes of concepts include not only diverse aspects of the cognitive
process, such as the perceived objects, the sense faculties, and the sense perceptions, but also
concepts of “time,” “numbers,” the “surrounding world” (bhājanaloka), and the “distinction
between oneself and others.” All these categories are assumed to be only modifications of the
mind. In order to explain how, under these circumstances, a subject experiences the perception
260
Asaṅga
of (external) objects, Asaṅga differentiates between two parts of the mind, the “image” (nimi-
tta) portion and the “vision” (darśana) portion.56 Thus, every moment of perception is consti-
tuted of three different kinds of mental presentations, namely (1) the mental presentation of
the sense faculty; (2) the mental presentation of the object – that is, the “image portion” of the
mind; and (3) the mental presentation of the perception of the object, that is, the perceiving
subject or the “vision portion” of the mind (Frauwallner 2010, 330).
An important topic of Asaṅga’s investigation of the Bodhisattva’s path to liberation is
“mental verbalization” (manojalpa). It is notable that Asaṅga treats mental verbalization as an
important part of the Bodhisattva’s path to awakening. Thus, mental verbalization is regarded
to be the basis for realizing true reality.57 The term “mental verbalization” is usually applied as
a synonym of mental perception (manovijñāna) in Yogācāra texts (Kramer 2020, 365) and as
such might also be assumed to be a hindrance on the path to liberation. In the third chapter of
the Summary of the Mahāyāna Asaṅga explains that the Bodhisattva penetrates “nothing but
presentation” by means of this very mental verbalization in the form of, for instance, the four
investigations ( paryeṣana) and the corresponding four kinds of knowledge.58
The tenth chapter of the Summary of the Mahāyāna also shows some peculiarities that might
be considered specific to Asaṅga. After a short overview of the three Buddha bodies – that is, the
manifestation body (nirmāṇakāya), the enjoyment body (sambhogakāya) and the dharma body
(dharmakāya) – this chapter deals extensively with the dharma body in most of its sections.59
In this part of the Summary of the Mahāyāna, we also find an explanation of the transformation
of the cognitive system into the four kinds of awareness ( jñāna) as known from the Ornament
of Mahāyāna Sūtras (chapter 9, verse 67), including the mirror-like awareness (ādarśajñāna),
awareness of equality (samatājñāna), discriminating awareness ( pratyavekṣājñāna), and the
all-accomplishing awareness (kṛtyānuṣṭhānajñāna). Notably, Asaṅga also mentions a less
known explanation of five kinds of “mastery” (vibhutva) resulting from a change of the five
constituents (skandha) of the person. He states that the dharma body is realized by means of the
transformation of all five constituents (Lamotte 1938, 85ff. [5]). With regard to the body, the
practitioner perfects the ability to manifest pure Buddha fields in all possible bodies, equipped
with the Buddha signs and with countless melodious voices. Regarding the Buddha’s feelings,
it is said that he has perfected the ability to remain in countless blissful states (sukhavihāra). His
ideation (saṃjñā) is changed so that he becomes a master of teaching through words (nāman),
phrases ( pada), and phonemes (vyañjana). The transformation of his impulses (saṃskāra)
leads to the mastery of the manifestation (nirmāṇa), transformation (bsgyur ba), and the accu-
mulation of positive factors (śukladharmasaṃgraha). Finally, his perception (vijñāna) changes
into the four types of knowledge ( jñāna) mentioned previously.60
In summary, Asaṅga was one of the most influential masters of the Yogācāra tradition.
Even though it remains controversial which works can definitely be attributed to him as their
original author, his two main texts – the Compendium of the Abhidharma and the Summary
of the Mahāyāna, whose authorship is considered fairly certain – offer a sufficient basis for
this statement. In the Compendium of the Abhidharma, Asaṅga creates a distinctive system of
Yogācāra Abhidharma and systematizes the ideas of the Foundation of the Yoga Practition-
ers. His Summary of the Mahāyāna became one of the most important works of the Yogācāra
school, summarizing all the main Yogācāra teachings, such as the “store mind” (ālayavijñāna),
the concept of “nothing but presentation” (vijñaptimātra), the three “natures” (svabhāva), and
the three bodies (kāya) of a Buddha. Asaṅga might not have been the originator of these central
Yogācāra concepts, but he doubtlessly contributed greatly to the creation of a coherent doctri-
nal system of the Yogācāra school.
261
Jowita Kramer
Notes
1 For details, see Watanabe (2000, 7).
2 See, e.g., Watanabe (2000, 5).
3 T 2049, fol.188a11ff. and 190c14ff.; Dalia (2002, 39ff. and 51). See also Delhey (2019, 73).
4 T 2087, fol. 896b20ff.; Beal I, 226ff.
5 For an overview of this discussion, see Kritzer (1999, 7–13).
6 Lokesh Chandra 103b4–104b2; Obermiller (1932, 137–39).
7 Lokesh Chandra 21a4ff. and Obermiller 1931, (54–56).
8 Lokesh Chandra 104b7ff. and Obermiller (1932, 140).
9 Lokesh Chandra 105b7ff. and Obermiller (1932, 143). The translation of this passage is quoted from
Chimpa and Chattopadhyaya (1970, 168) (i.e., the translation of Tāranātha’s History of the Doctrine
in India) as it is more precise than Obermiller’s rendering.
10 Schiefner (1868, 88,1–90,9); Chimpa and Chattopadhyaya (1970, 160–64).
11 Schiefner (1868, 88,1ff., 91,2–92,2 and 92,13–18) and Chimpa and Chattopadhyaya (1970, 160,
165–67).
12 Schiefner (1868, 99,8ff. and 105,18ff.) and Chimpa and Chattopadhyaya (1970, 177 and 186).
13 For a detailed overview of the various aspects and controversial points of this discussion, see Delhey
(2019, 74–77).
14 Lokesh Chandra 1971, 21a4ff.; Schiefner (1868, 88,7). The Foundation of the Yoga Practitioners is
also ascribed to Asaṅga in some late Indian sources (see Delhey 2013, 502; 2019, 75).
15 Lokesh Chandra 1971, 21a4ff.; Schiefner (1868, 88,7). Asaṅga is also mentioned as the author in the
colophons of the Tibetan translation in D 4048 and of the Chinese translation in T 1594, fol. 132c19.
16 Lokesh Chandra 1971, 21a4ff.; Schiefner (1868, 88,6ff.). Asaṅga is also mentioned as the author
in the colophons of the Tibetan translation in D 4049 and of the Chinese translation in T 1605, fol.
663a05.
17 Asaṅga is mentioned as the author in T 1602, fol. 480b13.
18 Asaṅga is mentioned as the author of the verses in T 1607, fol. 774a13, while the prose is attributed
to Asaṅga. See also Delhey (2019, 77).
19 Asaṅga is mentioned as the author in T 1607, fol. 774a13.
20 Asaṅga is mentioned as the author in T 1565, fol. 44c19.
21 Asaṅga is mentioned as the author in T 1514, fol. 885a06. For further references to studies discussing
the authorship of this work, see Delhey (2019, 78).
22 See Chimpa and Chattopadhyaya (1970, 394) and D 1115.
23 Asaṅga is mentioned as the author in the colophon of the Tibetan translation in D 3550.
25 Lokesh Chandra 1971, 105a1; Asaṅga is also mentioned as the author in the colophon of the Tibetan
translation in D 3981.
26 Asaṅga is mentioned as the author in the Tibetan colophons of these three works in D 3982, D 3983,
and D 3984.
27 The colophon of the Tibetan translation in D 4025 attributes it to Maitreya. As mentioned previously,
Bu ston also states in his History of the Doctrine that Asaṅga composed a commentary on the Distin-
guishing the Germ of the [Three] Jewels (Lokesh Chandra 1971, 105a1).
29 Schiefner (1868, 88,7ff.). Bu ston mentions this work as *Tattvaviniścaya (De nyid rnam nges;
Lokesh Chandra 1971, 105a1; on this title, see Brunnhölzl [2014, 103, n. 102]). According to Delhey
(2019, 75), Haribhadra also ascribes a commentary on the Ornament of Clear Realization to Asaṅga.
30 According to Delhey (2019, 74ff.), Ᾱryavimuktisena attributes this commentary to Asaṅga instead of
Vasubandhu.
31 For the various possibilities of retranslating the title of this text to Sanskrit, see Schmithausen (1987,
261, n. 99).
32 For bibliographical references of various modern studies on several of these texts, see Watanabe
(2000, 7–15).
33 For further details on the authorship of these three works, see Delhey (2019, 78), who points out that
it is probably impossible to exclude the possibility of Asaṅga’s authorship in the case of the Seventy
Stanzas on the Perfection of Wisdom in Three Hundred Stanzas.
34 See also Schmithausen (2018, 265ff.).
262
Asaṅga
35 For further details, see, e.g., Williams 179ff.

36 Radich (2007, 1187); Bayer (2010, 22).
37 For further details of this discussion, see Schmithausen (2018, 266).
38 According to Schmithausen (2018, 264), this is one of two strands from which the concept of “mind
only” emerged, the other being the Sūtra of the Concentration of the Direct Encounter with the
Buddhas of the Present (Pratyutpannabuddhasaṁmukhāvasthitasamādhisūtra), where cittamātra is
mentioned in connection with the visions of the Buddha Amitāyus.
39 The Sanskrit text of the Compendium of the Abhidharma was first edited by Gokhale in 1947 on the
basis of a fragmentary manuscript preserved in Tibet. In 1950, the whole Sanskrit text was published
by Pradhan, who filled in the passages missing in the manuscript with his own reconstructions of the
Sanskrit from the Tibetan translation of the Compendium of the Abhidharma (D 4049). For the Chi-
nese translation of the Compendium of the Abhidharma, see T 1605. Recently also a Sanskrit manu-
script of the Commentary on the Compendium of the Abhidharma (Abhidharmasamuccayavyākhyā),
which contains the whole of its root-text – that is, the Compendium of the Abhidharma – came to
light and is currently being edited by Li Xuezhu and Kazuo Kano (see Li 2012, 2013, 2014, 2015; Li
and Kano 2014). For a French translation of the Compendium of the Abhidharma, see Rahula (1980),
whose translation has been rendered into English by Sara Boin-Webb (2001).
40 Kritzer (1999, vii).
41 Bayer (2010, 22). As noted by Bayer, the idea of the apratiṣṭhitanirvāṇa is also present in the treatises
attributed to Maitreya, such as the Mahāyānasūtrālaṃkāra.
42 See Schmithausen (1972, 156); Bayer (2010, 21).
43 See Bayer (2010, 20ff.); Schmithausen (2014, 30ff.).
44 See Schmithausen (1987, 100); Schmithausen (2014, 302ff.).
45 The Sanskrit text of this passage is not extant. For the Tibetan version, see D 4049, 101b1.
46 The Sanskrit text of this passage is not extant. For the Tibetan version, see D 4049, 118a6ff.
47 Kritzer (1999, 5); Bayer (2010, 37ff.).
48 The Summary of the Mahāyāna has been translated to French in Lamotte (1938) and to English in
Brunnhölzl (2018).
49 Watanabe (2000, 16); Brunnhölzl (2018, 4ff.).
50 Schmithausen (1987, 403, n. 713); Bayer (2010, 32).
51 For a detailed summary of the ten chapters, see Brunnhölzl (2018, 7–21).
52 Lamotte (1938, 19ff., 45–49). See also Schmithausen (1987, 79ff.).
53 Lamotte (1938, 30ff., 14). For the Sanskrit text, see ASBh 41,21–42,15.
54 For details, see Kramer (2016).
55 Lamotte (1938, 25 [2] and 34ff. [20]). See also Frauwallner (2010, 329); Kramer (2018, 327, 329).
56 Lamotte (1938, 29 [11]). See also Frauwallner (2010, 329).
57 See also Watanabe (2000, 77ff.).
58 Lamotte (1938, 51ff. [7]). See also Brunnhölzl (2018, 13).
59 For a detailed study of this chapter of the Summary of the Mahāyāna, see Griffiths et al. (1989).
60 See also Brunnhölzl (2018, 237).
Bibliography
Bayer, Achim. 2010. The Theory of Karman in the Abhidharmasamuccaya. Tokyo: The International
Institute for Buddhist Studies.
Beal, Samuel. 1884. Si-yu-ki: Buddhist Records of the Western World, vol. 2. London: Trübner & Co.
Boin-Webb, Sarah. 2001. Abhidharma Samuccaya: The Compendium of the Higher Teaching (Philoso-
phy). Fremont: Asian Humanities Press (English translation of Rahula 1980).
Brunnhölzl, Karl. 2014. Gone Beyond: The Prajñāpāramitā Sūtras, The Ornament of Clear Realization,
and Its Commentaries in the Tibetan Kagyu Tradition. Boulder: Snow Lion Publications.
———. 2018. A Compendium of Mahāyāna, Asaṅga’s Mahāyānasaṃgraha and Its Indian and Tibetan
Commentaries, vol. 1. Boulder: Snow Lion Publications.
Chandra, Lokesh. 1971. The Collected Works of Bu-ston. Part 24 (ya). Śata-Piṭaka Series 64. New Delhi:
International Academy of Indian Culture.
Chimpa, Lama, and Alaka Chattopadhyaya. 1970. Tāranātha’s History of Buddhism in India. Simla:
Indian Institute of Advanced Study.
263
Jowita Kramer
Dalia, Albert A. 2002. “Biography of Dharma Master Vasubandhu.” In Lives of Great Monks and Nuns,
translated by Albert A. Dalia and Li Rongxi, 31–53. Berkeley: Numata Center for Buddhist Transla-
tion and Research.
Deleanu, Florin. 2006. The Chapter on the Mundane Path (Laukikamārga) in the Śrāvakabhūmi. Tokyo:
The International Institute for Buddhist Studies of the International College for Postgraduate Bud-
dhist Studies.
Delhey, Martin. 2013. “The Yogācārabhūmi Corpus Sources, Editions, Translations, and Reference
Works.” In The Foundation for Yoga Practitioners: The Buddhist Yogācārabhūmi Treatise and Its
Adaptation in India, East Asia, and Tibet, edited by Ulrich Timme Kragh, 498–561. Cambridge:
Harvard University Press.
———. 2019. “Asaṅga/Maitreya(nātha).” In Brill’s Encyclopedia of Buddhism. Vol. II: Lives, edited by
Jonathan Silk et al., 73–80. Boston: Brill.
Frauwallner, Erich. (1956) 2010. Philosophie des Buddhismus. Berlin: Akademie Verlag.
Gokhale, V. V. 1947. “Fragments from the Abhidharmasamuccaya of Asaṅga.” Journal of the Bombay
Branch of the Royal Asiatic Society 23: 13–38.
Griffiths, Paul et al. 1989. The Realm of Awakening: Chapter Ten of Asaṅga’s Mahāyānasaṃgraha. New
Kramer, Jowita. 2016. “Some Remarks on the Proofs of the ‘Store Mind’ (Ālayavijñāna) and the Devel-
opment of the Concept of Manas.” In Text, Philosophy, and History: Abhidharma across Buddhist
Scholastic Traditions, edited by Bart Dessein and Weijen Teng, 146–68. Leiden: Brill.
———. 2018. “Conceptuality and Non-Conceptuality in Yogācāra Sources.” Journal of Indian Philoso-
phy 46 (2): 321–38.
———. 2020. “Concepts of the Spiritual Path in the *Sūtrālaṃkāravṛttibhāṣya (Part II): The Eighteen
Manaskāras and the Adhimukticaryābhūmi.” In Mārga: Paths to Liberation in South Asian Buddhist
Traditions, edited by Cristina Pecchia and Vincent Eltschinger, 329–61. Wien: Österreichische Aka-
demie der Wissenschaften.
Kritzer, Robert. 1999. Rebirth and Causation in Yogācāra Abhidharma. Wien: Arbeitskreis für Tibetische
und Buddhistische Studien, Universität Wien.
Lamotte, Étienne. 1938. La somme du grand véhicule d’Asaṅga (Mahāyānasaṃgraha). Louvain: Uni-
versité de Louvain, Institut Orientaliste.
Li, Xuezhu. 2012. “ ‘Abhidharmasamuccayavyākhya no jobun ni tsuite’ Abhidharmasamuccayavyākhya
の序文について (‘Sanskrit text of the Opening Section of Sthiramati’s Abhidharmasamuccayavya-
khya’)”. Indogaku Chibettogaku Kenkyu インド学チベット学研究 (Journal of Indian and Tibetan
Studies) 16: 1–7.
———. 2013. “Diplomatic Transcription of Newly Available Leaves from Asaṅga’s Abhidharmasam-
uccaya.” Annual Report of the International Research Institute for Advanced Buddhology at Soka
University 16: 241–53.
———. 2014. “Diplomatic Transcription of Newly Available Leaves from Asaṅga’s Abhidharmasamuc-
caya – Folios 29, 33, 39, 43, 44.” Annual Report of the International Research Institute for Advanced
Buddhology at Soka University 17: 195–205.
———. 2015. “Diplomatic Transcription of the Sanskrit Manuscript of the Abhidharmasamuccayavyākhyā
– Folios 2v4–8v4.” Annual Report of the International Research Institute for Advanced Buddhology
at Soka University 18: 275–83.
Li, Xuezhu, and Kazuo Kano. 2014. “Restoration of Sanskrit Text in Missing Leaves (Fols. 2, 6, 7) of the
Abhidharmasamuccaya Manuscript on the Basis of the Abhidharmasamuccayavyākhyā Manuscript.”
China Tibetology 23 (2): 53–63.
Martin, Dan. 2002. “Gray Traces: Tracing the Tibetan Teaching Transmission of the Mngon-pa Kun-btus
(Abhidharmasamuccaya) Through the Early Period of Disunity.” In The Many Canons of Tibetan
Buddhism, edited by Helmut Eimer and David Germano, 335–57. Leiden: Brill.
Obermiller, Eugene. 1931. History of Buddhism by Bu-ston (Part I: The Jewelry of Scripture). Heidel-
berg: Harrassowitz.
———. 1932. History of Buddhism by Bu-ston (Part II: The History of Buddhism in India and Tibet).
Heidelberg: Harrassowitz.
Pradhan, Pralhad. 1950. Abhidharma Samuccaya of Asanga. Santiniketan: Visva-Bharati.
Radich, Michael. 2007. “The Somatics of Liberation: Ideas about Embodiment in Buddhism from Its
Origins to the Fifth Century C.E.” PhD diss., Harvard University.
264
Asaṅga
Rahula, Walpola. 1980. Le compendium de la super-doctrine ( philosophie) (Abhidharmasamuccaya)

d’Asanga. Paris: École Française d’Extrème Orient.
Schiefner, Anton. 1868. Tāranāthae de Doctrinae Buddhicae in India Propagatione Narratio. St. Peters-
burg: Academia Scientiarum Petropolitana.
Schmithausen, Lambert. 1972. “The Definition of Pratyakṣam in the Abhidharmasamuccayaḥ.” Wiener
Zeitschrift für die Kunde Südasiens 16: 153–63.
———. 1985. “Once Again Mahynasaṃgraha I.8.” In Buddhism and Its Relation to Other Religions:
Essays in Honour of Dr. Shōzen Kumoi, 139–60. Kyoto: Heirakuji Shoten.
———. 1987. Ālayavijñāna: On the Origin and Early Development of a Central Concept of Yogācāra
Philosophy. Tokyo: The International Institute for Buddhist Studies.
———. 2014. The Genesis of Yogācāra-Vijñānavāda: Responses and Reflections. Tokyo: The Interna-
tional Institute for Buddhist Studies.
———. 2018. “Some Remarks on the Genesis of Central Yogācāra-Vijñānavāda Concepts.” Journal of
Indian Philosophy 46 (2): 263–81.
Skilling, Peter 2013. “Vaidalya, Mahāyāna, and Bodhisatva in India: An Essay Towards Historical
Understanding.” In The Bodhisattva Ideal: Essays on the Emergence of Mahāyāna, edited by Bhik-
khu Nyanatusita, 69–119. Kandy: Buddhist Publication Society.
von Rospatt, Alexander. 1995. The Doctrine of Momentariness. Stuttgart: Franz Steiner Verlag.
Watanabe, Chikafumi. 2000. “A Study of Mahāyānasaṃgraha III: The Relation of Practical Theories and
Philosophical Theories.” PhD diss., University of Calgary.
Williams, Paul. 1989. Mahāyāna Buddhism: The Doctrinal Foundations. London: Routledge.
265
16
VASUBANDHU
Mainstream and Mahāyāna
Jonathan C. Gold
Introduction: Vasubandhu’s Identity Problem

Vasubandhu (fourth century CE) sits at the center of the Buddhist philosophy curriculum. His
Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya), especially its ninth
chapter, contains (among other things) many of the most compelling and influential arguments
in defense of the definitive Buddhist view of no-self.1 In India, this work was debated by Bud-
dhists of different schools, as well as between Buddhists and non-Buddhists, and its ideas were
fundamental to the development of Buddhist epistemology. Its influence expanded with Bud-
dhism into China and later Tibet, where it remains central to scholastic education. In recent
decades, it has found a new, admiring readership among Western philosophers such as Derek
Parfit. Similarly influential was Vasubandhu’s Twenty Verses (Viṃśikā), also called the Estab-
lishment of Appearance Only (Vijñaptimātrasiddhi), which contains the most widely repeated
arguments in defense of the Buddhist understanding that everything in experience is only mind,
like an illusion or a dream.2 Many other works, some of them also extremely influential, are
attributed to Vasubandhu (Skilling 2000). But just looking at these two, we see a strange fact:
The first is a work of mainstream Abhidharma philosophy, analyzing the nature of the founda-
tional elements (dharmas) found within the Buddha’s early teachings. The second, however,
expresses doctrines of the Great Way (Mahāyāna) and denies the ultimate reality of the very
elements decisively defended in the first. The works establish quite different perspectives.
Vasubandhu’s traditional biographies account for the difference between these works with a
story of his having converted from mainstream Buddhism to Mahāyāna late in his life, under the
influence of his elder brother, Asaṅga (Takakusu 1904). The scholar Erich Frauwallner (1898–
1974) proposed that the difference could be accounted for by the realization that there were,
in fact, two different scholars named Vasubandhu – an earlier one who wrote the Mahāyāna
treatises and commentaries and a later one who wrote the Treasury Commentary (Frauwallner
1951). This idea has drawn much discussion, and the “two Vasubandhus” hypothesis lives
on (Buescher 2013; Kapstein 2018). Yet, a key point for our purposes is that Frauwallner
was compelled to clarify his thesis and acknowledge that the two doctrinally conflictual texts
mentioned previously were, in fact, most likely written by the same person named Vasubandhu
(Frauwallner 1961, 131–32). According to the revised “two Vasubandhus” account, the texts
that the “earlier” Vasubandhu wrote are now thought to include commentaries on Mahāyāna
266 DOI: 10.4324/9781351030908-25

Vasubandhu
scriptures and treatises, but not all of the Mahāyāna treatises (and, notably, not the Twenty
Verses itself). The “two Vasubandhus” hypothesis, whatever its legitimacy, does not explain
why both mainstream and Mahāyāna positions are defended in the same author’s works.
To add to the problem, tradition ascribes Vasubandhu authorship of not just the Trea
sury’s Commentary, but also its root verses, the Treasury of Abhidharma Verses
(Abhidharmakośakārikā). This root text provides a comprehensive systematization of the
Abhidharma system according to the Vaibhāṣika (also called the Sarvāstivāda), a position
that the Commentary tends to challenge and critique from many perspectives, primarily one
that the Commentary calls Sautrāntika. So Vasubandhu apparently represented the Vaibhāṣika
and its opponents, respectively, in the root text and commentary. Traditional biographies
and surviving criticisms of his work attest that Vasubandhu was, indeed, considered a traitor
by Vaibhāṣikas (Park 2014). That means he was not just a flip-flopper from mainstream to
Mahāyāna; he switched within the mainstream of Abhidharma, as well, which is why his extant
work represents three distinct positions. The tradition that followed from Asaṅga, which his
Mahāyāna works defend, is called Yogācāra. So we have Vasubandhu writing as Vaibhāṣika,
Sautrāntika, and Yogācāra.
At the risk of yet further complication, we may note that when the term Yogācāra names a
system of Indian Mahāyāna philosophy, whether in traditional or contemporary scholarship, it
is commonly contrasted with Madhyamaka. But Vasubandhu never used the term “Yogācāra”
to refer to his own work, and he made use of some of the key argumentative moves that are
seen in the Madhyamaka works of Nāgārjuna. He never disparaged that system of thought,
except where it was misread as an overly zealous negation of traditional doctrines (a dispar-
agement that is contained within Nāgārjuna’s works, as well) (Gold 2015b). So of the four
standard schools of philosophy that Tibetans list in the introductory textbooks – Vaibhāṣika,
Sautrāntika, Yogācāra, and Madhyamaka – Vasubandhu appears to encompass them all.
A solution to the problem comes when we understand Vasubandhu not as a representative
of a particular school or even a set of schools but as a philosopher dealing in turn with a range
of inherited traditions. For in spite of the diversity of the various identities adopted therein, the
previously mentioned texts display a notable consistency of method and style. Other works
ascribed to Vasubandhu diverge in style but often extend a coherent set of ideas. We may
doubt whether one person wrote the various texts attributed to him – but a unified, and unify-
ing, intention resonates across them. When they are read together, they exhibit a distinctive
approach that weaves Mahāyāna through the full fabric of mainstream Buddhism.
Traditional biographies of Vasubandhu (e.g., Takakusu 1904) depict a brilliant and combat-
ive scholar whose career seems to have been animated by controversy. He was the middle of
three sons of a prominent brahmin in the city of Puruṣapura in Gandhara, all officially named
“Vasubandhu” – though his younger brother went by Viriñcivatsa and his elder by Asaṅga.
At the time, the Sarvāstivāda system of the Vaibhāṣikas was the pinnacle of Buddhist learn-
ing and it was officially taught only in Kashmir. Yet a handful of scholars had surreptitiously
brought these teachings out of the valley, and Vasubandhu was able to master them, possibly
through his teacher Buddhamitra. He composed a verse summary of the Vaibhāṣika system in
600 verses, which the elders of Kashmir apparently approved, but then wrote his commentary,
which they considered traitorous.
At one point, Buddhamitra was publicly humiliated by a defender of the Sāṃkhya school
while Vasubandhu was away, so after he returned, he wrote a devastating critique of the
Sāṃkhyas. This earned him praise and royal support, and he became the personal instructor of
the queen and the crown prince. Subsequently, after being challenged by a Vyākaraṇa (gram-
mar) scholar, he wrote a similarly ruinous critique of the Vyākaraṇas. This defeated enemy
267
Jonathan C. Gold
drew Vasubandhu to the attention of the Vaibhāṣika master Saṅghabhadra, who in turn wrote
two extensive criticisms of Vasubandhu’s Treasury Commentary and challenged Vasubandhu
to a debate. This time, Vasubandhu refused the contest since he believed that he had already
made a strong case and did not believe that his views had been undermined.
The final challenge to Vasubandhu’s views, however, came from his brother Asaṅga.
Asaṅga had spent his life propounding the Mahāyāna teachings he had solicited directly from
the future Buddha Maitreya in heaven, and he worried that his brilliant brother would compose
a treatise countering the legitimacy of the Mahāyāna. We are told that Asaṅga, feigning ill-
ness, sent for Vasubandhu and, when his brother arrived, personally taught him the Mahāyāna.
Vasubandhu saw the depth of the teachings and changed his perspective entirely, regretting his
previous doubts. He became, thenceforth, a defender of the Mahāyāna and spent his later years
composing his various Mahāyāna commentaries and treatises.
In this chapter, I will use a few of Vasubandhu’s key arguments to draw out a unifying
theme in his approach to causality and scripture. Vasubandhu’s preference for empirically
grounded causal stories over conceptual-linguistic abstractions, we will see, implies that the
words of the Buddha must often be considered figurative or under a secondary, special inten-
tion. Vasubandhu’s understanding of the Buddha’s indirect use of language is grounded in his
early reading of the doctrine of no-self as an expression of dependent origination, which, when
understood to be paradigmatic for all doctrines, provides him an opening to Mahāyāna in a
Yogācāra mode. Central to our reading of Vasubandhu’s view is his recognition that Yogācāra’s
distinctive position – the mind-only nature of reality – is both an ontological thesis about the
impossibility of perceived entities and a reading of how concepts generate false appearances
in the mind. Yogācāra in Vasubandhu’s hands never loses sight of the fact that even false con-
cepts are the most important and most effective method used by the Buddha to move disciples
forward on the path to enlightenment – that is to say, the concepts of the Dharma. To acknowl-
edge the mind-only nature of all dharmas is not to reject the Buddha’s teachings, but rather to
accept that all of the Buddha’s teachings are unified in the common project of revealing the
ineffable nature of reality, to whatever degree living beings are able to see it. As we will see,
an appeal to dependent origination as the causal basis of both reality and illusion is formalized
in the Yogācāra doctrine of three natures, but it was already implicit in the argument that the
false view of the self is superimposed on the aggregates.
The following two sections trace this argument first in Vasubandhu’s mainstream Abhi
dharma and then through his Mahāyāna works.
Perception, Causality, and Scripture in the Treasury Commentary

The final chapter of Vasubandhu’s Treasury Commentary begins by stating that it is a key
teaching of the Buddha that the self is a false construction imposed upon the ever-changing
continuum of five aggregates: forms (rūpa), feelings (vedanā), cognitions (saṃjñā), condition-
ings (saṃskāra), and consciousness (vijñāna). Then, in response to an objector’s request for
an explanation of how he can know this, Vasubandhu throws down the gauntlet by stating a
proof that there is no self:
And how is this to be understood, that the word “self ” indicates only the continuum
of aggregates, and does not apply elsewhere?
Because of there being neither perception nor inference. For where there are entities
(dharmas), their perception is apprehended when there is no obstruction. Such is the
case for the six sensory objects (viṣaya) and the mind. And there is an inference in the
268
Vasubandhu
case of the five sensory organs. Here the inference is that with a cause in place, when
another cause does not exist, no result is seen, and when it does exist the result too
exists, as in the case of a sprout. That is, with the cause in place that consists in the sen-
sory object with a particular appearance together with mental effort, the grasping of the
sensory object is seen to not exist and to exist, respectively, for people who are blind
or deaf, etc. and people who are not blind or deaf, etc. So the nonexistence and exist-
ence, respectively, of another cause are determined. And that other cause is the sensory
organ: that’s the inference. And this does not exist for the self, so there is no self.3
Vasubandhu declares that there are two means for ascertaining whether or not something is
real – perception and inference – and for each, he provides a model and a set of objects of
awareness. In one of his main categorical systems, the Buddha explained that all of experi-
ence was divisible into twelve so-called “spheres” or “domains” (āyatana): six sensory organs
and six sensory objects, six to include the five ordinary senses plus the mind and its objects.
Vasubandhu answers the question, then, by switching from the aggregate system to the sphere
system. All twelve spheres, Vasubandhu says, can be known by perception and inference,
whereas there is no such argument for the self. Thus, there is no self.
Perception is the direct acquisition of knowledge about perceptual objects. Vasubandhu
does not defend the idea that the self is unperceived, but this is an important premise of the
no-self view. On the one side, it is in line with prevalent non-Buddhist positions (“By what
means can one perceive the perceiver?” BAU 4.5.15, Olivelle 1996, 71), and on the other side,
it challenges potential opponents to define just what aspects of the self may be perceived, over
and above the aggregates (or, in this case, the spheres). He says that what is perceived consists
in, naturally, the sensory objects – but also, interestingly, the mind itself. This adds a twist to
an otherwise straightforward claim. The great Buddhist epistemologist Dignāga, who is said to
have been a follower if not a disciple of Vasubandhu, argued that the mind is intrinsically self-
perceiving (svasaṃvedana), a view that became a crucial topic in later Buddhist philosophy.
This is not Vasubandhu’s view; in order to understand what he is saying, we need to delve into
what he means by considering the mind as mental organ (manas).
As previously mentioned, the Buddha classified experience into twelve spheres, as well
as into five aggregates. In addition, he provided a third comprehensive classificatory system,
the eighteen “realms” (dhātu). Abhidharma analysis works to bring these three systems – and
many other subsidiary systems – into alignment. The “realms” classification keeps all twelve
of the twelve spheres and adds six consciousnesses: five sensory consciousnesses and one
mental consciousness. Now, an Abhidharma philosopher, charged with making sense of Bud-
dhist doctrine, needs to be able to explain: if the Buddha already said that all of reality may
be classified among the twelve spheres, how can there be six further elements in the realms
classification? Was the spheres classification not really comprehensive?
Vasubandhu addresses this issue in his discussion of the nature of the mind as mental organ
(I.17; Pradhan 1975, 11–12). What he says is that the six consciousnesses in the realms classi-
fication are just an expansion of the mind as understood in the spheres classification. The mind
as mental organ just is the sensory consciousnesses to the extent that they serve as a basis for
awareness. This, after all, is how sensation works: the eye comes into contact with a visual
object, which in turn generates a visual consciousness. If that visual consciousness is taken up
in a subsequent awareness (in a memory, for instance), it is playing the role of a “sensory organ”
for mental awareness, analogous to the eye with its visual object. So this is why the mind is
included among the objects of perception: not because the mind perceives itself in its every act of
awareness, but because sensory consciousnesses commonly appear to subsequent awarenesses.
269
Jonathan C. Gold
When a visual consciousness comes into being, there is perception of the visual object. When a
mental consciousness comes into being, there is a perception of a mental object, as when one has
a thought. When the mental object happens to be a previous mental consciousness – for instance,
when a visual consciousness is remembered – there is perception of the mind. This is how the
mind becomes an object of perception. It is not the case that all mental events are objects of men-
tal perception. But sometimes, the mind is perceived.
This accounts for the internal consistency of the Buddha holding two different classifica-
tory systems because they are two ways of speaking of the same thing. The realms classifica-
tion with its six consciousnesses allows you to supply a causal basis and object of awareness
for the mental organ. But if you do not need to detail its operations, the mental organ can be
considered under a spheres classification as including seven distinct realms (six conscious-
nesses and the mental organ). In fact, Vasubandhu allows that if one is considering only sub-
stantial entities, the realms classification ought to list only seventeen, not eighteen, because if
you list the six consciousnesses that are an expansion of the mental organ, you do not also need
to list the mental organ. It is only from the perspective that distinguishes the mental organ as
a support for subsequent awareness that you need to list them all. The Buddha’s classificatory
systems are all fully comprehensive, but they serve distinct discursive purposes.4
To return to the argument against the self under the spheres classification, then, if we have
covered all sensory objects and the mental organ, what is left unperceived are the five mate-
rial sensory organs. These, Vasubandhu says, can be established by inference. The inferential
structure he provides is original and consequential for our understanding of his approach to
reality. It is a causal inference: one may infer the reality of an unperceived entity if it is a nec-
essary component – the unseen cause – in a causal relation. We know that when a blind person
and a person who is not blind are put in the same visual situation, the blind person cannot see
what the other person can see. This is formalized as, “with a cause in place, when another
cause does not exist, no result is seen, and when it does exist the result too exists.”5 In such a
situation, the thing that either does not exist or does exist – in this case, the faculty of the visual
organ – is ascertained with certainty.
I am wary of those who profess Buddhism to be “science” as opposed to a religion; it is
not. But Vasubandhu’s emphasis on causal reasoning with a generally empirical bent resonates
with modern models for scientific reasoning. He emphasizes the evidential power of a “seen”
(dṛṣṭa) result even when inferring an unperceived object, and he only allows that causality
is inferable once proper controls are in place (quasi-experimental controls, because this is a
thought experiment). Most importantly, reality is founded, most basically, on causality. Vasu-
bandhu does not say that this is the only valid inferential means to an unperceived entity, but
he does say that all the unperceived spheres can be proven through this means, and across his
work, he often demands precise definitions and disallows reality ascription until an empirically
grounded causal argument is found. Vasubandhu also has his own version of the principle of
scientific reasoning commonly known as Occam’s Razor, or explanatory simplicity, which he
applies to eliminate causal overdetermination. He mocks an opponent who accepts a visible
cause but still holds onto an invisible one, saying, “For him, once his laxative is taken, the god-
dess causes the purge – so what else will he imagine?”6 Together, he uses these methods to dis-
prove the eternal self and other non-Buddhist eternal entities such as a creator God and atoms,
as well as a host of categories defended in the Vaibhāṣika system, from invisible materiality to
substantially existent past and future.
Vasubandhu’s requirement that you provide the causal matrix for every entity is also cen-
tral to his famous argument proving universal momentariness (Pradhan 1975, 193.2–8). If
everything is defined according to what causes it and what it causes, to exist is just to be a
270
Vasubandhu
cause for specifiable results. This means that a thing can only be said to properly exist when it
is producing its subsequent results. Consequently – and this is Vasubandhu’s innovation (Von
Rospatt 1995) – if a thing does not have it in its nature to pass out of existence, there is nothing
that could make it pass out of existence; conversely, a thing that has it in its nature to continue
to exist would have to always exist. The only thing that could stop it from continuing to exist
would be something that was the cause of its non-existence. However, a “non-existence” is
just the kind of thing that we think is real because of linguistic construction but does not really
exist. A non-existence is just what we think we see when we expect to see something that is not
present. So no entity could be defined as having a substantial nature that brings about another
entity’s non-existence; that would be like saying it causes a bachelor’s wife.
The passage we have been discussing initiates the chapter with the challenge to his opponent
that no self can be proven, and it is followed by a variety of responses that attempt to find a causal
justification for the self. Much of the chapter is taken up with answers to questions about how
the no-self view – that is, the view that all there is to the self is the aggregates – can account for
accepted empirical realities such as memory, karma, or the perception of an apparent person. But
although the refutation of self is Vasubandhu’s central purpose in the chapter, his opening point is
not just that there is no self; it is that the self is a false construction imposed upon the aggregates,
and that this was central to the Buddha’s teaching. I will address in turn these two crucial addi-
tions, which are important to Vasubandhu over and above the idea that there is no self: the self
as imposed upon the aggregates and the use of this truth as a guide to the Buddhist scriptures.
First, it is one thing to say that there is no essential reality to the self, but it is something else
entirely to say that the self is a false construction imposed upon the aggregates, spheres, and
realms. To say that is to admit more effective, more useful, measurable entities that deserve
our recognition because they can provide a truer reality than our false belief in the self. There
are substantial entities which, unlike the self, exist in a nexus of causes and results that we can
see and deduce. This nexus allows us to advance on the path to liberation and ultimately free
ourselves from the false construction which has no causal basis.
The key distinction here is that between how things appear, on the one hand, and how
things actually operate, causally, on the other. Now, this is a natural distinction that takes no
special instruction to affirm, but it will be of use to note that the two contrasting concepts
make use of two distinct frames of reference. Things look one way from a certain perspective,
and things are really happening, causally, in a different way from another perspective. It is
important to note that there is a hierarchy here between perspectives: the false construction is
grounded within the substantial perspective, whereas the opposite is not true. The belief in the
self is grounded in a misperception of the five aggregates – the body, first, and also the four
mental aggregates. These aggregated components are in constant flux; they have no unifying
“self ” or essence or core – but it is with reference to the aggregates that people imagine them-
selves to have a self. I sometimes call this a distinction between viewing-of and viewing-as-if.
One experiences the aggregates, but takes them to be a self.
Now, several paragraphs on from the section we have been discussing, Vasubandhu allows
an imagined opponent to question him on this: why, the opponent asks, are you so sure that it
is the aggregates that ground the mistaken idea of the self? This is no longer an argument that
there is no self; it is an argument that instead of the self, there is something else, specifically
the aggregates, which we mistake for a self. He answers as follows:
People of ordinary intellect come to believe, “I am white; I am dark; I am fat; I am

thin; I am old; I am young.” They identify themselves with these things. Souls are not
of this type. Therefore, the identification of the self has the aggregates as its object.7
271
Jonathan C. Gold
We all know that an eternal soul is not “old” or “young,” it is not “thin” or “fat,” it is not
“white” or “dark.” These are qualities that might be attributable to a body, but not a soul. If
a soul can be reborn, it can take on different kinds of bodies, so it could never be “white” or
“dark,” “fat” or “thin” in itself; and being “old” or “young” indicates not just a particular
body, but a particular time of life for a particular body. So Vasubandhu’s argument is simple
and straightforward: people do, of course, think of themselves as “white” or “dark” – he is
acknowledging that – but those are qualities of aggregates, not a soul. It is a mistake to attrib-
ute this kind of quality to a self. The fallacious construction of self, then, is fundamentally a
mistake whereby we grasp onto and identify with impermanent components – the aggregates.
Notice that the translator here, Charles Goodman, has made something of a clever choice in
rendering gaura and śyāma as “white” and “dark.” These are color terms, and they could just
be rendered “white” and “black”; but that translation would imply, in this context in English,
the racial identities associated with those terms in a way that would be anachronistic and cul-
turally inappropriate. In addition to their use as literal color terms, the words are used in San-
skrit to indicate lightness or darkness of skin color, both as terms of beauty. The gaura beauty
shines brightly like the moon, and the śyāma beauty has a glossy shine, like a raven. Needless
to say, in those contexts, they do not literally mean the colors “white” and “black,” and this is
what might motivate the alternate translation as “light” and “dark” here. Goodman’s render-
ing chooses one from each column: “white” and “dark.” For me, this elegantly foregrounds
the difference between our social setting and the classical Indian one while, at the same time,
displaying how in both cultural contexts, the false reification of skin color-related identity is
implicit in the terms themselves. This foregrounding of culturally generated biases might be
distracting, were it not the central point of the passage to argue that we delusively generate our
identities in dependence upon temporary, conditioned appearances.
The first several times I read this passage, I assumed that when he said we ground our
affection for ourselves in the aggregates, Vasubandhu was speaking entirely about the material
aggregate, the body. He mentions the aggregates all together, but the qualities seem to pick out
the material body. I was amazed to notice that Vasubandhu and other classical Buddhists were
aware of what we think of today as psychological constructions imposed upon the body. But
I had not seen just how insightful he was. Recently, thinking about this passage in the light of
the widely acknowledged social construction of race, it occurred to me that the form aggre-
gate – the material body itself – is unsuitable for identification. The error in which we experi-
ence (vijñāna) an affectionate identification with qualities of the material body (rūpa) requires
that we apply classifications (saṃjñā) that are based in inherited linguistic and conceptual pat-
terns (saṃskāra) and that are always associated with specific feelings and attitudes (vedanā).
Remarkably, Vasubandhu was aware of the psychological construction of identities (though
he was not aware of the social construction of identities) that the aggregates impose upon the
material body. He understood that the illusion of self involves all the aggregate groups.
Yet even if all of the aggregates are involved in the process of self-construction, the point
that Vasubandhu is making in this passage highlights the distinction between the aggregates as
actively constructing identities, which happens in the background (so to speak), and the mate-
rial body aggregate, with apparent qualities, with which the self is ostensibly identified. This
is, again, viewing-of and viewing-as-if: to say “I am white” is to see not just the aggregates, but
specifically the material body, as a self. The key difference between these two levels is that the
basis, the material aggregate, has its own causal story that has no need for the causal story of
the illusory view of self or the conceptual construction of “whiteness.” Their causal story – the
causal story of these views – requires that one import a description of the other aggregates:
karmic dispositions, emotional patterns and the imposition of false constructs. But when one
272
Vasubandhu
describes those causes, there is no longer anything one can literally describe as “self ” or
“whiteness.” They are just appearances.
Vasubandhu uses the distinction between useful causal descriptions (aggregates, spheres,
and realms) and false constructions not only to articulate the view of no-self, but also to provide
a method for analyzing Buddhist scripture. This is the second significant purpose in the chapter,
over and above the proof of no-self, that I mentioned previously. In fact, much of the chapter
is taken with scriptural citation, both in defense of the scriptural basis for no-self and in reply
to opponents’ interpretations of scriptures that challenge that notion. To take a wider view,
Buddhist philosophy is most fundamentally an attempt to make sense of Buddhist scripture.
Vasubandhu himself may have been more aware of this than most philosophers; his Proper
Mode of Exposition (Vyākhyāyukti) is the only known Indian Buddhist treatise dedicated to
scriptural interpretation. But the philosophical traditions in Buddhism are all, directly or indi-
rectly, derived from Abhidharma, which simply means “true Dharma” or “higher Dharma” –
that is, it is the practice dedicated to clarifying the reality known and expressed by the Buddha.8
But granting that Abhidharma – and, by extension, all Buddhist philosophy – must consider
scriptural evidence, Vasubandhu wants to distinguish between legitimate citations and those
that are foolish and misleading. Vasubandhu provides an extensive set of citations that lay out
the no-self view directly. But he is not a literalist in scriptural interpretation. On the contrary,
one of his key criticisms of his opponents is that they deduce the reality of entities based on
scriptural passages that are simply using conventional terminology to speak about complex
entities. That is, there is a tendency to take structures that are natural grammatical expressions
or common-sense expressions and read them at their surface level as reified entities.
To pick a telling example from earlier in the Treasury Commentary (I.42), this kind of
erroneous reification of entities appears to follow naturally from the Buddha’s statement that
“the eye sees, the mind cognizes,” which suggests that the eye is an agent of seeing. In truth,
as Vasubandhu argues, the eye is not an agent. It is a causal factor that, in the presence of the
appropriate kind of object, generates a visual consciousness. It would be more accurate to say
that it is the visual consciousness which “sees” – but even this is not entirely correct, for the
visual consciousness contains the visual information within it; it is not distinct from that visual
appearance. So it is not doing anything that could be considered “seeing.” What appears to us
as “seeing” is a causal event with several distinct elements operating in a series. As he argues
elsewhere, momentariness means that the visual object is actually already gone – it is in the
past – in the moment the visual consciousness arises, so there is no question of the awareness
acting on the object. When the opponent asks for clarification on whether, then, it is cognition
that “sees,” Vasubandhu reacts with exasperation, and points out that the opponent is seeking
precision with regard to merely conventional language:
On this, the Sautrāntika says: Why carve the ether? For, conditioned by the eye and
forms, the eye consciousness comes about. In that case, what sees, and what is seen?
For it is passive (nirvyāpāra), merely dharmas, and merely cause and effect. With
regard to this, figurative terms are used by choice with a conventional meaning: “Eye
sees, consciousness cognizes” – one should not be attached to them. For the Lord said,
“Do not be attached to the popular etymology, nor rush to accept the world’s ideas.”9
This appeal to the notion that what appears to be agent and action is really “merely dharmas, and
merely cause and effect” is one of Vasubandhu’s key philosophical moves for interpreting scrip-
tures in ways that generate unnecessary entities and relations out of grammatical conventions. Pos-
session, agency, and temporal identity are described as merely figurative, common-sense linguistic
273
Jonathan C. Gold
constructs that have been reified into ontologies. In chapter two of the Treasury Commentary,
Vasubandhu argues that if you start accepting elements this way, you end up reifying entities like
“number, size, separation, conjunction, disjunction, distance, nearness, and existence,” entities
that he insultingly designates “non-Buddhist conceptual constructions” (tīrthakaraparikalpita).10
In the Treasury Commentary’s final chapter, it is the notion of a real but ineffable “Person”
( pudgala) that is the target of this criticism. A variety of scriptures are passed back and forth
until the Personalist claims that one must acknowledge and make sense of how the Buddha
spoke of a person transmigrating and taking on the aggregates of a new life. According to
the Personalist, the “person” takes on the aggregates like a person “becomes a grammarian”
or “becomes a monk” by taking on a new status or quality. It is in response to this that Vasu-
bandhu cites the following quotation which proves, again, that the causal action of the aggre-
gates must be understood to be operating without any additional agent:
Monks, there is action and there is fruit, but no agent is perceived who casts off these
aggregates and appropriates other aggregates, because this is counter to the stipulated
meaning (saṃketa) of “dharma.”11
This very quotation is used in chapter three as a response to the very same issue (against a dif-
ferent opponent). There, he goes on to argue:
In this situation, the stipulated meaning of “dharma” is just what is dependently orig-
inated – elaborated as “when this exists, that arises.”12
Again, it is causality – the rule-governed, patterned transition from one set of elements in one
moment to another in the next – that is the foundation of the teachings and the world that the
teachings describe. Both are encompassed in the term “dharma.” For Vasubandhu, the doctrine
that all things are dependently originated serves as the foundational interpretive principle for
all of the Buddha’s teachings. In fact, since the teachings themselves are dependently origi-
nated, based in the many encounters the Buddha had with a wide variety of disciples, the lan-
guage is often not only “figurative” but requiring a secondary reading as containing a “special
intention” behind a basic teaching for a lesser disciple. This was why, for instance, the Buddha
sometimes spoke of a self being reborn – not because he did not really believe in no-self, but to
prevent people who were irreparably attached to the concept of a self from thinking that there
would be no afterlife consequences of their actions.
The principle of dependent origination, explained with the expression, “when this exists,
that arises,” brings us full circle to the argument at the opening of the chapter that established
the structure of inference and denied the possibility of self. With the nature of dependent origi-
nation in mind, we can read Vasubandhu’s opening argument to be saying not just that there
is no causal basis for a self, but that the very idea of a self is counter to the key doctrine of
dependent origination. To express things causally in a way that undermines false reification is
the quintessence of the Buddha’s Dharma.
Mind-Only, Three Natures, and Scripture (Again)

in Vasubandhu’s Yogācāra
Three principles that we have revealed in Vasubandhu’s Treasury Commentary point toward
named, thematized concepts in Mahāyāna. The causal view of perception, whereby sensory
274
Vasubandhu
organs and – especially – the mind are denied “agency” with regard to their respective objects
and are understood as dependently originated, is called “mind-only.” The distinction between
real causal entities on the one hand, and the false way that things appear – which is generated
by and superimposed upon those causal entities – is explained in terms of “three natures.” The
necessity that scripture be taken as largely “figurative,” with its truths only properly under-
stood through a reasoning that sees the dependently originated nature of all things, is referred
to by declaring all dharmas “figurative” and ultimate reality “ineffable” (acintya). These paral-
lels might make one think that Vasubandhu was a Mahāyāna advocate all along, and that is a
position held by some scholars today (Kritzer 2005). But Vasubandhu makes a point of arguing
for these positions, and this allows us to follow a process of development into Yogācāra.
The most important doctrinal innovation that distinguishes the Mahāyāna from mainstream
Buddhism, apart from the adoption of the bodhisattva path, is the claim that all entities (dhar-
mas) are as “empty” of reality as the self. Vasubandhu’s Twenty Verses on Appearance Only is
the most influential philosophical defender of this view in its Yogācāra form. At the opening of
the work, he states the position: “In the Mahāyāna the threefold world is defined as nothing but
appearance” – and appearance is said to be equivalent to “mind” or “consciousness,” together
with the qualities that minds can have.13 Everything non-mental is as unreal as the illusions
of hairs created by ailing eyes. The work is dedicated to proving this view and explaining its
consequences. I can only sketch the argument here, but we will see that the Twenty Verses
extends the ideas established in the Abhidharma analyses previously discussed and grounds
the conceptual system of the Yogācāra worldview within these principles.
At the start of the argument, Vasubandhu allows an objector to point out how, in various
ways, the world seems to be quite different from illusions. Specifically, things in the world
have spatiotemporal and intersubjective reality, and action is effective. That seems not to hap-
pen with illusions. Vasubandhu points out that all of this could still be true of an illusory world;
all of those qualities are sometimes true even in dreams. Changing tactics, then, the objector
turns to scripture and asks why, if everything is only a dream, the Buddha spoke of the material
spheres. In reply, Vasubandhu appeals to the notion of “special intention” mentioned earlier.
The aggregates and spheres are useful, he says, for countering the false view of self. Those
who are attached to the notion of self can see through this explanation that selfhood is merely a
superimposition on the spheres. But this is only the conventional, surface meaning of the doc-
trine. The Buddha spoke as if a real relation of perceiver and perceived obtained between sen-
sory organs and their objects. The real meaning of the sensory organs and objects – what the
Buddha was speaking of when he listed these pairs – can be known by advanced practitioners:
It is these two – the seed from which an appearance arises, and what appears in the
appearance – that the Sage spoke of as the twofold spheres.
(9)14
Sensory organs and their objects (“the twofold spheres”) are properly understood in a causal
sequence as, respectively, the “seed” of perception and the fruit that arises from that seed. The
formulation here is initially confusing. How is it that the sensory object – such as a visual object –
is the result in the causal story of sensation, rather than its source? Ordinarily, we would place
the visual object first, along with the visual organ, and the contact of these would cause a visual
consciousness. Here the key is noticing that the result named as the reality behind the Buddha’s
talk of an “object” is not a “thing” seen, but the visual object as it appears within an experienced
perception. The causes of a sensory experience, whatever they are, bring about some appearance,
275
Jonathan C. Gold
some consciousness, and that experience has content; it has what we called earlier the “informa-
tion” within a consciousness. The parallel with the view of the spheres in the “Why carve the
ether?” passage discussed previously is quite close. There he argued that it is not the eye but the
consciousness that “sees,” and an unspoken but implicit conclusion was that although perception
seems to be taking place in perceptual organs, in fact to the extent that there is actually percep-
tion, it is mental (see Gold 2015a, 80). Here, in an effort to explain the Buddha’s language of
sensation in a mind-only perspective, the argument is made that in reality there is no need for a
real sensory object appearing to awareness; perception only requires the appearance of a sensory
object. The sensory object is therefore the causal result of whatever dreams or illusions brought
it about. The metaphor of a seed as cause of mental events will be addressed in what follows.
It may be useful to pause to notice the parallels here between the reinterpretation of doctrine
under an indirect reading and the reduction of agency to causality. In the earlier passage, the
goal was to prove that there was no such thing as “agency” ascribable to an eye; the eye does
not “see.” And in order to prove that, Vasubandhu appealed to the “figurative” nature of the
Buddha’s scriptural language. Here, the example of the unitary causal line is used to justify the
idea that the Buddha’s apparently literal language about sensory “agency” is in fact concealing
a viable “special intention.” Among the more contentious questions in recent Buddhist philoso-
phy has been the debate over whether Yogācāra philosophers were making an ontological claim
with their mind-only doctrine, or simply an epistemological one. It is clear that Vasubandhu’s
Twenty Verses is an ontological argument; one of its central arguments disproves materiality.
Vasubandhu runs through a series of arguments that prove, to his satisfaction, that no consist-
ent account of materiality is possible, which proves, once again, that the Buddha could not
possibly have truly believed in materiality. But his parallel argument from the Treasury Com-
mentary discussed previously would seem to be arguing the epistemological version of the
argument, saying that even when we assume that there are material sensory organs and objects,
our experience of them is always only mental. It is possible, once again, to read the Treasury
Commentary as teeing up Abhidharma for the Twenty Verses’ drive into the Yogācāra position.
After the opponent’s defenses of materiality are spent, a final objection is raised as to the
karmic, which is to say moral, significance of reality under a mind-only perspective. Vasu-
bandhu says that there is no reason why causal relations and moral significance cannot be
effected in a mental mode. But in the end, he says, the reality of things under the Mahāyāna
perspective is known only by Buddhas; it is “beyond the scope of reason.”15 This is a remark-
able statement to find at the end of an elegant, compactly argued work by one of the great
philosophers of all time. (In this, it resembles nothing if not Wittgenstein’s Tractatus.) Yet,
it expresses not defeat but the doctrinal position of the ultimate ineffability of reality in the
Yogācāra. The conclusion of the Thirty Verses on Appearance (Triṃśikāvijñaptikārikā), which
is considered Vasubandhu’s companion work to the Twenty Verses, explains that the mind-only
position is only the first in a two-step process that leads to the attainment of Buddhahood.16
The final stage entails a relinquishment of all cognition: “It is without thought, without appre-
hension: and that is supramundane awareness.”17 The awareness of Buddhas is an inconceiv-
able state caused by, but beyond, the attainment of the state of mind-only.
Like the Twenty Verses, then, Vasubandhu’s Thirty Verses on Appearance (Triṃśikāvijñaptikārikā)
culminates in the ineffable. The Thirty Verses also opens like the Twenty Verses, indicating that
things are not as they seem. It says that selves and entities (dharmas) – that is, things we can
name – function to the extent that they do only within the illusory play of experience called
the “transformation of consciousness.” This seems to parallel the earlier discussions. But a key
aspect seems to have been added when the verse says that things exist figuratively:
276
Vasubandhu
A manifold figurative predication of “self ” and “things” is what functions with

respect to the transformation of consciousness.18
The Thirty Verses thus begins with a linguistic explanation of the doctrine of mind-only. Is this
something new?
In fact, it may be the Twenty Verses, with its express interest in arguments about the mate-
rial world that ground mind-only in Abhidharma disputes, that is new. Before turning to the
Thirty Verses, then, let us look at other, earlier Yogācāra scriptures and treatises – upon many
of which we have commentaries attributed to Vasubandhu. There we see that one of the most
important themes of this literature was to establish a view of language that would make sense
of the Buddha’s teachings in light of universal emptiness. One of the foundational Yogācāra
scriptures is the Scripture Clarifying the Intent (Saṃdhinirmocanasūtra), which introduces a
hermeneutic structure by which the Buddha’s teachings may be classified into “Three Turn-
ings of the Dharma Wheel.”19 The idea is that the Buddha taught doctrines in three categories
for beings of different capacities, and associated with the three turnings are, respectively, three
natures or three qualities of reality: the first turning is associated with the traditional reading of
doctrines as expressing a kind of surface meaning, which is still an expression of the emptiness
of ordinary reality; the second turning is associated with the Mahāyāna scriptures that declare
the emptiness, or negation, of the principles and entities that populate the first turning; and the
third and final turning is associated with a recognition that all realities, including the expres-
sions of the first and second turnings, are inconceivable.
Each of the three natures, then, represents a particular turn of the wheel of the doctrine, under-
stood as a distinct conventional expression of the same basic inexpressible reality, which is emp-
tiness: the Buddha taught the same emptiness in three different forms. This doctrine of the three
approaches to the nature of emptiness, also called the three natures, serves as an explanation of
the Buddha’s teachings as unified, but multi-faceted – like a three-sided crystal. The three natures
are three ways that the Buddha’s teachings have emerged, in answer to the three categories in
which the doctrinal needs of sentient beings can be satisfied. Since the Dharma has these three
natures, there is no single way to understand any given doctrine in ordinary language. A theme
of the Scripture Clarifying the Intent is that disputation over doctrines is always foolish, since
those who see things rightly (“practitioners of yoga,” yogācāras) know that all views are illusory.
In the treatises of the Yogācāra, this multi-faceted view of reality, of truth, provides a
method for interpreting the Dharma in all of its aspects. For instance, in the Commentary on
Distinguishing the Middle from the Extremes (Madhyāntavibhāgabhāṣya), one of those com-
mentaries attributed to Vasubandhu, the three natures serves exactly this purpose, as a method
for explaining and integrating different approaches to Buddhist doctrine – different facets of
each doctrinal module.20 Chapter three works this out in great detail under the rubric of “real-
ity” or “truth” (tattva). The fundamental root reality described in verse 3.3 explains the three-
fold nature in terms of existence: the first nature is truly non-existent, the second is existent but
not truly existent, and the third is truly both existent and non-existent.
The three natures should be understood as eternally non-existent, falsely existent,

and truly existent and non-existent.
(3.3b–d)21
After establishing this rubric, the chapter runs through every category and system in Buddhist
doctrine, explaining how all three of these aspects are present in every teaching. Just to give
277
Jonathan C. Gold
one example that will resonate with what we have discussed above, the Buddha’s teaching of
how the false view of self is projected onto the five aggregates is explained as enacting teach-
ings of all three kinds of emptiness, all three turnings:
It has the meanings of multiplicity, being thrown together, and being separated.
(3.17a–b)22
The Distinguishing Commentary explains “multiplicity” as the teaching of “non-existence”

in the first turning by saying that the teaching of the aggregates shows how the self lacks the
unity it appears to have. This makes sense of early Buddhist doctrine as a kind of teaching of
“emptiness” of the self. The second turning, expressed as “being thrown together,” is glossed
as providing a causal description that tells how the appearance of the self comes about from its
parts, and is therefore falsely existent. Here, dependent origination – specifically, dependent
origination of the appearance of self based in the aggregates, is a narrative of how false natures
come to appear to exist. The third turning, expressed as “being separated,” makes the point that
the separation of the aggregates, which follows from that fact that they came together through
causes and conditions, leaves a background fundamental nature of emptiness, which is the
existent truth of the non-existence of the self in each false construction. This is why the third
turning is said to be both existent and non-existent.23
This is complex and deserves further analysis; indeed, the whole chapter needs careful
analysis. But it is already clear from the present example that, when read in this mode, the
teaching of the aggregates, which is basic Buddhism, becomes not just a teaching on no-self
but also emptiness, and even more than that: when we take all three together, it becomes a
teaching on the inconceivable, simultaneous nature of appearance and emptiness.
To clarify why Buddhists would want to say that all of the teachings are, in this way,
ultimately inconceivable, I would note an interesting side point from this same commentary
(again, attributed to Vasubandhu) on the fifth verse of the fifth chapter. Chapter five explains
the bodhisattva path, and the fifth verse lists the ten perfections. The final perfection on the
tenth bodhisattva stage is awareness ( jñāna), and the commentary tells us that this awareness
consists in transcending the delusion that comes from adhering to the letter of the doctrine.24
Of course, we need to know the meaning and not the letter – but this is the last stage of the
path. How can a bodhisattva who never understood the meaning have ever gotten that far?
I hope the reader is already thinking of Vasubandhu’s view that doctrines need to be taken
as figurative, and an overly literal reading often leads one astray. In this case, though, on the
tenth bodhisattva level, we can see the culmination of this skepticism about the word of the
dharma expressed by the quintessential tenth-level bodhisattvas Avalokiteśvara and Mañjuśrī.
In the Heart Sutra, Avalokiteśvara speaks of his vision which denies the literal meanings of
the doctrine: there is “no eye, no ear, no nose, etc.” The final stage bodhisattva, with wisdom
perfected, sees that the dharma is not to be taken literally.
A key motivation for what we call early Yogācāra – the third turning of the dharma wheel –
was to counter the delusion that arises from an overly literal reading of scripture. This reha-
bilitates the doctrines of early Buddhism against the potential onslaught of nihilistic readings
that use Avalokiteśvara’s and Mañjuśrī’s scriptural negations of dharmas as a justification for
discounting their utility. Until we are Buddhas, we cannot hold all of these perspectives in our
awareness at once; until the tenth bodhisattva level, we are attached to the literal meanings of
the teachings, and we need to use language – in particular the language of the Dharma – as a
crutch. But that does not discount the fact that from the ultimate perspective, all of the Buddha’s
teachings are unified, emanating from the same, inconceivable source, which is the vision of
278
Vasubandhu
the true nature of reality. Yogācāra texts provide creative, alternative structures and strategies
to display the Dharma’s kaleidoscopic multiplicity of meanings and preclude any single, too
direct, reading of the Buddha’s teachings – whether it is an overly literal set of non-Buddhist
conceptual constructions, or a simplistic, nihilistic reading of the perfection of wisdom doctrine
that might lead one to neglect the beneficial “emptinesses” taught in the first turning.25
This is our context, then, for the opening statement in the Thirty Verses, which describes
everything as merely figurative (upacāra). Vasubandhu’s declaration ties the figurative nature of
scripture that we have seen to be a natural method for interpreting the teachings with the wider,
Yogācāra principle that all of the Buddha’s teachings serve their purposes within the nexuses of
constructed mental appearances of sentient beings. What seem to be selves and dharmas at the
level of appearance are actually based in an ever-changing flow of consciousness. From there, the
work lays out the key terms of Yogācāra doctrine under this rubric. Consciousness is constantly
transforming itself in, specifically, three different ways. The first kind of conscious flow is what is
called the store consciousness (ālayavijñāna). This is the subconscious and equanimous mentation
that, according to the text, flows like a river, and contains the seeds of the other transformations
(verses 2–4). I mentioned we would see the “seed” metaphor again, and here the “store conscious-
ness” is the “store of seeds” of the other kinds of consciousness. The second kind of transformation
is the mental consciousness (mano-vijñāna), which takes the first as its mental object, and is asso-
ciated with the afflictions and the view of the self. This is the kind of mental flow that is eliminated
by those at the stage of stream-enterer or above; in the transcendent (lokuttara) mind, there is no
mental organ (manas) (5–8a). When this “mental consciousness” is said to take the “seeds” of
the “storehouse consciousness” as its mental object, we have a direct recapitulation of the ninth
verse of the Twenty Verses discussed previously as an exemplification of the unitary causal line.
The fundamental error is here described as the activity of the organ that believes itself an agent of
perception, a subject of experience. The third kind of mental flow, finally, is the awareness of sense
objects – and in this context, Vasubandhu provides a catalog of the full efflorescence of mental
characteristics: virtuous and nonvirtuous, subtle and refined (kleśa and upakleśa) (8b–14).
Thus, the detailed analysis of the causes and conditions of the flow of consciousness pro-
vides the structure for the first half of the text. The middle of the text then integrates these three
kinds of mental transformation, and shows how they are interconnected. Verse 15, which is
the opening of the middle section, brings together the first and the third of the kinds of mental
transformation:
The five consciousnesses arise according to conditions in the root consciousness

Together or not, like waves in water.
(15)26
Sensory awarenesses can occur together or separately: you can have a smell and a taste together,
or just have one – and what you get depends upon what is happening in the root mind. The
currents in the water generate one or more waves. The analogy of waves in water expresses a
relationship between the root mind and sensation that is both cause-and-result and something
more. Because as much as waves come from the water, they also are the water. The waves are a
description of the state of the water itself – that it is wavy. At the very least, they are not separate
from it. So to say that sensory objects can be separate or together is to allow for different sensory
occurrences, but also to point out that sensation is a matter of perspective. Counting waves is an
act of constructive isolation, an activity of picking out and naming components from what is an
undifferentiated, ongoing flow. The five sensory consciousnesses, which are the awarenesses of
sensation, arise from – but are also a part of – the root mind. This interconnectedness is a key,
279
Jonathan C. Gold
sensory exemplification of the doctrine of non-duality in Yogācāra. It is an accurate description

of – a view of – what the deluded “mental consciousness” imagines and constructs about percep-
tion – the view as if – the mind were the “perceiver” and the sensation the “perceived.”
In the next two verses, Vasubandhu says that, unlike the sensory consciousnesses, the men-
tal consciousness is omnipresent in the ordinary waking mind and is always engaged in the
discrimination, the selective reification, of what appears. That is to say, mental consciousness
is always counting waves and failing to notice that the waves just are the water. This leads to
the next verse, where this same arising of appearances in the mind is given a more technical
analysis by calling it “construction” (vikalpa) and attributing it to the mutual influence of the
“seeds” that make up the root consciousness in its various transformations:
The consciousness that holds all the seeds is a transformation, this way and that,
Which proceeds through mutual influence, by which this and that construction
comes about.
(18)27
This verse is discussing the same thing as verse 15, saying both that there are causes that bring
about appearances, and that those appearances are not properly separate from the root mind of
which all of them consist. Here, the point is that all the seeds that make up the “store” of seed
consciousnesses (ālayavijñāna) undertake transformations by influencing one another in countless
ways, and it is this mutual influence itself that brings about constructions, however they appear.
The mutual influence of these consciousnesses just is the constructions. There is nothing else that
comes about over and above those consciousnesses. They are the currents and waves on the water,
interacting with one another, constantly forming new shapes. This is the point of the echoing “this
and that” (sa sa) which picks up the “this way and that” (tathā tathā) in the first line. They are two
different ways of talking about the same activity. However things come about causally, that is what
is coming about as appearances. These two levels of description are parallel and interconnected.
The next verse (verse 19) points out the karmic consequences of this activity of reifica-
tion enacted by the mental consciousness, saying that the “grasping” that separates out waves
from the water in our metaphor is what keeps saṃsāra going. It causes further maturations in
the root consciousness. This is a very important point because it reveals, conversely, how the
path to liberation is supposed to work. In order to get the grasping to stop, awareness has to
rest in the perfected nature, which is the awareness of appearance only: you have to see that
the waves just are the water. This is what happens when you, like Avalokiteśvara, attain the
transcendent awareness that constitutes a stream-enterer.
In the ending verses of the Thirty Verses, as we’ve already discussed, the point is made that
this awareness cannot be discursive – you can’t become a stream-enterer just by thinking the
thought, “this sensation is appearance only.” The transcendent (lokuttara) mind is without any
apprehension; it is devoid of thoughts (acintya). The final stage is beyond conceptual knowledge.
But before he gets there, Vasubandhu provides a very interesting line that suggests why dis-
cursive thought is always still tangled up with mistaken construction. The reason is that even
accurate thoughts are, like sensory appearances, waves on the water. In the last half of verse
22, Vasubandhu compares the three natures, which are interconnected, with the basic doctrines
of Buddhism. The complex, dizzying doctrines of Yogācāra may seem abstruse, but they share
yet another quality with the basic doctrines of Buddhism:
As is said of impermanence, etc., the one is not seen where the other is absent.
(22)28
280
Vasubandhu
Here the point is that impermanence, and the other central Buddhist doctrines of suffering and
no-self, only appear, when they do, along with the conditioning factors that they are intended
to describe and counteract. Impermanence needs to be impermanence of something – such
as a self. Impermanence is thus tied up with appearances; when no self is seen, neither is the
self’s impermanence. Impermanence is a wave on the water of the self. What we take to be a
Mahāyāna theme, the notion of non-duality, is here assimilated to three of the most central,
definitive Buddhist doctrines.
This, I take it, is one of the key points of Yogācāra thought and one that we may see expressed
throughout Vasubandhu’s work. Setting aside the worry about the unity of Vasubandhu, then,
we find common cause across these texts in the use of dependent origination understood in a
causal mode to streamline Buddhist ontology, legitimate and coordinate Buddhist scripture, and
sharpen Buddhist thinking on the interconnectedness of mind, language, and world. The project
of pointing behind the curtain of our delusions – the emptiness of our false conceptions – was
key to Buddhist doctrine all along. Every doctrine was meant to help us see through and coun-
teract this or that false construction; every doctrine was meant for our benefit, figuratively.
Notes
1 The ninth chapter of the Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya)
has been translated into English from the Sanskrit by Matthew Kapstein (2001a), James Duerlinger
(2003), and Charles Goodman (2009). The French translation of the full text from Chinese by La
Vallée Poussin (1923–1925) was translated into English by Leo Pruden (1988–1990) with reference
to the Sanskrit. See Gold (2011; 2015a) and sources noted therein for fuller treatments of Vasuban
dhu’s arguments mentioned in this chapter.
2 Translations of the Twenty Verses (Viṃśikā) in English are available by Kapstein (2001b), Anacker
(1998), and Silk (2016) and in French by La Vallée Poussin (1928–1929).
3 Pradhan (1975, 461.2–20).
4 In chapter nine of the Treasury Commentary, Vasubandhu in fact cites several passages to prove, against the
Personalist, that the Buddha declared these systems to be comprehensive (Pradhan 1975, 464.20–66.17).
5 Pradhan (1975, 461.8–9).
6 Ibid., 79.11–12.
7 This is Goodman’s translation (2009, 304).
8 I should acknowledge that I differ from most interpreters in intentionally neglecting the well-known dis-
tinction between the “Dharma” as the Buddha’s teachings and the dharmas as elements in lists in Abhi
dharma philosophical analyses. It is clear that both terms refer to the fundamental principles or realities
that the Buddha sees, understands, and articulates in various forms. To distinguish the Dharma from
dharmas is to hypothesize that the Buddha’s teachings might be expressions of something other than fun-
damental truths, or perhaps to confuse genres such as Sūtra and Vinaya genres with the Dharma expressed
through them. In any case, Vasubandhu’s definition of abhidharma (Pradhan 1975, 2.3–2.10) refers to the
fundamental reality and its analysis in treatises like his own, but not its expression in lists or teachings.
9 Pradhan (1975, 31.12–16).
10 Ibid., 79.24–25.
11 Ibid., 468.23–24.
12 Ibid., 129.11–12. The actual citation glossed there is Pradhan (1975, 129.10–11).
13 Silk (2016, 30).
14 Ibid., 67. My translations of the Twenty Verses and the Thirty Verses are greatly indebted to the ongo-
ing partnership of the Vasubandhu Translation Group (Mario D’Amato, Richard Nance, Parimal Patil,
Trina Jones, Dan Arnold, and myself).
15 Ibid., 145: tarkkāviṣaya.
16 For translations of the Thirty Verses, see Anacker (1998) for English, and Lévi (1932) for French.
17 Buescher (2007, 138): acitto ’nupalambho ’sao jñānaṃ lokottarañ ca tat |.
18 Ibid., 40: ātmadharmopacāro hi vividho yaḥ pravartate | . . . vijñānapariṇāme ’sau . . . Note: I am not
discussing Sthiramati’s commentary here.
281
Jonathan C. Gold
19 Translated into English in Powers (1995), into French in Lamotte (1935). My discussion focuses on
chapter seven.
20 See D’Amato (2012) for an English translation of the Distinguishing the Middle from the Extremes,
together with Vasubandhu’s commentary.
21 Nagao (1964, 38.1–2): asac ca nityaṃ sac cāpy atattvataḥ | sad-asat-tattvataś ceti svabhāva-traya
iṣyate || III.3
22 D’Amato (2012, 153); Nagao (1964, 45.1): anekatvābhisaṃkṣepa-paricchedārtha.
23 Nagao (1964, 45.2–5); D’Amato (2012, 153):
“With regard to the first” refers to the aggregates, understood according to three meanings. With
respect to the meaning of “multiplicity”: as it is said, “Form is dispersed across the past, present
and future”[;] hence it is not a unity. With respect to the meaning of “being thrown together”: all of
the aggregates are thrown together at once hence they are not a unity. And the aggregates may also
be understood in terms of the meaning of “being separated,” through laying out the characteristics
of form, feeling, conceptualization, dispositions, and consciousness according to their respective
fundamental natures wherein they would be seen as distinct.
24 D’Amato (2012, 176); Nagao (1964, 62.18): yathāruta-dharmma-saṃmohāpagamāt.
25 For a fuller exposition of the point of these two paragraphs, see Gold (2015b).
26 Buescher (2007, 102).
27 Ibid., 110.
28 Ibid., 126.
References
Anacker, Stefan. (1984) 1998. Seven Works of Vasubandhu, the Buddhist Psychological Doctor. Cor-
rected Edition. New Delhi: Motilal Banarsidass.
Buescher, Hartmut. 2007. Sthiramati’s Triṃśikāvijñaptibhāṣya: Critical Editions of the Sanskrit Text and
its Tibetan Translation. Philosophisch-Historische Klasse Sitzungsberichte 768. Vienna: Verlag der
Österreichische Akademie der Wissenschaften.
———. 2013. “Distinguishing the Two Vasubandhus, the Bhāṣyakāra and the Kośakāra, as Yogācāra-
Vijñānavāda Authors.” In The Foundation for Yoga Practitioners: The Buddhist Yogācārabhūmi
Treatise and Its Adaptation in India, East Asia, and Tibet, edited by Ulrich Timme Kragh, 368–96.
Cambridge, MA: Harvard University Department of South Asian Studies and Harvard University
Press.
D’Amato, Mario. 2012. Maitreya’s Distinguishing the Middle from the Extremes (Madhyāntavibhāga)
Along with Vasubandhu’s Commentary (Madhyāntavibhāga-bhāṣya): A Study and Translation. Treas-
ury of the Buddhist Sciences. New York: American Institute of Buddhist Studies.
Duerlinger, James. 2003. Indian Buddhist Theories of Persons: Vasubandhu’s “Refutation of the Theory
of a Self.” London: RoutledgeCurzon.
Frauwallner, Erich. 1951. On the Date of the Buddhist Master of the Law Vasubandhu. Rome: Istituto
Italiano per il Medio ed Estremo Oriente.
———. 1961. “Landmarks in the History of Indian Logic.” Wiener Zeitschrift für die Kunde des Süd- und
Ostasiens 5: 125–48.
Gold, Jonathan C. 2011. “Vasubandhu.” In Stanford Encyclopedia of Philosophy, edited by Edward N.
Zalta. https://plato.stanford.edu/entries/vasubandhu/.
———. 2015a. Paving the Great Way: Vasubandhu’s Unifying Buddhist Philosophy. New York: Colum-
bia University Press.
———. 2015b. “Without Karma and Nirvāṇa, Buddhism is Nihilism: The Yogācāra Contribution to the
Doctrine of Emptiness.” In Madhyamaka and Yogācāra: Allies or Rivals? edited by Jay L. Garfield
and Jan Westerhoff, 213–41. New York: Oxford University Press.
Goodman, Charles. 2009. “Vasubandhu’s Abhidharmakośa: The Critique of the Soul.” In Buddhist Phi-
losophy: Essential Readings, edited by William Edelglass and Jay L. Garfield, 297–308. New York:
Kapstein, Matthew T. 2001a. “Vasubandhu and the Nyāya Philosophers on Personal Identity.” In Rea-
son’s Traces, 347–91. Somerville, MA: Wisdom Publications.
———. 2001b. “Mereological Considerations in Vasubandhu’s ‘Proof of Idealism’.” In Reason’s Traces,
181–204. Somerville, MA: Wisdom Publications.
282
Vasubandhu
———. 2018. “Who Wrote the Trisvabhāvanirdeśa? Reflections on an Enigmatic Text and Its Place in
the History of Buddhist Philosophy.” Journal of Indian Philosophy 46: 1–30.
Kritzer, Robert. 2005. Vasubandhu and the Yogācārabhūmi: Yogācāra Elements in the
Abhidharmakośabhāṣya. Studia Philological Buddhiaca XVIII. Tokyo: International Institute for
Buddhist Studies.
La Vallée Poussin, Louis de. 1923–1925. L’Abhidharmakośa de Vasubandhu, 6 vols. Société Belge
D’Études Orientales. Paris: Geuthner.
———. 1928–1929. Vijñaptimātratāsiddhi: la siddhi de Hiuan-Tsang, traduite et annotée par Louis de
la Vallée Poussin. Paris: P. Geuthner.
Lamotte, Étienne. 1935. Saṃdhinirmocana sūtra: l’explication des mystères. Louvain: Université de
Louvain.
Lévi, Sylvain. 1932. “La Trentaine.” In Matériaux pour l’étude du système Vijñaptimātra, 61–123. Paris:
Librairie Ancienne Honoré Champion.
Nagao, Gadjin M., ed. 1964. Madhyāntavibhāga-bhāṣya: A Buddhist Philosophical Treatise Edited fror
the First Time from a Sanskrit Manuscript. Tokyo: Suzuki Research Foundation.
Olivelle, Patrick, trans. 1996. Upaniṣads. Oxford World’s Classics. New York: Oxford University Press.
Park, Changhwan. 2014. Vasubandhu, Śrilāta, and the Sautrāntika Theory of Seeds. Vienna: Arbeitskreis
für Tibetische und Buddhistische Studien, Universität Wien.
Powers, John, trans. 1995. Wisdom of Buddha: The Saṁdhinirmocana Mahāyāna Sūtra; Essential Ques-
tions and Direct Answers for Realizing Enlightenment. Berkeley, CA: Dharma Publishing.
Pradhan, Prahlad, ed. (1967) 1975. Abhidharmakośabhāṣyam of Vasubandhu. Tibetan Sanskrit Works 8.
Patna: K. P. Jayaswal Research Institute.
Pruden, Leo, trans. 1988–1990. Abhidharmakośa Bhāṣyam by Louis de la Vallée Poussin. Berkeley, CA:
Asian Humanities Press.
Silk, Jonathan A. 2016. Vasubandhu’s Viṁśikā (I): Sanskrit and Tibetan Critical Editions of the Verses
and Autocommentary; An English Translation and Annotations. Harvard Oriental Series 81. Cam-
bridge, MA: Harvard University Department of South Asian Studies and Harvard University Press
[Open Access Edition 2018].
Skilling, Peter. 2000. “Vasubandhu and the Vyākhyāyukti Literature.” Journal of the International Asso-
ciation of Buddhist Studies 23 (2): 297–350.
Takakusu, J., trans. 1904. “The Life of Vasu-bandhu by Paramārtha (A.D. 499–569).” T’oung Pao 5:
269–96.
Von Rospatt, Alexander. 1995. The Buddhist Doctrine of Momentariness: A Survey of the Origins and
Early Phase of This Doctrine Up to Vasubandhu. Alt- und Neu-Indische Studien 47. Stuttgart: Franz
Steiner Verlag.
283
17
DIGNĀGA
Early Innovator in Buddhist Epistemology1
Kei Kataoka
Introduction
The Buddhist logician Dignāga (c. 480–540/470–530 CE)2 is one of the foremost figures to
introduce full-fledged philosophical discussions of the means of valid cognition ( pramāṇa) in
Indian philosophy. A difficulty in studying Dignāga, however, is that many of his works are
not available in the original Sanskrit.3 Scholars therefore have studied his texts mainly through
Chinese and Tibetan translations, as well as some Sanskrit fragments. Portions of several of his
works have also been reconstructed using various forms of evidence. Dignāga’s final and most
influential work is the Compendium of the Means of Valid Cognition (Pramāṇasamuccaya,
PS, hereafter the Compendium), a verse treatise on which he also has written an autocom-
mentary (PSV). The recent discovery of Sanskrit manuscripts of the Indian sub-commentary
on the Compendium, namely that of Jinendrabuddhi (PSṬ), has been particularly fortunate in
that it has made possible the reconstruction of a good portion of Dignāga’s most famous work.
Among the Compendium’s six chapters, the first and the fifth chapters are now restored in
Steinkellner (2005) and Pind (2015), respectively.4
Dignāga himself describes Compendium as based on his own previous writings including
his Gateway to Logic (Nyāyamukha),5 a work on dialectics clearly modeled on his predecessor
Vasubandhu’s Method for Argumentation (Vādavidhi). A comparison of the chapter structure
of the Compendium to that of the Gateway indicates that Dignāga reorganized his philosophi-
cal system to prioritize the presentation of the means of valid cognition ( pramāṇa) – that is,
epistemology – rather than dialectics or the art of debating with others (vāda), which is con-
cerned primarily with proof (sādhana) and refutation (dūṣaṇa). This new focus can be seen
by considering the chapter structure in the Compendium, where the six chapters deal with the
following topics: (1) perception ( pratyakṣa); (2) inference-for-oneself (svārthānumāna); (3)
inference-for-others ( parārthānumāna); (4) example (dṛṣṭānta); (5) the exclusion of what is
other (anyāpoha); and (6) sophistic rejoinder ( jāti). The emphasis on the means of valid cogni-
tion is clear by their priority in the order of presentation.
In his Compendium, Dignāga emphasizes that he accepts only two means of valid cognition,
namely perception ( pratyakṣa) and inference (anumāna). He then further subdivides inference
into two kinds, inference-for-oneself and inference-for-others, thus enabling him to treat argu-
mentation under the heading of inference. He reduces verbal testimony (śabda) to inference
284 DOI: 10.4324/9781351030908-26

Dignāga
and thus denies its independent status as a separate means of valid cognition. Finally, he treats
polemical or sophistic techniques called jāti in an appendix to inference. In this way, Dignāga
unifies three domains in his Compendium: epistemology ( pramāṇa-tradition), dialectics (vāda-
tradition), and semantics (the theory of word-referents). Understanding proof and verbal testi-
mony as forms of inference constitute the hallmark of his philosophical genius and have had a
lasting influence in Buddhist philosophy. Beyond this, his particular contribution lies in treat-
ing important received ideas in a well-organized, consistent system of epistemology rather than
inventing insightful concepts out of nothing. Indeed, scholars have successfully located the
seeds of what seem, at first sight, to be Dignāga’s quite original ideas in earlier works of the Abhi‑
dharma, Sautrāntika, Yogācāra, and grammarian traditions. Vasubandhu was the most influen-
tial previous thinker for Dignāga, though this did not prevent Dignāga from critically engaging
with him.6 Some of Dignāga’s insights can also be traced back to the important grammarian
Bhartṛhari (Ogawa 2017); indeed, many of Bhartṛhari’s verses found their way into Dignāga’s
Investigation of the Three Times (Traikālyaparīkṣā) (Frauwallner 1959).
Given that its explicit goal is both to refute theories about the means of valid cognition
elaborated by other scholars and to demonstrate the correctness of his own theories, Dignāga’s
Compendium offers insight into the epistemological debates of his time. A textual analysis of
the structure of the first chapter, for example, illustrates Dignāga’s strategy in which he first
elucidates his own doctrine and then refutes other presentations. These other presentations
derive from, in order: (1) Vasubandhu’s Method for Argumentation (Vādavidhi); (2) the Apho-
risms on Logic (Nyāyasūtra) attributed to Akṣapāda Gautama; (3) the Aphorisms of Vaiśeṣika
(Vaiśeṣikasūtra) attributed to Kaṇāda; (4) the Manual of Sixty Principles (Ṣaṣṭitantra) attrib-
uted to Vārṣagaṇya together with the views of Dignāga’s elder contemporary, the Sāṁkhya
theorist Mādhava; and (5) the Aphorisms of Hermeneutical Inquiry (Mīmāṁsāsūtra) of Jai-
mini, together with the views of a “commentator,” probably Bhavadāsa. Dignāga’s Compen-
dium provoked strong positive and negative reactions from later scholars, whether from inside
or outside the Buddhist tradition; among works by the latter, Uddyotakara’s Commentary on
Logic (Nyāyavārttika) and Kumārila’s Commentary in Verse (Ślokavārttika) are the most rep-
resentative. Answering Kumārila’s severe criticism of Dignāga is probably one of the main
reasons why Dharmakīrti reorganized Dignāga’s system according to his own principles.7
Types of Cognition
Dignāga divides cognition into two basic types: nonconceptual and conceptual. The first of
these is equivalent to perception ( pratyakṣa), which Dignāga defines as cognition “devoid of
conceptualization” (kalpanāpoḍha). In the Compendium (PS I 4–6), he recognizes four dis-
tinct kinds of perception (Katsura 1984; Funayama 2000). These are: (i) perception by the five
sense-faculties; (ii-a) mental perception that deals with external objects such as color/form; (ii-
b) mental perception, a kind of self-awareness, that deals with internal feelings such as desire;
(iii) yoga practitioners’ pure intuitive perception that is not mixed with scriptural conceptu-
alization; and (iv) cognition’s self-awareness, including the case of conceptual cognition. As
for conceptual cognition, Dignāga recognizes three kinds, all of which are mental (mānasa),
meaning that they are primarily caused by a mental organ (manas). According to Dignāga, all
cognitions, whether nonconceptual or conceptual, are self-aware (svasaṃvedana) with regard
to themselves, because even a conceptual cognition of an inferred fire, for example, has direct
access to the cognition itself.
The Yogācāra scheme of the three natures (trisvabhāva) – i.e., (1) the perfected nature
( pariniṣpanna-svabhāva); (2) the dependent nature ( paratantra-svabhāva); and (3) the
285
Kei Kataoka
imagined nature ( parikalpita-svabhāva) – can be used as a framework for understanding

Dignāga’s classification of different types of cognition.8 This is true even though explicit ref-
erence to the three natures is absent in the Compendium, which can be regarded primarily as
a Sautrāntika work. The most crucial difference between the Yogācāra and the Sautrāntika
schools is that the latter accepts external reality, whereas the former holds that there is nothing
outside consciousness (i.e., everything is “consciousness-only,” or vijñāna-mātra). Addition-
ally, Sautrāntikas do not posit three natures but rather only two: (1) a substantially existent
(dravyasat) reality, which they regard as ultimately true; and (2) a nominally existent ( pra-
jñaptisat) reality, which they regard as only conventionally true or true merely as designation.
If we consider the Sautrāntika position in terms of the three natures, then, we might say that
Sautrāntikas distinguish the dependent (substantially existent) and the imagined (nominally
existent) natures but leave aside any explicit treatment of the perfected nature. Nevertheless,
while the three natures play only an implicit role in most of Dignāga’s writings,9 they still can
be a useful tool for considering his general approach to the types of cognition since we know
that Dignāga embraced a Yogācāra perspective in some of his work.
Here we can take some help from the later Buddhist philosopher Ratnākaraśānti (c. 1000
CE), who interprets the three natures respectively in his Instruction on the Perfection of Wis-
dom (Prajñāpāramitopadeśa, or PPU) as: (1) that which exists ultimately ( paramārthasat);
(2) that which exists substantially (dravyasat); and (3) that which exists merely as designation
( prajñaptisat).10 Although it may seem obvious that the three natures of the Yogācāra system
correspond to these three levels of existence beginning with paramārthasat, previous scholars
of the Buddhist logic-epistemological tradition probably felt hesitant to explicitly state this
equivalency due to the fact that only two levels are distinguished in the Sautrāntika system,
where the perfected nature, i.e., consciousness-only, has no room. Taking our inspiration from
Ratnākaraśānti, though, we can match the basic types of cognition recognized by Dignāga in
his Compendium (i–iv discussed previously) with the three natures as Ratnākaraśānti presents
them (1–3 discussed previously) as follows:
(1 = iv) Yoga practitioners’ intuitive perception of religious truths.11

(2a = i) Ordinary sense perception of color and form, smell, sound, taste, and touch.
(2b = ii-a) Mental perception of color/form and so on that immediately follows 2a.12
(2c = ii-b) Mental perception (i.e. self-awareness) of desire and so on.
(3abc) Various kinds of conceptual mental cognition (discussed in what follows).
When we classify Dignāga’s various types of cognition in terms of the three natures like this,
we can see that the border between (1) and (2) marks the difference between sages (“those
who have seen reality as it is”) and ordinary people; namely, the border between cognitions
which fall into Ratnākaraśānti’s first category (1), the perfected nature, and cognitions that
fall into the second category (2a, 2b, and 2c), the dependent nature. We can also see that all
the cognitions in the first and second categories are nonconceptual cognitions; i.e., they are
all perceptions. The line of demarcation between the second and third categories is therefore
the difference between nonconceptual and conceptual cognitions. As mentioned earlier, this
difference is the fundamental division in types of cognition for Dignāga.
Dignāga refers to three kinds of conceptual cognition, which we can consider as three divi-
sions within the imagined nature: (3a) various types of conceptual cognition that rely on previ-
ous experiences; (3b) the cognition of empirical reality (saṃvṛtisajjñāna) that superimposes
a different form onto something else; and (3c) erroneous cognition, e.g., mistaking a mirage
for “water.” In the first of these subcategories (3a), Dignāga lists inference together with its
286
Dignāga
result (ānumānika), memory, and desire. The second subcategory (3b), which does not rely on
a previous experience, can probably be called perceptual judgment (Katsura 1984, 116; 1986b,
226), e.g., the judgment “pot,” which superimposes general characteristics (sāmānyalakṣaṇa)
onto a particular characteristic (svalakṣaṇa), e.g., the absolutely unique instants and multiple
parts that are collectively subsumed under the judgment “pot.” The third subcategory (3c)
would consist of errors which do not reflect or correspond to real particular characteristics,
such as mistaking a rope for a snake.
An exceptional case of error – i.e., the erroneous cognition of what seems to be bundles
of hair and so on due to optical floaters or cataracts (taimira) – does not involve conceptual-
ization and therefore cannot be treated under the final subcategory (3c). This type of error is
mentioned by Dignāga as a type of pseudo-perception ( pratyakṣābhāsa). Later, Dharmakīrti
adds the word “non-erroneous” (abhrānta) to Dignāga’s previous definition of perception as
“devoid of conceptualization” (kalpanāpoḍha) in order to explicitly exclude this inconvenient
case from the category of perception (NB I 4).
Returning again to our classification of types of cognition according to the three natures (1, 2abc,
3abc), we can see that Dignāga groups valid inference together with invalid conceptual cognition,
such as memory, under the same category (3a). Furthermore, we note that he has difficulty handling
the exceptional case of nonconceptual erroneous cognition, which does not fit easily under the cat-
egory of 3c but also does not fit easily into the categories 2a, 2b, or 2c. These facts indicate all the
more that the demarcation line between nonconceptual and conceptual, i.e., the line between 2 and
3, has a binding force even in his (predominantly Sautrāntika) epistemology of valid cognition that
consists of perception and inference. In other words, one way of reading Dignāga is to see his main
framework as based on the old system of the three natures, in which conceptualization demarcates
the line between 2 and 3. But his new system of the means of valid cognition does not fit with this
old scheme, because inference – although valid in the new system – has to be grouped under the
third category of the old system, just like other types of invalid mental conceptualization, in as
much as it is a kind of conceptual cognition. The case of optical disease is similarly embarrassing
because its caused cognition is erroneous but nonetheless nonconceptual. We thus see that Dignāga
represents a transitional point in the development of Buddhist philosophy in which the primary
divisions of conceptual/nonconceptual and nominal/substantial do not easily line up.
Classifying Dignāga’s Works in Light of the Types of Cognition

By relating Dignāga’s hierarchical classification of the various types of cognition to the
Yogācāra scheme of the three natures, we can now shed light on his entire epistemological
project and arrange all of his works in a logical order, including those known only by name and
those extant in fragments or partial translation.13 Starting with Level 1, corresponding to the
perfected nature or the ultimately existent, as well as to the perception of sages, are his Sum-
mary of Perfection of Wisdom (Prajñāpāramitāpiṇḍārtha) and Entrance to Yoga (Yogāvatāra),
both of which deal with the path of cultivation for realizing the ultimate truth. Moving on to
Level 2, corresponding to the dependent nature or the substantially existent, as well as to the
various forms of ordinary perception, is the Lamp of Vital Points (Marmapradīpa), an abridg-
ment of Vasubandhu’s Commentary on the Treasury of the Abhidharma (AKBh), which deals
mainly with substantially existent entities.
Considered in terms of the transition from Level 1 to Level 2, we have the Investiga-
tion of the Three Times (Traikālyaparīkṣā), which borrows Bhartṛhari’s idea of the one brah-
man and its proliferation and which discusses how the absolute, i.e., cognition (vijñāna),
appears as diverse phenomena. This text explains a shift from the first to the second kind of
287
Kei Kataoka
cognition insofar as it moves from the perception of sages to that of ordinary persons. As for
the reverse transition, from Level 2 to Level 1, we have the Investigation of the Cognitive
Object (Ālambanaparīkṣā) with its autocommentary, which denies external reality (atoms and
their aggregate) in order to prove consciousness-only. We can regard the function of this text
as expounding a progression from the second level of cognition to the first insofar as it moves
from the perception of ordinary persons to that of sages.
The remaining texts all correspond to the imagined nature or the existent merely as designa-
tion, along with the various forms of conceptual cognition. Looking at Level 3a, which deals
primarily with inference and its object, there are the Two-Headed Drum of the Wheel of Reasons
(Hetucakraḍamaru) and the Gateway to Logic, as well as the Commentary on the Precepts of
Argumentation (Vādavidhānaṭīkā), and the Gateway to Logical Reason (Hetumukha). These texts
all treat inference (including argumentation) primarily from the perspective of the valid means of
cognition, whereas the Investigation of General Characteristics (Sāmānyalakṣaṇaparīkṣā) and
the Investigation of Universals (Sāmānyaparīkṣā) discuss general characteristics as word-refer-
ents. Turning to Level 3b, corresponding to cognition of empirical reality or perceptual judgment
is the Treatise on Dependent Designation (Upādāyaprajñaptiprakaraṇa), which handles what is
ordinarily taken to be empirically existent – e.g., the object of a conceptual cognition “pot” – in
order to justify the Buddha’s reference to such things in his teachings. Level 3c, corresponding
to erroneous cognition, is represented by the Hair in the Hand Treatise (Hastavālaprakaraṇa),
which discusses the famous case of the rope in a dark room erroneously taken to be a snake in
order to show the nature of conceptualization in general. The existence of the snake – i.e., the
object that erroneously appears to cognition – is denied and instead the existence of the rope is
affirmed; this corresponds to the transition from Level 3c to Level 3b. The existence of the rope,
in turn, is denied by reducing it into parts (aggregated atoms) that really exist in the second level,
corresponding therefore to the transition from Level 3b to Level 2. Ultimately, however, as the
Investigation of the Cognitive Object also proves, substantially existent entities (such as atoms)
can also be discarded as false, as is seen in the transition from Level 2 to Level 1. In this way, the
text covers the issues related to all three stages. The Compendium mainly deals with the issues
relevant to Level 2 and Level 3a, i.e., perception and inference, in addition to the articulation of
criticism of other schools as carried out previously in his now lost works such as the Investigation
of the Nyāya (Nyāyaparīkṣā), the Investigation of the Sāṁkhya (Sāṁkhyaparīkṣā), the Investiga-
tion of the Vaiśeṣika (Vaiśeṣikaparīkṣā), and the Investigation of Universals, judging from their
known titles.
Self-Awareness
The notion of self-awareness (svasaṃvedana/svasaṃvitti) refers to both the fact that and the
process by which cognition feels, senses, or is aware of itself by means of itself. It is one of the
characteristic features of Dignāga’s system. Modern scholars distinguish between two subcat-
egories of self-awareness: reflexive and intentional (Williams 1998; Kellner 2010), although
Dignāga himself does not explicitly distinguish between the two.
The idea of reflexive self-awareness, in which a cognition simultaneously grasps itself
as well as its targeted object, can be traced, before Dignāga, to the Buddhist school of the
Mahāsāṁghikas (Katsura 1976a; Yao 2005, 15). These earlier Buddhists argue that just as a
lamp simultaneously illuminates objects and itself, the Buddha’s cognition – in as much as
he is omniscient – can grasp everything, including cognition itself (T1545, 42c12–14). This
model of self-awareness is described in PS I 11cd (Kellner 2010). It is noteworthy that Dignāga
himself does not appeal to the example of the lamp when he explains the self-awareness
288
Dignāga
of ordinary cognition.14 However, his Mīmāṁsaka opponents, such as the “Commentator”

Vṛttikāra, who is quoted by Śabara and identifiable as Upavarṣa, and Kumārila, mention the
example of the lamp and fire, respectively, when they criticize the idea of self-awareness.15
Intentional self-awareness, in contrast to reflexive self-awareness, is the process by which,
within the same cognition, one aspect (ākāra) of cognition, called the grasper (grāhaka), appre-
hends another aspect, called the grasped (grāhya). This intentional structure is what makes it
possible to say that it is the same cognition that knows itself. This model fits Dignāga’s expla-
nation in PSV I 9a (Williams 1998). A precursor of this idea – namely, the notion of the two
internal aspects of cognition – is found in Yogācāra works before Dignāga. References to the
grasped and the grasper are very common, particularly when they refer to the phenomenal
world consisting of objects and subjects.16
What is the strongest motivation for Dignāga to claim that cognition feels or senses itself
by means of itself? A clue is given in PSV I 8cd, where Dignāga denies that cognition involves
any operation (vyāpāra), i.e., any dynamic process involving an action (kriyā) and factors
of action (kāraka). This idea of negating operation was explicitly advanced by Vasubandhu
(AKBh 31,12–15) when he proclaimed his Sautrāntika view that the cognitive process is not
an action carried out by an agent, but only a causal process in which “an optical cognition
arises depending on the eye and colors/forms.” In other words, Vasubandhu claims that a cog-
nitive process should be regarded as an instance of dependent origination ( pratītyasamutpāda)
among causal factors and their effect and not as a kind of action accomplished by its factors of
action (such as a separate agent, object, instrument, and so on).
According to the Vaibhāṣikas, mundane expressions such as “the eye sees” can literally be
interpreted as capturing the fact that the eye, i.e., the faculty of sight, is the agent of cogniz-
ing colors/forms. According to the Sautrāntikas, however, this is not a literal expression but
only a figurative one (upacāra). Dignāga accepts this and develops Vasubandhu’s Sautrāntika
view even further. In Dignāga’s theory of self-awareness, according to which cognition feels
or senses itself by means of itself, all elements of cognition are nothing other than cognition
and are not really distinct from each other. His arguments in support of this view can best be
understood as an attempt to convince his Brahmanical opponents – such as the Naiyāyikas – of
this fact. Naiyāyikas, for example, interpret a cognition as an action together with its factors of
action, such as the object and the instrument of knowing. Yet for Dignāga, ultimately speaking,
what happens in a cognitive process is not an action but merely the arising of an instance of
cognition endowed with two aspects, that of the grasper and that of the grasped.
The action-model of the Vaibhāṣikas, e.g., “The eye sees colors/forms,” should be revised,
as Vasubandhu states, to the dependent arising model of the Sautrāntikas, e.g., “An optical
cognition arises depending on the eye and colors/forms,” which in turn can be interpreted
according to the Yogācāra model of consciousness-only. Now, when speaking of visual cogni-
tion, all that is being asserted is that from a preceding cognition1, a subsequent instance of cog-
nition2 arises. Ultimately, in the Yogācāra view, there are only successive series of instances
of cognition; no factors external to cognition are necessary to explain it. What people call
“apprehending” or “grasping” something is not an operative process with a subject and an
object, but the mere origination of an instance of cognition. Adopting the position common to
the latter two models (which is called sākāra-vāda or the theory that a cognition is necessarily
endowed with the form/image of an object), Dignāga criticizes the first model in which the
object of cognition is outside of cognition (and which is therefore classified as nirākāra-vāda,
or the theory that a cognition is not endowed with the form/image of an object).
Thus, in contrast to the typical Brahmanical framing of the process producing a valid cog-
nition as involving a means of valid cognition ( pramāṇa), its object ( prameya), and valid
289
Kei Kataoka
cognition ( pramiti) as a result ( phala),17 Dignāga offers two different analyses of the cognitive
process based on two different models, the Yogācāra and the Sautrāntika. From the standpoint
of the Yogācāra, when the internal grasped-aspect of cognition is the object,18 the grasped-
aspect of cognition is self-cognized by its grasper-aspect, and the result is self-awareness (PS
I 9ab, 10). From the standpoint of the Sautrāntika view, however, i.e., when an external object
is the object, cognition cognizes it by “having its appearance” (viṣayābhāsatā). Sautrāntrikas
understand that it is simply this fact of a cognition having the form or appearance of an exter-
nal object that is the “means” by means of which valid cognition of an external object is
accomplished as the result (PS I 9cd).
On this second view, the fact of self-awareness, in the intentional sense that a cognition cog-
nizes an aspect of itself and does not directly cognize an external object, is disregarded, Dignāga
claims, because the form that enters cognition and the form of an external object are similar.19
One can legitimately say that an external object is cognized, although – even in the Sautrāntika
view – in reality, it is the internal aspect of cognition and not an external object that is directly
cognized. These two are ontologically different, though alleged to be similar, due to the fact
that the image is understood to be the direct effect of the external object. Thus, while cognition
actually knows an aspect of itself, it is stated to know an external object in the Sautrāntika view.
Kumārila (c. 600–650 CE), one of the prominent Mīmāṁsā philosophers, revealed Dignāga’s
trick and severely criticized him by pointing out that in Dignāga’s Sautrāntika view, the means
( pramāṇa) and the result ( phala) would have different objects (bhinnārthatva).20 This is because
the former deals with an external object, but the latter, in fact, deals only with the internal aspect
of cognition, at least according to Dignāga. As we shall see, it is only with Dharmakīrti’s further
revision of Dignāga’s theory that Kumārila’s criticism was avoided (Kataoka 2016).
As explained previously, it seems that Dignāga’s introduction of the notion of “self-
awareness” is motivated by his desire to present his own Buddhist views using the Brahmani-
cal terminology of the three factors of action (means, object, and result), which ultimately go
back to the grammatical notions of action and its factors. But as Vasubandhu stated, the fram-
ing of cognition as an action is not apt for the Buddhist. Mundane expressions of cognitive
operation such as “sees” and “cognizes” are only figurative. The cognitive process should be
explained not as an action but rather in terms of the Buddhist framing of dependent origina-
tion. Dignāga wholeheartedly agrees with Vasubandhu in this regard and he himself makes it
explicit that while cognition seemingly has an operation, in fact all the elements of a cognition
are devoid of operation (nirvyāpāra). Nonetheless, in order to clarify his views in contrast to
those of his opponents, Dignāga provisionally stands on the same ground as his opponents,
most typically that of the Naiyāyikas. For this reason, Dignāga engages the language of means,
object, and result – even while holding that these are ultimately not separate entities.
Kumārila is harsh in his condemnation of this trick, and he also severely criticized Dignāga’s
view of self-awareness21 by pointing out effectively that transitive and causative verbs, such as
“to grasp,” “to illuminate,” “to cause to shine forth,” and so on, which require a grammatical
object, necessarily imply a distinct subject and object (ŚV Śūnya 65–67ab). Insofar as transi-
tive and causative verbs such as “to feel” (vedayate) clearly require an object, one cannot but
admit the operation or action toward something else when using them. Likewise, cognition
cannot know or “grasp” itself but must cognize or “grasp” something separate from itself.
This is where Dharmakīrti’s revision – changing the model of self-awareness from transi-
tive to intransitive, from dual to nondual – proved to be critical. For Dharmakīrti, it is not the
case that cognition causes something else to shine forth ( prakāśayati) (PV III 329ab, PSṬ
I 70,12–16); rather, it spontaneously shines forth (PV III 478d: svayam eva prakāśate). The
self-luminosity or self-revelation of a cognition endowed with content is the actual nature
290
Dignāga
of the cognitive process (PV III 478, 480). The notion of self-luminosity (svayaṁprakāśa/
svaprakāśa) clarifies his view regarding cognition endowed with a form (sākārajñāna) with-
out referring to the action-model. Dharmakīrti gives up figurative expressions such as “grasps
(itself)” and “cognizes (itself)” to describe his own views and paves the way to a distinctively
Buddhist epistemology oriented to consciousness-only by introducing the notion of self-lumi-
nosity. Cognition endowed with a form spontaneously shines forth without requiring a distinct
object to be illuminated. Dignāga’s earlier attempts to articulate a Buddhist theory of self-
awareness in which the three factors of action are not different were clearly critical in allowing
Dharmakīrti to take the next step to a vision of self-awareness as pure luminosity.
Inference
As explained previously, Dignāga’s system of inference, typically exemplified by the inference
of fire from smoke,22 includes both argumentation (e.g., where one proves the property of being
transient from the property of being produced in the case of sounds and other products23) and
verbal testimony (e.g., where a hearer understands a cow from the word “cow”). As previously
noted, Dignāga applies the same method of analysis to the three domains of epistemology,
dialectics, and semantics, which in earlier Buddhist traditions had been conceived as distinct
from each other. Dignāga is not completely free from the tradition in which the analogical –
and therefore empirical – character of inference is retained, and this emphasis on the analogical
and empirical can be seen in his continued emphasis on the necessity for similar and dissimilar
examples in the presentation of a valid inference. At stake in these conversations is how to
supply a warrant for the validity of inference. How, in other words, can we be certain that a
particular reason property (e.g., being produced) is a reliable indication of a property we are
trying to prove (e.g., being transient)? The emphasis on similar and dissimilar examples allows
us to recognize Dignāga’s basic empirical orientation, which will implicitly become the target
of criticism by Dharmakīrti, who favors a more analytical approach to inferential reasoning.
The empirical attitude of Dignāga is, in fact, shared by other scholars prior to Dharmakīrti.
Dharmakīrti also severely criticizes the empirical attitude of Kumārila, who appeals to repeated
observations (darśana) for establishing the relation between reason and consequent. At the
same time, Dharmakīrti equally criticizes the empiricism of the lesser-known Buddhist thinker
Īśvarasena, who appeals to mere non-observation (adarśana) as a warrant for the absence of
a given entity (Steinkellner 1997). From the perspective of Dharmakīrti, however, one cannot
analogically infer x on the basis of thus-far repeated experiences of x; nor can one negate x just
because one has never perceived x so far (Tillemans 2004). The development of further criteria
beyond the analogical and empirical for establishing a warrant for valid inference is one of the
hallmarks of Dharmakīrti’s contribution to Buddhist argumentation. Īśvarasena, who is placed
after Dignāga and before Dharmakīrti, accepts nonperception (anupalabdhi/anupalambha) as
an independent means of valid cognition, whereas Dignāga posits only two means of valid
cognition, i.e., perception and inference. We should probably regard Īśvarasena’s revision to
Dignāga’s theory, therefore, as an attempt to make manifest the implicit presupposition of his
predecessors.24 As we will see, nonperception does play a critical role in Dignāga’s system of
inference, in particular regarding the ascertainment of invariable concomitance (vyāpti). That
is, from the perspective of Dignāga, one can conclude the absence of x if x has never been
perceived.25 A deviating counterexample, such as smoke without fire, certainly does not exist
because such a case has never been observed. This non-observation of deviation is the key to
understand Dignāga’s system of inference (Katsura 1992). In developing his theory of nonper-
ception as a separate means of valid cognition, Īśvarasena was probably trying to make further
291
Kei Kataoka
sense of Dignāga’s own theory of inference. Dharmakīrti’s further refinements would not have
been possible without the pioneering contributions of these two predecessors.
The Three Conditions of an Inferential Reason

Dignāga’s contribution to logic lies first in his sharpening of the three conditions of an inferen-
tial reason already well established in his time.26 These three conditions are:
1. The reason property (e.g., smoke) must belong to the subject (e.g., a mountain). This is
called the pakṣadharmatā, or possession of the reason property.
2. The reason property occurs only in similar examples (e.g., a kitchen). This is called the
anvaya, or coexistence.
3. The reason property does not occur in dissimilar examples (e.g., a lake). This is called the
vyatireka, or co-absence.
Taking these three conditions into account, Dignāga lists nine types of reasons and makes a
complete chart in his earliest logical work, the Two-Headed Drum of the Wheel of Reasons
(Katsura 1984, 140; 1998, 262). His adding the word “only” to the second condition is the
most crucial and original point in his analysis of a genuine reason.
Inference of fire from smoke requires the prior establishment of invariable concomitance that
wherever there is smoke, there is fire. This invariable concomitance is certified by conditions 2 and
3 in the preceding list, the so-called coexistence and co-absence. Dignāga is the first scholar who
thematized invariable concomitance through the concept of pervasion (vyāpti) so as to highlight
that a genuine reason must showcase that the domain of the reason property is pervaded by the
domain of the target property or the property to be proved. In other words, the domain of smoke
must be pervaded by that of fire if smoke is to be a legitimate reason for inferring fire. Dignāga’s
introduction of the concept of pervasion implies that he does not allow a mere analogical inference
called śeṣavat (“just like the rest” or “possessing a remainder”), i.e., a probabilistic inference based
on induction. The Essentials of Method (Fang-pien-hsin-lun) illustrates the śeṣavat type of infer-
ence with the example of inferring the saltiness of all sea water from tasting only a bit of it (T1632,
25b13–14). Another example, inferring that all the rice in a pot is cooked from sampling just a sin-
gle grain, is mentioned in Piṅgala’s commentary (T1564, 24b10) on Nāgārjuna’s Middle Way and
criticized later by Dharmakīrti as śeṣavad anumānam (in PV I 13–14; see Katsura 1986a, 41–42).
As we will see in Dignāga’s semantic theory, Dignāga prefers co-absence (vyatireka), e.g.,
the fact that there is no smoke without fire, to coexistence (anvaya), e.g., the fact that smoke
occurs with fire, as the primary method for establishing invariable concomitance. This is
because the latter, coexistence, cannot cover all infinite cases, whereas he believes the former,
co-absence, can. As mentioned previously, Dignāga presupposes, like other scholars in his time,
that a counterexample, such as the place where we find smoke without fire, does not exist if it
has never been perceived. He refers to the invariable concomitance of the reason property with
the property to be established, e.g., of smoke with fire, by means of a double negative expres-
sion avinābhāva (literally “nonexistence without”). The idea is that without fire, there can be
no smoke, thus solidifying the invariable nature of this concomitance. Dignāga inherits this
preference from Vasubandhu’s Method for Argumentation (Vādavidhi) (Katsura 1986a, 52–60),
and we see here a structural similarity to the semantic theory of exclusion of what is other
(anyāpoha), to be discussed in what follows.
In this connection, it is necessary to distinguish between contingent coexistence and invari-
able coexistence. Dignāga (PSV IV 4) refers to the former as “a reason merely existing in
292
Dignāga
(a limited number of) similar examples” (sapakṣe sadbhāvamātram) and the latter as “a reason
existing only in similar examples (and not in dissimilar examples)” (sapakṣe eva sadbhāvaḥ).27
The former does not imply that a reason is excluded from dissimilar examples, whereas the latter
does do so. The particle “only” (eva) highlights a concomitance that it is invariable and neces-
sary (avaśyam), such as the case that wherever there is smoke, there is fire.28 He also etymologi-
cally interprets (in PSV IV 2) a probans’s “following” (anugama) a probandum as its “going
everywhere” (sarvatra gamaḥ). In other words, smoke’s co-occurrence with fire is universal.
In contrast to the case of coexistence, Dignāga does not distinguish between the two types
of co-absence. He considers only invariable co-absence as the vyatireka and not contingent
co-absence. For Dignāga, the idea that “without fire, there is no smoke” indicates a universal
relationship and not a single case of co-absence by chance. It is necessarily the case that in
the absence of a probandum there is no probans (sādhyābhāve’sattvam eva). In order to estab-
lish an invariable concomitance, Dignāga adopts the negative method and rejects the positive
method based on limited experiences of coexistence, because the former method is universal.
As mentioned earlier, Dignāga makes a chart in his Two-Headed Drum of the Wheel of Rea-
sons to illustrate the existence or absence of a reason in similar and dissimilar examples either
everywhere or partially (sarvatra ekadeśe vā). Altogether he lists nine (3 × 3) possible types
of reasons (Katsura 1984, 140; 1998, 262), of which only the second and the eighth are valid.
For example, as the ninth type, he refers to amūrtatva, or formlessness, as the reason (wrongly)
applied to prove the eternality of sounds. This reason, being formless, exists in a portion of simi-
lar examples, such as ether, and it also exists in a portion of dissimilar examples, such as motion.
It is to be noted that in this chart, the coexistence that Dignāga intends is contingent and
not invariable. He does not intend existence of formlessness only in similar examples because
he clearly accepts it as existing in dissimilar examples such as ether, too. In other words, the
list of nine possible reasons presents an asymmetry between coexistence that is contingent and
co-absence that is invariable. This asymmetry is the main reason that prompts him to give pref-
erence to (invariable) co-absence and to disregard (contingent) coexistence. But it also entails
an exegetical problem that either of the two conditions of a proper reason, i.e., its existence
in similar examples or its absence in dissimilar examples, well established in the Buddhist
tradition, would be useless. As mentioned previously, he adds “only” to the second condition
in his last work (PSV IV 4) when expressing invariable concomitance, e.g., of smoke with
fire: smoke exists only in similar examples. For him, this expresses invariable co-absence,
i.e., smoke never exists without fire. But then, the third condition, invariable absence in dis-
similar examples (vipakṣe’sattvam eva), e.g., that smoke never exists wherever fire does not
exist, would be useless, because it would be already understood by implication (arthāpatti).
That smoke exists only where fire exists logically implies that smoke never exists without fire.
We can depict Dignāga as caught between the traditional account of the three conditions of
the reason and his own original contribution of adding “only” to the formulation of the second.
Table 17.1 Dignāga’s nine possible types of reasons from the Two-Headed Drum of the Wheel of Reasons.
In Dissimilar Examples
In Similar Examples existent everywhere absent partially existent
existent everywhere I II III
absent IV V VI
partially existent VII VIII IX
293
Kei Kataoka
His earlier Two-Headed Drum of the Wheel of Reasons still retains the old notion of contingent
coexistence, whereas the later Compendium already shows a new idea of invariable coexist-
ence, but Dignāga does not totally reject the traditional account of the three conditions. He
tries to find some use for either of the last two conditions of a proper reason, i.e., its existence
in similar examples and its absence in dissimilar examples, by considering five different cases
in PSV IV 4 as described in what follows (Kitagawa 1965, 251ff.; Katsura 2004, 159ff.). He
first divides the cases into two: (1) when the third condition (i.e., absence in dissimilar exam-
ples) is demonstrated; and (2) when the second condition (i.e., existence in similar examples)
is demonstrated:29
1. When the third condition is demonstrated,

1a. A similar example too should be stated to exclude a pseudo-reason.
1b. Stating a similar example is not required because invariable coexistence is implied.
2. When the second condition is demonstrated,
2a. Stating a dissimilar example is not required because invariable co-absence is implied.
2b1. The third condition should be additionally stated to confirm invariable concomitance.
2b2. The third condition is interpreted differently.
Let us examine these cases one by one.

1a. In the first case, i.e., when the third condition is demonstrated with a dissimilar exam-
ple to show the invariable concomitance, a similar example should also be stated in order
to exclude a pseudo-reason called “unique-inconclusive” (asādhāraṇa-anaikāntika). This
pseudo-reason does not prove a probandum, because it is unique and found only in the subject.
In the case of the fifth type of the nine reasons listed previously, audibility (śrāvaṇatva) is
(wrongly) applied to prove the eternality of sounds. The inference is formulated in this way:
“sounds must be eternal, because they are audible, just like x.” But audibility is found only in
sounds, the very target of the proof, and not in similar examples. At the same time, the reason
fulfills the third condition, because anything transient, except the disputed sounds, is inaudi-
ble. (Only sounds are regarded as audible in this argument.) This proof, therefore, fulfills the
first and the third conditions, but not the second. Although invariable concomitance is stated in
reference to a dissimilar example, it is also necessary to mention at least one similar example
in order to exclude this type of exceptional pseudo-reason. In this way, Dignāga shows some
use of the second condition (contingent coexistence) with consideration for the tradition.
1b. He simply admits that it is not necessary to state a similar example, because the third con-
dition that confirms invariable co-absence implies, via implication, the second one, i.e., invari-
able coexistence. In other words, 2a is the same as 1b. When a similar example is stated to show
invariable coexistence, invariable co-absence is implied – and therefore, stating a dissimilar
example is not necessary. This also means that Dignāga accepts contraposition. Either of the two
conditions would be superfluous. In both cases, Dignāga presupposes invariable coexistence.
2b1. When the second condition is demonstrated with a similar example, invariable con-
comitance cannot in fact be shown, because coexistence – if following the traditional idea – is
contingent. Therefore, the third condition should be additionally stated to confirm invariable
concomitance.
2b2. Following the new idea, however, invariable concomitance can be confirmed by the
second condition. In other words, invariable coexistence is stated with a similar example. Then,
the third condition, if normally interpreted, would be useless. Therefore, Dignāga changes the
294
Dignāga
position of “only” and tries to apologetically show the necessity of the third condition.30 His
forceful explanation, which objectively speaking is nonsense, all the more indicates that he
tries to rescue the third condition from being useless.
In his PSV, Dignāga is well aware of contraposition between the second and the third condi-
tions that entails redundancy of one of them. But he also gives consideration to the traditional
idea that both are needed, and he tries to some extent to show some use when one of them seem-
ingly becomes redundant. Again, it is only with Dharmakīrti that a paradigm change occurs
with the introduction of the completely new idea of natural binding (svabhāvapratibandha) as
the basis of a phenomenal invariable concomitance. This allows Dharmakīrti to claim that only
one of the two conditions is sufficient to show the invariable concomitance if the reason in
question fulfills the objective criteria. Without Dignāga’s innovations, however, it is unlikely
that Dharmakīrti would have been able to conceptualize his breakthrough ideas.
Antinomy
The science of dialectic had a long history before Dignāga, as gauged from Buddhist and non-
Buddhist logical works. In actual debates, there occurs a situation where neither of the two
sides – the proponent and the opponent – wins over the other. This is a case of antinomy, in
which a proponent’s proof encounters an equally strong counterproof. It results in a dialectical
impasse. In PS III 23, Dignāga claims that a proper reason should not consist of two mutually
contradicting reasons. In other words, a proper argument should not allow an equally strong
counterargument. Dignāga designates this type of inconclusive reason as “not deviating from
a contradictory reason” (viruddhāvyabhicārin), i.e., a pseudo-reason that has an equally strong
counter-reason. Effectively, he recognizes the possibility of antinomy (Kitagawa 1965, 193ff.;
Ono 2010).
Dignāga’s successor Īśvarasena clarifies Dignāga’s intention by adding three more condi-
tions to the three previously mentioned. Among them, the fifth is the condition that a proper
reason should not have a counterproof of equal argumentative weight. Debates showcase
situations when a reason, even when invalid objectively, can remain “valid” until a counter-
reason actually is presented by an opponent. Of course, this sort of subjective validity is
inadmissible for philosophers like Dharmakīrti, who insist on the discovery of truth, and not
mere victory, as the only legitimate goal of debate. In his Drop of Reasons (HB 37,15–38,
3), Dharmakīrti implicitly criticizes Dignāga’s empirical attitude when he explicitly criti-
cizes Īśvarasena’s notion of the fifth condition. For Dharmakīrti, if the proponent’s reason
passes the required criteria, it is always valid and never changes its value. If it is invalid
because of not fulfilling the required criteria, then it is invalid from the beginning and does
not have to wait for an opponent’s counterproof to render it so. In this matter too, then, we
can confirm Dignāga’s empirical attitude that a counterproof does not exist if it has not been
witnessed so far. And we can see as well how Dignāga’s ideas become the stepping stones for
Dharmakīrti’s further refinements.
The Exclusion of What Is Other

Non-observation plays a key role also in Dignāga’s semantics, i.e., in his theory of the exclu-
sion of what is other (anyāpoha), which he claims is parallel to inference (PS V 1). For
Dignāga, a word denotes its own referent just as an inferential reason property (probans) indi-
cates a target property (probandum). Dignāga’s claim implies that one establishes the connec-
tion of the word “cow” with its referent, just as one establishes the invariable concomitance of
295
Kei Kataoka
a probans, such as being produced, with a probandum, such as transiency. This connection or
invariable concomitance is based on non-observation. One has never seen something produced
without transiency. Similarly, one has never seen the word “cow” applied to non-cows. Co-
absence, which he glosses as non-occurrence in the absence of a similar case (atulye’vṛttiḥ) or
the absence of the reason property in the absence of the target property,31 can be shown only
through non-observation (adarśanamātreṇa), as Dignāga claims in PSV V 34.
Dignāga refutes four semantic theories of realist opponents and establishes his own by
demonstrating that his theory fulfills five criteria necessary for any semantic theory to func-
tion. The five criteria are as follows:
1. One can connect a word with its referent. In other words, learning the word-referent rela-
tionship is possible.
2. There is non-deviation of a word from its own referent. A word is always learned with
respect to its referent, and thus does not arouse any doubt when it is applied. It denotes its
own referent with certainty.
3. The theory must be able to explain the linguistic phenomenon of coreferentiality
(sāmānādhikaraṇya),32 such as the fact that the same thing can receive the qualifications
“an existing substance” (sad dravyam) and “an existing pot” (san ghaṭaḥ).
4. Lower concepts (i.e., subcategories) must be implied (ākṣepa), included (vyāpti), and not
excluded (anapohana). For example, the word “existing,” which denotes existing entities,
includes and implies its lower concepts such as substances and pots.
5. A word should operate toward its referent directly (sākṣādvṛtti) and not via something
else. Indirectness or dependency should be avoided so as to secure the characteristic fea-
ture of (primary) denotation that is distinguished from secondary or metaphorical usage.
Dignāga utilizes these five criteria to criticize the following four semantic theories of his real-
ist opponents: (1) the theory of particulars (bheda); (2) the theory of universals ( jāti); (3) the
theory of relationship (sambandha); and (4) the theory according to which a word denotes
anything that has a real universal (tadvat). We shall look at these each in turn.
The first semantic theory, which claims that words denote particulars, does not meet the
first and the second of the stated five criteria. It is impossible to actually connect the word
“cow” with every individual cow due to their infinite number. The word “cow,” if connected
with individual cows, would cause a doubt when applied, because a hearer would not be sure
which particular cow the word denotes.33 The second and the third theories, which claim that
words denote universals and their relationships, respectively, are both refuted by way of the
third criterion, because neither theory allows coreferentiality. For instance, according to the
second theory, the word “existing” denotes a property, namely “existence,” whereas the word
“substance” denotes not a property but rather the locus of a property.34
“existing” → existence
|
“substance” → substance
The two denotations cannot occur at the same level in a single entity co-referred to by the
two words as, for example, an “existing substance” (sad dravyam, both in the nominative).
Rather, the two denotations would have to be coordinated in an expression such as the “exist-
ing of a substance” (sad dravyasya, where the latter is in the genitive).
296
Dignāga
Having rejected the first three theories, Dignāga undertakes in PS V 4 to deny the fourth
theory, which claims that a word denotes anything that has a real universal (tadvat/jātimat) and
which also fulfills the first three criteria. The holder of this last view is no doubt the strongest
opponent for Dignāga because the structure of this theory is parallel to Dignāga’s own theory
of apohavat (literally “having exclusion”),35 i.e., the view that a word denotes anything that has
the exclusion of what is other. For example, the word “cow” expresses any cow that is excluded
from non-cows. In this way, Dignāga gets rid of the real universal cowness in his semantics.
Dignāga refutes the fourth view first by appealing to the fifth criterion, showing that in this
view, a word would be dependent on a real universal when denoting its own referents. In this
view, for example, the universal “existence” would intervene between the word “existing” and
actual existing entities. Therefore, the word would not denote its referents directly but medi-
ately or secondarily through the universal (PS V 4b: upacārāt). Furthermore, due to this media-
tion, the word “existing” cannot imply the lower concepts (subcategories), e.g., substances and
pots. Thus, the fourth view does not fulfill the fourth condition, either. By contrast, Dignāga’s
own view does not exhibit these issues. The word “existing” directly denotes existing entities
and implies substances and pots. Exclusion, which functions as a common property, is unreal
and, ontologically speaking, an absence. Thus, because exclusion is ontologically nothing, co-
referential expressions can be explained in Dignāga’s exclusion theory, whereas they cannot
be explained in the fourth theory due to the reality of the universal that causes intervention.
Another important reason why the exclusion theory is superior to the fourth theory is its
different conception of the relation (sambandha) between the word and its referent and its
resulting different conception of denotation (abhidhāna). Whereas Dignāga presupposes, so to
speak, a weak relation and a weak denotation based on the negative method of co-absence (e.g.,
the word “cow” is not applied to non-cows), his realist opponents presuppose a strong relation
and a strong denotation based on the positive method of coexistence. According to the posi-
tive theory, the word “existing” is connected only with its referent, and it is neither connected
with higher general concepts such as something knowable ( jñeya) nor with lower particular
concepts such as substances (dravya), earthen things ( pārthiva, being composed of the earth
element), trees (vṛkṣa), and so on, in the ontological hierarchy.36 A word has a strong relation
with its own definite referent, e.g., something that has the universal existence, and therefore
it targets only a particular referent and does not go beyond that boundary. Consequently, this
theory cannot explain co-referential expressions, because according to this theory, the word
“existing” does not imply a pot. It denotes only existing entities as such.
Furthermore, this theory cannot explain the linguistic phenomenon (i.e., the fourth cri-
terion) that a word brings about a certainty with regard to higher general concepts and an
uncertainty with regard to lower particular concepts. When one hears the word “tree,” for
example, one is sure that a tree is earthen, a substance, existing, and so on; but the same per-
son is uncertain whether the tree is a North Indian rosewood or a parrot tree. According to the
realist account, however, there would be no adequate explanation for why we can know with
certainty what a tree means in general but lack certainty about which particular individual tree
actually is being denoted in any given utterance of the word “tree.” The realist theory of strong
denotation cannot explain this phenomenon, because according to their theory, the denoting
power of words reaches only as far as its own directly connected referent. According to the
realist account, therefore, one would indeed have certainty about the two types of referent of
the word “tree,” i.e., both the higher general and the lower particular concepts. Rather than
being uncertain about whether the particular referent is a North Indian rosewood or a parrot
tree, then, one would – due to the strong connections between words and particular referents –
be able to ascertain that “The tree is surely a North Indian rosewood and a parrot tree,” just
297
Kei Kataoka
as one ascertains, “The tree is certainly earthen, a substance, existing, and so on.” But this is
rather absurd.
Dignāga, on the other hand, does not meet the same difficulty, because the relation and
denotation that he presupposes are weak in that they are based on co-absence confirmed by
non-observation. The word “tree” has never been seen applied to non-earthen things, and so
on. Therefore, the word can cover the complement domain (remnants; e.g., earthen things and
so on) left after the exclusion of non-earthen things and so on. The higher concepts of tree are
not excluded and thus may be included as the target domain of the word. The negative method
of confirming the semantic relationship leaves a wide semantic field possible as the referents
of a word. Thus, it can explain why the word “tree” brings about a certainty with regard to the
higher general concepts. It can also explain how it leaves in pace an uncertainty with regard to
the lower particular concepts, because the word is indifferent to (upekṣate), i.e., neither totally
excludes nor includes, its subtypes.
For Dignāga, denotation (abhidhāna) is based on the non-exclusion (anapohana) of refer-
ents. This non-exclusion is possible only in the negative semantic theory, because according to
this theory, the referent of a word is what remains after the exclusion of other referents. In other
words, it is a remainder. Non-exclusion is the basis for the inclusion of a certain referent. Due
to inclusion, a certain referent is denoted, or, more correctly speaking, it is made manifest by a
word. Dignāga himself uses the term “make manifest” (dyotayati) instead of “denote” (abhid-
hatte) in PSV V 1. This word choice indicates Dignāga’s awareness of the difference between the
realist type of strong denotation and his own theory of denotation that can also be called “making
manifest.” He probably implies that a word makes manifest and specifies a potentially existing
meaning among many that an object has, just as a prefix (upasarga) such as pra- makes manifest
a particular meaning of a verbal base that has many potential meanings. It is not the case that
a word conveys all aspects of an object.37 The word “existing” highlights a certain aspect by
excluding other referents, i.e., nonexisting entities. A word highlights only a part of an object.
A Concluding Remark
In the three domains – epistemology, dialectics, and semantics – dealt with in the Compen-
dium, nonperception plays a central role. Dignāga is well aware of the defects of limited and
contingent experiences of coexistence and instead emphasizes the function of invariable co-
absence that can cover an infinite number of cases. Later, Dharmakīrti revises Dignāga’s view
by introducing a new criterion of an inferential reason and adding a qualification “of a percep-
tible entity” to nonperception, criticizing Dignāga’s empirical attitude by way of criticizing
Īśvarasena. Dignāga’s innovations and developments, difficulties and considerations, and incon-
sistencies and limitations, are all thrown into relief when we compare him with his predecessors,
such as Vasubandhu and Bhartṛhari; his opponents, such as Kumārila; and his faithful or critical
successors, such as Īśvarasena and Dharmakīrti. Ongoing studies of the six chapters of his Com-
pendium will further elucidate his exact place in Indian intellectual history. Due to the limitation
of primary material, the present chapter has to remain a preliminary overview of his contribu-
tion. What remains abundantly clear, however, is the critical role that Dignāga played in setting
the stage for the further innovations of Dharmakīrti, his most famous intellectual successor.
Notes
1 I thank Hugo David, William Edelglass, Pierre-Julien Harter, Sara McClintock, Shinya Moriyama,
John Taber, Toshikazu Watanabe, and Alex Watson for their comments.
298
Dignāga
2 For chronology, see Frauwallner (1959); Hattori (1961a; 1968).

3 For Dignāga’s works, Frauwallner (1959) is the first reference. Hattori (1968) contains an exhaustive
list of his works. Up-to-date bibliographical details are given in the website of the EAST (Epistemol-
ogy and Argumentation in South Asia and Tibet) project (http://east.uni-hd.de/). Tsukamoto, Mat-
sunaga, and Isoda (1990) is also helpful for bibliographical information. For Dignāga’s philosophy,
Katsura’s articles are most helpful.
4 Of these, the first and the second chapters of the PSṬ are published as PSṬ I and PSṬ II respectively.
PSṬ V is being prepared by Horst Lasic and Patrick McAllister. A Japanese group headed by Shōryū
Katsura (and including Motoi Ono, Yasutaka Muroya, and Toshikazu Watanabe) is preparing the
third, fourth, and sixth chapters of PS(V) and PSṬ. Their recent articles and interim reports based on
the ongoing projects offer new insights.
5 Dignāga’s system of the Gateway to Logic thrives in east Asia via Xuanzang’s Chinese translation of
and Kuiji’s commentary on Śaṅkarasvāmin’s Entry into Logic (Nyāyapraveśa) (Radich and Lin 2014).
6 Dignāga is often critical of Vasubandhu’s Vādavidhi and prefers the system of the Vādavidhāna,
another work attributed to Vasubandhu.
7 As a result of this reorganization, Jinendrabuddhi later had to make huge efforts to reinterpret
Dignāga’s inconvenient passages in accordance with Dharmakīrti’s new system. See Pind (2015,
I xxxii); Kataoka (2016, 235).
8 In the case of an erroneous cognition of a rope as a snake, a snake is the imagined nature, whereas the
aggregated atoms of the rope is the dependent nature; and the absence of a snake in the aggregated
atoms of the rope is the perfected nature, i.e., the ultimate truth.
9 We also see an implicit reference to the three natures in some of Dignāga’s other works, includ-
ing his Hair in the Hand Treatise (Hastavālaprakaraṇa) and Treatise on Dependent Designation
(Upādāyaprajñaptiprakaraṇa) (Hattori 1961b),
10 See § 3.4.1 of the synopsis by Katsura (1976b). Ratnākaraśānti’s PPU offers us a systematic overview
of various traditions, both Buddhist and non-Buddhist, from the Yogācāra perspective as the highest
point.
11 Dignāga’s description of yogins’ pure intuitive perception as “not mixed with the teacher’s teaching”
and “not mixed with scriptural conceptualization” (PS[V] I 6cd) implies that it is primarily concerned
with religious matters.
12 For the exact nature of this mental perception, see Funayama (2000).
13 Here, my perspective is synchronic. For a diachronic analysis of his works, see Frauwallner (1959).
14 This implies that Dignāga does not understand the process of cognition with the metaphor of illumina-
tion. For Dignāga’s possible intention in introducing the notion of self-awareness, see Kataoka (2017).
15 ŚBh ad 1.1.3–5 (Frauwallner 1968, 28, 20–30, 1); ŚV Śūnya 65–67ab.
16 It is to be noted that for the Yogācāras the cognitive process (the action of grasping) is not the
focus; rather, the emphasis lies on the point that the two categories, the grasped (grāhya) and the
grasper (grāhaka), represent everything, i.e., the entire universe. Dignāga also seems to avoid men-
tioning the dynamic operation of grasping. He merely states that cognition is “self-felt by itself ”
( jñānasvasaṃvedya).
17 NBh 1,12–15. Vātsyāyana also mentions pramātṛ or the agent of valid cognition. In Buddhism, how-
ever, the agent, i.e., self (ātman), is naturally omitted. For the date of NBh, see Franco and Pre-
isendanz (1995).
18 For interpretations of this passage, see Kellner (2010); Kataoka (2016).
19 PSV I 9cd (4,13–14), Kataoka (2016, 232). Thus, the cognition of an external object fulfills the two
criteria of a genuine perception prescribed in Ālambanaparīkṣā 1–2a.
20 ŚV Pratyakṣa 79cd (cf. Taber 2005 for translation). See Kataoka (2016).
21 “A tip of finger cannot touch itself by means of itself. Similarly, cognition too cannot cognize itself
by means of its part.” A quote from Kumārila’s lost work, the Bṛhaṭṭikā (Schmithausen 1965, 133, n.
40). Cf. LAS X 568. See also T1545, 43a26–28, T1552, 953c16–17, and T1564, 17a10.
22 An ontological perspective is particularly necessary to understand Indian theories of inference.
The probans (hetu, sādhana) and the probandum (sādhya) – e.g., smoke and fire, respectively – are
regarded as properties (dharma) of the property-bearer (dharmin), e.g., the mountain, which is also
referred to as the subject ( pakṣa). Different terms originally coming from different disciplines and
traditions are used to refer to the same concepts as a consequence of unification.
23 “A sound is transient, because it is produced, just like a pot.” This is a typical example of inference
for others with regard to the two aspects of the same object which are not causally connected like
299
Kei Kataoka
fire and smoke. The invariable concomitance presupposed here is that whatever is produced has been
experienced as transient, as demonstrated in familiar examples like pots.
24 We find the same attitude in Vasubandhu. See Kellner and Taber (2014, 709): “We suggest that Vasu-
bandhu develops in the Viṃśikā an extended argumentum ad ignorantiam where the absence of exter-
nal objects is derived from the absence of evidence for their existence.” Kumārila’s pramāṇābhāva,
i.e., the sixth means of valid cognition in his system, retains a similar character. Ogawa (2014)
attempts to trace Dignāga’s adarśanamātra to the Pāṇinian notion of adarśanaṁ lopaḥ (A1.1.60).
25 Dharmakīrti’s view of the non-perception of perceptible entities (dṛśya-anupalabdhi), however, directly
contradicts Dignāga’s position. For Dharmakīrti, one cannot assert the nonexistence of x just because x is
not perceived, because there are cases when x, though existing, is not perceived due to its imperceptibility
in terms of place, time, and nature. Non-perception of x works only when x is perceptible. Non-perception,
if valid, must be qualified as “that of a perceptible entity.” Dharmakīrti rejects the old idea of mere (unqual-
ified) non-perception (adarśanamātra) of x, which does not consider whether x is perceivable or not.
26 The three conditions are known to Asaṅga in his Conforming to the Middle Way (Shun-chung-lun,
T1565, 42a24). The Ru-shi-lun (T1633, 30c20–21), which predates the Vādavidhi, is another impor-
tant Buddhist work before Dignāga that also mentions them (Katsura 1986a).
27 The sandhi in this phrase has been omitted for clarity.
28 PSV II 10: yatra dhūmas tatra agniḥ.
29 I thank Shōryū Katsura for generously sharing his unpublished reconstruction of PSV IV.
30 He interprets the third condition as sādhyābhāve eva asattvam (= asati eva nāstitā) and not sādhyābhāve ’sat-
tvam eva so that “only” functions to restrict the meaning of the negative particle to one of its three meanings,
i.e., absence, by excluding other two meanings, i.e., other and contrary. For the details, see Lasic (2009).
31 PS IV 2 (Katsura 2004, 141): sādhyābhāve ca nāstitā.
32 For coreferentiality, see Taber and Kataoka (2017).
33 Remember that in this first theory, the word denotes (ungrouped) individual cows and not their aggre-
gate. Claiming “all (individual) cows” in a group is equivalent to claiming either the second or the
fourth theory.
34 Dignāga does not take into consideration substance-ness in this context, probably for the sake of
simplicity.
35 The designation apohavat is found in Kumārila’s critique. Dignāga’s own expression is
arthāntaranivṛttiviśiṣṭa, or “qualified by the negation of other referents.”
36 For the Vaiśeṣika hierarchy of universals that Dignāga presupposes, see Katsura (1979).
37 Pind (2015, II 3, n. 14).
Bibliography and Abbreviations

Primary Sources
AKBh Abhidharmakośabhāṣya. Pradhan, Prahlad, ed. 1967. Abhidharma Kośabhāṣya of
Vasubandhu. Patna: Kashiprasad Jayaswal Research Institute.
A Aṣṭādhyāyī. Katre, Sumitra M., ed. 1989. Aṣṭādhyāyī of Pāṇini. Delhi: Motilal
Banarsidass.
JIBS Journal of Indian and Buddhist Studies.
T Taishō Shinshū Daizōkyō. The SAT Daizōkyō Text Database. http://21dzk.l.u-tokyo.
ac.jp/SAT/.
NB Nyāyabindu. Malvania, Dalsukhbhai, ed. 1971. Paṇḍita Durveka Miśra’s
Dharmottarapradīpa, Being a Sub-commentary on Dharmottara’s Nyāyabinduṭīkā,
a Commentary on Dharmakīrti’s Nyāyabindu. Patna: Kashiprasad Jayaswal
Research Institute.
NBh Nyāyabhāṣya. Thakur, Anantalal, ed. 1997. Gautamīyanyāyadarśana with Bhāṣya
of Vātsyāyana. New Delhi: Indian Council of Philosophical Research, 1997.
PPU Prajñāpāramitopadeśa.
PV Pramāṇavārttika. PV III: Tosaki (1985).
300
Dignāga
PS(V) Pramāṇasamuccaya(vṛtti).
PS(V) I Steinkellner (2005).
PS(V) V Pind (2015).
PSṬ Pramāṇasamuccayaṭīkā.
PSṬ I Steinkellner, Ernst, Helmut Krasser, Horst Lasic, and Anne MacDonald, eds. 2005.
Jinendrabuddhi’s Viśālāmalavatī Pramāṇasamuccayaṭīkā Chapter I. Part I: Critical
Edition. Part II: Diplomatic Edition. Beijing, Vienna: China Tibetology Publishing.
PSṬ II Lasic, Horst, Helmut Krasser, and Ernst Steinkellner, eds. 2013. Jinendrabuddhi´s
Viśālāmalavatī Pramāṇasamuccayaṭīkā Chapter 2: Part 1: Critical Edition. Part
2: Diplomatic Edition. Beijing, Vienna: China Tibetology Publishing House and
Austrian Academy of Sciences Press.
LAS Laṅkāvatārasūtra. P. L. Vaidya, ed. 1963. Saddharmalaṅkāvatārasūtram. Darb-
hanga: The Mithila Institute.
ŚBh Śābarabhāṣya. See Frauwallner 1968.
ŚV Ślokavārttika. Dvārikadāsa Śāstrī, ed. 1978. Ślokavārttika of Śrī Kumārila Bhaṭṭa with the
Commentary Nyāyaratnākara of Śrī Pārthasārathi Miśra. Varanasi: Tara Publications.
HB Hetubindu. Steinkellner, Ernst, ed. 2016. Dharmakīrti’s Hetubindu. Beijing, Vienna:
China Tibetology Publishing House and Austrian Academy of Sciences Pres.
Secondary Literature
Franco, Eli, and Karin Preisendanz. 1995. “A Note on Bhavadāsa’s Interpretation of Mīmāṃsāsūtra 1.1.4
and the Date of the Nyāyabhāṣya.” Berliner Indologische Studien 8: 81–86.
Frauwallner, Erich. 1959. “Dignaga, sein Werk und seine Entwicklung.” Wiener Zeitschrift für die Kunde
Süd- und Ostasiens 3: 83–164.
———. 1968. Materialien zur ältesten Erkenntnislehre der Karmamīmāṃsā. Wien: Hermann Bölaus Nachf.
Funayama, Toru. 2000. “Kamarashīra no Chokusetsu Chikaku ron niokeru I niyoru Chikaku (mānasa).”
Tetsugaku Kenkyū 569: 105–32.
Hattori, Masaaki. 1961a. “Digunāga oyobi Sono Shūhen no Nendai. Fu Sanji no Kōsatsu Wayaku.”
Bukkyō Shigaku Ronshū, Tsukamoto Hakase Shōju Kinen, 79–96. Kyoto: Tsukamoto Hakase Shōju
Kinenkai.
———. 1961b. “Digunāga ni okeru Kashō to Jitsuzai.” FAS 60: 16–28.
———. 1968. Dignāga, on Perception. Cambridge, MA: Harvard University Press.
Kataoka, Kei. 2016. “Dignāga, Kumārila and Dharmakīrti on the Potential Problem of pramāṇa and
phala Having Different Objects.” Journal of Indian Philosophy 44 (2): 229–39.
———. 2017. “Jikoninshiki no Seisei, Haikei, Henshitsu.” South Asian Classical Studies 12: 191–214.
Katsura, Shōryū. 1976a. “Sarvālambanajñāna ni tsuite.” Journal of International Buddhist Studies 24 (2):
678(160)–679(161).
———. 1976b. “A Synopsis of the Prajñāpāramitopadeśa of Ratnākaraśānti.” Journal of International
Buddhist Studies 25 (1): 487(38)–484(41).
———. 1979. “TheApoha Theory of Dignāga.” Journal of International Buddhist Studies 28 (1): 493(16)–489(20).
———. 1984. “Digunāga no Ninshikiron to Ronrigaku.” In: Kōza Daijō Bukkyō, edited by Akira
Hirakawa, Yūichi Kajiyama, and Jikidō Takasaki, vol. 9, 104–52. Tokyo: Shunjūsha.
———. 1986a. “Indo Ronrigaku ni okeru Henjū Gainen no Seisei to Hatten.” Hiroshima Daigaku Bun-
gakubu Kiyō 45: 1–122.
———. 1986b. “Dharmakīrti’s Theory of Truth.” Journal of Indian Philosophy 12: 215–35.
———. 1992. “Dignāga and Dharmakīrti on adarśanamātra and anupalabdhi.” Asiatische Studien 46/1: 222–31.
———. 1998. Indojin no Ronrigaku. Tokyo: Chūō Kōron sha.
———. 2004. “The Role of dṛṣṭānta in Dignāga’s Logic.” In The Role of the Example (Dṛṣṭānta) in Clas-
sical Indian Logic, edited by Shōryū Katsura and Ernst Steinkellner, 135–73. Vienna: Arbeitskreis für
Tibetische und Buddhistische Studien Universität Wien.
301
Kei Kataoka
Kellner, Birgit. 2010. “Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti.”

Kellner, Birgit, and John Taber. 2014. “Studies in Yogācāra-Vijñānavāda Idealism I.” Asiatische Studien
68 (3): 709–56.
Kitagawa, Hidenori. 1965. Indo Koten Ronrigaku no Kenkyū. Tokyo: Suzuki Gakujutsu Zaidan.
Lasic, Horst. 2009. “Dignāga’s Description of a Logical Mark in Pramāṇasamuccaya 2.5cd.” Interna-
tional Journal of Buddhist Thought & Culture 13: 7–22.
Ogawa, Hideyo. 2014. “Pānini Bunpōgaku Kotoba no Ryōikigai Fushiyō no Gensoku ni tsuite.” Indo
Ronrigaku Kenkyū 7: 53–78.
———. 2017 “The Qualifier-Qualificand Relation and Coreferentiality.” In Reading Bhaṭṭa Jayanta on
Buddhist Nominalism, edited by Patrick McAllister, 83–151. Wien: Verlag der Österreichischen Aka-
demie der Wissenschaften.
Ono, Motoi. 2010. “Sōiketsujō (viruddhāvyabhicārin) o megutte.” Indo Ronrigaku Kenkyū 1: 125–43.
Pind, Ole Holten. 2015. Dignāga’s Philosophy of Language: Pramāṇasamuccayavṛtti V on Anyāpoha. 2
parts. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Radich, Michael, and Chen-kuo Lin. 2014. “Introduction.” In A Distant Mirror, edited by Chen-kuo Lin
and Michael Radich, 15–31. Hamburg: Hamburg University Press.
Schmithausen, Lambert. 1965. Maṇḍanamiśra’s Vibhramavivekaḥ. Wien: Hermann Böhlaus Nachf.
Steinkellner, Ernst. 1997. “Kumārila, Īśvarasena, and Dharmakīrti in Dialogue. A New Interpretation
of Pramāṇavārttika I 33.” In Bauddhavidyāsudhākaraḥ: Studies in Honour of Heinz Bechert on the
Occasion of His 65th Birthday, edited by Petra Kieffer-Pülz and Jens-Uwe Hartmann, 625–46. Swist-
tal-Odendorf: Indica et Tibetica Verlag.
———. 2005. Dignāga’s Pramāṇasamuccaya, Chapter 1. A Hypothetical Reconstruction of the San-
skrit Text with the Help of the Two Tibetan Translations on the Basis of the Hitherto Known Sanskrit
Fragments and the Linguistic Materials Gained from Jinendrabuddhi’s Ṭīkā. https://www.oeaw.ac.at/
fileadmin/Institute/IKGA/PDF/forschung/buddhismuskunde/dignaga_PS_1.pdf
Taber, John. 2005. A Hindu Critique of Buddhist Epistemology. London and New York: Routledge.
Taber, John, and Kei Kataoka. 2017. “Coreference and Qualification: Dignāga Debated by Kumārila and
Dharmakīrti.” In The Oxford Handbook of Indian Philosophy, edited by Jonardon Ganeri, 255–71.
Tillemans, Tom. 2004. “Inductiveness, Deductiveness and Examples in Buddhist Logic.” In The Role of
the Example (dṛṣṭānta) in Classical Indian Logic, edited by Shoryu Katsura and Ernst Steinkellner,
251–75. Vienna: Wiener Studien zur Tibetologie und Buddhismuskunde.
Tosaki, Hiromasa. 1985. Bukkyō Ninshikiron no Kenkyū (ge). Tokyo: Daitō Shuppansha.
Tsukamoto, Keishō, Yūkei Matsunaga, and Hirofumi Isoda, eds. 1990. Bongo Butten no Kenkyū III Ron-
shohen. Kyoto: Heirakuji shoten.
Williams, Paul. 1998. The Reflexive Nature of Awareness. A Tibetan Madhyamaka Defense. Richmond:
Curzon Press.
Yao, Zhihua. 2005. The Buddhist Theory of Self-Cognition. London and New York: Routledge.
302
18
DHARMAKĪRTI
Philosopher and Defender of the Faith
John Taber
Dharmakīrti is considered one of the co-founders of the so-called logico-epistemological

school of Buddhist philosophy, together with his predecessor Dignāga. His contributions,
however, are not limited to epistemology and logic. He was, in fact, one of the great sys-
tematizers of Buddhist doctrine at a time in Indian history when Buddhism was coming under
increasing attack from competing traditions (Eltschinger 2014, 35–92). His works, especially
his magnum opus, Commentary on the Means of Knowledge, represent collectively a system-
atic, rational defense of the main teachings of the Buddha such as they were understood by
Dharmakīrti in his intellectual milieu, though at first glance they often appear concerned with
problems of logic, language, ontology, and psychology that have little or nothing to do with
soteriology. It is the underlying apologetic concern of his thought that I shall emphasize in this
chapter by focusing on his Commentary.1
Life and Works

Dharmakīrti’s dates are uncertain. Erich Frauwallner proposed 600–660, but other scholars
have pushed his dates back by as much as seventy years (Krasser 2012, 580–87). The only
reliable information we have about his life is that he was a famous teacher at the Buddhist
university of Nālandā, but this fame may have been posthumous. Dharmakīrti himself laments
in the final verse of his Commentary that his achievements were not duly appreciated by his
contemporaries, and this is confirmed by reports of Chinese pilgrims who visited Nālandā in
the seventh century. The delay in his renown may have had something to do with the difficulty
of his works. He is considered one of the most challenging authors of the Sanskrit canon. His
verses are extremely terse, practically like puzzles; his prose style is convoluted and his dic-
tion idiosyncratic. The explanations of commentators can vary considerably, though there is a
consistent line of interpretation from Devendrabuddhi through Śākyabuddhi, Karṇakagomin,
and Manorathanandin. Tibetan scholars felt freer to see problems in his theories and to suggest
improvements. Modern Dharmakīrti scholarship has proceeded carefully, with much effort
still devoted to recovering and editing his works. At this time, only portions of the Commen-
tary have been translated into European languages, and much of it remains poorly understood.
The summary offered here must therefore be considered provisional, subject to change as our
knowledge of Dharmakīrti advances.
DOI: 10.4324/9781351030908-27 303

John Taber
The chronological sequence of Dharmakīrti’s works proposed by Frauwallner (1954) has been
generally accepted. They are: the Commentary on the Means of Knowledge (Pramāṇavārttika,
PV, consisting of verses), his prose autocommentary on the first chapter of the same work
(Pramāṇavārttikasvavṛtti, PVSV), the Ascertainment of the Means of Knowledge (Pramāṇaviniścaya,
PVin), The Drop of Logic (Nyāyabindu), The Drop of Logical Reasons (Hetubindu), and The Rules of
Debate (Vādanyāya). Two other works, The Examination of Relations (Sambandhaparīkṣā) and Proof
of the Existence of Other Minds (Santānāntarasiddhi), belong to the same period as the Vādanyāya.
Although probably his earliest work, the Commentary is his most ambitious. It covers the
entire range of his interests and reveals how they are interrelated and arise out of his basic con-
cern to defend the central teachings of Buddhism as rationally sound. In composing the Commen-
tary, Dharmakīrti naturally built on the work of predecessors, as well as recent developments in
Indian Buddhism. Besides Dignāga, he is clearly indebted to Vasubandhu, especially the system
of Abhidharma that author works out in his Treasury of Abhidharma (Abhidharmakośa). Vasu-
bandhu, indeed, was the giant on whose shoulders both Dignāga and Dharmakīrti stood. More
broadly, Dharmakīrti was influenced by the movement of Indian Mahāyāna Buddhism known
as Yogācāra, to which Vasubandhu also contributed,2 for he tends toward an idealist form of
antirealism, which is distinctive of this “school.” Whether Dharmakīrti was really an adherent
of a school called Yogācāra, however, is a matter of controversy, since much of the time he can
simply be seen as defending the teachings of Buddhism in general, or as interpreted by more
traditional, mainstream (Śrāvakayāna and Ābhidharmika) Buddhist teachers. In crafting his
philosophical theories, he often set himself against established Brahmanical positions and criti-
cisms of Buddhist doctrines, especially those of Nyāya-Vaiśeṣika and Mīmāṃsā. His archrival
appears to have been the great Mīmāṃsā philosopher Kumārila – though he never mentions
his opponents by name – who was probably an older contemporary. Dharmakīrti’s views were
actively debated by his Buddhist and non-Buddhist successors in India up through the eleventh
century. He is considered to have been one of India’s greatest philosophers.
The Commentary, being ostensibly a commentary on Dignāga’s Compendium of the Means
of Knowledge (Pramāṇasamuccaya), is organized in four chapters corresponding roughly to the
contents of the Compendium. The question of precisely how Dharmakīrti intended it to be organ-
ized is confusing, however, due to the fact that different commentators sequence its chapters differ-
ently and none of the proposed sequences exactly matches the Compendium. The second chapter,
according to what was probably the original order, treats just the initial dedicatory verse of the
Compendium, which praises the Buddha using terms that express different aspects of his nature
as a universal savior. The third chapter of the Commentary, on perception, corresponds to the first
chapter of the Compendium, while the first chapter of the Commentary, on inference, roughly cor-
responds to the second of the Compendium and the fourth, on “inference-for-others” or proof, to the
first eight verses of the third.3 Why the chapters of the Commentary were originally sequenced dif-
ferently from the chapters of the Compendium, even though it is a commentary on it, whether in fact
Table 18.1 Organization of the chapters of Dharmakīrti’s Commentary
Compendium Commentary
Dedicatory verse Chapter 2, “The Establishment of the Means of

Knowledge”
Chapter 1, “Perception” Chapter 3, “Perception”
Chapter 2, “Inference for Oneself ” Chapter 1, “Inference for Oneself ”
Chapter 3, “Inference for Others,” vv. 1–8 Chapter 4, “Inference for Others”
304
Dharmakīrti
the above sequence is even the correct one, and why only the chapter corresponding to Dignāga’s
chapter on “inference for oneself ” (the second chapter of the Compendium) is accompanied by
Dharmakīrti’s own commentary – these are matters of speculation going back to classical times.
Buddha as a Means of Knowledge

The first of five epithets that Dignāga applies to the Buddha in his dedicatory verse is “tan-
tamount to a means of knowledge” ( pramāṇabhūta).4 The second chapter of Dharmakīrti’s
Commentary, which analyzes Dignāga’s dedicatory verse, attempts to demonstrate this claim:
the Buddha was fully awakened, knew the path to salvation, and taught it to his followers. We
can be confident, therefore, that he was and, through the scriptures that preserve his teachings,
continues to be a reliable source of knowledge with regards to not only how to escape suffer-
ing but also everything else he talked about. Dharmakīrti appropriately begins this chapter by
defining “means of knowledge” ( pramāṇa), a concept that already had a long history but had
not yet been given a general definition by a Buddhist author. We shall return to Dharmakīrti’s
theory of pramāṇa in what follows. After dispensing with the idea that God (īśvara) could be
a means of knowledge (2.10–28), he asserts that the “proof ” that the Buddha is a means of
knowledge is his compassion, cultivated over many lifetimes (2.34a); for it was his desire to
alleviate the suffering of others that drove him to practice “in many ways, without interrup-
tion, over a long time” the means of eliminating his own suffering, in particular, the perception
of selflessness (2.135–36). Having become acquainted, through his own experience, with the
various practices for eliminating suffering – that is, conditioned existence or saṃsāra – he was
eminently qualified to teach the way to others (2.131cd–132ab).
The very idea of cultivating compassion and practicing the means of eliminating suffering
over many lifetimes, however – along with the notion that the Bodhisattva (the Buddha prior
to enlightenment) was actually successful in completely and irreversibly eliminating the cause
of suffering in his own case – raises certain philosophical questions. Does desire, as the Bud-
dha claimed, really cause suffering? Is it possible to overcome desire and other moral defects
such as aversion and delusion completely? Does the vision of selflessness really eliminate it?
Indeed, is there really no self? And is there such a thing as rebirth (even though there is no self),
so that it is possible for the Bodhisattva to have carried out his practice over many lifetimes?
The second chapter is taken up with answering these and other questions, which concern
some of the underlying presuppositions of the Four Noble Truths and the Buddhist path as a
whole. Dharmakīrti suggested already in the first chapter of the Commentary that we can infer
the truth of all that the Buddha taught by confirming the “main topic” of his teaching – that
is, the Four Noble Truths (1.217). We shall consider this strategy for justifying belief in the
Buddhist scriptures in what follows. The second chapter, then, can be seen as establishing for
ourselves, using reasoning, the central teaching of the Buddha, from which we may infer the
truth of all that he said as it has been transmitted to us in the Buddhist scriptures. To be sure,
Dharmakīrti does not attempt to prove the Four Noble Truths directly; he does not attempt
to prove, for instance, that everything conditioned is really suffering. But he does attempt to
show that they cannot be overturned by considerations of psychology or metaphysics.
Critique of Physicalism
From PV 2.34–79, Dharmakīrti considers doubts about rebirth based on materialist assump-
tions, arguing that it is more reasonable to believe that cognitions – not the body – are the
cause of cognitions. In effect, he embarks on an ambitious critique of physicalism, which will
305
John Taber
continue to 2.175. At 2.34b’c, a materialist philosopher objects, “Since cognition depends

on the body, repeated practice of compassion is not established for more than one lifetime.”
Indian materialists, called Lokāyatikas or Cārvākas, believed that consciousness arises from
the elements, like the power of intoxication arises from fermentation. Moreover, they believed
that there is no world beyond this one; hence, “liberation is nothing but death.” When refuting
materialism, Dharmakīrti proceeds primarily negatively – he does not offer any formal proof
of rebirth of his own, but instead points out weaknesses in the materialist position. If con-
sciousness could arise just from the body, then “there would be absurd consequences” (2.35–
36a’), by which he seems to mean that everything would be conscious. Life and consciousness
indeed appear in a wide variety of kinds of matter. Of living beings, according to ancient
Indian belief, there are the egg-born, the sprout-born, the womb-born, and the moisture-born
(e.g., maggots coming forth from rotting food). It seems, then, that all matter is potentially
sentient. But, then, why does sentience actually appear only in certain kinds of matter and not
everywhere (2.37–38)? If one maintained that it is due to some transformation ( pariṇāma)
of the matter in those bodies in which it appears, why does that transformation occur only in
those forms of matter and not others?
There is also a lack of correlation between the body and consciousness, Dharmakīrti argues.
Cognitions are sequential; they arise successively, not all at once. But the body appears to
be relatively static, and hence unsuitable as the cause of something constantly changing. If
one were tempted to attribute change (one might say, plasticity) to the body as well in order
to allow for the possibility of a causal relation, why do that when every cognition is always
preceded by some other cognition (2.43–44)? What we observe, actually, is one stage of the
body endowed with cognition and various faculties producing an immediately following stage
endowed with cognition and faculties, which implies a causal correlation between the same
types of entities (2.36, 41ab). “Let only what is seen to be the cause always be the cause!”
Dharmakīrti urges (2.44cd). Nor could changes in the humors of the body – bile, phlegm, and
wind – account for the required plasticity (2.54 ff.).
Dharmakīrti also rejects the idea that the body could be the “support” (āśraya) of cogni-
tions, analogous to the way a plate is the support of the berries resting on it, on the grounds that
the very idea of a support is incoherent. “[The body] is not the support, because both existing
and nonexisting [things] do not rest on [anything as their support]” (2.63ab’). A non-existing
thing obviously could not have a support since it does not exist, but an existing thing would not
need one as the cause of its continuity, either, since it already exists and its continuity is noth-
ing different from its existence. Something can be called a “support” only insofar as it causes a
series of momentary entities to occur in the same place, not insofar as it props up or assists the
continued existence of an already existing entity (2.66cd–67; Franco 1997, 139–40). Rejec-
tion of the notion of a support has ramifications beyond the debate about cognition. Brahmin
philosophers, who also were not materialists, believed the self (ātman) to be the support or
substratum of mental states and, more generally, believed a substance to be the support of its
properties. There could conceivably be a support of things like water in the sense of a “con-
tainer” that prevents it from flowing away, but how could there be a container of motionless
things like qualities, universals, and actions (2.68)?
Rebirth is not the only thing necessary for the Buddha to have cultivated truly extraordinary
compassion over many lifetimes, unadulterated by opposing tendencies. Especially if it is a
matter of compassion toward all living beings, it must be able to increase indefinitely. Is that
even possible? Most qualities can only increase up to a certain limit. No matter how much
one practices jumping, one will never be able to jump a mile, and water can only be heated up
to the boiling point before it begins to evaporate. Dharmakīrti argues, however, that mental
306
Dharmakīrti
qualities like compassion have no limit because as one cultivates them, they become part of
one’s very nature and reproduce themselves without effort. Every additional effort, therefore,
will contribute to their growth so that they can increase indefinitely. Jumping, on the other
hand, requires effort at every stage just to sustain the level that has been reached; additional
efforts yield diminishing returns, while the accumulation of heat in water is limited by the
instability of its medium (2.120–31).
The cause of suffering is desire or thirst, so the elimination of suffering depends on the
elimination of desire. But what causes desire? One of the main themes of PV 2 is that desire
is caused by the false belief in a self. One overcomes suffering, then, ultimately by abandon-
ing the belief in a self (2.135, 140cd–141a, 185, 191, 199–201). This commits Dharmakīrti to
refuting all other naturalistic theories of desire – in particular, that it is a property (dharma)
of the humors (2.147cd–157) or of the elements (2.158–75). Thus, his critique of physicalism
continues. Desire, aversion, and delusion are not properties of the humors “because of dis-
crepancy” (2.147cd’). There is no correlation between dominance of a particular humor and
the manifestation of some moral defect. And if desire were a property of all the humors col-
lectively, then everyone would have the same desire. What actually causes desire to increase is
pleasure (2.151c’d). Whatever physiological state one is in when one feels desire, one cannot
at that time also experience aversion simply because desire and aversion are opposed, but
“there is no restriction in regard to the humors” (2.156). Even when phlegm predominates, one
can experience anger; even when bile predominates, one can experience desire.
By similar reasoning, Dharmakīrti says, desire, etc., cannot be a property of the elements.
Yet he introduces other arguments. “Since they have different appearances,” desire, etc., can-
not be identical with or inseparable from the elements, as the power of intoxication is insepa-
rable from liquor (2.160–61; cf. 169cd–171ab). One cannot even attribute the arising of desire
to a particular object one perceives (for instance, someone of the opposite sex), since whether
it arises or not has to do with how we conceive of the object (2.174cd).
Critique of Brahmanical Soteriology

Thus, for Dharmakīrti, desire, the cause of suffering, arises from a cognition with a particu-
lar content, ultimately from the erroneous apprehension of a self and what belongs to a self.
Hence, desire can be eliminated – the error can be corrected – and suffering really can come
to an end (2.176–78)! In PV 2, however, Dharmakīrti does not attempt to prove that there is
no self. In fact, nowhere in his corpus does Dharmakīrti work out a proof of the non-existence
of a self, per se. What he offers instead, as we shall see, is a proof that all phenomena are
momentary, which implies the non-existence of anything alleged to be permanent including,
besides a self, God, words, and universals. Elsewhere in the Commentary and in other works,
he also takes aim in passing at questionable inferences introduced by Brahmin philosophers as
proving a self – for instance, the Nyāya argument from the fact that the body is breathing (see
what follows) and the Sāṃkhya argument from the fact that the body is composite (since what
is composite is “for the benefit of another”) (Eltschinger and Ratié 2013, 117–86).
In PV 2, however, Dharmakīrti is not concerned with the question of whether the self exists,
but rather with the question of whether any notion of a self is compatible with the elimina-
tion of desire. Brahmin philosophers argue that it is not the idea of a self that is responsible
for rebirth, but attachment to one and especially to “what belongs to the self ” – the body,
mind, senses, feelings, and cognitions – or else confusion of the self with primordial nature
( pradhāna) or matter ( prakṛti), specifically, the intellect (buddhi), which is a transformation
of primordial nature. Liberation from the cycle of rebirth, which Brahmin philosophers also
307
John Taber
consider suffering, is gained by no longer mistaking body, mind, senses, and so on for the
self or by discriminating (viveka) the self from everything material, including the intellect. In
short, bondage is due not to belief in the existence of a self but to confusing self and not-self,
which engenders attachment to the latter as one’s own self, whose precarious existence one
seeks to preserve and enhance (Eltschinger and Ratié 2013, 188–99).
Dharmakīrti challenges this view of liberation at length (2.220–56) on the grounds that
attachment cannot be eliminated so long as one believes one has a self at all. One is attached
to something because one sees good qualities (guṇas) in it. Now, existence is a good quality;
therefore, so long as one believes that the self exists, and especially if one regards it as pure
and free of blemishes, one will be attached to it (2.221–22). In that case, however, one will
also be attached to the body and so forth because they are means of providing enjoyment for
the self. More basically, though, one is attached to anything one believes is one’s own. We are
not attached to the hair clippings that fall to the floor, but we love our hair when it is “ours,”
still on our heads, simply for that reason. We are even attached to a limb bitten by a snake or
a defective sense faculty so long as we consider it our own (2.226cd–228, 243–44). In short,
we do not consider body, mind, and senses to be ours because we see good qualities in them,
but we see them as having good qualities and are fond of them just because we consider them
ours. Therefore, so long as there is belief in the existence of any kind of self, attachment to
what belongs to the self cannot be broken – and suffering will continue.
Dharmakīrti further criticizes Brahmanical accounts of liberation because they envision the
self as completely dissociated from body, mind, and senses, and hence as unable to experience
anything. A self that is not an experiencer, however, is not really a self. Thus, the self-theorists
are inconsistent. They “absurdly defend the existence of a self while seeking to reach a state
where the self as they define it ceases to exist” (Eltschinger and Ratié 2013, 282). Better just
to accept that there is no self in the first place!
Knowledge that there is no self brings about a complete “transformation” of one’s moral
state, the destruction of all moral defects, including desire, together with the mental impres-
sions that give rise to them (2.205ab). This transformation is irreversible, Dharmakīrti argues.
An opponent – probably a Mīmāṃsaka, for Mīmāṃsā was highly skeptical of the claim that
there can be morally perfect, omniscient beings such as the Buddha – suggests that even some-
one for whom the practice of compassion and right conduct has become natural could relapse
and re-acquire moral defects such as desire, just as someone ruled by the defects can become
compassionate and begin to practice the Path (2.205cd’). Dharmakīrti responds that the aware-
ness of selflessness, which undermines desire and the other defects, is inherently stable. It is of
the nature of cognition to know things as they are. The mind is “naturally luminous; impurities
are adventitious” (2.208ab). One only stops knowing something if one’s belief is overturned
by another cognition that reveals things to be otherwise. You cannot “unknow” something you
know unless presented with evidence that contradicts it – and there is certainly no conclusive
evidence that there is a self (Pecchia 2015, 219–37).
The Authority of Scripture

In summary, the Buddha is a pramāṇa because, moved by compassion for all living beings,
he diligently practiced the various methods of eliminating suffering over many lifetimes until
he hit upon the right one, which finally enabled him to realize that there is no self. This in turn
purified him of all moral defects, especially desire, all of which qualifies him as a knowledge-
able, compassionate, and truthful guide for those seeking to eliminate suffering for themselves
(2.280–82; Vetter 1984, 169–71). Some of the details of this story, however, derive from the
308
Dharmakīrti
Buddhist scriptures. How do we know that they are a reliable means of knowledge, a pramāṇa?
If one answered “Because they are the word of the Buddha,” then one would obviously be
reasoning in a circle. In the first chapter of the Commentary, Dharmakīrti in fact discusses
different ways in which one can justifiably be confident in the truth of a scripture. Scripture is
distinguished from other types of discourse in that it purports to tell us about things beyond
experience; however, scripture also contains statements that we are able to verify for our-
selves. Portions of the Veda, for instance, contain magical incantations (e.g., the Atharvaveda)
and medical prescriptions (the Ayurveda). By verifying the truth of those portions of the Veda,
by trying out the prescriptions and incantations to see if they work, we can be confident in
the reliability of the author or authors of the Veda and from that infer that the entire Veda is
true. This is actually the strategy for defending the authority of the Vedic scriptures that was
adopted by Nyāya.
Dharmakīrti modifies it for his own purposes in PV 1. There, he suggests that there are two
main ways we can determine if a scripture is reliable. One is to confirm that it is right about
empirical matters; specifically, that what it says is perceptible is in fact perceptible and what it
says is not perceptible is in fact not perceptible; and that what it says is inferable independently
of scripture (such as the Four Noble Truths) is indeed so and what it says is not inferable (God,
the self) is, likewise, indeed not so. It should also be free of internal contradictions (1.215).
Dharmakīrti, in the first chapter, argues that the Veda fails to meet these criteria.
Another method is simply to confirm the reliability of a scripture in regard to “the main
topic” of its teaching relating to “what is to be obtained and what is to be avoided and the
means thereto,” that is to say, the path to liberation, from which one may infer its reliability in
regard to all transcendent matters (1.217). As we have seen, PV 2 attempts to corroborate the
Four Noble Truths not by proving them directly, but by showing that they – or at least the phil-
osophical-psychological presuppositions on which they are based – withstand rational scrutiny.
Therefore, we can be confident that everything in the Buddhist scriptures, even pronounce-
ments about matters beyond our experience, is true, just because the Buddha is their author.5
When discussing how scripture can be reliable, however, Dharmakīrti makes a surprising
concession. The apparent inference by which one concludes from the truth of one thing a per-
son says the truth of everything he says is actually a fallacy; it is the same sort of fallacy one
commits when one draws the conclusion that all of the rice in a pot is cooked from sampling
just a couple of grains. Hence, strictly speaking, scripture is not a pramāṇa, Dharmakīrti says,
or at least not a “real” one. People resort to it – justifiably – because they have no other way
of finding out about important transcendent matters such as the workings of karma (PVSV
108,1–6; 109,19–22; 167,23–168,3; 173,22–174,2). This opinion may have led to tension
between Dharmakīrti and his coreligionists – here he is supposed to be defending Buddhism,
but now he says the Buddhist scriptures are not a bona fide means of knowledge! – and could
have been one of the reasons why he chose, in his later writings, to focus on more technical
logical, epistemological, and metaphysical questions. It is his contributions to these fields – for
which he is most famous – to which we now turn.
Definition of Means of Knowledge and Levels of Reality

Despite all the discussion and debate surrounding Dharmakīrti’s concept of pramāṇa in clas-
sical and modern times, he actually says very little about it in PV 2. There he offers two
definitions, which are probably intended to be equivalent (Krasser 2001). Since ancient times,
Indian scholars had discussed the kinds of evidence that should be cited in support of theses
advanced in debate or scientific treatises. These were referred to as pramāṇas, literally, “means
309
John Taber
of knowledge”; the discussion initially focused on how many legitimate pramāṇas there are
and their individual definitions. The influential Nyāya tradition, for instance, recognized four
pramāṇas – perception, inference, scripture (more generally, verbal testimony), and compari-
son or analogy – and they offered precise definitions of each one. The Sāṃkhya school, on the
other hand, recognized only perception, inference, and scripture, and gave different definitions
of them. Dignāga, meanwhile, in his Compendium accepted only perception and inference – he
regarded scripture as a form of inference. We have already seen how Dharmakīrti, following
Dignāga, also considered the authority of scripture to be based on an inference, albeit a falla-
cious one, and again offered his own definitions while critiquing the definitions of the other
schools. The nature of a pramāṇa in general, however, had been neglected. It seems that prior
to Dharmakīrti, only the Mīmāṃsā philosopher Kumārila had ventured to offer a substantive
general definition as opposed to a mere enumeration. The definition of pramāṇa became a
matter of greater urgency now that Dharmakīrti was keen on defending Dignāga’s claim that
the Buddha was “tantamount to a pramāṇa.”
Dharmakīrti’s first definition of pramāṇa is “a nondeceptive cognition” (2.1ab’), whereby
“nondeceptive” means reliable or not leading one astray. Dharmakīrti then immediately clarifies
that non-deceptiveness consists in “the stability of the causal efficacy of the object” (2.1b’c’).
That is to say, a cognition is a pramāṇa if, acting on it, one experiences the desired effect of the
object it represents. For instance, a visual perception of a glass of water is non-deceptive if, by
prompting one to go over, pick it up, and drink it, it leads one to experience the quenching of
one’s thirst, the typical desired “causal efficacy” of water. The continuity or stability (sthiti) of
the causal efficacy of the object may refer to the fact that, having initially produced a cognition
of itself, the object of a pramāṇa (or, more precisely, a causal successor of it in the same series
of momentarily existing objects) will later produce an experience of its causal efficacy.6 Some
commentators, however, have specifically taken sthiti to mean the “capacity” ( yogyatā) of a
cognition to be a pramāṇa. Even if one does not act on the cognition, it is a pramāṇa if it would
lead one to experience the causal efficacy of its object if one did act on it. Thus, my visual per-
ception of the jewels on the neck of the queen can be a pramāṇa, even though I will never get
close enough to touch them. I know that my perception is a pramāṇa in this case – that it would
result in an experience of the causal efficacy of its object – because my perception is similar in
nature to others I have had in the past that have always been confirmed. In general, any percep-
tion of a familiar kind of object under normal circumstances will be known to be a pramāṇa (for
instance, my perception of the glass of water on the table just in front of me) without actually
having to experience the causal efficacy of its object (i.e., by picking up the glass and drinking
from it). Thus, while the paradigm case of a means of knowledge for Dharmakīrti is a cognition
which is confirmed “extrinsically” by acting on it and experiencing its object’s causal efficacy,
we know many of our cognitions to be pramāṇas “intrinsically,” according to his commentators.
Besides the “frequently repeated” type of cognition just mentioned, the cognition of the
causal efficacy of an object itself does not require further confirmation in order to be (known to
be) a pramāṇa; in fact, this is obviously required to avoid an infinite regress. For Dharmakīrti,
all our knowledge is ultimately anchored in cognitions of the causal efficacy of objects whose
status as means of knowledge is known intrinsically. In this way, he can be seen as a kind of
foundationalist, though the dichotomy foundationalism – anti-foundationalism was not explic-
itly thematized in Indian philosophy: all empirical knowledge is rooted in cognitions which,
though they may not be “infallible,” do not require justification by further cognitions in turn.
Scholars have often depicted Dharmakīrti as holding a pragmatist theory of justification, how-
ever, since his view also suggests that a cognition is a pramāṇa if one may act on it success-
fully. Finally, inferential cognitions are also known to have the status of means of knowledge
310
Dharmakīrti
“intrinsically” (that is, without having to confirm them), since in order for an inferential cog-
nition to arise at all, the knower must apprehend a necessary connection between reason and
consequent, which guarantees its truth. An inferential cognition is indeed devoid of any doubt.
Dharmakīrti’s second definition of pramāṇa is “the presentation of a previously unknown
object” (2.5c). We can see this formulation as making explicit an essential feature of a pramāṇa
that is only implicit in the first definition. In order for a cognition to be non-deceptive by virtue
of the constancy of the causal efficacy of its object, it must be able to motivate one to try to
obtain its object, at least potentially. It is, however, one’s initial apprehension of something
that motivates one to obtain it, not the second or the third apprehension; hence, a pramāṇa will
present a new object, one that is not already known. What Dharmakīrti is particularly keen
to rule out as a pramāṇa here is the “perceptual judgment” (adhyavasāya) – the conceptual
identification of an object – that arises immediately upon perceiving it. (In the case of the first
definition, this had to be ruled out by an additional phrase, PV 1.3ab’; here, in the case of the
second definition, its exclusion is implied by the definition itself.). Most Indian philosophers
believed perceptual experience to be a two-stage process: first, there is a bare awareness of the
object, a “non-conceptual perception,” then a judgment that identifies the object as a particu-
lar type of thing (a “cow”), having a certain property (“brown”) or motion (“walking”), and
so forth. According to realist philosophers – Mīmāṃsakas and Naiyāyikas – this perceptual
judgment is actually a “conceptual perception”; that is to say, it is itself a variety of percep-
tual experience that apprehends a real property of the object, hence a means of knowledge,
since perception is acknowledged by everyone to be a pramāṇa. We shall see, however, that
Dharmakīrti thinks that conceptual awareness pertains to the domain of thought, but not per-
ception; conceptual awareness removes one from an immediate experience of the object as it
is. Moreover, the properties attributed to things by thought are imagined, not real. A concep-
tual judgment, then, is not a pramāṇa by the second definition because it apprehends an object
that has already been made known by perception;7 it is in fact a variety of memory. One might
think that this second definition fails to capture the non-deceptiveness of a pramāṇa, since it
does not explicitly say that a pramāṇa is a cognition that is confirmed by the experience of
causal efficacy. But commentators suggest that this what is implied by the word “object”: an
object in the fullest sense is real, hence something whose causal efficacy can be experienced.
Dharmakīrti, following Dignāga, recognizes only two distinct pramāṇas: perception and
inference. Most Indian philosophers before Dignāga and Dharmakīrti recognized at least
scripture in addition to these two, and some identified several others (for example, compari-
son, abductive inference, and cognition of absence). Dharmakīrti justifies the reduction of the
number of pramāṇas to only two at the beginning of the third chapter of his Commentary by
declaring that there are only two kinds of object to be known: the “particular characteristic”
(svalakṣaṇa) and the “universal characteristic” (sāmānyalakṣaṇa). A particular characteristic
(or simply, particular) is that which is capable of causal efficacy and is what exists “ulti-
mately” ( paramārthasat). In contrast, what lacks causal efficacy exists only in a dissimulated
or “concealed” way (saṃvṛtisat), or as it usually rendered, “conventionally,” and the univer-
sal characteristic (or simply, universal) belongs to this domain (PV 3.3). While Dharmakīrti
largely presupposes this ontology as he works out the details of his epistemology, we shall see
that he eventually abandons it. Already, immediately after introducing the distinction between
conventionally and ultimately existing, he appears ready to concede to an opponent that causal
efficacy, the criterion of the ultimately existent, itself exists only conventionally (PV 3.4).
Indeed, in his Ascertainment of the Means of Knowledge, he will say that the theory that
defines a means of knowledge in terms of non-deceptiveness and distinguishes perception
from inference pertains only to everyday practice (sāṃvyavahārika), not ultimate reality.
311
John Taber
The purpose of his epistemology, therefore – apart from clarifying the status of the Buddha
as tantamount to a pramāṇa, and the nature of the means by which one properly reflects on his
teachings – is to refute mistaken views of non-Buddhist thinkers whose epistemological theo-
ries promote a wholly different picture of reality from the one implied by Buddhist teachings
and, hence, ultimately prevent the overcoming of suffering. At the heart of their conception of
reality are: (1) belief in the existence of real universals, which we apprehend in true judgments
as the properties of particulars; (2) belief in the existence of entities that endure through time –
besides universals, selves and other substances; and (3), most fundamentally, belief in the
existence of a world external to and independent of cognition. We shall see how Dharmakīrti’s
epistemology renders these beliefs untenable in what follows. In general, Dharmakīrti’s epis-
temology has a dominant polemical tone, at least in the Commentary, rooted in his apologetic
purpose (Eltschinger 2010, 405–6).
Universals, Exclusions, and Momentariness

Although inference, according to Dharmakīrti, does apprehend universal characteristics inso-
far as it apprehends the properties of individuals (for instance, the fieriness of a mountain
from the smoke seen rising from it), it is crucial to recognize that it only apprehends pseudo-
universals – that is to say, conceptually constructed universals, not real ones. Dharmakīrti
develops a sustained, relentless critique of real universals throughout his Commentary. Two
passages in particular, however, stand out: PV(SV) 1.143–84 and PV 3.11–50.8 Based princi-
pally on the latter, one may reconstruct his main argument as follows (cf. Franco and Notake
2014). A universal is not real, because it is the referent of a word. If it were the referent of a
word and real, then the sense faculties would become superfluous, for if a universal were real,
it would pertain to the unified, partless nature of the object that “has” it. In that case, having
already apprehended the nature of the object by a verbal cognition arising from an utterance
such as “A cow is standing in the pasture,” the senses could not tell us anything new about it;
as a result, perception would not be a pramāṇa (3.11ab).
If one maintained that the senses still could have a purpose because the word reveals just
the universal by itself – say, cowness – while the senses comprehend the individual cow, then it
must be pointed out that any awareness of a common aspect cannot really be an apprehension
of a real universal, since a universal could be neither the same as nor different from the indi-
vidual. It could not be different – this was the view of Nyāya-Vaiśeṣika; namely, a universal is
distinct from the individuals in which it “inheres” – for, just insofar as it is distinct from those
individuals, it could not be their nature and so could not really be the universal of them (PVSV
25,3–4). Nor could it be the (distinct) universal of those individuals insofar they are its support
(āśraya) or substratum. We have already seen that Dharmakīrti rejects the notion of a support
as incoherent. Moreover, something could only be called the support of something else if it
caused it to occur continuously in the same location. A universal, however, is by hypothesis
eternal; it cannot be “helped,” that is, caused to occur, by something else. Nor, for similar
reasons, could the universal be “manifested” by the individuals characterized by it (3.19–26).
If the universal, on the other hand, were not different from the individual – this was the
Mīmāṃsā and Sāṃkhya view – then all of the individuals falling under a universal would
have the same nature; namely, that of the universal. Thus, they would not really be distinct
individuals; but a universal by definition is a universal of many individuals (3.41ab; PVSV
24,24–25,3). And if an individual had both a common and a distinct nature, its nature would
not be cohesive or united (3.41cd). Thus, neither the same nor different from the individual
that “has” it, a universal is undefined; it is “without nature” (niḥsvabhāva, 3.27ab). Most
312
Dharmakīrti
significantly, Dharmakīrti insists, something having the form of a universal simply does not
appear in experience (3.48ab; cf. PVSV 39,21–23). Indeed, it is difficult to see how a univer-
sal – being eternal – could cause a cognition of itself at all.
Therefore, the universal characteristic that is known by inference is not a real universal but
something else, an “exclusion” (apoha), which is a kind of pseudo-universal. An exclusion is a
difference shared by many particulars, in themselves quite distinct from each other, from other
particulars that lack certain distinctive effects. All cows, though individually distinct from
each other, give milk; they also produce a certain cow-like cognition. The exclusion “cow” is
the difference of cows from other things that do not have such effects – horses, humans, and so
on. It is on the basis of such an exclusion that we consider cows to be the same kind of thing,
though they do not share a common nature. Quite distinct things can have the same effects,
Dharmakīrti points out, citing the example of different herbs that can reduce fever (1.74). The
fact that several things have the same effect(s) has simply to do with their individual essences
and nothing they have in common. Discerning an exclusion, however, one erroneously con-
ceives a unique particular to be the same as others; one believes there is cowness inhering in,
or not different from, all particular cows. Similarly for other properties, brownness, walking,
and so forth. However many differences a thing is conceived to have from other entities,
Dharmakīrti says, it is thought to have as many properties (1.40–42).
The attribution of common properties to particulars, which obscures their uniqueness, is
the essence of “concealment” (saṃvṛti) – that is to say, ignorance (avidyā). Concealment
defines the “conventionally existent” (saṃvṛtisat), the world of everyday activity (PV[SV]
1.68–70). Conceptual thought has other distorting effects besides obscuring uniqueness. When
one attributes multiple properties to a single particular one veils its inherently unified, part-
less nature, as one does when one distinguishes the particular as a property-bearer from its
properties (PVSV 34,23–35,9; Eltschinger et al. 2018, 63–65). The theory of apoha, then,
can perhaps be seen as the first step Dharmakīrti takes in the direction of antirealism. It calls
into question the world of objects – chairs and tables, dogs and cats, which other philosophers
consider (enduring) substances of certain types possessing multiple properties – that we take
for granted in everyday activity. It interprets them as, essentially, imagined.
The theory of exclusion was developed by Dignāga in the fifth chapter of his Compendium
as a Buddhist theory of the meaning of general terms designed to counter Brahmanical theories
of words as referring to real universals. Dharmakīrti expands and modifies the theory to meet
criticisms of Dignāga’s account in a lengthy passage of PV 1 (PV[SV] 1.40–185; Eltschinger
et al. 2018). There, he analyzes exclusions both as the meanings of words and as objects of
thought, and he explains in detail the psychological process through which exclusions are con-
ceived on the basis of experiences of particulars (cf. Dunne 2004, 116–26). Hence, although
an exclusion itself is not real – as a difference, it is essentially just a negation – it ultimately
relates to real particulars. It is for this reason that an inferential cognition, e.g., a judgment
such as “the mountain is on fire” derived from the observation that it is smoking, can be “non-
deceptive,” hence a pramāṇa, even though it actually apprehends the mountain in a distorted
way by ascribing to it a common nature which obscures its uniqueness; for to think that some-
thing is on fire is ultimately just to associate it with other (distinct) particulars that would pro-
duce the experience of certain effects under the right circumstances. In other words, it is just to
predict that, if one approached it, it would cause a sensation of heat. From this, incidentally, we
are able to see that the concept of being a means of knowledge or “epistemic instrumentality”
( prāmāṇya) does not precisely map onto the notion of truth.
Dharmakīrti also insists repeatedly throughout his discussion of apoha that exclusions are
assigned as the meanings of words by convention. There is no natural, eternal relation of a
313
John Taber
word with its meaning; rather, it is completely arbitrary. By contrast, Mīmāṃsā philosophers,
who sought to ground the authority of the Veda on its authorlessness, were committed to the
view that word, meaning, and the relation between them are eternal. A considerable portion of
PV 1 (1.213–68; Eltschinger 2007) is devoted to refuting the Mīmāṃsā doctrine of an author-
less scripture, along with the various linguistic theses that underpin it, while the final passage
of the chapter (1.312–40) seeks to demonstrate that, even if the Veda were authorless, we
could never know how to interpret it; we could never know what it really means (Eltschinger,
Krasser, and Taber 2012).
We have seen that Dharmakīrti asserts that particulars are “ultimately real” because they
have causal efficacy. The capacity to causally influence things is an entirely plausible criterion
of reality that other philosophers have embraced. Yet Dharmakīrti employs it to demonstrate a
central thesis of the Buddhism of his day that is intuitively much less plausible, the momentari-
ness of all phenomena. That is to say, things lack any temporal duration; they cease to exist
as soon as they arise. Dharmakīrti actually offers two proofs for this thesis. One is based on
the idea, accepted by all Buddhists, that things are impermanent or perishable, together with
the claim that their perishing could not be caused by outside factors. If something ceased to
exist due to outside factors, it could turn out to be eternal if those factors did not materialize.
Moreover, ceasing to exist is commonly held to be a kind of “non-existence” (abhāva), and a
non-existence cannot be caused. Things must, therefore, perish spontaneously, of themselves;
transience is part of their very nature. In that case, they will perish as soon as they arise (PVSV
98,4–22; 100,8–19; Steinkellner 1968/1969). The other proof turns on the idea that phenomena
are momentary due simply to the fact that they exist, which implies that they are causally effica-
cious. Dharmakīrti focuses on establishing the contrapositive of the statement, “All that exists
is momentary”; namely, “All that is non-momentary is nonexistent.” He argues that something
that is not momentary could not produce an effect either gradually or all at once. Something
non-momentary that is capable of producing an effect would not delay in producing it; if it did
not produce it as soon as it existed, insofar as its causal capacity is part of its nature, it would
not do so later, since that would involve a change in its nature. Hence, it would not produce its
effect gradually. But if it could produce its effect in a single moment, it would do so continu-
ously, as long as it exists (which is never observed). Since, then, a non-momentary entity could
not produce an effect in either of these ways, it would not exist (PVin 79,1–4; 80,1–6).
The proof of momentariness is often seen as the cornerstone of Dharmakīrti’s metaphys-
ics. It supplements his refutations of universals and the various eternalist linguistic theses that
underlie the doctrine of an authorless Veda, and he may have felt that it made any explicit
proof of the non-existence of a self unnecessary.
Perception
In addition to a general definition of a means of knowledge, Dharmakīrti, of course, provides
individual definitions of each of the two pramāṇas he recognized. In defining perception, he
stays close to Dignāga’s definition, “a cognition free from conceptual construction,” adding
only the qualifying expression “non-erroneous” to rule out cognitions arising from a sense fac-
ulty that is diseased or defective, but which are nevertheless non-conceptual: “Perception is a
cognition that is free from conceptual construction and non-erroneous” (PVin 1.4ab’). The core
idea is that a perception is a bare awareness of a particular as it is, which takes in all aspects of
its nature – which is nonetheless “unified,” in that its various aspects are somehow inseparable
from each other – without identifying it in any determinate way according to a genus, quality,
motion, accessory (as in the determination “a man with a stick”), or proper name.
314
Dharmakīrti
While Dignāga understands “conceptual construction” (kalpanā) to be the association of a

word with the object, Dharmakīrti takes it to be a cognition that represents something capable
of being associated with a verbal designation (abhilāpasaṃsargayogya). As already mentioned,
non-Buddhist philosophers believed an initial “non-conceptual” perception to be followed imme-
diately by a “conceptual” one, which is caused by the former and merely articulates (typically,
by the application of words) real features of the object that is non-conceptually apprehended.
Obviously, Dharmakīrti cannot allow this second, conceptual awareness to be a perception. He
has already attempted to rule it out, if only in an ad hoc way, by defining pramāṇa (explicitly in
his second formulation, implicitly in the first) as a cognition that presents a previously unknown
object. More fundamentally, however, we have seen previously that Dharmakīrti considers con-
cepts to misrepresent or “conceal” the true nature of particulars. While it may be possible for
inference to misrepresent its object by apprehending it conceptually, yet still be a pramāṇa inso-
far as it ultimately leads to an experience of its causal efficacy, there must be some mode of
awareness that puts us directly and faithfully in touch with things (i.e., that reveals how they
really are) for us to be confident that empirical knowledge connects us up with the world.
In PV 3.123–191ab and PVin 1.5–14, however, Dharmakīrti raises additional considera-
tions to show that the conceptual awareness or “perceptual judgment” that arises immediately
upon the (non-conceptual) perception of a particular lacks the qualities we would normally
associate with a perception. First, Dharmakīrti maintains that it is self-evident that perceptions
are devoid of conceptual content. As we shall see, he believes that all cognitions are “self-
aware”; they cognize themselves along with their objects. Thus, every cognition reveals its
own character as a cognition as it occurs. A perceptual cognition immediately reveals itself
as the awareness of some object-form devoid of conceptualization. This is confirmed by the
experience of meditation. Even when thought has completely ceased, one can still have visual
perceptions (presumably, meditating with the eyes open) (PV 3.124–25). Second, as we have
seen, conceptual awareness involves the application of an expression to what is being cog-
nized, at least potentially. Since a verbal expression cannot refer to a temporally and spatially
restricted particular – for a linguistic convention could not be established in regard to such an
ephemeral thing – conceptual awareness cannot have a particular as its object. Therefore, a
perception, which apprehends a particular, cannot be conceptual (PV 3.126–28). If one held,
however, that perception of a particular triggers a memory of the linguistic convention that
relates a certain expression to particulars of that type, then the conceptual awareness would
still be “removed” from the object. It would not be an immediate, direct awareness of it, but
rather an awareness processed by the mind (PVin 1.5–7). Finally, and perhaps most funda-
mentally, perceptual experiences are vivid, while conceptual awarenesses – occupied not with
particulars but the meanings of words – are not (PVin 1.15; PV 3.130). Dharmakīrti advances
many other arguments in support of this thesis.
Dharmakīrti identifies four types of perception. Besides sense perception there is mental
perception, of somewhat mysterious status (it seems to be a sort of non-conceptual mental
image of the next moment of the object apprehended by a sensory perception); the self-aware-
ness of any kind of cognition (even a conceptual awareness), already mentioned; and yogic
perception, the direct experience of an object, which is the culmination of meditation on it
and which presents it in a vivid, non-conceptual way (3.285). The latter is of particular impor-
tance for Dharmakīrti. He notes that anything we think about intensely and long enough will
eventually present itself as if it were immediately in front of us – even an “unreal” (abhūta)
object, such as an imagined corpse one is meditating on in order to reduce desire for the body.
Dharmakīrti held that the Buddha achieved such a direct perception of the real features of
things – selflessness, momentariness, and suffering, as well as the Four Noble Truths – through
315
John Taber
constant meditation on them, having first known them through insight gained by understand-
ing the meaning of scripture and rational reflection (Eltschinger 2010, 422–23).
Antirealism
Dharmakīrti’s epistemology, as presented thus far, secures a world of causally efficient, hence
momentary, particulars which are in themselves utterly unique, that is, completely distinct
from each other, but having properties attributed to them on the basis of the observation of
their causal powers, which define common “differences” among them. Such an account is
consistent with the world-view of certain traditional (mainstream Buddhist) orientations that
Dharmakīrti wished to provide room for in his system; it is, incidentally, also potentially con-
sistent with certain modern versions of empiricism. In the latter half of the third chapter of his
Commentary, however, he embraces a Mahāyāna perspective, rejecting the reality of a world
of external particulars and advocating instead the distinctive Yogācāra teaching that every-
thing is just consciousness (vijñaptimātra). There are three components to his demonstration
of consciousness-only: (1) a deconstructive analysis of the object of perception; (2) proof of
the thesis that the object of a cognition is not distinct from the cognition; and (3) development
of the implications of the idea that all cognition is self-aware.
(1) When discussing, at PV 3.197–219, the object of perception, which presents itself as
something spatially extended – a tree, a piece of fruit, and so forth – Dharmakīrti initially
entertains the idea that it is a collection of atoms which, though individually invisible, become
visible when they arise in close proximity to each other. (For Dharmakīrti, following Abhid-
harma thought, atoms are not eternal, partless substances but momentary, autonomous proper-
ties – solidity, fluidity, and so forth.) In that case, the object is something variegated; it consists
of many visible atoms, like a pile of sesame seeds. How, then, does it appear as one thing with
a single, extended form – a tree or a cow?
Rejecting the Nyāya proposal that the object is a single whole (avayavin) inhering in many
parts, he also questions whether its oneness can be accounted for by the cognition that appre-
hends it. In order to cognize an object the cognition must somehow bear the image, or assume
the form, of the object it represents. If the object is variegated, the cognition, too, must be
variegated. But if the cognition itself is variegated, it cannot be one; nor can it really even
be variegated (3.210ab) – here his precise argument is obscure, but his commentator Mano-
rathanandin suggests that it has to do with the fact that its diverse aspects are not able to
perceive each other, “like the cognitions of many [different] persons.” The cognition, then,
can be neither one nor many; nor, by implication, can the object be either one or many. Alas,
Dharmakīrti says, “What the wise say is forced upon us by the facts: the more we reflect on
things, the more they dissolve!” (3.209).
When we press the question, then, neither the object nor the cognition can have extension,
yet they both appear to be extended. Hence, neither is real. Now, consciousness is structured
by the opposition between an object and the cognition which apprehends it. If one of them is
unreal, so is the other (and in this case, both are). Hence, the reality of cognition is “the empti-
ness of duality” of subject and object (3.211–13). The distinction of entities, however, which is
fundamental to the empirical world as Dharmakīrti analyzes it, is grounded on the dichotomy
of subject and object. If there are no objects standing over against cognition, there can be no
distinct entities (3.214). Moreover, the fundamental constituents of reality identified by the
Buddha, the “aggregates,” are either objective (matter) or mental (sensation, etc.) in nature.
Entities devoid of the characteristics of being either an object or a cognition or aspect of cog-
nition would be without essence (3.215). Thus, all the entities the Buddha talked about when
316
Dharmakīrti
teaching the Four Noble Truths, selflessness, and so on – namely, the aggregates (skandhas), the
“bases” (āyatanas), and the “elements” (dhātus) – are in fact without essence, that is, ultimately
unreal (3.216). Although himself cognizant of the true non-dual nature of cognition – hence, the
non-existence of what we take to be external things – the Buddha spoke as if there were external
objects for the benefit of those whose minds are afflicted by ignorance, “like an elephant closing
one eye,” choosing just what is on one side of her (3.217–19) (Dunne 2004, 396–411).
(2) We have seen that, according to the epistemological assumptions Dharmakīrti is work-
ing with, a cognition must somehow bear the form of the object it cognizes; it must take the
object up into itself, so to speak. In a passage beginning with PV 3.320, Dharmakīrti asks
whether this having the form of an object (tādrūpya) can establish that perception is the appre-
hension of external objects. In this passage, along with others later in PV 3, Dharmakīrti raises
various problems with attempting to infer the existence of an external object from the apparent
resemblance of a cognition with an object. On one interpretation, such reasoning would be
inconclusive simply because we have so many cognitions which seem to be of external objects
but which turn out not to be cognitions of external objects at all – dreams, for instance, or opti-
cal illusions (Kellner 2017, 110–11). But at PV 3.387–390ab, Dharmakīrti introduces what
is allegedly a direct proof of the fact that one is aware only of an aspect of cognition itself,
an “object-appearance” in cognition, whenever one seems to be aware of an external object.
This proof, referred to as the “invariable co-apprehension” (sahopalambhaniyama) argument,
can be stated very simply: cognition and object are not different, because they are invariably
perceived together. As Dharmakīrti puts it in PVin 1.54ab, “Because they are invariably appre-
hended together, blue and the cognition of it are not different.”
The major premise of the argument, as one might reconstruct it, is that two things that are
invariably, or perhaps necessarily, found together are not different (consider: the commander-
in-chief of the U.S. Armed Forces and the president of the United States). The minor premise
of the argument is that one is not aware of an object without being aware of a cognition of the
object, and vice versa. This premise expresses the principle that all cognition is self-aware.
Both premises can be defended against prima facie objections; together, they entail the con-
clusion that cognition and object are not different. It is a corollary of the argument that every-
thing exists in consciousness or that there are no objects external to consciousness. This was
seen as a formidable argument for consciousness-only by non-Buddhist philosophers after
Dharmakīrti and provoked intense debate (Taber 2020).
(3) A third line of reasoning that denies the reality of the external world by transposing
the experienced world within consciousness starts with the consideration of cognition as self-
illumination. My own tentative reconstruction of this argument, first articulated in the Ascer-
tainment, is as follows.9 Cognition must present something that only belongs to itself, because
it is inherently self-aware. Every cognition is not just an awareness of some object-form, but
also awareness of itself aware of that object-form. Thus, self-awareness is of the very essence
of cognition, in which case the awareness of an object-form itself must also be a kind of self-
awareness; that is to say, an awareness by the cognition of (an aspect of) itself. A cognition
cannot be aware of anything outside itself.
A fuller understanding of Dharmakīrti’s antirealism awaits closer study of the so-called
“self-awareness section” of PV 3.425–539. Although he nowhere – to my knowledge – makes
the Berkeleyan point that we cannot comprehend how a material thing could even impose an
imprint of itself upon consciousness, we have seen a resistance in his thinking to any sugges-
tion that matter and consciousness causally interact. An underlying theme of his critique of
physicalism was that it is easier to explain the arising of cognitions from other cognitions than
from the body. Here, he seems to be implying that the best explanation for why a cognition
317
John Taber
contains a certain object-form is that it assumes that form of itself, for this is consistent with its
nature as a kind of self-awareness. “Cognition is by nature just an intransitive ‘appearing-in-a-
certain way,’ as opposed to a transitive apprehension of something else” (Kellner 2017, 113).
Theory of Inference
Given Dharmakīrti’s concern to demonstrate the flaws of the epistemological and metaphysi-
cal theories of his non-Buddhist opponents, it is natural that he would be interested in patterns
of reasoning as such. A separate entry would be required to do justice to his contributions to
Indian logic. I must confine myself here to certain aspects of his treatment of inference that
have particular bearing on his refutation of key Brahmanical teachings.
The history of logic in India prior to Dharmakīrti was largely taken up with the quest for an
adequate characterization of the relation between ground and consequent in a sound inference.
By Dharmakīrti’s time, the most widely accepted view, to which Dignāga subscribed, was that
a reason is a good one if it exhibits a certain pattern of occurrence. It must (1) be true of the
subject of inference; (2) occur together with the property-to-be-proved in other cases; and (3)
not occur where the property-to-be-proved is absent. Thus, in the inference, “The mountain is
on fire, because it is smoking,” “smoking” is the reason, “mountain” the subject of inference,
and “on fire” the property-to-be-proved of the subject. If the mountain really is smoking, the
first condition is satisfied. The second condition is satisfied because other things that give off
smoke are on fire, and the third is satisfied because nothing that is not on fire gives off smoke.
Dignāga introduced refinements of this theory – in particular, he tightened the formulations
of the second and third conditions so that they define a necessary connection between the
reason and the property-to-be-proved – but stayed within its basic framework. Nyāya critics of
Dignāga, however, pointed out certain problems with it.
One question that was raised was whether in fact both the second and third conditions must
be satisfied for a sound inference. It was generally understood that the statement “whatever
is smoking is on fire,” which is captured by the first condition, is equivalent to “whatever is
not on fire is not smoking,” expressed by the third condition. Why, then, for a reason to be a
good one, do both conditions have to be satisfied, as Dignāga maintained? Why would it not
be sufficient if we knew that just one of them is? In particular, the Nyāya author Uddyotakara
pointed out that there seem to be inferences that are sound but have reasons that only satisfy
the third condition. Take, for instance, the following inference for the existence of a self from
breathing, mentioned previously: “the body must be possessed of a self because it is breath-
ing.” One cannot establish a necessary connection between the reason “breathing” and the
property-to-be-proved, “possessed of a self,” in a positive way by pointing to uncontroversial
cases of things that breathe and have selves; a Buddhist, who denies the existence of a self, will
naturally reject all such cases. Thus, the second condition cannot be satisfied. But all parties to
the debate about the self will agree that things like rocks, furniture, and automobiles, which do
not breathe, are without a self. Even the Buddhist would accept this. Hence the “negative con-
comitance” (vyatireka) between reason and property-to-be-proved – that is, the fact that there
is nothing that lacks a self that breathes (or “everything that lacks a self is without breath”),
which represents the third condition and is equivalent to the second condition – seems to be
established, thereby rendering the argument sound. Thus an argument can be sound that does
not satisfy both conditions (2) and (3), but only condition (3).
It was partly because of such inferences that Dharmakīrti introduced a more stringent
requirement for the relation between reason and property-to-be-proved than that the reason
318
Dharmakīrti
exhibit the pattern of occurrence reflected by conditions (1), (2), and (3). He proposed that a
reason must, more fundamentally, be related to the property-to-be-proved by a “natural con-
nection” (svabhāvapratibandha), which explains the satisfaction of conditions (2) and (3).
Dharmakīrti recognizes two kinds of natural connection: causation and essential identity. The
non-occurrence of the reason in the domain where the property-to-be-proved is absent is signifi-
cant only if it is the result of one of two deeper conditions: either the reason is the effect of the
property-to-be-proved, or else it is an essential property (svabhāva) of the subject of inference
that is identical with another essential property of it.10 Otherwise, if neither of these conditions
is met, the non-occurrence is not significant. Smoke does not occur in things that are not on fire
because fire causes smoke. Simply never observing something that is not on fire to be smoking
is not by itself sufficient to establish that everywhere there is smoke, there is fire. Similarly,
never observing that nothing that is without a self breathes does not by itself establish that
everything that breathes will have a self. One must determine that there is a causal connection
between breathing and having a self, and a causal connection between X and Y, in general, can
be established only by observing that, given X, Y arises and, in the absence of X, Y does not
arise. Since a self is imperceptible, however, this is not something one could ever observe for
a self and breathing.
While it would be wrong, then, to say that Dharmakīrti revised the logical theory of his day
with the express intent of refuting certain non-Buddhist arguments, it seems plausible that he
arrived at logical insights as a result of reflecting on certain Brahmanical arguments, in particu-
lar, and becoming convinced that the theory he inherited failed to expose their flaws. Another
argument Dharmakīrti criticizes as fallacious is one that was current among Mīmāṃsakas
that impugned the authority of the Buddha, which, as we have seen, is Dharmakīrti’s central
preoccupation in PV 2. The argument in question is: “The Buddha must have been possessed
of desire because he spoke.” As we have seen, Dharmakīrti argues in PV 2 that the Buddha
eliminated desire by eliminating its cause, the false view of a self, as a result of practicing
the Path over many lifetimes. The Buddha’s freedom from desire is not just an indication of
his enlightenment but also what partly justifies him as a trustworthy guide for others seeking
to overcome suffering. Now, the Mīmāṃsā tradition, which was skeptical of any claims of
superhuman attainment made by followers of non-Vedic religious traditions on behalf of their
founders, questioned the Buddha’s special status on the grounds (among other reasons) that he
spoke – which of course he had to do in order to teach. One only speaks out of some desire, the
Mīmāṃsakas plausibly claimed. The Buddha spoke; therefore, he had to have been afflicted
by desire – just like the rest of us.
Dharmakīrti mentions this argument when developing another key point of his theory of
inference at PV 1.12: a negative concomitance – the absence of the reason in the domain
where the property-to-be-proved is absent – cannot be established merely by non-observation.
Dignāga, along with Dharmakīrti’s teacher Īśvarasena, held that simply the non-observation of
any counterexample to the “pervasion” of the reason by the property-to-be-proved is all that is
needed to establish that the reason is never found where the property-to-be-proved is missing,
i.e., condition (3). Dharmakīrti, however, was aware that just because one never encounters
a counter-example does not mean that one does not exist – and that, in fact, is the case when
it comes to speaking and having desire. One may never have observed someone speak who
did not have some desire, but that does not mean that there could not ever be such a person.
Indeed, the Buddha was such a person. We must not draw the conclusion that something is
never the case from not having, so far, observed it to be the case. Negative concomitance can-
not be established simply by non-observation.
319
John Taber
Once again, for there to be a natural connection between reason and property-to-be-proved,
the reason cannot occur without the latter. This is founded on there being either a causal
relation or an identity between the two properties. In the case of a causal relation, the cause
must be a necessary and sufficient condition of the effect.11 While desire, to be sure, is one
of the causes of speaking, it is not the only cause. Compassion could also be the cause of
teaching others the way to overcome suffering, as it allegedly was in the Buddha’s case. The
cause which underlies both the positive and negative concomitance between the reason and
property-to-be-proved of a sound inference must be a necessary condition for the arising of
its effect.
Notes
1 In taking this approach, I have been influenced by the groundbreaking research of Vincent
Eltschinger.
2 This is assuming there was only one Vasubandhu, not two (see Chapter 16 of this volume).
3 The correspondences between the chapters of Dharmakīrti’s Commentary and Dignāga’s Compen-
dium, as proposed by his earliest commentators, Devendrabuddhi (630–690) and Śākyabuddhi (660–
720), are presented in Table 18.1.
From Prajñākaragupta (ca. 800) on, commentators began to consider the Commentary as beginning
with the “The Establishment of the Means of Knowledge,” often referred to as “The Pramāṇasiddhi
chapter,” followed by the chapter on perception, the chapter on inference for oneself, and the chapter
on inference for others, in that order. This brings the Commentary more in line with the organization
of the Compendium (even though one of Prajñākaragupta’s commentators, Jayanta, denied that it was
intended as a commentary on the Compendium). The Tibetan translation of the Commentary has the
original sequence; that is to say, it begins with the chapter on inference for oneself.
4 Hattori (1968, 23, slightly emended):
Saluting Him, who is tantamount to a means of cognition, who seeks the benefit of all living
beings, who is the teacher, fully accomplished, the protector, I shall for the purpose of establish-
ing the means of knowledge compose the Compendium, by uniting here my theories scattered in
many treatises.
5 Alternatively, one may infer the Buddha’s knowledge from the truth of the main part of his teaching,
the Four Noble Truths, independently established. From that, one may infer his compassion. From
his knowledge and compassion taken together, his authority and trustworthiness are inferred, on the
basis of which one may be confident in the truth of the non-essential parts of the scriptures authored
by him (Franco 1997, 68).
6 However, sthiti is also interpreted simply as “cognition.”
7 Dharmakīrti clarifies this point with PV 2.6.
8 See also PV 3.145–73.
9 See Kellner (2017, 112–13) for a much more nuanced interpretation.
10 For instance, “being produced” is identical with “being impermanent,” and hence can serve as evi-
dence for the latter, because both are essential properties of a thing that arise from the same causal
complex. The property-to-be-proved, in an inference based on identity, “follows upon the mere exist-
ence of the reason,” without need of any further causal factors.
11 A cause is a sufficient condition of the effect, given that all auxiliary factors are present and their
operation is unobstructed. Since, however, the auxiliary factors are not always in place and obstruc-
tions can always occur, one cannot reliably infer an effect from its cause, but only the cause from the
effect, as its necessary condition.
References
Dunne, John. 2004. Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom Publications.
Eltschinger, Vincent. 2007. Penser l’autorité des Écritures. Vienna: Verlag der Österreichischen Akad-
emie der Wissenschaften.
320
Dharmakīrti
———. 2010. “Dharmakīrti.” Revue Internationale de Philosophie 64 (3): 397–440.

———. 2014. Buddhist Epistemology as Apologetics. Vienna: Verlag der Österreichischen Akademie
der Wissenschaften.
Eltschinger, Vincent, Helmut Krasser, and John Taber. 2012. Can the Veda Speak? Dharmakīrti Against
Mīmāṃsā Exegetics and Vedic Authority. Vienna: Verlag der Österreichischen Akademie der
Wissenschaften.
Eltschinger, Vincent, and Isabelle Ratié. 2013. Self, No-Self, and Salvation: Dharmakīrti’s Critique of
the Notions of Self and Person. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
Eltschinger, Vincent, John Taber, Michael Torsten Much, and Isabelle Ratié. 2018. Dharmakīrti’s Theory
of Exclusion (apoha), Part I: On Concealing. Studia Philologica Buddhica 36. Tokyo: International
Institute for Buddhist Studies.
Franco, Eli. 1997. Dharmakīrti on Compassion and Rebirth. Wiener Studien zur Tibetologie und Bud-
dhismuskunde 38. Vienna: Universität Wien.
Franco, Eli, and Miyako Notake. 2014. Dharmakīrti on the Duality of the Object. Leipziger Studien zu
Kultur und Geschichte Süd- und Zentralasiens 5. Leipzig: LIT Verlag.
Frauwallner, Erich. 1954. “Die Reihenfolge und Entstehung der Werke Dharmakīrti’s.” In Asiatica. Fest-
schrift Friedrich Weller, edited by Johannes Schubert and Ulrich Schneider, 142–54. Leipzig: Otto
Harrassowitz.
Hattori, Masaaki. 1968. Dignāga, On Perception. Harvard Oriental Series 47. Cambridge, MA: Harvard
University Press.
Kellner, Birgit. 2017. “Proofs of Idealism in Buddhist Epistemology.” In Indian Epistemology and Meta-
physics, edited by Joerg Tuske, 103–28. London: Bloomsbury.
Krasser, Helmut. 2001. “On Dharmakīrti’s Understanding of pramāṇabhūta and His Definition of
pramāṇa.” Wiener Zeitschrift für die Kunde Südasiens 45: 173–99.
———. 2012. “Bhāviveka, Dharmakīrti and Kumārila.” In Devadattīyam. Johannes Bronkhorst Felicita-
tion Volume, edited by François Voegeli et al., 535–94. Bern: Peter Lang.
Pecchia, Cristina. 2015. Dharmakīrti on the Cessation of Suffering. Brill’s Indological Library 47.
Leiden: Brill.
PV. Dharmakīrti’s Pramāṇavārttika with Commentary by Manorathanandin, edited by Rāhula
Sāṅkṛtyāyana. Journal of the Bhandarkar Oriental Research Institute 24–26 (1938–1940), Appendix.
PVin. 2007. Dharmakīrti’s Pramāṇaviniścaya. Chapters 1 and 2. Edited by Ernst Steinkellner. Vienna:
Austrian Academy of Sciences Press.
PVSV. 1960. The Pramāṇavārttikam of Dharmakīrti. The First Chapter with the Autocommentary.
Edited by Raniero Gnoli. Rome: Istituto Italiano per il Medio ed Estremo Oriente.
PV(SV). PV together with PVSV.
Steinkellner, Ernst. 1968/1969. “Die Entwicklung des Kṣaṇikatvānumānam bei Dharmakīrti.” Wiener
Zeitschrift für die Kunde Südasiens 12–13: 361–77.
Taber, John. 2020. “Philosophical Reflections on the Sahopalambhaniyama Argument.” In Reverbera-
tions of Dharmakīrti’s Philosophy. Proceedings of the Fifth International Dharmakīrti Conference,
edited by Birgit Kellner et al., 441–62. Vienna: Austrian Academy of Sciences Press.
Vetter, Tilmann. 1984. Der Buddha und Seine Lehre in Dharmakīrtis Pramāṇavārttika. Wiener Studien
zur Tibetologie und Buddhismuskunde 12. Vienna: Universität Wien.
321
PART 5
Early-Period Commentators
(Fifth–Seventh Century)
EARLY-PERIOD
COMMENTATORS
Indian Buddhist philosophers regarded as the founders of philosophical schools – such as the
figures in the preceding part of this volume – often wrote texts that are terse and sometimes
so obscure as to be incomprehensible without a commentary. Because the root texts were
composed with the expectation of oral or written commentary to explain them, commentaries
were not simply a supplement to an otherwise sufficient work; they were essential to the root
text. This is why some Indian Buddhist philosophers wrote auto-commentaries on their own
texts. In many instances, it would be reasonable to consider the root text and its commentary as
together constituting a kind of hybrid text, and this is one way in which authorship is rendered
complicated in Indian Buddhism. Yet, it was through these hybrid texts that different schools
were constructed as distinct and coherent. It was through the commentarial literature that these
schools developed and became more sophisticated as commentators drew on the resources of
their traditions and the indeterminate meaning of the texts they were explaining to introduce
novel interpretations and ideas in response to new philosophical challenges. By the sixth cen-
tury, developments in logic and epistemology contributed to the rise of debates across Indian
philosophical traditions. Commentators often understood their role to include not just defend-
ing their own view but also demonstrating the deficiency of other Buddhist and non-Buddhist
philosophical positions. As with other South Asian traditions, then, even as many Indian Bud-
dhist philosophers were creative and original thinkers, they often presented their thought as
explication and practiced philosophy though commentary on earlier works.
Buddhaghosa’s commentaries on Buddhist scriptures and his formulation of a Buddhist
life in the Path of Purification (Visuddhimagga) played a significant role in the formation of
what today is called the Theravāda tradition. At the heart of Chakravarthi Ram-Prasad’s chap-
ter is what he terms Buddhaghosa’s “hermeneutic of contextualization,” which is especially
attentive to the context in which the Buddha offers his teaching. Applying the hermeneutic
of contextualization to Buddhaghosa’s own texts, Ram-Prasad argues, suggests that he is not
primarily concerned with the ontological analysis of objects of experience. Rather, Buddha
ghosa practices a phenomenological method aimed at transforming human experience. Draw-
ing on Buddhaghosa’s hermeneutics and phenomenological method, Ram-Prasad shows how
his account of understanding in the Path of Purification is a form of phenomenological practice
that prepares the practitioner for systematic, methodical, and morally informed meditation.
DOI: 10.4324/9781351030908-29 325

Early-Period Commentators (Fifth – Seventh Century)
In the next chapter, Karl-Stéphan Bouthillette shows how Bhāviveka (sixth century) drew on
recent developments in epistemology and argumentation to interpret Nāgārjuna’s philosophy of
the Middle Way. Bhāviveka’s interpretation – using the three-part inference (anumāna) devel-
oped by Dignāga and the epistemological framework of valid means of cognition or knowing
( pramāṇa) – became the dominant view of Nāgārjuna in Indian Buddhist scholarship. Interpret-
ing and defending Nāgārjuna’s philosophy with the fifth- and sixth-century methods of critical
reasoning and debate firmly established it as a system that could compete with other Indian phil-
osophical schools. Indeed, Bhāviveka devoted considerable resources to critiquing the views of
other Indian philosophical traditions, setting up a doxographical hierarchy that became a model
of much Buddhist philosophy. In this hierarchy, philosophy is paramount and wrong ideas need
to be refuted because they are obstacles to liberation. Bhāviveka’s work, then, emphasizes the
importance of rational argument, and he himself was an original and rigorous thinker who was
able to synthesize different philosophical approaches into a coherent Madhyamaka system.
Like Bhāviveka, Dharmapāla (sixth century) employed Dignāga’s logic and epistemol-
ogy to interpret and defend views first articulated before the fifth- and sixth-century devel-
opments in argumentation and reasoning and to critique other Buddhist and non-Buddhist
schools. In his chapter on Dharmapāla, Ching Keng suggests that one of his opponents was in
fact Bhāviveka and that the debate between the two thinkers is the beginning of the debates
between Madhyamaka and Yogācāra views that became important for many later Buddhist
philosophers. While Dharmapāla’s commentaries are conservative, hewing closely to the root
texts, works attributed to him also include original thinking on a variety of important Yogācāra
themes, especially meta-cognition. Keng suggests that perhaps this tension can be resolved by
thinking of “Dharmapāla” as including the Indian scholar and his students and followers and
that it was their combined work that had such a significant influence on Buddhist philosophy
in India and especially in East Asia. Authorship, in this context, works more like the label on
a “workshop” rather than a designation of a single individual.
As Roy Tzohar and Jowita Kramer show in the next chapter, Sthiramati (ca. sixth cen-
tury) used his commentaries on Yogācāra texts to make novel philosophical contributions. In
the first half of the chapter, Kramer provides an overview of current research on Sthiramati,
including the philosophical themes he explored in his work. In the second half, Tzohar gives a
close reading of Sthiramati’s commentary on the beginning of Vasubandhu’s Treatise in Thirty
Verses (Triṃśikā), emphasizing Sthiramati’s innovative understanding of meaning, metaphor,
and language. Sthiramati’s linguistic figurative theory of meaning complements the Yogācāra
account of the causal activity of consciousness resulting in a more systematic philosophy.
Indeed, it was through his commentaries that Sthiramati synthesized a variety of texts into one
coherent Yogācāra system.
Alexander Yiannopoulos, in the following chapter, focuses on Devendrabuddhi (seventh
century) and Śākyabuddhi (seventh century), the two earliest commentators on Dharmakīrti’s
Commentary on the Means of Knowledge (Pramāṇavārttika). Yiannopoulos reflects on how
these commentators interpret Dharmakīrti and explores particular distinctions and themes that
exemplify their approach to the root text, such as Devendrabuddhi’s account of “unmediated
instrumental effect” and Śākyabuddhi’s analysis of reflexive awareness. Because Dharmakīrti’s
Commentary on the Means of Knowledge is so terse, it requires additional commentary to
make the arguments explicit. Indeed, Dharmakīrti’s text is not just philosophically obscure;
the commentaries are often necessary even to make sense of the syntax and grammar. Given
how thoroughly interdependent this particular root text and its earliest commentaries are, Yian-
nopoulos suggests that we should consider them as together constituting one “hybrid text.”
326
Candrakīrti (seventh century) is widely regarded as one of the most significant Madhy-
amaka philosophers. This is especially true in Tibet, where his commentaries on Nāgārjuna’s
texts are often taken to be authoritative on Madhyamaka philosophy. Focusing on Candrakīrti’s
understanding of the two truths, the importance of his approach to language, and his overall
philosophy as a form of anti-essentialism, Mattia Salvini gives an overview of Candrakīrti’s
writings and an analysis of his main ideas. In addition, Salvini emphasizes three aspects of
Candrakīrti’s work that are often neglected in English language scholarship. The first is how
Candrakīrti’s philosophical practice is grounded in traditions of Sanskrit grammar, which is
key to understanding his anti-essentialist philosophy of language. Second, Salvini shows how
Candrakīrti is attentive to the ways in which Nāgārjuna himself was working in an abhidharma
philosophical context and how Candrakīrti employs abhidharma analysis in his own works.
The third aspect is Candrakīrti’s inclusive hermeneutics, as he finds arguments for emptiness
in both Mahāyāna and non-Mahāyāna sources and presents his own ideas without relying
exclusively on the authority of any one circumscribed set of Buddhist sūtras.
327
19
BUDDHAGHOSA
Phenomenology, Hermeneutics, and
Understanding
Introduction
Buddhaghosa (fifth century) is the most prominent commentator on Theravāda Buddhist scrip-
tures. He is also the author of what is widely regarded as the definitive Theravāda formulation
of a Buddhist life, the Path of Purification (Visuddhimagga), which is grounded in his exegesis.
Together, the commentaries and the Path of Purification constitute perhaps the most system-
atic Theravāda presentation of the Buddha’s teachings after the early Buddhist texts, and they
played an influential role in the very formation of what today we think of as Theravāda Bud-
dhist traditions. (The term “Theravāda” meant something very different in the fifth century,
when Buddhaghosa would have identified with the Mahāvihāra, or Great Monastery, tradition.)1
According to legend, Buddhaghosa was born in India and ordained as a Buddhist monk so
he could study Abhidhamma. His prodigious intellectual powers were quickly recognized in
the monastery. Eventually, Buddhaghosa was called to Sri Lanka to work on the massive pro-
ject of editing and translating the ancient Sinhala commentaries on the Pāli Buddhist scriptures
into the more widespread language of Pāli. He composed the Path of Purification, it is said,
to demonstrate his scholarly skills and prove his worthiness for the project. While he could
not have written all the works attributed to him, it would seem that some major commentar-
ies on canonical literature, such as The Expositor (Atthasālinī) and The Dispeller of Delu-
sion (Sammohavinodinī), so closely resemble the ideas and content of what is undeniably his
magnum opus, The Path to Purification, that we can talk about them as his works even if they
were written under his close supervision.2 In addition to determining the final form of the Pāli
Canon, this work also reinforced the importance of Pāli in Theravāda traditions.
This chapter draws on a way of thinking about Buddhaghosa that has been argued for
by Maria Heim (2014; 2018) and which I have worked on myself (Ram-Prasad 2018, chap-
ter 3), and also jointly with her (Heim and Ram-Prasad 2018). Our basic contention is that
Buddhaghosa is primarily concerned with a phenomenological approach that he regards as
conducive to transforming experience in contrast to the interpretation of Buddhaghosa as
primarily interested in metaphysics and the ontological tabulation of the objects of experi-
ence (Heim and Ram-Prasad 2018, 1087). We are at pains to point out that this is not at all
to deny that he has doctrinal commitments to how things truly are, as conveyed in the Bud-
dha’s teachings.
328 DOI: 10.4324/9781351030908-30

Buddhaghosa
Doctrinal commitments do not automatically imply metaphysical views, and such is the
case with Buddhaghosa as Buddhist. Moreover, we should be prepared to treat Buddhaghosa
as a philosopher for reasons other than whether he does or does not deploy a metaphysics and
defend an ontology in his texts; we may look at his contributions in ways that, in comparativist
perspective, locate him in other areas of philosophy. We will explore these points through a
reading of some of his work in this chapter.
Drawing on Maria Heim’s work (2018) on Buddhaghosa’s practices of reading text, I will
suggest that his interpretive theory gives us a hermeneutics with a quite general significance,
one that can be called a hermeneutic of contextualization. Drawing out this hermeneutics will
be the dominant task of this chapter. I will, however, then turn to that area of his work that
most naturally has been assimilated into a larger history of metaphysics, specifically, atomistic
ontology, as is held to be found in the teachings on the abhidhamma.3 I will suggest that, given
his hermeneutic of contextualization, if we properly contextualize his study of the supposed
ontological constituents of the person, the aggregates (khandhas), then an ontological reading
of the aggregates becomes inconsistent with his purposes.
Buddhist philosophers share the same beginning and end: the teachings of the Buddha and
the cessation of suffering, respectively. But what goes on in between is not always argument
about the structures of reality and the pursuit of an epistemic vindication of one’s argument.
In Buddhaghosa, we have a figure who is original in many ways within Buddhist thought but
should also be taken seriously for his contribution to global philosophy.
Hermeneutics: Buddhaghosa’s Contextualization Strategies

Buddhaghosa’s way of reading text matters philosophically in two ways. One is that his meth-
ods can enrich hermeneutics as a cross-cultural philosophical discipline so that we are not
tempted to see his practices as limited to a distant Buddhist past, of interest only to a history of
ideas. The other is that taking his hermeneutics seriously has consequences for how we locate
his treatment of abhidhamma categories as something other than an atomistic (and, from the
hostile perspective of Madhyamaka opponents, a conveniently simplistic) ontology.
If there is one word that sums up Buddhaghosa’s approach to Buddhist texts, it is “contex-
tualization.” As we will see, this is a paradoxical approach, one in which attention to detail
does not limit but rather opens up inquiry – a fact routinely missed in treatments of Buddhag-
hosa as someone bogged down in endless disquisitions on the minutiae of monks’ conduct.4
One specific aspect of Buddhaghosa’s approach to texts makes it particularly apt when
thinking of modern hermeneutics as a discipline framed by its response to Christian theology:
the “canonical” nature of the texts with which he was dealing. This can sometimes be forgotten
in a pan-Buddhist approach to hermeneutics. In a path-breaking work, Robert Thurman makes
the point that Buddhist hermeneutics must concern itself with an incredibly vast array of texts
and teachings and yet treat them as somehow uniformly authentic (1978, 22). But developing
this thought from a Mahāyāna perspective, Thurman almost casually sets aside the “simplistic
approach” of the “Hīnayāna Abhidharma.”5 He then focuses, for its vastness, on various lin-
guistic traditions of the Mahāyāna. Referring to Thurman, Lopez points to “the sheer bulk of
the canon,” explaining it by saying “the Buddhists seemed to have simply written a new sūtra
or interpolated new material into an existing text” (1998, 2). Two different points are elided
here. The quantitative issue is one thing, but Lopez expands on it by talking about something
else, Mahāyāna Buddhism’s qualitatively different way of treating sacred text by generating
new, authoritative ones. But the vastness of the “Pāli canon” does not equate to the method
of generating new material; Buddhaghosa is very concerned with defending the anomalous
329
character of the Abhidhamma, for example. His strategies are comparable in this specific sense
with Biblical hermeneutics, in having to deal with types of reading and consequently find
meaning in a fixed set of materials.6
It is contextualization of a fixed textual range of teachings that offers depth to his herme-
neutic practice, but if we are to suggest what that practice offers to a cross-cultural discourse
on hermeneutics, we must start with acknowledging his location as a Buddhist. Yet, as we
will see, the theoretical originality he would bring to such a discourse cuts across the divide
between theological exegesis and critical interpretation that defines the (Western) history and
discipline of hermeneutics.
Framing Buddhist Truth: Discovered Omniscience

Let us begin with an outline of Buddhaghosa’s hermeneutics, both its framework and contex-
tualizing practice, following Heim’s detailed study (2018). Obviously, we have the straight-
forward consideration that he is committed to faith in the Buddha, as any commentator within
a religious community would be to sacred texts. This requires acceptance of doctrines as ulti-
mately holding true, as in other religious traditions. But their holding true is due to a striking
combination in Buddhism: these doctrines articulate what was always true but came to be
taught only because insight into them was attained by the Buddha. This stands in contrast to
doctrines being true due to revelation by an always already divine source – a view that was
only just gaining some traction in Hindu theisms in the time of Buddhaghosa. The ambiguities
in the canonical material about whether the Buddha himself claimed omniscience are deci-
sively settled by the later consensus that he was, although interpreted in very different ways by
Mahāyāna systems (Griffiths 1989; Perrett 1989). It is important to note that Buddhaghosa’s
hermeneutic is driven by his conception of the Buddha’s omniscience (Heim 2018, 44–49). It
is a dynamic conception and draws attention to a continuous quality of being enlightened. He
comments on a canonical discussion of the term “Buddha” (Heim 2018, 48):
Here, just as it is said in the world that someone has “understood” because of “under-
standing,” this is the case too for one who is Awakened (“Buddha”), who has discovered
the Truths. . . . Awakened because of knowing all things: “awakened” is said because
of a wisdom capable of discovering [literally, “attaining knowledge of,” bujjhana] all
things. Awakened because of seeing all things: “awakened” is said because of a wis-
dom capable of knowing [as also teaching, bodhana] all things. Awakened because of
being unguided by another: “awakened” is said because of being awakened only by
himself, not woken up by another. Awakened because of expanding: “awakened” is
said in the sense of expanding like a lotus blossoming with many qualities.
(Paṭisambhidāmagga-aṭṭhakathā, ii.485, Pāli Text Society)
What we should keep in mind here is that the buddhahood of the Buddha consists in his attaining
omniscience and being “capable” of or “able” to make such discoveries and teach them. There
are two consequences that are relevant to our study, one regarding hermeneutics, the other per-
taining to phenomenological method. First, in as much as the Buddha discovered “the Truths,” a
Buddhist hermeneutics makes the same commitment to ultimacy as any Christian hermeneutics.
But because the Buddha discovered “the Truths,” his omniscience stands in a different relation-
ship to the Buddhist hermeneuticist (and thus to us) than does the status of truths conveyed by
divine revelation. This is important because, in the West, hermeneutics as a discipline grew out
of the dialectic between Christian commitment to a particular understanding of ultimate truth
330
Buddhaghosa
and the secular rejection of it in favour of context. Buddhaghosa’s framing of the Buddha’s
authority as consisting in his attainment of omniscience encloses a conception of hermeneutics
as the constant contextualization of our understanding of such knowledge. He therefore offers
us a way of re-working the relationship between truth and context, as we will see shortly. Relat-
edly, Buddhaghosa’s understanding of the Buddha’s omniscience is dynamic and ongoing: it is
the activity of awakening, it is a potency lying in his ever-ready capacity to transmit his teach-
ings, and it expands like a lotus, an ever-unfolding capacity (Heim 2018, 48–49).
According to Buddhaghosa, this is demonstrated by, amongst other things, the Buddha’s
methods (naya) of teaching how to read experience endlessly. It is Buddhaghosa’s reading of
the nature and content of the Buddha’s omniscience and the Buddha’s means of teaching such
knowledge that together show how Buddhaghosa’s hermeneutics points to his phenomenological
method, free from ontological concerns. This understanding of the teachings is expressed through
Buddhaghosa’s elaboration of the idea that the teachings (as contained in the three piṭakas) are an
“ocean of methods” (nayasāgara) in The Expositor. Of the Abhidhamma, he says:
And endless joy and happiness arise for those monks who are Abhidhammikas [that
is, specialists] reflecting on the Abhidhamma scripture thus: “the Teacher taught us
by analyzing the name-and-form dhammas, making this or that grouping, this or that
division, as though arranging the stars in the sky, which is an abstruse and subtle
teaching that has classified into what is form and what is formless the manifold aggre-
gates, bases, elements, faculties, powers, limbs of awakening, kamma and its results.”
(Heim 2018, 241)
Characteristically, he says this should be understood through an incident in which an Elder,

contemplating the ocean on a voyage, asked which had the greater strength, the ocean or
the methods, and concluded joyfully that, whereas the ocean was eventually bound by land,
there could be no circumscription to the method. The entire classificatory method of the Abhi
dhamma here is given an organization of various groupings, like discerning constellations as
patterns in the sky. This is the practice that has been taught, and its enactment is phenomeno-
logical. So we see again the folding in of how the Buddha taught with how we should learn
what he taught: the infinitude of methods by which he taught keys us to approach those teach-
ings through these infinite methods. Note, too, that the experiencing of the practice of endless
classificatory analysis is joyful, as we see with what the Elder undergoes. So the joy within the
teaching becomes the joy of those who learn.
Contextual Teaching as Guide to Contextual Reading

In order to understand the significance of Buddhaghosa’s hermeneutic of contextualization, we
must first notice how salient a contrast to certain Western conceptions of sacred authority it is
to say that the Buddha’s authority came through his discovery of truth. Jussi Backman deline-
ates the nature of classical Western and Christian hermeneutics thus (2016, 29):
For the ontotheological ethics of Aristotle, firmly inscribed within the ancient and
medieval ethical tradition reaching from Parmenides to Thomas Aquinas, the culmi-
nation of human activity and intellectual being is thus the escape of thinking from the
situatedness of “merely” human existence to the autarkic beholding of the divine and
the absolute. Philosophical hermeneutics, by contrast, is precisely the philosophical
embrace of this irreducible mortal situatedness of humanly accessible meaning, and
331
in this respect it indeed marks an entirely “new philosophical epoch” with regard to
antiquity.
For Buddhaghosa, the Buddha is indeed “autarkic” (for he is “unguided by another”), and we
can think of his omniscience as “absolute” in the specific sense of being unlimited relative to
what needs knowing. Yet, at the same time, omniscience consists precisely in the Buddha’s
specific orientation to the situatedness of existence. Buddhaghosa’s insight is that because the
Buddha taught according to context, we must see the Buddha’s teaching through attending to
context. “Situatedness,” in other words, is itself the vector of the “absolute” truths. Let us see
an example of how this is so.
A key concept to consider here is the nidāna: It is “a context, origin, source, occasion,
introduction, or causal condition. When used to name part of a sutta or other text it is the
introduction or narrative context in which the text occurred in the Buddha’s teaching career”
(Heim 2018, 116). The nidāna is a device that can be used to frame teachings at the canonical
level but also be used by the commentators. The first point is important because it establishes
the nidāna as integral to the very existence of the body of teachings and not a mere conveni-
ence for the commentators. But the second is relevant because it shows how Buddhaghosa
was able to utilize the form of the nidāna ingeniously. Heim points out that modern scholar-
ship has thought of nidāna as primarily serving to show the authenticity of the teachings, and
Buddhaghosa’s sub-commentator Dhammapāla sees it exactly as performing this solemn task
(Heim 2018, 118, 121):
For a teaching connected to a time, place, teacher, story, and the recipient of the
Dhamma is a long-lasting, unconfused doctrine, and credible, like a legal contract
provided with notations of place, date, maker, and witnesses.
(Dīgha ṭīkā i.70)
Heim contrasts this concise presentation with Buddhaghosa’s approach, which is lav-
ish in describing the purpose of nidāna and which Dhammapāla queries for its “poeticity”
(kabyaracanā). But I want here to focus on the specific issue of contextualization and Bud-
dhaghosa’s reading of it hermeneutically. It is not that he deviates from the standard under-
standing that a nidāna is about context and involves situating a given teaching in a moment
and time and in relation to a specific audience (see Heim 2018, 117). Indeed, Buddhaghosa is
perfectly consistent with the tradition’s reading of the nidāna as demonstrating that the Bud-
dha taught in a very contextual way. And in this, Buddhism markedly contrasts with the com-
mitment to the ontotheological universalism that marks Biblical teaching.7
Yet this is only one half of the principle of contextualization. Buddhaghosa’s achievement,
I think it can be argued, is that he turns the fact that the Buddha taught according to context
into a hermeneutic principle. Heim points out the aesthetic role a nidāna plays for Buddhag-
hosa “in preparing the imagination for entering the sutta perspective to the beauty and pleas-
urable delight” of faith in the Buddha (Heim 2018, 121). And this is evident in a passage she
quotes in extenso. Let us look at the first part of this passage.
Thus far, the nidāna, adorned with time, place, teacher, narrative, assembly, and
region, is spoken by Venerable Ᾱnanda, and the commentary on its meaning is com-
plete. It is spoken for the sake of the ease or pleasure of entering this sutta, which is
perfect in meaning and phrasing and indicates the power of the Buddha’s qualities.
(Sv.i.50; Ps.i.15; Pv-a.iii.536)
332
Buddhaghosa
This nidāna, Buddhaghosa says, is like a crossing place (tittha), a staircase (sopāna), and a
great door (mahādvāra), and he describes each metaphor in exuberantly, lavishly poetic terms,
evoking pearls and blue water lilies, golden creepers and delicately carved ivory, sweet laugh-
ter and the tinkling of bracelets. It is clear that he wants to suggest that the nidāna is meant
to delight and prepare the reader. But I also think that it is crucial to notice that the nidāna is
meant for entering into (avagāhati) the sutta. In other words, it is a method for interpretation.
If we are to understand the content of the teaching that is to come, we are to get to it through
context. This explains the densely phenomenological approach he takes to the nidāna, in con-
trast to the precise epistemological reading of its function that Dhammapāla gives (and which,
in a different form, is taken to be its aim in modern scholarship). The nidāna is important
because the context it provides through time and place and the rest is integral to how we are to
receive the teaching we will encounter. It prepares us to read the teachings. It delights but also
informs. This information is not only about the nature of the teaching itself (which, of course,
he takes for granted) but about the way to learn the teaching. My suggestion is that Buddhag-
hosa uses the tradition’s invocation of the Buddha’s contextualization of teaching to himself
develop a hermeneutic of contextualization. Let me draw this out with an example.
Heim offers a detailed reading of Buddhaghosa’s commentary on the Discourse on the
Root of All Things (Mūlapariyāya Sutta) of the Middle Length Discourses (Majjhima Nikāya)
(Heim 2018, 136–41). The sutta itself has no elaborate nidāna but ends with the enigmatic note
that, after the Buddha delivered a sermon, “the monks did not delight” in what he had said.
Buddhaghosa goes to great lengths to explain the context: the monks were former brahmins
who were conceited due to their knowledge of both Vedic and Buddhist texts. The Buddha’s
teaching was to “break down the conceit” (mānabhañjanatthaṃ) of these monks: it shows
the subtle ways in which conceptuality (sañña) distorts experience and generates “craving,
conceit and wrong views.” Buddhaghosa reads the dismay of the monks as being due to their
overgrown learning – which actually reinforced their conceit – impeding their recognition of
the truth of the teachings. Strikingly, Buddhaghosa increases the contextualization. When his
followers ask why those monks were dismayed, the Buddha says that they were brahmins in
a past birth who were similarly stymied by the Buddha in a previous life and that in the future
they will become realized after hearing another of his teachings.
Heim explores this commentary in terms of the Buddha’s omniscience, as well as to make
the point that Buddhaghosa narrates the enactment of the sutta’s teaching so that the “frame
story performs the sermon” (Heim 2018, 139). I want to extend the analysis to say that this is
a striking demonstration of the double-aspect contextualization which I suggest is at the heart
of Buddhaghosa’s hermeneutics. The double aspect is that context is both the intrinsic feature
of the events (name, place, occasion, etc.) surrounding the truthful teaching that is now being
read, but it is also the technique by which the truth is read. His hermeneutic is a form of nar-
rative reasoning, arguing through the story by which doctrine is conveyed, and, if the story is
not in the canon, developing an explanatory version himself. If we disregard this strategy as
over-imaginative piety so that we can isolate the “philosophy” of conceptualization, then we
find Buddhaghosa both prolix where it does not seem to matter and didactic when he should
be making an argument. From this perspective, Buddhaghosa asserts the perpetual truth that
conceited conceptualization of selfhood impedes felicitous freedom not through arguing for
the epistemic and moral incoherence that follows from the conceit but through the narrative
of why and at what point(s) and to whom the Buddha taught. The Buddha’s teaching does not
need expatiation; it is our reckoning of its truth that does.
This double-aspect contextuality makes Buddhaghosa’s hermeneutic attitude to the text
radically different from how modern hermeneutics has understood religious reading. Felix Ó
333
Murchadha cites Ricoeur’s assumption with regard to the challenge which a critical hermeneu-
tic faces in its relationship to earlier religious hermeneutics:
[H]ermeneutics bears the mark of this interpretive effort of understanding enigmatic,

apparently inscrutable, divine communications. . . . [I]t has a text‐like quality in the
sense of being preserved as meaningful beyond the situation of its utterance. Crucial
here is the relation between tradition and experience, mediating between a past event
of divine communication expressed in written form and the present experience of
those reading and applying that text.
(Ó Murchadha 2016, 77, referring to Ricoeur 1995)
The significance of the nidāna, by contrast, is that the teachings require exegesis not because
they are inscrutable or enigmatic but rather because their meaning is encountered and under-
stood in every context that could occur (for the Buddha is omniscient about the contexts to
which his teachings should be directed).
Jussi Backman (2016, 26) broadly defines hermeneutics as “discursive contextuality,”
which works well to define Buddhaghosa’s approach to canon. But for Backman, this means
that philosophical hermeneutics “abandons the prospect of attaining an ultimate, pre- or supra-
contextual point of reference.” Buddhaghosa cannot exactly subscribe to this understanding.
The Buddha’s teachings are indeed an “ultimate” point of reference; they are not “ultimate” in
the sense of a higher-order ontology beyond the objects of ordinary experience but in the sense
of being that beyond which there is no further source for Buddhist doctrine, since they are
rooted in the omniscient Buddha’s authority. But this point of reference is precisely not “pre-
contextual,” as we have just seen. So when it comes to a Buddhaghosan hermeneutic, we must
tease apart some of the assumptions which, in contemporary Western theories of hermeneutics,
are bundled together. As for “supra-contextual” points of reference, even the abstract teachings
of the Buddha are not really supra-contextual. It is that insight that will lead us from the her-
meneutics to a non-metaphysical phenomenology of abhidhamma. But we are getting ahead
of ourselves. Let me quickly outline some of Buddhaghosa’s contextualizing techniques; these
are dealt with in detail in Heim (2018).
Broadly, contextual hermeneutics in Buddhaghosa proceeds through offering paired dis-
tinctions, the choice between each member of the pair being keyed by the manner in which
the Buddha presented the teachings, with that choice then opening up each teaching accord-
ing to our understanding of that manner. We have space only for a few of these double-aspect
contextualization techniques. One pair is “meaning” (attha) and “phrasing” (byañjana), each
containing a “toolkit of hermeneutical devices” (Heim 2018, 74). It would take a different
enterprise to demonstrate how exactly these tools could contribute to a cross-cultural global
hermeneutics. But I draw attention to the potential for precision of analysis offered by the list-
ing under “meaning” of “syllables, words, phrasing, modes, language, descriptions” (akkharaṃ
padaṃ byañjanaṃ ākāro nirutti niddeso; in Nettiprakaraṇa 9, which Buddhaghosa draws
on), and “showing, making known, opening up, distinguishing, making clear, and denoting”
(saṅkāsanā pakāsanā vivaraṇā vibhajanā uttānīkammaṃ paññatti) under “phrasing.” Another
pair is “definitive” and “interpretable,” which we encountered earlier. A consequential pair
that plays an important role in seeing that Buddhaghosa utilizes language common to him and
Buddhist metaphysicians, but in a different way, is that between “conventional” (sammuti or
vohāra) and “further sense” ( paramattha) teachings. The term paramattha here is not “ulti-
mate truth” in the sense of being what holds in a higher-order ontology behind ordinary objects
of experience or as a proposition that is unsublatably or ineliminably true in contrast to those
334
Buddhaghosa
that are only provisionally true. Rather, paramattha is “ultimate” in the doctrinal sense of what
is of highest value, as when nibbāna is called the highest truth (sacca), or the final authority, as
in the Buddha’s insight. Heim’s exegetical demonstration that “paramattha” should be treated
as “further sense” (Heim 2018, 85–94) when it comes to the Buddha’s teachings (desanā) is
consistent with the latter but does not directly pertain to it. She demonstrates that Buddhaghosa
explains at length that the distinction between “conventional” and “further” is about the con-
text of who needs teaching what. There are contexts that require teaching about persons, gods,
and so on and others that require language utilizing concepts of suffering, no-self, aggregates,
elements, and similar categories. The former are not lesser in relation to a higher but a type of
discursive analysis pertaining to what is ordinarily presented, namely persons and the rest, as
opposed to analysis that goes further to what requires more sustained types of contemplative
inquiry. The equal legitimacy of conventional language – both in terms of spiritual perfection
and of exegetical analysis – is manifestly evident in The Wish Fulfiller (Manorathapūraṇi),
the commentary (aṭṭakathā) of the Numerical Discourses (Aṅguttaranikāya) (Heim 2018, 86):
If someone, having heard a conventional teaching, is able to attain distinction, hav-

ing abandoned delusion and comprehended the meaning, then the Lord teaches a
conventional teaching.
(i.94)
And the same is said of the further teaching. This critical fact should enable us to see that
the Buddhaghosan distinction pertains to hermeneutic context and not to metaphysical levels.
Heim drives this home through analyzing several passages, a particularly extraordinary one
being a comparison of the contextual application of one or the other discursive strategy by the
Buddha with pedagogical code-switching: The Buddha is like a skilful teacher of the Vedas
who gives commentary on them according to the language of the students, be it the Tamil
language or the Andhra language (Heim 2018, 88). If the Buddha taught according to context,
then our reading of his teaching should likewise make the contextual distinction. Again, how
the Buddha taught and how we read his teachings are folded together; hermeneutic context
follows pedagogical context. I believe that this drawing out of techniques from what a text
itself enacts is a critical insight that can be applied outside the Buddhist context: the text yields
methods even as it performs its tasks.
The Continuous Horizon

Now, the idea at the very heart of modern hermeneutics is the temporal connection between
textual context and the reader’s context that Gadamer famously called “the fusion of horizons.”
As has been pointed out, Gadamer’s hermeneutics is deeply involved with the texts of the past,
since his programme assumes that the ancients have something to teach us (Weinsheimer 1985,
133). He constantly warns against the spurious separation from sacred text by which a supposed
“objective knowledge” is gained (Weinsheimer 1985, 183). Scripture is not meant merely for its
own time and place, but it should also apply to situations in time to come since it is revelation.
We can straightforwardly locate Buddhaghosa within this approach. But then Gadamer says
that the consequence of the sacred text’s applicability for the future is that the circumstances
in which its truths may be applied are unforeseen and unknowable (2004, 309). The text must
continue to apply, but in these circumstances, we need a fusion of horizons: “the horizon is the
range of vision that includes everything that can be seen from a particular vantage point” (Gad-
amer 2004, 302). The text’s continued applicability beyond the original temporal boundaries
335
comes through its having an alterity in relation to the reader’s current perspective (“prejudices,”
to use the Gadamerian term) but yet, at the same time, becoming part of that current perspective
and thereby altering it. Understanding therefore consists in the fusion of horizons, in hermeneu-
tically bridging the gaps between then and now. However, Gadamer rejects the notion that in
order to “understand” another historical period, one needs to enter into it, to transpose oneself,
as it were, to become one with the mindset of another time and place. Rather, fusion implies a
common understanding at a level of a “higher universality.” True (traditionary) understanding
in Gadamer’s view consists in the bridging or the “fusing” of horizons; in particular, the present
horizon, which one maintains, and the past horizon, out of which the present horizon develops
(Knotts 2014, 239). As Gadamer clearly indicates, this is particularly and powerfully the case
when the text stands in the relationship of (in)forming the tradition of the reader (2004, 306).
I want to locate the Buddhaghosan hermeneutic here in conversation with Gadamer. While
wrestling with the same issue of the temporal distance and the transtemporal applicability of
sacred text, Buddhaghosa’s framework of contextualization offers a different mode of treating
the horizon of meaning. For him, contextualization is not the objectification of the past that
Gadamer rightly criticizes. Rather, it is the very means through which the text is present, as
also at the same time the means by which the reader makes it present. How, then, does Bud-
dhaghosan contextualization change the terms of the hermeneutic relationship between text
and commentator or between text and reader when they are separated by time (including, in
the case of the last, the cultural rupture of modernity)? If we can discern the outlines of an
answer to that, then we can have an idea of how such a hermeneutic can make a philosophical
contribution to hermeneutics.
Contextualizing to situations was an integral part of how the Buddha taught in the suttas.
So, it seems, he anticipated the experience of Buddhaghosa’s reading. While, in turn, Bud-
dhaghosa did not anticipate our experience and context as such (for he claims no universality
for himself), his method does anticipate our experience in the sense of always calling for con-
textualization. We get an idea of Buddhaghosa’s understanding of the temporal relationship
to the truthful content of the Buddha’s insightful teaching – the Dhamma – in his section on
the contemplative practice called the recollection (anussati) of the Dhamma. The Dhamma,
it is taught by the Buddha, is both “visible here and now” (sandiṭṭhika), and it is “timeless”
(akālika) (VII.76–80 of Ñāṇamoli; Heim 2018, 226). He explains that the Dhamma is “visible
here and now” in the sense that the qualified person can see it for himself when free of lust and
of injurious intent, and free of mental suffering and unhappiness.
Furthermore, the nine-fold world-transcending Dhamma is visible here and now for,
in whatever manner it is attained, it is to be regarded as that reflective understanding
that is to be approached through benefiting from faith in another.
(VII.77)
For the world-transcending Dhamma is the seeing that leads away from the fear of
the cycle of births, through the comprehension from cultivation and the comprehen-
sion from realization. Just as the one who is clothable is worthy of being clothed, so
too the “visible here and now” is what is worthy of being seen here and now.
(VII.79)
There is no time between it and the bearing of its fruit, so it is “timeless.” The “time-
less” takes no time. It does not give of its fruit after a gap of, say, five days or seven
336
Buddhaghosa
days, for it means that it bears fruit immediately after its own occurrence. On the
other hand, what takes time is that which needs some distant time before it bears fruit.
What is that? The good Dhamma of this world.
(VII.80)
While the Dhamma holds beyond the world (lokuttara, “world-transcending”) and its under-
standing is dependent on faith (saddha), it is also the reflective understanding ( paccavek
khaṇañāṇa), attained through attention or consideration, of the recipient of its textual form.
In the final section of this chapter, we will look more carefully at Buddhaghosa’s use of the
term “understanding” ( paññā) and its implications for our interpretation of his philosophical
approach. But for now, it is sufficient to point out that this understanding is not primarily about
epistemic grasp of propositional content; as we saw earlier, reflection ( paccavekkhaṇa) is said
to bring the Abhidhamma monks endless joy and happiness (ananta pītisomanassa).
Now (as it were), VII.79 has it that “here and now” is a quality of transtemporal worth,
not a temporal index. What the Buddha said was here and now at the time of his teaching
(the Dhamma) is the here and now of Buddhaghosa’s commentarial attention and the here
and now of our current reader. This idea, I think, is not so strictly limited in concept (even if
it is in source) to the teaching of Dhamma. Rather, we may think of a sacred text’s ultimacy
as consisting not in its temporal horizon needing to fuse with ours but in its constantly being
what is present here and now.
But we must note that what is transtemporal is not atemporal, for if it were, we would
lose traction on the very discipline of hermeneutics, for we are creatures of time such that a
text beyond time is uninterpretable. Buddhaghosa’s reading of ‘timeless” (akāla) at VII.80
shows this. If we translated akāla literally as “without-time,” we would indeed be in that
quandary. But his reading is that the Dhamma is akāla in the sense that the attainment of
meaning is an instant occurrence. Adroitly, he points to “timeless” being the obverse of
“here and now”; obverse not in the sense of “opposite” but as in the other side of the same
coin. The here and now of readability is also the instant of realizability. He distinguishes
this temporal punctuation from the longer drawn-out span of the consequences in this world
of having understanding of the Dhamma. This, again, is of quite general applicability. If the
continuity of horizon were an explanatory metaphorical alternative to the fusion of horizons,
we may ask how we are to account for the understanding that occurs at the point of fusion
in the Gadamerian paradigm. The Buddhaghosan account points to an answer – the under-
standing is the temporal point of attaining reflective knowledge. It does the work that fusion
does in the Gadamerian account.
Hermeneutics and/to Phenomenology

I have argued for seeing Buddhaghosa as a hermeneuticist, primarily to bring out his cross-
cultural significance for this contemporary discipline of “philosophy” rather than metaphysics
and epistemology, which have for decades been seen as appropriate areas within which to
bring Buddhist thought into conversation with the West. Furthermore, this sketch enables me
to shift to a different angle of inquiry, to offer the interpretation that Heim and I have defended,
that Buddhaghosa has a phenomenology that is methodic all the way through and not directed
towards the identification of ontological irreducibles. What is this angle? It is that we must use
his contextualizing hermeneutics to approach the non-ontological nature of his phenomenol-
ogy. Here again, a Gadamerian formulation offers striking comparison when it comes to the
337
relationship between the contextualized reading of an abhidhamma text and the contemplative
practice of discerning its teachings in Buddhaghosa.
Gadamer himself, in the foreword to the second edition of Truth and Method, explains
that the work is “phenomenological in its method”. . . . By referring to his “method,”
Gadamer means, not a set of procedural rules, but rather the discipline of attending to
things. By calling his method “phenomenological,” he does not thereby subscribe to
Husserl’s account of the phenomenological method but rather indicates that the task
of the enterprise is descriptive – in this case, descriptive of the human experience of
understanding, i.e., of hermeneutical experience.
(Dostal 2002, 251)
As I have indicated, Maria Heim and I have argued at length that the phenomenology of
Buddhaghosa is a descriptive methodology: rigorously structured (“Not any list goes”; Heim
2018, 153) but with potentially endless, ramifying practices of attentively analyzing expe-
rience so that their appropriate classification under the different valences as given in the
canonical material – the good (kusala), the bad (akusala), and the indeterminate (abyākata)
– purifies the life of the practitioner on the path to perfection (see especially Heim 2018,
163–80). The determination of the various dhamma classifications is a continuous practice of
phenomenological analysis that does not call for a reading of the dhammas as elements of an
ontology. Here, rather than reiterating these findings, I want to ask: what does being sensitive
to the hermeneutics of Buddhaghosa demonstrate about the incoherence of then going on to
treat abhidhamma categories ontologically? I suggest that this may thereby demonstrate in a
different way the nature of his phenomenological method. The conclusion is that if we recog-
nize Buddhaghosa as an hermeneuticist of contextuality, we will have a better understanding
of the way he treats phenomenological analysis as an endless exploration and classification
of experience leading to perfection.
Let me show this through a key passage in The Path of Purification, at the start of the
lengthy exposition on the aggregates (Chapter XIV in Ñāṇamoli), which is the first chapter of
the third module of The Path to Purification, namely understanding ( paññā).8 Buddhaghosa has
already taken the monastic reader through virtue (sīla) and contemplation (samādhi), although
the three are always continuously interrelated in the life of the practitioner along the path. Bud-
dhaghosa’s application of understanding will start with the aggregates and proceed to various
Buddhist doctrines, such as suffering, dependent origination, and name-and-form, all of which
follow the programmatic framing of the concept of understanding that we are considering.
The aggregates – form (rūpa), cognition (viññāṇa), feeling (vedanā), conception (saññā),
and syntheses (saṅkhāra) – are the five focal points of the analysis of the human being, meant
to demonstrate that they together generate the conception of a unitary self (attā) where there
is none. I have previously argued that Buddhaghosa’s treatment of the aggregates there, espe-
cially and crucially concerning the first aggregate, “form” (rūpa), is a methodical analysis and
clarification of phenomenological features and not the determination of irreducible ontologi-
cal entities (Ram-Prasad 2018, 125–39). The focus is on breaking down the sense of self into
constituent phenomenological dimensions; it is not an argument for an atomistic ontology to
replace the metaphysics of selfhood. Now, I want to look more closely at Buddhaghosa’s fram-
ing of his description of the aggregates by inquiring into the kinds of understanding ( paññā)
that must be clarified before one embarks on the study of the aggregates.
To lay my cards on the table, if we look at the context of his analysis of the aggregates, it
is clear that what Buddhaghosa undertakes is not an epistemological enterprise by which to
338
Buddhaghosa
determine the means of knowledge that are required to elucidate the irreducible features of the
world. Rather, it is a project of purifying the phenomenology of the practitioner by understand-
ing the valence of experiential features (as taught by the Buddha) in order to transform them.
By attending to how Buddhaghosa frames the reading of that core abhidhamma category,
the aggregates, we can see the purpose of that reading. In short, a hermeneutic sensitivity to
context, which I hope to have shown is the defining feature of Buddhaghosa’s philosophical
approach, brings out the undertaking as one about phenomenological method and not onto-
logical determination.
If we take hermeneutics as one result of the recognition in high modernity that specific-
ity (primarily historical, secondarily cultural) informs the interpretation of what is given as
knowledge, then we are not surprised that Jürgen Habermas thought of interpretive under-
standing as “postmetaphysical thinking” (1992, 34) in the context of a condition under which
religion (by which of course he means European Christianity) has lost its tethering as an
authoritative account of reality but which yet continues to offer spiritual sustenance.
After metaphysics, the nonobjective whole of a concrete lifeworld, which is now

present only as horizon and background, evades the grasp of theoretical objectifica-
tion . . . what has, following the disintegration of metaphysical and religious world-
views, been divided up on the level of cultural systems under various aspects of
validity, can now be put together – and also put right – only in the experiential con-
text of lifeworld practices.
(Habermas 1992, 50–51)
Of course, from Habermas’s perspective, this is because metaphysics as the understanding of

reality as a whole is required for the authority of religion, an authority that cannot be granted
through any expert epistemology in modernity. But I am arguing that we should not saddle
Buddhaghosa’s doctrinal commitments with metaphysics, and therefore I am in agreement
with Habermas – albeit in a way he would not have anticipated – when he says that,
Viewed from without, religion, which has largely been deprived of its worldview
functions, is still indispensable in ordinary life for normalizing intercourse with the
extraordinary. . . . [R]eligious language is the bearer of a semantic content that is
inspiring and even indispensable.
(Habermas 1992, 51)
In effect, what we may claim for a Buddhaghosan hermeneutic is that it bypasses the histori-
cal need to rediscover experiential meaning without metaphysics, for it already functions as a
purely phenomenological method for the exploration of meaning.
Clearly, the relationship between metaphysics and hermeneutics in Buddhaghosa is quite
different than the historical shift that Habermas delineates in the West. Nevertheless, the argu-
ment for postmetaphysical thinking can be read as being about an ametaphysical thinking,
such as we find in Buddhaghosa. By this I mean that Buddhaghosa’s rigorous adherence to
the doctrines taught by the Buddha structures the way he frames various practices of attention,
whether textual or contemplative (or, in fact, both simultaneously). I propose to explore how he
talks about understanding before he proceeds to his analysis of the aggregates (and thereafter,
many other key doctrinal features of the Buddha’s teachings). The location of this description,
before the analysis begins, is clearly meant to prime the reader/practitioner in preparation for
the analysis of key doctrines. Given the scrupulous contextualization that marks his approach
339
to the study of doctrines, what he has to say about understanding will teach us what he means
to do with his subsequent analysis.
What is understanding? Understanding is of many sorts and has many aspects. An

answer that attempted to explain it all would accomplish neither its intention nor its
purpose, and would, besides, lead to distraction; so we shall confine ourselves to the
type intended here, which is understanding consisting in the knowledge from intro-
spective meditation associated with good awareness.
(XIV. 2)
Is this perhaps an oblique reference to – and dismissal of – the flourishing epistemological tra-
ditions in Sanskrit that he is said to have left behind in India? In any case, this is an emphatic
signalling of a tightly focused and pragmatic programme of explication that will take the trans-
formation of phenomenology and the cultivation of valence as the objective of understanding.
Only if we ignore this explicit orientation would we begin looking for the epistemological
pursuit of the constituents of reality.
In what sense is it understanding? It is understanding ( paññā) in the sense of being an

act of understanding ( pajānana). What is this act of understanding? It is cognizance
( jānana) in a pre-eminent mode, distinct from conceptualizing (sañjānana) and
cognizing (vijānana). For although conception and cognition and understanding are
equally of the nature of cognizance ( jānana-bhāva), nevertheless conception is only
the mere conceiving of a phenomenal object (ārammaṇa) as, say, “blue or yellow”; it
cannot bring about the penetration into its characteristics as “impermanent, painful,
and not-self.” Cognition has cognizance of objects as “blue or yellow,” and attains
penetration into its characteristics, but it cannot bring about, by ascetic practice, the
appearance of the path. Understanding has cognizance of objects in the way already
stated, it brings about the penetration of the characteristics, and it also brings about,
by ascetic practice, the manifestation of the path.
(XIV.3)
Understanding is a phenomenologically dynamic treading of the path. Notably, too, Buddha

ghosa is not in the least dismissive of ordinary epistemic functions: conception and cognition
determine our apprehension of the objects of experience. But he moves swiftly on from that to
understanding as an ascetic teleology, the disciplined mental activity of taking what is experi-
enced and directing it towards Buddhist perfection.
However, it is not always to be found where conception and cognition are. But when
it is, it is not separate from those states. And because it is not possible to separate
it thus: “This is conception, this is cognition, this is understanding,” its distinc-
tion is consequently subtle and difficult to see. Hence the venerable Nāgasena said:
“A difficult thing, O King, has been done by the Blessed One.” – “What, venerable
Nāgasena, is the difficult thing that has been done by the Blessed One?” “The dif-
ficult thing, O King, done by the Blessed One was the defining of the formless states
of awareness and its concomitants, which occur with a single object, and which he
declared thus: ‘This is contact, this is feeling, this is perception, this is intention, this
is awareness.’ ” (XIV. 6)
(Milindapañha 2.7.16)
340
Buddhaghosa
This is a place where we see why it is best to think of “ārammaṇa” not merely as “object”
but as “phenomenal object,” that which occurs in experience. What is relevant here are the
phenomenal features, that is, the subjective states that constitute the experience of an object.
For what the Buddha did was take a phenomenal occurrence and tease apart its subtle analytic
components: the Teacher as First Phenomenologist.
We must pay attention to what Buddhaghosa presents as the types of understanding,
because it is an elaborate and ramifying list, and what is – and what is not – in it should
tell us something about how to work with his interpretive exposition (niddesa) of that core
abhidhamma category, the aggregates. He lists twelve types, one that is unitary, five that are
two-fold, four that are three-fold, and two that are four-fold. We cannot go through them
all (one needs to read the whole of The Path of Purification to know what Buddhaghosa is
doing, such is his rhizomatic presentation of that path). But let us look at a few of them and
draw out their implications for the phenomenological method. Some are self-evidently about
spiritual stages, ways of being on the path: one dyad is that which is subject to “cankers”
(sāsava) and free from it, the cankers being the persistence of ideas deleterious to progress
(XIV.3). Another is understanding of the realm (bhūmi) of seeing (dassana) and development
(bhāvana) (XIV.13). We may well begin with the expectation that what is said about under-
standing (or indeed, as happens midway through, about “knowledge” [ñāṇa]) is going to be
epistemological, but it turns out to be descriptions of analytic reflexivity, a fine-grained and
methodical paying of attention to phenomenal states. The reflexivity is not only cognitive; it
is also affective: the fourth dyad has to do with understanding in the mode of good awareness
(kusalacitta) involving meditative attainments (the jhānas) “accompanied by joy” (somanas-
sasahaghata) and likewise good awareness “accompanied by equanimity” (upekkhāsahagata)
(XIV.5).
An important classification of understanding that drives home the phenomenological
method has to do with understanding attained purely by thought (cintāmayā).9 Here, perhaps,
in the middle of this elaborate location of understanding as an integrated way of being, is an
epistemological kernel that might carry some of the flavour of the famous Sanskritic pramāṇa
system of the means of valid, true cognition. He quotes the Analysis (ibhaṅga 768).
Then it is said, what is this understanding only by thought? In organized applica-

tion in the sphere of action, in organized application in the sphere of the crafts, in
the organized application of the eight wisdoms, adaptation, forbearance, perspective,
inclination, opinion, wish, pondering intently on things, with regard to the owner-
ship of action, or which is in conformity with the truth, or is such a type as conforms
with the teachings, “form is impermanent, or feeling . . . conception . . . syntheses, or
cognition is impermanent,” which one attains without hearing it from others – that is
called understanding only by thought.
(XIV.14)
Buddhaghosa goes on to say that understanding attained through others is understanding by

only hearing, and finally, understanding purely through development (bhāvanāmayā) is all
the understanding of one who has attained what is to be attained (XIV.14). Pure reasoning for
Buddhaghosa is pragmatic, dynamic, reflexive on the teachings. There is no structured state-
ment on how to know objects, let alone irreducible particulars such as the factors (dhammas),
as is supposed to be the case in the metaphysical reading of abhidhamma.
Buddhaghosa also gives a triadic definition of understanding that concerns skill in improve-
ment (āya), detriment (apāya), and means (upāya). His description of the understanding that
341
consists of improvement – which is “twofold, the elimination of the harmful and the arous-
ing of the good” (duvidhā anatthahānito atthuppattito ca) (XIV. 16) – clearly shows that
his analysis is going to be oriented to moral cultivation, the transformation of the valence of
experience. He quotes the Analysis.
Here, what is skill in improvement? When one brings factors to mind, and bad fac-
tors that have not yet arisen do not occur, and the bad factors that have arisen are
abandoned by him; or when he brings factors to mind, and good factors that have not
arisen do occur. And good factors that have arisen increase, become abundant, are
perfected in him. This is what is understanding, the act of understanding.
(XIV.16; Analysis, 771)
This is an important consideration when coming to grips with the hermeneutic of Buddhag-
hosa’s analyses of various core Buddhist doctrinal principles such as the aggregates: He treats
understanding as being about valence. If the salience of the understanding of dhammas lies
in the disciplining of thought so as to abandon bad ones and cultivate the good, it becomes
incoherent to think that the dhammas are, for Buddhaghosa, the entitative foundation of real-
ity. There are two reasons for this. One is that, if they are the building blocks of reality, then it
would make no sense to think they can be stopped from occurring, whereas it makes perfect
sense that they can be disciplined into not occurring if we thought of them as experiential
states whose arising can be controlled through meditative skill. The other is that, to put it in
the classic terms of modern Western philosophy, an ontological reading of Buddhaghosa here
would mysteriously treat what “is” in terms of what “ought” to be the case. In other words,
there would be an abyss between what Buddhaghosa says understanding of the dhammas is for
and what the dhammas are held to be on the ontological view. And yet again, to see them as the
factors that structure phenomenology is immediately to make sense of their role as valences to
be abandoned or cultivated on the path to perfection.
In the fourth triad, understanding through introspective meditation that is initiated by

apprehending one’s own aggregates is the interpretation of the subjective. That which
is initiated by apprehension of another’s aggregates outside or forms not bound up
with sense-faculties is the interpretation of the external.
(XIV.19)
Buddhaghosa is not interested in a metaphysical distinction between subject and object; he

simply takes for granted the features of experience by which the subjective and the external are
presented and builds contemplative strategies upon them. This is manifest in his requirement
that understanding here is through vipassana or introspective meditation on what he already
takes for granted is presented in certain ways. Once more, if we are looking for someone
hunting for an ontology, it makes no sense that the subject-object distinction is made without
any further comment on their metaphysical status. Furthermore, the means of attending to the
aggregates is meditative rather than epistemological, for no mode of validating the determina-
tion of ontological constituents is given, whereas the rigorous phenomenological method of
vipassana meditation is given as the mode of interpreting how experience occurs. This is, in
its own striking way, a hermeneutic phenomenology.
Proceeding with his discursive contextualization of the analyses to come, Buddhaghosa
also lists two tetradic types of what he now shifts to calling “knowledge” (ñāṇa): these turn
out to be entirely about suffering (dukkha), its origin, its cessation, and the way leading to it.
342
Buddhaghosa
He talks of the fourfold modes of discrimination ( paṭisambhida): discrimination of mean-

ing (attha), the ordered teachings (Dhamma), linguistic analysis (nirutti), and utter clarity
( paṭibhāna) (XIV. 21), and goes on to state that the Abhidhamma analysis maps each of the
Noble Truths onto the four discriminations. (XIV.24).
In all this, there are five aspects ( pañcahākārā) in common with the understanding of the
one who is still learning and the one who has already learnt perfectly (sekha and asekha):
Here, what is called attainment is the reaching of arahantship. What is called profi-
ciency is proficiency in the word of the Buddha. What is called hearing is hearing
the Dhamma carefully and finding its value. What is called questioning is the inves-
tigation of knotty and explanatory passages in the canonical text, commentaries, etc.
What is called prior discipline is application to the doctrines of previous Buddhas
right up to the stages of conformity and change-of-lineage, by one who has cultivated
going and coming back with the meditation subject.
(XIV.28)
By now we should be well attuned to seeing Buddhaghosa’s exploration of understanding prior

to analysis of categories such as the aggregates as the hermeneutic priming of the practitioner
for disciplined, systematic, and morally configured meditation. Understanding is the organic
term for the entire practice of this phenomenological method. We must pay attention to the
centrality of his hermeneutic contextualization of all the teaching of the Buddha, as the exami-
nation of his techniques for reading show. And once we pay such attention to the context within
which he locates the analysis of dhammas, we notice that he has given us no scope to derive
an epistemological system through which to determine the dhammas as ontological entities.
It would be perfectly possible to read Buddhaghosa’s techniques in isolation and apply
them to our contemporary, cross-cultural philosophical interests, but then we could not claim
to be representing what he was actually doing with Buddhist categories. While I do not deny
the legitimacy of such a comparativist undertaking as the application of Buddhaghosa’s ideas
to concerns far removed from his, I think it is important to see how he can contribute to a
cross-cultural philosophy more directly, from closer to his own interests, through following
his own words across his oeuvre. In taking this path between philosophy and Indology, I have
tried to show that Buddhaghosa’s great contributions to the emerging global philosophy would
be in what we can discern as a highly original hermeneutic phenomenology.
Notes
1 For more on the complex history and meaning of the term Theravāda, see Skilling et al. 2012.
2 For the historical discussion of what is and what is not the work of the fabled fifth-century figure Bud-
dhaghosa, see works by Lance Cousins (2015), Tosiichi Endo (2013), and Oskar von Hinüber (2008).
A relatively recent outline of Buddhaghosa’s oeuvre is given by Ole Pind (1992).
3 See Heim in this volume for the argument that the canonical materials themselves, as found in the
Dhammasaṅgaṇī and Vibhaṅga, should not be interpreted as offering an ontology of irreducible entities.
4 Damien Keown, for example, disparages what he takes to be “dry and disconnected classifications” on
virtue (sīla) (1996, 61).
5 He mentions its concern with the distinction between “definitive meaning” (nītārtha in Sanskrit) and
“interpretable meaning” (neyārtha) and passes over it as being “merely” about explicit and implicit
meaning (Thurman 1978, 25–26). We may contrast this with Heim’s remark about the Pāli tradition
that, while the Buddha sometimes spoke in nītattha and at other times in a neyyattha manner, “often
the first interpretive task is to discern which is which” (2018, 7), an undertaking which is surely not a
simple matter, let alone a simplistic one.
343
6 The Hindu traditions combine these through the distinction between a fixed set of “heard” (śruti) and
an expansive category of “remembered” (smṛti) texts that call for different exegetical strategies.
7 For a spirited defence of ontotheology as a sustainable approach to Christian thought, see Adams 2014.
8 I use the Chapter.Paragraph style of reference given in Ñāṇamoli’s translation. Although I have ben-
efited from this major work, translation in the rest of this chapter is mine.
9 The Pali Text Society’s Dictionary offers the strict meaning of cintāmayā as “consisting of pure
thought” but then, in a leap of lexical faith, adds as an alternative “metaphysical.”
Bibliography
Adams, Marilyn. 2014. “What Is Wrong with the Ontotheological Error?” Journal of Analytic Theology
2: 1–12.
Backman, Jussi. 2016. “Hermeneutics and the Ancient Philosophical Legacy. Hermēneia and Phronēsis.”
In The Blackwell Companion to Hermeneutics, edited by Niall Keane and Chris Lawn, 22–33. Oxford:
Blackwell.
Cousins, Lance. 2015. “The Case of the Abhidhamma Commentary.” Journal of the Association of Bud-
dhist Studies 36/37: 389–422.
Dostal, Robert J. 2002. “Gadamer’s Relation to Heidegger and Phenomenology.” In The Cambridge Com-
panion to Gadamer, edited by Robert J. Dostal, 247–66. Cambridge: Cambridge University Press.
Endo, Toshiichi. 2013. Studies in Pāli Commentarial Literature: Sources, Controversies, and Insights.
Hong Kong: Centre of Buddhist Studies.
Gadamer, Hans-Georg. 2004. Truth and Method. Translated by Joel Weinsheimer and Donald Marshall.
2nd ed. London: Continuum.
Gold, Jonathan. 2015. Paving the Great Way: Vasubandhu’s Unifying Buddhist Philosophy. New York:
Griffiths, Paul J. 1989. “Buddha and God: A Contrastive Study in Ideas About Maximal Greatness.”
Journal of Religion 69 (4): 502–29.
Habermas, Jürgen. 1992. Postmetaphysical Thinking: Philosophical Essays. Translated by William Mark
Hohengarten. Cambridge, MA: MIT Press.
Heidegger, Martin. 1962. Being and Time. Translated by John Macquarrie and Edward Robinson. New
York: Harper & Row.
Heim, Maria. 2014. The Forerunner of All Things: Buddhaghosa on Mind, Intention, and Agency. New
———. 2018. Voice of the Buddha: Buddhaghosa on the Immeasurable Words. New York: Oxford Uni-
versity Press.
Heim, Maria, and Chakravarthi Ram-Prasad. 2018. “In a Double Way: Nāma-Rūpa in Buddhaghosa’s
Phenomenology.” Philosophy East and West 68 (4): 1085–115.
Jaini, Padmanabh S. 2001. “On the Sarvajñatva (Omniscience) of Mahāvīra and the Buddha.” In Col-
lected Papers on Buddhist Studies, edited by Padmanabh S. Jaini, 97–122. New Delhi: Motilal
Banarsidass.
Keown, Damian. 1996. “Morality in the Visuddhimagga.” Journal of the International Association for
Buddhist Studies 6 (1): 61–75.
Knotts, Matthew W. 2014. “Readers, Texts, and the Fusion of Horizons: Theology and Gadamer’s Her-
meneutics.” Acta Universitatis Carolinae Theologica 4 (2): 233–46.
Lopez, Donald S. 1998. Elaborations on Emptiness: Uses of the Heart Sutra. Princeton: Princeton Uni-
versity Press.
Matilal, Bimal. 1986. Perception: An Essay on Classical Indian Theories of Knowledge. Oxford: Clar-
endon Press.
Mills, Ethan. 2018. Three Pillars of Skepticism in Classical India: Nāgārjuna, Jayarāśi Bhatta and Śrī
Harṣa. Lanham: Lexington Books.
Ñāṇamoli, Bhikkhu, trans. 1975. The Path of Purification (Visuddhimagga). Kandy: Buddhist Publica-
tion Society.
———. 1987. The Dispeller of Delusion. Translation of Sammohavinodanī. Revised by L. S. Cousin,
Nyanaponika Mahāthera, and C. M. M. Shaw. London: Routledge and Kegan Paul.
Ó Murchadha, Felix. 2016. “Religion.” In The Blackwell Companion to Hermeneutics, edited by Niall
Keane and Chris Lawn, 77–85. Oxford: Blackwell.
344
Buddhaghosa
Perrett, Roy W. 1989. “Omniscience in Indian Philosophy of Religion.” In Indian Philosophy of Religion,
edited by Roy W. Perrett, 125–42. Dordrecht: Kluwer Academic Publishers.
Pind, Ole Holten. 1992. “Buddhaghosa: His Works and Scholarly Background.” Bukkyō kenkyū 21:
135–56.
Ram-Prasad, Chakravarthi. 2001. Knowledge and Liberation in Classical Indian Thought. Basingstoke:
Palgrave Macmillan.
———. 2018. Human Being, Bodily Being: Phenomenology from Classical India. Oxford: Oxford Uni-
versity Press.
Ricoeur, Paul. 1995. “Manifestation and Proclamation.” In Figuring the Sacred: Religion, Narrative, and
Imagination, edited by Mark I. Wallace, 48–67. Minneapolis, MN: Fortress Press.
Skilling, Peter, Jason A. Carbine, Claudio Cicuzza, and Santi Pakdeekham, eds. 2012. How Theravāda is
Theravāda? Exploring Buddhist Identities. Chiang Mai: Silkworm Books.
Thurman, Robert A. F. 1978. “Buddhist Hermeneutics.” Journal of the American Academy of Religion
XLVI (1): 19–39.
Von Hinüber, Oskar. 2008. A Handbook of Pali Literature. New Delhi: Munshiram Manoharlal.
Weinsheimer, Joel C. 1985. Gadamer’s Hermeneutics: A Reading of Truth and Method. New Haven, CT:
Yale University Press.
345
20
BHĀVIVEKA
Madhyamaka Dialectic, Doxography, and
Soteriology
The Author and His Context

The name “Bhāvaviveka” was first introduced to a Western audience by Eugene Burnouf (1801–
1852), designating an early Buddhist Madhyamaka philosopher whose life and seminal work
generated substantial scholarly interest (Burnouf 1844, 499).1 The exact recording of the name,
however, varies depending on sources.2 Tibetan sources mainly gave the name Bhavya, though it
is not found in any known Sanskrit text. Chinese translations consistently use a set of characters
matching the name Bhāviveka, from the Sanskrit bhā, for ‘light,’ ‘brightness,’ and so on, and
viveka, for ‘discrimination’ and ‘discernment.’ Hence, what looks like a laudatory title refers
to a man of ‘brilliant intellect’; it is a suitable name for the sharp dialectician who was the first
Indian philosopher to systematically compile, synthesize, and criticize all the prominent philo-
sophical views of his time, including those alien to his Buddhist tradition, within a coherent
dialectical compendium, known in English as the Verses on the Essence of the Middle Way, the
Madhyamakahṛdayakārikā (MHK). This doxographical3 text is the central focus of the present
discussion.
Although seven other titles are attributed to the famous dialectician, only two more are gen-
erally acknowledged to be from his hand: the Lamp of Wisdom (Prajñāpradīpa, PP)4 and the
Jewel in the Hand Treatise (Karatalaratna, KR) (Iida 1980, 12–19; Watanabe 1994, 9–15).
The first is Bhāviveka’s commentary on Nāgārjuna’s Fundamental Verses on the Middle
(Mūlamadhyamakakārikā, MMK), which was commented upon by Avalokitavrata and is now
available only in Tibetan and Chinese. The second is Bhāviveka’s final and shortest of the three
works mentioned and appears to be a concise and more accessible version of the MHK, summa-
rizing Bhāviveka’s arguments against the Sāṃkhya, Vaiśeṣika, and Yogācāra. It is available only
in Chinese (Dàchéng zhǎng zhēn lùn/大乘掌珍論).5 Thus, the MHK alone is extant in Sanskrit,
while its commentary, the Blaze of Reasoning (Tarkajvālā, TJ), is preserved in Tibetan only.
A shorter text is sometimes added to the list of Bhāviveka’s accepted oeuvre, as in Eckel and
Eltschinger (2019), A Summary of the Meaning of the Madhyamaka (Madhyamakārthasaṃgraha,
MAS), which focuses on Bhāviveka’s understanding of the two levels of truth in Buddhism:
conventional truth (saṃvṛtisatya) and ultimate truth ( paramārthasatya).
Compared with some other Indian philosophical figures, putting a date on the life of
Bhāviveka is a relatively easy task because his works engage many well-known Buddhist
346 DOI: 10.4324/9781351030908-31

Bhāviveka
thinkers such as Dharmapāla, Buddhapālita, Dignāga, and Sthiramati. Today, scholars believe
Bhāviveka lived around 490–570 or 500–570 CE (Kajiyama, 1968–1969). Pinpointing a precise
location for the author’s life is a less certain enterprise. Chien Y. Hsu concluded, from a review
of the many sources at hand, that “Bhāviveka was born into a royal family in Dhānakataka,
which is presently a western neighbor of Amarāvatī” (Hsu 2013, 14). By the second century
CE, Amarāvatī was one of the greatest Buddhist centers of the eastern Deccan (Verardi 2011,
74). By the sixth century, when Bhāviveka was active, it was one of the three major centers of
Buddhist studies in India, along with Nālandā in the east and Valabhī in the west.
Bhāviveka was thus active in the agonistic context of sixth-century India. By the second
half of the fifth century CE already, we notice the beginning of a sustained philosophical con-
frontation between competing groups in the form of a systematic approach to debate (vāda)
and critical examination ( parīkṣā), making use of new developments in the field of argumen-
tative reasoning (hetuvidyā), which were to develop into a fully epistemological framework
( pramāṇa), used and debated throughout the Indian cultural sphere. It is within this context
that Bhāviveka began his intellectual work.
But why would philosophers suddenly feel obliged to strengthen their position and to
engage with the views of others, and not only with those circulating within their own religious
tradition? Although the trend was already set in the second to third centuries by the works
of Nāgārjuna and Āryadeva, for example, it reaches a new systematic height in Bhāviveka’s
oeuvre. Vincent Eltschinger suggests that these innovations, though introduced by new devel-
opments in thought, were also influenced by a given sociohistorical context: “a dramatic
increase of hostility towards Buddhism as the Gupta dynasty starts to crumble” (Eltschinger
2014, 72). If doctrinal issues could fuel such animosity, one can safely assume that, as else-
where, Indian religions and politics became intimately intertwined in an explosive mix. Eltsch-
inger convincingly argues that the social dynamics of the sixth century onwards brought about
two major innovations in Indian Buddhism: Buddhist esotericism6 (tantra) and epistemology
( pramāṇa). He notices how both “legitimate themselves in a structurally homologous way,
and in quite martial terms, as that which, by defeating the outsiders, removes the obstacles to
the path towards liberation” (Eltschinger 2014, 174). The work of Bhāviveka gains in depth
and historical significance when framed within that context: the emergence of a new apolo-
getic strategy of Buddhism, making use of epistemology to defend itself and to attack the posi-
tions of others. Debate, indeed, preoccupied the sixth-century dialectician. However, assessing
Bhāviveka’s MHK, one notices how this dialectical endeavor was not a purely worldly pursuit
after fame and scholarly career, nor a mere apologetic strategy, but it was conceived as an
essential dimension of the spiritual practice leading to the highest realization of religious life.
Few attempts have been made to reflect on Bhāviveka’s work in the light of the turbu-
lent sociohistorical context of the sixth century CE. There has been ample discussion of the
reasons that led Tibetan doxographers to classify Bhāviveka as a proponent of the trend of
Madhyamaka which favors independent (svatantra) inference (anumāna) in debate, a school
known as Svātantrika-Madhyamaka, in opposition to the school championed by Candrakīrti,
known as Prāsaṅgika-Madhyamaka, which rather favors reducing its opponent’s assertions to
their absurd ‘consequences’ ( prasaṅga). More precisely, since Bhāviveka’s methodology also
seems to favor direct references to the texts (sūtras) attributed to the Buddha rather than resort-
ing to the later scholastic materials of the Abhidharma literature, some Tibetan doxographers
classified him as a Sautrāntika-Svāntantrika-Madhyamaka thinker. They thus considered him a
Mādhyamika commentator who relies on the sūtras and on independent inference (svatantra-
anumāna) in support of his arguments. This classification stems from Bhāviveka’s logical
method deployed to elucidate Nāgārjuna’s Fundamental Verses on the Middle Way in his Lamp
347
of Wisdom. Critical of the previous Buddhapālita’s conventional method of argumentation on the

same verses, relying strictly on the exposition of absurd consequences ( prasaṅga), Bhāviveka
rather promotes a rendering of Nāgārjuna’s arguments into the form of a standard three-part
inference inherited from the famous Buddhist logician Dignāga (480–540 CE). Candrakīrti
eventually came to the defense of Buddhapālita, whence the Tibetans saw in this opposition
the birth of two separate lines of interpretation: the Svātantrika branch of Madhyamaka inau-
gurated by Bhāviveka and the Prāsaṅgika branch defended by Buddhapālita and Candrakīrti.
While much attention has been given to the philosophical implications of both positions, lit-
tle consideration has been given to the sociohistorical context Bhāviveka was facing, such
as the rising antagonism between Buddhism and Brahmanism, an ideological confrontation
which likely contributed to his stance on inference (anumāna). As Giovanni Verardi rightly
stressed, “Indian Buddhism should not be studied per se, but in counterpoint with Brahmanical
theorisations” (2011, 106).7 Taking into consideration the pan-Indian context of philosophical
disputation (vāda) where one is required to assert his own position prior to engaging in the
fray of debate, a prerequisite already stated in the Brahmanical Nyāya-Sūtra-s of Gautama, and
knowing how crucial debates had become for the financial and physical security8 of the debater
and of his following, one might better appreciate the desire of Bhāviveka to elaborate on and
defend a solid dialectical position of his own in the name of the Middle Way (Madhyamaka).
Status of Bhāviveka’s Work in Buddhist Philosophy

For the sake of the overall coherence of the present Handbook, Bhāviveka’s chapter has been
placed within the section on ‘philosophical commentators.’ As attested by his earliest work,
the Lamp of Wisdom, Bhāviveka was indeed concerned with the interpretation of Nāgārjuna’s
thought, which laid the foundations of Madhyamaka Buddhism. Yet categorizing him as a
philosophical commentator should not obscure the fact that Bhāviveka was an original thinker
in his own right, as well as an acute synthesizer and staunch apologist. These salient features of
his thought are observable in his magnum opus, the Verses on the Essence of the Middle Way.
Bhāviveka’s interpretation of his revered Madhyamaka master, Nāgārjuna, quickly became
authoritative. But, like most commentators, he was not satisfied with repeating verbatim the
stances of his predecessors. For example, as shall be discussed more in depth, his logical
method of argumentation required him to finely distinguish the two levels of truth, bring-
ing forth a more nuanced application of this distinction than that articulated by Nāgārjuna.
Bhāviveka also extended the scope of Nāgārjuna’s dialectical method to encompass the philo-
sophical developments of his days, such as the three-part inference (anumāna) inherited from
Dignāga and the epistemological framework of the standard means of knowing ( pramāṇa).
In India, according to reports of Chinese pilgrims, Bhāviveka’s influence was so important
that he came to be considered the legitimate successor of Nāgārjuna. Nāgārjuna and his pupil
Āryadeva (second–third century CE) did not seem to use the term “Mādhyamika” to desig-
nate their philosophy. It was mainly due to Bhāviveka that it acquired its name and became
an established philosophical system, equipped with a rigorous logical and dialectical meth-
odology to engage with opposition – something which was missing when the Mādhyamikas
resorted to prasaṅga alone. Modern scholars have argued that the rigor and the norm of the
Mādhyamika system owes a great debt to Bhāviveka and that most later Indian Mādhyamika
commentators followed in his footsteps (Iida 1980, 19–23).
Prior to Bhāviveka, as Mahāyāna philosophy developed based on a new set of sūtras, nota-
bly the ‘Perfection of (Transcendent) Wisdom’ literature (the Prajñāpāramitā sūtras, composed
somewhere between the first century BCE and the sixth century CE), proponents of the new
348
Bhāviveka
movement, including Mādhyamika and Yogācāra scholars, made a particular effort to distinguish
their commentaries from those of the Ābhidharmikas. Such treatises dialectically engaged with
the views of Buddhist and non-Buddhist thinkers.9 Bhāviveka’s MHK is an excellent example
of this trend, which emphasizes debate and logical arguments. Āryadeva, an earlier proponent
of Nāgārjuna’s thought, in his Catuḥśataka and Śataśāstra, had already critically engaged the
views of non-Buddhist schools such as Sāṃkhya, Vaiśeṣika, Jaina, Lokāyata, and even Vaiṣṇava
and Śaiva (Warder 1970, 368–69). Bhāviveka, however, was the first Mādhyamika to adopt the
method of the ‘logical science’ (hetuvidyā) developed by Dignāga. He established his own strat-
egy of presenting independent (svatantra) logical demonstrations – inference (anumāna) – in
defense of the Madhyamaka doctrine, in contrast to prior Madhyamaka methodology, which
had employed reductio ad absurdum arguments ( prasaṅga)10 against the opponent’s thesis
without presenting any counter-thesis. Bhāviveka’s major inspiration for the composition of
a doxographical treatise (the MHK), according to Olle Qvarnström, may have been Dignāga’s
Compendium of Valid Cognition (Pramāṇasamuccaya, PS), which addresses Nyāya, Vaiśeṣika,
Sāṃkhya, and Mīmāṃsā, drawing on the methods of the epistemological ( pramāṇa) tradition.
Bhāviveka’s further innovation would thus have been to supplement Dignāga by criticizing from
the Mādhyamika perspective the various schools known to him which discuss epistemological
and logical claims ( pramāṇa), adding the Vedānta system, for example, to the ones found in
the Pramāṇasamuccaya (Qvarnström 1988, 6). Moreover, I have argued that Bhāviveka’s pres-
entation of each system of thought within the MHK follows a systematic dialectical teleology
meant to establish his philosophical position as the sole valid soteriological method, a rhetorical
device which inspired subsequent Indian doxographies (Bouthillette 2020).
The Doxographical11 Method and Its Function Within the MHK

The Sanskrit manuscript of the Verses on the Essence of the Middle Way was first discovered
by the adventurer Pandit Rāhula Sāṅkṛtyāna in August 1936 in the archive of Shalu (zhwa lu)
monastery. Iida informs us that the manuscript, ascribed to Bhavavadviveka, was written in
Rañjanā script.12 The MHK consists of roughly 927 anuṣṭubh verses (of two lines, each with
16 syllables) and is divided into 11 chapters. It is one of the first comprehensive and system-
atic study of the doctrines (siddhānta-s) of all major schools of Indian philosophy of the time
and was composed when these philosophies were still in the process of formation. Hence, it is
an invaluable historical document, even if its purpose is dialectical and not historiographical.
At the same time, or rather in the first place, as an independent work, it presents an original
articulation of Bhāviveka’s own Madhyamaka system. In that sense, it is also a genuine philo-
sophical reflection, with numerous witty metaphors and poetic allusions to a wide array of
pan-Indian literary and cultural lieu communs.
If we locate the MHK as the first example of a historically extended and broad literary
genre of systematic philosophical writings, generically referred to as ‘Indian doxography,’
then the overall structure of the text becomes a key element to interpret its philosophical
method and the way its worldview is articulated in relation to competing claims. Hence, it
is significant to notice that the structure of the whole text develops out of a primary section
presenting the author’s own position elaborated through an emptying out of Abhidharmic cat-
egories. This is to say that the introductory section of the MHK, Chapter Three, discusses core
topics in scholastic Abhidharma, highlighting how every constituent (dharma) of any mental
and physical event listed within the Abhidharma account of sentient experience is devoid of
‘intrinsic nature’ (svabhāva) and is thus empty (śūnya). This initial treatment of Abhidharma
categories allows Bhāviveka to establish the position of his own Madhyamaka system without
349
actually enunciating an independent set of propositions in its name. This method stands in
sharp contrast to the doxographies composed by philosophers defending other systems. Wil-
helm Halbfass observes, for example, that the structure of Advaita Vedānta doxographies,
such as the ones doubtfully attributed to Śaṅkara and Mādhava, tends to present a hierarchi-
cal classification at whose apex stands the Vedānta (Halbfass 1988, 351). In the case of the
MHK, every view is refuted from the beginning based on the no-view principle of the Madh
yamaka, expressing the perspective of the ultimate truth, which itself barely admits being a
view. Indeed, the Madhyamaka subscribes to an attitude found early on in the discourses of
the Buddha, in the Collection of Discourses (Sutta Nipāta), for example, to the effect that the
speculative and religious views (dṛṣṭi) of non-Buddhists are tainted mental states obscuring
the immediate perception of reality. For this reason, the Madhyamaka presents its dialecti-
cal methodology as an antidote devised to alleviate the affliction of those who grasp onto
views. Hence, it does not elaborate a view of its own but rather negates the views of others, as
medicine removes diseases to restore health, in order to open their eyes on the true nature of
things, the ultimate ‘emptiness’ (śūnyatā) or ‘dependent co-arising’ ( pratītyasamutpāda) of all
phenomena. Medical metaphors alluding to the healing function of the Madhyamaka pervade
the MHK. For example, in Chapter Nine, Bhāviveka argues that
Then, it should be granted that, since it is the antidote to that ignorance, liberation
comes from knowledge, just as freedom from sickness comes about through medi-
cine, because it is the antidote to illness.
(MHK 9.21, Bouthillette 2020, 37)
Accordingly, Madhyamaka is not a mere theoretical view but a remedy to the various forms
of mental grasping caused by speculative philosophies. This is what Bhāviveka expresses in
the following passages from the MHK, translated by Chikafumi Watanabe, where the eye is
clearly identified as the locus of application for the Madhyamaka’s remedy:
When the eye-disease passes away, one whose eye becomes completely clean and
pure does not see the hair (keśa), mosquito (maśaka), two moons (dvicandra) and the
eye in a peacock’s plumage (śikhicandraka).
Likewise, when the eye-disease of darkness which envelopes what is to be known
and defilements passes away, a wise man whose eye has become pure by means of
proper knowledge does not see anything.
(MHK 3.251–2, Watanabe 1994, 85)
Not seeing anything (superimposed over reality) metaphorically captures the highest wisdom
gained through the Mādhyamika negational analysis of reality. Thus, the Madhyamaka is said
to be a non-view. It does not add a view to one’s mind; it removes existing ones.
As Jan Westerhoff notes, the general position of the Madhyamaka is not foundational, nor
does it develop out of the “theory-free or unsophisticated conventionalities of the ordinary
man on the street” (2014, 123). Rather it works out of the intricate analytical framework of
the Abhidharma, which it accepts as a somewhat reliable depiction of the conventional world,
once properly deconstructed to be made svabhāva-free, free of any notion of ‘intrinsic nature.’
In order to point out the ultimate, then, Mādhyamikas like Bhāviveka do not elaborate their
own systematization of conventional categories but analyze and refute existing ones. In brief,
they deconstruct the conventions of others to assert the ultimate truth, the no-view of emp-
tiness. Bearing this method in mind, doxography, within the MHK, becomes an ingenuous
350
Bhāviveka
soteriological and propaedeutic device where ‘poison’ is used as ‘medicine.’ It indicates that
all doctrines presented within the MHK are taken as more or less problematic depictions of the
conventional world to be analyzed and ultimately relinquished by those seeking ultimate truth.
The sequence of presentation of each view within the MHK suggests that the further away
a system stands from the initial exposition of the author’s own stance, the more misguided it is
and the more it stands in contrast to the Madhyamaka.13 This overall scheme confirms Wester-
hoff’s observations to the effect that:
[W]hen the Mādhyamika deliberates between different systems of categories at the

conventional level he is considering two main criteria: (a) its general philosophical
usefulness and (b) the extent to which the system can be made ‘svabhāva [intrinsic
nature]-free.’ This allows him to rank such systems and provides criteria to choose
between systems, none of which have any claim to truth at the absolute level.
(2014, 131)
This might explain the particular virulence of Bhāviveka’s criticism towards the Mīmāṃsakas
in the last chapter, a group which he discourteously introduces as the ‘shameless’ (anapatrapa)
who value ritual acts above meditation and knowledge.
The MHK opens on a brief discussion about safeguarding the ‘aspiration towards awak-
ening’ (bodhicitta-aparityāgaḥ) in Chapter One, followed by a reflection on the significance
of monastic vows (muni-vrata-samāśrayaḥ) in Chapter Two and a revision of foundational
Abhidharma doctrinal expositions examined in a way which reveals the ‘no-view’ approach of
the Madhyamaka propounded by Bhāviveka in Chapter Three. Thus, the third chapter, titled the
Quest for the Cognition of Reality (tattva-jñāna-eṣaṇā), employs the analytical method initiated
by Nāgārjuna, submitting familiar Abhidharmic ontological categories to an analysis which
shows that none of them has intrinsic nature (svabhāva), a hermeneutical strategy found in the
text In Praise of the Supramundane One (Lokātītastava), the Verses on the Heart of Depend-
ent Origination (Pratītyasamutpādahṛdayakārikā), and similar texts. V. V. Gokhale suggests
that these first three chapters were originally composed by Bhāviveka, under the title Penetra-
tion into the Essence of Ultimate Reality (Tattvāmṛtāvatāra), to form the core of the MHK,
while the remaining chapters were added later (1985, 78). However, since the whole MHK is
accepted as coming from the hand of the same author, Gokhale’s remarks need to be moderated.
There are numerous allusions to non-Buddhist tropes within these three initial sections, which
suggests that the author already had in mind engaging with non-Buddhist materials.
In these first three chapters, it is argued that reality (tattva) is to be discerned through the
eye of wisdom ( prajñā), a higher intelligence essential for liberation, which Bhāviveka under-
stands as that which is in conformity with the highest truth ( paramārtha), the nectar (amṛta)
of the Madhyamaka. It is that which wise men seek. The introductory verses of Chapter Three,
translated by Iida Shotaro, explain that
Whoever has the eye consisting of knowledge, has the eye; the other [eye] is not [it].
Therefore the wise man should be dedicated to the quest for knowledge of ultimate
reality (tattva-jñāna).
(MHK 3.1)
One bereft of intelligence is as if eyeless even with a thousand eyes. Because he does
not see the real path to heaven and release.
(MHK 3.3, Iida 1980, 56–57)
351
Ultimately, perceiving reality (tattva) amounts to seeing the buddha through nonconceptual
cognition. In MHK 4.2, for example, Bhāviveka explains, that, “[t]he Teacher’s body is not the
locus of nonconceptual cognition, because it is a body, like the body of a cowherd.”14 Thus, as
explained in the commentary (TJ), to have nonconceptual cognition of the Buddha amounts
to, for example, apprehending the four noble truths and so on. To ‘truly’ see the Buddha means
to nonconceptually cognize the truth of the teachings, not merely to see the teacher’s body. In
brief, the expression ‘seeing the Buddha’ is but a mere linguistic expedient. Enjoying para-
doxes and apophatic statements, Bhāviveka describes how:
Seeing the Buddha is something which belongs to lower truth (saṃvṛti). [It is said to come]
from removing the dust of afore-mentioned deficiency [and] from lighting the torch of
true dharma (saddharmolkāśana). But this is a conventional [expression] as an assistance
towards activity (kriyopakārarūpa). Because what is similar to illusion never takes place,
is not truth, from the standpoint of [the highest] truth, that (seeing) is not really seeing. It
[really seeing the Buddha]15 is beyond inference (apratarkya), is indiscernible (avijñeya),
cannot be demonstrated (anirūpya), does not have an illustrative example (anidarśana),
is without any mark, is devoid of any image (nirābhāsa), is beyond mental constructions
(nirvikalpa), is beyond words (nirakṣara) and is to be awakened by the mind of the viewer
( paśyato buddhiboddhavya) and is an act of seeing from the standpoint of non-vision.
(MHK 3.244–46, Watanabe 1994, 82–83)
In paradoxical words, the philosophy of the Madhyamaka is a non-philosophy in the sense

that it is not a doctrinal ‘view’ detailed and supported by an independent system of reasoning.
Its highest truth is not established through mere syllogisms and arguments, nor through mere
scriptures. Madhyamaka rather makes use of such means to get rid of views altogether, since
they prevent one from ‘seeing’ reality as it is. Logical devices are skillful expedients on the
path, and Bhāviveka makes full use of them every time he analyzes down a view presented by
an opponent or, again, when it comes to analyzing the Buddhist teachings. But, by themselves,
these expedients are not establishing some sort of highest truth. They are removing obstacles
to seeing, like medicine removes diseases.
In the remaining chapters of the MHK, after having presented his own position, Bhāviveka
launches his doxographical conquest. To highlight the dramatic potential of the text, I read
Bhāviveka’s philosophical venture like the conquest of a ‘World Monarch’ (cakravārtin) of
philosophy, as he begins his dialectical campaign on his own turf by directing his attacks at the
doctrines of previous Buddhists, which he regroups under the label of the ‘hearers’ of the Bud-
dha’s teachings (the śrāvakas) (MHK chapter 4: śrāvaka-tattva-viniścayāvatāra) and of the
proponents of Yogācāra or the proponents of consciousness (vijñānavādins; MHK chapter 5:
yogācāra-tattvaviniścayāvatāra). Through his corrective critique, Bhāviveka’s MHK strategi-
cally unifies the realm of Buddhism under his banner before marching into Brahmanical lands.
His criticism is meant to stress that not only are the teachings of the Great Vehicle (mahāyāna)
– in opposition to those of the hearers – authoritative but that their presentation articulated
within his Madhyamaka conveys their ultimate meaning – a reply to the exclusivity claim of
the Yogācāras (Eckel 2008, 62–81). Malcolm David Eckel summarizes Bhāviveka’s critic of
the two groups of Buddhist opponents as follow:
In the case of the Yogācāras, the error takes two forms: they fall into the extreme of
“improper denial” (apavāda) by denying the reality of imagined identity, and they
352
Bhāviveka
fall into the extreme of “improper reification” (samāropa) by affirming the reality of
dependent and absolute identity. In the case of the Śrāvakas, the mistake has to do
principally with an act of reification (they take the eightfold path as if it were a real
thing).
(2008, 80–81)
Bhāviveka invites these competing schools of Buddhism to abandon their extreme views and
to join ranks and adopt what he considers the best approach to interpreting Buddhist scriptures,
the Madhyamaka.
Following these intra-Buddhist skirmishes, Bhāviveka expands his criticism to the non-
Buddhist school of the Sāṃkhya (MHK chapter 6: sāṃkhya-tattvāvatāra), portrayed as enter-
taining two different theories concerning the process of knowledge, which, according to Olle
Qvarnström, are “reminiscent of the doctrine of two truths (dvisatya)” found in Madhyamaka
Buddhism and in Advaita-vedānta, among others (2012, 398). In the following chapter (MHK
chapter 7: vaiśeṣika-tattvāvatāra), the theory of liberation of the Vaiśeṣika becomes the main
target.16 Then, Bhāviveka becomes the first-ever Buddhist critic of Vedānta (MHK chapter 8:
vedānta-tattvaviniścaya) as such, particularly criticizing its theory of the self as having an
intrinsic nature and suggesting that all that is well spoken in Vedānta was taken from Bud-
dhism.17 Finally, the author kept his utmost wrath for the Mīmāṃsakas (MHK chapter 9:
mīmāṃsā-tattvanirṇayāvatāra). This last chapter, however, presents a hodgepodge of Brah-
manical views going beyond what came to be traditionally designated as Mīmāṃsā, ridiculing
the belief in such deities as Śiva and Kṛṣṇa, for example. In the end, after having trampled
the Brahmins, Bhāviveka finally turns his gaze towards the notion of omniscience (sarvajña-
siddhinirdeśa), pivotal in Jainism, Yoga, and Śaivism, on which he will spend a few verses, as
in a postscript, before setting the final boundary of his doctrinal conquest. This overall dialecti-
cal strategy is meant to conquer the entire realm of philosophy that preoccupied Bhāviveka. It
sets the tone for later Indian philosophical compendia.
The eloquence of the MHK, gilded with subtle allusions and humor, at times incredibly
scathing, also makes it a literary masterpiece. On several occasions, Bhāviveka is able to skill-
fully appropriate non-Buddhist tropes and repurpose them in a Buddhist context. For example,
the concluding verses of the second chapter present the purpose of the MHK, its method, and
its target: Bhāviveka calls upon his brethren and his opponents to investigate their own beliefs
through sound reasoning ( pratisaṃkhyāna) in order to realize the truth which lies therein, the
resplendent Sāvitrī of dependent origination. This reference to a famous Brahmanical devo-
tional prayer would immediately be recognized by Bhāviveka’s audience, identifying Sāvitrī
with the solar deity found in the Ṛg-veda and made famous by the celebrated verses RV. iii,
62.10, also called gāyatrī, particularly used during the initiations of the “twice-born” or brah-
mins. Here, Bhāviveka suggest that the real light of wisdom (remember Bhāviveka’s earliest
work, the Lamp of Wisdom or Prajñāpradīpa) is the one shining in emptiness (śūnyatā). This
was the impetus of the first Indian systematic philosophical doxography.
Thus, the MHK is not a mere historiographical account of sixth-century Indian philoso-
phy but the exposition of a dialectical training aimed at converting the views of those who
hold erroneous ones; at eradicating doubts in minds confused by competing claims; and at
consolidating the insight of the others who, having given birth to knowledge already, seek
to see ‘no-thing.’18 Within philosophy, Bhāviveka is chiefly concerned with soteriology. The
fundamental problem in each of the ideas he refutes is that they are impediments to liberation.
The rest is method.
353
Nature of the Path and Purpose of the MHK

In the eyes of Bhāviveka, wisdom ( prajñā) is gained through the analytical process of nega-
tion, which leads one to see through ‘not seeing’ (adarśana), ultimately leaving one with no
view. This is made clear by his definition of ultimate wisdom:
Ultimate wisdom negates the entire network of concepts, and it moves without mov-
ing into the clear sky of reality, which is peaceful, directly known, nonconceptual,
nonverbal, and free from unity and diversity.
(MHK 3.10–11, Eckel 2008, 41–42)
As already elucidated by Nāgārjuna, ultimate reality ( paramārtha), in Bhāviveka’s system,

relies on a corresponding relative (saṃvṛti) truth.19 In terms of Madhyamaka practice, the crux
of the matter is to know and explain how to perceive relative views and similarly attain ulti-
mate reality, the non-view, ‘liberation’, in other words, how to move from a produced and thus
impermanent conventional knowledge to a permanent nonconceptual one. For that purpose,
Bhāviveka innovated by integrating formal logical demonstrations, inherited from Dignāga,
into his therapeutic methodology. To better understand how this process is supposed to work,
one must become familiar with the three possible ways Bhāviveka understood the meaning of
‘ultimate reality.’ Conveniently, Hsu has summarized how, for Bhāviveka,
ultimate reality contains three connotations: 1) from an ontological perspective,

“ultimate reality” means the ultimate object which is understood as the true nature
of things; 2) from an epistemological perspective, ultimate reality is nonconceptual
wisdom, which refers to how the Buddhist sages view reality; 3) the teachings in
accord with non-arising is ultimate reality.20
This compact summary can be traced back to the TJ commentary breaking down Bhāviveka’s
first syllogism within the MHK. Watanabe provides the translation of the syllogism and of
Bhāviveka’s own commentary:
(proposition) from the standpoint of the highest truth ( paramārthataḥ) the earth, etc. do not
have the gross elements as their own-beings (intrinsic nature),
(reason 1) because they are things which are made, or
(reason 2) because they are things which have cause, etc., (instance) just like knowledge
( jñāna).
(MHK 3.26, Watanabe 1994, 18)
Commentary: In [the word] paramārtha, artha is what is to be obtained ( pratipattavya)

and what is to be understood (adhigantavya) because artha is the object to be known
( jnātavya). Paramārtha means21 “the most excellent.” The compound parama-artha
[can be interpreted in three ways].
(1) It means “the most excellent object” because it is the object and the most excellent
(karmadhārya compound)
(2) Or, it is the object of the most excellent. That is, because it is the object of the most
excellent knowledge that is beyond discrimination (nirvikalpa-jñāna), it means the
object of the most excellent (tatpuruṣa compound)
354
Bhāviveka
(3) Or, it is “comformable to paramārtha” ( paramārthānukūla). That is, since there

is that paramārtha in wisdom ( prajñā) which is conformable to obtaining of
paramārtha, it means “one which is in conformity with paramārthaḥ” (bahuvrīhi
compound).
Paramārthataḥ [in syllogisms] is taken as the third type of compound (the bahuvrīhi com-
pound), i.e., paramārthataḥ, in the ultimate reality.
(TJ; Watanabe 1994, 18–19)
These kinds of subtle definitions demonstrate what is meant when scholars say that Bhāviveka
brought forth a much more nuanced application of the distinction between the two truths than
that articulated by Nāgārjuna. Here, Bhāviveka begins his syllogism by stating the perspective
from which it is to be understood. This perspective is that of the highest truth ( paramārthataḥ),
in opposition to that of the relative (saṃvṛti) truth. The commentary explains that the highest
truth can be understood in three cumulative ways, which correspond to the summary given by
Hsu previously. The first two definitions refer to truth itself, which lies beyond concepts and thus
cannot be understood through any conventional means. However, when it comes to the highest
truth mentioned in a syllogism, it is to be understood as the third interpretation: that which is in
conformity with the highest truth, which is the wisdom ( prajñā) found in Buddhist teachings,
such as those of Bhāviveka. Thus, the third definition involves concepts and language, used in
such a way as to approach truth itself. It allows Bhāviveka to formulate a logical position on
topics of a conventional nature (the earth, etc.) by delimiting its perspective (that of a specific
understanding of the absolute) to avoid contradictions (the absolute, merely understood as being
nonconceptual, cannot be referred to in the analysis of concepts). Watanabe fleshes out the syl-
logism further, but, for our purpose, it suffices to notice that it has the characteristic of a three-
part inference (anumāna) with (1) a proposition (the earth, etc. do not have the gross elements
as their intrinsic nature), (2) a reason (reasons 1 and 2), and (3) an illustration (like knowledge).
This last part, the illustration, can help us shed more light on Bhāviveka’s understanding of the
function of logic in matters of liberation. To one who might ask in what manner knowledge is
( jñāna) a produced thing, like the earth and so on, Bhāviveka’s auto-commentary replies: “it is
accomplished by logic (tarka) which aims at heaven (svarga) and nirvana (apavarga)” (Watan-
abe 1994, 24). Clearly, Bhāviveka gives a soteriological function to logic (tarka). That function
must be explained in relation to Bhāviveka’s understanding of ultimate reality.
By introducing these three nuances within his definition of the highest reality, Bhāviveka allows
for a twofold nature of ultimate reality, one devoid of volitional concepts (anabhisaṃskārapravṛtti)
and another which involves them (sābhisaṃskārapravṛtti) (Watanabe 1994, 23). The third defi-
nition of the ultimate truth allows conventional means such as logic (tarka) within its scope,
whereas the other two definitions forbid such means of access to truth. This volitional aspect
of wisdom ( prajñā) can now act as a lever to pass from the conceptual to the nonconceptual,
through the door of the Buddhist teachings. To understand how this works, this volitional wis-
dom can be further divided into three stages. The first is obtained from hearing teachings and is
hence called the ‘wisdom acquired from hearing’ (śrutamayīprajñā). At this level, conceptual-
ity and language remain throughout. Then, through a critical analysis of scriptures, which have
been ‘heard,’ a second wisdom is produced, intellectual in nature, called the ‘wisdom acquired
through reflecting’ (cintāmayīprajñā) on the teachings, which could be described as an internal
and critical dialogue with the teachings, an analytical contemplation of their meaning by means
of logic. At this point, conceptuality remains but becomes more abstract. This second and subtler
355
conceptual activity is meant to support a third and last layer of wisdom, a nonconceptual one this
time, a direct experience of contemplative insight called the ‘wisdom born out of cultivation’
(bhāvanamayīprajñā), often rendered as meditative wisdom. Thus, the divide between the con-
ceptual and the nonconceptual is said to be progressively crossed over by a bridge whose entry
is rooted in firm concepts (heard – śruta – from scriptures) and that ends in pure nonconceptual
perception, tied at its center by the ever-subtler junction ( yoga) of reasoning and analysis.
This theoretical framework, outlining the pivotal role of reasoning and hermeneutics in
Mahāyāna religious praxis, was already enunciated in the well-known third-century Treatise
on the Foundation for Yoga Practitioners (Yogācārabhūmiśāstra).22 Bhāviveka’s account of
ultimate truth ( paramārtha) is informed by this framework, which strengthens the soteriologi-
cal resonance of his Dignāga-inspired practice of inference. This gradual threefold mode of
insight into the nature of reality has also been asserted by other prominent Buddhist think-
ers, such as Dignāga, Dharmakīrti,23 and Kamalaśīla. In order to cultivate the non-view, they
argued that worldly conventional views had to be removed, like cataracts from the eye, with
the appropriate critical antidote. This is precisely the task of the MHK.
In other words, the existential disease and its remedy are found within the conventional
realm. It is through conceptual analysis, directed by logic (tarka), that one can finally move
beyond conceptuality altogether. The conventional alone – through language, logic, and so on –
allows the meaning of scriptures to indicate the ultimate, which, in and by itself, has no indicator
and nothing to indicate. Scriptures are composed according to conventions and must thus be
understood through conventional means (inference). The wisdom born from reasoning is thus a
crucial step on the stairway to the conceptual palace of liberation, a metaphor dear to Bhāviveka:
The nectar of wisdom, bestowing satisfaction, is a lamp of unobstructed light, the

stairway to the palace of liberation, the oblation-fire which burns afflictions.
(MHK 3.6, Bouthillette 2020, 37)
Notes
1 Although not exhaustive, the following list covers the most important scholarly contributions to date,
in relation to Bhāviveka and the MHK: Ames 2019; Bahulkar 1994; Eckel 1992, 2008; Ejima 1980b,
1980a; Eltschinger 1998; Gokhale 1958, 1972; Ham 2016; He 2014; Heitmann 1995, 2004; Hira-
bayashi and Iida 1978; Honda 1967; Hoornaert 1999, 2001, 2003, 2004; Hsu 2011, 2013; Iida 1966,
1980; Jones 2011; Kawasaki 1973, 1976; Krasser 2011, 2012; La Vallée Poussin 1932–1933, 1933;
Lindtner 1982, 1986, 1995, 2001a, 2001b; Miyasaka 1954a, 1954b, 1958; Nakada 1972; Nakamura
1950, 1955, 1958, 1968; Potter 2003; Qvarnström 1989, 1999, 2012, 2015; Saitō 2005; Ruegg 1990;
Warder 1970; Watanabe 1994, 1998, 2012; Yamaguchi 1941 [1964].
2 One can find: Bhavya, Bhavyaka, Bhavyakāra, Bhavyaviveka, Bhavyavivikta, Bhavyaviveka,
Bhāvivika, Bhāvivikta, Bhagavadviveka, and Bhavavadviveka. Shotaro Iida presents a table asso-
ciating the various names with their textual reference Iida 1980, 5. On Bhāviveka’s name, see Ejima
1980a; Hsu 2013, 10–12.
3 By ‘doxography,’ I understand a text recognizable through the conjunction of six main features.
A doxographical text is (1) either a whole text, or a part of a text; (2) where competing views of phi-
losophers or schools are presented following a division of topics organized into sets and sub-sets with
specific differences to which a name-label is attached in most cases; (3) where the original argument
supporting such views may or may not be given; (4) where the author’s own view and arguments may
or may not be criticized; (5) where the philosophical content consists either in literal or in non-literal
renderings of sources; and (6) where the overall concern is primarily systematic, dialectic, and with
little or no historiographical character. See Bouthillette 2020, 13–14.
4 For translations and commentaries on the PP, see Ames 1986, 1993, 1999, 2000, 2019.
5 For more information on the two other texts of Bhāviveka, see Hsu 2013, 40–43.
6 Ronald Davidson (2002) was the first to fully theorize such ideas in relation to Buddhist esotericism.
356
Bhāviveka
7 Jonardon Ganeri also made a similar argument in the context of the scholastic tradition; see Ganeri 2001.
8 Verardi’s work on the downfall of Buddhism in India gives ample illustrations of the many kinds of
physical tortures and death penalties allegedly befalling those losing public debates in medieval India.
9 For a general discussion of the issue, see Hsu 2013, 9. See also Eltschinger 2014 on the engagement
of Buddhist scholars with non-Buddhists.
10 Shoryu Katsura argues that Nāgārjuna succeeded in formulating the formal proof of reductio ad
absurdum ( prasaṅga) probably for the first time in India (Katsura 2007, 70). The prasaṅga argumen-
tative method basically rests on pointing out the ‘consequence’ ( prasaṅga) that results from any posi-
tion or thesis operating with the idea of the real existence of an entity (bhāva). A general introduction
to the topic can be found in Ruegg 1981, 36.
11 I discuss the topic of doxography, Indian and Classical, at length in Bouthillette 2020.
12 For more details on the MHK manuscript, see Iida 1980, 52–53 and, more recently, Watanabe 1994,
14–15. Concerning the Rañjanā script, see Hartmann 1998.
13 Thus, I disagree with Nicholson’s remarks to the effect that, within Bhāviveka’s MHK, “there is no
apparent order in the sequence of doctrines he refutes” (Nicholson 2010, 152).
14 The translation is from Eckel 2008, 105.
15 The brackets here are my own insertion to facilitate the reading.
16 For more information on this chapter, including a critical evaluation, see He 2011.
17 On the vedāntatattvaviniścaya chapter of the MHK, see Qvarnström 2015.
18 Similarly, Eltschinger observed that “according to Sthiramati, the yukti/hetu-vidyā enterprise was
expected to shape arguments in order (1) to found and support key Buddhist doctrines, (2) to defeat
the non-Buddhist intellectuals’ hostility towards Buddhism; (3) to convert them to the Buddhist sad-
dharma or śāsana, (4) strengthen the coreligionists’ adherence to Buddhism” (Eltschinger 2010, 562).
19 In reference to Nāgārjuna’s MMK, XXIV.9–10: “Those who do not understand the distinction
between these two realities do not realize the profound truth embodied in the Buddha’s doctrine.
Without relying on conventional [reality], an ultimate [reality] cannot be taught. Without understand-
ing ultimate reality, nirvāṇa is not realized.” The translation is from La Vallée Poussin 1903–13, 494.
20 See Hsu 2013, ii–iii, 61, 143–45.
21 I am correcting here what appears to be a syntactic mistake in Watanabe’s translation. His origi-
nal line reads as follow: “Paramārtha which means ‘the most excellent’.” Since the reading is odd,
I removed the “which.”
22 Its first appearances are noticed as early as the Saṅgītisutta (Dīghanikāya III 219) and the Vibhaṅga
(324).
23 On cintāmayīprajñā in Dharmakīrti, see Eltschinger 2010.
Bibliography
Ames, William Longstreet. 1986. “Bhāvaviveka’s Prajñāpradīpa: Six Chapters.” PhD diss., Department
of Asian Languages and Literature, University of Washington.
———. 1993. “Bhāvaviveka’s Prajñāpradīpa. A Translation of Chapter One: ‘Examination of Causal
Conditions’ (Pratyaya).” Journal of Indian Philosophy 21: 209–59.
———. 1999. “Bhāvaviveka’s Prajñāpradīpa: A Translation of Chapters Three, Four, and Five, Examin-
ing the āyatanas, Aggregates, and Elements.” Journal of Buddhist Literature 1: 1–119.
———. 2000. “Bhāvaviveka’s Prajñāpradīpa: A Translation of Chapter Six, Examination of Desire and
the One Who Desires, and Chapter Seven, Examination of Origination, Duration, and Cessation.”
Journal of Buddhist Literature 2: 1–91.
———. 2019. The Lamp of Discernment: A Translation of Chapters 1–12 of Bhāvaviveka’s Prajñāpradīpa.
Moraga: Numata Center for Buddhist Translation & Research.
Bahulkar, S. S. 1994. “The Madhyamaka-Hṛdaya-Kārikā of Bhāvaviveka: A Photographic Reproduction of
Prof. V. V. Gokhale’s Copy.” Nagoya Studies in Indian Culture and Buddhism Saṃbhāṣā 15 (1–4): 1–49.
Bouthillette, Karl-Stéphan. 2020. Dialogue and Doxography in Indian Philosophy: Points of View in
Buddhist, Jaina, and Advaita Vedānta Traditions. Dialogues in South Asian Traditions: Religion,
Philosophy, Literature and History. London and New York: Routledge.
Burnouf, Eugène. 1844. Introduction à l’histoire du Bouddhisme Indien. 1876 Reprint ed. Paris: Imprim-
erie Royale.
Davidson, Ronald M. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New
York: Columbia University Press.
357
Eckel, Malcolm David. 1992. To See the Buddha: A Philosopher’s Quest for the Meaning of Emptiness.
Princeton: Princeton University Press.
———. 2008. Bhāviveka and His Buddhist Opponents. Harvard Oriental Series. Cambridge, MA: Har-
vard University Press.
Eckel, Malcolm David, and Vincent Eltschinger. 2019. “Bhāviveka.” In Brill’s Encyclopedia of Bud-
dhism: Volume Two, Lives, edited by Jonathan A. Silk, Richard Bowring, Vincent Eltschinger, and
Michael Radich, 81–84. Leiden: Brill.
Ejima, Yasunori 江島恵教. 1980a. “Bhāvaviveka/Bhavya/Bhāviveka [in Japanese].” Indogaku
Bukkyōgaku Kenkyū 印度學佛教學研究 38: 98–106.
———. 1980b. Chūgan shisō no tenkai – Bhāvaviveka kenkyū 中観思想の展開： Bhāvaviveka 研究
[An Exposition of Madhyamaka Thought. A Study of Bhāvaviveka]. Tokyo: Shunjūsha 春秋社.
Eltschinger, Vincent. 1998. “Bhāvaviveka et Dharmakīrti sur āgama et contre la Mīmāṁsā.” Asiatische
Studien/Études Asiatiques 52: 57–84.
———. 2010. “Studies in Dharmakīrti’s Religious Philosophy: The Cintā-mayī-prajñā.” In Logic and
Belief in Indian Philosophy, edited by Piotr Balcerowicz, 553–91. New Delhi: Motilal Banarsidass.
———. 2014. Buddhist Epistemology as Apologetics: Studies on the History, Self-Understanding and
Dogmatic Foundations of Late Indian Buddhist Philosophy. Beiträge Zur Kultur Und Geistesge-
schichte Asiens. Wien: Ôsterreichisch Akademie Der Wissenschaften.
Ganeri, Jonardon. 2001. Philosophy in Classical India: Proper Work of Reason. London and New York:
Routledge.
Gokhale, V. V. 1958. “The Vedānta-Philosophy Described by Bhavya in the Madhyamakahṛdaya.” Indo-
Iranian Journal 2: 165–80.
———. 1972. “The Second Chapter of Bhavya’s Madhyamakahṛdaya.” Indo-Iranian Journal of Indian
Philosophy 14: 40–45.
Gokhale, V. V., and S. S. Bahulkar. 1985. “Madhyamakahṛdayakārikā Tarkajvālā, Chapter I.” In Miscel-
lanea Buddhica, edited by Chr. Lindtner, 76–108. Copenhagen: Akademisk Forlag.
Halbfass, Wilhelm. 1988. India and Europe: An Essay in Understanding. New York: State University of
New York Press.
Ham, Hyoung Seok. 2016. “Buddhist Critiques of the Veda and Vedic Sacrifice: A Study of Bhāviveka’s
Mīmāṃsā Chapter of the Madhyamakahṛdayakārikā and Tarkajvālā.” PhD diss., Asian Languages
and Cultures, University of Michigan.
Hartmann, Jens-Uwe. 1998. “The Ranjana Script, Translated from German by Prof. Dr. Ursula Heise.”
The Fifth Seal Calligraphic-Icons/Kalligraphikons, The Radheshyam Saraf Art Collection at Hotel
Yak & Yeti, Kathmandu, Nepal.
He, Huanhuan. 2011. “Bhavya’s Critique of the Vaiśeṣika Theory of Liberation in the Tarkajvālā.” Stud-
ies in Indian Philosophy and Buddhism (インド哲学仏教学研究) 18: 23–38.
———. 2014. “Further Notes on Bhāviveka’s Principal Oeuvre.” Indo-Iranian Journal 57: 299–352.
Heitmann, Annette L. 1995. “Momentane (ekakṣaṇa) Gnosis ( jñāna) im Sinne der Madhyamakahṛdaya
kārikā und der in tibetischer Übersetzung vorliegenden Tarkajvālā I – IV.” Asiatische Studien/Études
Asiatiques 49: 391–427.
———. 2004. Nektar der Erkenntnis: Buddhistische Philosophie des 6. Jahrhunderts: Bhavyas Tarka-
jvala I-III. 26. Aachen: Shaker Verlag.
Hirabayashi, Jay, and Shotaro Iida. 1978. “Another Look at the Mādhyamika vs. Yogācāra Controversy
Concerning Existence and Non-Existence.” Edward Conze Festschrift (Berkeley Buddhist Studies
Series) 1: 341–60.
Honda, Megumu. 1967. “Sāṃkhya Philosophy Described by His Opponent Bhavya.” Journal of Indian
and Buddhist Studies 16 (1): 33–38.
Hoornaert, Paul. 1999. “An Annotated Translation of Madhyamakahṛdayakārikā/Tarkajvālā V.1–7.”
Kanazawa Daigaku Bungakubu Ronshū Kōdō Kagaku. Tetsugaku hen 金沢大学文学部論集行動科
学・哲学篇沢大学文学部論集行動科学・哲学篇 [Bulletin of the Faculty of Letters of Kanazawa
University, Behavioral Sciences and Philosophy] 19: 127–59.
———. 2001. “An Annotated Translation of Madhyamakahṛdayakārikā/Tarka-jvālā V.27–54.” Kanaz-
awa Daigaku Bungakubu Ronshū Kōdō Kagaku. Tetsugaku hen 金沢大学文学部論集行動科学・
哲学篇 21: 149–90.
———. 2003. “An Annotated Translation of Madhyamakahṛdayakārikā/Tarka-jvālā V.85–114.” Kanaz-
awa Daigaku Bungakubu Ronshū Kōdō Kagaku. Tetsugaku hen 金沢大学文学部論集行動科学・
哲学篇 23: 139–70.
358
Bhāviveka
———. 2004. “The Dharmapāla-Bhāvaviveka Debate as Presented in Dharmapāla’s Commentary to

Catuḥśataka XVI.23.” Kanazawa Daigaku Bungakubu Ronshū Kōdō Kagaku. Tetsugaku hen 金沢大
学文学部論集行動科学・哲学篇 24: 119–49.
Hsu, Chien Yuan. 2011. “Bhāvaviveka’s Syllogism as an Initial Step to Enlightenment (正觀第五十九
期/二Ο一一年十二月二十五日).” Satyābhisamaya: A Buddhist Studies Quarterly 59: 57–92.
———. 2013. “Bhāviveka’s Jewel in the Hand Treatise: Elucidating a Path to Awakening Utilizing For-
mal Inference.” PhD diss., Department of Religious Studies, University of Calgary.
Iida, Shotaro. 1966. “Āgama (Scripture) and Yukti (Reason) in Bhāvaviveka.” In Kanakura Kinenronbun-
shû (Kanakura Festschrift), edited by Kanakura Hakushi Koki Kinen Ronbunshū Kankōkai, Enshō
Kanakura, and Yukio Sakamoto, 77–96. Kyoto: Heirakuji Shoten.
———. 1980. Reason and Emptiness: A Study in Logic and Mysticism. Tokyo: Hokuseido Press.
Jones, Richard H. 2011. Indian Madhyamaka Buddhist Philosophy After Nagarjuna, Volume I: Plain
English Translations and Summaries of the Essential works of Aryadeva, Rahulabhadra, Buddha-
palita, and Bhaviveka. New York: Jackson Square Books.
Kajiyama, Yuichi. 1968–1969. “Bhāviveka, Sthiramati, and Dharmapāla,” WZKSO 12–13: 193–203.
Katsura, Shoryu. 2007. “How Did the Buddhists Prove Something? The Nature of Buddhist Logic.”
Pacific World 3 (9): 63–84.
Kawasaki, Shinjo. 1973. “Bhavyano tsutaeru Mīmāṃsā shisô (The Mīmāṃsā thought transmitted by
Bhavya).” In Indo Shisô to Bukkyô (Nakamura Festschrift), 71–86. Tokyo: Shunjûsha.
———. 1976. “The Mīmāṃsā Chapter of Bhavya’s MHK – Text and Translation (1) Pūrva-pakṣa.” Stud-
ies 1976, Institute of Philosophy, The University of Tsukuba: 1–15.
Krasser, K. 2011. “How to Teach a Buddhist Monk to Refute the Outsiders – Text Critical Remarks on
Some Works by Bhāviveka.” Dhīḥ 51: 49–76.
———. 2012. “Bhāviveka, Dharmakīrti and Kumārila.” In Devadattīyam. Johannes Bronkhorst Felicita-
tion Volume, edited by F. Voegeli et al., 535–94. Bern: Peter Lang.
La Vallée Poussin, Louis de, trans. 1903–1913. Mūlamadhyamakakarikas de Nagarjuna avec la
Prasannapadā Commentaire de Candrakīrti. Bibliotheca Buddhica, vol. 4. St. Petersburg: l’Académie
Impériale des Sciences.
———. 1932–1933. “Madhyamaka, L’auteur du Joyau dans la main; 3, traduction du Joyau dans la
main.” MCR 2: 60–138.
———. 1933. “Madhyamaka.” Mélanges Chinois et Bouddhiques II.
Lindtner, Chr. 1982. “Adversaria Buddhica.” Wiener Zeitschrift für die Kunde Südasiens 26: 167–94.
———. 1986. “Bhavya, the Logician.” Adyar Library Bulletin 50 (Golden Jubilee Volume): 58–84.
———. 1995. “Bhavya’s Madhyamakahṛdaya (Pariccheda Five) Yogācāratattvaviniścayāvatāra.” Adyar
Library Bulletin 59 (C. KunhanRaja Birth Centenary Volume): 37–65.
———. 2001a. Bhavya on Mīmāṃsā, Mīmāṃsātattvanirṇayāvatāraḥ, with English Translation by
Christian Lindtner. Adyar: The Adyar Library and Research Centre.
———. 2001b. Madhyamakahrdayam of Bhavya. Adyar: The Adyar Library and Research Center.
Miyasaka, Yûshô. 1954a. “Bhavya to Vaiśeṣika gatusetsu (Bhavya and the Vaiśeṣika doctrines).” Shûkyô
Kenkyû 133: 173–77.
———. 1954b. “Shôben inyô no Vaiśeṣika tetsugakusetsu (The Vaiśeṣika philosophy quoted by
Bhavya).” Bunka 18 (3): 24–40.
———. 1958. “Ronri no honoho ni okeru Vaiśeṣika Tesugaku (Vaiśeṣika philosophy in the Tarkajvālā).”
Kôyasan Daigaku Ronsô 1 (1): 1–37.
Nakada, Naomichi. 1972. “The Sanskrit Text of the MHK and the Tibetan text of the MHK-vṛtti Tarkajvālā
Ṣaṣṭaḥ-paricchedaḥ/Sāṃkhya-tattvāvatāraḥ // (Part 1. Pūrvapakṣa).” Tsurumi Jyoshidaigakubu Kiyô
6: 145–46.
Nakamura, Hajime. 1950. “Shoki no Vedānta Tetsu-gaku (The Early Vedānta Philosophy).” Tôkyô:
238–332.
———. 1955. “Upaniṣadic Tradition and the Early School of Vedānta, as Noticed in Buddhist Scripture.”
Harvard Journal of Asiatic Studies 18 (1–2): 74–104.
———. 1958. “The Tibetan Text of the Madhyamaka-hṛdaya-vṛtti-tarkajvālā.” Indo-Iranian Journal 2:
181–90.
———. 1968. “Shoki Vedānta-tetsugaku kenkyû no hatten (The Progress Report on the Study of the
Early Vedānta.” Journal of Indian and Buddhist Studies 16 (2): 10–22.
Nicholson, Andrew J. 2010. Unifying Hinduism: Philosophy and Identity in Indian Intellectual History.
New York: Columbia University Press.
359
Potter, Karl H., ed. 2003. Buddhist Philosophy from 350 to 600 A.D. Encyclopedia of Indian Philoso-
phies, vol. 9. New Delhi: Motilal Banarsidas Publishers.
Qvarnström, Olle. 1988. “Space and Substance. A Theme in Madhyamaka-Vedānta Polemics.” Studies in
Central and East Asian Religions 1: 3–34.
———. 1989. Hindu Philosophy in Buddhist Perspective. The Vedantatattvavinishcaya Chapter
of Bhavya’s Madhyamakah. 4 vols. Lund Studies in African and Asian Religions, vol. 4. Lund:
Almqvist & Wiksell.
———. 1999. “Haribhadra and the Beginnings of Doxography in India.” In Approaches to Jaina Stud-
ies: Philosophy, Logic, Rituals and Symbols, edited by Narendra K. Wagle and Olle Qvarnström,
169–210. Toronto: Center for South Asian Studies.
———. 2012. “Sāṃkhya as Portrayed by Bhāviveka and Haribhadrasūri.” Journal of Indian Philosophy
40 (4): 395–409.
———. 2015. Bhāviveka on Sāṃkhya and Vedānta, The Sāṃkhya and Vedānta Chapters of the
Madhyamakahṛdayakārikā and Tarkajvālā. Harvard Oriental Series, vol. 78. Cambridge, MA: Har-
vard University Press.
Ruegg, David Seyfort. 1981. The Literature of the Madhyamaka School of Philosophy in India. Wies-
baden: Otto Harrassowitz.
———. 1990. “On the Authorship of Some Works Ascribed to Bhāvaviveka/Bhavya.” In Earliest Bud-
dhism and Madhyamaka, edited by David Seyfort Ruegg and Lambert Schmithausen, 59–71. Leiden:
E. J. Brill.
Saitō, Akira 斎藤明. 2005. “Chūgan shinron’ no shōmei to sono seiritsu o meguru shōmondai
『中観心論』の書名とその成立をめぐる諸問題 [Some Problems with the Title of the
‘Madhyamakahṛdayaśāstra’ and Its Formation].” Indogaku Bukkyōgaku Kenkyū 印度學佛教學研究
53: 832–38.
Verardi, Giovanni. 2011. Hardships and Downfall of Buddhism in India. Singapore: ISEAS Publishing.
Warder, Anthony Kennedy. 1970. Indian Buddhism. 3rd Revised ed. New Delhi: Motilal Banarsidas.
Watanabe, Chikafumi. 1994. “Bhā-viveka (A.D. c. 490–570)’s Madhyamaka-hṛdaya-kārikā,
Tattvajñānaiṣanā, verses 137–266, An English Translation and Explanation.” Master of Arts, Depart-
ment of Asian Studies, University of British Columbia.
———. 1998. “A Translation of the Madhyamakahrdayakarikd with the Tarkajvala III. 137–146J.” Jour-
nal of the International Association of Buddhist Studies 21 (1): 125–55.
———. 2012. “Madhyamakahṛdayakārikā III. 147–158.” In Saṁskṛta-sādhutā: Goodness of Sanskrit.
Studies in Honour of Professor Ashok N. Aklujkar, edited by Chikafumi Watanabe, Michele Des-
marais and Yoshichika Honda, 545–51. New Delhi: D. K. Printworld.
Westerhoff, Jan. 2014. “Madhyamaka: Conventional Categories in Madhyamaka Philosophy.” In Cat-
egorisation in Indian Philosophy: Thinking Inside the Box, edited by Jessica Frazier, 115–31. Surrey
and Burlington: Ashgate.
Yamaguchi, Susumu. 1941 [1964]. Bukkhyô ni okeru Mu to U no Tairon. (Buddhist controversies con-
cerning existence and non-existence). Tokyo: Sankibô.
360
21
DHARMAPĀLA
A Janus-Faced Interpreter of Yogācāra?1
Ching Keng
Dharmapāla (active in the sixth century)2 was an important Yogācāra scholar whose work exerted
a significant influence on the development of Buddhist philosophy in India and in East Asia.
Dharmapāla was active between two productive periods of Indian Buddhist thought: the time of
Vasubandhu (c. 400–480) and Dignāga (c. 480–540) and the time of Dharmakīrti (c. 600–660).
Thus, his works provide insight into Indian Buddhist philosophy during the transitional period
of the sixth century. According to the Chinese tradition, there were quite a few interpreters of
Yogācāra after Vasubandhu, but most of their works are no longer extant. Dharmapāla is an
exception, as are Dignāga and Sthiramati (sixth century). Dharmapāla is also a major figure of
Indian Buddhism given the lacuna of logico-epistemological ( pramāṇa) Buddhist texts from the
period between Dignāga and Dharmakīrti. Dharmapāla was among the earliest Buddhist schol-
ars to employ Dignāga’s logic (hetuvidyā) in his writings, including applying Dignāga’s logic to
the interpretation of root texts as well as to debates with opponents. In addition, Dharmapāla is
famous for disagreeing with a certain Mahāyāna master who held a very different interpretation
of emptiness (śūnyatā). According to the Chinese interpretation, this opponent was the Madh
yamaka thinker Bhāviveka (sixth century). In this regard, Dharmapāla and Bhāviveka were
probably the first participants in the debates between Madhyamaka and Yogācāra, which have
marked so profoundly the shape of Indian Buddhist philosophy and its interpretations in Central
and East Asia. Finally, Dharmapāla left a significant impact on the Yogācāra school in East Asia,
the so-called School of Dharma-characteristics, known as the Faxiang School (faxiang zong 法
相宗) in China and Hossō School (hossō shū 法相宗) in Japan, which is centered around the
Yogācāra teachings of A Treatise for the Establishment of Consciousness-Only (Cheng weishi
lun 成唯識論; T1585, henceforth abbreviated as CWSL), compiled by the Chinese pilgrim and
erudite scholar monk Xuanzang (玄奘; 602? – 664).
Despite his importance, Dharmapāla remains insufficiently studied. This is partly because
all the works attributed directly or indirectly to him are preserved only in Chinese translations.
Moreover, these texts are extremely challenging to interpret even for scholars with expertise
in classical Chinese Buddhist texts. None of his works survive in Sanskrit or Tibetan. This
casts doubts on whether he was as influential as claimed by his Chinese interpreters. Another
hindrance to our study of Dharmapāla is that he seems to have two very different identities.
In the eyes of the Faxiang School, Dharmapāla was an original and creative thinker. He holds
distinctive opinions about almost every important issue, according to the commentary on the
DOI: 10.4324/9781351030908-32 361

Ching Keng
CWSL by Master Ji (基法師, also known as Kuiji 窺基; 632–682), Xuanzang’s famous disci-
ple. The significance of his positions is manifest by the fact that Master Ji always treats them as
definitive and authoritative. As shown in the following, philosophically original and insightful
doctrines were attributed to Dharmapāla, such as the strategy for resolving the tension regard-
ing the issue of meta-cognition (svasaṃvitti or svasaṃvedana). This provided a template for
defending Yogācāra idealism philosophically, an important aspect that is largely missing in
post-Dignāga Buddhist philosophy and also largely neglected in contemporary philosophical
discourse. However, judging from the works reliably attributed to him, Dharmapāla actually
appears to be a rather conservative interpreter. He does not seem to offer particularly original
interpretations or creative doctrines beyond what the root texts seem to already mean.
The fact that Dharmapāla has two different identities casts doubt on whether the doctrines
attributed to him actually originated with him, which then casts further doubt on the historical
image of Dharmapāla constructed by the Faxiang School. Toward the end of this chapter, I will
suggest a tentative way to deal with this question of authorship. In addition, I aim to highlight
the distinctive features of Dharmapāla’s works and their potential contributions to encourage
more research on this insufficiently studied figure.
Dharmapāla’s Life and Works

As with almost all figures in Indian Buddhism, we know very little about Dharmapāla’s life.3
The only biography we have is preserved in Chinese sources.4 It indicates that Dharmapāla
came from the city of Jianzhibuluo 建志補羅, the capital of Daluopitu 達羅毘荼 (*Draviḍa?)
in southern India. He was the son of a high-ranking minister and became a monk to avoid hav-
ing to marry the king’s daughter. The biography stresses his mastery of Buddhist doctrine and
that he died very young. The historical accuracy of this biography is impossible to confirm.
Given that Dharmapāla adopted the system of logic devised by Dignāga (c. 480–540) and was
referred to by Vinītadeva5 (690–750, see Funayama 2001), we can infer that he must have
lived sometime in the late sixth or seventh century.
According to Xuanzang’s disciples, Dharmapāla composed four texts, one of which is a
lengthy work on grammar that has long been lost (Tillemans 2008, 11–12). The other three,
preserved only in Chinese translations, are all commentaries. In my discussion in the follow-
ing, when I refer to “works reliably attributed to Dharmapāla,” I am referring to these three
commentaries, as follows:
1. A Commentary on the Four-Hundred Verses (Dasheng guang bailun shilun 大乘廣百論

釋論; T1571, 10 fascicles; hereafter Four-Hundred Commentary): Dharmapāla’s com-
mentary on the Four-Hundred Verses (Catuḥśataka) of Āryadeva (c. third century).
2. Jewel-Arising Treatise on the Establishment of Consciousness-Only (Cheng weishi
baosheng lun 成唯識寶生論; T1591, 5 fascicles; hereafter Twenty Verses Commentary):
Dharmapāla’s commentary on Vasubandhu’s Twenty Verses (Viṃśikā).
3. A Commentary on the Investigation of Cognitive Objects (Guan suoyuan lun shi 觀所緣
論釋; T1625, 1 fascicle; hereafter Investigation Commentary): Dharmapāla’s commen-
tary on Dignāga’s Investigation of Cognitive Objects (Ālambanaparīkṣā).
The Four-Hundred Commentary was translated by Xuanzang. Compared to the Twenty Verses
Commentary and Investigation Commentary, which were both translated by Yijing (義淨;
635–713), it is relatively accessible. Moreover, both the Twenty Verses Commentary and the
362
Dharmapāla
Investigation Commentary are incomplete. Verses 11–15 (the refutation of the Realists’ theory
of atoms [ paramāṇu]) and verse 22 (the closing verse) in the Twenty Verses Commentary are
missing, as well as verses 7–8 (the last two verses) in the Investigation Commentary.6
In addition to Dharmapāla’s works mentioned previously, the CWSL is a collection of
commentaries on Vasubandhu’s Thirty Verses (Triṃśikā) compiled by Xuanzang, in which
we find references to doctrines professed by Dignāga, Sthiramati, and Dharmapāla, among
others.7 But doubts have been raised: both Lusthaus (2002, Chapter 15 in particular) and
Sakuma (2006 [2008]) have challenged Master Ji’s attribution of certain doctrines to particu-
lar Indian masters. These doubts are magnified by the fact that in the works reliably attributed
to Dharmapāla, he appears to be a rather conservative commentator and interpreter, while the
positions Master Ji attributes to him portray him as quite creative and bold. This has led some
scholars to suspect that Xuanzang and Master Ji in fact crafted their own ideas in the name of
Dharmapāla. As I shall discuss in the following, I agree that there are legitimate doubts about
whether any of the doctrines cited by the CWSL represent the actual teachings of Dharmapāla.8
Features of Dharmapāla’s Works

A common method employed by Dharmapāla and other Indian authors in their commentar-
ies is to lay out carefully the content of an argument and explain why the author of the root
text (that is, the work being commented upon) phrases his argument in the way that he does.
Dharmapāla shares with other Indian commentators a widespread principle of Buddhist and
Indian scholastic exegesis: every word in the root texts counts, and not a single word is unim-
portant or redundant.9
Another feature that Dharmapāla shares with other Indian commentators is that he is rather
conservative in his commentaries on the root texts. Often, he does not go much beyond the
philosophical arguments made in the original text. In his Twenty Verses Commentary, for
instance, Dharmapāla simply repeats what has already been said in Vasubandhu’s Twenty
Verses, with some re-arrangement of the arguments for the sake of refuting different opponents
(see following discussion).
A distinctive feature that really sets Dharmapāla apart from other commentators is that he
often tries to make sense of Yogācāra texts by formulating the argument in terms of Dignāga’s
logic. A clear example can be seen in Dharmapāla’s reading of the first verse of the Twenty
Verses Commentary,10 which reads:
This world is consciousness-only (vijñaptimātra)

Because of the appearance (avabhāsana) of non-existent external objects
As when someone with a timira disease sees non-existent nets of hair and the like.
(T31.1591.79c12–14)11
Dharmapāla takes this statement to be a well-formulated inference (anumāna) proposed by

Vasubandhu in order to establish the thesis ( pratijñā) of the entire Twenty Verses. Dignāga’s
logic requires that a well-formulated inference consist of three parts: a thesis, which itself
consists of a subject (dharmin) and a predicate (dharma); a reason (hetu); and an example
(dṛṣṭānta).12 According to Dharmapāla, Vasubandhu’s inference can be formulated as follows:
Thesis: cognitions that have matter (rūpa), etc., as their cognitive object (ālmabana)
do not cognize actual external objects;
363
Ching Keng
Reason: because the cognitions merely have the appearance (avabhāsana) of non-
existent objects;
Example: like a person with a timira disease who sees a net of hair.
In the following, I discuss why Dharmapāla feels the need to construe this verse as establish-
ing an inference in Dignāga’s sense. Dignāga, except in the Investigation of Cognitive Objects,
rarely employs his own logical paradigm and epistemology to prove a specific Yogācāra doc-
trine. Dharmapāla, on the other hand, is one of the first Buddhist philosophers to apply Dignāga’s
logic extensively to interpret Yogācāra texts. Similarly, his younger contemporary, Bhāviveka,
uses Dignāga’s logical system in his interpretations of Madhyamaka texts. In fact, this is a
movement that can be observed in the period following Dignāga, in which Buddhist commen-
tators started to use the logical tools Dignāga elaborated in the service of their own particular
positions. It effectively amounted to a marriage of epistemology ( pramāṇa) and Yogācāra, in
the case of Dharmapāla, and of epistemology and Madhyamaka, in the case of Bhāviveka.13
What We Can Learn From Dharmapāla’s Works

Although Dharmapāla’s commentaries may not be very original, much can be learned about
the evolution of Indian Buddhist philosophy from them. Here, I highlight three areas in which
Dharmapāla’s commentaries illuminate significant philosophical developments in his era.
Post-Vasubandhu Developments: How to Win a Debate

First, Dharmapāla’s commentaries illustrate the ways that views about how to win a debate
changed in the period between Vasubandhu and Dharmapāla. As mentioned earlier, Dharmapāla
maintains that Vasubandhu’s Twenty Verses expresses a thesis in the first verse. In contrast,
Kellner and Taber (2014) insist that in the Twenty Verses, Vasubandhu simply refutes the objec-
tions of his opponents without providing a positive thesis. Namely, having refuted the validity
of all realist objections (including those of both the Nyāya-Vaiśeṣika and the Sarvāstivāda),
Vasubandhu implicitly resorts to an “argument from ignorance” in order to reach the conclusion
that things external to mind do not exist. That is to say, he first demonstrates that arguments
establishing the existence of external objects fail. From this absence of sound argument, he then
concludes that there can be no rational basis for believing in the existence of external objects.
Dharmapāla, by contrast, maintains that Vasubandhu lays out a positive thesis beyond this
“argument from ignorance.” According to Hu (2018, 120–21), Dharmapāla’s main reason for
insisting that Vasubandhu has a thesis is that he sees a simple refutation of the opponents’
argument as inadequate for establishing one’s own position. Rather, both the refutation of the
opponent’s thesis and the establishment of one’s own thesis are necessary. In the Investigation
Commentary, Dharmapāla says, “It is not the case that one’s own thesis can be established
simply by pointing out the fallacy in the opponent’s thesis” (T31.1625.889c27–28). And in
the Four-Hundred Commentary, he says that one must equally establish one’s own thesis and
refute the opponent’s thesis (T30.1571.243b8–11). In the Investigation of Cognitive Objects,
Dignāga at least implicitly thinks both are necessary,14 as does Vinītadeva in his commentary
on the Investigation of Cognitive Objects.15
One of Dharmapāla’s reasons for thinking that both establishment and refutation are
necessary is that pointing out an opponent’s fallacy is insufficient: while it demonstrates that
the opponent’s thesis is not necessarily right, it does not prove that it is definitely wrong.16 For
this reason, establishing one’s own thesis via a three-partite argument is also necessary for
364
Dharmapāla
proving that one’s thesis is right and hence the opponent’s thesis is wrong (see the Investiga-
tion Commentary, T31.1625.890a3–5).
If Kellner and Taber’s reading of the Twenty Verses is correct, it would imply that, for Vasu-
bandhu, simply dismissing the opponent’s challenges on the basis of an implicit resort to the
argument from ignorance is sufficient for establishing one’s own thesis. For Dharmapāla, this
is not the case. In light of the different strategies adopted by Vasubandhu and Dharmapāla, we
may suspect that the mode of constructing argumentation underwent dramatic changes in the
period between Vasubandhu and Dharmapāla, particularly with respect to whether refutation
of an opponent’s objection is sufficient for establishing one’s own thesis.
Non-Buddhist Indian Philosophy at the Time of Dharmapāla

Dharmapāla’s commentaries also contain much information about non-Buddhist schools. As
a conservative interpreter, Dharmapāla primarily repeats and rearranges the root texts upon
which he is commenting. This rearrangement of an argument may not be philosophically
significant, but it can be very historically informative. For example, according to the report
by Xuanzang’s disciple, Wengui (文軌, c. seventh century), Dharmapāla, in Chapter One of
his Four-Hundred Commentary, refutes in turn the atomist theories defended by three oppo-
nents, namely the Vaiśeṣikas, the pre-Vasubandhu17 Vaibhāṣikas, and the post-Vasubandhu
Vaibhāṣikas (see Keng 2019). Here, Dharmapāla’s rearrangement of Vasubandhu’s argument
can provide useful information about the philosophical debates that emerged between Vasu-
bandhu’s intellectual milieu and Dharmapāla’s.
Several times in his Four-Hundred Commentary, Dharmapāla cites and harshly criticizes
non-Buddhist as well as Buddhist philosophical schools. He critiques non-Buddhists (tīrthika),
the six masters, and then more specifically Vaiśeṣika, Sāṃkhya, Lokāyata or Cārvāka, Nihil-
ism, Jainism, and Grammarians. Dharmapāla also criticizes the mainstream Buddhist schools
such as Sarvāstivāda and Sautrāntika, but, interestingly, he refers to them less frequently than
to non-Buddhist schools.18 In this respect, Dharmapāla resembles his predecessor Dignāga and
his younger contemporary Bhāviveka to the extent that they all engage in dialogue and debate
with several contemporary Indian Buddhist and non-Buddhist schools in their philosophical
treatises. This indicates the existence of a sort of “neutral” philosophical platform where phi-
losophers from different traditions and backgrounds were able to meet and argue.
Dharmapāla’s Engagement With Bhāviveka and Madhyamaka

The Four-Hundred Commentary offers for Dharmapāla an opportunity to stage an important
debate with an opponent about the reality of the three natures (tri-svabhāva) (T30.1571.247b22–
249b3).19 The Yogācāra theory of three natures holds that our phenomenal world can be exam-
ined from three different perspectives: the imagined nature, dependent nature, and perfected
nature. The imagined nature ( parikalpita-svabhāva) refers to reified concepts such as “cups,”
which are generic names or misconceptualizations superimposed upon particulars (svalakṣaṇa)
that defy names and concepts. The dependent nature ( paratantra-svabhāva) refers to those
particulars to the extent that they arise and cease in accordance with the principle of dependent
origination ( pratītyasamutpāda). Most Yogācāra texts subscribe to a kind of metaphysical
idealism and treat particulars as merely mental representations (nimitta or ākāra) rather than
things external to the mind. Finally, the perfected nature ( pariniṣpanna-svabhāva) refers to
the aforementioned principle of dependent origination, usually named thusness (tathatā) in the
sense that it is “the way things are,” and it governs the dependent nature of arising and ceasing
365
Ching Keng
particulars. The Yogācāra position is that the imagined nature is not real, but the dependent
nature and the perfected nature are real. The dependent nature is real because Yogācāra holds
the foundationalist view that, while a cup as a generic and conceptualized name (the imagined
nature) is not real, there must be something real (the dependent nature) – that is, the mental
representation as a particular – that serves as the foundation for the name “a cup.” The per-
fected nature is real because it refers to the principle of dependent origination according to
which the mis-named mental representation arises and ceases.
In the commentary itself, Dharmapāla does not mention the name of his opponent, but
Xuanzang’s disciple Wŏnch’ŭk (圓測; 613–696) identifies this opponent as Bhāviveka (Qing-
bian 清辯) (X21.369.338b4–339b1, T33.1708.360b5–361a28). The opponent claims that the
dependent nature does not ultimately exist, but Dharmapāla maintains that it does. Dharmapāla
refutes the opponent as follows:
In many places where the Tathāgata talks about the three natures, he always claims
that the imagined nature does not exist, but the dependent nature and the perfected
nature do. Hence, we know that when the Tathāgata teaches emptiness, his teachings
about emptiness have a special intent. One must not deny the existence of all dharmas
by taking at face value the words of the Tathāgata. If people understand his intent by
taking his words at face value, then they would be criticizing the Mahāyāna.
(T30.1571.248a29 – b3; see also Keenan 1997, 111)
If people deny the existence of the dependent nature, then they are denying the exist-
ence of both defiled and undefiled dharmas. They should be named as holding a
wrong view about emptiness (durgṛhītā śūnyatā) and doing harm to both their own
position and the position they are refuting.
(T30.1571.248c5–6; see also Keenan 1997, 114)
Here Dharmapāla’s point is to insist that the dependent nature and the perfected nature are
not empty (literally meaning “zero,” and hence “non-existence”). If people insist that all three
natures are empty, then they would endorse nihilism. Moreover, if the dependent nature is
empty, then both defiled and undefiled dharmas would have no foundation, and hence neither
defiled dharmas (i.e., misconceptions such as “a cup”) nor undefiled dharmas (to see a mental
representation as it really is, i.e., devoid of the misconception of “a cup”) would exist, and
Buddhist cultivation would have no point. Hence, Dharmapāla insists, the doctrine of empti-
ness preached by the Buddha must not be understood as the non-existence of all three natures
but as the non-existence of merely the imagined nature.
For Dharmapāla, the imagined nature does not exist at all, but, for his opponent Bhāviveka,
ordinary things do exist at the conventional level. For Dharmapāla, the dependent nature
and the perfected nature ultimately exist. In contrast, Bhāviveka accepts the existence of the
dependent nature only at the conventional level (Madhyamakahṛdaya verse V.71 and com-
mentary, Eckel 2008, 274) but strongly denies the ultimate existence of unconditioned dhar-
mas such as thusness or emptiness.20
There is much to say about this debate. First, many major Yogācāra texts assert that, at the
stage of buddhahood, the dependent nature ceases.21 This implies that it is not ultimately real.
Interestingly, this is Wŏnch’ŭk’s basis for thinking that Bhāviveka is close to Paramārtha since
they both assert that dependent nature is ultimately unreal (T33.1708.360b11–16). That is to
say, the dependent nature must be eliminated in buddhahood. In contrast, CWSL, following the
Treatise on the Scripture of the Buddhaland (Fodi jing lun 佛地經論; *Buddhabhūmivyākhyā)22
366
Dharmapāla
(T26.1530.302c1–3), says that when one becomes a buddha, a dramatic transformation (the
so-called “transformation of the basis” [āśraya-parivṛtti or āśraya-parāvṛtti]) takes place so
that the previous defiled function of ordinary consciousness is transformed into an undefiled
cognitive function (T31.1585.56b2–3).
Thus, according to the CWSL, when Yogācāra texts mention the elimination of the depend-
ent nature, they do not mean to eliminate the dependent nature as a whole but simply to elimi-
nate the defiled aspect of the dependent nature (i.e., the cognitive function derived from the
so-called “storehouse consciousness” [ālayavijñāna], which is a sort of unconscious reposi-
tory of previous actions, defilements, and tendencies). The undefiled cognitive function sub-
sists even in buddhahood.
From philosophical and “buddhological” perspectives, the issue of whether the dependent
nature ultimately ceases concerns how we conceive of the mind of a buddha. If the dependent
nature ceases after the attainment of buddhahood, a buddha will be left with no mind because
all cognitive functions belong to the dependent nature. This would mean that the activity of
a buddha’s saving sentient beings would not imply that there is any cognitive function and
would resemble the activity of what Mark Siderits calls a “Robo-Buddha” (Siderits 2011).
In contrast, if dependent nature does not cease, a buddha will continue having a mind. The
only difference between ordinary people and a buddha would be that ordinary minds always
involve false discrimination (vikalpa), which generates afflictions, whereas a buddha’s mind
does not. This debate is important because it points to a major difference between Dharmapāla
and Bhāviveka on the one hand and Dharmapāla and his Yogācāra predecessors (Paramārtha,
for example) on the other.23
This disagreement between Dharmapāla and Bhāviveka is probably the first substantial
debate between Madhyamaka and Yogācāra in the history of Indian Buddhism. It is preceded
only by brief critical remarks in Yogācāra texts, including the disparagement in the Stages
of Yogācāra (Yogācārabhūmi) of “those who misunderstand emptiness (durgṛhītā śūnyatā)”
(Wogihara 1971, 47; T30.1579.488c22–28) and the denigration of Madhyamaka implied by
the passage in the Scripture on Unpacking the Profound Meaning (Saṃdhinirmocana-sūtra)
on the three turnings of the dharma wheel (Lamotte 1935, 85–86, 206–7). Both passages are
quite brief and do not explain exactly how Yogācāra differs from Madhyamaka.
Apart from this debate, Dharmapāla criticizes Bhāviveka in his commentary on the Four-
Hundred Verses by Āryadeva, allegedly a Madhyamaka master. This suggests that around
the time of Dharmapāla, the boundary between Madhyamaka and Yogācāra was more fluid
than many later Buddhist thinkers and contemporary scholars generally take it to be. In fact,
I would suggest that we consider Dharmapāla’s engagement with Bhāviveka the beginning of
the divide between Madhyamaka and Yogācāra.24
The Cognition of Self-Cognition25

As mentioned, some doctrines in the CWSL are attributed to Dharmapāla by Xuanzang’s disci-
ples. Among these, some of the most distinctive doctrines do not appear in any of the commen-
tarial works attributed to Dharmapāla. Hence, scholars have expressed doubts whether they
are authentic teachings of Dharmapāla. However, despite uncertainty about their authenticity,
these doctrines can tell us something about Indian Buddhist philosophy in the sixth and seventh
centuries. One such example can be picked from CWSL’s discussion about meta-cognition, that
is, about how to explain our ability to (sometimes) report on the occurrence of a cognition.
The Yogācāra tradition maintains that cognition consists of two aspects (*bhāga, ākāra,
ābhāsa, rūpa; fen 分): the image (nimitta) aspect and the seeing (dṛṣṭi or darśana) aspect. This
367
Ching Keng
conception is probably derived from the Sautrāntika representationalism, according to which

we cognize external objects indirectly through mental images (ākāra). This idea is only one
step away from the Yogācāra idealism, which claims that mental images originate not from
external objects but from the internal processes of consciousness.
According to the account given by the CWSL, it was Dignāga who developed this two-
aspect theory into a theory of three aspects in his Corpus of the Valid Means of Cognition
(Pramāṇasamuccaya).26 In addition to the image and the seeing aspects, there is an aspect of
self-cognition (svasaṃvedana or svasaṃvitti; zizheng fen自證分), namely the cognition that the
seeing aspect is cognizing the image aspect, for example, the awareness of seeing blueness.27
Somewhat surprisingly, the CWSL further maintains that cognition consists of four
aspects, not three (CWSL, T31.1585.10b17–28; Treatise on the Scripture of the Buddha-
land, T26.1530.303b13–20). The reason for establishing a fourth aspect, the cognition of
self-cognition (zhengzizheng fen 證自證分; *svasaṃvitti-saṃvitti), is that since self-cog-
nition cognizes and hence testifies to the functioning of the seeing aspect, another aspect
is required to cognize and testify to the functioning of self-cognition. In order to avoid the
problem of infinite regress, the CWSL claims that self-cognition, the third aspect, plays a
special role: it cognizes and testifies to the functioning of both the second and the fourth
aspects. Admittedly, this way out of the difficulty of infinite regress may sound too easy
to be persuasive. We can still ask: if the cognition of self-cognition (fourth aspect) can be
cognized by the self-cognition (third aspect) and hence a fifth one is in no need, then why
not make it that the self-cognition (third aspect) can both cognize and be cognized by the
seeing aspect (second aspect) and hence do away with the cognition of self-cognition (fourth
aspect)?28 In what follows, I aim not so much to defend this position as to contextualize it in
the debates between two models for meta-cognition in Abhidharma.
Master Ji’s commentary on the CWSL attributes the four-aspect theory to Dharmapāla
(T43.1830.242a24–26 and T43.1830.320c20–22). But the theory of four aspects does not
appear in any of the commentarial works attributed to Dharmapāla. Instead, it is found in the
Treatise on the Scripture of the Buddhaland, which is attributed to multiple authors, headed
by *Bandhuprabha (Qinguang 親光, dates unknown), who seems to be a contemporary of
Dharmapāla (X50.819.284a12–13). The aforementioned passage from the CWSL appears ver-
batim in the Treatise on the Scripture of the Buddhaland. This shows that, at least according to
Master Ji, Dharmapāla and *Bandhuprabha maintain similar doctrinal positions. In a few other
places, Master Ji also groups Dharmapāla and *Bandhuprabha together in contrast to other
masters (T43.1830.230b19–26; T45.1861.252c8–253a5; X48.806.377a11 – b4). One possible
explanation is that this theory originated not from Dharmapāla himself but from the group
headed by Dharmapāla, which also includes *Bandhuprabha and other authors of the Treatise
on the Scripture of the Buddhaland. So when Master Ji attributes this theory to Dharmapāla,
he is actually referring to the group headed by Dharmapāla rather than to Dharmapāla him-
self. Another possibility is that given the fact that this four-aspect theory is not attested in any
Indian sources, it is likely a Chinese invention (see Funayama 2020). Currently, we have no
way to confirm either explanation. However, regardless of its origin, this theory can provide
useful information about the evolution of Indian Buddhist philosophy.
As was pointed out by Yao (2005), in Abhidharma, there were already two models for
meta-cognition. According to Sarvāstivāda Abhidharma, meta-cognition cannot be achieved
by the cognition itself. A cognitive event can become the object of meta-cognition (e.g., the
awareness of cognizing blueness) only by means of another cognitive event. In other words,
a cognition can only be cognized by a subsequent cognition to achieve meta-cognition. Thus,
meta-cognition is similar to memory: it is a later cognition that brings an earlier cognition into
368
Dharmapāla
awareness. The typical metaphor provided to support this idea is that a knife does not cut itself,
and thus another knife is necessary to cut the first knife. In contrast, the Mahāsāṃghika school
of Abhidharma claims that meta-cognition is achieved not by a subsequent cognition because
every cognition is self-cognizing. After Dharmakīrti, the Mahāsāṃghika model became the
mainstream position (Yao 2005: 149).
The four-aspect theory can be interpreted as an interesting and ingenious synthesis of the
two different models proposed by the Sarvāstivāda and Mahāsāṃghika schools. As the Trea-
tise on the Scripture of the Buddhaland concludes,
Based on such reasoning, we conclude that cognition, despite being one, consists of
multiple aspects. These aspects are neither identical nor separate from each other.
Taken together, they cognize both what is inside and outside of cognition and free
from the fallacy of infinite regress.
(T26.1530.303b18–20)
This means that each cognition is itself also meta-cognition. But cognition consists of four
aspects, each being made aware by another aspect of the cognition. From the viewpoint of
each individual cognition, this theory adopts the Mahāsāṃghika model, namely that each cog-
nition is also meta-cognition. But, from the viewpoint of each aspect within a cognition, it
follows the Sarvāstivāda model; that is, each aspect is cognized by another aspect. In my
opinion, this theory clearly aims at reconciling the two conflicting models. Thus, whether this
four-aspect theory was formulated by Dharmapāla or not, it shows that there was a tension
between Sarvāstivāda and Mahāsāṃghika in their understandings of cognition and that crea-
tive solutions were in need.
Dignāga’s theory of self-awareness (svasaṃvedana) differs from contemporary theories in
the sense that for Dignāga, any cognition must at least implicitly include meta-cognition of
itself. Dignāga argues, based on memory, that if a cognition does not include meta-cognition
of itself, then at a later moment, one can never possibly recall that experience of the cogni-
tion in question.29 But some contemporary scholars think “beliefs and desires can be activated
unconsciously” (Carruthers and Gennaro 2020, 5). That is, there are mental states that we are
not aware of. Only those states of which we are aware are conscious states (Carruthers and
Gennaro 2020, 5). The distinction between unconscious and conscious mental states also moti-
vates, at least partly, the so-called higher-order theories of consciousness, of which we find a
counterpart in Dharmapāla’s theory of four aspects.
Higher-order theories in general hold that a mental state is not in itself conscious. It is
another mental state, that is, the higher-order mental state, that makes conscious the first-order
mental state. There are three major types of higher-order theories. The higher-order percep-
tion (HOP) theory claims that it is another perception (achieved by an inner sense scanning
the first-order cognition) that makes the first-order mental state conscious. The higher-order
thought (HOT) theory claims that it is a conceptualized belief that makes the first-order mental
state conscious, and there is some sort of causal power (either actual or dispositional) of the
first-order mental state that brings up the higher-order thought. What is common to these two
theories is that one of the first-order or the higher-order mental states can occur without the
other, despite some causal connection between the two. In contrast, the third theory, the self-
representational higher-order theory, holds that the first-order mental state possesses in itself a
higher-order intentional content that makes itself conscious.30
Due to the limitation of space, I cannot review various versions of the higher-order theories
and their respective strength and weakness. But I would like to draw attention to the similarities
369
Ching Keng
between Dharmapāla’s theory and Peter C. Carruthers’s “dual-content theory,” a subtype of

the self-representational higher-order theory (Carruthers and Gennaro 2020, 32).31 These two
are similar in two aspects. First, both theories hold that it is one and the same mental state that
is both first-order and higher-order (Carruthers and Gennaro 2020, 31). Despite Dharmapāla’s
claims that there are four aspects, some first-order and some higher-order, all four aspects
taken together constitute one single cognitive event. Second, in contrast to the part-whole
self-representational account (another subtype of the self-representational higher-order theory
according to which a conscious perception is always partially conceptual), Carruthers insists
that the higher-order intentional content, that is, a “seeming of the first-order content of the
state,” can be entirely non-conceptual (Carruthers and Gennaro 2020, 35). This agrees with
both Dignāga and Dharmapāla, who hold that self-cognition together with the cognition of
self-cognition must be non-conceptual. The question of self-cognition is one of multiple areas
in which Dharmapāla’s thought can engage contemporary philosophical debates.
A Solution to the Problem of the Janus-Faced Image of Dharmapāla

We now have a Janus-faced image of Dharmapāla: the commentarial works reliably attributed
to him make him appear quite conservative, but the theories attributed to him in the CWSL
depict him as an original thinker. To make matters more complicated, there is little evidence to
suggest that the positions cited in the CWSL are actually Dharmapāla’s.
Given the conflicting evidence, I suggest that the attributions of several positions to
Dharmapāla in the CWSL indicate that they were held by members of Dharmapāla’s lineage,
including *Bandhuprabha and Xuanzang, but not by Dharmapāla himself alone. Even if the
opinions attributed to Dharmapāla were actually Xuanzang’s or Master Ji’s, as suggested by
Funayama and others, they preserve invaluable information about theories of cognition in
sixth- and seventh-century India. Further comparison of doctrines attributed to Dharmapāla by
Master Ji with other extant sources will help us better judge the reliability of the attributions.32
Dharmapāla’s Legacy in East Asia

After Xuanzang returned to China in 643, he established a scholarly lineage that was later
labeled School of Dharma-Characteristics in China and Japan. In this tradition, the CWSL was
treated as the definitive doctrinal text. But Xuanzang’s lineage began to decline in the Tang
dynasty, and some of the main texts, including Master Ji’s commentary on the CWSL, were
lost in China. After that, interest in the CWSL largely receded, except for a brief revival in the
late Ming dynasty.33
In the history of Chinese Buddhism, the Indian masters Dharmapāla and Paramārtha have
long been considered to hold opposing views. For Dharmapāla, the mind (here understood as
the cognitive function originating from the seeds in the storehouse consciousness) is defiled
in nature, but for Paramārtha, the mind is fundamentally pure. Disputes arose surrounding
these different views that were expressed in terms of the opposition between the School of
Dharma-Characteristics (faxiang zong 法相宗 or simply xiang zong 相宗) and the School
of Dharma-Nature (faxing zong 法性宗 or simply xing zong 性宗). In the late Ming dynasty,
when interest in the CWSL revived, there were attempts to reconcile these two schools. The
disagreement concerning which school represents the authentic teachings of Indian Yogācāra,
Dharmapāla/School of Dharma-Characteristics or Paramārtha/School of Dharma-Nature, reap-
peared in early twentieth-century Chinese Buddhism. Regarding this millennium-long debate,
I have shown (Keng 2022) that the doctrinal source behind the School of Dharma-Nature was
370
Dharmapāla
not Paramārtha himself but the Chinese reinterpretation of Paramārtha’s original teachings
through the lens of the Awakening of Faith in Mahāyāna (Dasheng qixin lun 大乘起信論;
T1666). In reality, the doctrinal divide between Dharmapāla and Paramārtha may not have
been as substantial as the tradition holds it to be.
Conclusion
We thus have a bifurcated image of Dharmapāla: his commentarial works make him appear
rather conservative, but the positions ascribed to him in the CWSL make him seem quite
innovative. I have suggested that a tentative way to make sense of this bifurcation is to treat
Dharmapāla not as a single person but as a group composed of Dharmapāla and his students
and followers. Furthermore, even if these doctrines were Chinese inventions falsely attributed
to Dharmapāla, they still provide useful information about the dynamics at play within Indian
Buddhist philosophy in the sixth and seventh centuries.
The study of Dharmapāla contributes to a better understanding of Indian Buddhist philoso-
phy in at least the following aspects. First, Dharmapāla shows how to employ Dignāga’s logic
to defend Yogācāra philosophy, a strategy we find neither in pre-Dignāga Yogācāra thinkers
and only rarely in Dignāga. Second, Dharampāla’s reference to and criticism of his contempo-
rary Indian schools, both non-Buddhist and Buddhist, preserve invaluable information about
Indian philosophy around the sixth century. Third, Dharmapāla’s debate against Bhāviveka
provides a perspective about the relation between Yogācāra and Madhyamaka on the one hand
and between early Yogācāra and later Yogācāra on the other. Finally, even if some doctrines
were suspiciously and retroactively attributed to Dharmapāla, they might still teach us about
significant issues lying beneath the surface of Indian Buddhist philosophy around the sixth
century, such as the tension between the Sarvāstivāda and the Mahāsāṃghika models for
meta-cognition.
Finally, I close with encouragement for further research on Dharmapāla. Based on what
I have shown previously, it would not be an exaggeration to say that without adequate studies
of Dharmapāla, we will remain far from reaching a clear picture of Indian Buddhist philoso-
phy during the sixth and seventh centuries, a crucial transitional period between Vasubandhu
and Dignāga on the one hand and Bhāviveka, Candrakīrti, and Dharmakīrti on the other.
Notes
1 I am grateful to Mark Siderits, Funayama Tōru, Michael Radich, Tony Cheng, Lok-Chi Chan, and
Chih-Chiang Hu for their very useful comments and suggestions.
2 Most sources give 530–61 as Dharmapāla’s exact dates, but Funayama (2000) has doubts about this
dating. It would be safe to simply say that he was active in the sixth century.
3 Tillemans (2008, 8 ff.) gave a detailed description of Dharmapāla’s life and works. For a more recent
yet brief overview, see Moriyama (2019).
4 They are mostly brief and scattered in various Chinese sources. Here I rely on the brief biography
from the Biography of Xuanzang entitled The Biography of the Tripiṭaka Master of the Great Ci’en
Temple of the Great Tang dynasty (大唐大慈恩寺三藏法師傳); see T50.2053.241c11–24.
5 Cross-references to Vinītadeva can be extremely helpful for studying Dharmapāla. As mentioned
in the following, the three works reliably attributed to Dharmapāla are all commentaries. It should
be noted that Vinītadeva also composed commentaries on two of the same texts, namely Vasuban
dhu’s Twenty Verses and Dignāga’s Investigation of Cognitive Objects. The Japanese translations of
Vinītadeva’s two commentaries can be found in Yamaguchi and Nozawa (1953). Recently Duckworth
et al. (2016, 78–104) provide an English translation of Vinītadeva’s commentary on the Investiga-
tion of Cognitive Objects. In addition, Yamaguchi and Nozawa (1953) have translated Sthiramati’s
371
Ching Keng
commentaries on Vasubandhu’s Thirty Verses into Japanese. An examination of Sthiramati’s com-

mentary is necessary to ascertain the accuracy of Master Ji’s ascription of certain doctrines to Sthi-
ramati and Dharmapāla in the CWSL.
It is interesting to note that Vinītadeva refers to Dharmapāla’s commentary and observes that
Dharmapāla’s interpretation “might be a correct understanding, but his idea is too profound for me to
comprehend” (Duckworth et al. 2016, 80). In contrast, Vinītadeva says that the goal of his commen-
tary on the Investigation of Cognitive Objects is merely “to delight beginners” (ibid., 103). Despite
this difference, Yamaguchi and Nozawa (1953, 418) remark that the two masters generally agree.
6 As for contemporary scholarship on the Four-Hundred Commentary, Tillemans (2008 [1990])
includes an excellent study and annotated English translation of Chapter 4 (Refutation of Heretical
Views) and Chapter 5 (Refutation of the Sense-Organs and Their Objects). Keenan (1997) provides
an English translation of Chapter 10, which contains an important debate between Dharmapāla and
his opponent (probably Bhāviveka, see subsequent discussion); also see Hoornaert (2004), which,
however, is not a complete translation of the chapter. There is no English translation of the Twenty
Verses Commentary. Ui (1963) offers a kundoku 訓讀 (Japanese way of reading Chinese texts) with
footnotes. Similarly, no English translation of the Investigation Commentary is available. And, again,
Ui (1951) provides a kundoku with footnotes.
7 Traditionally, the CWSL was said to be compiled by Xuanzang, who collected from the ten commen-
taries on the Thirty Verses. Recently, Yamabe (2020) argues that this genre of collecting the views of
various masters into one volume may have already existed before Xuanzang’s time, which suggests
that the CWSL might have been based on a single Indian original source.
8 As for contemporary studies of the CWSL, the best available is still the French annotated translation
by Louis de la Vallée Poussin (1928–29), which was completely translated into English by Gelong
Lodrö Sangpo and Gelongma Migme Chödrön (2017). Francis Cook (1999) provides another com-
plete English translation.
9 For example, in a few places in his commentaries, Dharmapāla tries to answer the objection that
a particular statement in Dignāga’s root text is pointless (anartha) or serves no purpose (e.g.,
T31.1625.890a13–15).
10 There are doubts about whether these first lines are verse or prose. In addition, Kellner and Taber
(2014, 735–36) argue that this is neither the first verse nor an inference. My assertion that Dharmapāla
takes this to be a well-formulated inference owes much to Hu (2018, 16 ff.).
11 Unless otherwise cited, all translations are mine.
12 For more details, please see the chapter on Dignāga by Kei Kataoka in this volume.
13 For Bhāviveka and the culture of debate, see Eckel (2008, 9ff.).
14 That is, Dignāga refutes three of the realist opponent’s theories in verses 1–5 and then proposes his
own theory in verses 6–8.
15 Vinītadeva thinks that Dignāga’s argument consists of two parts: “It refutes an external view of the
percept, and it affirms an internal one” (Duckworth et al. 2016, 52).
16 For example, if we show that the opponent’s argument commits the fallacy of indeterminacy
(anaikāntika; buding 不定), this simply shows that the reason does not necessarily lead to the conclu-
sion, not that the opponent’s thesis is definitely wrong.
17 To be more precise, “pre-Vasubandhu” and “post-Vasubandhu” mean before and after the text
Abhidharmakośabhāṣya.
18 More work should be done to examine what the major philosophical disagreements are between
Dharmapāla and his opponents, whether Dharmapāla’s references to his opponents are faithful, and
how Dharmpāla’s refutation agrees or differs from those found in Bhāviveka and his other contempo-
rary Buddhist scholars.
19 For more details about the three natures, please see the chapter on Vasubandhu by Jonathan C. Gold
in this volume.
20 An example of this can be found in the inference proposed by Bhāviveka in his Jewel in the Hand
(*Mahāyāna-hastaratna-śāstra): “From the ultimate perspective ( paramārthatas), conditioned dhar-
mas are empty, like illusion, because they originate in dependence [on something else]. Unconditioned
dharmas do not exist because they do not arise, like flowers in the sky” (T30.1578.268b21–22;).
I cannot go into detail here. Readers should consult recent studies by Keenan and Eckel. For a brief
discussion, see Moriyama (2019, 170).
21 Such texts include Madhyāntavibhāga I.4, III.9; Mahāyānasaṃgraha II.2; and Triṃśikā 5a. For the
Mahāyānasūtrālaṃkārabhāṣya XIX.51 on the destruction of the dependent nature, see D’Amato
(2005, 197).
372
Dharmapāla
22 No Sanskrit survives for this text, unfortunately. For information about its authors, see subsequently
under the section “The Cognition of Self-Cognition.”
23 I would like to suggest that the Yogācāra thinkers living before Dharmapāla, including Asaṅga and
Vasubandhu, all side with Bhāviveka’s view that the dependent nature does not ultimately exist; that
is, it does not exist as the buddha’s cognitive function. But there is no space here to prove this point.
24 For example, Tāranātha (1575–1634) once remarked, “Before the arrival of these two masters
(Buddhapālita and Bhāviveka), all adherents of the Mahāyāna remained under the same teaching”
(Eckel 2008, 11).
25 Here I follow Funayama (2020) and translate the Chinese term zheng zizheng fen 證自證分 and the
reconstructed Sanskrit term *svasaṃvitti-saṃvitti as “cognition of self-cognition.” The issue I am
dealing with in this section is better known among Anglophone Buddhologists under the designa-
tion “self-awareness.” I did not choose the more popular translation “self-awareness” for svasaṃvitti
or “awareness of self-awareness” for *svasaṃvitti-saṃvitti because it has now become standard to
use self-awareness to refer to Dignāga’s specific way of explaining how meta-cognition is possible.
According to Dignāga, meta-cognition is achieved without the mediation of any higher-order cogni-
tion. But, as shown subsequently, for Dharmapāla, meta-cognition is achieved through a combination
of first-order and higher-order mental states.
26 As far as I am aware, most scholars in the Indian epistemological tradition ( pramāṇa) do not agree
that, for Dignāga, self-cognition refers to some third aspect that cognizes and testifies to the function-
ing of the seeing aspect. In their interpretation, self-cognition simply means that any cognition, which
consists of two aspects, is reflexive or self-cognizing in nature. See footnote 28. Thanks to Mark
Siderits for reminding me to make clarifications on this point.
27 For Dignāga’s establishment of the two aspects and the reasons behind it, see the chapter on direct
perception ( pratyakṣa) in the Corpus of the Valid Means of Cognition, especially verses I.11–12.
Note that Dignāga’s terminology is different: he uses the term svābhāsa for the seeing aspect and
arthābhāsa for the image aspect. The CWSL’s interpretation of Dignāga as holding a theory of three
aspects may have been based on Dignāga’s distinction among “what is being measured” ( prameya),
“the means for measuring” ( pramāṇa), and “the result of measuring ( pramāṇa-phala)” in verse I.9–
10. Given the Buddhist notion of no self, self-cognition should be formulated as “awareness of seeing
blueness” rather than “awareness that I am seeing blueness.”
28 Note that the CWSL does offer an explanation for why the second aspect does not cognize the third
aspect, and hence the fourth aspect is necessary. The reason is that what cognizes and testifies to
the function of the self-cognition must function in the manner of direct perception (xianliang 現量;
pratyakṣa), but the second aspect does not always function in that manner. See T31.1585.10b20–22.
Indeed, the post-Dignāga Indian epistemologists, Dharmakīrti in particular, did take this approach to
claim that self-cognition does not refer to any aspect within a cognition but simply means that any
cognition, which consists of two aspects, is reflexive or self-cognizing in nature. A cognition does
not become aware because of a separate or higher-order cognition. To borrow the stock example,
cognition is self-illuminating, just like a lamp illuminates the objects as well as itself. For a brief
summary of the views about meta-cognition from Dignāga and Dharmakīrti, see section five, “Bud-
dhist Accounts of Meta-Cognition,” in the general introduction of Siderits, Keng, and Spackman
(2020).
29 See Corpus of the Valid Means of Cognition, verse I.11d (Hattori 1968, 30). For a brief explanation
of Dignāga’s memory argument for self-awareness, see section five, “Buddhist Accounts of Meta-
Cognition,” in the general introduction of Siderits, Keng, and Spackman (2020).
30 For a survey of the higher-order theories, see Carruthers and Gennaro (2020).
31 Here I aim only to draw attention to the similar aspects between Dharmapāla’s and Carruthers’ views.
I don’t mean to suggest that these two are similar in every aspect. For example, Carruthers holds that
a state’s being conscious is a dispositional property, but Dharmapāla would not accept this. Again,
many thanks to Mark Siderits for drawing my attention to this difference between Carruthers and
Dharmapāla.
32 This is precisely the path taken by Sakuma (2006 [2008]), where he provides an excellent compari-
son between the doctrines attributed to Dharmapāla in the CWSL and Sthiramati’s position regarding
three issues: the correspondences between the buddhas’ four kinds of cognition and ordinary sentient
beings’ eight kinds of consciousness; the correspondences between the buddhas’ four kinds of cogni-
tion and the buddhas’ three bodies; and the formation of the five-gotra system.
33 For a very brief description of the reception of Yogācāra Buddhism in China and its revival in the late
Ming dynasty, see Duckworth et al. 2016, xviii–xx.
373
Ching Keng
Bibliography
T: Taishō shinshū Daizōkyō 大正新脩大藏經 (Tōkyō: Daizōkyō kankō kai, 1924–35).
X: Shinsan Dai Nihon zoku Zōkyō 新纂大日本續藏經 (Tōkyō: Kokusho kankō kai, 1975–1989).
Carruthers, Peter, and Rocco Gennaro. 2020. “Higher-Order Theories of Consciousness.” In The Stanford
Encyclopedia of Philosophy, Fall 2020 ed., edited by Edward N. Zalta. https://plato.stanford.edu/
archives/fall2020/entries/consciousness-higher/.
Cook, Francis. 1999. Three Texts on Consciousness Only. Berkeley: Numata Center.
D’Amato, Mario. 2005. “Three Natures, Three Stages: An Interpretation of the Yogācāra Trisvabhāva-
Theory.” Journal of Indian Philosophy 33: 185–207.
de La Vallée Poussin, Louis. 1928–29. Vijñaptimātratāsiddhi: La siddhi de Hiuan-Tsang. Paris: Paul
Geuthner.
Duckworth, Douglas, Malcolm David Eckel, Jay L. Garfield, John Powers, Yeshes Thabkhas, and Sonam
Thakchoe. 2016. Dignāga’s Investigation of the Percept: A Philosophical Legacy in India and Tibet.
Eckel, Malcolm D. 2008. Bhāviveka and His Buddhist Opponents. Cambridge: Harvard University Press.
Funayama, Tōru 船山徹. 2000. “Two Notes on Dharmapāla and Dharmakīrti.” Zinbun 35: 1–11.
———. 2001. “On the Date of Vinītadeva.” In Le Parole e i Marmi, edited by Raffaele Torella, 309–25.
Roma: Istituto Italiano per l’Africa e l’Oriente.
———. 2020. “The Genesis of *Svasaṃvitti-saṃvitti Reconsidered.” In Buddhist Philosophy of Con-
sciousness: Tradition and Dialogue, edited by Mark Siderits, Ching Keng and John Spackman,
209–24. Leiden and Boston: Brill-Rodopi.
Hoornaert, Paul. 2004. “The Dharmapāla-Bhāvaviveka Debate as Presented in Dharmapāla’s Com
mentary to Catuḥśataka XVI.23.” Kanezawa daigaku bungakubu ronshū: Kōdō kagaku tetsugaku hen
金沢大学文学部論集. 行動科学‧哲学篇 24: 119–49.
Hu, Chi-Chiang 胡志強. 2018. “Zhijue, tazhe yu luoji: Hufa Cheng weishi baosheng lun zhi zhexue
yanjiu 知覺、他者與邏輯：護法《成唯識寶生論》之哲學研究” (“Perception, Others, and Logic:
A Philosophical Study on Dharmapāla’s Cheng weishi baosheng lun”). PhD diss., National Chengchi
University, Taiwan.
Keenan, John P. 1997. Dharmapāla’s Yogācāra Critique of Bhāvaviveka’s Mādhyamika Explanation
of Emptiness: The Tenth Chapter of Ta-Ch’eng Kuang Pai-Lun Shih Commenting on Āryadeva’s
Catuḥśataka Chapter Sixteen. Lewiston: Edwin Mellen Press.
Kellner, Birgit, and John Taber. 2014. “Studies in Yogācāra-Vijñānavāda Idealism I: The Interpretation of
Vasubandhu’s Viṃśikā.” Asiatische Studien/Études Asiatiques 68 (3): 709–56.
Keng, Ching. 2019. “Three Senses of Atomic Accumulation – An Interpretation of Vasubandhu’s
Viṃsika Stanzas 12–13 in Light of the Abhidharmakosabhaṣya and Dharmapala’s Dasheng Guang-
bailun Shilun.” Journal of Indian Philosophy 47 (3): 565–601.
———. 2022. Toward a New Image of Paramartha: Yogacara and Tathagatagarbha Buddhism Revis-
ited. London, UK; New York, NY: Bloomsbury Academic.
Lamotte, Étienne. 1935. Samdhinirmocana sutra, l’explication des mystères: texte tibétain édité et tra-
duit par Étienne Lamotte. Louvain: Bureaux du recueil Bibliothèque de l’Universite.
Lusthaus, Dan. 2002. Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Buddhism
and the Ch’eng Wei-shih lun. London and New York: RoutledgeCurzon.
Moriyama, Shinya. 2019. “Dharmapāla.” In Brill’s Encyclopedia of Buddhism, vol. II, 168–72. Leiden
and Boston: Brill.
Sakuma, Hidenori. 2006 (2008). “On Doctrinal Similarities Between Sthiramati and Xuanzang.” Journal
of the International Association of Buddhist Studies 29 (2): 357–82.
Sangpo, Gelong Lodrö, and Gelongma Migme Chödrön, with the assistance of Alexander Leonhard
Mayer, trans. 2017. Vijñapti-mātratā-siddhi: a commentary (Cheng weishi lun) on Vasubandhu’s
Triṃśikā by Xuanzang, translated from Chinese into French and Annotated by Louis de La Vallée
Poussin. New Delhi: Motilal Banarsidass Publishers.
Siderits, Mark. 2011. “Buddhas as Zombies: A Buddhist Reduction of Subjectivity.” In Self, No Self?
Perspectives from Analytical, Phenomenological, & Indian Traditions, edited by Mark Siderits, Evan
Thompson, and Dan Zahavi. Oxford: Oxford University Press.
Siderits, Mark, Ching Keng, and John Spackman. 2020. Buddhist Philosophy of Consciousness: Tradi-
tion and Dialogue. Leiden and Boston: Brill-Rodopi.
Tillemans, Tom J. F. 2008. Materials for the Study of Āryadeva, Dharmapāla, and Candrakīrti: The
Catuḥśataka of Āryadeva, Chapters XII and XIII with the Commentaries of Dharmapāla and
374
Dharmapāla
Candrakīrti: Introduction, Translation, Notes, Sanskrit, Tibetan and Chinese Texts, Indexes. New
Delhi: Motilal Banarsidass Publishers.
Ui, Hakuju 宇井伯壽. 1951. Jinna chosaku no kenkyū 陳那著作の研究 (Studies of the Works of
Dignāga). Tokyo: Iwanami Bookstore.
———. 1963. “Jō yuishiki hōshōron kenkyū 成唯識寶生論研究 (A Study of the Cheng weishi baosheng
lun)”. In Daijō butten no kenkyū 大乘佛典の研究 (Studies of Mahāyāna Buddhist Texts), 607–811.
Tokyo: Iwanami Bookstore.
Wogihara, Unrai 荻原雲來. 1971. Bodhisattvabhūmi; A Statement of Whole Course of the Bodhisattva
(being fifteenth section of Yogācārabhūmi). Tōkyō: Sankibō.
Yamabe, Nobuyoshi 山部能宜. 2020. “A Hypothetical Reconsideration of the ‘Compilation’ of Cheng
Weishi Lun.” In From Chang’an to Nālandā: The Life and Legacy of the Chinese Buddhist Monk
Xuanzang (602? – 664), edited by Shi Ciguang, Chen Jinhua, Ji Yun, and Shi Xingding. Singapore:
World Scholastic Publishers.
Yamaguchi, Susumu 山口益, and Jōshō Nozawa 野沢静証. 1953. Seishin yuishiki no genten kaimei 世
親唯識の原典解明 (Explication of the Yogācāra Primary Texts by Vasubandhu). Kyoto: Hōzōkan.
Yao, Zhihua 姚治華. 2005. The Buddhist Theory of Self-Cognition. London and New York: Routledge.
375
22
STHIRAMATI
A Yogācāra Commentator and Innovator
Sthiramati1 (circa sixth century CE) was an Indian Buddhist scholar of the Yogācāra, best
known for his extensive commentaries on some of the most important treatises of this tradi-
tion. What often goes less noticed, however, is his role in the very creation of that tradition, a
function of his particular position within the school’s textual development.2 Relative to previ-
ous thinkers like Asaṅga and Vasubandhu, Sthiramati operated under a much more defined
notion of the Yogācāra as a distinct school, or at least a more defined textual tradition; accord-
ingly, his interpretive challenge – and contribution – consisted in synthesizing a varied textual
corpus into a coherent and consistent worldview, adding to it in the process some original and
strikingly innovative insights.
In this chapter, we seek to provide the necessary background to his writings and life and to
present a brief survey and evaluation of his philosophical contributions and their significance.
The first section, authored by Jowita Kramer, presents the state of current research on what
is known regarding his life, dates, and authorship and aims to point out some common philo-
sophical and stylistic traits among the works ascribed to him by providing a comparative and
systematic consideration of his intertextual realm. The concluding section, authored by Roy
Tzohar, shows Sthiramati as an individual thinker by examining, as a case study, his lengthy
comments on the opening verse of Vasubandhu’s Treatise in Thirty Verses (Triṃśikā), high-
lighting their innovativeness and unique contribution to Yogācāra thought.
Sthiramati’s Life, Authorship, and Intertextual Realm

Most of the information about Sthiramati’s life and work available to us today comes from
Chinese and Tibetan sources whose historicity is uncertain.3 A datable Indian reference to
the name Sthiramati can be found in an inscription of King Guhasena (r. ca. 553–569) from
Valabhī in western India, in which a monastery built by an abbot with this name is men-
tioned (Bühler 1877, 9 and plate II, lines 3–4; Silk 2009, 384–85). In an inscription of
Śīlāditya III (661/2), Sthiramati is also said to have founded a monastery in Valabhī.4 Fur-
thermore, the Sanskrit colophons of the Commentary on the Treatise on the Five Constituents
(Pañcaskandhakavibhāṣā) (Kramer 2014a, 133), the Commentary on the Distinguishing of the
Middle from the Extremes (Madhyāntavibhāgaṭīkā) (Yamaguchi 1934, 165), the Commentary
on the Treatise on Mental Presentation in Thirty Verses (Triṃśikāvijñaptibhāṣya) (Buescher
376 DOI: 10.4324/9781351030908-33

Sthiramati
2007, 142), and the Commentary on the Treasury of Abhidharma [Called] True Reality
(Tattvārthā Abhidharmakośaṭīkā) (Matsuda 2013, 49) name Sthiramati as the author of these
commentaries. However, these do not allow us to determine his date (Kramer 2019, 456f.).
Based on the Valabhī inscription of Guhasena and on the assumption that Sthiramati and
Guhasena were contemporaries, Erich Frauwallner proposed Sthiramati’s lifetime as 510–570
(Frauwallner 1961, 136–37). However, taking into account that the inscription states only that
the monastery was erected by Sthiramati (and not for Sthiramati), we cannot know with certainty
that Sthiramati was still alive at the time of Guhasena’s donation. Some scholars have dated
Sthiramati to 470–550 (Nguyen 1990, 21) because, according to Chinese sources, one of the
commentaries attributed to Sthiramati in the colophon to its corresponding Tibetan translation,
namely the Da Baoji jinglun (大寶積經論, T. 1523), corresponding to the Commentary on the
Kāśyapa Chapter (*Kāśyapaparivartaṭīkā), was translated into Chinese already between 508 and
535 (by Bodhiruci) (Silk 2009, 382). It has also been suggested in modern studies that there may
have been more than one author named Sthiramati or that works that have actually been com-
posed by an author (or authors) with a different name over time have (erroneously) been assigned
to one single author, namely Sthiramati (Silk 2009, 385; Silk 2015, 149; Kramer 2016a, 71).
All together, seventeen works have been ascribed to (an author called) Sthiramati. They
are extant in Sanskrit, Tibetan, Chinese, Uighur, and/or Mongolian and cover a wide range of
topics. The ten commentaries among them refer to both Yogācāra and non-Yogācāra root texts
(see also Kramer 2019, 457). These commentaries include
1. the Explication of the Compendium of the Abhidharma (Abhidharmasamuccayavyākhyā;

available in Sanskrit, Tibetan, and Chinese)
2. the Commentary on the Treasury of Abhidharma [Called] True Reality (Tattvārthā
Abhidharmakośaṭīkā; Sanskrit [chapter 3 and parts of chapters 2 and 4 missing], Tibetan,
Chinese [fragments], Uighur [fragments], Mongolian)
3. the Commentary on the Treatise on the Five Constituents (Pañcaskandhakavibhāṣā; San-
skrit, Tibetan)
4. the Commentary on the Treatise on Mental Presentation in Thirty Verses
(Triṃśikāvijñaptibhāṣya; Sanskrit, Tibetan)
5. the Commentary on the Distinguishing of the Middle from the Extremes
(Madhyāntavibhāgaṭīkā; Sanskrit, Tibetan)
6. the Commentary on the Comments to the Ornament of [Mahāyāna] Sūtras
(*Sūtrālaṃkāravṛttibhāṣya; Tibetan)
7. the Commentary on the Kāśyapa Chapter (*Kāśyapaparivartaṭīkā; Tibetan, Chinese, per-
haps more correctly *Ratnakūṭasūtraṭīkā?)
8. the Commentary on the Mahāyāna Madhyamaka (Dasheng zhongguan shilun 大乘中觀
釋論; Chinese)
9. the Commentary on the Exposition of Akṣayamati (*Akṣayamatinirdeśaṭīkā; Tibetan;
Braarvig 1993, cxxiii and cxxviii)
10. a commentary on the Collection of Means of Knowledge (Pramanasamuccaya; lost; Mat-
suda 2013, 52).
The remaining seven works associated with the name Sthiramati seem to belong to the Tantric
genre and are only mentioned in Tibetan sources (Cordier 1909, 138ff., 151 and 158; Kramer
2016a, 49).
Sthiramati’s work has received far less attention in modern studies than the writings of other
Yogācāra authors, such as Asaṅga or Vasubandhu, most likely because he is perceived as a
377
commentator and not as an original author in his own right. As shown in Kramer 2012, 2013a,
2013b, 2016a, 2016b, this negligence is unjustified, as some of the commentaries that are regarded
as Sthiramati’s main works, such as the Commentary on the Treatise on the Five Constituents and
the Commentary on the Treatise on Mental Presentation in Thirty Verses, clearly contain creative
and innovative thought (as will be seen in the second part of this chapter). It is also important to
note that the creative energy of commentators such as Sthiramati manifests not only in the crea-
tion of entirely new ideas but also in the reorganization of previous teachings (Kramer 2015, 327).
We may assume with some certainty that the tantric works ascribed to Sthiramati were com-
posed by a different author than the philosophical commentaries. At the same time, the Com-
mentary on the Treatise on the Five Constituents, the Commentary on the Treatise on Mental
Presentation in Thirty Verses, and the Commentary on the Treasury of Abhidharma [Called] True
Reality appear very closely related and were probably composed by the same person. The three
texts share many parallel passages, which make their interdependence obvious. For instance, a
whole section on the mental (caitasika) factors of the Commentary on the Treatise on the Five
Constituents is reused in the Commentary on the Treatise on Mental Presentation in Thirty Verses.5
In the Commentary on the Treasury of Abhidharma [Called] True Reality, Sthiramati
provides an extensive commentary on the Commentary on the Treasury of Abhidharma
(Abhidharmakośabhāṣya). Kazunobu Matsuda and other scholars working on the edition of
the Sanskrit manuscript of the text have found a great number of parallels between the Com-
mentary on the Treatise on the Five Constituents and the first chapter of the Commentary on
the Treasury of Abhidharma [Called] True Reality, showing the close relationship of the two
texts.6 At the same time, the Commentary on the Treasury of Abhidharma [Called] True Real-
ity generally follows the traditional Abhidharmic system without referring to specific Yogācāra
terminology and doctrines. However, the exploration of this text, the Sanskrit original of which
came to light only recently, is still in its infancy; thus, at the present time, no definitive state-
ments about its nature and its philosophy can be made.7
In contrast, the Commentary on the Treatise on the Five Constituents represents a compre-
hensive compendium of Yogācāra Abhidharma, discussing topics such as the “store mind”
(ālayavijñāna) extensively. By means of the Treatise on the Five Constituents (Pañcaskandhaka),
Vasubandhu produced a handy manual on the five constituents of the person (skandha) as viewed
from the perspective of the Yogācāras. According to Sthiramati, the reasons for the conciseness
of the work, covering merely seven manuscript folios, are to be found in the intention to meet the
needs of its potential readers: the householders who do not have enough time to read extensive
treatises because of their various duties and the contemplating monks who should not be dis-
tracted by reading lengthy works (Kramer 2014a, xx). In contrast to the briefness of Vasubandhu’s
work, Sthiramati produced an extensive commentary ten times the length of the root-text.
The Commentary on the Treatise on the Five Constituents may be subdivided into seven
main parts: (1) introductory remarks; (2) explanation of the five constituents; (3) explanation
of the twelve realms (āyatana); (4) explanation of the eighteen elements (dhātu); (5) reasons
for explaining the constituents, realms, and elements; (6) list (mātṛkā) of qualities; (7) con-
cluding matter. The main and longest section of the text is the second one, in which the five
constituents (skandha) are described and analyzed in detail. The two most notable and impor-
tant parts of this section are the explanations of “matter” (rūpa) and of “perception” (vijñāna).
Although Sthiramati mostly does not explicitly mention his sources in his comments on mat-
ter, his positions are closely related to the earlier Yogācāra texts, the Commentary on the Treas-
ury of Abhidharma and the Compendium of the Abhidharma (Abhidharmasamuccaya).8 At the
same time, he occasionally develops his own views, for instance, when characterizing objects
of the sense faculties, such as sounds and “information” (vijñapti), that is, karmic action that is
378
Sthiramati
perceivable by others.9 Another site of Sthiramati’s innovation is his discussion of whether there
are any other types of matter that are comparable to “non-information” (avijñapti), an invisible
and penetrable kind of matter related to karmic action that does not reveal itself to others. Here,
he puts forth ideas that do not appear to be mentioned in the texts upon which he is comment-
ing, and the way he discusses this problem indicates that the boundary between material and
mental factors was controversial among the authors of Abhidharmic texts. Especially Yogācāra
authors seem to have had some difficulties in distinguishing material from mental entities, when
they tried to categorize objects which appeared to be “borderline cases” between matter and
mind, such as the matter of “non-information” (avijñapti) and the matter of atoms or visualized
images. The reason for these difficulties among the Yogācāras in particular might be found in
the fact that it was in this milieu that the existential status of matter weakened increasingly in
favor of the position of mind – a condition which finally resulted in the evolving of the theory
that all external phenomena exist only as mental images (vijñaptimātratā).
It is notable that Sthiramati rejects a view that clearly has its origin in the Compendium of
the Abhidharma, namely the division of the invisible and penetrable kind of matter into five
categories. These five are classified in the Compendium of the Abhidharma as belonging to the
Abhidharmic category of “factors” (dharma), which are the object of the mental faculty (manas),
that is, can only be perceived by thinking. The fivefold categorization of the Compendium of
the Abhidharma includes compressed (ābhisaṃkṣepika) matter, matter of space (ābhyavakāśika),
matter of commitment (sāmādānika), imagined ( parikalpita) matter, and matter produced by
those with meditative power (vaibhūtvika). Sthiramati opposes the view that the categories other
than avijñapti (appearing as “matter of commitment” [sāmādānika] in the Compendium of the
Abhidharma) are to be classified as material. According to Sthiramati, the first two kinds, the
compressed matter and the matter of space, are both matter of atoms and thus included in “color,”
which is part of visible matter and an object of the faculty of sight. The objects of the last two cat-
egories, the imagined matter and the matter produced by those with meditative power, are nothing
other than mental images and are therefore to be regarded as being part of the mind (and not mat-
ter). Remarkably, Sthiramati seems to generally follow a position here that in the Commentary
on the Treasury of Abhidharma is ascribed to the Vaibhāṣika-Sarvāstivāda tradition. According to
the latter, only the non-information (avijñapti) is to be regarded as invisible and penetrable matter.
The contrary view, namely that meditative images are to be considered this kind of matter, too,
was held by some early Yogācāras (as represented by some portions of the Foundation of Yoga
Practitioners [Yogācārabhūmi]) and the “Sautrāntikas” (Kramer 2013b, 93), a group within the
Sarvāstivāda tradition known for frequently opposing the positions of the Vaibhāṣikas.
However, it is also important to note that, unlike the Vaibhāṣikas, Sthiramati rejects the
view that non-information exists as a real entity (dravya). At the same time, he points out
that this nonexistence of non-information as a real entity does not lead to the nonexistence
of the various types of restraint (saṃvara). The concepts of restraint and non-information are
inextricably linked in the Abhidharmic context. Therefore one could (wrongly) conclude that
Sthiramati’s rejection of non-information as a real entity means that restraint is not really exist-
ent. The background of this statement is probably the discussion between the Vaibhāṣikas, who
object the claim that non-information cannot really exist, and the “Sautrāntikas,” who defend
their view that non-information cannot be a real entity. Notably, Sthiramati’s explanation of
his own view partly relies on the assumption of the “store mind” (ālayavijñāna) and thus goes
beyond Vasubandhu’s line of reasoning. Sthiramati understands the concept of “restraint” as
an intention to restrain oneself from committing wrong deeds. This intention is produced at
the time when the ordination ceremony takes place, imprinting a seed in the store mind, which
constitutes the basis for future intentions of the same kind (see Kramer 2013b, 93).
379
Also in contrast to the Compendium of the Abhidharma, Sthiramati, in a few instances, even
incorporates into the Commentary on the Treatise on the Five Constituents some aspects of
the position “that everything is a (mental) presentation only” (vijñaptimātratā) – for instance,
when he mentions that color, from the viewpoint of the highest reality ( paramārthatas), is
not the object of the faculty of sight, because perception (vijñāna) does not have any external
objects. According to Sthiramati, from the viewpoint of highest reality, the existence of color
atoms, like that of shape, is not possible.10
The greatest part of the section on perception (vijnāna) of the Commentary on the Treatise
on the Five Constituents is devoted to a description of the “store mind” (ālayavijnāna). Sthi-
ramati comments extensively on the three characteristics, mentioned already in the Treatise
on the Five Constituents, which mark the differences between the store mind and the actual
perception ( pravṛttivijñāna) (Kramer 2014b, 315):
1. The object of the ālayavijnāna and the mode in which it is apprehended is not clearly
determined (aparicchinnālambanākāra).
2. The ālayavijnāna is of one kind (ekajātīya).
3. The continuity of the ālayavijnāna is not interrupted (santānānuvrtti).
Sthiramati moreover explains in detail the four arguments (also already stated by Vasubandhu)
that make the existence of a store mind besides the mental stream of continuously fluctuating
sense perceptions necessary.11 These four proofs are of particular interest because they differ from
the eight proofs presented in the Yogācārabhūmi and the Commentary on the Compendium of
the Abhidharma (Abhidharmasamuccayabhāṣya), as well as from the arguments provided in the
Summary of the Mahāyāna (Mahāyānasaṃgraha).12 Of particular interest is also the Commentary
on the Treatise on the Five Constituents’ explanation of the process beginning in the moment
of actually perceiving an object until the moment in which intention (cetanā) leaves an imprint
(vāsanā) in the store mind (Kramer 2016b, 152ff.). The Yogācāra concept of the “afflicted notion
[of ‘I am’]” (kliṣṭaṃ manas) is treated more briefly in the Commentary on the Treatise on the Five
Constituents than the store mind (ālayavijnāna). It is notable that neither Vasubandhu nor Sthi-
ramati provide any proofs for the afflicted notion of “I am” comparable to those for the store mind.
It appears likely that the same author who composed the Commentary on the Treatise on the
Five Constituents and the Commentary on the Treatise on Mental Presentation in Thirty Verses
also produced the Explication of the Compendium of the Abhidharma. This commentary on the
Compendium of the Abhidharma is a compilation of the root text and another commentary, namely
the Commentary on the Compendium of the Abhidharma. Neither of the colophons of the Sanskrit
versions of the Commentary on the Compendium of the Abhidharma and the Explication of the
Compendium of the Abhidharma mention the author. However, both commentaries are ascribed
to *Jinaputra in Tibetan sources. At the same time, the Chinese tradition attributes the compilation
of the Explication of the Compendium of the Abhidharma to Sthiramati. The idea that Sthiramati
was involved in the production of the Explication of the Compendium of the Abhidharma appears
possible, insofar as the Commentary on the Compendium of the Abhidharma (or the Explication
of the Compendium of the Abhidharma) plays a central role in his Commentary on the Treatise on
the Five Constituents (Kramer 2015, 295–98). However, it is not very probable that he composed
the Commentary on the Compendium of the Abhidharma itself, as this commentary seems to be
older than the Commentary on the Treatise on the Five Constituents (see Kramer 2016a, 50).
The same person who authored the Commentary on the Treasury of Abhidharma [Called]
True Reality (as well as the Commentary on the Treatise on the Five Constituents and the Com-
mentary on the Treatise on Mental Presentation in Thirty Verses) may have also composed a no
380
Sthiramati
longer extant commentary on the Collection of Means of Knowledge. This has been suggested
by Kazunobu Matsuda on the basis of a remark to be found in the Commentary on the Treas-
ury of Abhidharma [Called] True Reality (Matsuda 2013, 52). There the reader is referred to
a commentary on the Collection of Means of Knowledge (Pramāṇasamuccaya-upanibandha).
A similar term, namely Pañcaskandhaka-upanibandha, is used in the Commentary on the
Treatise on Mental Presentation in Thirty Verses when Sthiramati refers the reader to his Com-
mentary on the Treatise on the Five Constituents. Therefore, Matsuda suggests understanding
Pramāṇasamuccaya-upanibandha in a parallel way, that is, in the sense of a commentary on
the Collection of Means of Knowledge composed by Sthiramati himself.
In general, the author of the Commentary on the Treatise on the Five Constituents and the
Commentary on the Treatise on Mental Presentation in Thirty Verses appears to be a well-
educated and innovative scholar. He very often offers philosophically complex explanations
of high quality and not just definitions of individual terms. His comments are concise and
not repetitive and rarely contain redundant sentences and phrases. Unlike the author of the
Commentary on the Comments to the Ornament of [Mahāyāna] Sūtras, he hardly gives any
examples for the concepts he is explaining. As the commentarial style of the Commentary on
the Comments to the Ornament of [Mahāyāna] Sūtras differs strongly from that of the Com-
mentary on the Treatise on the Five Constituents and the Commentary on the Treatise on Men-
tal Presentation in Thirty Verses, it appears doubtful that the Commentary on the Comments
to the Ornament of [Mahāyāna] Sūtras was composed by the same author. The explanations
of the Commentary on the Comments to the Ornament of [Mahāyāna] Sūtras, a rather strict
word-by-word commentary, are philosophically less independent and elaborate, are often
much shorter than the comments in the Commentary on the Treatise on the Five Constituents
and the Commentary on the Treatise on Mental Presentation in Thirty Verses, and are partly
repetitive. The author of the Commentary on the Comments to the Ornament of [Mahāyāna]
Sūtras appears less creative and inspired than the commentator of the other two texts.
The authorship of the Commentary on the Kāśyapa Chapter and the Commentary on the
Mahāyāna Madhyamaka remains doubtful and requires further examination of the contents and
style of these texts.13 The fact that comments on the Kāśyapa Chapter (Kaśyapaparivarta) that
are found in Sthiramati’s Commentary on the Distinguishing of the Middle from the Extremes dif-
fer strongly from parallel explanations in the Commentary on the Kāśyapa Chapter as well as the
Commentary on the Kāśyapa Chapter‘s rather early translation into Chinese (between 508 and
535 by Bodhiruci) raise doubts that the same author wrote both texts (see also Kramer 2016a, 50).
The Commentary on the Exposition of Akṣayamati (composed by Vasubandhu, according to the
Tibetan tradition) has been attributed to Sthiramati by Jens Braarvig on the basis of some passages
in the Commentary on the Exposition of Akṣayamati that have parallels in parts of the Commentary
on the Comments to the Ornament of [Mahāyāna] Sūtras, the Commentary on the Treatise on the
Five Constituents, and the Commentary on the Treatise on Mental Presentation in Thirty Verses
(Braarvig 1993, cxxiii and cxxviii). The parallels referred to by Braarvig all deal with explana-
tions of the mental (caitasika) factors, which indeed seem to be related in the three commentaries.
We do not know with certainty whether this fact proves that the Commentary on the Exposition of
Akṣayamati and the two other commentaries were composed by the same person (since the possi-
bility cannot be ruled out that the author of the former simply copied the definitions from the latter).
Innovation and Original Contribution

The overwhelmingly commentarial nature of Sthiramati’s work makes pointing out his contri-
bution as an independent thinker – beyond mere local innovations – something of a challenge;
381
yet, as I hope to demonstrate in this section, it is an undertaking well worth pursuing. His
reliance on a great number of sources, from which he quotes often and freely; his habit of
juxtaposing a spectrum of views (often without stating the preferred one); and a general Indian
scholastic inclination to tone down innovation in the name of tradition – all these make it hard
to pinpoint his own views and indeed give the impression of works that in their greater part
are meant to explicate rather than innovate.14 And yet a close reading of his oeuvre, one that is
attuned to the accumulative and dispersed manner in which his ideas appear as the commen-
tarial genre dictates, reveals not just striking and innovative insights but also a substantive and
profound underlying philosophical outlook.
One distinct display of this outlook and of Sthiramati’s innovativeness as a thinker is found
in his commentary on the opening verse of Vasubandhu’s Treatise in Thirty Verses (Triṃśikā).
Here, arguably, Sthiramati presents what may be seen as a broadly conceived theory of mean-
ing – one that is applicable, in his words, “both in the world and in texts” – according to which
all language use is figurative (upacāra). The term upacāra, it should be noted, is mentioned by
Vasubandhu in the Triṃśikā only once. Sthiramati’s disproportionally lengthy commentary on
this topic indicates the significance he accords to this issue but also suggests that he seems to
see this as an opportunity to set the philosophical framework for venturing into the Triṃśikā‘s
unfolding claims and, by extension, for the Yogācāra worldview.
To achieve this, Sthiramati incorporates ideas introduced by his predecessors and con-
temporaries, both Buddhists and non-Buddhists, into a novel philosophical synthesis that ties
together the Yogācāra understanding of the causal activity of consciousness with a linguistic
theory of sense. Space does not allow me to delve into the question of his verified and possible
sources of influence for making these claims, as well as into the textual and philological sup-
port for their interpretation (I have dealt with these extensively elsewhere,15 and the reader is
referred to these sources in the notes). Instead, what follows is more of a speed-tour through
the essentials of Sthiramati’s arguments, with an eye to the extent and nature of their innova-
tiveness and their ramifications for and overall contribution to Yogācāra thought.
Vasubandhu’s opening verse of the Triṃśikā is as follows:
The metaphor (upacāra) of ‘things’ (dharma) and of ‘self’ in its various forms, which
is set in motion, [Sthiramati: “that is to say in the world and in treatises”], that is with
reference to the transformation of consciousness.
(Buescher 2007, 43)16
Sthiramati’s ensuing commentary does many kinds of philosophical work, much of it ventur-
ing beyond the focus or scope of the root verse. Put briefly, the commentary connects a rather
formulaic scholastic understanding of metaphor – under a more general discussion of theories
of meaning – with the Yogācāra model of the activity of consciousness, that is, its “transforma-
tion,” and then utilizes this model to refute two “extreme views” (one referring apparently to
the Madhyamaka), culminating with a view of all language use as figurative.
Sthiramati begins his commentary by clarifying that the verse uses the terms “self ” and
“things” (“dharma”) as generic terms for all substantive entities (Buescher 2007, 40:4–8; Tzo-
har 2018, 157 n. 6). But why are these referred to as “metaphors”? His explanation points out
that the “self ” and “things” are mere mental appearances (nirbhāsa), understood under the
Yogācāra view by which all phenomenal objects do not exist as they appear and can be known
or discussed solely in terms of mental representations (vijñapti). The commentary clarifies
that whereas in ordinary language, we understand the words “self ” and “things” to refer to
objectively existent phenomena, in fact we can only actually relate to the constructed mental
382
Sthiramati
appearances of a self and of things (which appear as if they were objectively and externally
existent; the paradigmatic analogy here is to an eye disease, timira, that causes one to see
nonexistent objects). These mental appearances, Sthiramati further explains, are an outcome
of the transformation of consciousness (vijñānapariṇāma), defined by him as a causal mental
development, identical to the Buddhist notion of dependent arising ( pratītyasamutpāda; in the
following, I provide a more detailed explanation of this term).17
Sthiramati then proceeds to explain what renders the words “self ” or “things” metaphors.
For this purpose, he uses a formulaic śāstric definition of metaphor according to which a word
is used figuratively when it indicates “something that is not there,” or, in more formal terms,
when a word’s primary referent is absent from the locus of reference.18 So, in the example
used by Sthiramati (Figure 22.1, right-hand side), when a person of the Bāhīkas is called an
“ox” (because he is similarly strong, slow witted, etc.), the word “ox” directly denotes its
primary referent (an ox, whether as an individual or an implicit generic property), which is,
however, absent from the locus of reference (i.e., from that which the word actually describes:
the Bāhīka person); therefore, the word is said to designate the Bāhīka person only in a second-
ary, figurative way.
Accordingly, the same applies to the terms “self ” and “things” (Figure 22.1, left-hand side):
they are figurative because their primary referents (the self, things) are absent from their locus
of reference, identified here as the “transformation of consciousness.” However, unlike in the
śāstric stock example, “self ” and “things” – like “ox,” or any other word for that matter – are
always mere figures, since according to the Yogācāra, they never exist in the manner in which
they appear (being mere mental appearances) and hence are always absent from their locus of
reference. Furthermore, because the self and things, we are told, stand for the entire range of
subjective and objective phenomena, all of which do not exist as they appear, it follows – as
Sthiramati asserts explicitly – that all language use is necessarily figurative, and hence his
theory can be characterized as “pan-figurative.”
Bracketing for now the ramifications of this view, we may ask to what extent was it
grounded in the work of Sthiramati’s predecessors and to what extent is it innovative, and
more specifically, what are the particular points of influence and difference between his works
In Ordinary Language
“Bāhīka” (as well as
“self” and “things”, “Ox”
“ox”, etc.) Figurative-indirect
denotation
Figurative-indirect
denotation Direct
denotation
Transformations of
Consciousness
(causal, mental)
An appearance Direct
of a Bāhīka denotation
person “Bāhīka”
Figure 22.1 Sthiramati’s pan-metaphorical claim in the Triṃśikābhāṣya.
383
and those of his predecessors? First in line, of course, is the question of how Sthiramati’s ideas
stand in relation to those of Vasubandhu. As I mentioned previously, the term upacāra in the
sense of figurative application appears only once in the Triṃśikā. It does figure prominently in
Vasubandhu’s Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya), though
there, it is implemented mostly as a hermeneutical device for indicating assertions or phrases
that need to be interpreted as carrying an implicit intent. Within this largely hermeneutical
context, however, some cases in the Abhidharmakośabhāṣya seem to consider that named
objects may be figures insofar as they are underlined by mental events (Gold 2015, 268 n.47),
and in this restricted sense, Sthiramati may perhaps be explicating what is already implicit
in Vasubandhu’s work. Nonetheless, it is only in Sthiramati’s commentaries – not just on the
Triṃśikā but also on the Abhidharmakośabhāṣya – that figures are explicitly tied, theoretically
speaking, to their underlying causal mental phenomena (Tzohar 2018, 126–36).19 Another
important difference – if not a full-blown theoretical leap – between Sthiramati’s account and
Vasubandhu’s involves the way in which the former develops the discussion of figures beyond
the idea of involving a mere cognitive imposition of “things” and “self ” on the transformation
of consciousness into a full-fledged linguistic figurative theory of meaning.
Sthiramati facilitates this move by explicitly and consciously situating his arguments about
metaphor against the background of a non-Buddhist Indian philosophical conversation on
denotation (in which metaphor is distinctively linguistic),20 culminating his own discussion
with some general observations on language.21 Within this framework, metaphors serve as
much more than mere indications of a cognitive mistake, as they are utilized to explain how
meaning is created within language.
But why was this linguistic context so important to Sthiramati? A possible explanation
is found when we consider his particular position within the development of the Yogācāra
textual tradition. Relative to the earlier textual corpus – ascribed to Asaṅga, Maitreya, and
Vasubandhu – Sthiramati operated under a much more defined notion of the Yogācāra as a
textual tradition and a doxographically distinct school. This position brought with it a new set
of philosophical challenges in need of addressing, a dynamic that is distinctly expressed in
Sthiramati’s rebuttal – at the heart of his exposition of metaphor – of two “extremist” views
(ekāntavāda), one of which is ostensibly that of the Madhyamaka school.
For his move against the Madhyamaka, Sthiramati, putting new wine into old wineskins,
unmistakably paraphrases an argument originally found in the Bodhisattvabhūmi (Dutt 1978,
31 ll. 16–20), which in turn has been cited and explicitly identified by Bhāvaviveka as a
Yogācāra critique of the Madhyamaka (Tarkajvālā on the Madhyamakahṛdayakārikāḥ 5.82–
5.83ab; Eckel 2008, 281–82, 431–32). Arguably aware of this polemical history, Sthiramati
uses this argument to take the Yogācāra debate with the Madhyamaka to the next stage, draw-
ing on his figurative theory of meaning.22
To understand how his argument proceeds, we need first to clarify in more detail how Sthi-
ramati’s pan-figurative claim is translated into a theory of meaning, and of what kind. Briefly put,
Sthiramati’s pan-figurative view manages to salvage the meaningfulness of ordinary language
by presenting what amounts to a figurative theory of sense (by which the meaning of language
is understood primarily as a function of its use; see Figure 22.2).23 According to Sthiramati,
all words are figurative because they denote objects only indirectly, referring directly to their
appearances in their capacity as the transformation of consciousness. As I mentioned, the trans-
formation of consciousness is not a static state of affairs but a dynamic causal process – indeed, it
is the very Buddhist notion of dependent origination as a causal nexus underlying all phenomena.
According to this scheme, a word designates not an object but rather the entire causal mental
process by virtue of which the object appears in consciousness. Accordingly, the reference of the
384
Sthiramati
Meaning of a word as function

of its use. Knowledge of the
use pertains to the meaning. “Pot”
(“Transformations of Sense
Consciousness so there is
an appearance of an
external pot along with its
designation”)
Transformations of Reference
Consciousness so
there is an
appearance of an
external pot along
with its designation
Figure 22.2 The Yogācāra figurative theory of meaning.
word “person” is the underlying mental causal nexus that brought it about (its mental appear-
ance), while the sense of the word is roughly the description of this causal process.24
The benefits for the Yogācāra from this creative understanding of meaning are multiple.
First, it creates a kind of scholastic meta-language that “cleanses” ordinary discourse of its
essentialist and reifying quality by understanding it in terms of causal mental descriptions – but
does so without changing its basic vocabulary. Second, this framework allows the Yogācāra to
distinguish between different levels of discourse and posit a hierarchy of meaning within the
conventional realm of language. This last point requires some elaboration.
Recall that under the Yogācāra theory of meaning, what a term comes to mean is consti-
tuted by an underlying causal description. These causal descriptions, however, are not static.
As seen, for instance, in the Triṃśikā, causal mental explanations of phenomena presuppose
an evolving description, one that increases in intricacy and subtlety and which embodies a
variety of points of view on the continuum between the ordinary, deluded point of view and
the complete, direct knowledge of a Buddha or an accomplished bodhisattva. So, for example,
while for the simpleton (Figure 22.3), the word “pot” refers to an objectively existent pot, for
an adept informed by śāstra and practice, it comes to mean the activity of the eight modes of
consciousness and so on, and this continues until an exhaustive direct knowledge of causal
relations (as the dependent nature) is reached, which, according to Sthiramati, marks the point
of view of an advanced bodhisattva.25
Note that this account renders meaning (1) dependent upon point of view and (2) hierar-
chically organized, since insofar as some speakers’ causal descriptions are more accurate or
insightful than others, their respective use of language is also more meaningful.26 Thus while
the Yogācāra accept that all discourse is indeed conventional, its figurative nature and objec-
tive causal underpinning imply that some claims are more meaningful than others.
This layered view of discourse works well, structurally speaking, with both the Yogācāra
soteriology and system hermeneutics,27 but in terms of our present concern, it serves Sthiramati
primarily by enabling him to defend the meaningfulness of Yogācāra metaphysical claims from
the threat posed by Madhyamaka radical conventionalism.28 Sthiramati’s line of defense is to
point out that the latter’s view of language as a self-referential closed system (1) is incoherent
385
“Pot”: direct
knowledge of all
causal relations that
brought bout the
appearance,
“Pot”: “Transformations of under the
Consciousness so there is tatpṛṣṭhalabdhajñāna
an appearance of a pot
along with its designation”
“Pot”: Externally
objectively
existent pot
Figure 22.3 The Yogācāra conception of discourse.
and (2) carries a heavy price tag for Buddhists. First, while language does not truly represent
reality in any respect, its denotative function and the very conditions of its meaningfulness
require us to presuppose extra-linguistic underlying conditions as its “support.” This is because
language, insofar as it is meaningful, must be intentional – words have to be about something;
that is, they have to operate within a certain referential framework that distinguishes between
designations and their alleged extra-linguistic referents. This does not mean that the referents
of words are real or exist as they are designated; but it does suggest that the very principle of
conceptual differentiation allowed by referential language – the principle according to which
certain names designate certain objects and not others – hinges on the presence of underly-
ing extra-linguistic restricting conditions. In light of this claim, the Madhyamaka position
undermines the very conditions that render the school’s own discourse meaningful – thereby
implicating it in a pragmatic contradiction. The Yogācāra critique of the Madhyamaka stance,
therefore, is not (as it has often been interpreted) that it is nihilistic but that it is senseless.29
Second, though it may be polemically powerful, the Madhyamaka conventionalism entails
a flattening of conventional discourse insofar as this view renders all claims – whether they are
about ultimate reality or about medium-sized dry goods – equally meaningful (since they are
all referenceless and hence empty). Given this, the Madhyamaka view of language is patently
self-undermining also because it cannot provide any criterion in light of which the meaning-
fulness of their own claims might be distinguished from those of any other conventional lan-
guage use (including, for instance, uninformed, naïve assertions about the “self ” or “things”).
This conception of conventional discourse, however, cannot serve the Yogācāra, which pre-
sents explicit metaphysical claims – about an inexpressible underlying ontology and its causal
mental modifications – and therefore needs to defend the meaningfulness of this discourse,
that is, to defend the possibility of describing reality in a way that is not merely negative. This
need is met by the school’s figurative theory of meaning, which allows it to present a notion
of meaning and discourse according to which some descriptions of reality, although merely
conventional, are more meaningful than others. Apart from responding to the Madhyamaka
critique, this allows the Yogācāra to carve out a position that is exceptional in the Buddhist
386
Sthiramati
landscape: one that views ordinary language as incapable of representing or reaching reality
but at the same time justifies the meaningfulness of the school’s own metaphysical discourse.
The significance of Sthiramati’s figurative theory of meaning, however, does not end with
this and carries broader ramifications for our understanding of the Yogācāra conception of expe-
rience and of the school’s soteriology. In the remainder of this section, I will point out, if only
schematically, some deep structural affinities between Sthiramati’s understanding of linguistic
meaning and his understanding of experience, particularly of intersubjective experiences of the
external word as well as the experience of the enlightened mind while it abides in saṃsāra.
This allows us to identify several fundamental themes running through his thought that, taken
together, imply a broadly conceived theory of meaning that is not merely linguistic but also
perceptual and is designed to address some of the fundamental concerns of the Yogācāra.30
An important benefit of the figurative theory of meaning as characterized previously is that
it makes it possible to account for successful communication between diverse speakers – for
instance, between an ordinary person and an enlightened one – despite deep discrepancies in
their understanding. Specifically, this is enabled by seeing conventional discourse as incorpo-
rating different modes of language use, layered in accordance with their respective standing
vis-à-vis knowledge of causality. Above all, this ability turns on an important feature that the
figurative theory of meaning shares with contemporary causal theories of reference, namely
that it allows for the reference of a term to remain fixed even as its meaning changes from
one speaker to another. As such, it overcomes the danger of incommensurability – namely the
relativistic claim that vast differences in our conceptual frameworks entail that the meaning of
a common term (and insofar as meaning determines it, also its reference) may radically change
between frameworks.
Within the Yogācāra tradition, a version of the challenge of incommensurability – in a lin-
guistic but also in a broader experiential sense – is palpably present in the need to account for
the way in which beings at different ends of the Buddhist path (say, a fulfilled bodhisattva who
is required to operate within saṃsāra and a simpleton) can converse and interact in a meaning-
ful way despite the epistemic abyss that separates them. This issue is very clearly expressed in
the Yogācāra discussions of how to bridge the bodhisattva’s nonconceptual (nirvikalpa) expe-
rience of reality on the one hand and her use of ordinary language within the realm of saṃsāra
on the other. To this end, various Yogācāra thinkers, Sthiramati included, introduced a division
of labor between two types of knowledge: nonconceptual or nondiscriminatory awareness
(nirvikalpajñāna), and the “awareness obtained subsequent to it” (tatpṛṣṭhalabdhajñāna) – a
different kind of insight that the advanced bodhisattva is said to attain following nonconcep-
tual awareness. Sthiramati explicitly identifies the latter awareness with the dependent nature
and as the state in which ultimate knowledge of causality is attained.31 This knowledge of
causality, we are told by various Yogācāra sources, is what enables the bodhisattva to remain,
operate, and communicate effectively in the phenomenal world. This awareness appears to
enable a unique kind of conceptual (!) activity, which is, however, very different from ordinary
conceptual awareness insofar as it is non-reifying (Tzohar 2018, 183–88; Tzohar 2020).
Note here the similarities with Sthiramati’s figurative theory of meaning. In both cases, a
certain epistemic/linguistic inherent gap – which is presupposed by the Buddhist worldview or
path – is bridged by appealing to an underlying casual reality and by positing the possibility of
conceiving of this underlying reality from a number of perspectives simultaneously (i.e., the
dependent nature, seen from the side of the imagined in essentialist terms and from the side of
the perfected without essentializing or reifying).
The structural affinities between these schemes are quite striking, and once we discern them,
they seem to be at work in other instances of Yogācāra accounts of experience as well. A distinct
387
example is the school’s – and pronouncedly Sthiramati’s – overall explanation for the possibil-
ity of intersubjective experiences under a representation-only framework (i.e., without accept-
ing externality as objectively existent). Briefly put, these experiences are explained by first
reducing them to an underlying mental causal nexus and then tracing the distinction between
intersubjective and private kinds of experiences to various common and uncommon causal
(karmic) strands, respectively. The focus here is indeed perceptual, but the governing principle
is the same as in the case of linguistic meaning – in both cases, the reduction to causal and men-
tal reality establishes a shared constructed realm of experience while still accounting for acute
differences and discrepancies between each individual experience (Tzohar 2017c, 343–45).
An analysis of Sthiramati’s treatment of these instances shows them all to be founded on
a similar analytic structure. In all these cases, both agreement and discrepancy, the shared
and the private, are explained in causal mental terms, yielding what can be described as a
broadly conceived understanding of meaning – in the experiential and existential senses of the
word – that is not merely linguistic but also perceptual. Common to all these accounts is an
understanding of meaning as inevitably relational, shaped by the confluence of perspectives
of a community of speakers and experiencers and implying a view of language as inherently
polysomic and of experience as inevitably perspectival. This understanding of meaning, I have
argued, is utilized by Sthiramati for various purposes: to defend the meaningfulness of the
Yogācāra metaphysical discourse, to show the possibility of intersubjective experiences under
a “mere representation” view, and to allow the school to overcome its soteriological version
of the problem of incommensurability – that is, of how a bodhisattva may operate in saṃsāra
and in particular how her nonconceptual experience and everyday experience and language
use may be bridged (Tzohar 2018, Chapter 6).
Seen against this background, Sthiramati’s lengthy commentary on the opening verse of the
Triṃśikā and his account of metaphor emerges as more than just a local philosophical move: it
functions as a broad strategy devised, rather ingeniously and innovatively, to deal with broader
and more fundamental Buddhist concerns. Venturing beyond merely clarifying a case of cog-
nitive or linguistic imposition (these are intertwined, his analysis tells us), his figurative theory
of meaning introduces what can be seen as a “master” template for the Yogācāra concep-
tion of meaning – one that is applicable “in the world and in texts” and one that he applies
throughout his commentary and develops in other works as well. Set within this context and
the Yogācāra’s overall understanding of experience and meaning, his pan-metaphorical claim
emerges as not merely internally coherent and polemically effective but indeed essential to the
school’s salvific vision.
Notes
1 Chinese: Anhui (安慧), Jianhui (堅慧) (Silk 2015, 150–51); Tibetan: Blo gros brtan pa.
2 See Tzohar 2017b for a discussion of Sthiramati’s conception of the Yogācāra identity as a distinct
school in the Madhyāntavibhāgaṭīkā, as reflected in his use of commentarial techniques for support-
ing a certain traditional narrative of the Yogācāra lineage of textual transmission.
3 See especially the writings of the Chinese pilgrim Xuanzang (玄奘, 600?–664) and the Tibetan histo-
rians Bu ston (1290–1364) and Tāranātha (1575–1634).
4 Diskalkar 1925, 37–40; Silk 2009, 385, n. 13. See also Kramer 2019, 456.
5 See Kramer 2015, 330–47. In some cases, Sthiramati presents the definitions in an abbreviated form
in the Commentary on the Treatise in Thirty Verses, probably extracting only those phrases from
his Commentary on the Treatise on the Five Constituents that appeared most relevant to him. But
even though the text is not identical in these instances, it is still obvious that there is a direct relation
between the comments (see also Kramer 2015, 300ff.).
388
Sthiramati
6 For studies on the Commentary on the Treasury of Abhidharma [Called] True Reality, see, for exam-
ple, Matsuda 2013; Kazuo and Kramer 2020; Odani 2021.
7 Until recently, the text of the Commentary on the Treasury of Abhidharma [Called] True Reality was
only available in a rather bad Tibetan translation and some Chinese and Uyghur fragments. One of
the notable aspects of the Sanskrit manuscript of the Commentary on the Treasury of Abhidharma
[Called] True Reality is that it seems to be one of the very oldest manuscripts preserved in Tibet. It
has been dated to the eighth or ninth century by Kazunobu Matsuda based on its script. For more
details, see Kazuo and Kramer 2020.
8 For details, see Kramer 2013a.
9 For details, see Kramer 2008.
10 See Kramer 2013b, 89. However, it should be noted that he does not use the term vijñaptimātratā
explicitly.
11 The four arguments may be summarized as follows (Kramer 2016b, 146f.): (1) Actual percep-
tion ( pravṛttivijñāna) reappears after a person has risen from unconscious states; (2) actual per-
ceptions appear in different modes ( prakāra) depending on different kinds of object conditions
(ālambanapratyaya); (3) actual perception reappears after it has been interrupted by sleep (middha)
or a swoon (mūrchā); (4) without the store mind, an individual could not (a) arise in and (b) be liber-
ated from saṃsāra, because, on the one hand, the process of rebirth would not be possible and, on the
other, the contaminations (kleśa) could not be removed.
12 For a detailed comparison, see Kramer 2016b, 146f.
13 Nguyen presents several arguments that from his point of view make it plausible that only one Sthi-
ramati composed all but the tantric works attributed to him in traditional sources (1990, 23–83).
14 See Delhey 2016, 16ff. for an assessment of Sthiramati’s commentarial approach and style.
15 The present chapter is based on arguments that I elaborate on in greater detail in Tzohar 2017a,
2017b, 2017c, 2018.
16 All translations from the Triṃśikā and Triṃśikā-bhāṣya are my own. Translations from the first two
are from Buescher’s Sanskrit and Tibetan critical edition (2007).
17 “And this metaphor of two kinds (of self and things] is applied in reference to the transformation of
consciousness and not to a primary actual self or things. Why so? Since these do not exist outside of
the transformation of consciousness. What is it that is termed ‘transformation’? The state of being
otherwise than what it previously was. Here ‘transformation’ – which is the acquisition of the effect
as something distinct from the moment in which the cause exists and simultaneous with the moment
of the passing away of the cause – is a mental construct, an appearance of self, etc., and the appear-
ance of matter, etc., all arising from the storehouse consciousness because of the incubation of latent
impressions through the conceptual construct of a self, etc., and the incubation of latent impressions
through the conceptual construct of matter, etc. Hence, the metaphor of things such as matter, etc.,
and the metaphor of self, etc., having its base in the appearance of matter, etc., and the appearance
of self, etc., which due to a mental construction appear as if they were external, occurs from time
immemorial, even without an external self and things, like in the case of the metaphor of net-like
apparitions experienced by a person suffering from an eye-disease. And that which is not existent
there in the locus of reference is figuratively designated, like when one calls the Bāhīka person an
ox. . . . And thus it should be accepted that all objects are in fact unreal, due to their intrinsic nature
being a fabrication. Consciousness, on the other hand, because of its being dependently originated,
exists as a substrate (dravyataḥ). And again, consciousness’s state of being dependently originated is
made known by the use of the term transformation’ ” (Tzohar 2018, 157–59).
18 This principle, which is constitutive of Sthiramati’s understanding of metaphor, is in fact based on the
formula of the Nyāya-sūtra’s 2.2.61. . . . atadbhāve ’pi tadupacāraḥ / Tarkatirtha (1936–1944, 662).
19 A possible source for Sthiramati’s conception of this relation may be found in the Laṅkāvatārasūtra.
See Tzohar 2018, 137–44.
20 Thus, in a lengthy section, Sthiramati pits his arguments about secondary denotation against the argu-
ments of opponents from non-Buddhist schools and alludes clearly (perhaps even with parodic intent)
to similar discussions in the Nyāya-sūtra and in the works of grammarians (Buescher 2007, 46:3–
48:11; Tzohar 2018, 161–66). Furthermore, Sthiramati’s account of metaphor appears to owe greatly
to Bhartṛhari’s discussion of the issue, to which it shows close thematic as well as stylistic similarities
(Tzohar 2018, Chapter 2). It is also possible that Sthiramati’s knowledge of these polemics was gained
through the mediation of Dignāga’s works (especially the fifth chapter of the Pramāṇasamuccaya;
389
see Tzohar 2018, 144–49). Finally, an early and important Yogācāra source of influence on Sthirama-
ti’s thought, one that discusses upacāra in a distinctively linguistic polemical context, is the Chapter
on the Meaning of Reality (Tattvārthapaṭalaṃ) of the Levels of the Bodhisattva (Bodhisattvabhūmi)
and its accompanying sections in the later Collection of Clarifications (Viniścayasaṃgrahaṇī), both
included in the Levels of Spiritual Training (Yogācārabhūmi) corpus. See Tzohar 2018, Chapter 3.
21 “Thus, since there is no correlation of an object with a word, language and knowledge do not exist. Thus,
because of the non-existence of an expression and of a thing to be expressed, a primary object does not
exist. And all that is merely figurative (gauṇa) and not a primary meaning” (Buescher 2007, 48:6–8).
22 “And because of the impossibility of a metaphor without a support (nirādhāra), the occurrence of the
metaphors of the self and things should be necessarily regarded as actually being the transformation
of consciousness. Due to this, the following [Mādhyamika’s] view would not withstand reasoning:
‘That the object of consciousness, just like consciousness, exists only conventionally (saṃvṛtitaḥ)
and not ultimately ( paramārthataḥ).’ This view is not logically tenable since it would result in the
nonexistence of both of them even conventionally, because convention without a basis (upādāna)
does not make sense. In light of this, the extremist views of both types must be abandoned because
they are illogical, so said the ācārya [Vasubandhu].” (Buescher 2007, 42:9-15)
23 An alternative and viable reading takes it as a figurative theory of reference, whose main purpose is a
Yogācāra defense against the realist attack on its supposed Mentalism. See Gold 2007.
24 Sthiramati provides a dual characterization of the transformation of consciousness – as an underly-
ing causal reality on the one hand and a conceptual construct on the other (for the latter, see Bue-
scher 2007, 108:4–109:13; Tzohar 2018, 166–69) – suggesting a basic distinction between real causal
activity and its imagined products, or, with an eye to their philosophical presentation, between inef-
fable real causal facts and constructed casual descriptions.
25 According to Sthiramati, such knowledge is attained by the bodhisattva under a “subsequent mundane
pure awareness” (tatpṛṣṭhalabdhaśuddhalaukikajñāna), which follows from his previous nonconcep-
tual transcendent awareness (nirvikalpalokottarajñāna). See Sthiramati’s commentary on Triṃśikā
verse 22; Tzohar 2018, 180–82.
26 Note that under this framework – this figurative theory of sense – discourse is layered in terms of its
degrees of meaning but not in terms of its truth value. This is because, within this framework, the
truth of propositions was understood to be a function of the relations between expressions and their
alleged referents. The latter’s ultimate non-existence on the one hand and the utter conventionality of
discourse on the other hand render all propositions false from the ultimate point of view. However,
the picture of conventional discourse as incorporating different modes of language-use, layered in
accordance with their respective standing vis-à-vis knowledge of causality, results in some proposi-
tions being more meaningful than others. See also Tzohar 2017b, section 2.
27 A common feature to all these schemes is their progressional structure: they all facilitate a dialecti-
cal movement of understanding through which erroneous accounts of the same reality are gradually
replaced by more accurate ones, up to the final knowledge of a Buddha, thereby enabling the school
to correlate structurally between more accurate levels of discourse on the one hand and the gradual
progress along the path on the other.
28 See note 22 for Sthiramati’s argument. It presents a set of parallel observations: the transformation of
consciousness is to the metaphors of “self ” and “things” what a “support” is to a linguistic metaphor
and what a “basis” (upādāna) is to the “conventional.” To a large extent, the understanding of Sthi-
ramati’s critique of the Mādhyamika turns on what we take him to mean by these parallel observa-
tions and these specialized terms. For an analysis of this argument, see Tzohar 2018, 173–77.
29 This reading of Sthiramati’s argument is further reinforced by Vinītadeva’s later explanation of this
passage in his sub-commentary, the Sum Cu Pa’i ’grel Bshad (Triṃśikā-ṭīkā*) TD. 4070 6b.3–6.
30 The essence of this move appears in Tzohar 2018, Chapter 6.
31 Sthiramati on the Triṃśikā verse 22 (Buescher 2007, 126:7ff.). For the translation and analysis of this
section, see Tzohar 2018, 180–82.
Bibliography
Braarvig, Jens. 1993. Akṣayamatinirdeśasūtra. Oslo: Solum.
Buescher, Harmut. 2007. Sthiramati’s Triṃśikāvijñaptibhāṣya. Wien: Verlag der Österreichischen Akad-
390
Sthiramati
Bühler, Georg. 1877. “Further Valabhî Grants.” Indian Antiquary 6: 9–21.

Cordier, Palmyr. 1909. Catalogue du fonds tibétain de la Bibliothèque Nationale. Deuxième partie: Index
du Bstan-ḥgyur (tibétain 108–179). Paris: Imprimerie Nationale.
Delhey, Martin. 2016. “The Indian Yogācāra Master Sthiramati and His Views on the Ālayavijñāna Con-
cept.” International Association for Buddhist Thought and Culture 26: 11–35.
Derge, T. 199 – ?. 1986. Bstan ’Gyur Sde Dge’i Par Ma: Commentaries on the Buddha’s Word by Indian
Masters (Electronic Edn, 11 Cds). [Reproduced from Editions of the Individual Sections Published in
Delhi at the Delhi Karmapae Chodhey, Gyalwae Sungrab Partun Khang, Itself a Reproduction from
Prints from the 18th Century Sde-Dge Blocks], 213 vols. New York: Tibetan Buddhist Resource Center.
Diskalkar, D. B. 1925. “Some Unpublished Copper-Plates of the Rulers of Valabhī.” Journal of the Bom-
bay Branch of the Royal Asiatic Society 1: 13–64.
Dutt, N. 1978. Boddhisattvabhūmiḥ, 2nd ed. Patna, India: K. P. Jayaswal Research Institute.
Eckel, M. D. 2008. Bhāviveka and His Buddhist Opponents. Harvard Oriental Series. Cambridge, MA:
Harvard University Press.
Frauwallner, Erich. 1961. “Landmarks in the History of Indian Logic.” Wiener Zeitschrift für die Kunde
Süd- und Ost-Asiens 5: 125–48.
Gold, Jonathan C. 2007. “Yogācāra Strategies Against Realism: Appearances (Ākṛti) and Metaphors
(Upacāra).” Religion Compass 1 (1): 131–47.
———. 2015. Paving the Great Way: Vasubandhu’s Unifying Buddhist Philosophy. New York: Columbia
University Press.
Kazuo, Kano, and Jowita Kramer. 2020. “The Fourth Chapter of the Tattvārthā Abhidharmakośaṭīkā:
on Forbidding Intoxicating Liquor.” In Sanskrit Manuscripts in China III, edited by Birgit Kellner,
Xuezhu Li, and Jowita Kramer, 107–37. Beijing: China Tibetology Publishing House.
Kramer, Jowita. 2012. “Descriptions of ‘Feeling’ (Vedanā), ‘Ideation’ (Saṃjñā), and ‘the Unconditioned’
(Asaṃskṛta) in Vasubandhu’s Pañcaskandhaka and Sthiramati’s Pañcaskandhakavibhāṣā.” Rocznik
Orientalistyczny 65 (1): 120–39.
———. 2013a. “A Study of the Saṃskāra Section of Vasubandhu’s Pañcaskandhaka (with Refer-
ence to Its Commentary by Sthiramati).” In The Foundation for Yoga Practitioners: The Buddhist
Yogācārabhūmi Treatise and Its Adaptation in India, East Asia, and Tibet, edited by Ulrich Timme
Kragh, 920–70. Harvard Oriental Series. Cambridge, MA: Harvard University Press.
———. 2013b. “Notes on the Rūpa Section of the Pañcaskandhakavibhāṣā.” China Tibetology 21 (2):
86–99.
———. 2014a. Sthiramati’s Pañcaskandhakavibhāṣā. Beijing and Vienna: China Tibetology Publishing
House/Austrian Academy of Sciences Press.
———. 2014b. “Indian Abhidharma Literature in Tibet: The Section on Vijñāna in Sthiramati’s
Pañcaskandhakavibhāṣā.” In Buddhism Across Asia: Networks of Material, Intellectual and Cultural
Exchange, edited by Tansen Sen. Singapore: Institute of Southeast Asian Studies.
———. 2015. “Innovation and the Role of Intertextuality in the Pañcaskandhaka and Related Yogācāra
Works.” Journal of the International Association of Buddhist Studies 36/37: 281–352.
———. 2016a. “Some Remarks on Sthiramati and his Putative Authorship of the Madhyāntavibhāgaṭīkā, the
*Sūtrālaṃkāravṛttibhāṣya and the Triṃśikāvijñaptibhāṣya.” Buddhist Studies Review 33 (1–2): 47–63.
———. 2016b. “Some Remarks on the Proofs of the ‘Store Mind’ (Ālayavijñāna) and the Development
of the Concept of Manas.” In Text, Philosophy, and History: Abhidharma Across Buddhist Scholastic
Traditions, edited by Bart Dessein and Weijen Teng, 146–68. Leiden: Brill.
———. 2019. “Sthiramati.” In Brill’s Encyclopedia of Buddhism, Volume II: Lives, edited by Jonathan
A. Silk, 456–57. Leiden: Brill.
Matsuda, Kazunobu. 2013. “Sanskrit Manuscript of Sthiramati’s Commentary to the
Abhidharmakośbhāṣya.” China Tibetology 20 (1): 48–52.
Nguyen, Cuong Tu. 1990. “Sthiramati’s Interpretation of Buddhology and Soteriology.” PhD diss., Har-
vard University.
Obermiller, Eugene. 1931. History of Buddhism by Bu-ston (I. Part: The Jewelry of Scripture). Heidel-
berg: Harrassowitz.
———. 1932. History of Buddhism by Bu-ston (II. Part: The History of Buddhism in India and Tibet).
Heidelberg: Harrassowitz.
Odani, Nobuyicho. 2021. “An Introduction to Sthiramati’s Tattvārthā Chapter I.” In Illuminating the
Dharma: Buddhist Studies in Honour of Venerable Professor KL Dhammajoti, edited by Toshiichi
Endo, 295–306. Hong Kong: Centre of Buddhist Studies, The University of Hong Kong.
391
Sakuma, Hidenori. 2006. “On Doctrinal Similarities Between Sthiramati and Xuanzang.” Journal of the
International Association of Buddhist Studies 29 (2): 357–82.
Schiefner, Anton. 1868. Tāranāthae de Doctrinae Buddhicae in India Propagatione Narratio. St. Peters-
burg: Academia Scientiarum Petropolitana.
Silk, Jonathan A. 2009. “Remarks on the Kāśyapaparivarta commentary.” In Pāsādikadānaṃ: Fest-
schrift für Bhikkhu Pāsādika, edited by Martin Straube et al., 381–97. Marburg: Indica et Tibetica.
———. 2015. Buddhist Cosmic Unity: An Edition, Translation and Study of the Anūnatvāpūrṇatvanirde
śaparivarta. Hamburg: Hamburg University Press.
Tarkatirtha, T. N., and A. M. Tarkatirtha, eds. 1936–1944. Nyāyadarśanam with Vātsyāyana’s Bhāṣya,
Uddyotkara’s Vārttika, Vācaspati Miśra’s Tātparyaṭīkā, and Viśvanātha’s Vṛtti. Calcutta Sanskrit
Series. Calcutta, India: Metropolitan Print and Publishing House.
Tzohar, Roy. 2017a. “Does Early Yogācāra Have a Theory of Meaning? Sthiramati’s Arguments on Meta-
phor in the Triṃśikā-Bhāṣya.” Journal of Indian Philosophy 45 (1): 99–120.
———. 2017b. “Indian Yogācāra Understanding of Āgama-Pramāṇa.” In E. Freschi (guest ed.), “There
in only ‘Philosophy:’ The Case of Testimony”. Special issue of Kervan: International Journal of Afro-
Asiatic Studies 21: 261–77.
———. 2017c. “Imagine Being a Preta: Early Indian Yogācāra Approaches to Intersubjectivity.” Sophia
56 (2): 337–54.
———. 2018. A Buddhist Yogācāra Theory of Metaphor. Oxford and New York: Oxford University
Press.
———. 2020. “Contemporary Non-Conceptualism, Conceptual Inclusivism, and the Yogācāra View of
Language Use as Skillful Action.” Philosophy East and West 70: 638–60.
Yamaguchi, Susumu. 1934. Madhyāntavibhāgaṭīkā. Nagoya: Librairie Hajinkaku.
392
23
DEVENDRABUDDHI AND
ŚĀKYABUDDHI
Dharmakīrti’s First Commentators
Devendrabuddhi (fl. c. 650) and Śākyabuddhi (fl. c. 675) were the first two authors – apart from
Dharmakīrti himself – to compose commentaries on Dharmakīrti’s (fl. c. 625) Commentary on
the Means of Knowledge (Pramāṇavārttika, henceforth PV). Devendrabuddhi is not known
to have composed any additional works. Śākyabuddhi is listed as the author of two additional
texts in the Tibetan Tengyur, commentaries on the Ten Stages Sūtra (Daśabhūmika Sūtra,
Toh 3999) and the Summit of Mount Gayā Sūtra (Gayāśīrṣa Sūtra, Toh 3992); his authorship
of these commentaries is generally accepted. Although Devendrabuddhi’s and Śākyabuddhi’s
commentaries are distinct works, they are so closely intertwined that, for reasons which will
be elaborated upon in the following, it is best to consider them parts of the same “hybrid text.”
Tibetan tradition holds Devendrabuddhi to have been a direct disciple of Dharmakīrti (Chimpa
and Chattopadhyaya 1970, 239), and this is widely accepted by contemporary scholars as well.
The precise nature of the relationship between Śākyabuddhi and Devendrabuddhi is less certain,
which is to say that it is unclear whether or to what extent Śākyabuddhi may also have been taught
directly by Dharmakīrti or how closely Devendrabuddhi and Śākyabuddhi collaborated on their
closely intertwined commentaries. In other words, while Śākyabuddhi is typically portrayed as
the disciple of Devendrabuddhi, this should not necessarily be taken to mean that Śākyabuddhi
had no direct contact with Dharmakīrti. It should similarly not be assumed that Devendrabud-
dhi’s Explanation of the Commentary on the Means of Knowledge (Pramāṇavārttikapañjikā,
henceforth PVP) was composed strictly prior to and entirely independently of Śākyabuddhi’s
Commentary for the Commentary on the Means of Knowledge (Pramāṇavārttikaṭīkā, henceforth
PVṬ). That is, although it is common to regard “commentaries” as works composed indepen-
dently of and posterior to the “root text” upon which they are commenting, the commentarial
relationship is significantly more complex in the present case: at the very least, it must be under-
stood that the root text of the PV was never intended to be even grammatically intelligible in the
absence of some sort of commentary, and Devendrabuddhi may well have intended the PVP to
be similarly dependent upon the PVṬ for broader exegesis and analysis. As for the texts them-
selves, the PV is fortunately well attested in Sanskrit. Unfortunately, and by contrast, although a
few Sanskrit manuscript fragments of the PVP and PVṬ are still extant, for the most part, both
texts are currently only known to be extant in Tibetan translation.
In general, Devendrabuddhi’s PVP is written in an extremely close “word commentary”
(tshig ’grel) style, for the most part only clarifying the literal meaning of Dharmakīrti’s terse
DOI: 10.4324/9781351030908-34 393

Sanskrit, which on its own is frequently obscure to the point of sheer incomprehensibility.
Devendrabuddhi only very rarely ventures outside the bounds of providing semantic and gram-
matical glosses to Dharmakīrti’s root text. Śākyabuddhi’s PVṬ, by contrast, contains much
more by way of extrapolation and interpretation. But given the complex relationships among
Dharmakīrti’s PV, Dharmakīrti’s own Autocommentary for the Commentary on the Means of
Knowledge (Pramāṇavārttikasvopajñavṛtti; henceforth PVSV), Devendrabuddhi’s PVP, and
Śākyabuddhi’s PVṬ, these texts may be considered to constitute a type of aggregated “hybrid
text” (McClintock 2010, 2) rather than a series of isolated or independent works. In order to
understand this point, however, it is first necessary to examine these closely intertwined rela-
tionships among the four texts in question: the PV, PVSV, PVP, and PVṬ.
Commentarial Strata and the PV

The genre of South Asian scholastic commentarial literature is intertextual to an extraordinary
degree. Even the pretense of original work hardly exists; nearly all intellectual labor is per-
formed in terms of the twin projects of (1) commenting upon the predecessors in one’s own
tradition and (2) rebutting those in other traditions – and sometimes those in one’s own – who
have rebutted one’s predecessors in one’s own tradition and so on (Ganeri 2010). It is thus
a mistake, albeit a common one, to consider the underlying verses of the “root” (mūla) text
complete and self-contained with the commentary as a mere supplement – that is, something
superfluous, ultimately inessential, or strictly unnecessary, which is added to the purportedly
original unity of the root text. While it is certainly arguable that there exist examples of this
kind of relationship in South Asian literature, in the context of Dignāga and Dharmakīrti’s
works specifically, such a view would fundamentally mischaracterize the relationship between
the root verses and their commentaries.
Consider, for example, the relationship between Dignāga’s Encyclopedia of the Means of
Knowledge (Pramāṇasamuccaya, henceforth PS) and his Commentary on the Encyclopedia
of the Means of Knowledge (Pramāṇasamuccayavṛtti, henceforth PSV). The root verses of
the PS function more as a mnemonic aid, or as a skeleton to be fleshed out by the commen-
tary, than as a series of grammatically intelligible Sanskrit sentences – the verses primarily
serve to facilitate the memorization of arguments that are developed at least somewhat more
clearly in the commentary, and the two (verse and commentary) were undoubtedly composed
contemporaneously (Steinkellner 2005, IIIn1). To the extent that any “original unity” exists,
then, it is located not in the verses of the PS itself, but rather in the complex textual interplay
between the PS and the PSV: a “hybrid text” that may be designated as the “PS(/V).” The
relationship between Dharmakīrti’s PV and his autocommentary, the PVSV, should be under-
stood along these same lines. That is to say, in a manner precisely analogous to the PS(V),
the PV and the PVSV were deliberately composed as a hybrid text: the “PV(/SV).” The key
difference between the PS and PV in this regard is that Dignāga composed his own autocom-
mentary to each of the six chapters of the PS, while Dharmakīrti’s PVSV only deals with the
first chapter of his root text, PV 1.1
To repeat: Dharmakīrti’s Autocommentary for the Commentary on the Means of Knowledge
only comments upon the first chapter of his Commentary. Devendrabuddhi’s Explanation of
the Commentary on the Means of Knowledge thus defers to Dharmakīrti’s own Autocom-
mentary with respect to this first chapter. That is to say, Devendrabuddhi’s PVP does not com-
ment upon the first chapter of Dharmakīrti’s PV but rather deals only with its second, third,
and fourth chapters. Devendrabuddhi thus rhetorically positions his Explanation as occupy-
ing the exact same commentarial stratum as Dharmakīrti’s own Autocommentary, effectively
394
Devendrabuddhi and Śākyabuddhi
serving as an extension or equivalent of Dharmakīrti’s own contribution. And this is precisely

how Śākyabuddhi’s secondary commentary treats Devendrabuddhi’s work: Śākyabuddhi
comments on Dharmakīrti’s Autocommentary with respect to the first chapter of the PV and
Devendrabuddhi’s Explanation with regard to its second, third, and fourth chapters. In this
way, Śākyabuddhi’s methodological choice places Devendrabuddhi’s commentary on an equal
footing with Dharmakīrti’s Autocommentary. Like Devendrabuddhi’s methodological choice
to eschew composing a commentary on the chapter of the PV that Dharmakīrti had himself
already commented upon, this clearly indicates that Dharmakīrti’s immediate successors –
and quite possibly Dharmakīrti himself – considered the PVP the functional equivalent of a
commentary composed by Dharmakīrti himself. Hence, despite the relative lack of scholarly
attention paid to this intricate commentarial structure to date, in more or less the same way that
Dignāga’s PS and PSV should be regarded as a single (though bipartite) work, the PV, PVSV,
PVP, and PVṬ should similarly be regarded as a single (though hybrid) textual structure.
With this understanding in mind, we now turn to an examination of some key issues raised
by Devendrabuddhi and Śākyabuddhi in their respective commentaries on the PV. These points
are noteworthy both for their philosophical relevance in terms of drawing out the underlying
logic and broader implications of Dharmakīrti’s system and for their exemplification of the
commentators’ general hermeneutical approach.
Unmediated Instrumental Effect

As mentioned previously, Devendrabuddhi largely refrains from presenting additional content
on top of a basic explanation of Dharmakīrti’s root text. For the most part, his commentary
only provides syntactic and semantic glosses on Dharmakīrti’s extremely elliptical Sanskrit.
Perhaps the most dramatic exception to this general rule concerns Devendrabuddhi’s introduc-
tion of a terminological distinction between the “mediated” (vyavahita) versus “unmediated”
(avyavahita) instrumental effect ( pramāṇaphala),2 the latter being defined as the “resulting”
( phala) knowledge that comes about due to the “instrumental activity” (kriyā) or operation of
an “instrument of knowledge” ( pramāṇa).3 But to the extent that Devendrabuddhi’s innova-
tion here serves only to resolve an otherwise unexamined ambiguity in Dharmakīrti’s episte-
mological system, it illustrates Devendrabuddhi’s general approach as a commentator: namely
that of maximal self-effacement. Unlike later commentators on Dharmakīrti such as Dharmot-
tara (c. 775–850) and Prajñākaragupta (c. 800–875), who (not uncommonly for the genre) took
the task of commentary as an opportunity to engage in substantively constructive projects of
their own, Devendrabuddhi’s commentarial “voice” is difficult to discern. As with his decision
to forego any commentary on the PV’s first chapter – seemingly because Dharmakīrti himself
had already provided it – even Devendrabuddhi’s introduction of novel terminology is charac-
terized by maximal deference to Dharmakīrti.
The underlying ambiguity addressed by Devendrabuddhi’s introduction of the “(un/)medi-
ated instrumental effect” concerns a tension or ambiguity between two meanings of the cen-
trally important Sanskrit term artha, which in this context may mean “object” or “real thing”
on one hand and “purpose,” “aim,” or “goal” on the other hand. This second meaning can be
understood specifically in terms of the type of “human aim” ( puruṣārtha) that is the goal of
practical activity (vyavahāra). These two meanings of artha in turn bear upon two possible
interpretations of the compound arthakriyā, a key concept in Dharmakīrti’s system, which
combines the word artha with the word kriyā, meaning activity. The compound may thus
be interpreted as either the “causal functionality” possessed by real objects or the “purpose-
oriented activity” engaged in by humans and other sentient beings.
395
The ambiguity between these two different senses of artha and arthakriyā may be illus-
trated in the distinction between two different types of sensory-cognitive events: the aforemen-
tioned sensory perception of fire and the spurious perception of myodesopsic “floating hairs”
(keśa) or the “double moon” (dvicandra) that arises due to having crossed eyes. In the first
case, fire is the object (artha) of a sensory perception; hence, when perceiving fire for the pur-
pose (artha) of seeking out warmth, the causal functionality (arthakriyā) of the fire, consist-
ing in the fact that the fire is warm by nature, backstops the trustworthiness of the perception
of fire. “In other words, an entity’s causal characteristics, while independent of a perceiver’s
expectations, are conceptualized by the perceiver within the context of expectations concern-
ing some goal” (Dunne 2004, 259): the perception of fire is trustworthy because real fire actu-
ally does possess the causal functionality (arthakriyā) of fire, entailing that one’s purposeful
activity (arthakriyā) of seeking out warmth will result in the attainment of one’s goal (artha).
But the ambiguity in and of itself does not necessarily constitute a philosophical problem in
this case. Indeed, the overlapping senses of the terms artha and arthakriyā on display here are
almost certainly part of Dharmakīrti’s intended point: that the practical utility of a perceptual
event derives from the perceptual object’s fulfillment of some practical goal. However, the
situation is rather different in the case of the cognition of floating hair or a double moon. There,
because it is not the case that there really exists any floating hair or second moon, the cognition
technically has no artha – that is, neither any real object nor any practical purpose. As Dunne
(2004, 257–58) explains in a passage that includes an excerpt from Devendrabuddhi’s PVP ad
PV 3.1 (note that, in this translation, Dunne renders arthakriyā as “telic efficacy”):
Although artha as “aim” and “object” may thus be distinguished, even when the
term artha is best rendered as “object” it retains some connection to the notion of a
purpose or goal. Such is the opinion of Devendrabuddhi, as is apparent when he com-
ments upon Dharmakīrti’s statement at the outset of PV 3 that the illusory “hairs” or
“flies” perceived by a person with cataracts cannot be considered “objects” (arthas).
As noted earlier, the crux of the problem here is that Dharmakīrti might be obliged to
posit not only particulars and universals as objects of knowledge ( jñeya), but also a
third type, namely, illusory objects. Commenting on this problem, Devendrabuddhi
raises an objection and offers his reply:
Opponent: “Well, in the awareness of a person with defective vision there are
false appearances such as hairs, flies, two moons and so on. Those false appear-
ances do not accomplish any purpose (artha) at all; hence, they are not counted as
particulars. But even though they are devoid of telic efficacy, they are not included
among universals because they appear clearly and because they are not distributed
over anything. Thus, since they are not subsumed under the categories of particulars
and universals, they are another kind of object. Hence, it is not correct that there are
just two kinds of objects.”
Devendrabuddhi: “No, it is not the case that our view is not correct. To be specific,
hairs and so on are not objects (arthas). Why? Because they are not considered to be
objects (PV 3.1d). That is, persons engaged in practical action (vyavahartṛ) do not
consider them to be objects. The intention of this statement is as follows. If the flies
and so on that are perceived by a person with cataracts and so on were to be objects,
then one would investigate the situation, asking, ‘Is it a particular, or is it the other
[i.e., a universal]?’ But they are not objects in that fashion because, with regard to
an awareness in which there is the appearance of flies and so on, persons engaged
in practical action do not have the intention, ‘This is the object of that awareness.’ ”
396
The crux of the problem here is that Dharmakīrti only admits two types of epistemic objects,
particulars and universals, which are respectively the object of nonconceptual (i.e., percep-
tual) and conceptual (i.e., inferential) awareness. Dharmakīrti furthermore insists that only
particulars are actually real because only particulars possess causal functionality (arthakriyā).
According to Dharmakīrti, only particulars can be the cause of a sensory cognition because
a sensory cognition is the direct and immediate effect of the causal interaction between the
sensory faculties and the sensory object. Universals are not real in this way because they do
not possess causal functionality. However, Dharmakīrti also goes to great lengths to explain
that the appearance of floating hair or the double moon is in fact nonconceptual (Yiannopoulos
2020, 129–56). But how can a nonconceptual cognition arise without a real object (artha)
to serve as its cause? What, in other words, accounts for the “vivid” (spaṣṭa) appearance of
nonconceptual sensory errors, as opposed to the nonvivid appearance of conceptual sensory
errors, such as when a mirage is mistaken for water? In the latter case, the mirage does not
truly have the appearance of water and may be correctly identified as a mirage; in the former
case, however, there is indeed a visual cognition with the vivid appearance of two moons.
While this is an enormous topic that touches upon nearly every aspect of Dharmakīrti’s
ontology and epistemology, the short answer is that there are two ways to analyze such non-
conceptual errors. These two methods of analysis ultimately correspond to Devendrabuddhi’s
distinction between the mediated (vyavahita) and the unmediated (avyavahita) instrumental
effect ( pramāṇaphala), the latter being defined as the knowledge that is the “result” ( phala) of
the causal operation of the epistemic instrument ( pramāṇa). Analyzed in terms of “human aims”
( puruṣārtha) and “practical activity” (arthakriyā), such nonconceptual errors are in fact “non-
cognitions” (ajñāna; Yiannopoulos 2020, 276) because they have no artha. They have no artha
since they only arise due to an internal defect within the sensory faculties (PV 3.293). Analyzed
in terms of sheer “causal functionality” (arthakriyā), however, it must be admitted that the cog-
nition of floating hair or a double moon does arise directly from its causes (the aforementioned
defect in the sensory faculties), and it does produce effects. As such, it meets Dharmakīrti’s
requirement that all real phenomena are causally efficacious by nature, and vice versa.4 In this
latter sense, in fact, every cognition may be understood to possess arthakriyā, just insofar as
every cognition is a real entity that arises from causes and produces effects. Thus, in the first
case, a cognition (such as the perceptual cognition of fire) is “instrumental” insofar as it will
eventually allow for the attainment of some practical goal, while in the second case, cognition is
only “instrumental” with respect to itself as a cognition, just in terms of the fact that it is occur-
ring at all, without any reference to any further goal or aim. As Dunne (2004, 261–62) explains,
On Devendrabuddhi’s view, Dharmakīrti proposes two alternative interpretations of

what constitutes the instrumental effect. He claims that “. . . there are two kinds of
instrumental effects: one called a ‘human aim ( puruṣārtha),’ which is a mediated
(chod pa = vyavahita) effect, and a distinctive one (khyad par can = viśiṣṭa) that is
not mediated.” . . .
[An] effect is mediated (vyavahita) in that there is an “interval” (vyavadhāna)
between the functioning of the instrument (the karaṇa, i.e., the pramāṇa) and the
effect. In other words, the effect is separated (vyavahita) from the instrument by
virtue of other intervening causes and conditions that must be in place for the effect
to occur. The effect is thus “remote” (vyavahita) relative to the instrument, and since
some obstruction can therefore occur between the functioning of the instrument and
the production of the effect, the effect may not necessarily occur, even if the instru-
ment functions correctly. . . .
397
In contrast to the mediated effect, the unmediated effect is “distinctive” in that, as

Śākyabuddhi notes, it “necessarily occurs.” This is so because the effect is not sepa-
rated (avyavahita) from the instrument; it is, in fact, identical to the instrument itself.
As such, it is not remote (avyavahita); instead, the effect is actually simultaneous
with the instrument. In other words, the instrumental cognition ( pramāṇa) is itself
the instrumental effect ( pramāṇaphala).
In Devendrabuddhi’s account, in other words, the mediated instrumental effect – construed in terms
of attaining a human aim ( puruṣārtha) – involves a temporal delay between the operation of the
instrumental cognition ( pramāṇa) and the occurrence of the instrumental effect ( pramāṇaphala).
For example, in this account, the initial sensory perception of fire would be the instrumental cogni-
tion, and the subsequent experience of its warmth would be the instrumental effect. This, indeed, is
basically the only way to make sense of Dharmakīrti’s model of “instrumentality” ( prāmāṇya) in
practical terms, since the attainment of practical goals in the world is the overarching context for
pramāṇa discourse generally (Dunne 2004, 45–49), and the mere sensory perception of fire does
not yet constitute the fulfillment of any practical goals (such as getting warm).
However, there is a problem: this interpretation is irreconcilable with Dignāga’s and
Dharmakīrti’s insistence that the instrumental cognition (i.e., the pramāṇa) is ontologically
identical to the instrumental effect (i.e., the phala). In the Buddhist system as articulated by
Dignāga and defended by Dharmakīrti, it is not permissible to assert any ontological distinction
or temporal delay between the operation of the pramāṇa and the existence of the phala.5 Hence,
the potential problem here is that, as Dunne (2004, 268 n. 78) notes, the underlying issue is
that “Dharmakīrti’s extensive treatment of the instrumental effect ( pramāṇaphala) in PV 3 [in
the context of reflexive awareness] does not correspond well to his account of instrumental
cognition in PV 2.1–6 [in the context of practical human aims].” This is why Devendrabuddhi
felt the need to make the distinction between the mediated and unmediated instrumental effect
in the first place. In Devendrabuddhi’s account, only the unmediated instrumental effect is
ontologically identical with the instrumental cognition; the mediated instrumental effect, being
temporally subsequent, is ontologically distinct from the instrumental cognition.
As with the case of floating hairs, however, in general, any cognition that is analyzed in
terms of the unmediated instrumental effect cannot be understood to possess any causal func-
tionality or telic efficacy (arthakriyā) other than its own momentary existence. When analyz-
ing a cognition in terms of the unmediated instrumental effect, in other words, by definition,
there can be no reference to any subsequent acquisition of what is desired, since any such
acquisition would be “mediated” in precisely the sense at stake. This type of analysis is cru-
cially important to Śākyabuddhi’s account of reflexive awareness, to which we now turn.
 vedanamātra) and False

“Mere Reflexive Awareness” (svasam
Imagism
Reflexive awareness (svasaṃvitti or svasaṃvedana) is among the most important key terms
in Dharmakīrti’s system, the topic of nearly half of the Commentary’s “Perception Chapter”
( pratyakṣapariccheda). And yet nowhere does Dharmakīrti provide an explicit definition of
this centrally important term. Nevertheless, the basic idea is simple: reflexive awareness des-
ignates that feature of a cognition which immediately and reflexively (which is to say, non-
transitively) presents the contents of cognition to itself.
Grammatical reflexivity is central to the philosophical issue here: svasaṃvitti has not infre-
quently been translated as “self-awareness,” but this translation is at best misleading and at
398
worst simply incorrect. The confusion centers around the reflexive pronoun sva, which does not
mean “self ” in the sense of a “soul” or “ego” (more commonly designated by ātman), but only
designates that the accusative object of the verbal action is the same entity as the nominative
subject; as in: “He holds himself well,” or “I hurt myself.” More rarely, Dharmakīrti does use
ātman in compound with the root √vid (i.e., ātmavit) to designate reflexive awareness; but here
as well, ātman is only functioning as a reflexive pronoun, rather than in any way that might
imply some type of metaphysical “self ” that is aware of “itself.” Indeed, Dharmakīrti takes great
pains to construct his epistemological system in such a way so as to avoid postulating any such
agentive “self,” which would be absolutely incompatible with any type of Buddhist philosophy.
Thus, the meaning of reflexive awareness is not “self-awareness” in the sense of the mind
somehow taking itself for a transitive object or deliberately reflecting upon its own contents.
As MacKenzie (2007) notes, reflexive awareness is not at all the same thing as reflective aware-
ness. On the contrary, reflexive awareness designates the fact that the contents of cognition are
always immediately, which is to say “unmediatedly” (avyavahita), present to the cognizing
mind. Indeed, it is precisely by virtue of being unmediated in this exact way that the present
moment of awareness constitutes the “unmediated instrumental effect” described previously.
As Dunne (2004, 276) explains:
Reflexive awareness . . . is an exception to the rule that the object of a perception

( pratyakṣa) is the cause of the image in the perceptual cognition. Hence, if it is to
be considered instrumental, it clearly cannot be considered in terms of a mediated
instrumental effect, since this would require an appeal to causality that is not pos-
sible in the case of reflexive awareness. And it likewise [is] difficult to imagine how
reflexive awareness can fit into a scheme of human goals; hence, unlike a cognition in
which appears the accomplishment of one’s aim, one cannot appeal to some unmedi-
ated experience of telos, except the restricted telos of knowing that one knows. . . .
Thus, if we are to accept reflexive awareness as a trustworthy cognition . . . we must
again turn to an alternative interpretation of arthakriyā, namely, that reflexive aware-
ness is reliable in that it reveals the mere fact of experience, which is the same as
saying that it reveals the mere causal efficiency (arthakriyā) of awareness.
In other words, the “instrumentality” of pure reflexive awareness can no longer be understood
in terms of the accomplishment of some ordinary human goal, precisely because (in this rarefied
context) reflexive awareness is understood to be entirely unmediated. That is to say, because in
this specific context epistemic instrumentality ( prāmāṇya) is no longer construed in terms of a
cognition’s (i.e., a pramāṇa’s) ability to facilitate the acquisition or avoidance of some extra-
mental “epistemic patient” or “object of knowledge” ( prameya) at some later time, but rather is
strictly construed in terms of the cognition’s ability in the present moment to bear upon itself just
as a cognition, there is no longer any question of the cognition being practically useful for any
goal besides providing information about its own contents. Thus, as Dunne (2004, 317) explains:
If we trust Śākyabuddhi’s opinion, the ultimate pramāṇa would be the pure, nondual,
reflexive awareness of the mind itself. But while this ultimate instrumental cognition
is the means to Dharmakīrti’s final soteriological goal, it is not useful for practical
action in the world (i.e., saṃsāra). If the ultimate instrument of knowledge is indeed
some pure form of reflexive awareness, then there are no longer external objects –
or even mental content – on which to act. Hence, it would seem that conventional
perceptions and inferences are eventually left behind, but in terms of cognitions for
399
practical action in the world, these conventional cognitions are as trustworthy and
irreproachable as any such cognition could ever be.
The question or problem here arises because, according to Dharmakīrti, every ordinary cog-
nition appears dualistically, such that the “object-image” or “aspect of the apprehended”
(grāhyākāra) is necessarily accompanied by the first-person sense of subjectivity or “aspect of
the apprehender” (grāhakākāra). Dharmakīrti refers to this necessity as the “rule that phenom-
enal subject and object must appear together” (sahopalambhaniyama). But in keeping with the
Yogācāra tradition, and with Buddhist philosophy more generally, Dharmakīrti also repeatedly
asserts that phenomenological duality as such constitutes a fundamental error or distortion
(bhrānti) within the nature of cognition. This, then, is the issue: given that any dualistic cogni-
tion is necessarily and by nature erroneous, but also that any objective cognitive content (i.e.,
any grāhyākāra) must by necessity appear dualistically as part of a “structure of apprehended
and apprehender” (grāhyagrāhakasthiti), how is it possible for any cognitive content to exist
in the final analysis? Put slightly differently: if, as would seemingly have to be the case, the
absence of phenomenological duality necessitates the absence of what would ordinarily be
meant by “phenomenal content” tout court, how or in what sense could such a nondual cogni-
tion be useful for ordinary practical action in the world?
Historically, the major dividing line within the post-Dharmakīrti Indian Buddhist epistemo-
logical tradition ran through the answer to this question. On one side were the “True Imag-
ists” (satyākāravādins), who maintained that cognitive “images” or “content” (ākāra) could
somehow remain, even in the absence of duality. On the other side were the “False Imagists”
(alīkākāravādins), who argued to the contrary that the absence of duality entails the absence of
cognitive content, leaving only “mere reflexive awareness” (svasaṃvedanamātra). Although this
debate was not fully developed until centuries after Dharmakīrti, perhaps not until as late as 1000
CE, the False Imagist perspective may be observed in embryonic form within Śākyabuddhi’s
remarks. In particular, Śākyabuddhi’s extensive comments concerning the extremely important
verse PV 3.212 are a rich and important source for this perspective. This verse reads:
This part of cognition – which is situated as though external – is different from the
internal determination (i.e., the part that is situated as though internal). Hence, the
appearance of difference in an actually undifferentiated awareness is cognitive distor-
tion (upaplava). || 212 ||
Devendrabuddhi’s comments to this verse are sparse. This should come as no surprise, since,
as discussed previously, Devendrabuddhi’s overriding concern as a commentator was fidelity to
Dharmakīrti’s root text. In the case of the mediated versus unmediated instrumental effect, for
example, Devendrabuddhi attempted to smooth out – with only the bare minimum of original
contributions – an issue that constitutes perhaps the most glaring potential contradiction within
Dharmakīrti’s epistemological system taken as a whole. But the issue here does not really pre-
sent the same type of contradiction, because Dharmakīrti frequently specifies that duality just
is cognitive error. Dharmakīrti’s silence on this topic may be frustrating to his readers, but does
not in and of itself necessarily constitute an exegetical dilemma, insofar as there is no contradic-
tion or incoherence to sort out. Accordingly, Devendrabuddhi apparently did not feel the need to
introduce any new theoretical elaborations or terminological distinctions concerning this issue.
Śākyabuddhi, on the other hand, in keeping with his penchant for extrapolation from and
interpretation of Dharmakīrti’s root text, takes this extremely important verse as an opportu-
nity to make more extensive comments. To begin with, he notes that “the bodhisattvas who
400
have realized that phenomena are selfless only know mere reflexive awareness (rang rig pa
tsam = svasaṃvedanamātra), which is devoid of duality” (Dunne 2004, 406). In other words,
according to Śākyabuddhi, all that remains following the collapse of phenomenological dual-
ity is mere contentless reflexive awareness. Furthermore, in keeping with what would later
become the False Imagist position, Śākyabuddhi also asserts in the same passage that “dualis-
tic awareness is not real; rather, it is established through cognitive error.”
Śākyabuddhi then expands upon these brief remarks with a very interesting and lengthy
digression on the nature of nonduality, subjectivity, and reflexive awareness. This excursus is
framed as a response to a common recurring objection against the False Imagist perspective –
namely that if neither phenomenological duality nor dualistic cognitions are actually real,
then beings should constantly be experiencing “suchness” (tathatā) or nondual reality-as-such,
since this would be the only thing that is in fact available to experience (Dunne 2004, 407):
Someone objects: “If the object and subject do not exist, then what would be left
but the suchness of awareness itself? Cognitively myopic beings do not experience
anything but the objective and subjective cognitive images. If they were to experi-
ence something else, they would see suchness. That being the case, beings would be
effortlessly liberated.”
Śākyabuddhi answers this objection – which, it should be remembered, occurs in the context
of comments to a verse that does not directly reference reflexive awareness at all – by articu-
lating a distinction (only ever implied in Dharmakīrti’s root text) between reflexive awareness
and subjectivity:
[With] the word “subject” we do not mean to express reflexive awareness – the inter-
nal cognition that arises in various forms such as the pleasant and the unpleasant –
such that by expressing it with the term “subject” we would be saying that it does
not exist. Rather, we mean the following. Cognitive appearances such as blue seem
to be external to awareness; [however,] there is ultimately no object that is distinct
from awareness itself, and since that object does not exist, we say “the subject does
not exist”. . . . Since an agent and its patient are constructed in dependence upon
each other, these two (i.e., subject and object) are posited in dependence on each
other. The expression “subject” does not express mere reflexive awareness, which
is the essential nature of cognition itself. The essential nature of cognition is not
constructed in mutual dependence on something else because it arises as such from
its own causes. The essential nature of cognition is established in mere reflexive
awareness. Since it is devoid of the above-described object and subject, [reflexive
awareness] is said to be nondual.6
A complete analysis of this remarkable passage would lie outside the scope of this brief treat-
ment; for a more thorough consideration of Śākyabuddhi’s distinction between reflexive aware-
ness and subjectivity, see Yiannopoulos (2020, 388–413). In short, however, it should be noted
that the key point here is the fact that reflexive awareness is “the essential nature of cognition
itself ” – what Dharmakīrti elsewhere (PV 3.302) terms the “similar nature” (sadṛśātman) of
every cognition, just insofar as it is a cognition. By contrast, phenomenal subjectivity is just an
erroneous artifact of ignorance, posited in mutual dependence upon phenomenal objectivity.
In concrete terms, what Śākyabuddhi is saying here in response to the opponent’s objection is
that, so long as appearances appear, those appearances must be understood as fundamentally
401
distorted or mistaken, precisely insofar as their objective appearance necessitates the concomi-
tant appearance of subjectivity; because, even when they are properly understood within the
idealistic framework of Yogācāra as “mind only” (cittamātra), these appearances are necessar-
ily bound up with the fundamentally distorted structure of duality. “Mere reflexive awareness”
(svasaṃvedanamātra), on the other hand, does not present any such properly objective phe-
nomenal content. Indeed, later False Imagists such as Ratnākaraśānti (c. 1000 CE), drawing
heavily upon Śākyabuddhi’s line of reasoning, would explicitly name this state “luminosity-
only” ( prakāśamātra), as “luminosity” ( prakāśa) is a synonym for reflexive awareness, even
in Dharmakīrti’s Commentary (see, in particular, PV 3.327–32).
Hence, even though every cognition is identical in terms of having the essential, lumi-
nous nature of nondual “mere reflexive awareness,” it is not the case that cognition ordinarily
appears as mere nondual luminosity. On the contrary: despite not actually having a dualistic
nature, cognition ordinarily appears as though there were a discrete phenomenal subject and
phenomenal object. Therefore, in the absence of nondual yogic practice (i.e., yogācāra) aimed
at eliminating this distorted structure of duality, it is impossible to have a definite determina-
tion or ascertainment (niścaya) of the suchness of one’s own experience, which is to say, of
the fact that experience as such is by nature devoid of duality. As Ratnākaraśānti succinctly
summarizes in his Pith Instructions for the Ornament of the Middle Way, referencing the exact
same objection raised by Śākyabuddhi previously (Yiannopoulos 2012, 227–28):
One might ask: “Since this emptiness of duality is beginningless, doesn’t that mean
that we should have been enlightened from that time, at the very beginning?”
This does not follow, because even though one has some kind of experience, still
the emptiness of duality on the part of that experience is not ascertained.
In this way, we may see that Śākyabuddhi’s line of interpretation, venturing rather further
afield from Dharmakīrti’s root text than Devendrabuddhi’s extremely conservative approach,
proved quite influential for the later Indian Buddhist epistemological tradition.
While a detailed examination of Śākyabuddhi’s comments would lie outside the scope of
this brief treatment, it is worth sketching out a few key takeaways from the preceding discus-
sion. First, and arguably foremost, it must be emphasized that Dharmakīrti’s discussion of the
nondual nature of awareness dovetails – almost certainly deliberately – with Mahāyāna Bud-
dhist approaches to contemplative practice. Indeed, it would not overstate the matter to note
that Dharmakīrti’s presentation of “luminosity” ( prakāśa), particularly as elaborated upon by
Śākyabuddhi, provides a theoretical and epistemological foundation for advanced nondual
contemplative practices such as Mahāmudrā. From this perspective, the nascent False Imag-
ist view that may be discerned in the text is not simply an epistemological stance, but also
provides an ontological and epistemological account of the key point of meditative equipoise
(samāhita), and to some extent even an account of the nature of Buddhahood itself, as phe-
nomenally contentless (i.e., nirākāra) luminosity. We may conclude on this note by highlight-
ing yet again how, far from merely representing the opinion of two subsequent authors, the
commentaries of Devendrabuddhi and Śākyabuddhi should be regarded as integral to the PV
itself, forming together with it a complex hybrid-textual structure. This is not necessarily to
say that the False Imagist perspective is the only possible interpretation of Dharmakīrti, but
(per Devendrabuddhi) it is certainly the most grammatically and syntactically straightforward
reading of Dharmakīrti’s Sanskrit, as well as (per Śākyabuddhi) the most hermeneutically
concise and consistent interpretation of it.
402
Notes
1 The order of the chapters in the PV has been a matter of some controversy in India and Tibet as well as
in the contemporary scholarly literature. For a discussion of this issue, see Yiannopoulos 2020, 11–14.
2 See PVP ad PV 2.3b – d, translated in Dunne 2004, 383–85.
3 For an overview of pramāṇa theory in terms of the relation between “means” and “result,” see Patil
2009, 37–40; Dunne 2004, 49–50; Yiannopoulos 2020, 159–65.
4 See Dunne 2004, 84–97.
5 For more detail concerning this point, see Yiannopoulos 2020, 168–72.
6 Translated in Dunne 2004, 407.
Bibliography
Devendrabuddhi. 2002. “Pramāṇavārttikapañjikā (tshad ma rnam ’grel kyi ’grel pa).” In Bstan ’gyur
(dpe bsdur ma), edited by krung go’i bod kyi shes rig ’jug lte gnas kyi bka ’bstan dpe sdur khang,
translated by Subhutiśrī and Dge ba’i blo gros, vol. 98 (tshad ma che), 3–862. Beijing: krung go’i bod
kyi shes rig dpe skrun khang.
Dharmakīrti. 1960. “Pramāṇavārttikasvopajñavṛtti.” In The Pramāṇavārttikam of Dharmakīrti: The
First Chapter with the Autocommentary, edited by Raniero Gnoli. Roma: Istituto Italiano per il Medio
ed Estremo Oriente.
———. 1979. “Pramāṇavārttika.” In Bukkyō ninshikiron no kenkyū: Hōshō cho “Puramāna Vārutika”
no genryōron, edited by Hiromasa Tosaki. Tōkyō: Daitō Shuppansha.
———. 2007. “Pramāṇaviniścaya.” In Dharmakīrti’s Pramāṇaviniścaya: Chapters 1 and 2, edited by
Ernst Steinkellner. Sanskrit Texts from the Tibetan Autonomous Region. Wien and Beijing: Verlag der
Österreichischen Akademie der Wissenschaften; China Tibetology Publishing House.
Chimpa, Lama, and Alaka Chattopadhyaya. 1970. Tāranātha’s History of Buddhism in India. Delhi:
Ganeri, Jonardon. 2010. “Sanskrit Philosophical Commentary.” Journal of the Indian Council of Philo-
sophical Research 27: 187–207.
MacKenzie, Matthew. 2007. “The Illumination of Consciousness: Approaches to Self-Awareness in the
Indian and Western Traditions.” Philosophy East and West 57 (1): 40–62.
McClintock, Sara. 2010. Omniscience and the Rhetoric of Reason: Śāntarakṣita and Kamalaśīla on
Rationality, Argumentation, and Religious Authority. Somerville, MA: Wisdom Publications.
Patil, Parimal. 2009. Against a Hindu God: Buddhist Philosophy of Religion in India. New York: Colum-
Ratnākaraśānti. 2002. “Madhyamakālaṃkāropadeśa (dbu ma rgyan gyi man ngag).” In Bstan ’gyur (dpe
bsdur ma), edited by krung go’i bod kyi shes rig’jug lte gnas kyi bka’bstan dpe sdur khang, translated
by Śāntibhadra and Śākya’od, vol. 78 (sems tsam hi), 604–25. Beijing: Krung go’i bod kyi shes rig
dpe skrun khang.
Śākyabuddhi. 2002. “Pramāṇavārttikaṭīkā (tshad ma rnam ’grel kyi ’grel bshad).” In Bstan ’gyur (dpe
bsdur ma), edited by krung go’i bod kyi shes rig ’jug lte gnas kyi bka ’bstan dpe sdur khang, translated
by Subhutiśrī and Dge ba’i blo gros, vols. 98–99 (tshad ma je; nye), 955–1802 and 1–714. Beijing:
Krung go’i bod kyi shes rig dpe skrun khang.
Steinkellner, Ernst, ed. 2005. Dignāga’s Pramāṇasamuccaya, Chapter 1: A Hypothetical Reconstruc-
tion of the Sanskrit Text with the Help of the Two Tibetan Translations on the Basis of the Hith-
erto Unknown Sanskrit Fragments and the Linguistic Materials Gained from Jinendrabuddhi’s Ṭīkā.
www.oeaw.ac.at/fileadmin/Institute/IKGA/PDF/forschung/buddhismuskunde/dignaga_PS_1.pdf.
Yiannopoulos, Alexander. 2012. “Luminosity: Reflexive Awareness in Ratnākaraśānti’s Pith Instruc-
tions for the Ornament of the Middle Way.” M.A. Thesis, Rangjung Yeshe Institute at Kathmandu
University.
———. 2020. “The Structure of Dharmakīrti’s Philosophy: A Study of Object-Cognition in the Per-
ception Chapter ( pratyakṣapariccheda) of the Pramāṇasamuccaya, the Pramāṇavārttika, and Their
Earliest Commentaries.” PhD diss., Emory University.
403
24
CANDRAKĪRTI
Gardener of Sky-Flowers1
Mattia Salvini
The Teacher, Gautama, and the perfect Buddhas

that abide in the same Dharma-nature;
the entire host of the Victors’ sons,
and the Dharma they have spoken –
to these I prostrate, and then bow to him,
the eye for embodied beings to see
the Buddha’s Words without end,
he who composed this Madhyamaka,
out of compassion – Nāgārjuna.2
Candrakīrti (attributed),
Praise of the Madhyamaka Treatises
Introduction
Candrakīrti (ca. seventh century) is a prime exponent of Madhyamaka thought, the Bud-
dhist “Middle Way” system free from the extremes of existence, nonexistence, eternalism,
and materialist denial of rebirth. He presents his philosophy primarily in the context of com-
menting upon Nāgārjuna’s work, to defend his Mahāyāna view through reason and appeal to
authoritative texts, against an array of opposing interpretations. He moves with clarity and
ease between scholastic technicalities and pithy instructions to point at the inconceivable
nature of things and infuses his arguments with literary beauty and suggestion.
Candrakīrti speaks like a debater, a monastic, and a meditator. Little independent histori-
cal evidence regarding his life survives, but these features of his personality are represented
in later hagiographical accounts (Chimpa, Chattopadhyaya, and Chattopadhyaya 1990, 198–
209). Blurring the boundaries between exegesis, literature, philosophy, and advice to those on
the path, his writings present a vision of reality through metaphors, images, and other literary
devices as well as through explicit argumentation. In this vein, Candrakīrti’s clarity lies in
how he prepares and affects the interlocutor’s mind, making it receptive to key Madhyamaka
themes presented in striking and ingenious ways.
404 DOI: 10.4324/9781351030908-35

Candrakīrti
Candrakīrti’s Nāgārjuna
The “Madhyamaka” system (literally, “relating to the Middle”) is named after Nāgārjuna’s Root
Stanzas of the Madhyamaka (Mūlamadhyamakakārikā). Candrakīrti is a Mādhyamika philo
sopher who presents his work as the elucidation of Nāgārjuna’s intent. He poetically praises
Nāgārjuna by explaining the two parts of his name: he is a water-dragon (nāga) who “gained
birth in the ocean of the wisdom of the Perfect Buddha”; that is, he is a bodhisattva with direct
insight into reality. Like the epic hero Arjuna, he is an impeccable archer, vanquishing saṁsāra
with the arrows of his insightful speech (MacDonald 2015, Vol. II, 7; Salvini 2016, 419–21).
Candrakīrti writes that Nāgārjuna composed his work in order to explain the system of
the Perfection of Wisdom, an important set of Mahāyāna sūtras (MacDonald 2015, Vol. I,
116–17), out of compassion for sentient beings who would have difficulty understanding it.
This determines the focus of Candrakīrti’s philosophy on the harmony between basic Buddhist
categories relating to sentient experience and the illusory nature of that very experience –
the recurrent, perhaps even main, theme of the Perfection of Wisdom literature. This litera-
ture has been misunderstood, according to Candrakīrti, by Yogācāra philosophers (Li 2015).
Candrakīrti agrees with his opponents that the mind has causal prominence when compared to
non-sentient causal factors. On the other hand, he argues that the Yogācāra view of an onto-
logical primacy of the mind is an untenable reification of awareness, incompatible with the
intent of either the sūtras themselves or of Nāgārjuna’s treatises.
Candrakīrti’s main philosophical works

The Introduction to the Madhyamaka (Madhyamakāvatāra) (de La Vallée Poussin 1912; Hun-
tington 1989; Li 2015; Lasic, Lee, and MacDonald 2022), together with its autocommentary, is
a comprehensive outline of the bodhisattva path, organized in accordance with the Mahāyāna
gradation of ten levels (bhūmi) – from the entrance into the state of “Noble Individuals” on the
first level (when a practitioner is said to see reality directly for the first time) to complete bud-
dhahood. Within this text, the sixth chapter, the longest of all, is devoted to the cultivation of the
Perfection of Wisdom and constitutes an influential presentation of Candrakīrti’s Madhyamaka
understanding. Its definitions of the two truths is of particular significance and informs the writ-
ing of later Madhyamaka authors who followed in its wake, such as Prajñākaramati and, argu-
ably, Śāntideva. While the Introduction to the Madhyamaka is not quite literally a commentary
on any of Nāgārjuna’s works, it is still an exegesis of his thought, as Candrakīrti implies at the
beginning of chapter six. This section contains a long refutation of the Yogācāra view of mind.
According to Candrakīrti, Yogācāra philosophers mistook the “-only” in the Buddha’s expres-
sion “mind-only” (citta-mātra) to be a refutation of materiality, implying an ontology wherein
only mental entities are real. In contrast, Candrakīrti argues, that expression was meant by the
Buddha as a refutation of any other “agent” (kartr.) apart from the mind, especially, in terms
of an agent of creation, that is, an omnipotent singular creator. For Candrakīrti, the mind is the
agent of creation, that is, the creator; at the beginning of each cosmogonical cycle, the collec-
tive accumulated karma of sentient beings in that particular universe determines the formation
of the environment that they will inhabit – a standard Buddhist perspective on cosmogony.
His commentary In Clear Words (Prasannapadā; de La Vallée Poussin 1903–13; Ruegg
2002; MacDonald 2015) is a detailed gloss on all the chapters of Nāgārjuna’s Root Stanzas on
the Madhyamaka. The title is an indication of Candrakīrti’s preferred style. The term prasanna
may refer to clear, well-settled water, unagitated by “the wind of speculation” (tarka-anila) –
an allusion to his opponent Bhāviveka’s fondness for presenting arguments in a manner clearly
405
Mattia Salvini
influenced by Dignāga’s innovations in the formalized articulation of inferences. Candrakīrti

did not endorse the new format, as he rather preferred Buddhapālita’s own focus on pointing
out unwanted consequences (a sort of reductio ad absurdum). The extent to which this diffe
rence in argumentative style may have any broader ontological implications is a matter of con-
siderable debate in the later Madhyamaka tradition (see, for example, Dreyfus and McClintock
2003). This commentary refers to the Introduction to the Madhyamaka several times, which
therefore must have been composed earlier.
The commentary on Nāgārjuna’s Sixty Verses on Reasoning (Yuktiṣaṣṭikāvr․ tti; Scherrer-
Schaub 1991; Loizzo and The AIBS Translation Team 2007; Li and Ye 2014) is a helpful
complement, offering several additional clues to better understand Candrakīrti’s thought on
a number of important and difficult points in the interpretation of Nāgārjuna. In particular,
Nāgārjuna’s text contains a verse (34) that has been read by a number of Yogācāra authors
as supporting an interpretation of Nāgārjuna according to their own philosophy. Candrakīrti
offers a completely different understanding of this key passage, in no way suggesting that it
could support a “mind-only” reading.
The Commentary on the Seventy Verses on Emptiness (Śūnyatāsaptativr․ tti; Erb 1997) is a
relatively short commentary. In the root text, Nāgārjuna clarifies his understanding of how con-
ventional dependent arising ( pratītyasamutpāda) relates to ultimate non-arising (anutpāda) and
is thus of great importance in assessing the sense in which Nāgārjuna declares that there is no
difference between saṁsāra and nirvāṇa. Candrakīrti’s commentary engages with Abhidharmic
matters and shows, for example, his acquaintance with the doctrine of six causes (Erb 1997, 242).
Āryadeva is one of the early great masters of Madhyamaka, reputed to be a direct student
of Nāgārjuna. Candrakīrti composed an elaborate commentary on his Four Hundred Verses
on the Yoga Practice of the Bodhisattvas (Bodhisattvayogācāracatuḥśatakaṭīkā; Shastri 1914;
Suzuki 1988, 1989, 1994; Tillemans 1990; Lang 2003; Johnson 2012). Since the root text is
itself a very wide-ranging philosophical treatise, this commentary offers insights into aspects
of Candrakīrti’s thought that go beyond more habitual topics: kingship and its likely downfalls,
the importance of contemplating death, the necessity of avoiding harmful actions, the unreli-
ability of romantic attraction, and so on. It also offers additional glimpses of Candrakīrti’s vast
erudition and establishes once again his familiarity with Vasubandhu’s works.
The Treatise on the Five Aggregates (Pañcaskandhaprakaraṇa; Lindtner 1979; Li and Kanō
2015) is an independent short manual of Abhidharma. It follows the classification of entities
into five realities (discussed subsequently) and yet does not perfectly match usual Sarvāstivāda
presentations. This treatise engages with Abhidharmic analysis to show that, when brought to
its logical conclusion, it culminates in the type of wisdom that establishes emptiness. A some-
what neglected work in contemporary scholarship (and some even doubt its authorship), it
offers important clues in reconstructing Candrakīrti’s overall philosophical perspective, espe-
cially in order to understand his approach towards other strands of Buddhist thought.
The Seventy Verses on Refuge (Triśaraṇasaptati) (Sorensen 1986; Skilling and Saerji 2012;
Kanō and Li 2014) differs slightly from other works of Candrakīrti in terms of its main empha-
sis and confirms his familiarity with several Buddhist schools and his willingness to engage
with their position. This work is also important for the Tibetan understanding of the meanings
of the Three Jewels (the Buddha, the Dharma, and the Saṅgha) and of taking refuge in them,
which is the foundational Buddhist vow common to lay practitioners and monks. This shorter
treatise complements the Introduction to the Madhyamaka by offering a more basic discussion
on the initial stages of Buddhist practice.
Last, at least two important texts on Tantra are ascribed to Candrakīrti (Chakravarti 1984;
Campbell 2009, 149–54; Hong and Tomabechi 2009). While this attribution is disputed by
406
Candrakīrti
modern scholars, it had a significant impact on the later reception and contextualization of his
philosophy.
Candrakīrti’s philosophical grammar

It is not indeed the case that words, as if they were people with clubs and snares, cause
the person who speaks them to lose one’s freedom: rather, when they have the proper
capacity, they function in conformity with what the speaker wishes to express.3
Candrakīrti, In Clear Words
Buddhist philosophers in South and Southeast Asia were active participants in the intellectual
environment of their times (Skilling 2019, 214), wherein not only was Sanskrit the idiom
of scholarly communication, but Sanskrit traditional grammar (vyākaraṇa, “analysis”) also
defined a common conceptual language, being the first example of how to reflect on sound
logical connections. Grammar affects the manner of philosophical exposition and argumenta-
tion and offers basic templates to conceptualize and express causal processes: sentential syntax
is analyzed as expressing primarily an activity and the factors that participate in it (the agent,
the object, the instrument, etc.). Thus, when the nature of causality and the limits of language
are discussed, traditional grammar becomes relevant.
Sanskrit traditional grammar involves linguistic analysis in a sense broader than in our
common usage of the term “grammar”. The first role of Sanskrit grammar, when employed
in philosophical texts, is to offer pertinent etymologies, clarifying the intent of sentences
that have “great meaning” (mahārtha), while discarding incompatible exegetical positions.
Another role of grammar, in Candrakīrti, is to discuss the implied presuppositions that make
a linguistic expression comprehensible. This type of discussion often relies on the dialectic
between an activity and the factors that make it possible (the agent, the instrument, the object,
etc.), according to the analysis Sanskrit grammarians apply to sentences and even single parts
of speech. One instance of such analysis is when Candrakīrti examines a sentence describ-
ing commonly perceived events in the world, such as motion. He shows that these events
are illusory by discussing the grammatical implications of their formulation. For example,
Nāgārjuna examines motion as follows: does motion necessitate an agent of motion in order
to be comprehensible? If so, is that agent a “goer” or a “non-goer”? If the agent is a “goer,”
by virtue of which action of going is the agent a “goer”? Is it the same action of going that he
is then performing? Or a different one, that establishes the existence of a “goer” who is actu-
ally not going, but may then “go”? and so on. Candrakīrti expands on Nāgārjuna’s analysis
by highlighting some of the conceptual underpinnings of the arguments, such as the insight
that factors such as “agent” are “capacities” (śakti), that is, roles with respect to single specific
actions rather than things (dravya) that may relate to multiple actions. This distinction, taken
from the grammatical tradition, makes better sense of Nāgārjuna’s argument, where the root
text states that once there are two actions of going, it follows that there must be two agents (an
argument found in the second chapter of the Stanzas on the Madhyamaka, verse 6). If we took
the term “agent” in the verse to refer to a continuing entity rather than to a role in respect to a
specific action, Nāgārjuna’s argument would have no force.
Alternatively, Candrakīrti may take up an instance of “bad philosophy,” wherein an oppo-
nent contradicts the shared presuppositions that make an expression comprehensible, in the
hope of evading analysis and positing an irrefutable entity. Such “bad” linguistic expres-
sions are akin to writing “this sentence is not in English,” hoping to avoid any possible criti-
cism of the sentence’s content, by claiming that it is not analyzable any further. This is how
407
Mattia Salvini
Candrakīrti criticizes Dignāga’s account of particulars or, more literally, “own-characteristics”

(svalakṣaṇa), according to which an ultimate momentary entity is simultaneously the instru-
ment and the object of its own definition. The literal sense of “particular” implies that any par-
ticular is characterized by itself only and not by a property shared with anything else, since such
momentary entity is literally called “own-characteristic,” and is absolutely unique. However,
an expression like “own-characteristic,” when used in Dignāga’s sense, contradicts the very
grammar that would make it comprehensible, since its form implies instrument and/or object,
while Dignāga implies that its referent escapes any analysis in terms matching such grammati-
cal categories. That is, Candrakīrti points out that Dignāga is employing words that imply syn-
tactical roles while denying that he is using them as expressing those roles (“agent,” “object,”
“instrument,”etc.; MacDonald 2015, Vol. I, 265; Vol. II, 262–63; see especially 262 n. 500).
Candrakīrti’s usage of grammar to uncover presuppositions may be a distant echo of the
grammarian’s principle that usage is taken “from the world” (lokataḥ, Joshi and Roodber-
gen 1986, 25). Candrakīrti appeals to expressive cogency, that is, to the necessary coherence
between what the linguistic forms presuppose and what they are intended to express. This
principle of linguistic coherence relies directly on his view that whatever exists can exist only
in relation to something else. Furthermore, this principle is what makes such a view of real-
ity expressible; it offers precious clues to understand what, for Candrakīrti, is the “virtuous”
circle allowing the conceptual elements of philosophical reflection to function in support of an
understanding of their own empty nature.
Candrakīrti also appeals to grammatical categories when explaining the intent of Madhy-
amaka refutations. He does so by distinguishing between a negation that implies the acceptance
of its opposite ( paryudāsa) versus a negation that simply negates without positing anything
( prasajya). This distinction originates in Sanskrit grammar and is used by other philosophers,
too. Candrakīrti introduces it in the context of a broader discussion about the nature of lan-
guage, which, according to him, does not oblige the speaker to uphold any particular ontologi-
cal view. He uses this grammatical distinction of two types of negation in an ingenious way.
He applies it to the context of philosophical debate to explain how a Madhyamaka thinker
should refer to entities or positions that are the subject matter of refutation without being
in any way committed to their existence. Sanskrit grammarians (some more explicitly than
others) apply a methodological bracketing regarding the ontological status of the referents of
those linguistic expressions that they are analyzing. Candrakīrti appeals to this general princi-
ple of linguistic analysis to highlight that a philosopher may argue without being forced into
committing to any counter-position. Moreover, he employs the distinction between two types
of negation to highlight that negating one type of ontological position does not imply being
committed to an alternative one, even to the extent that negating that “something exists” does
not entail upholding that “it doesn’t exist” – nor that “I am uncertain as to whether it does or
it does not exist.” In the quote at the beginning of this section, Candrakīrti extends an insight
derived from traditional grammar to bear upon the nature of language in general and more
specifically of philosophical and argumentative language, making it possible to speak of real-
ity and unreality in a uniquely Madhyamaka way. In this sense, Sanskrit traditional grammar
is the key to understanding Candrakīrti’s views (or “non-views”) on ontology.
Candrakīrti’s Abhidharma
Where the Guide explains the act of seeing as being due to an assemblage,
The intelligent say that it is on the level of an approximation of the ultimate.4
From the Sūtra on Moving within Existence
408
Candrakīrti
Abhidharma is the Buddhist phenomenological investigation into entities (dharmas), wherein

causally efficient entities are primarily categorized as either mental or material. Abhidharma
furthermore discusses the relationship between those two spheres and analyzes how concep-
tualization of spatial and temporal extension arises in the mental continuum. “Abhidharma”
is a very broad term, often referring, very generally, to the employment of a recurrent set of
analytical categories within Buddhist texts of any genre; in other contexts, it refers to Buddhist
philosophical schools who drew specific ontological conclusions regarding such categories.
Candrakīrti quotes the above verse to elucidate the role of Abhidharmic categories – some
of which he just explained to be ultimately illusory. The passage describes an intermediate step
between more ordinary/common-sense ways of speaking, where one may say “I see through
the eye,” and the ultimate truth of emptiness wherein there is no “seeing” to be spoken of. The
verse represents the description of causal processes in terms of the Abhidharma: a moment of
visual consciousness comes about due to a causal assemblage of the object, the sense-faculty,
and a prior moment of consciousness; it does not involve a temporally extended “person” as the
agent of seeing. This account is described in the verse as the “level of an approximation of the
ultimate”: it is not an ultimate description, since it is not from the perspective of emptiness and
still involves entities, and yet it is not simply tantamount to ordinary and common conventions.
In other words, for Candrakīrti one goes from “I see something” (ordinary speech) to “visual
consciousness arises due to causes and conditions” (Abhidharma account) to “neither seeing
nor non-seeing could ever possibly take place” (ultimate view). While the first two steps may
not represent ultimate reality in Madhyamaka terms, they are not treated as equal. That the last
two steps are connected is also implicit in the Mahāyāna terminology of “two-fold selfless-
ness”: “selflessness of person” refers to the level of Abhidharmic analysis, ascertaining that
there is no personal identity beyond the impermanent constituents of mind and body; “selfless-
ness of dharmas” means that these constituents are also illusory. Both are called “selflessness”
and both go beyond the ordinary conventions, conceptualizing persons performing activities.
Candrakīrti offers a rationale for viable conventions, applicable to both the ordinary and the
more refined types, as discussed previously. For Candrakīrti, comprehensible linguistic expres-
sions consist of mutually dependent elements, not isolated or self-referential items to be poten-
tially coordinated later. This parallels his stance towards those defining traits that are employed
in philosophical treatises in order to categorize entities. In Buddhist Abhidharma, definitions
make entities recognizable for the purposes of “insight meditation” (vipaśyanā), a form of
phenomenological refinement that for Candrakīrti culminates in the insight that those defini-
tions are, in ultimate analysis, only provisionally viable and context bound. Thus, Abhidharma
is linguistically viable, just like ordinary speech, but, being a more refined form of analysis,
Abhidharma also goes beyond ordinary speech and offers the basic rationale for Candrakīrti’s
investigations into ontology. This dual viability maps onto the Buddhist tradition’s general view
that the Buddha spoke in terms of “two truths” conventional and ultimate. In the Abhidharma,
the two truths are interpreted as pertaining to two levels of existence: existence only by way of
a convention and ultimate existence that does not depend on conceptually created conventions.
The two levels of reality are sometimes distinguished in terms of analysis. For instance,
according to Vasubandhu, what can be broken into constituent parts, either in actuality (like a
pot) or through analysis (like minuscule groups of atoms arising together), is only convention-
ally existent; what cannot be analyzed into more basic constituents (like atoms of materiality
or moments of mind) is ultimately existent (Pradhan 1967, 334). This idea is so central to
the Abhidharma project that in some ābhidharmika schools the term “part” (avayava) is a
synonym for an ultimately existent entity, to be contrasted with the “part-possessor” or whole
(avayavin) that is only a concept referring to those ultimately existing entities and does not
409
Mattia Salvini
exist on its own. This is unlike, for example, in the non-Buddhist Nyāya-Vaiśeṣika system,
where the part-possessor is regarded as having a reality beyond and besides the parts.
It is important to note that, according to the Abhidharma, when a part-possessor, such as a
“table” is taken to be something real, this is not a perceptual mistake, that is, not the result of some
sensory malfunction, but rather it is the result of a conceptual mistake, that is, the habit, whether
innate or cultivated through inaccurate philosophical views, of taking one’s recurrent concep-
tual syntheses as matching something “out there.” Thus, while the Nyāya-Vaiśeṣika system may
seem to be in harmony with a more “common sense” view of reality, this, for the Buddhist, is
only in as much as it reifies the habitual mismatches between sensory and conceptual activity,
and does not therefore represent “ordinary perception” in the sense of the way images, sounds,
and so on appear to the mind of even ordinary, untrained people. Rather, it represents a theoreti-
cal solidification of concepts that ordinary people already may have about those appearances.
The Abhidharma discusses the more fundamental level of reality in terms of defining traits,
regarded to be ultimately valid. For example, materiality includes “four great elements” the first
among which is the earth-element; the defining trait of the earth-element is “hardness.”5 The
question then arises as to the ontological status of those defining traits: are they identical to (first
solution) or different from (second solution) the entity that they characterize? Both solutions
offered in the Abhidharma presuppose the reality of the characterized entity. Candrakīrti applies
Nāgārjuna’s examination of the characteristic and the characterized entity to offer a solution
discarding the ultimate reality of either the characteristic or the entity it characterizes (de La
Vallée Poussin 1903–13, 172–73). Thus, Candrakīrti’s solution is that, while conventionally we
may need to speak of both the earth-element and the hardness that characterizes it, it is precisely
this need to conceptualize both that shows how they are correlatives with no ultimate, inde-
pendent existence. They are not just two ways of speaking of the same entity: they are mutually
dependent concepts that make our experience of entities possible while showing that there is
no real entity in the first place once we cognize that they are not conceivable independently.
Abhidharma analyses reach a level where further analysis requires taking into consideration
the mutual dependence of the elements under analysis, which are therefore, for the Madhya
maka, proven unfindable as separate independent entities and thus existent only by way of our
concepts. This last step differs from mereological analysis of larger entities, that is, coarser
conventions, which can take several steps before reaching the type of impasse that discloses
the unreality, rather than the just the reducibility, of the object under examination.
This gradation from ordinary speech to Abhidharma and eventually to ultimate emptiness
can be observed, for example, in Candrakīrti’s analysis of the locus of motion, meant to show
that motion is illusory. He starts by showing that the foot of someone who walks is not a real
singularity, but rather, it can be analyzed into parts, starting from dividing it into toes and so
on down to infinitesimal elements of existence. Those infinitesimal parts may be supposed to
be the ultimate elements that make up the foot, and this is where the Abhidharma would stop.
For the Madhyamaka, however, they can be further analyzed by applying a conceptual alter-
native: is it really partless? Then how does it produce the extended objects that we seem to
perceive? Indeed, if it is partless, it is unextended, like a geometrical point, and it is therefore
difficult to establish extension on the basis of a collection of unextended entities. This alterna-
tive relates to a more fundamental level than the conceptual proliferation of extended objects
(de La Vallée Poussin 1903–13, 93). It deals with the conception of minimal entities existing
without extension and thus not analyzable into parts. When collected into larger groups, these
entities would then be the basis for the perception of extension, according the Abhidharma.
To put it differently, the special feature of Abhidharmic categories (of the type that Candrakīrti
endorses as valid analytical conventions) is that they are only one step removed from the
410
Candrakīrti
final step of analysis before the resultant “non-perception” by the mind. Concepts of coarser
extended entities such as persons or tables are several steps removed from that finer step: they
do not involve interdependence between parts or moments, and they do not even match the
“selflessness of persons,” the level at which one speaks of causal processes but not anymore of
personal identities. I can analyze a table first as a collection of parts of a table, such as the upper
surface, the lower surface, the legs, and so on down to minimally extended entities in space –
that is, partless elements (atoms in the Buddhist philosophical sense). At that point, however,
the analysis will have to shift to a mode unavailable to the Abhidharma, revealing that the atom
is untenable: analysis has exhausted itself in not finding the initial object of analysis anywhere.
Another way to express this is to say that (most) Abhidharmic categories are viable ways
to understand the causal elements of experience. On the other hand, the categories of the non-
Buddhists are invalid analyses of ordinary perceptions and speech, since they rely on explana-
tory elements (such as a permanent “self ”) that exist neither conventionally nor ultimately,
being posited as permanent and substantially real. Said another way, the non-Buddhists con-
fuse sensory data and conceptual synthesis. In other words, non-Buddhist categories start from
common experience and go in the wrong direction of reifying those experiences. Abhidharmic
analysis, on the other hand, goes in the right direction of dissolving concepts about the reality
of entities. While the Abhidharma schools may not be willing to continue the analysis until the
very ideas of “perception” and “entities” are shown to be illusory, in some sense they have,
according to Candrakīrti, done half of the Madhyamaka job.
Candrakīrti highlights how the person-less categories of Abhidharma, the most basic of
which are called “aggregates, bases, and entrances” (skandha, āyatana, and dhātu) are a
necessary didactic tool in disclosing the nature of things:
Just like, although they had destroyed a sense of “I” and “mine,” the Buddhas, the
Bhagavats, in accordance with the world, said “I” or “mine,” for in this way they
could communicate the meaning that they wished to express; in the same way the
aggregates, bases, and entrances were spoken of due to a task to be accomplished.
For, without distinguishing those, there would be no means to introduce the world to
reality (tattva).
(Li and Ye 2014, 142, my translation)
The last sentence is particularly important. Although the analysis of a “person” relies on bodily
and mental constituents that are regarded as ultimately illusory, they need to be employed so
as to show precisely that they, too, and not only the “person,” are illusory.
In addition to miscellaneous passages where he takes a position in Abhidharmic debates,
Candrakīrti composed an independent work offering an overview of Abhidharma, the Treatise
on the Five Aggregates. The overall framework of this treatise is the Sarvāstivāda classification
of entities into “five realities”: form/materiality; mind; mental states; causal factors that belong
neither to materiality nor to the mind; and unproduced, permanent entities. Candrakīrti is
acquainted with several Abhidharmic traditions. He can thus argue that even when Nāgārjuna’s
analysis mentions only the categories familiar to a specific school, its applicability goes far
beyond. For example, when Nāgārjuna analyzes causality in terms of “four conditions” in the
first chapter of his Root Stanzas of the Madhyamaka, Candrakīrti explains that other classifica-
tions of causes and conditions can be subsumed under this four-fold scheme (Salvini 2014).
This point is, I think, worthy of note: Candrakīrti is attentive to the specific technicalities of
Nāgārjuna’s Abhidharmic underpinnings, and yet he is keen to show that the Madhyamaka
refutations are meant to apply far more broadly. In the previous example, this implies that the
411
Mattia Salvini
refutation of the “four conditions” should be taken as a template to refute any variant of the
view that a real effect arises from a real cause different from itself.
Candrakīrti and the sūtras

Without understanding the signless, you said, there is no liberation:
Therefore, that has been taught by you, in its entirety, in the Mahāyāna.6
Nāgārjuna, Praise of the One Beyond the World
In Clear Words is filled with quotes from both non-Mahāyāna (“sūtras from the Āgama”) and
Mahāyāna sūtras (de La Vallée Poussin 1903–13, 548–49). Candrakīrti considers “emptiness” the
ultimate intent of the Buddha’s teachings and argues that it has been taught in both Mahāyāna and
non-Mahāyana texts. While the Mahāyāna sūtras explicitly use the term “emptiness,” Candrakīrti
argues that the sense of non-Mahāyāna passages describing all things as illusory is precisely the
same. Showing familiarity with non-Mahāyāna sūtras from different schools, Candrakīrti, like
Nāgārjuna, is keen to offer a common ground for the Madhyamaka perspective. After quoting
a very famous passage that he interprets as teaching emptiness, Candrakīrti remarks that “this
is found in the treatises and the sūtras of all the schools” (de La Vallée Poussin 1903–13, 549).
Candrakīrti mentions different ways in which the Buddha’s words had been collected by
different schools (Sorensen 1986, 50–51). By taking into account how sūtras were preserved
in different Buddhist lineages, his work displays his work displays an inclusive hermeneutics
in his approach to texts from other Buddhist traditions. Writing with the awareness of a var-
ied audience, Candrakīrti prioritizes rational argumentation over authoritative claims, and his
hermeneutics allows for alternative positions on important topics, such as, for example, dif-
fering Abhidharma accounts of experience. This flexibility is likely connected to his attitude
towards his sources – hermeneutics and reflection on the nature of experience become closely
intertwined.
Candrakīrti’s exegetical standard regarding the Buddha’s words relies on the familiar
dichotomy between the interpretable meaning and the definitive meaning. He quotes two
Mahāyāna sūtras to the effect that whichever passages teach emptiness are to be considered as
articulating the ultimate meaning, while others, taught for the sake of introducing to the path
and speaking in terms of persons, living beings, and so on, require interpretation (MacDonald
2015, Vol. I, 207–9). This exegetical principle determines his use of both Mahāyāna and non-
Mahāyāna sources when supporting arguments in favor of emptiness. For instance, the Rice
Stalk Sūtra contains an elaborate discussion of dependent arising. Candrakīrti relates this sūtra
to the overall structure of the Root Stanzas on the Madhyamaka (Salvini 2011), according to
the following progression: first, the misunderstanding about causality and entities must be
cleared away, which then allows dependent arising to be taught as the template to understand
repeated birth, and this, in turn, results in the relinquishing of views. These three steps are
therefore, for Candrakīrti, a common feature of the sūtras and of the foundational treatise of
his system, which highlights a key feature of Madhyamaka thought: the proper understanding
of causality goes hand in hand with a proper understanding of its illusory nature.
Two truths
All entities bear two natures,
obtaining their state by force of right seeing and false seeing;
412
Candrakīrti
the domain of those who see rightly is reality,

the domain of those who see falsely is explained as conventional truth.
Even those with false seeing are, indeed, accepted as being of two types:
those with clear faculties and those with defective faculties.
Compared to the cognition of those with good faculties,
the cognition of those with defective faculties is considered to be false.
What the world obtains as perceptible for the six faculties

when they are not damaged, that is truth from the perspective of the world itself,
while the rest, that is wrongly imagined, is false from the perspective of the world itself.
Things thoroughly imagined on their own account by the Tīrthas,

who are bewildered by the sleep of ignorance,
and things imagined in illusions, mirages, and so forth
have indeed no reality even according to the world.7
Candrakīrti, Introduction to the Madhyamaka
As we have seen, the theme of two truths is well developed in Buddhist traditions preceding
Candrakīrti. Not surprisingly, he relies on Abhidharmic discussions to interpret Nāgārjuna’s
own two-truths system. According to this system, conventional truth must be relied upon in
order to teach ultimate truth. I understand Candrakīrti’s presentation of the two truths as hav-
ing three salient features:
1. only ultimate truth is truth in a primary sense, while conventional truth is truth only in a
secondary and circumscribed sense (i.e., for those who cannot see ultimate truth);
2. Abhidharmic categories are valid conventions and a stepping stone towards ultimate
truth; non-Buddhist categories are invalid even as conventions;
3. ultimate truth does not entail any positing of existence or non-existence (or of any other
properties) as real properties.
By applying Nāgārjuna’s refutations of entities, Candrakīrti demonstrates that if we accept the

Abhidharmic principle that what cannot resist analysis exists only conventionally and not ulti-
mately and apply it consistently, it follows that nothing exists ultimately. There is nothing that
resists analysis: we saw this point earlier in the discussion about the interdependence between
a characteristic and the entity that it characterizes, as well as in the examination of the locus
of motion and the supposedly ultimate elements of materiality that make up a foot. It therefore
follows that ultimate truth is not ultimate existence, and what is more, any predication of exis
tence or non-existence makes sense only within a conventional context – one which relies on
a network of concepts, and on the presupposition that perception really occurs and validates
something. Thus, conventional “truths” are always based on a false superimposition.
Candrakīrti harmonizes Nāgārjuna’s statements about the ultimate non-reliance on exist-
ence or non-existence with an ontological principle taken from the Abhidharma. This pro-
cedure goes beyond the limits of Abhidharmic analysis to show that while we may talk of
“ultimate truth,” talking of “ultimate existence” is, at best, an occasional metaphor employed
for a specific purpose; thus, the whole idea that conceptual conventions are predicated upon
real existent entities collapses. The two truths are truths in a very different sense – the conven-
tional truth is “true” for those who have not understood its falsity, whereas the ultimate truth
413
Mattia Salvini
is true for those who have realized the falsity of the first, and its truth consists precisely in
realizing the falsity of conventional “truth.” Thus, it is first of all useful to recognize that only
ultimate truth is, properly speaking, truth.
The second point, that is, distinguishing between ordinary conventions, Abhidharmic con-
ventions, and non-Buddhist conventions, is where Candrakīrti goes far beyond Nāgārjuna’s
text that only speaks of the “truth of the conventions of the world” and the “ultimate truth.”
Candrakīrti offers a more elaborate system, spelling out the reasons for such a subdivision.
In Clear Words presents a detailed analysis of the expression “truth of the conventions of
the world.” The term here translated as “conventions” (saṁvr. ti) is analyzed in three different
ways: it can mean “complete obscuration” of the nature of reality; mutual dependence between
entities; and convention, very broadly understood as linguistic expression, transactional usage,
and behavior/attitude towards something (all possible meanings of conventions of the world or
vyavahāra) (Salvini 2019). What I have translated as “world” (loka) is also a semantically rich
term, and Candrakīrti clearly employs different nuances that are not easily captured by any
single English term. First, loka may mean the “world” of ordinary experience. But it can also
mean a “person” in that world, conceptually imputed upon the “five aggregates,” that is, the
finer elements into which Buddhist texts analyze the mental and material constituents of living
beings. Last, Candrakīrti takes loka to mean a “perceiver,” and this allows him to introduce a
distinction not explicitly found in Nāgārjuna: some people in the world do not qualify even as
loka, since they are not genuine “perceivers,” and these “false-perceivers” are of two types –
those with impaired sense faculties and those with distorted views.
We know from related passages on “false conventions” in the Introduction to the Madhya
maka quoted previously that “those with distorted views” refers to non-Buddhist philosophers.
Candrakīrti reiterates in several places that the scholastic definitions and categories used by
non-Buddhist thinkers are to be refuted even conventionally, unlike (most of) the Abhidharmic
definitions, that must be accepted as valid conventions. This point is raised even in the con-
text of criticizing Bhāviveka’s preferred method of argumentation, by which the latter refutes
certain principles of non-Buddhist views by arguing that they cannot exist “ultimately.” For
Candrakīrti, the qualifier “ultimately” is not only redundant but misleading, for Mādhyamikas
should refute non-Buddhist categories even at the level of conventional truth. The reasons
for this have been highlighted previously: non-Buddhist categories go in the wrong direction,
reifying ordinary perceptions instead of refining them towards an understanding of emptiness.
For instance, neither the Vaiśeṣika nor the Sāṁkhya systems can make sense of continued
causality involving sentient “persons” without positing a permanent self, which is a mistake
at the level of conventions themselves. For the Vaiśeṣika, the self is not sentient, but it is the
agent of all actions. For the Sāṁkhya, the self is not active, but it is the only sentient element,
which prompts everything else into activity. In both cases, the entire system, and many of the
categories within it, stand in some necessary relationship with the permanent self and are in
fact geared towards a direct realization of the permanent self on the part of the Vaiśeṣika or
Sāṁkhya practitioner. This means that without a permanent self, these two systems would
collapse. Since a permanent self is a mistaken concept even conventionally, all the other cate
gories in their systems cannot be accepted even conventionally.
The refutation of the Cārvāka system, which also does not accept a permanent self, works
differently. The Cārvāka are “materialist” in the specific sense that they consider sentience
derivative and dependent upon the material constituents of the body, so that the continuum of
consciousness for them is “cut-off ”at death – from which they take the appellation, in Bud-
dhist texts, of “upholding cutting-off ” (ucchedavādin). This is one of the two extremes that
Buddhist teachings steer away from, even according to the very first line of the Stanzas on the
414
Candrakīrti
Madhyamaka, and indeed is often cited as the worst possible mistake. Candrakīrti imagines an
opponent raising the following objection against the Madhyamaka: if, for the Madhyamaka,
in ultimate analysis, there is no rebirth, no karma, no person, and so on, isn’t Madhyamaka a
type of “nihilism” (nāstika), that is, not different from the Cārvāka? Candrakīrti answers by
explaining that the Madhyamaka accepts rebirth and so on at a conventional level, and the
same statement “there is no rebirth” has therefore a completely different sense when uttered by
a nihilist as opposed to a Madhyamaka. Furthermore, if someone were to say that “ultimately,
it is the same” since neither birth nor re-birth exist ultimately, Candrakīrti would respond that
this does not matter, as there is a difference in the cognizer. The example he offers is that of two
people who both report that a third person has committed theft; one of them has witnessed the
theft, and the other has not. Thus, although what the sentence corresponds to may be true, one
of the speakers is uttering a lie, for the context of the utterance implies that they have directly
witnessed the event. Those who deny rebirth both ultimately and conventionally resemble the
liar of that story: although they have not experienced the nature of things, they imply that they
have by speaking about it. When we consider that for Candrakīrti, the level of rebirth and so
on is the level of Abhidharmic analysis, we can see that the passage once more emphasizes the
necessity of going through the Abhidharmic step on the way to the ultimate analysis. We could
also use a different example: stating the conclusion that entities are empty without understand-
ing their interdependence in Abhidharmic terms is akin to writing the correct result of a mathe
matical problem by chance, due to a wrong procedure and a series of calculation mistakes.
Candrakīrti does not spell out the reasons for rejecting the nihilist view. He even states that
Nāgārjuna does not go into a detailed refutation of the nihilist view, because it had already
been successfully refuted in the Abhidharma (Scherrer-Schaub 1991, 121–22). This may indi-
cate one of the reasons Candrakīrti does not offer any detailed refutation of the materialist
view at the conventional level; even most of his own opponents would not have taken that
view very seriously, and Abhidharmic analysis is perfectly sufficient to discard materialist
misconceptions. It should also be noted that a materialist perspective would not be easily har-
monized with the Madhyamaka view of conventional reality as consisting of ultimately empty
concepts. The materialist (in the sense delineated previously) gives an epistemic and ontologi-
cal priority to the material substratum of concepts, making all concepts existentially dependent
upon the material substratum, and this may not square well with Madhyamaka views on inter-
dependence. In other words, this type of materialism could be accused by the Mādhyamika
of falling into a type of essentialist reification of matter, avoided by the Abhidharma dualist
classifications and by its refinement in phenomenological observation.
Another way to understand the two truths is in terms of mistaken perception, unmistaken per-
ception, and non-perception. Mistaken perception, such as the perception arising due to defec-
tive sense-faculties, or the conceptual superimpositions of the non-Buddhist (a self, a primordial
“nature,” etc.), is a wrong type of convention and thus does not even qualify as a conventional
“truth.” Unmistaken perception, such as the perception that arises from well-functioning sense-
faculties, as well as ordinary speech and the analytical categories of the Abhidharma, that are in
harmony with unmistaken perception, constitutes conventional truths. At the level of ultimate
truth, however, perception is itself discarded as a mere convention that cannot withstand analy-
sis, and thus Madhyamaka authors (including Candrakīrti) often speak of “non-perception” and
“non-seeing” as the end result of Madhyamaka analysis. This gradualism may allow us to make
better sense of the role of the Abhidharma within Candrakīrti’s system: the Abhidharma offers
the most refined possible “truth” as long as truth relates to perception, and going beyond the
Abhidharma means discarding perception as a parameter of truth altogether. Thus, while it can-
not be said that Abhidharmic categories are ultimate, their status cannot be compared to that
415
Mattia Salvini
of the ordinary conceptualization of extended entities, such as chariots and persons, which are
liable to be analyzed into constituent parts even within the context of perception.
Abhidharma, furthermore, plays a key role in Candrakīrti’s epistemology of the two truths.
It should be noted that Candrakīrti distinguishes between the functioning of the sense facul-
ties and further conceptualization. It should be clarified, moreover, that Candrakīrti counts
the “sense faculties” as six rather than five, that is, the five well-known sense perceptions
and mental perception or the faculty for the mind to know objects. Mental perception and
further conceptualization, therefore, should be distinguished, and Candrakīrti does so by dis-
tinguishing between the “object of perception” (grāhya) and the “object of conceptualization”
(vikalpita). This helps us understand in which sense Abhidharmic categories can still be said
to relate to (mental) perception, while non-Buddhist categories (such as the part-possessor,
or the “self ”) are excluded from it (despite apparently conforming and expanding upon what
ordinary people may take to be “perception”). In other words, it seems that for Candrakīrti,
there is a clear difference between “the perception of ordinary people” and “what ordinary
people may take to be perception,” since ordinary people may have very little ability to dis-
tinguish between the cognitions that result from the functioning of the six faculties and the
further layers of conceptualization that occur on that basis. We should also not assume that for
Candrakīrti all “ordinary people” are exactly at the same level in all respects; the term “ordi-
nary people,” rather, here refers to anyone who does not have a direct realization of reality,
irrespective of the refinement and soundness of their view.
There is another sense in which Abhidharmic analysis, as opposed to ordinary speech, may
be said to be “on the level of an approximation of the ultimate”: it emphasizes dependent
arising. Candrakīrti’s own definition of conventional truth, as we have seen, is presented in
Abhidharmic terms, relying on the scheme of six sense-faculties. This scheme, as Candrakīrti’s
commentary clarifies, is closely connected with the analysis of causality: in this case, the object
(visible entities) and the sense-faculty (the eye) are the cause for the arising of a specific cogni-
tion (visual consciousness). This framework is an “approximation to the ultimate” because it
is precisely dependence, when extended so as to include the mutual dependence of conceptual
correlatives, that discloses the emptiness of all entities, that is, their unfindability as independ-
ent ultimates. Abhidharmic categories are the results of the analysis of experience in terms of
dependent arising based on perception, while Madhyamaka analysis is based on an extended
application of the same principle, wherein the distinction between notional and existential
dependence collapses. Thus, for the Madhyamaka, it is not only true that “effects depend on
causes” but also that “causes depend on effects,” since there could be no notion of a cause
without having in sight the notion of an effect. For example, a father is not a “father” until we
know that he has children. This, once again, makes causes and effects viable as conventions
precisely because they allow for a step of analysis that shows their mutual dependence and thus
their ultimate emptiness. If causes or effects were to be non-empty, it would entail their having
a non-contingent essence precluding their functioning within causal processes through time.
Accepting the categories of the Abhidharma as a finer level of analysis does not entail dis-
carding all expressions referring to spatio-temporally extended entities; these would remain
“valid fictions,” that is, valid conventional truths, as their bases remain a finer level of per-
ception at the border with non-perception rather than a philosophical reification of sensory
objects. Ordinary speech is not in contradiction with Abhidharmic categories; rather, it is a
helpful shorthand in ordinary interactions.
Thus, Mādhyamikas argue that for something to be dependently arisen, it must be empty
of independent existence, that is, of the type of ultimate existence that entails fixity of nature
and precludes participation in causal change. More directly, Candrakīrti says that the term
416
Candrakīrti
“emptiness” refers to dependent arising rather than to non-existence. He suggests that some-
one may even say that emptiness “exists,” by a conventional superimposition, in order to help
those listeners who may otherwise become alarmed (de La Vallée Poussin 1903–13, 264).
I understand this “conventional superimposition” as being based on a common feature of emp-
tiness and existence: both are not non-existence. Candrakīrti’s metaphorical qualification of
emptiness forcefully clarifies that his view of ultimate truth does not entail non-existence.
Rather, emptiness or non-arising is the very nature of all entities, since “it is never otherwise,”
“it does not depend on something else,” “is not fabricated,” and this ultimate nature “neither
exists nor doesn’t exist,” for, as we have seen, existence and non-existence cannot be ultimate
properties of ultimately ascertainable entities.
Candrakīrti’s epistemology ( pramān a)

Candrakīrti discards Dignāga’s system of two-fold epistemic validation (direct perception
and subsequent inferential validation). Instead, he endorses a system of four valid sources of
knowledge: direct perceptual evidence, subsequent inferential validation, tradition, and com-
parison (MacDonald 2015, Vol. I, 274–75). These four epistemic sources are also accepted as
valid in the Nyāya system. We should, however, be cautious in understanding this as implying
that Candrakīrti is adopting non-Buddhist philosophical categories. This epistemic scheme
may have been perceived to be at a mundane enough level to be acceptable by a number of
different philosophical traditions (as they would all accept, for example, the medical system
of tridoṣa, occasionally referred to in philosophical treatises).8 Furthermore, Dignāga did not
exactly discard tradition and similarity but rather showed how they should be understood as
instances of inferential validation. He explained “subsequent inferential validation” as the
conceptual ascertainment that a certain perceived logical mark is concomitant with a non-
perceived, inferable entity. Thus, the inference that “there is fire on that hill, because there is
smoke” is “subsequent validation,” because it rests on a prior perceptual ascertainment of the
concomitance between smoke and fire. Tradition can function as a valid source of knowledge
only when this rests on a valid inference regarding the reliability of the source of that tradi-
tion: for example, the Buddha must be proven to offer reliable reports even when facts can’t
be independently ascertained by direct perception, and if he cannot be proven thus reliable, his
textual tradition will not be a valid source of knowledge.
Dignāga’s reduction of other epistemic means to just two, that is, direct perception and
inferential validation, is predicated upon an ontology of the particular (lit. “own-characteris-
tic”) vs the universal (lit. “commonality-characteristic”) rejected by Candrakīrti on grounds
discussed earlier. Accordingly, the reduction of all possible epistemic means to two and only
two is not justified, and it should not be surprising that Candrakīrti would re-instate two addi-
tional epistemic sources, in agreement with some pre-Dignāga Buddhist texts that also fol-
lowed a scheme of four epistemic means (MacDonald 2015, Vol. II, 289–91, n. 541). Even
though Candrakīrti rejects Dignāga’s epistemic framework, reflections on the nature of know
ledge play an important role in his philosophy. We have seen that Candrakīrti’s own definition
of valid conventional truths focuses on perception and is primarily concerned with unimpaired
cognition. Nonetheless, Candrakīrti discards the possibility that the perception of ordinary
people could function as a reliable epistemic means to validate ultimately existing entities.
Some later Tibetan Mādhyamikas interpret Candrakīrti so as to harmonize his system with
a significant part of Dignāga’s epistemological innovations. This is achieved by pointing to
passages where Candrakīrti speaks of two possible objects of knowledge, direct and indirect,
and thus qualifies his refutation of a two-fold epistemic scheme as pertaining only to the
417
Mattia Salvini
ensuing ontological claims and not to the two epistemic means in themselves (Ruegg 2000,
272–81; Vose 2009, 9; Thakchoe 2013).
Conclusions
This overview of Candrakīrti’s thought offers tentative solutions to some difficult points in the
interpretation of his system and situates itself among long-standing debates about Candrakīrti’s
intent. Several aspects of Candrakīrti’s philosophy deserve more attentive representation in
English, starting from his engagement with the Abhidharma and with non-Mahāyāna sources
in general. To some extent, both later tradition and modern scholarship have tended to focus on
Candrakīrti’s relationship with a specific epistemological tradition rather than unpacking his
employment of the Abhidharma and its basic categories. This may not be a significant concern
for a readership well trained in Abhidharmic details and their implications, but modern rea
ders may need the Abhidharmic underpinnings of Candrakīrti’s system to be more explicitly
spelled out. This is even more so in the case of Sanskrit grammar and for the likely implica-
tions of his literary style.
Candrakīrti’s philosophical concerns may fruitfully overlap with those of some contempo-
rary Western philosophers and be (partly) comprehensible in terms of ontology, epistemology,
or logic. Yet, Candrakīrti’s intellectual project has a different horizon, different priorities, and
a very different conceptual idiom, not always fitting so easily and naturally into those categori-
zations. While he offers cogent and sophisticated insights into the nature of reality, he does so
independently of, and very differently from, Western intellectual modes (especially scientific
empiricism) and may be called philosophical in a rather different sense, or at least a broader
one, than what many contemporary readers may have come – perhaps somewhat narrowly – to
expect.
Thanks to his massive erudition and literary refinement, Candrakīrti is a bright window into
the world of South-Asian thought. Candrakīrti delves into the conceptual grammar of Buddhist
thought, at the limits of what we can speak of, primarily with respect to “existence” and “non-
existence,” perception and non-perception, and their mutual dependence. As an especially
lucid and engaging exponent of the Madhyamaka, his critique of essentialism has, arguably,
few matches in any tradition. His reflections on the nature of language and conceptualization
are counterintuitive, provocative, and deeply rewarding.
Notes
1 Acknowledgements: I thank Guru Bhikṣu Vāgindraśīladhvaja for important clarifications on key
points of the Madhyamaka view. Part of this work was funded by the Australian Government through
the Australian Research Council Discovery Projects funding scheme, within the project “Ancient
Today” (DP DP220100370); I thank the ARC, and my team-mates – especially, Yasmin Haskell.
2 śāstāraṃ praṇipatya gautamam ahaṃ taddharmatāvasthitān | sambuddhān sakalaṃ jinātmajagaṇaṃ
dharmaṃ ca tair bhāṣitam | cakṣurbhūtam anantabuddhavacanasyālocane dehinām | yo ‘muṃ
madhyamakaṃ cakāra kṛpayā nāgārjunas taṃ name || Madhyamakaśāstrastuti 14, Pandey 1994, 155.
3 na hi śabdāḥ dāṇḍapāśikā iva vaktāram asvatantrayanti, kiṃ tarhi satyāṃ śaktau vaktur vivakṣām
anuvidhīyante || MacDonald 2015, 163.
4 sāmagryā darśanaṁ yatra prakāśayati nāyakaḥ | prāhopacārabhūmiṁ tāṁ paramārthasya buddhimān
|| Bhavasaṁkrāntisūtra, de La Vallée Poussin 1903–13, 120.
5 See, for example: kharaḥ pr. thivīdhātuḥ | Abhidharmakośabhāṣya, Ejima 1989, 12; bhūtānāṁ catuṣṭ
vakhakkhaṭādilakṣaṇāvadhāraṇāt sūtre | Abhidharmakośabhāṣya, Ejima 1989, 37; Candrakīrti also
refers to hardness as the defining trait of the earth element: iha bhāvānām anyāsādhāraṇam ātmīyaṁ
yat svarūpaṁ tat svalakṣaṇaṁ, tadyathā pṛthivyāḥ kāṭhinyaṃ . . . MacDonald 2015, Vol 1, 248.
418
Candrakīrti
6 animittam anāgamya mokṣo nāstīti tvam uktavān | atas tvayā mahāyāne tat sākalyena deśitam ||
Lokātītastava 27, Lindtner 1982, 138.
7 samyagmṛṣādarśanalabdhabhāvaṃ rūpadvayaṁ bibhrati sarvabhāvāḥ | samyagdṛśāṁ yo viṣayaḥ sa
tattvaṁ mṛṣādṛśāṁ saṁvr. tisatyam uktam || mṛṣādṛśe ’pi dvividhā matā hi spaṣṭendriyā doṣavadindriyāś
ca | sadindriyajñānam apekṣya mithyā jñānaṁ mataṁ doṣavadindriyāṇām || vinopaghātena yad
indriyāṇāṁ ṣaṇṇām api grāhyam avaiti lokaḥ | satyaṁ hi tal lokata eva śeṣaṁ vikalpitaṁ lokata eva
mithyā || ajñānanidrāpracalāyamānais tīrthair yathāsvaṁ parikalpitā ye | māyāmarīcyādiṣu kalpitāś
ca teṣām atattvaṃ khalu lokato ’pi || Madhyamakāvatārakārikā 6.34 – 26, Li 2015, 7 and see Salvini
2019, 666 n. 12 (containing an emendation to Li’s edition also followed here).
8 See, for example, Abhidharmakośabhāṣya 3.44.
References
Campbell, John R. B. 2009. “Vajra Hermeneutics: A Study of Vajrayāna Scholasticism in the Pradīpoddyotana.”
PhD diss., Columbia University.
Chakravarti, Chintaharan, ed. 1984. Guhyasamājatantrapradīpoddyotanaṭīkāṣaṭkoṭivyākhyā. Tibetan
Sanskrit Works Series, no. 25. Patna: Kashi Prasad Jayaswal Research Institute.
Chimpa, Lama, Alaka Chattopadhyaya, trans., and Debiprasad Chattopadhyaya, ed. 1990. Tāranātha’s
History of Buddhism in India. New Delhi: Motilal Banarsidass (Reprint: First Edition, Simla 1970).
de La Vallée Poussin, Louis, ed. 1903–13. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna
avec la Prasannapadā Commentaire de Candrakīrti. St. Petersbourg: Académie Impériale des
Sciences.
———. 1912. Madhyamakāvatāra par Candrakīrti. Traduction tibétaine. Bibliotheca Buddhica 9. St.-
Pétersbourg: Académie impériale des sciences.
Dreyfus, Georges B. J., and Sara L. McClintock, eds. 2003. The Svātantrika-Prāsaṅgika Distinction.
Boston: Wisdom Publications.
Ejima, Yasunori, ed. 1989. Abhidharmakośabhāṣya of Vasubandhu, Chapter 1: Dhātunirdeśa. Tokyo:
Sankibo Press.
Erb, Felix, ed. and trans. 1997. Śūnyatāsaptativr.tti. Stuttgart: Franz Steiner Verlag.
Hong, Luo, and Toru Tomabechi, eds. 2009. Candrakīrti’s Vajrasattvaniṣpādanasūtra (Vajrasattvasādhana).
Beijing and Vienna: China Tibetology Publishing House/Austrian Academy of Sciences.
Huntington, C. W., trans. 1989. The Emptiness of Emptiness. Honolulu: University of Hawaii Press.
Johnson, Dennis A. 2012. Refuting the Conditioned: The Saṁskr. tārthapratiṣedha of Candrakīrti’s
Catuḥśatakaṭīkā: Introduction, Translation and Summary together with a Critical Edition of the
Tibetan Translation. Wien: University of Vienna (unpublished Magister der Philosophie thesis).
Joshi, S. D., and J. A. F. Roodbergen, eds. and trans. 1986. Patañjali’s Vyākaraṇa-Mahābhāṣya.
Paśpaśāhnika. Introduction, Text, Translation and Notes. Pune: University of Poona.
Kanō, Kazuo, and Xuezhu Li. 2014. “Sanskrit Verses from Candrakīrti’s Triśaraṇasaptati Cited in the
Munimatālaṁkāra.” China Tibetology 1: 4–11.
Lang, Karen C., trans. 2003. Four Illusions: Candrakīrti’s Advice for Travelers on the Bodhisattva Path.
Lasic, Horst, Li, Xuezhu, and MacDonald, Anne, ed. 2022. Candrakīrti’s Madhyamakāvatārabhāṣya
Chapters 1 to 5. Critically and diplomatically edited by Horst Lasic, Xuezhu Li & Anne MacDonald
on the basis of preparatory work by Helmut Krasser. Beijing/Vienna: China Tibetology Publishing
House/Austrian Academy of Sciences Press.
Li, Xuezhu. 2015. “Madhyamakāvatāra-kārikā: Chapter 6.” Journal of Indian Philosophy 43 (1): 1–30.
Li, Xuezhu, and Kazuo Kanō. 2015. “Critical Edition of the Sanskrit Text of the Munimatālaṁkāra,
chapter 1 (fol. 48r4–58r5): Explanation of skandha, dhātu, and āyatana based on Candrakīrti’s Pan-
caskandhaka.” The Mikkyo Bunka 234.
Li, Xuezhu, and Shaoyong Ye. 2014. Yuktiṣaṣṭikākārikā: Editions of the Sanskrit, Tibetan and Chinese
Versions, with Commentary and a Modern Chinese Translation. Shanghai: Zhongxi Book Company.
Lindtner, Chr, ed. 1979. “Candrakīrti’s Pañcaskandhaprakaraṇa.” Acta Orientalia 40: 87–145.
Lindtner, Chr, ed. and trans. 1982. Nagarjuniana: Studies in the Writings and Philosophy of Nagarjuna.
Copenhagen: Institut for indisk filologi.
Loizzo, Joseph J., and The AIBS Translation Team, trans. 2007. Nāgārjuna’s Reason Sixty with
Candrakīrti’s Reason Sixty Commentary. Edited by Robert A. F. Thurman, Thomas F. Yarnall, and
419
Mattia Salvini
Paul G. Hackett. Critical Editions by Joseph John Loizzo and Paul G. Hackett. New York: American
Institute of Buddhist Studies, Columbia University.
MacDonald, Anne, ed. and trans. 2015. In Clear Words: The Prasannapadā, Chapter One. Vienna: Aus-
trian Academy of Sciences Press.
Pandey, J. S., ed. 1994. Bauddhastotrasaṁgraha. New Delhi: Motilal Banarsidass.
Pradhan, Prahlad, ed. 1967. Abhidharma-Koshabhāṣya of Vasubandhu. Patna: K. P. Jaisawal Research
Institute.
Ruegg, David S. 2000. Three Studies in the History of Indian and Tibetan Madhyamaka Philosophy.
Wien: Universitat Wien.
———, trans. 2002. Two Prolegomena to Madhyamaka Philosophy. Wien: Universitat Wien.
Salvini, Mattia. 2011. “The Nidānasamyukta and the Mūlamadhyamakakārikā: Understanding the Middle
Way Through Comparison and Exegesis.” Thai International Journal of Buddhist Studies 2: 57–95.
———. 2014. “Dependent Arising, Non-Arising, and the Mind: MMK1 and the Abhidharma.” Journal
of Indian Philosophy 42 (4): 471–97.
———. 2016. “A Clear and Learned Guide in Reading Candrakīrti’s Prasannapadā.” Bulletin de l’École
française d’Extrême-Orient 102: 415–35.
———. 2019. “Etymologies of What Can(not) Be Said: Candrakīrti on Conventions and Elaborations.”
Scherrer-Schaub, Cristina A., ed. and trans. 1991. Yuktiṣaṣṭikāvr. tti. Commentaire à la soixantaine sur le
rasoinnement ou du vrai ensaignement de la causalité par la Maître indien Candrakīrti. Bruxelles:
Institut Belge de Haute Études Chinoises.
Shastri, H., ed. 1914. “Catuḥśātikā of Āryadeva.” Memoirs of the Asiatic Society of Bengal 3 (8): 448–514.
Skilling, Peter. 2019. “Many Lands of Gold.” In The Golden Land Suvarnabhumi. The New Finding for
Suvarnabhumi Terra Incognita, edited by Bunchar Pongpanich and Somchet Thinapong, 191–217.
Bangkok: GISTDA and BIA.
Skilling, Peter, and Saerji. 2013. “Candrakīrti and the Pūrvaśailas: A Note on Triśaraṇasaptati v. 51.”
Annual Report of the International Research Institute for Advanced Buddhology at Soka University
for the Academic Year 2012 16: 267–72.
Sorensen, Per K. 1986. Candrakīrti, Triśaraṇasaptati: The Septuagint on the Three Refuges. Wien: Wie-
ner Studien zur Tibetologie und Buddhismuskunde.
Suzuki, Kōshin. 1988. “A Study of Candrakīrti’s Bodhisattvayogācāracatuḥśatakaṭīkā II, A Critical Edi-
tion of the Tibetan Text with the English Translation.” Sankōbunka-kenkyūjo-nenpō 20: 1–26.
———. 1989. “A Study of Candrakīrti’s Bodhisattvayogācāra-catuḥśatakaṭīkā I, A Critical Edition of
the Tibetan Text with the English Translation.” Taishōdaigaku Daigakuin Kenkyūronshū 13: 1–12.
———, ed. 1994. Sanskrit Fragments and Tibetan Translation of Candrakīrti’s Bodhisattvayogācāraca
tuḥśatakaṭīkā. Tokyo: Sankibo Press.
Thakchoe, Sonam. 2013. “Prāsaṅgika Epistemology: A Reply to Stag tsang’s Charge Against Tsong-
khapa’s Uses of Pramāṇa in Candrakīrti’s Philosophy.” Journal of Indian Philosophy 41: 535–61.
Tillemans, Tom J. F., ed. and trans. 1990. Materials for the Study of Āryadeva, Dharmapāla and
Candrakīrti. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
Vose, Kevin A. 2009. Resurrecting Candrakīrti. Disputes in the Tibetan Creation of Prāsaṅgika. Somer-
420
PART 6
Middle-Period Commentators
(Eighth–Ninth Century)
MIDDLE-PERIOD
COMMENTATORS
Following the development and articulation of Madhyamaka and Yogācāra as distinct and
coherent philosophical schools, with thinkers from both traditions engaging Dignāga and
Dharmakīrti, in the eighth century, we see efforts to take stock of the diversity of philosophical
positions. These efforts took different approaches, some critical and some syncretic, to address
the hermeneutical and philosophical issues that such diverging orientations and interpretations
posed to the coherence of the Buddhist tradition. At times, this meant presenting a path leading
through a hierarchy of views from the lowest to the most liberating wisdom. At other times, it
meant offering new interpretations that could bend the meaning of root texts in creative ways.
It could also involve the creation of new conceptual and exegetical tools aimed at solving
problems posed by Buddhist and non-Buddhist opponents.
We see one version of a philosophical synthesis together with the development of original
philosophical positions in Śubhagupta (eighth century), who draws on the atomic theories
found in Vaibhāṣika and Sautrāntika traditions even as he is firmly rooted in the logic and
epistemology of Dignāga and Dharmakīrti. Serena Saccone explores how Śubhagupta, while
mostly interpreting and elaborating on Dharmakīrti’s thinking, develops his own ontological
and epistemological views. This is especially true in his Verses on the Demonstration of Exter-
nal Objects (*Bāhyārthasiddhikārikā), in which he defends a form of externalism in contrast
to the ontological idealism of Dignāga and Dharmakīrti even as he employs their methods of
argument. In the same text, Śubhagupta argues against both the Sautrāntikas and the idealists
and defends the view that sense cognitions are not shaped by the image of their objects.
Śubhagupta’s student, Dharmottara (eighth century), wrote his own independent treatises,
but he is primarily known for his commentaries on Dharmakīrti’s writings on epistemology.
David Nowakowski, in his chapter, focuses on Dharmottara’s commentary on Dharmakīrti’s
Epitome of Reason (Nyāyabindu). While Dharmottara presents himself as faithfully interpret-
ing Dharmakīrti’s thought, in resolving tensions and constructing a more coherent and sys-
tematic account, Dharmottara nonetheless introduces novel ideas. In particular, Nowakowski
spotlights how Dharmottara’s epistemology and philosophy of mind blur Dharmakīrti’s clear
distinction between perception, inferential reason, and inference-for-others. While he draws
on Dharmakīrti’s resources, Dharmottara’s epistemology is much more cognizer relative, as
he emphasizes the ways in which both perception and inference are themselves dependent
on the particular conditions of individual cognitive agents. This rejection of Dharmakīrti’s
DOI: 10.4324/9781351030908-37 423

Middle-Period Commentators
more impersonal account of cognitive processes proved influential for some later Buddhist
philosophers.
Ryusei Keira, in his chapter, discusses Jñānagarbha (eighth century), author of the Com-
mentary on the Distinction Between the Two Truths (Satyadvavibhaṅgavṛtti). Following
Bhāviveka and informed by Dharmakīrti’s epistemology, Jñānagarbha is perhaps the first
Indian Buddhist philosopher to explicitly integrate Yogācāra views into a Madhyamaka frame-
work. For example, Jñānagarbha employs the idea of “mind-only” (cittamātra) to argue for the
selflessness of external objects. Keira argues that this moment of Yogācāra thought illustrates
Jñānagarbha’s gradualism, as it is only one stage in what will ultimately be a Madhyamaka
account that shows the selflessness of mind. Keira is particularly interested in Jñānagarbha’s
account of the two truths, how it is different from the account of the two truths found in other
prominent Madhyamaka thinkers, and the diverse Tibetan interpretations of Jñānagarbha’s
distinction between the conventional and the ultimate.
Śāntarakṣita (ca. 725–788) is said to have been Jñānagarbha’s student and wrote a subcom-
mentary ( pañjikā) on his teacher’s Commentary. He also employs Dignāga and Dharmakīrti’s
logic and epistemology and is the first prominent Mādhyamika philosopher to produce a com-
mentary on a text by Dharmakīrti. Allison Aitken, in her chapter on Śāntarakṣita, traces his
syncretic method that was to become influential for much Indian and Tibetan Buddhist phi-
losophy. This syncretic method employs and then rejects one view after another, culminating
in the Madhyamaka relinquishing of all views. For example, in his Ornament of the Middle
Way (Madhyamakālaṃkāra), Śāntarakṣita rejects the direct realist theory of perception asso-
ciated with Vaibhāṣikas by using arguments from the Sautrāntika representationalist theory
of perception. But then he adopts the Yogācāra idealist ontology – which, together with the
Yogācāra understanding of cognition as reflexively aware, he believes to be conventionally
true – to critique that same Sautrāntika representationalism. In the end, while the Yogācāra
position may be helpful for freeing us from misguided views of the reality of external objects,
Śāntarakṣita employs it instrumentally and leaves it behind with his famous articulation of
the neither-one-nor-many argument for the Madhyamaka ultimate truth of the emptiness of
intrinsic nature, including the emptiness of the Yogācāra concept of mind.
Continuing this lineage, Śāntarakṣita’s prolific student Kamalaśīla (ca. 740–795) shows
even greater enthusiasm for integrating Yogācāra and Madhyamaka thought, often bolstered
by elaborate arguments rooted in Dharmakīrti’s epistemological theories. In the following
chapter, Sara McClintock presents an overview of Kamalaśīla’s oeuvre, showcasing his range
from philosophical treatises to sūtra commentaries to meditation and monastic training manu-
als to epistles. Like his teacher Śāntarakṣita, Kamalaśīla is unusual in this volume in that we
have historical records testifying to his existence as the result of his sojourn in Tibet, where he
traveled at the invitation of the king and where he engaged in teaching and debate. Some of
his works appear to have been written in Tibet and may also therefore represent his attempt to
transmit the full range of scholastic knowledge of Nālandā, the famed monastic university in
North India with which he is associated. Throughout his many writings, however, a consistent
theme is the importance of understanding the conventional realities that allow reason to func-
tion. McClintock argues that Kamalaśīla sees even wisdom as a “necessary illusion” which the
bodhisattva must embrace if they are to be of help to the world.
Another of Śāntarakṣita’s students, Haribhadra (eighth century), also draws on diverse
Buddhist philosophical traditions in a syncretic fashion. Pierre-Julien Harter, in his chapter,
focuses on Haribhadra’s engagement with the Perfect Wisdom (Prajñāpāramitā) tradition. Har-
ibhadra exemplifies those commentators who want both to appear subservient to texts and to
manifest a creativity that opens new philosophical horizons. According to Harter, we see this
424
in the alliance Haribhadra forges between the sūtras of Perfect Wisdom and the Ornament of
Realizations (Abhisamayālaṃkāra), of which he became a champion. This exegetical innova-
tion is at the same time a philosophical innovation since it updates the Perfect Wisdom by
giving it what Harter calls a “voice” and by giving to the Buddhist path a treatise (śāstra) and
a place within philosophical debates.
Śāntideva (eighth century) is the Indian Buddhist author whose work is most widely read
by contemporary scholars as a contribution to moral philosophy. Both of Śāntideva’s surviving
texts – the Introduction to the Practice of Awakening (Bodhicaryāvatāra) and the Compen-
dium of Trainings (Śikṣāsamuccaya) – make significant contributions to understanding the
moral dimensions of mindfulness, concentration, virtue, well-being, and the relation between
metaphysics and ethics. Stephen Harris, in his chapter, is particularly interested in Śāntideva’s
analysis of the virtues of generosity, patience, and wisdom and its role in the cultivation of
the other virtues of a bodhisattva. He also focuses on bodhicitta, or “the mind dedicated to
full awakening” and the vow to become a bodhisattva. Harris emphasizes the ways in which,
throughout Śāntideva’s account of cultivating the virtues of a bodhisattva, we see the conver-
gence of altruism and self-interest, that attending to the good of the other is simultaneously to
attend to our own most fundamental good.
In the final chapter in this section, Shinya Moriyama turns to another commentator on
Dharmakīrti, Prajñākaragupta (eighth–ninth century). Moriyama focuses on the most distinc-
tive elements of Prajñākaragupta’s thought in ontology, epistemology, and philosophy of reli-
gion. He begins with Prajñākaragupta’s theory of “backward causation,” according to which
future entities function as causes of present entities. Moriyama then discusses Prajñākaragupta’s
ontological idealism, which equates existence with being perceived and which informs his
epistemology. Then Moriyama turns to Prajñākaragupta’s expansive account of the Buddha’s
omniscience, including the knowledge of all past and future entities. He ends with a discussion
of Prajñākaragupta’s doctrine of nonduality, the culmination of the Buddhist path of wisdom.
425
25
ŚUBHAGUPTA
An Externalist Outsider Within the
Dharmakīrtian Tradition
Serena Saccone
Introduction
Śubhagupta (Tib. Dge [b]srung[s]) was an eighth-century philosopher in what is referred to as
the Buddhist tradition of logic and epistemology ( pramāṇa). Much of his intellectual work is
related to the discussion of Dharmakīrti’s theories. However, unlike some figures in this tradi-
tion, he is not strictly an exegete of the master’s works. Rather, he systematizes, elaborates on,
and, in some cases, even refutes Dharmakīrti’s doctrines.
Śubhagupta’s thinking is distinguished by his syncretic method of drawing on the views
of different Buddhist traditions to conceive his own original ontological and epistemological
theories. This may be why his doctrinal affiliation was, and has continued to be, debated.1
Śubhagupta’s innovative views are especially apparent in the Verses on the Demonstration of
External Objects (*Bāhyārthasiddhikārikā), his magnum opus, which remains his most valu-
able work by far. Particularly, in this text, he tries to validate, within the “logico-epistemolog-
ical” tradition, a form of externalism and a form of nirākāravāda (namely the doctrine that
cognitions are not shaped by the images of their objects). It is this specific theory on external-
ism (bāhyārthavāda) on the ontological level (as well as the nirākāravāda standpoint on the
epistemological level) that is his most distinctive position.
Śubhagupta’s novelty as an author is also reflected in his adopting the genre of short digests,
arguably for the first time within the Buddhist epistemological tradition. He composed brief
“manuals for debate,” organizing them by single subjects that were relevant at the time. These
writings are indeed synopses of arguments and theories that are already found (in part or in
whole) in works by previous Buddhist authors, especially Dharmakīrti. However, Śubhagupta
articulates (and reworks) them as effective proofs to use against non-Buddhist opponents and
to establish Buddhist truths. Most likely, these compendia were intended for pedagogical rea-
sons, in order to instruct the monks. At the same time, they could also have been compiled
with the aim of fostering the monks’ own rational conviction in the truths defended therein.
This holds true especially with regard to Śubhagupta’s four other extant works, though less so
for the Demonstration of External Objects. This “literary genre” was continued by subsequent
authors, such as Jitāri (tenth century).
As for specific contributions to the logico-epistemological tradition, Śubhagupta also inno-
vated by stimulating further refinement of the theories of other Buddhist philosophers, both
DOI: 10.4324/9781351030908-38 427

Serena Saccone
contemporaries and successors, who responded to his criticisms. One example of such innova-
tion can be found in the “heterodox” views that are expounded in the Demonstration of Exter-
nal Objects. This text serves to refute some of the principal doctrines of what we could call
the “mainstream” logico-epistemological tradition, particularly the “idealistic” (vijñānavāda)
turn represented by the Vasubandhu (ca. 350–430) -Dignāga (ca. 480–540) -Dharmakīrti (ca.
550–660?) lineage. However, his other extant works contain little innovation in this regard
and appear to follow the mainstream ideas of this lineage and the Dharmakīrtian tradition of
the period.
The Demonstration of External Objects has been the focus of study and investigation by a
few contemporary scholars, but his four other works have garnered very little attention. This
is probably the reason Śubhagupta is regarded as a fierce adversary of the views of the “main-
stream” logico-epistemological lineage and particularly of Dharmakīrti. And yet, it would be
a mistake to overlook his doctrinal debt to the latter.
Biography
Little is known about Śubhagupta’s life. He certainly lived after Dharmakīrti and was most
likely an earlier contemporary of Śāntarakṣita (ca. 725–788), as he refers to the former’s views
and is quoted by the latter. There are different conjectures among contemporary scholars
regarding his exact dates (on this, see Saccone 2018, 2019a). However, the dates provided by
Frauwallner (1961, 147), that is, 720–80, are the most likely, because they take into account the
possibility that he was Dharmottara’s (ca. 740–800) teacher. The Tibetan historian Tāranātha
(1575–1634) reports that Śubhagupta lived at the time of Dharmapāla (ed. Schiefner 1868,
166, 15–17), whose reign has been dated ca. 775–812 (Sanderson 2009, 87). Accordingly, he
most likely lived in the mid-eighth century, perhaps in northeastern India.
Intellectual Activity
Singularity Within the Tradition: Śubhagupta’s Relation

to Other Authors
Based on Pocock’s theory of “contextualism” in defining an author’s intellectual activity,
I believe that Śubhagupta’s singularity as a thinker must be determined primarily through the
way his works and ideas were received, interpreted, and used within the tradition in which he
was active (i.e., the Buddhist tradition of logic and epistemology). In other words, his intellec-
tual biography requires delineating his unique contributions to the history of Buddhist thought.
Such contributions are particularly evident in analyzing the writings of his contemporar-
ies, such as Śāntarakṣita and Kamalaśīla (ca. 740–795). The two authors appear to follow in
Śubhagupta’s footsteps, both – to some extent – doctrinally and formally, in their extensive
and doxographical magna opera, the Compendium of Truths (Tattvasaṃgraha) and the Com-
mentary on the Compendium of Truths (Tattvasaṃgrapañjikā), respectively.
For instance, it seems plausible that, with reference to doctrines connected to the debates
against Brahmanical opponents, Śāntarakṣita and Kamalaśīla followed Śubhagupta’s elabora-
tion of arguments. This is the case with the proof of the possibility of omniscience for human
beings, which is connected with the reliability of scriptures written by human authors. Omnis-
cience here is construed as the knowledge of those truths that are the object of scriptures and
are extrasensory. Śāntarakṣita and Kamalaśīla’s debt to Śubhagupta holds true also with regard
to the refutation of a creator God understood as a real and permanent entity.
428
Śubhagupta
The number of parallel arguments emerging from the comparison of the Compendium and
the Commentary and some of Śubhagupta’s writings, such as the Verses on the Demonstration
of the Omniscient One (*Sarvajñasiddhikārikā) and the Verses on the Refutation of God (*Īśva
rabhaṅgakārikā), is remarkable. If, as I hypothesize (Saccone 2019b, Forthcoming), all of
Śubhagupta’s works were composed before those two, then Śāntarakṣita and Kamalaśīla owe
a great deal to their earlier contemporary Śubhagupta. This would be all the more evident in
Kamalaśīla’s work. The latter thinker explicitly refers to Śubhagupta by name many times,
calling him “venerable” (bhadanta) and “master” (*ācārya/slob dpon) in two different works.
Some of the formulations and overall ideas that are present in the three authors are, neverthe-
less, part of a common background for many Buddhist thinkers; certainly, most of them are to be
found, either condensed or in full, in some previous Buddhist authors, particularly Dharmakīrti.
However, to the best of my knowledge, the specific way Śubhagupta articulates these notions to
design proofs against non-Buddhist opponents is unique (Saccone 2019b, Forthcoming). This
is particularly the case for the demonstration of the Buddha’s omniscience (Saccone Forthcom-
ing). Accordingly, I believe that the parallel arguments that are found in Śāntarakṣita’s and
Kamalaśīla’s works betray the direct influence of Śubhagupta’s ideas.
On the other hand, Śāntarakṣita and Kamalaśīla also make a point of refuting the epistemo-
logical and ontological views that are found in the Demonstration of External Objects. They do
so especially in the Examination of External Objects chapters of the Compendium and the Com-
mentary. Śāntarakṣita and Kamalaśīla are the representatives of a mainstream current within the
logico-epistemological school that defended the view of “consciousness-only” (vijñaptimātratā).
According to the latter standpoint, everything can be explained as the mere occurring of con-
sciousness, one after the other. Śubhagupta, on the other hand, endorses an externalist position
on ontology and epistemology, being the obvious target of their criticism. Kamalaśīla, in particu-
lar, shapes his defense of the Dharmakīrtian “invariable co-apprehension” (sahopalambhaniya-
ma) argument (see next section and chapters of this volume on Dharmakīrti, Ratnakīrti, and
Devendrabuddhi and Śākyabuddhi) as a refutation of Śubhagupta’s attacks on it, thus testifying
to the significance of those attacks in the period. This provides additional evidence of how
Śubhagupta’s ideas stimulated the further refinement of theories within the tradition.
Śubhagupta’s main intent with his digests appears to be that of compiling “manuals for
debate” – namely instructions on how to defeat the opponents as well as (perhaps) develop-
ing a personal and rational conviction in Buddhist truths. In this sense, he sets a precedent for
the chapters of the Compendium and the Commentary, with their similar scope. That this was
Śubhagupta’s goal may be confirmed by the third verse of his Demonstration of the Omnisci-
ent One. This verse serves as a declaration of intent informing at least four of his works, if
not also his magnum opus: “And that fool who says ‘everything is produced by God (Īśvara)’
should be clearly told the following arguments, in an assembly in order to dispel ignorance.”2
Here, Śubhagupta is explicitly introducing his text as a summary of arguments to use against
opponents in an assembly in order to dispel ignorance (perhaps to be intended as both that of
the opponents and that of the Buddhist monks themselves).
Among the other authors within the tradition of logic and epistemology, Śubhagupta is also
doctrinally associated with Arcaṭa and Dharmottara. Dharmottara describes Śubhagupta (and
Dharmākaradatta, another name for Arcaṭa) as a celebrated master and declares that his know
ledge is, to some extent, the source of all well-conceived views expounded in his Commentary
on the Ascertainment of the Means of Knowledge (Pramāṇaviniścayaṭīkā).3 To the best of my
knowledge, there is no source of information regarding Śubhagupta’s relation with Arcaṭa.
Regarding Śubhagupta’s association with Dharmottara, there are a few ancient sources that
testify to their connection to a certain extent (on these sources, see Saccone 2019a, 459). In his
429
Serena Saccone
History of Buddhism, for instance, Bu ston (1290–1364) states that Dharmottara was the pupil
of Arcaṭa and Śubhagupta.4 In modern scholarship, the subject of the relationship between
Śubhagupta and Dharmottara as well as Arcaṭa, though it merits analysis, has not been thor-
oughly investigated (for a more extensive bibliography on this, see Saccone 2018, 2019a). In
fact, Matsumoto (1980, 278–77) regards Śubhagupta’s and Dharmottara’s teacher-pupil (guru-
śiṣya) relationship as impossible due to Dharmottara’s use of the same proof as Kamalaśīla
while defending the “invariable co-apprehension” argument against Śubhagupta’s refutation.
Writings
Only five works, all of which are found in the Tibetan Canon, are explicitly ascribed to Śubhagupta.
These are the (1) Verses on the Demonstration of the Omniscient One (*Sarvajñasiddhikārikā)
(Tōhoku no. 4243), (2) Verses on the Demonstration of External Objects (*Bāhyārthasiddhikārikā)
(Tōhoku no. 4244), (3) Verses on the Investigation of the Vedic Revelation (*Śrutiparīkṣākārikā)
(Tōhoku no. 4245), (4) Verses on the Examination of the Exclusion from Others (*Anyāpoha-
vicārakārikā) (Tōhoku no. 4246), and (5) Verses on the Refutation of God (*Īśvarabhaṅgakārikā)
(Tōhoku no. 4247). They are preserved in Tibetan only, but many Sanskrit fragments have been
found in the texts of other authors, particularly in Kamalaśīla’s Commentary. Śubhagupta’s extant
works are all written in verse (likely providing the monks a mnemonic device for debates) and
center on single themes. Those were arguably the most controversial of their time: Vedic revela-
tion (śruti), the existence of an omniscient being, the reality of God, and the theory of exclusion
(apoha) from others. Another work attributed to him, namely the Demonstration of Selfless-
ness (*Nairātmyasiddhi) (Frauwallner 1957, 100), is now lost. Steinkellner (1985, 216–18) also
regards the Demonstration of the Other World (*Paralokasiddhi) and its commentary, both lost, as
his. Based on the existence of prose passages in Kamalaśīla’s Commentary that are introduced as
direct quotations of Śubhagupta, Frauwallner (1933) had surmised the existence of another non-
extant work of his, that is, an auto-commentary on the Demonstration of External Objects, which
was never translated into Tibetan. I have argued against that assumption (Saccone 2014, 2018).
The chronology of Śubhagupta’s works has not been established with certainty. Three works
of his, the Examination of the Exclusion, the Investigation of the Vedic Revelation, and the
Demonstration of the Omniscient One, appear to form a trilogy. Each of them begins with the
word “therefore” (de’i phyir), referring to topics that had been treated in an earlier writing. Due
to the character of the arguments they contain, as well as the comparison with analogous argu-
ments in some chapters of the Compendium and the Commentary, the chronological succession
of the three could be in the order listed previously.5 The Verses on the Refutation of God and
the Demonstration of External Objects are Śubhagupta’s only writings that are not connected
to his earlier works. Accordingly, one may conjecture that the Verses on the Refutation of God
is the earliest (among those that are known to us). The Demonstration of External Objects, in
turn, being his magnum opus and his most original writing, is arguably later, if not the latest.
The “Orthodox” Works

As mentioned previously, the Refutation of God starts with a verse explicitly introducing it as a
summary of arguments to use against opponents in an assembly in order to dispel ignorance. This
work is constructed as a compendium of proofs against those who maintain that the universe is
created by God. Śubhagupta’s opponents can be identified as representatives of the Brahmani-
cal traditions of Nyāya and Vaiśeṣika, as well as of some Śaiva sources, while God is intended
mainly as Śiva.6 Śubhagupta does not show any striking doctrinal originality in this work; he
430
Śubhagupta
primarily refers back to some argumentations and views already developed by other Buddhist
philosophers in the past – such as those of Dharmakīrti (especially in the Commentary on the
Means of Knowledge) and Vasubandhu (in the Commentary on the Treasury of the Abhidharma).
At the same time, the specific use and articulation of the evidence can be viewed as his own origi-
nal contribution. Many parallel formulations of these arguments are also found in Śāntarakṣita’s
Compendium and Kamalaśīla’s Commentary. Here, Śubhagupta is mainly interested in rebutting
those properties of permanence and unity that his opponents attribute to God – a God that creates
a universe that is impermanent and manifold. The creation of the universe involves a succession
of different phases, and a permanent and unitary being is, ex hypothesi, always identical to itself;
accordingly, such succession would not be possible. Another undesirable consequence would
be the impossibility of destroying ignorance, which, being created by an eternal cause, would
be permanent. This would make any spiritual path towards liberation, whether Buddhist or non-
Buddhist, aimless. The upholder of the existence of a creator God tries to avoid the contradiction
of a permanent God creating an impermanent universe by affirming that God’s will is not perma-
nent. However, Śubhagupta responds that a quality (i.e., will) of a substratum that is permanent
must itself be permanent. Otherwise, the substratum itself cannot be admitted as such, since,
having that quality only momentarily and not always, it would change based on its presence or
absence. Issues arising with the concept of God’s will are also analyzed. If God is admitted by
the opponents as permanent and always identical to itself, it is impossible to explain why his will
to create arose at a certain point in time and was not always there since the beginning. Moreover,
Śubhagupta argues, if God has a purpose, he is lacking something and accordingly is not perfect.
If he does not have a purpose, on the other hand, he acts irrationally.
According to the hypothetical sequence of works, the Refutation of God was followed by
the “trilogy.” These three works center on ideas related to the philosophy of language and con-
cept formation; they elaborate greatly on Dharmakīrti’s views as found in the Commentary on
the Means of Knowledge. This is evident in the first text, the Examination of Exclusion, which
discusses the theory of exclusion (apoha), the theory of linguistic and conceptual exclusion
already found in Dignāga and elaborated upon in Dharmakīrti. Some things, though ultimately
being different from each other, produce the same effects, such as similar cognitions; this is
why they are all determined through the same concept (and denoted by the same word) – not
because they have the same nature.
The Investigation of the Vedic Revelation7 is intended as a refutation of the Vedic revela-
tion (as synonymous with veda, see Eltschinger 1999, 48), in other words, a refutation of
the authority of the Brahmanical scriptures. Exactly as in the Demonstration of the Omnis-
cient One, Śubhagupta reuses (with slight modifications) some of the linguistic theories
expounded in the chapter entitled Inference for Oneself (Svārthānumāna) in Dharmakīrti’s
Commentary on the Means of Knowledge. This is done in order to refute some Brahmani-
cal (namely Mīmāṃsā’s) linguistic conceptions that lie at the core of the justification of the
Veda’s authority. The Mīmāṃsā school defends the idea that the Veda does not have any author
(apauruṣeyatā) and its words, being eternal, are connected to their object by their own nature.
Against this view, Śubhagupta argues that words are potentially fit to refer to any object. They
refer to the speaker’s intended meanings, which are concepts and not external objects. The
understanding of a thing by means of a word arises due to a convention, not because of any
natural relation between that thing and the word. Thus, since words found in the Veda, like
all words, are dependent on human-based conventions and intentionality, the Veda cannot be
viewed as reliable by itself, that is, merely because it is not produced by human beings.
Similarly, in the first part (vv. 1–12)8 of the Demonstration of the Omniscient One, which is
devoted to the proof of the Buddha’s omniscience, Śubhagupta employs some of Dharmakīrti’s
431
Serena Saccone
linguistic theories. He aims to provide arguments against the same Brahmanical opponents
(exponents of the Mīmāṃsā) and their theory of the absence of a human author of the Veda.
Śubhagupta links the authoritativeness ( prāmāṇya) of the Veda to the omniscience of its
author. He aims at demonstrating the necessary relationship between the truth of scriptures
and the perception of extrasensory objects (atīndriyārthadarśana) by their author. Scriptures,
being based on language, must necessarily have an author, since the meanings of words are
just concepts intended by the mind of a speaker and are understood through human conven-
tion. Accordingly, their reliability cannot but be grounded in their being produced (kṛtakatva)
by an author and in that author being a person ( puruṣatva). That person must also be admitted
as being omniscient (sarvajña) in the sense of seeing truths that are beyond ordinary percep-
tion and are, ipso facto, precisely the object of scriptures. In part 2 (vv. 13–25),9 Śubhagupta
introduces an actual proof of the Buddha’s omniscience. He is established as being the only
omniscient person among many other omniscient beings who are admitted by other traditions
as the authors of their scriptures. This is because: (i) he is the only one teaching about selfless-
ness (and doing so efficaciously, that is, by adjusting his teachings to various people’s spiritual
needs); and (ii) he knows and teaches true, that is, effective, ritual utterances (mantras).
Śubhagupta employs the knowledge (and teaching) of the power and effects of mantras as
evidence that the Buddha has direct perception of extrasensory things. Since the power of the
mantras cannot be perceived or inferred, only someone who has that type of supersensible
perception can know about them. Śubhagupta is the first to attest to the use of the teaching of
mantras as proof for the Buddha’s omniscience. Before him, Dharmakīrti had already intro-
duced that type of knowledge as a demonstration of extraordinary perception in human beings.
However, Śubhagupta is the first (and one of the few) philosophers to revisit that proof in his
writings and explicitly relate it to the Buddha’s omniscience.
In the last verse, he states that, even when moral faults are purged, a zealous cultivation is
not enough to attain the omniscience related to the knowledge of mantras. The source of that
superior knowledge, Śubhagupta argues, is the Buddha, who teaches it and is established as
possessing perception of extrasensory truths. In this sense, I believe that this text presents itself
as the attempt of an author mostly concerned with logic and epistemology ( pramāṇa) to put
the knowledge of mantras into relation (albeit indirectly) with the Buddhist soteriological path.
The “Heterodox” Work: The Verses on the Demonstration

of External Objects
The most original and innovative theories of Śubhagupta, however, are found in the Dem-
onstration of External Objects.10 This work is intended to demonstrate the reality of external
objects of cognitions against a form of “idealism,” that is, the Vijñānavāda (literally the doc-
trine of the [reality] of consciousness) tradition of Buddhism and, particularly, the doctrine of
“consciousness-only,” that is, the true existence of consciousness alone (vv. 2–3). The trade-
mark doctrines in this work are a form of externalism (bāhyārthavāda) as well as the notion
that the images of sense objects are external and do not shape their cognitions (nirākāravāda).
Śubhagupta’s externalism is construed as the conception that material things, consisting of
atoms, are truly existent and are the support of sense cognitions. This is based on an atomic
theory that draws on both the Buddhist Vaibhāṣikas’ and Sautrāntikas’ views on atoms. In the
first part of the Demonstration, Śubhagupta is interested in proving the reality of infinitesimal
material particles that are singular (*eka), that is, devoid of parts, and ultimately (“substan-
tially,” *dravyataḥ) existent. These atomic particles are perceived by direct perception but
432
Śubhagupta
never singly. However, once they are apprehended, one superimposes some concepts, such as
indigo and other conceptual images, on them. Accordingly, the images that manifest when we
experience sense cognitions, such as a multicolored rug, and appear as if they were external
are conceptual in nature. They are not verily apprehended by those sense cognitions, which
are ultimately not shaped by the image of their objects (vv. 37–44). The proof of atoms’ sin-
gularity (45cd–56) is formulated as a refutation of Vasubandhu’s attacks on the atomic theory
attributed to the Vaibhāṣikas of Kashmir by the author himself in his Twenty Verses (vv.
12–14 and commentary). Atoms are finally established to be unitary, that is, devoid of parts
and identical to themselves. This is the conditio sine qua non for their aggregation, because
the multiplicity of entities is admitted as tenable only if those entities are established as units.
They are continuous and proximate but do not touch each other. Like the Vaibhāṣikas from
Kashmir, Śubhagupta sees their continuity as their having nothing between them (such as
light) and as being in opposition to each other. After demonstrating this, he turns to proving
that atoms are real substances (dravya), in the sense that they are causally efficient. Atoms
come to existence as aggregating since only with other atoms do they have the distinctive
characteristic of supporting each other in performing one activity, that is, in being causally
efficient and causing their own cognition. Atoms are “tied” to each other through this power
of being a substance; accordingly, they cause their own cognition and are apprehended as
such (vv. 57–58; see Saccone 2015). Śubhagupta’s position on how atoms aggregate was
likely influenced by Dharmakīrti’s view in the Commentary on the Means of Knowledge and
its developments in Devendrabuddhi’s and Śākyabuddhi’s commentaries.11
Śubhagupta’s theory that sense cognitions are not shaped by the image of their objects runs
counter to both the Sautrāntikas and the Vijñānavādins but is particularly tailored to oppos-
ing the tenets connected with the “invariable co-apprehension” argument. This is an argument
advanced by Dharmakīrti in the Ascertainment of the Means of Knowledge (particularly v. 1.54)
but is already present in a different shape in the Commentary on the Means of Knowledge.12 It is
meant to prove the nondifference between a cognition and its object since they are always neces-
sarily perceived together. The latter, being proof of “consciousness-only,” is largely defended by
the subsequent authors within the “mainstream” Buddhist logico-epistemological tradition. In
contrast to this, Śubhagupta argues that, on the one hand, a cognition cognizes its object because
it has the nature (svabhāva) of apprehending it, and, on the other, no sense cognition can occur
without an external object. The nature (svabhāva) of a certain cognition is intended as the nature
that belongs to that specific elementary particle (dharma) and nothing else. A cognition limits
itself to illuminating the object without assuming its form. Since these two, object and cogni-
tion, arise in continuity, they are always found together, but this is not because they are identical,
as the idealists would argue (vv. 66, 81, 82, 89). The original Sanskrit fragments of verses 66,
81, and 89 are quoted by Kamalaśīla in the part of his Investigation of External Objects chapter
where he intends to defend the “invariable co-apprehension” argument. Kamalaśīla appears to
view his proof of the latter argument especially as a response to Śubhagupta’s criticism.
Thus, according to Śubhagupta, the cognitive process requires two elements in order to
occur: a cognition, which is the only apprehender, and an object (*viṣaya), which possesses a
form to be apprehended. Cognitions are compared to pure “light” since they have the nature
of making their objects known. For this reason, a cognition and its object depend on each
other and are part of the same causal complex, namely that particular perceptual process.
Śubhagupta concludes that this is why a cognition and its object are necessarily perceived
together; the invariability (niyama) is due to their causal relationship, not to their nondiffer-
ence. In the Demonstration of External Objects, we read:
433
Serena Saccone
“ There is no other apprehender besides cognition, [and] there is no sensory cognition

without objects. And therefore, the [fact of] being aware [of them] together is not
because of the nondifference of[, for example,] an indigo [thing] and its cognition.”
(v. 66)
But it is precisely the preceding [causal] complex that can make the instant of the
object cognized, in the same way that a visual form [is lit up] by light. It is by this
that there can be [their] being brought to awareness together.
(v. 81)
Since cognitions and [their] objects always arise in continuity, then, the term
[“co-perception”] is [used], because it is like that; however, in reality there is no
co-perception [of them].
(v. 82)
In Kamalaśīla’s corpus, particularly in the Commentary on the Compendium of Truths and in

the Commentary on the Ornament of the Madhyamaka, one finds many accounts of Śubhagupta’s
viewpoints, which Kamalaśīla presents in prose passages as quotations of his. These accounts
are historically relevant insofar as they testify to the reception of Śubhagupta’s views in the
tradition; this is useful in order to assess the intellectual distinctiveness of Śubhagupta himself.
A rather exhaustive account of Śubhagupta’s position on the nature of cognitions and
the cognitive process is provided in Kamalaśīla’s Commentary on the Ornament of the
Madhyamaka, and is “copied” verbatim in the “Chapter on Suchness” from Haribhadra’s
Abhisamayālaṃkārālokā.13 The text reads,
Śubhagupta says: “And every cognition of an ordinary person occurs as devoid of

that impression that is the image of an indigo [thing]. Regarding this [cognition,] the
undesirable consequence of the loss of [its] singularity does not [arise], since it does
not [truly] have a variegated form. The fact that this [cognition] is determined as hav-
ing the nature of experiencing an indigo [thing] and so on, is because it has the nature
of experiencing an indigo [thing] and so on; however, [it is] not because it is endowed
with the form of an indigo [thing] and so on. To explain: an image (*ākāra) is said
to be only [the] aspect (*prakāra) of apprehending the object-support (*ālambana),
but not [to be] the form of that object-support [Abhisamayālaṃkārālokā: but not the
similarity (of the cognition with the object)]. [That] indigo [thing] and so on, which
is cognized as appearing as if it were external, is not an image of cognition. However,
the cognizer, experiencing a cognition [that is] the awareness of an indigo [thing]
and so on, conceptually determines the indigo [thing] and so on with such an image
[Abhisamayālaṃkārālokā: due to ignorance, (the cognizer) conceptually determines
an indigo (thing) and so on, like this, with an external image].”14
Here, Śubhagupta is reported as defining an image (ākāra) in cognition as the cognition’s

own aspect ( prakāra) of apprehending its object-support (ālambanagrahaṇaprakāra), in con-
trast with an interpretation that takes the cognitive image to be the form of that object shaping
a cognition. This appears to be an echo of Vasubandhu’s Commentary on the Treasure of the
Abhidharma (Abhidharmakośabhāṣya) (Vasubandhu 1975, 401 [ad 7.13b]). In his sub-com-
mentary, Yaśomitra attributes this position to the Sautrāntikas (Yaśomitra 1936, 629). How-
ever, Śubhagupta’s position differs significantly from that of the Sautrāntikas. Unlike him,
434
Śubhagupta
they think that, for a cognition, apprehending an object means precisely its having the image of
that object. According to Śubhagupta, instead, the cognition has the characteristic, that is, the
nature, of cognizing an object, and, in this sense, prakāra should be intended as its “manner,”
“mode of being.” Therefore, the image (ākāra) of a cognition, for Śubhagupta, is not an image
in the cognition but rather the characteristic, that is, the mode of being itself, of the cognition
as apprehending the object. This could be compared to the light having the characteristic of
illuminating objects without taking their images or being shaped by them.
In the Demonstration of External Objects, at least two ways of construing an aspect-image
(ākāra) can be found: aspect-image in the sense of prakāra, that is, the aspect or mode of
the cognition that apprehends a specific object, and aspect-image in the sense of the physi-
cal form that belongs to the external object only and does not shape the cognition internally.
Śubhagupta expresses this in v. 84:
A cognition is said [to be] brought to awareness since it is endowed with the nature of
being an awareness. The object is brought to awareness because it generates a cogni-
tion having [its] appearance [as an object].
Against the “invariable co-apprehension” argument, Śubhagupta argues that there is not only one
aspect-image when someone cognizes an object, but two. Those two aspect-images (one belonging
to cognition, the other to the object) are both experienced, separately, one after the other. Verses
77, 78, and 79 explicitly counter the idea of the existence of a single image. If there were only one
image, it would have to be either that of the object only or that of the mere cognition only. However,
Śubhagupta argues that neither possibility is logical. Instead, in perception, we are aware of both the
aspect-image ( prakāra) of the cognition as cognizing the object and of the aspect-image (ākāra)
of the object. This is still purely at a perceptual level. Since these two are experienced one after the
other, they are wrongly determined as one aspect-image, although they are not (v. 82). Moreover,
for Śubhagupta, such apprehending (grahaṇa) of the object-support (ālambana) is certainly not an
activity. In v. 92, he implicitly reaffirms the claim that when the sūtras say “cognition cognizes,”15
and so on, they do not mean that cognition is an agent or does something. Cognition is an appre-
hender because that is its nature but not by virtue of any activity. A cognition, like every elementary
particle (dharma), is in fact devoid of activity and only conventionally talked about as having one.
A very similar concept was advanced in Vasubandhu’s Commentary on the Treasury of the
Abhidharma and reported as a Sautrāntika idea. According to the latter view and Śubhagupta’s,
a visual cognition arises depending on sight and visual forms. However, this visual cognition
is devoid of activity. It consists only of elementary particles and can be explained through the
relation of cause and effect. Metaphors like “eye sees, cognition cognizes” are thus not literally
true but are used intentionally for the sake of everyday activity (vyavahāra). Nevertheless, one
should not get attached to them (Vasubandhu 1975, 31 [ad 1.42]).
Thus, once again, Śubhagupta elaborates on a Sautrāntika principle and reworks it to fit
his personal standpoint. His take on the cognitive process is evidently an instance of his syn-
cretic reuse of different standpoints from diverse traditions. A cognition is devoid of activity
but has the nature of being the cognition of a specific object-support (ālambana). However,
it does not assume the image of that object-support, which comprises the atoms serving as
the actual cause of their own perception and, as such, being grasped by sense cognitions.
The image (ākāra), in the sense of the physical form of those atoms, remains external. In the
moment immediately following the non-conceptual direct perception, conceptual construction
“interprets” the atoms through a conceptual image, such as that of an extended object, which
is superimposed on them (v. 36). This conceptual image appears as if it were external, does
435
Serena Saccone
not manifest in sense cognitions, and is manifold. However, sense cognitions are not manifold,
since they do not ultimately possess those images.
In sum, with the Demonstration of External Objects, Śubhagupta breaks with the tradition
by re-elaborating upon it and proposing original ideas deviating from its baseline doctrine. In
this sense, while he employs some ideas of Vasubandhu, Dignāga, and Dhamakīrti, he refutes
their “idealistic” standpoints.16
Conclusions
Though largely overlooked by contemporary scholarship, Śubhagupta represents a pivotal and
innovative figure within the Buddhist tradition of logic and epistemology. His influence and
importance in the eighth century are manifested by Śāntarakṣita and Kamalaśīla, two promi-
nent figures of the tradition, who attribute great significance to the refutation of his external-
ism and his theory of the absence of images in cognition. They do so in the chapters of their
Compendium of Truths and Commentary that are devoted to the demonstration and defense of
“consciousness-only.” In spite of the presence of important and valuable non-Buddhist oppo-
nents, such as Kumārila Bhaṭṭa, they prefer to address their criticism to a contemporary, and
fellow Buddhist, scholar-monk. This is particularly true with regard to Kamalaśīla.
Three, perhaps four, of Śubhagupta’s works were conceived as one set of texts. All of them
are deeply indebted to the Dharmakīrtian tradition and drew on the works of previous Buddhist
authors. Assuming that all his works were composed prior to Śāntarakṣita and Kamalaśīla’s
Compendium of Truths and Commentary, Śubhagupta was most likely the first Buddhist author
within the logico-epistemological tradition to adopt the genre of digests – a genre that would
prove popular among later authors. Moreover, the specific statements of argumentation and
elaboration of doctrines (predominantly taken from previous Buddhist authors) that are found
in his digests demonstrate his unique innovations and contributions to the field.
Nevertheless, the most groundbreaking element of his thought remains his attempt to inte-
grate views from the tradition in a syncretic fashion. This was done in order to reaffirm a more
“externalist” account of reality in contrast to the mainstream Buddhism idealism. From a care-
ful analysis of Śubhagupta’s corpus in its entirety, we thus uncover the figure of an innovative
thinker who played a crucial role in the history of the Buddhist tradition of logic and episte-
mology both by systematizing it and by subverting it.
Notes
1 Śubhagupta’s doctrinal affiliation has been defined in different ways in ancient sources and among
contemporary scholars. In particular, it is debated whether he was a Vaibhāṣika or a Sautrāntika. On
this, see Saccone 2018, 2019a.
2 On this verse, see Saccone Forthcoming; Eltschinger and Ratié Forthcoming.
3 On this verse, see Steinkellner 2006, 201, 204.
4 See Bu ston 1988, 162. Bu ston is almost certainly referring to the previously mentioned colophon of
the Commentary on the Ascertainment of the Means of Knowledge.
5 On this, see Saccone 2019b, 457.
6 On this, see Eltschinger and Ratié Forthcoming.
7 On this work, see Eltschinger 1999.
8 On the first part of this work, see Saccone 2019b.
9 On the second part of this work, see Saccone Forthcoming.
10 For an edition, English translation, and analysis of parts of this work, see Saccone 2018. For its first
English translation, see Shastri 1967.
11 See Pramāṇavārttika, chapter on perception, 195–96; Saccone 2015, 126–28.
436
Śubhagupta
12 See Pramāṇavārttika, chapter on perception, 387–390abc.

13 This is introduced by Kamalaśīla as a prose quotation of Śubhagupta and by Haribhadra as the
objection of an unidentified opponent. The passage is not found in any extant work that is explicitly
ascribed to Śubhagupta and neither explicitly refers to nor is connected with any verses of the Dem-
onstration of External Objects, which is the only work devoted to epistemological issues with regard
to sense cognitions. This is not the only instance where Haribhadra appears to be reusing materials
from the Commentary on the Ornament of the Madhyamaka. Accordingly, one is led to believe that
Haribhadra is copying Kamalaśīla here, not quoting a lost work of Śubhagupta.
14 See Kamalaśīla 1985, 163; Haribhadra 1932–1935, 632–33.
15 See yat tarhi “vijñānaṃ vijānāti” iti sūtra uktaṃ kiṃ tatra vijñānaṃ karoti (Vasubandhu 1975, 473–
74 [Pudgalavādapratiṣedha chapter]) 2).
16 The targets of his refutation are mostly the Twenty Verses and its commentary by Vasubandhu and the Inves-
tigation of the Object-Support and its commentary by Dignāga, as well as the “invariable co-apprehension”
argument expressly formulated by Dharmakīrti, especially in the Ascertainment of the Means of Knowledge.
Bibliography
Arcaṭa. 1949. Hetubinduṭīkā of Bhaṭṭa Arcaṭa with the Sub-Commentary Entitled Āloka of Durvekamiśra,
edited by Sukhlal Sanghavi. Baroda: Baroda Oriental Institute.
Bu ston. 1988. Bu ston chos ’byung, edited by Rdo rje rgyal po. Beijing: Krung go’i bod kyi shes rig dpe
skrun khang.
Dharmakīrti. 1979 [1985]. Pramāṇavārttika Pratyakṣa. Bukkyōninshikiron no kenkyū. edited by Hiro-
masa Tosaki, 2 vols. Tokyo: Daitōshuppansha.
Eltschinger, Vincent. 1999. “Śubhagupta’s Śrutiparīkṣākārikā (vv. 10cd – 19) and Its Dharmakīrtian
Background.” In Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceed-
ings of the Third International Dharmakīrti Conference (Hiroshima, November 4–6, 1997), edited by
Shōryū Katsura, 47–61. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Eltschinger, Vincent, and Isabelle Ratié. Forthcoming. “A Buddhist Refutation of the Existence of a
Creator God: Śubhagupta’s Īśvarabhaṅgakārikā.” In Gedenkschrift Helmut Krasser, edited by Vin-
cent Eltschinger, Parimal Patil, Leonard Van der Kuijp, and Chizuko Yoshimizu. Hamburg Buddhist
Studies. Hamburg: Hamburg University Press.
Frauwallner, Erich. 1933. “Dignāga und anderes.” In Festschrift für Moriz Winternitz, edited by Otto
Stein and Wilhelm Gambert, 237–42. Leipzig: Otto Harrassowitz.
———. 1957. “Zu den buddhistischen Texten in der Zeit Khri-sroṅ-lde-bstan’s.” Wiener Zeitschrift für
die Kunde Süd- und Ostasiens 1: 95–103.
———. 1961. “Landmarks in the History of Indian Logic.” Wiener Zeitschrift für die Kunde Süd- und
Ostasiens 5: 125–48.
Funayama, Toru. 1995. “Arcaṭa, Śāntarakṣita, Jinendrabuddhi, and Kamalaśīla on the Aim of a Treatise
(Prayojana).” Wiener Zeitschrift für die Kunde Südasiens 39: 181–201.
Haribhadra. 1932–1935. Abhisamayālaṃkār’ālokā Prajñāpāramitāvyākhyā (Commentary on the
Aṣṭasāhasrikā Prajñāpāramitā) by Haribhadra. Edited by Unrai Wogihara. Tōyō Bunko Publications
Series D, vol. 2. Tokyo: The Tōyō Bunko.
Hermann-Pfandt, Adelheid. 2008. Die lHan kar ma – Ein früher Katalog der ins Tibetische übersetzten
buddhistischen Texte. Kritische Neuausgabe mit Einleitung und Materialien. Wien: Verlag der Öster-
reichischen Akademie der Wissenschaften.
Kamalaśīla. 1985. Madhyamakālaṃkāra of Śāntarakṣita with His Own Commentary or Vṛtti and with the
Subcommentary or Pañjikā of Kamalaśīla. Edited by Masamichi Ichigō. Kyoto: Buneidō.
Matsumoto, Shiro. 1980. “Sahopalambhaniyama.” Sōtōshū kenkyūin kenkyūsei kenkyū kiyō 12: 298–65.
Saccone, Serena. 2014. “Śubhagupta on the Cognitive Process.” Journal of Indian Philosophy 42 (2/3) . In Birgit
Kellner, Sara McClintock (eds.): ākāra in Buddhist philosophical and soteriological analysis, pp. 377–399.
———. 2015. “The Conception of Atoms as Substantially Existing in Śubhagupta.” Journal of the Inter-
national Association of Buddhist Studies 38: 107–37.
———. 2018. On the Nature of Things. A Buddhist Debate on Cognitions and Their Object. Wien: Arbe-
itskreis für Tibetische und Buddhistische Studien Universität Wien.
———. 2019a. “Śubhagupta.” In Brill’s Encyclopedia of Buddhism, vol. II, edited by Jonathan A. Silk,
R. Bowring, V. Eltschinger, and M. Radich. Leiden and Boston: Brill: 458–462.
437
Serena Saccone
———. 2019b. “Of Authoritativeness and Perception: The Establishment of an Omniscient Person
(Against the Mīmāṃsakas).” In Wind Horses. Tibetan, Himalayan and Mongolian Studies, edited by
Andrea Drocco, Lucia Galli, Chiara Letizia, Giacomella Orofino, Carmen Simioli. Napoli: Università
degli Studi di Napoli “L’Orientale”, Dipartimento Asia, Africa e Mediterraneo, Series Minor LXXX-
VIII: 455–484.
———. Forthcoming. “Apology for Omniscience. An Eighth-Century Demonstration of the Buddha’s
Sarvajñatva.” In: Gedenkschrift Helmut Krasser, edited by Vincent Eltschinger, Parimal Patil, Leon-
ard Van der Kuijp, and Chizuko Yoshimizu. Hamburg Buddhist Studies. Hamburg: Hamburg Univer-
sity Press.
Sanderson, Alexis. 2009. “The Śaiva Age – The Rise and Dominance of Śaivism During the Early Medi-
eval Period.” In Genesis and Development of Tantrism, edited by Shingo Einoo, Tokyo: Institute of
Oriental Culture, University of Tokyo: 41–350.
Schiefner, Anton. 1868. Rgya gar chos ’byung (Tāranātha): Tāranāthae de Doctrinae Buddhicae in India
Propagatione Narratio. Contextum tibeticum e codicibus petropolitanis edidit Antonius Schiefner. St.
Petersburg: Academia Scientiarum Petropolitana.
Shastri, N. A. 1967. “Bāhyārthasiddhikārikā.” Bulletin of Tibetology 4 (2): 1–96.
Steinkellner, Ernst. 1985. “Paralokasiddhi-Texts.” In Buddhism and Its Relation to Other Religions:
Essays in Honour of Dr. Shozen Kumoi on His Seventieth Birthday. Kyoto: Heirakuji Shoten: 215–224.
———. 2006. “Miszellen zur erkenntnistheoretisch-logischen Schule des Buddhismus IX: The Colophon
of Dharmottara’s Pramāṇaviniścayaṭīkā.” Wiener Zeitschrift für die Kunde Südasiens 50: 199–205.
Ui, H., M. Suzuki, Y. Kanakura, and T. Tada, eds. 1934. Chibetto-Daizō-kyō Sōmokuroku (A Complete
Catalogue of the Tibetan Buddhist Canons). Sendai: Tōhoku Imperial University.
Vasubandhu. 1975. Abhidharmakośabhāṣyam of Vasubandhu. Edited by P. Pradhan. Patna: K. P. Jayas-
wal Research Institute (first ed. 1967).
Yaśomitra. 1936. Sphuṭārthā Abhidharmakośavyākhyā by Yaśomitra. Edited by Unrai Wogihara (1932–
1936), 2 vols. Tokyo: The Publishing Association of Abhidharmakośavyākhyā.
438
26
DHARMOTTARA
Systematic and Innovative Commentator
David Nowakowski
Dharmottara, who flourished circa 740–800 CE, is best known as a commentator on

Dharmakīrti’s (c. 600–660 CE) epistemological works, though several of his independent
(non-commentarial) treatises also survive in Tibetan translation. While biographical informa-
tion is scant, we have evidence that Dharmottara moved to Kashmir during the reign of King
Jayāpīḍa, a great patron of scholars and builder of three large images of the Buddha, who
reigned in the latter decades of the eighth century. Roughly a century later, Dharmottara would
gain the distinction of being one of the only Buddhist philosophers to appear as a character
in classical Sanskrit drama: a Buddhist monk named Dharmottara appears in the first act of a
play by the Kashmiri Nyāya philosopher Jayanta Bhāṭṭa, though as we might expect from a
self-styled defender of Vedic traditions, Jayanta’s portrayal of Dharmottara is quite far from
flattering.1
The present chapter will focus especially on Dharmottara’s commentary on Dharmakīrti’s
Epitome of Reason (Nyāyabindu),2 examining the ways in which Dharmottara works out the
consequences of key tenets of Dharmakīrti’s philosophy in an integrated and systematic way.
In general, Dharmottara’s work as a systematist pushes him to comment in ways that go well
beyond the surface meaning of individual passages in Dharmakīrti’s text taken on their own,
prompted not by misunderstanding or a desire for novelty but by the need to resolve apparent
contradictions or tensions between various aspects of the system. Specifically, Dharmottara’s
exposition of Buddhist epistemology and philosophy of mind,3 as seen in his commentary on
the Epitome of Reason, seems to blur Dharmakīrti’s bright lines between perception, the inter-
nal process of inferential reasoning known as “inference-for-oneself” (svārthānumāna), and
the interpersonal dialogue and dialectic known as “inference-for-others” ( parārthānumāna).
And in doing so, Dharmottara tends to make all of these much more dependent upon the state
and circumstances of particular individual cognizers, as contrasted with a more objective,
impersonal account of the cognitive-epistemic processes.
Of course, as a good classical Indian commentator, Dharmottara presents all of this as if it
were already present – if not on the page of Dharmakīrti’s texts, then at least in his illustrious
predecessor’s mind as the author of those texts. Yet at a minimum, we can confidently say,
consistently with Dharmottara’s own self-presentation, that Dharmottara’s work constitutes a
marked shift in emphasis, such that cognizer-relativity takes center stage in a way that it does
not in Dharmakīrti’s root texts.
DOI: 10.4324/9781351030908-39 439

David Nowakowski
In what is perhaps that same spirit, the present chapter itself is not particularly innova-
tive. It simply draws together three threads which have emerged separately in the scholarly
literature on Dharmottara in order to suggest the interconnections of these three themes and
from there to draw some tentative lessons about thinking of Dharmottara as a systematist, as
an innovator on Dharmakīrti, and as a stepping-stone to later Buddhist epistemologists such
as Jñānaśrīmitra and Ratnakīrti, who will both adopt Dharmottara’s innovations and further
innovate in the same areas.
The first section will examine what is perhaps Dharmottara’s most significant contribution
to the subsequent tradition: the additional complications he brings to Dharmakīrti’s account of
the differences between perception and inference, making these two cognitive processes appear
much more similar in their structure and operation and (especially for perception) much more
dependent on contributions made by the perceiver herself. The second section will consider
Dharmottara’s account of “the non-apprehension of the perceptible” in the Buddhist theory of
inference, examining how perceptibility is understood as being relative to some particular cog-
nizer’s circumstances. The third section will address Dharmottara’s presentation of the formal
conditions which define genuine inferential reasons, where Dharmottara uses resources from
his predecessor’s account of debate and dialectic to elucidate an account of inferential reason-
ing which likewise depends upon the epistemic state of each individual cognizer, where the
cognizer’s doubt or certainty about the relations between various properties plays a key role in
making it the case that one property constitutes a reason for inferring another.
Larger systematic considerations drawn from Dharmakīrti’s philosophy will play a role in
motivating Dharmottara’s arguments in each section: universal momentariness, basic causal
theory, and the basic account of doubt, in the first through third sections, respectively.
The Processes and Objects of Cognition

The issue on which Dharmottara most clearly, and perhaps also most radically, innovates on
Dharmakīrti’s work is in defining and distinguishing perception and inference, both in terms
of the processes by which these two means of knowledge operate and in terms of the cogni-
tive objects which each of them make known. In the wider Indian philosophical tradition of
which the work of Dharmakīrti and Dharmottara is a part, the two questions which organize
and structure debates in epistemology are: (1) what are the instruments, or causal processes, by
which that subset of awareness-events which count as knowledge are generated? and (2) what
sorts of objects can be known by means of each of those instruments or processes?
In answer to the first question, the standard claim found throughout the work of Dignāga
and Dharmakīrti, and which all of their Buddhist successors feel bound to uphold in some way,
is that there are exactly two such instruments of knowledge: perception and inference. Other
Indian philosophical schools such as the Nyāya or Mīmāṃsa would countenance a longer list
of instruments, such as testimony, analogical reasoning, and the awareness of absence. But for
the Buddhist epistemologists, all of these latter candidates will either be subsumed as special
cases of perception or inference or else written off as misleading (and thus not invariably reli-
able means of knowledge at all).
The answer to the second question gets more complicated, and these complications point
directly to Dharmottara’s elaboration or revision of Dharmakīrti’s account. In Dharmakīrti’s
account of cognition, perception and inference each have a single object (i.e., what it is that
the cognition is a cognition of), and that single object is of a different kind in perception and in
inference: a unique particular is the object of a perceptual awareness, while a universal is the
object of an inferential awareness. Thus, perception and inference each range over a distinct
440
Dharmottara
domain of objects – a claim which serves to distinguish Dharmakīrti’s account of knowledge

from that of rival, non-Buddhist schools, who argue that all of the instruments of knowledge
range over a single, unified domain of objects.4
While Dharmottara will strive to maintain some clear difference between perception
and inference, he will argue that, in contrast to Dharmakīrti’s model, each type of veridical
awareness has two objects: one which is grasped (grahya) and another which is determined
(adhyavaseya). The implications of this shift get at some of the central commitments of the
Buddhist epistemological tradition. This section will briefly present the core of Dharmakīrti’s
model before examining the motivations for Dharmottara’s revisions and the implications of
those changes for Indian Buddhist theories of mind and cognition more generally.
Throughout his corpus, Dharmakīrti consistently follows Dignāga in distinguish-
ing two kinds of objects of knowledge: particulars (svalakṣaṇa), which exist ultimately
( paramārthasat), by their nature are utterly distinct from every other particular, and are nec-
essarily momentary (i.e., existing for only a single, smallest instant of time); and universals
(sāmānya), which exist only conventionally, are characterized by a common nature shared
over multiple instances, and are for those reasons non-momentary (i.e., persisting through
time). While this distinction is at work throughout Dharmakīrti’s corpus, an especially clear
presentation and defense of the view occurs at the beginning of the chapter on perception in his
Pramāṇavārttika (PV 3, trans. Franco and Notake 2014). The chapter begins with the assertion
that “the instruments of knowledge are of two kinds, because the objects of knowledge are of
two kinds” (PV 3.1a–b).5 The direction of this explanatory relation – the duality of objects giv-
ing rise to, and justifying, Dharmakīrti’s postulate of the duality of instruments – is repeated
in PV 3.63, which concludes the relevant section of that text: “Therefore, the duality of the
instruments of knowledge is accepted because of the duality of the objects of knowledge.”
Dharmakīrti follows this arrangement in setting out the argument: first presenting four argu-
ments for the distinction between particulars and universals (PV 3.3–50) and then arguing on
the basis of that distinction for a corresponding distinction between perception and inference
(3.55–62). The main argument (3.3–6) is that particulars are unlike universals because only the
former possess the capacity to bring about effective action (arthakriyā). That is, because only
particulars ultimately exist, it is only by acting toward particulars that we can achieve real results
in the world, and so it is only when we act toward a particular that the desire which motivated
our action will be satisfied. Since universals do not ultimately exist, acting only toward a univer-
sal (without a real particular somehow being implicated) will necessarily leave us unsatisfied.
Thus, both the distinction between particulars and universals and the corresponding dis-
tinction between perception and inference on that basis are central tenets of Dharmakīrti’s
epistemology and philosophy of mind. Dharmottara will radically revise his predecessor’s
account of both.
Commenting on the perception chapter of Dharmakīrti’s Epitome of Reason, Dharmottara
explicitly asserts that perception has not one but two objects: the grasped object, defined as
“that whose image is produced (in awareness),” and the determined object, defined as that
which can be attained. Dharmottara explains:
The grasped object is one thing and the determined is something else, since for per-
ception, what is grasped is a single moment, but what is determined – through a
judgment that arises by the force of perception – can only be a continuum. And only
a continuum can be the attainable object of perception, because a moment cannot be
attained.
(NBṬ on NB 1.12, trans. McCrea and Patil 2006, 325)
441
David Nowakowski
As Dharmottara’s argument clearly shows, his revision to the basic account of perceptual
awareness – itself a fundamental doctrine in the Buddhist epistemological tradition – is
motivated by systematic concerns arising from the closely related and equally fundamental
doctrine of momentariness: the thesis that whatever ultimately exists exists only for a single
smallest instant of time (i.e., a single “moment,” kṣaṇa), after which it immediately ceases
to exist.
There are two interconnected parts to Dharmottara’s reasoning. The first part follows
directly from the theory of momentariness. Given that any particular exists for only a single
instant of time, by the time we act in response to our perception of it, it is no longer there to
be attained. If, as the Buddhist epistemologists assert, every purposeful action is motivated by
desire, and enabling us to attain the objects of our desire is what distinguishes veridical from
non-veridical awarenesses, then the particular entity which we grasp in perception cannot pos-
sibly be what we desire, since due to its momentariness, it is inevitably unattainable. In other
words, because of momentariness, we can never obtain any particular which we perceive.
In some way, then, something which is not a particular must be involved both as a cognitive
object of awareness and as an intentional object of action.
The second part is based on the distinction between conceptual and non-conceptual aware-
nesses. As Dreyfus notes (1997, 355), effective action (arthakriyā) requires cognition which
is veridical, or non-deceptive. Such a cognition, in turn, necessarily involves accurately
identifying the object of thought and action. And this need for accurate identification and
categorization straightforwardly means that conceptual awareness must be involved. A cogni-
tion which is purely non-conceptual will not suffice. Yet, as Dreyfus then remarks, “such an
epistemic coordination between perception and conception, however, is in principle outside
Dharmakīrti’s system” (1997, 358).
Thankfully for Dharmottara, these two lines of reasoning converge on a single solution,
since in the standard Dharmakīrtian system, non-conceptual awarenesses deal always and only
with unique, momentary particulars, while conceptual awarenesses deal always and only with
repeatable, non-momentary universals. Where Dharmakīrti limited his account of perception
to involving only the former, Dharmottara need only find a way to include, within his revised
account of perception itself, some role for the latter.
Dharmottara thus posits a second object of perception, in addition to the grasped object
which is “given” from the world: a continuum (santāna), which the cognizer determines. The
continuum is an ultimately unreal entity, composed of the various real momentary entities
which exist in a causal series. And this continuum is not something simply “found” in the
external world, nor simply “given,” but rather is constructed by the cognizer herself based on
her own interests, desires, habits, and latent mental traces. Thus Dharmottara’s act of deter-
mination, as McCrea and Patil note, “makes available to us constructed objects which are not
directly presented to our awareness” (2006, 331). The very suggestion that the perceptual
process itself involves such conceptually constructed, not directly presented objects is a break
with Dharmakīrti, the radical nature of which cannot be understated.
To give a concrete example, according to the theory of momentariness, what we conven-
tionally but inaccurately take to be a single pot persisting through time is really a series of
distinct pot-moments, each of which is among the causes of the subsequent pot-moment.
These pot-moments are real particulars, while the continuum comprised of all the sequential
pot-moments is an unreal, conceptually constructed universal.6 While it lacks ultimate real-
ity, the pot-continuum does have the virtue of persisting through time. And it is that temporal
persistence which enables us to perceive, form the desire for, act toward, and finally obtain it.
Without the persisting, determined object, we would never be able to attain what we perceive,
442
Dharmottara
and so we would have no basis for distinguishing veridical from non-veridical cognitions;
rather, all cognitions would be non-veridical.
Immediately following the passage quoted previously, Dharmottara goes on to explain that
inference, too, has both a grasped and a determined object but in a way that is reversed from
the perceptual case. In inference, it is “an imposed thing,” “a non-entity” – that is, a uni-
versal – which is grasped, while it is a particular that is determined. In both perception and
inference, the grasped object is given or presented to us, and the determined object is what we
act upon or toward. In order to ward off the obvious questions, Dharmottara accounts for the
apparent divergence between himself and his teacher by explaining that in the passage of the
Epitome of Reason upon which he is commenting (and, by implication, at every other point in
Dharmakīrti’s corpus where the subject arises), Dharmakīrti is concerned merely with explain-
ing the grasped objects of each mode of awareness.
Yet this should not suggest that the term “determination” (adhyavasāya) is absent from
Dharmakīrti’s corpus. Quite the contrary. This is a term employed by Dharmakīrti but radically
re-purposed by Dharmottara. Dharmakīrti uses “determination” to refer to a proper subset
of conceptual (i.e., verbal or inferential) awareness; the term plays no role in his account of
perception as such. For Dharmottara, as we have seen, “determination” shifts to become one
component part in every instance of both perceptual and inferential awareness.
Thus, we find two important shifts from Dharmakīrti’s presentation of the basic model
of awareness to that of Dharmottara. First, as McCrea and Patil (2006, 332) observe, while
Dharmakīrti could distinguish perception from inference not merely by their objects but also by
the quite distinct processes involved in each, for Dharmottara, the parallel processes of grasping
and determination in both cases mean that perception and inference can be distinguished only
by their objects. The sharp opposition between the two epistemic instruments in Dharmakīrti is
thus significantly worn down. Second, we see Dharmottara bringing conceptual awareness into
the process of perception itself, which Dharmakīrti has presented as purely non-conceptual.
The process of concept-formation is highly dependent upon each individual cognizer.
Here we see an important case of Dharmottara as innovator and as systematic thinker, for it
is on the basis of reconciling momentariness with cognitive theory that he makes these major
revisions to Dharmakīrti’s basic account of perception and inference. And while it is spelled
out less explicitly in the present passages, we can also see Dharmottara shifting toward a more
cognizer-relative account of perception and inference thanks to the constructed nature of the
determined object.
As the first to articulate such a dual-object model, Dharmottara leaves some challenges
to be worked out, a task which is taken up enthusiastically by Jñānaśrīmitra (fl. 975) and his
student Ratnakīrti (fl. 1000), who adopt Dharmottara’s dual-object account while continuing
to revise the details of exactly how precisely to characterize those two objects.7
Perceptibility and Cognizer-Relativity

Dharmottara also functions as an (apparently) innovative commentator, expanding on small
hints latent in Dharmakīrti to move Buddhist epistemology in a cognizer-relative direction, in
his discussion of the theory of inference.
Dharmakīrti famously sets out a classification of inferential reasons (hetu) into exactly
three legitimate varieties, or styles of cases, which invariably produce knowledge of what
is to be inferred. Of these, the most complicated type, normally discussed last, involves the
non-apprehension of what would be, if it were present, perceptible (dṛśyānupalabdhi), such
as when someone sees a floor in all its full and vibrant detail but does not see a pot which
443
David Nowakowski
would be perceptible were it to be present and infers the absence of a pot on that floor. Here,
the problem of interest to Dharmottara is this: what does it mean for the pot (or whatever other
object) to be “perceptible” (dṛśya), especially given that it is not present and perhaps even
non-existent? And, correspondingly, what would it mean for something to be imperceptible
(adṛśya) in the relevant sense?
Just as with “determination” in the preceding section, here again, we find Dharmottara
deploying a conceptual (and terminological) resource that is present in a very limited way in
Dharmakīrti but is used by Dharmottara much more extensively. This is the notion of being
“remote in space, time, or kind” (deśakālasvabhāvaviprakṛṣṭa). While Dharmakīrti mentions
remoteness a number of times, the full term with its triple classification appears only once
in the Epitome of Reason (NB 2.27), where it is given without elaboration, and only rarely
elsewhere. The term rises to greater prominence in Dharmottara’s commentary and takes on a
great significance in other contexts for subsequent members of the Buddhist epistemological
tradition, including Ratnakīrti, for whom a rigorous account of remoteness in space, time, and
kind plays an important role in the refutation of the existence of Īśvara, the omniscient demi-
urge who figures prominently in later Nyāya philosophy, and of the existence of other minds.8
As we will elaborate on in the following, the term is intended to capture all of the relevant
ways in which an object could be too far away to be perceived by some particular cognizer: in
spatial distance, temporal distance, or metaphysical distance (for instance, certain demons can
be perceived only by their fellow demons or advanced yogins but not by ordinary humans).
The Buddhist epistemologists’ basic principle, already found in Dharmakīrti, is that there
are no unactualized causal capacities. Any entity is only appropriately called “cause of (some
effect) X” if at that very moment it is causing X.
In the case of perceiving some object, such as a pot, what are the causes of that perceptual
awareness? And, correspondingly, what are the causes of the pot’s (or other object’s) being
perceptible? Such causes will include the pot itself, along with the cognizer’s sense-organ(s),
sufficient light (or similar conditions, for sense-modalities other than sight), and a variety of
other factors. For our purposes, following Dharmottara’s discussion in commenting on NB
2.14 (translated and discussed by Kellner 1999, 199–200), we need only note the presence of
an extensive list of causes, where the object itself – the distinct entity which is perceived – can
be considered the primary cause, and all the others can be grouped together as the complete set
of additional causal factors. Now, in the case of two ordinary material objects which actually
are situated side by side or one atop the other like the pot and the floor, the complete group of
additional factors which, together with the floor itself, produce a perceptual awareness of the
floor will be exactly the same as the complete group of causal factors which, together with
the pot itself, produce a perceptual awareness of the pot. Given the principle that there are no
unactualized causal capacities, it follows that in this case, it is impossible to perceive the floor
without simultaneously perceiving the pot or to perceive the pot without the floor. For if each
distinct object (pot and floor) is itself present, the only way for that object not to be perceived
is if one (or more) of the additional factors were lacking. But since the additional factors are
the same for both, then the absence of any factor required for pot-perception would also pre-
clude floor-perception, and vice versa.
By contrast, in the case where the floor is present but the pot is not, a cognizer who is aware
of perceiving the floor but not the pot, and who knows that the additional factors for pot-
perception and floor-perception are the same, can readily conclude that since all the additional
factors are present (evidenced by their bringing about a perception of the floor) the lack of
pot-perception must be due to the absence of the pot itself. Reflecting appropriately on all of
this, the cognizer infers the absence of a pot here on the floor.
444
Dharmottara
In this latter example, it is the presence of all the other additional causal factors which in
some sense justifies calling the pot “perceptible” (dṛśya). And in his discussion of this issue,
Dharmottara explicitly states that the account of perceptibility is “given with regard to one
particular cogniser.”9
While the foregoing discussion makes clear the basic insight which the Buddhist episte-
mologists are attempting to capture in the non-apprehension inference, the complicated sub-
junctive phrasing – what properties a thing would have if it were to be present in a certain
way – goes both far beyond a literal translation of the Sanskrit and beyond what Dharmakīrti
and Dharmottara’s principle regarding the necessary actualization of causal capacities will
permit. The Sanskrit term dṛśyānupalabdhi means merely “the non-apprehension of the per-
ceptible,” while the causal principle seems to exclude the possibility of something’s simulta-
neously being both perceptible and not perceived. Moreover, the “perceptible” object (or at
least, the cognizer’s conception of such an object), precisely insofar as it is not apprehended
(and thus does not generate a perceptual awareness), is the cause of the inferential awareness
of the absence of the pot. So there must be something here which has real causal capacity with
respect to producing that inferential awareness as its effect.
As Dharmottara observes in an earlier passage of the same commentary (on NB 2.12), it is
only in some sense that the pot (or whatever absent object) is appropriately called “perceptible.”
His hesitation about describing the absent object as perceptible in an absolute sense arises from
the same principle which underwrote this class of inferences in the first place: the requirement
that all genuine causal capacities be immediately actualized. By that principle, we cannot prop-
erly call something “perceptible” unless it is actually perceived – which the absent pot is not.
Thus, while the inference depends upon causal capacity, “being perceptible” cannot, strictly
speaking, be a genuine causal capacity that really belongs to the object whose absence is inferred.
If there were a being, like the Naiyāyikas’ Īśvara, who was actually (and not merely poten-
tially) omniscient and who therefore perceived the absent pot (along with all other beings)
at every moment, this might seem to resolve the problem: the pot’s perceptibility would be
actualized with respect to that special omniscient cognizer. But even if the Buddhist episte-
mologists were to accept the existence of such a being (which they do not), that would be
insufficient to explain what is at work in the relevant class of inferences. Dharmottara would
explain this in terms of the object’s being “remote in kind” (svabhāvaviprakṛṣṭa), analogously
with the other two ways of being remote: in space or in time.
Among the three ways of being remote, the spatial case is easiest to grasp. A pot that is
located far from me on some remote mountain may very well be visible to – and, therefore,
actually seen by – someone standing beside it on that mountain, but that does nothing to make
it perceptible to me. Temporal remoteness will behave similarly: a pot which once existed in
the past may well have been seen by someone at that past time, but this too does nothing to
make it visible now. For remoteness in kind, the classic Indian example is of a demon who can
be perceived by other demons and also by accomplished yogic practitioners (whose perceptual
abilities far outstrip those of ordinary humans) but who is imperceptible to ordinary humans.
A more contemporary example would be a microorganism or subatomic particle, present at
the current place and time, but too small to be seen by the unaided human eye (though again,
perhaps, by the accomplished yogin).
In all the cases, what matters is not that the pot is perceptible to someone or other but that
for the particular cognizer who is seeing the floor, the pot would also be perceptible to her with
no additional supporting factors being required. It is for this reason that Dharmottara explains
that the pot “is called ‘perceptible,’ even though it does not exist, because ‘being perceptible’
is superimposed. ‘Being perceptible’ is superimposed on the pot, even though it does not exist
445
David Nowakowski
there, when the assumption is made that, if it existed here, it would have to be perceived.”10
Thus, the inference depends not on a property possessed objectively by the pot but rather on a
property which is inaccurately but effectively superimposed, based on the epistemic standing
of some particular cognizer with respect to the absent or imagined pot – in other words, a prop-
erty which is both provided by, and relative to, the cognizer. As Kellner (1999, 201) observes,
this doctrine of superimposition is associated with Dharmottara in particular by later writers in
the Buddhist epistemological tradition.
Doubt and Certainty in Inference-for-Oneself

The two preceding sections examined how Dharmottara connects issues involving the cog-
nitive/epistemic processes of perception and inference in his commentaries on the first two
chapters of Dharmakīrti’s Epitome of Reason. This final section will briefly address a bridge
Dharmottara builds between the second and third chapters of the Epitome of Reason, which
connects inference-for-oneself (the inner cognitive process by which each person attains
knowledge) with inference-for-others, which appears in the earlier Buddhist epistemologists
primarily as a theory of effective debate. This bridge thus connects the philosophy of mind
with argumentation and debate theory, again via a move to the cognizer-relative.
In theorizing the cognitive process of inference, Dignāga11 famously originates the “triple
conditions” meant to specify the conditions under which awareness of one property – the
inferential reason (hetu) – invariably gives rise to inferential knowledge of another property,
the target property (sādhya). The conditions are revised and more precisely formalized by
Dharmakīrti both in the second chapter of the Epitome of Reason and at much greater length
in the Hetubindu. While the formal analysis of Dharmakīrti’s triple conditions is well beyond
the scope of this section,12 we can briefly point to one key aspect of the interpretive shift from
Dharmakīrti to Dharmottara in terms of two questions: First, what qualifies as an inferential
reason? Second, when does inferential knowledge occur for some cognizer?
Dharmakīrti makes clear that the triple conditions state purely formal aspects of how the
inferential reason relates to the property to be inferred in various groups of property-bearers.
A genuine reason is found only in loci where the target is also found and excluded from every
locus in which the target is absent while also being found throughout the locus (or loci) in
which the target is to be inferred. The conditions thus provide the answer to the first question
but not the second. After stating the three conditions themselves in NB 2.5, Dharmakīrti then
goes on to state that the cognizer must be certain (niścita) that the conditions obtain, in order
for her to infer. Only then does knowledge result. In terms of the stock example of inferring
fire from smoke, in Dharmakīrti’s model, the fact of smoke being present only where fire is
also present and absent from every place in which fire is absent makes it the case that smoke
is in some fairly objective sense an inferential reason for fire. But any particular cognizer will
only have the inferential awareness of fire when she is certain that smoke is an inferential rea-
son for fire, even though smoke is (and has been) such an inferential reason for fire all along.
Where Dharmakīrti mentioned the need for certainty only once, following the list of the
triple conditions, Dharmottara by contrast adds the requirement for certainty to each of the
three conditions themselves. In his account, then, a property’s being an inferential reason at all
is dependent upon the cognizer’s being free from doubt regarding its status vis-à-vis the other
components of the inference. As applied to the stock example, this means that smoke being
an inferential reason for fire at all is dependent upon each particular cognizer. And so because
certainty is built in to being an inferential reason, upon having such a reason, the resulting
inferential awareness of the property to be inferred will result immediately.
446
Dharmottara
This is in keeping with the principle of the necessary actualization of causal capacities,
discussed in ”Perceptibility and Cognizer-Relativity.” The term hetu, which in inferential
contexts is standardly translated as “(inferential) reason,” also has the meaning of “cause”
in a variety of more mundane senses. And it is widely accepted by classical Indian epis-
temologists, both Buddhists and their rivals, that the inferential reason is in some sense
causally responsible for the resulting awareness. While he is not entirely explicit about
this point, by making the triple conditions which define the inferential reason depend upon
each cognizer’s certainty in this way, Dharmottara solves a parallel problem to that of per-
ceptibility presented in the previous section. Because being an inferential reason depends
upon a particular cognizer’s being certain, and therefore inevitably inferring on its basis,
the reason’s causal capacity will always be actualized. Neither in the case of inferential
reasons nor in the case of perceptible objects will Dharmottara be confronted with causes
which fail to produce their effects.
How do we know that Dharmottara is making such a shift? In his commentary on the
third chapter of the Epitome of Reason (on “inference-for-others”), Dharmottara works out an
exhaustive classification of pseudo-reasons (hetvābhāsa): properties which resemble genuine
inferential reasons by virtue of meeting some of the triple conditions but which fall short of
actually being such reasons insofar as they fail to satisfy one or two of the conditions. To the
extent that Dharmakīrti has the beginnings of such a classificatory system, its presence in
the chapter on debate and reasoning for others suggests that for him, it is primarily a tool for
identifying and critiquing illegitimate moves in a debate or conversation rather than being
primarily about the internal cognitive process of inference-for-oneself.
While Dharmottara expands on the classification of pseudo-reasons in the third chapter –
he is, after all, a commentator, committed to following the root text insofar as possible – he
deploys the pseudo-reasons in his formal analysis of the triple conditions in the second chapter
of the Epitome of Reason on inference-for-oneself.
A standard technique for analyzing definitions in Indian philosophical texts is to show how
each word in the definition excludes a particular group of things which fail to meet the defini-
tion and which are thus appropriately and effectively removed from the definition’s scope.
When all of the words in a single definition are taken together, we see that having excluded
everything which does not belong, we are left with a definition of exactly the right scope. In
analyzing the triple conditions according to this method, Dharmottara explains that each term
excludes particular ways of being a pseudo-reason. And since his classification is exhaus-
tive, when all the types of pseudo-reasons have been ruled out, we are left with only genuine
inferential reasons. Importantly, the term “certain” (niścita) plays such a role just as much as
every other term, and it plays a different role, excluding different pseudo-reasons, in each of
the three conditions. The need for certainty is thus built into the defining statements of what
it is to be an inferential reason. Being such a reason, on Dharmottara’s account, is clearly and
explicitly cognizer-relative.
Dharmottara’s method of analyzing the triple conditions would survive as at least one
accepted account of inference-for-oneself through the end of the Buddhist epistemological
tradition in India, as attested by its prominent presence in Mokṣākaragupta’s Tarkabhāṣā,13
albeit with some confusion on the part of the latter thinker.
Conclusion
In all three of the topics discussed in this chapter, we have seen Dharmottara as a deeply
systematic commentator, expanding and perhaps radically revising Dharmakīrti’s epistemic
447
David Nowakowski
theories in order to resolve tensions between various core positions in the Buddhist epistemo-
logical tradition. In so doing, Dharmottara displays a consistent tendency toward defining key
positions in terms of specific individual cognitive agents and away from more absolute and
impersonal accounts. Through all of these contributions, he sets the stage for later Buddhist
epistemologists, including Jñānaśrīmitra, Ratnakīrti, and Mokṣākaragupta, all of whom will
adopt and modify important parts of Dharmottara’s work on these topics.
With this creative and dynamic approach to philosophical commentary, we see Dharmot-
tara, like so many other Buddhist and non-Buddhist scholars in classical India, inhabiting a
thought-world. Dharmottara is not simply telling us what Dharmakīrti himself explicitly said
or taught; rather, he is thinking himself fully into this system of thought. Once there, he can
explore the implications and interconnections of the various parts of Dharmakīrti’s system and
also use the resources of that system to respond to new challenges, problems, and objections.
In so doing, the commentator engages with the philosophical system not as a museum piece
but as a living tradition.
Notes
1 Jayanta’s play, the Āgamaḍambara, is translated by Dezső (2005); for the historical Dharmottara’s
migration to Kashmir, see Dezső (2005, 270) and the references cited therein.
2 Dharmakīrti’s Nyāyabindu, literally “The Drop (or Epitome) of Reasoning,” will be abbreviated NB;
Dharmottara’s commentary, the Nyāyabinduṭīkā, will be abbreviated NBṬ. Both survive in Sanskrit;
the only complete English translation is the now extremely dated work of Stcherbatsky (1962).
3 The reference to “epistemology and philosophy of mind” is, from the perspective of the Sanskrit
traditions, a bit redundant, insofar as the issues raised under the single heading of pramāṇavāda, the
study of the instruments of knowledge, include within a single domain many of the topics covered by
both these distinct labels in Anglophone traditions.
4 In the technical language of these debates, Dharmakīrti’s view is known as pramāṇaviplava, “the
distinction of the instruments of knowledge,” while the rival view is called pramāṇasaṃplava, “the
mixing-up, or running-together, of the instruments of knowledge.”
5 Both this and the following translation are slightly modified from Franco and Notake (2014), with my
italics.
6 In the language of later Indian scholars, each pot-continuum is a “vertical universal” composed of
pot-moments, while the potness common to all pot-continua is a “horizontal universal” composed of
vertical universals.
7 Jñānaśrīmitra and Ratnakīrti are particularly (though not exclusively) engaged with this question
in their treatises on “the exclusion of others” (anyāpoha). Jñānaśrīmitra’s monograph on exclusion
has been translated by McCrea and Patil (2010), along with an admirably clear and accessible intro-
duction that explicitly discusses the evolution of the theory from Dharmakīrti to Dharmottara to
Jñānaśrīmitra. For Ratnakīrti’s work on exclusion and the dual objects of cognition, again in explicit
comparison with Dharmottara, see McAllister 2014, 2017.
8 For Ratnakīrti’s application of “remoteness in space, time, or kind” in the context of challenging the
Naiyāyikas’ Īśvara-inference, see his Īśvarasiddhidūṣaṇa, translated in Patil 2001, 307–83. For his
application of this concept regarding other minds, see his Santānāntaradūṣaṇa, translated and dis-
cussed in Kajiyama 1965.
9 NBṬ on NB 2.14, translated in Kellner 1999, 199.
10 NBṬ on NB 2.12 (translation heavily modified from Kellner 1999, 201).
11 Dignāga (c. 480–540) was the founder of the Buddhist epistemological tradition. Dharmakīrti’s own
masterwork, the Pramāṇavārttika, is presented as a commentary upon Dignāga.
12 One such technical analysis, with reference to the precise differences between various Buddhist and
Nyāya epistemologists, is in Nowakowski (2017).
13 The relevant section of Mokṣākaragupta is translated by Kajiyama (1998, 65–70). For the relation of
Mokṣākaragupta and Dharmottara on the triple conditions, see Nowakowski (2017, 350–52).
448
Dharmottara
Bibliography
Dezső, Csaba, ed. and trans. 2005. Much Ado About Religion by Bhatta Jayanta. New York: New York
University Press.
Dreyfus, Georges B. J. 1997. Recognizing Reality: Dharmakīrti’s Thought and Its Tibetan Interpreta-
tions. Albany: State University of New York Press.
Franco, Eli, and Miyako Notake, eds. and trans. 2014. Dharmakīrti on the Duality of the Object:
Pramāṇavārttika III 1–63. Berlin and Zurich: LIT Verlag.
Kajiyama, Yuichi. 1965. “Buddhist Solipsism: A Free Translation of Ratnakīrti’s Saṁtānāntaradūṣaṇa
[sic].” Journal of Indian and Buddhist Studies (Indogaku Bukkyogaku Kenkyu) 13 (1): 9–24 (= 435–
420). Reprinted in Yuichi Kajiyama, Studies in Buddhist Philosophy: Selected Papers, edited by Kat-
sumi Mimaki et al., 401–16. Kyoto: Rinsen Book Co., 2005.
———, trans. 1998. An Introduction to Buddhist Philosophy: An Annotated Translation of the Tarkabhāṣā
of Mokṣākaragupta. Reprint, with corrections in the author’s own hand. Wiener Studien zur Tibe-
tologie und Buddhismuskunde 42. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien
Universität Vien.
Kellner, Birgit. 1999. “Dharmottara’s Views in the Dṛśya in Dṛśyānupalabdhi.” In Dharmakīrti’s
Thought and Its Impact on Indian and Tibetan Philosophy: Proceedings of the Third International
Dharmakīrti Conference, edited by Shoryu Katsura, 193–208. Vienna: Österreichischen Akademie
der Wissenschaften.
McAllister, Patrick. 2014. “Ratnakīrti and Dharmottara on the Object of Activity.” Journal of Indian
Philosophy 42 (2–3): 309–26.
———. 2017. “Competing Theories of Conceptual Cognition: Dharmottara and Trilocana vs. Ratnakīrti?”
In Reading Bhaṭṭa Jayanta on Buddhist Nominalism, edited by Patrick McAllister, 291–321. Vienna:
Verlag der Österreichischen Akademie der Wissenschaften.
McCrea, Lawrence J., and Parimal G. Patil. 2006. “Traditionalism and Innovation: Philosophy, Exegesis,
and Intellectual History in Jñānaśrīmitra’s Apohaprakaraṇa.” Journal of Indian Philosophy 34 (4):
303–66.
———. 2010. Buddhist Philosophy of Language in India: Jñānaśrīmitra on Exclusion. New York:
Nowakowski, David. 2017. “Reasons and Doubt in Dharmottara and His Critics.” Asian Philosophy 27
(4): 340–68.
Patil, Parimal G. 2001. “Necessity, Naming, and the Existence of Īśvara.” PhD diss., University of
Chicago.
Shastri, Dwarika Das, ed. 1985. Nyāyabindu of Acharya Dharmakīrti. Bauddha Bharati 18. Varanasi:
Bauddha Bharati.
Stcherbatsky, Theodore. 1962 [c. 1930]. Buddhist Logic. Volume II. New York: Dover Publications.
449
27
JÑĀNAGARBHA
Two Truths Theory, Gradualism, and
Mādhyamika Philosophy
Ryusei Keira
There have been at least three figures with the name Jñānagarbha in the history of Indo-Tibetan
Buddhism. The first was the author of the Commentary on the Distinction Between the Two
Truths (Satyadvayavibhaṅgavṛtti; hereafter Commentary) in the early eighth century. The sec-
ond Jñānagarbha translated mainly Madhyamaka texts into Tibetan in the early ninth century.
The third taught the Secret Assembly Tantra (Guhyasamājatantra) to the Tibetan translator
Marpa (1012–1096) in the eleventh century (Matsumoto 1978, 109–10; Seyfort Ruegg 1981, 69
fn. 224). In this chapter, we will focus on the first Jñānagarbha, the author of the Commentary.1
According to Tibetan sources, Jñānagarbha was originally from Oḍiviśa (contemporary Orissa
in eastern India) and was a teacher of Śāntarakṣita (ca. 725–788) (Schiefner 1869, 198–99; Mat-
sumoto 1978, 111; Seyfort Ruegg 1981, 69 fn. 225). Śāntarakṣita and his disciple Kamalaśīla
(ca. 740–795) were strongly influenced by Jñānagarbha’s Commentary and adopted his two
truths theory to establish their own Mādhyamika philosophies. Jñānagarbha is also important
for further developing the ideas of the Mādhyamika scholar Bhāviveka (sixth century), who
first adopted three interpretations of the term “ultimate” ( paramārtha). Bhāviveka, relying on
Dignāga’s theory of valid cognition ( pramāṇa), employed the method of “autonomous reason-
ing” (svatantrānumāna), proving the voidness of all dharmas by means of inference (anumāna).
Like Bhāviveka, Jñānagarbha is often categorized within Tibetan Buddhism as a member of
what was retrospectively termed the Svātantrika school, which adopted autonomous reasoning
as the means of proving voidness. In contrast to Bhāviveka, however, who relied upon Dignāga’s
thought and thoroughly rejected Asaṅga and Vasubandhu’s Yogācāra philosophy, Jñānagarbha’s
work, drawing on Dharmakīrti’s theory of valid cognition, constitutes the first transition to the
syncretic thinking of the eighth century when Śāntarakṣita and Kamalaśīla integrated Vaibhāṣika,
Sautrāntika, and Yogācāra philosophies into one coherent Mādhyamika philosophical system.
Jñanagarbha, Śāntarakṣita, and Kamalaśīla are referred to by Tibetans as “the three teachers from
the East” (rang rgyud (gyi) shar gsum), that is, present-day Bengal, and Tibetan scholars regard
Jñānagarbha as one of the three most important teachers of later Indian Mādhyamika philosophy.
Jñānagarbha’s Works
The Commentary is generally regarded as Jñānagarbha’s main work. The Verses on the
Distinction Between the Two Truths (Satyadvayavibhaṅgakārikā; hereafter Verses) is
450 DOI: 10.4324/9781351030908-40

Jñānagarbha
also attributed to Jñānagarbha, but contemporary scholars do not believe that it was
actually a distinct text composed by Jñānagarbha himself; instead, they believe it was
extracted when the Commentary was edited (Matsumoto 1978, 135 n. 2; Akahane 2013,
93–94 n. 2). The Verses and the Commentary are extant only in Tibetan translations and
can be found in the Dergé (sDe dge) edition (D) of the Tibetan canon, though not in the
Peking edition (P). Śāntarakṣita wrote a subcommentary ( pañjikā) on the Commentary,
which is also extant only in Tibetan.
In the Dergé edition, in addition to the Verses (D3881) and the Commentary (D3882), there
are six other works attributed to Jñānagarbha. Of these six, however, two are tantric texts –
Śmaśānavidhi (D1282, P2404) and Caturdevatāparipṛcchāṭīkā (D1916, P2779) – which were
most likely composed by the Jñānagarbha who was Marpa’s teacher and not the author of the
Commentary. According to the Cordier catalogue, the tantric commentary Guhyasamājaṭīkā is
also attributed to Jñānagarbha. This Jñānagarbha is also most likely the one who was Marpa’s
teacher (Lalou 1933, 13 and 146; Akahane 2003, 34 n. 5, 56 n. 6). Of the six additional texts
attributed to Jñānagarbha in the Dergé edition of the Tibetan canon, then, four may have been
composed by the author of the Commentary.
Two of these four texts are included in the tantra section of the Tibetan canon and are extant
only in Tibetan translation. One is the Verses on the Explanation of the Sacred Dhāraṇī of the
Perfection by the Infinite Gates (Ārya-anantamukhanirhāradhāraṇīvyākhyānakārikā, D2695,
P3519). A second is an autocommentary on these verses (Ārya-anantamukhanirhāradhāra
ṇīṭīkā, D2696, P3520). There are several verses in these two texts that have parallels in the
Commentary. For this reason, the Verses on the Explanation of the Sacred Dhāraṇī and the
autocommentary may very well have been composed by the same author as the Commentary
(Eckel 1987, 61 n. 29; Akahane 2003, 34–37).
The third text, the Path for the Practice of Yoga (Yogabhāvanāmārga, D3909, D4538,
P5305, P5452), presents a Mādhyamika account of the path (mārga), outlining steps on the
gradual path to awakening. The Path for the Practice of Yoga is distinguished by its inter-
pretation of the five phrases (tshig lnga po) explained in the Discourse on the Enlightenment
of Mahāvairocana (Mahāvairocanasūtra, D494, P126) as showing the Mādhyamika account
of a gradual path (Yogabhāvanāmārga D4a6–7; Mahāvairocanasūtra D256b3–4, P221b3–4,
Taisho 18, 46b23–25). A similar type of gradualism is also presented in the Commentary.
Although Jñānagarbha’s definition of relative or conventional truth (saṃvṛtisatya) in the Com-
mentary – namely “what is appearing” ( yathābhāsa/yathādarśana) – is absent from the Path
for the Practice of Yoga, there is no compelling reason to deny that the same author composed
both texts (Eckel 1987, 34).
It is difficult to determine whether the fourth text, the Commentary on the Maitreya Chapter of
the Saṃdhinirmocanasūtra (Ārya-saṃdhinirmocanasūtre ārya-maitreyakevalaparivartabhāṣya,
D4033, P5535), was composed by the author of the Commentary. The sūtra that it comments
upon, Clarifying the Hidden Intent (Saṃdhinirmocanasūtra), is an important text primarily asso-
ciated with thinkers identifying with the Yogācāra school such as Asaṅga and Vasubandhu. These
thinkers regard the sūtra as a primary source of their philosophy, particularly its accounts of
definitive meaning (nītārtha), interpretative meaning (neyārtha), the three natures (trisvabhāva),
and the three kinds of non-nature (trividhā niḥsvabhāvatā). Their understanding of these features
inspired the doctrinal conflict between what later became two distinct schools, the Mādhyamika
and the Yogācāra. The author of the Commentary on the Maitreya Chapter does not seem to pre-
sent an interpretation consistent with the Mādhyamika philosophy of the Commentary. Rather,
he seems to present a standard Yogācāra interpretation. At this point, there is not sufficient reason
for determining whether they were both composed by the same author (Eckel 1987, 31–34).
451
Ryusei Keira
Two Truths Theory

While Jñanagarbha addresses a number of questions in Buddhist philosophy, he is most famous for
his account of the two truths. In the first two verses of the Commentary, Jñānagarbha grounds his
account in the words of the Buddha and Nāgārjuna (ca. 150–250 CE). In the Fundamental Verses
on the Middle Way (Mūlamadhyamakakārikā), Nāgārjuna distinguishes between two truths:
Relying upon two truths, buddhas teach the Dharma: the conventional truth of the
world and the truth from the ultimate perspective. Those who do not understand the
distinction between these two truths do not understand what is true in the buddhas’
profound teachings. The ultimate is not taught without depending on the conven-
tional. Without understanding the ultimate, one will not attain nirvāṇa.
(XXIV.8–9)
According to Jñānagarbha, however, even though the Buddha and Nāgārjuna distinguished
the two truths, not only non-Buddhists but also his coreligionists such as Dharmapāla (ca.
530–561), Sthiramati (ca. 510–570), Devendrabuddhi (seventh century), and Śākyabuddhi
(ca. 660–720) have misunderstood the distinction. This is why, he says, he is motivated to
write the Commentary. For, according to Jñānagarbha, those who understand the distinc-
tion do not misunderstand the Buddha’s teachings; can accumulate an assemblage of both
merit and wisdom ( puṇyajñānasaṃbhāra); can accomplish what benefits themselves and
others (svaparārthasampat); and will achieve the goal, that is, becoming a buddha. From
Jñānagarbha’s perspective, understanding the distinction between the two truths is necessary
to achieve the Mahāyāna ideal of actually becoming a buddha. In what follows, I will focus on
this distinction, which Jñānagarbha held to be so important, between ultimate truth and rela-
tive, or conventional, truth.
Ultimate Truth (paramārthasatya)
Ineffable Ultimate Truth

According to his commentators, Bhāviveka and Candrakīrti (seventh century), Nāgārjuna
explains the nature of ultimate reality (*paramārthatattva) or ultimate truth ( paramārthasatya)
in his Fundamental Verses on the Middle Way:
Not to be attained by means of another, peaceful, not fabricated through conceptual

fabrication ( prapañcair aprapañcita), devoid of conceptualization, not having dif-
ferent meanings – this is the nature of reality.
(XVIII.9)
Nāgārjuna argues that the ultimate is understood individually and directly and is not attained
dependently on anyone else. It also does not have many different meanings because ultimately
entities have the same nature, that is, voidness. For Nāgārjuna, the most important aspect of
the ultimate seems to be that it is free from conceptual diversification or fabrication ( pra-
pañca). Because it is free from conceptualization, it is peaceful and beyond language: it is
ineffable. On the level of the ultimate, Nāgārjuna argues, because the domain of objects of
consciousness is not established, there is nothing to be expressed – like nirvāṇa, the nature of
things (dharmatā) is not produced, nor does it cease (XVIII.7).
452
Jñānagarbha
Corresponding to Nāgārjuna’s account of ultimate truth, Jñānagarbha says that reality (tat-
tva), that is, the ultimate truth, cannot be grasped conceptually and is free of the net of all
conceptualizations (11b). He also notes in the additional verse (antaraśloka) 5ab on verse 11:
In reality, there is nothing that can be expressed at all.
And in the Commentary on verse 7, he quotes the following line from the Akṣayamatinir-
deśasūtra:
The reality is that in which even minds do not function.
Jñānagarbha thus seems to accept Nāgārjuna’s account of ultimate truth as free from con-
ceptual diversification or conceptualization. However, Jñānagarbha also emphasizes that “the
ultimate does not appear at all in any cognition” (v. 5cd). According to the commentary on
verse 5cd (D4b1; cf. Eckel 1987, 157, 15–17), the ultimate does not appear even to the wisdom
of the omniscient one (sarvajñājñāna). Therefore, it is said in the Dharmasaṅgītisūtra, “Not
seeing anything, this is the seeing of reality.”
Jñānagarbha’s view in verse 5cd, that “the ultimate does not appear at all in any cognition,”
is the primary reason for his rejection of the Yogācāra interpretation of this line of the sūtra. The
Yogācāra interpretation is undertaken on the basis of their theory of the three natures of entities:
the imagined nature ( parikalpitasvabhāva), dependent nature ( paratantrasvabhāva), and thor-
oughly established nature ( pariniṣpannasvabhāva).2 Yogācārins, such as Asaṅga, Vasubandhu,
Sthiramati, and Dharmapāla, interpret “not seeing anything, this is the seeing of reality” to be
showing that the nonseeing of the imagined nature is the seeing of the ultimate reality of enti-
ties. And this, they argue, is the thoroughly established nature, which is defined in Yogācāra
texts as the dependent nature being free from the imagined nature. Yogācārins claim that the
appearance of the thoroughly established nature is the seeing of ultimate reality. Against the
Yogācāra interpretation, Jñānagarbha holds that the ultimate does not appear at all. According
to him, the ultimate reality of entities is the nonappearance of any nature of entities, that is, the
nonproduction of all entities. Therefore, the appearance of the thoroughly established nature is
precisely not the seeing of the ultimate reality (Keira 2004, 99 fn. 151 and 47–86).
In his Subcommentary on the Distinction (Satyadvayavibhaṅgapañjikā, D18b4; hereafter
Subcommentary), Śāntarakṣita calls the ultimate reality consisting in the nonappearance “the
ineffable ultimate” (aparyāyaparamārtha).
The Ultimate Taken as Wisdom

While Jñānagarbha accepts the Nāgārjunian account of the ultimate as free of conceptualiza-
tion, he also draws on Bhāviveka’s method to understand and have access to the ultimate by
means of wisdom arising from reasoning, nonconceptual meditation, and so forth. Jñānagarbha
interprets this wisdom as a secondary meaning of the ultimate.
Jñānagarbha accepts the three interpretations of paramārtha, a compound consisting of
parama and artha, that was first adopted in Mādhyamika thought by Bhāviveka. According
to Bhāviveka, paramārtha can be interpreted as: (1) an appositional (karmadhāraya) com-
pound, (2) a case-determined (tatpuruṣa) compound, and (3) a possessive (bahuvrīhi) com-
pound.3 The first interpretation is that parama-artha is ultimate ( parama) and is also the object
(artha) to be understood as the truth of selflessness (nairātmya). According to the second
interpretation, paramārtha is the object (artha) of the ultimate ( parama), that is, supreme
453
Ryusei Keira
nonconceptual wisdom. According to the third interpretation, it is the wisdom which has the
ultimate ( paramārtha) as its object. The third interpretation allows Bhāviveka to say that not
only the supreme nonconceptual wisdom which has the ultimate as its object but also the wis-
dom arising from hearing, reflection, and meditation is regarded as the ultimate because the
wisdom conducive to the understanding of the ultimate is consistent with the ultimate, due to
its object being ultimate (Prajñāpradīpa, D228a3–6). Although any wisdom itself is regarded
as conventional, there is a secondary ultimate that is taken as wisdom because it leads one to
understanding and having access to the ineffable ultimate.
In the Commentary, commenting on verse 4ab, Jñānagarbha applies the third interpretation.
Thus, he argues that correct reasoning (nyāya), that is, inferential cognition of the ultimate
truth consisting in the nonproduction of entities, ought to be regarded as paramārtha. In his
Subcommentary, Śāntarakṣita writes that the reasoning or inferential cognition which has the
ultimate truth as its object is called the ultimate, which is conducive to the understanding of
the ultimate (don dam pa dang mthun pa’i don dam pa).
With the establishment of the third interpretation as a precondition, commenting on verse
4ab, Jñānagarbha then argues that the object, that is, the concept of the nonproduction of enti-
ties and so forth, which is determined by means of correct reasoning, is also regarded as the
ultimate. This is because correct reasoning is ultimate ( parama) and can also be the object
(artha), just as a perception ( pratyakṣa) can be a cognition and also be an object. This can
be regarded as the first of the three previous interpretations, the appositional interpretation of
parama-artha, which presupposes the establishment of its possessive interpretation.
Moreover, in verse 9ab, Jñānagarbha claims that although the nonproduction, that is, nega-
tion of the production, which is determined by means of correct reasoning, is regarded as
the ultimate or reality (tattva) because it is conducive to the understanding of the ultimate or
reality, that negation itself is only conventional when one examines it by means of reasoning.
Indeed, when the object to be negated, that is, the production of entities, is not existent, it is
clear that the negation of it is also not existent in reality (v. 9cd), since it is absurd that negation
would have no object to be negated. Since the production of entities is completely imagined,
the negation of the completely imagined production is also imagined, and therefore, because
the negation is not existent in reality, it is just conventional (v. 10abc).
On the Qualifier “Ultimately”

Jñānagarbha uses the qualifier “ultimately” ( paramārthatas) or “really” ( yang dag par) in his
arguments aiming to prove the Mādhyamika position that all entities do not ultimately have any
intrinsic nature. Bhāviveka is regarded as the first scholar who adopted this qualifier in articulating
Mādhyamika proof-statements. However, Bhāviveka did not explicitly state which terms in the
proof-statements should be qualified by this qualifier. This qualifier was rejected by Candrakīrti,
who regarded it as incomprehensible. When Candrakīrti critiqued Bhāviveka, he interpreted the
qualifier in Bhāviveka’s reasonings as qualifying the negation, that is, “non-having,” or “being
without.” According to Candrakīrti, if the intrinsic nature of all entities is ultimately negated, it
would follow absurdly that the conventional could not exist and that therefore the Mādhyamika
inferences would either invalidate the non-erroneous perceptions of conventional entities or
would themselves be invalidated by the perceptional experiences of such conventional things.
Jñānagarbha, however, does not hold that the qualifier should qualify the negation of intrinsic
nature. According to Jñānagarbha, the qualifier qualifies “having production” (utpāda) or “hav-
ing intrinsic nature” (svabhāva). That is, he aims to prove the proposition that all entities do
454
Jñānagarbha
not have ultimately existing intrinsic natures, that is, intrinsic natures superimposed upon con-
ventional things by the opponents as being ultimately existent (Keira 2004, 30–38). He writes:
We, i.e., the Mādhyamikas, negate only those things imagined on the basis of philo-
sophical texts and the like. That is, we negate aspects that do not appear to cognitions
to which things like visible matters (rūpa) appear, aspects such as real production
and mental appearances, primordial matter ( pradhāna) and the transformations
( pariṇāma) of the elements. When we negate those imagined things, we do not incur
any invalidation by direct perception and the like at all.
(Commentary on v. 29)
We negate the primordial matter and so forth which appear to conceptualization and
have been imagined by others as being really the cause of entities. When we negate
the things imagined by others, the qualifier “really” does not qualify the negation.
Therefore, how would the conventional be nonexistent?
(additional verses 1–2 on v. 30)
In the Mādhyamika proof-statements, the objects to be negated are the opponents’ superimpo-
sitions or conceptual imaginations. Therefore, they are not existent in reality. Since the nega-
tion of the opponents’ superimpositions is also imagined, the negation is not existent in reality.
Therefore, the negation is also just conventional. However, as explained earlier, the wisdom
that arises from the reasoning which proves the negation is regarded as a secondary ultimate
because it is conducive to the understanding of the ineffable ultimate.
Conventional, or Relative, Truth (sam

 vrtisatya)
According to Jñanagarbha, only those things appearing ( yathābhāsa/yathādarśana) in con-
sciousness are conventionally true (commenting on verse 3cd). This is his definition of
conventional truth. A thing like a pot, as everyone sees, is established as being true in the con-
ventional sense because, corresponding to one’s seeing of a pot, one understands the external
pot (Commentary on v. 3cd).
Jñānagarbha distinguishes between two kinds of conventional truth. The first is “correct
conventional truth” (*tathyasaṃvṛtisatya). The second can be understood as precisely not the
correct conventional truth because calling something the “incorrect (atathya) . . . truth” (i.e.,
the truth which is incorrect) would not make sense (*atathyasaṃvṛtisatya, ad v. 8).
This type of distinction between two kinds of conventional truth can also be found in
Candrakīrti’s Introduction to Madhyamaka (Madhyamakāvatāra, VI.24–26), Commentary on
the Introduction to Madhyamaka (Madhyamakāvatārabhāṣya, La Vallée Poussin 1907–1912,
103, 7–106, 1), and In Clear Words (Prasannapadā, La Vallée Poussin 1903–1913, 493, 1–5).
Bhāviveka, on the other hand, does not seem to have distinguished two kinds of conventional
truth. Still, he believes that there can be something that is conventionally false in contrast to
what is conventionally true. For example, Bhāviveka argues that it is impossible even con-
ventionally that Īśvara, the creator God of some Indian traditions, could be the cause of the
production of the whole world (Prajñāpradīpa, D51b2).
Jñānagarbha and Śāntarakṣita use “the excluded from the correct convention” ( yang dag
pa ma yin pa’i kun rdzob; *atathyasaṃvṛti) in their works (Akahane 2013, 87–89). The term
“false convention” (log pa’i kun rdzob; *mithyāsaṃvṛti) does not appear in the Jñānagarbha’s
455
Ryusei Keira
Commentary. Nor does it appear in Śāntarakṣita’s Subcommentary, his Ornament of the Mid-
dle Way, or his Auto-Commentary on the Ornament of the Middle Way.4
Correct Conventional Truth vs. Imagined Things

Jñānagarbha gives an account of what he calls “correct conventional truth” in verse 8abc.
A mere entity (vastumātra) which is free from imagined things and arises dependently, he
writes, should be known as the correct conventional truth. Commenting on verse 8abc, he
argues: “Imagined things mean ‘real arising’ and so forth, ‘real appearance in consciousness,’
‘transformation of the primordial matter’ ( pradhāna), ‘real elements’ (bhūta), and so forth.”
That is, “imagined things” – here, primarily the ideas of non-Buddhist philosophers – are a
kind of false thing excluded from the correct conventional truth; it may be the case that some
groups of philosophers believe that these ideas are true, but they are false. In contrast to these
things imagined by other philosophers, which he regards as false even conventionally,
Mere entities are free from these imagined things. This is because the mere entities
free from imagined things have the capacity for causal efficacy (arthakriyāsamartha)
to produce an entity, which capacity corresponds to appearances. Mere entities which
arise dependent on causes and conditions should be known as the correct conven-
tional truth. That is to say, it is correct that all entities (artha) which causally appear
commonly in the consciousnesses of people from the infantile on up, are the correct
conventional truth, because external entities are existent corresponding to the appear-
ances in people’s consciousnesses.
(Eckel 1987, 160, 10–16)
According to Śāntarakṣita, in the Subcommentary, correct conventional truth must meet the
following two conditions (D24a1–2):
(1) It must be a thing appearing in consciousness, that is, a thing which appears/arises
dependently.
(2) It must be a mere entity which has the capacity for causal efficacy.
The first condition for correct conventional truth, then, is that a thing appears in conscious-
ness. This is in contrast to the ultimate truth, which is precisely not the appearing of things in
consciousness: “Not seeing anything, this is the seeing of reality.” Furthermore, Śāntarakṣita
argues in his subcommentary (D38b6) that what has the nature of appearance is acceptable
when it is not analyzed (avicāraramaṇīya/avicāraikaramya) from the ultimate perspective.
That is, an appearance can be accepted as true, but when it is analyzed from the ultimate per-
spective, we can realize that it is not as we had thought it to be.
According to the second condition, correct conventional truth is established when it meets
the first condition of appearing to consciousness and when the external entities which cor-
respond to the appearances in consciousness are established. These entities are established
by having a capacity for causal efficacy. They must also be free from the imagined nature.
An entity free from the imagined nature – that is, a mere entity – can appear in the direct
nonconceptual perceptions of many people. When the common appearance of an entity is not
established, people’s common understanding of it cannot be established. One cannot start a
debate when nothing appears in common to both parties (Commentary on vv. 18–19; Eckel
1987, 173, 23–174, 1).
456
Jñānagarbha
In contrast to mere entities, imagined things, such as primordial matter (namely things
superimposed in reliance upon doctrinal systems), cannot appear in direct nonconceptual cog-
nition and therefore cannot be the basis for a common understanding. If these imagined enti-
ties were to appear commonly to the minds of participants in debates, and the proponent and
opponent were to share a common perception, there would never be any difference of opinion
regarding doctrine and therefore no debate. If the participants had a debate, their debate would
be invalidated by means of direct perception and so forth (Commentary on v. 8d; Eckel 1987,
160, 19–24). Jñānagarbha thus concludes in verse 8d:
Imagined things are not the correct conventional truth.
Imagined things do not meet the first condition of appearing to consciousness, nor do they
meet the second condition of being causally efficacious. Imagined things – for Jñānagarbha,
typically the things postulated in doctrinal systems by philosophers – are thus not mere entities
and therefore cannot be even conventionally true.
Moreover, conventional entities, although equal in clear appearance, should be divided into
correct and incorrect entities on the basis of their capacity or incapacity for causal efficacy
(v. 12). When one sees a mirage (marīci) and believes that one is looking at water, although
the mirage is clearly appearing in direct nonconceptual cognition, one will not be able to get to
the illusory water because mirages have no capacity for causal efficacy to actually get to water
and are thus deceptive (visaṃvāda). By this logic, mirages, double moons in the perception
of people suffering from eye disease, and so forth do not meet the second condition of being
causally efficacious and are therefore outside of correct conventional truth even though they
may meet the first condition of appearing to consciousness.
How can one distinguish between what is deceptive and what is non-deceptive (avisaṃvāda)
with reference to causal efficacy? Discussing verse 12 in the Commentary, Jñānagarbha argues
that we should distinguish between what is deceptive and what is non-deceptive according to
what is commonly acknowledged to be deceptive or non-deceptive with reference to causal
efficacy because the efficacy is also ultimately without intrinsic nature. Thus, because there
is no ultimate grounding of conventional truth, one should follow the conventional consensus
of the world, based on empirical observation and reasoning, in determining what is casually
efficacious and what is not.
The Conventional (sam

 vr ti) as an Erroneous Cognition
Candrakīrti was the first Mādhyamika scholar to argue that the conventional (saṃvṛti) is a form
of ignorance (ajñāna), a claim which characterizes all possible entities that appear in conscious-
ness (In Clear Words on vv. 24–28; La Vallée Poussin 1903–1913, 492, 10–11). According to
the Great Commentary Concerning the Teachings (Abhidharmamahāvibhāṣā), this idea that all
conventional cognition was a form of ignorance was developed by Sanskrit grammarians: “The
grammarians say that this conventional cognition is the ignorance which covers things. It is just
like, for example, the bowls which cover the things inside the bowls. Therefore, it is named
saṃvṛti (covering)” (Taisho vol. 27, 548b21–22). The same interpretation can be found in the
Commentary. Jñānagarbha writes, for example, in verse 15ab that saṃvṛti is considered the
cognition by which or in which reality is covered. In the Commentary (Eckel 1987, 171, 3–4),
he argues that the cognition covering reality, which is commonly acknowledged in the world
(lokapratīti), is regarded as saṃvṛti (Wogihara 1932, 976, 3–4). Quoting the Sūtra on the Descent
into Laṅkā (Laṅkāvatārasūtra) and the Perfection of Wisdom Sūtra (Prajñāpāramitāsūtra),
457
Ryusei Keira
Jñānagarbha characterizes saṃvṛti as an erroneous cognition (bhrānti) or mistaken cognition

(viparyāsa) (Keira 2016, 47–48). According to Jñānagarbha, then, it is precisely this erroneous
and mistaken cognition that is commonly acknowledged in the world (lokapratīti).
However, this does not mean that saṃvṛti is necessarily that which is acknowledged by
the majority of the world. Jñānagarbha notes in the commentary on verse 15cd (Eckel 1987,
171, 18–20) that mistaken cognition is lokapratīti, and this lokapratīti means that it is true in
the realm where valid reasoning can be properly employed. In this respect, his account differs
from Candrakīrti’s; Candrakīrti accepts what is acknowledged by the majority of the world
(lokaprasiddha) even though it is unacceptable when examined by means of valid cognition
(La Vallée Poussin 1903–1913, 67, 7–69, 5, 1907–1912, 179, 14–181, 7).
Gradualism on the Buddhist Philosophical Path

Asaṅga, Vasubandhu, Sthiramati, Dharmapāla, and other Yogācāra philosophers interpret
the term “mind-only” (cittamātra) as meaning the non-existence of external objects. Neither
Bhāviveka nor Candrakīrti, however, accept the Yogācāra interpretation of this term. In his
Verses on the Heart of Madhyamaka (Madhyamakahṛdayakārikā), Bhāviveka argues that
cittamātra means that there is no agent (kartṛ) and enjoyer (bhoktṛ) other than mind (citta)
(V.28cd). And in the Commentary on the Introduction to Madhyamaka, Candrakīrti states that
cittamātra simply means that “mind alone” is preeminent among the dharmas (on VI.87; Eckel
1987, 185, 6–20 and 190, 5–8). Thus, although Bhāviveka and Candrakīrti both reject the
standard Yogācāra interpretation of cittamātra, they have different understandings of the pro-
cess of establishing the wisdom of the selflessness of all dharmas.
In the Lamp of Wisdom (Prajñāpradīpa), Bhāviveka argues that one should simultaneously
understand the selflessness or nonexistence of both external objects and of cognition or mind
(D247a1–3; Prajñāpradīpaṭīkā D353a7–354a8). On the other hand, Candrakīrti presents an
account of the path as gradual, involving first understanding the selflessness of external objects
and then later understanding the selflessness of mind. According to Candrakīrti, the Yogācāra
understanding of cittamātra in the sūtras as meaning the nonexistence of external things
depends on the interpretative meaning of the teaching and not on its definitive meaning. In his
Introduction to Madhyamaka (V.94cd) and the commentary, Candrakīrti argues that the Buddha
taught the doctrine of cittamātra to help people remove their attachment to external objects (La
Vallée Poussin 1907–1912, 195, 1–9). The Buddha only gradually (rim gyis; *krameṇa) leads
people to an understanding and experience of the absence of intrinsic nature (niḥsvabhāvatā),
that is, selflessness of all dharmas. Thus, the Buddha first taught the selflessness of external
things and then taught the selflessness of mind or cognition so that those on the path could more
easily understand the selflessness of mind (Commentary on the Introduction to Madhyamaka on
VI.96; La Vallée Poussin 1907–1912, 199, 1–11). Candrakīrti seems to accept the interpretative
or provisional meaning of cittamātra as a helpful means to understand the selflessness of mind.
Jñānagarbha presents an account of the path of understanding as gradual that is similar to
Candrakīrti’s, although it is not clear whether Jñānagarbha knew of Candrakīrti’s work. For
example, in the Commentary, Jñānagarbha writes on verse 32:
The Buddha dispels people’s ideas of real entities by means of a gradual teaching of
five aggregates (skandha), eighteen realms (dhātu), twelve spheres (āyatana), mind-
only, and the selflessness of all dharmas, according to the mental capacities of his
listeners, and taught people the liberation from bondage.
(Eckel 1987, 183, 18–21)
458
Jñānagarbha
In this gradual philosophical path, cittamātra seems to mean the nonexistence of external
objects, although it is not clear whether he uses it in its interpretative meaning. Just after his
comments on verse 32, in verse 33ab he quotes the Yogācāra commentator Sthiramati (sixth
century), who, in his interpretation of Vasubandhu in the Commentary on the Thirty Verses on
Representation (Triṃśikābhāṣya), writes:
Imagination (rtog pa) is the mind and mental factors which belong to the three worlds
and which function with the superimposed forms of external objects.
(Buescher 2007, 108, 8)
As Sthiramati notes, this imagination is explained as the imagination of what is unreal

(abhūtaparikalpa) in Distinguishing the Middle from the Extremes (Madhyāntavibhāga), one
of five texts attributed to Maitreya, the future Buddha, who along with Asaṅga and Vasu-
bandhu was regarded as a founder of the Yogācāra tradition. In the Commentary on the Thirty
Verses on Representation, this sentence serves as the reason for stating that everything – all
cognizable objects – is the mind representing to itself (vijñaptimātra); this is the doctrine of
mind-only. It is significant that Jñānagarbha quotes this sentence from the Commentary on the
Thirty Verses on Representation because it implies that he accepts Vasubandhu’s justification
for the establishment of the cittamātra, or the nonexistence of external objects.
According to Śāntarakṣita in his Subcommentary, this imagination, which is the mind and
mental factors in the three worlds, is what is appearing, that is, conventional (D46b7–47a1).
Ultimately, then, it is without intrinsic nature.
Neither Jñānagarbha’s Commentary nor Śāntarakṣita’s Subcommentary gives a compre-
hensive explanation of the gradual path itself. We can, however, draw on other texts attributed
to Jñānagarbha, especially the Path for the Practice of Yoga, which presents the bodhisattva
path in some detail. There, Jñānagarbha claims that first, by relying upon the Buddhist śrāvaka
teachings of aggregates, realms, and spheres associated with Vaibhāṣikas and Sautrāntikas, one
can transcend the non-Buddhist understanding of permanent entities such as a self (ātman).
Next, by relying upon the teaching of the mind-only associated with Yogācāra figures such as
Vasubandhu and Asaṅga, one can transcend the teaching of śrāvakas. Finally, relying upon
the Mādhyamika teaching of the selflessness of all dharmas, one can transcend the mind-only
framework and directly understand the nonproduction and voidness of all external and internal
dharmas. In Path for the Practice of Yoga, Jñānagarbha’s account of Buddhist gradualism is
not fundamentally different from the system of his successors, Śāntarakṣita and Kamalaśīla,
whose work also draws on multiple Buddhist traditions (Kajiyama 1978).
Jñānagarbha’s Philosophical Position and Future Research

Tibetan scholars have proposed a wide array of interpretations of Jñānagarbha’s two truths
theory, especially his philosophical position on convention (saṃvṛti); as a result, Jñānagarbha
is regarded as representative of several different Tibetan doxographical categories. I do not
take this variety of interpretations to suggest that these Tibetan accounts are somehow wrong.
Rather, I believe that Jñānagarbha’s text actually lends itself to each of these multiple and
competing Tibetan interpretations.
One interpretation that can be found in the Commentary, later associated with what was
termed the Sautrāntika-Mādhyamika position, conventionally accepts the existence of exter-
nal objects and the cognition in which the object appears can be established. Many Geluk
scholars, such as Jamyang Zhepa (’Jam dbyangs bzhad pa, 1648–1722), count Jñānagarbha
459
Ryusei Keira
as a Sautrāntika-Mādhyamika. Their interpretation is justified by Jñānagarbha’s comments

on verses 3cd and 8abc, where he argues that external entities exist because they correspond
to appearances in people’s consciousnesses. A second interpretation, the position of the
Mādhyamika who proceeds in accordance with what is acknowledged in the world, can also
be justified by Jñānagarbha’s text. Commenting on verse 15, which characterizes saṃvṛti as an
erroneous cognition, Jñānagarbha argues that erroneous cognition is commonly acknowledged
in the world (lokapratīti). The Kadampa scholar Üpa Losel (dBus pa blo gsal; ca. 1265–1355)
interpreted Jñānagarbha in this way (Mimaki 1982, 27–54). A third interpretation, associated
with what came to be called the Yogcācāra-Mādhyamika position, can also be established on
the basis of the Commentary. As Jñānagarbha writes in reference to verse 23,
When one examines atoms in terms of their separate parts, one understands that they
too do not exist at all. So there can be no real cause for the designation of something
like a tree. Therefore, saṃvṛti is nothing other than appearances and does not depend
on any real causes. This should be accepted.
(Eckel 1987, 176, 17–21)
Here, negating the existence of atoms, Jñānagarbha argues that appearances in consciousness
do not depend on any external real causes like atoms. And, as we saw earlier, at the conven-
tional level, he accepts the mind-only doctrine explained in verses 32–33ab and the commen-
tary. This interpretation of Jñānagarbha as a Yogācāra-Mādhyamika is defended by the Sakya
scholars Gorampa Sonam Senge (Go rams pa dSod nams seng ge; 1429–1489) and Büton
Rinchendrup (Bu ston rin chen grub; 1290–1364). Note that Jñānagarbha seems to accept
conventionally the establishment of reflexive cognition (sva-saṃvitti, -saṃvid, -saṃvedana),
which means that cognition is cognized by itself and does not rely upon another cognition to
cognize itself (Akahane 2004, 100, 12–102, 16).
These differing Tibetan interpretations of Jñānagarbha’s philosophical position on saṃvṛti
can all be justified by passages in the Commentary, even as he very clearly defines conven-
tional truth as that which appears to consciousness. One might be tempted to argue that rather
than any of the more elaborate Tibetan interpretations, Jñānagarbha’s insistence on just the
“appearance in consciousness” is the most important characteristic of his account of saṃvṛti.
His account of lokapratīti also should be understood as the appearance of erroneous cognition
covering reality. Therefore we should not regard lokapratīti as a fundamental characteristic of
Jñānagarbha’s presentation of saṃvṛti.
After proving that there is no valid cognition to establish real entities (antaraśloka 1 on v.
13) and that those entities are invalidated by valid cognition (v. 14), Jñānagarbha argues in
verse 15ab and the commentary that saṃvṛti is erroneous cognition. From verse 15ab, his view
of external entities seems to shift from the Sautrāntika notion that these external entities exist
to the Yogācāra account that these apparently real entities grasped by erroneous cognition are
completely nonexistent (v. 15cd). Thus, appearances do not depend on any external real causes
like atoms (Commentary on v. 23), physical things like visible matter (rūpa) are the appear-
ances of the cognition free from imagination (rtog pa) (v. 30cd), and external entities are
imagined by the mind (vv. 32–33ab). Taking the position that saṃvṛti is appearance, he seems
to explain two levels of saṃvṛti: the Sautrāntika realistic level and the Yogācāra idealistic or
mind-only level, the latter of which corresponds to the gradually ascending steps to remove
attachment to real external entities. When we recognize that Jñānagarbha understands saṃvṛti
on two different levels, we can be justified in suggesting that all of the Tibetan interpretations
of the two truths in the Commentary may be only partial.
460
Jñānagarbha
It will be up to future research to further clarify Jñānagarbha’s philosophy. Jñānagarbha

scholarship will benefit from greater understanding of how he was influenced by Dharmakīrti’s
refutation of external objects and what has been called a “sliding scale of analysis” (Kellner
2011), according to which one can have multiple ontological accounts which are hierarchically
ordered. Clarifying Jñānagarbha’s philosophical position will also enable us to understand dif-
ferences and commonalities in the philosophies of Jñānagarbha, Śāntarakṣita, and Kamalaśīla
and thereby help us better understand the history of late Indian Mādhyamika.
Notes
1 According to the Nyāyabinduṭīkāṭippaṇī, there was a scholar named Jñānagarbha who developed an
interpretation of the theory of mental perception (mānasapratyakṣa), although it is not clear whether this
is the same Jñānagarbha as the author of Commentary on the Distinction or not (Mimaki 1988, 248–49).
2 For more on the Yogācāra framework of the three natures, see Jonathan C. Gold’s chapter in this
volume.
3 For more on Bhāviveka’s account of the ultimate, see Karl-Stéphan Bouthillette’s chapter in this volume.
4 In his Abhisamayālaṃkārālokā Prajñāpāramitāvyākhyā, the eighth-century Buddhist philosopher
Haribhadra does not use mithyāsaṃvṛti. He uses atathyasaṃvṛti, although in his text, one can find log
pa as a Tibetan translation of atathya (Wogihara 1932, 594, 21; P267a1). Kamalaśīla may have been
the first Buddhist author who started to use the term mithyāsaṃvṛti (Madhyamakāloka D230a4–5;
Kano and Li 2017, 154, 9–10 for Munimatālaṃkāra).
Bibliography
Akahane, Ritsu (赤羽律). 2003. “Nendai-kakutei no shihyō toshite no avicāraramaṇı̄ ya” (年代確定の指
標としての avicāraramaṇı̄ ya). Nanto Bukkyō (南都仏教) 83: 33–59.
———. 2004. “Kyūkyokuteki-shinri to sezokuteki-shinri: Junyānagaruba no ni-shinri-setsu to chibetto
ni okeru shisōteki-tachiba” (究極的真理と世俗の真理: ジュニャーナガルバの二真理説とチベ
ットにおける思想的立場). Tetsugaku Kenkyū (哲學研究) 577: 80–114.
———. 2013. “The Two Truth Theory of Jñānagarbha.” Memoirs of the Research Department of the
Toyo Bunko 71: 69–107.
Buescher, Hartmut. 2007. Sthiramati’s Triṃśikāvijñaptibhāṣya: Critical editions of the Sanskrit text and
its Tibetan translation. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
Eckel, Malcolm David. 1987. Jñānagarbha’s Commentary on the Distinction Between the Two Truths:
An Eighth Century Handbook of Madhyamaka Philosophy. Albany, NY: State University of New
York Press.
Inagaki, Hisao. 1987. Anantamukhanirhāra-dhāraṇī Sūtra and Jñānagarbha’s commentary: A study and
the Tibetan text. Kyoto: Nagata Bunshōdō.
———. 1999. Amida Dhāraṇī Sūtra and Jñānagarbha’s commentary. Kyoto: Ryukoku Gakkai.
Kajiyama, Yūichi. (1978) 1989. “Later Mādhyamikas on Epistemology and Meditation.” In Mahāyāna
Buddhist Meditation: Theory and Practice, edited by Minoru Kiyota, 114–43. Honolulu: University
of Hawai’i Press. Reprinted in Y. Kajiyama, Studies in Buddhist Philosophy: Selected papers, edited
by Katsumi Mimaki, 99–128. Kyoto: Rinsen Book Co.
Kanō, Kazuo, and Li, Xuezhu. 2017. “Bonbun-kōtei ‘Muni-ishu-shōgon’ Daiisshō (fol. 59v4–61r5):
‘Chūgan-kōmyō’ sezoku no teigi-kasho-itsubun” (梵文校訂『牟尼意趣荘厳』第一章 (fol.
59v4–61r5): 『中観光明』世俗の定義箇所佚文). Mikkyō Bunka (密教文化) 239: 141–60 (7–26).
Keira, Ryusei. 2004. Mādhyamika and epistemology: A study of Kamalaśīla’s method for proving the
voidness of all dharmas. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien.
———. 2016. “ ‘Chūgan-kōmyō-ron’ (Madhyamakāloka) Kōshuchō-daiisshō ‘Seiten ni yoru issaihō-
mujishōsei no shōmei’ no Kenkyū (1); Wayaku, Chūkai, Chibettogo-koutei-tekisuto” (『中観光明
論』 (Madhyamakāloka) 後主張第1章「聖典による一切法無自性性の証明」の研究 (1)：和
訳・註解・チベット語校訂テキスト). Acta Tibetica et Buddhica 9: 1–121.
Kellner, Birgit. 2011. “Dharmakı̄ rti’s Criticism of External Realism and the Sliding Scale of Analysis.”
In Religion and Logic in Buddhist Philosophical Analysis, ed. H. Krasser et al., 291–8. Vienna: Verlag
der Österreichischen Akademie der Wissenschaften.
461
Ryusei Keira
Lalou, Marcelle. 1933. Répertoir du tanjur d’après le catalogue de P. Cordier. Paris: Bibliothèque
Nationale.
Matsumoto, Shirō (松本史朗). 1978. “Jñānagarbha no ni-tai-setsu” (Jñānagarbha の二諦説). Bukkyōgaku
(佛教學) 5: 109–37.
Mimaki, Katsumi. 1982. Blo gsal grub mtha’: Chapitres IX (Vaibhāṣika) et XI (Yogācāra) édités et chapi-
tre XII (Mādhyamika) édité et traduit. Kyoto: Zinbun Kagaku Kenkyūsho.
———. 1988. “Kyōryōbu (経量部).” In Iwanami-kōza Tōyō-shisō 8: Indo-Bukkyō 1 (岩波講座東洋思
想 8: インド仏教 1), edited by Gadjin M. Nagao et al., 226–60. Tokyo: Iwanami-shoten.
La Vallée Poussin, Louis de. (1903–1913) 1970. Mūlamadhyamakakārikās (Madhyamikasūtras) de
Nāgārjuna: Avec la Prasannapadā, commentaire de Candrakı̄rti. St. Petersburg: Imprimerie de
l’Académie Impériale des Sciences. Reprint, Osnabrück: Biblio Verlag.
———. (1907–1912) 1970. Madhyamakāvatāra par Candrakı̄rti: Traduction tibétaine. St. Petersburg:
Imprimerie de l’Académie Impériale des Sciences. Reprint, Osnabrück: Biblio Verlag.
Schiefner, Anton. 1869. Tāranātha’s Geschichte des Buddhismus in India aus dem tibetischen Übersetzt.
St. Petersburg: Buchdruckerei der Kaiserlichen Akademie der Wissenschaften.
baden: O. Harrassowitz.
Tauscher, Hermut. 1999. Phya pa chos kyi seng ge, dBu ma shar gsum gyi stong thun. Vienna: Arbeitskreis
für Tibetische und Buddhistische Studien.
Wogihara, Unrai. (1932) 1973. Abhisamayālaṃkārālokā prajñāpāramitāvyākhyā: The work of Harib-
hadra. Tokyo: Toyo Bunko. Reprinted in 1973, Tokyo: Sankibo.
462
28
ŚĀNTARAKS․ITA
Climbing the Ladder to the Ultimate Truth
Allison Aitken
Introduction to Śāntaraksita’s Life and Works

The scholar-monk and prolific author Śāntarakṣita (c. 725–788)1 left a lasting and significant
impact on both Indian and Tibetan Madhyamaka Buddhist philosophy. He is known for his
synthesis of Nāgārjuna’s (c. second century) Madhyamaka with elements of Dignāga (c. 480–
540 CE) and Dharmakīrti’s (c. seventh century) tradition of logic and epistemology as well
as Yogācāra idealist ontology. Śāntarakṣita’s works are characterized by an emphasis on the
indispensable role of rational analysis on the Buddhist path as well as serious and systematic
engagement with competing Buddhist and non-Buddhist schools of thought.
Śāntarakṣita is said to have been abbot of the great monastic university of Nālandā in
Magadha (present-day Bihar, India) and counted Kamalaśīla (c. 740–795) and Haribhadra
(late eighth century) among his most prominent students. Yet aside from this, few biographi-
cal details about Śāntarakṣita’s life in India remain. He did, however, play a central role in the
early transmission (snga dar) of Buddhism to Tibet, and numerous semi-legendary reports
of his activities there survive. The earliest accounts agree that, upon receiving an imperial
invitation to Tibet from King Trisong Detsen (khri srong lde btsan) (742–797?), Śāntarakṣita
oversaw the establishment of the first Tibetan Buddhist monastery at Samyé (bsam yas),2 serv-
ing as its abbot and ordaining the first Tibetan Buddhist monks into the Mūlasarvāstivāda
monastic order (c. 779), whereupon he became known in Tibet as the “Khenpo (mkhan po)/
Ācārya Bodhisattva,” or “Abbot Bodhisattva.” According to Tibetan sources, Śāntarakṣita’s
own ordination lineage proceeds as follows: Śāriputra → Rāhula → Nāgārjuna → Bhāviveka
(c. sixth century) → Śrīgupta (c. seventh century) → Jñānagarbha (early eighth century) →
Śāntarakṣita.3 This lineage also reflects philosophical affinities among these authors, with the
later figures influenced by Bhāviveka’s interpretation of Nāgārjuna’s Madhyamaka (which
would later come to be known as *Svātantrika-Madhyamaka).4
A sizeable corpus is attributed to Śāntarakṣita, spanning a range of genres and subject matters,
including Madhyamaka metaphysics, logic and epistemology, Buddhist path literature, tantra,
as well as several praises. His two most important independent treatises are the Compendium of
True Principles (Tattvasaṃgraha, hereafter Compendium = TS) and the Ornament of the Mid-
dle Way (Madhyamakālaṃkāra, hereafter Ornament = MA) together with an autocommentary
(Madhyamakālaṃkāravṛtti = MAV). Kamalaśīla authored lengthy commentaries ( pañjikā-s)
DOI: 10.4324/9781351030908-41 463

Allison Aitken
on both the Compendium (TSP) and the Ornament (MAP). In practice, Śāntarakṣita’s basic
texts (particularly the Compendium) are standardly read with the aid – and thus through the
lens – of Kamalaśīla’s commentaries, such that Śāntarakṣita’s thought and that of his principal
student are often inextricable.
Several notable commentaries are also attributed to Śāntarakṣita, including one on
Jñānagarbha’s The Distinction Between the Two Truths and its Autocommentary
(Satyadvayavibhaṅga and Satyadvayavibhaṅgavṛtti = SDV and SDVV, Satyadvayavibhaṅgap
añjikā = SDVP), as well as a commentary on Dharmakīrti’s The Logic of Debate (Vādanyāya)
titled Commentary on the Logic of Debate: Elucidation of Its Meaning (Vādanyāyaṭīkā
Vipañcitārthā).5 Of all the works attributed to Śāntarakṣita, only the Compendium, his com-
mentary on Dharmakīrti’s The Logic of Debate, and a tantric-cum-epistemological work,
Establishing the Truth (Tattvasiddhi),6 survive in Sanskrit. Śāntarakṣita’s references to his own
works yield the following chronology of composition: Ascertainment of the Ultimate, Com-
pendium, Commentary on the Logic of Debate, Commentary on the Ornament. This suggests
that the Commentary on the Ornament represents his most mature thought and is his definitive
work on Madhyamaka.7
As noted previously, Śāntarakṣita’s place in the history of Madhyamaka philosophy is per-
haps most remarkable for his synthesis of Nāgārjuna’s Madhyamaka with elements from the
Dignāga-Dharmakīrti tradition of logic and epistemology together with Yogācāra idealist onto
logy. Though not the first Mādhyamika to be influenced by Dignāga and Dharmakīrti, Śāntarakṣita
looks to be the first Mādhyamika to author a commentary on one of Dharmakīrti’s works. The
influence of the Dignāga-Dharmakīrti tradition of epistemology, according to which testimony
as a source of knowledge is reduced to a form of inference, is reflected in Śāntarakṣita’s empha-
sis on the central role of rational analysis in the gradual progression toward a correct metaphysi-
cal view. As Śāntarakṣita repeatedly suggests, the ideal reader of his works and the ideal trainee
on the Buddhist path is a discerning person ( prekṣāvat), that is, a rational epistemic agent.8
Prior to Śāntarakṣita, Śrīgupta is noteworthy for integrating the Dignāga-Dharmakīrti
tradition of logic and epistemology into his presentation of Madhyamaka, but Śrīgupta
rejects Yogācāra ontology without qualification.9 And while Jñānagarbha subsequently
alludes to Yogācāra conceptual frameworks in his presentation of the Madhyamaka theory
of two truths (satyadvaya), viz. the conventional truth (saṃvṛtisatya) and the ultimate truth
( paramārthasatya) (see, e.g., SDVV ad SDV 30), it is Śāntarakṣita who explicitly formalizes
the incorporation of Yogācāra ontology into his presentation of Madhyamaka, though rele-
gated to the domain of conventional truth. Śāntarakṣita accepts as conventionally true not only
the Yogācāra doctrine that apparently external objects are merely mental in nature (cittamātra)
(MA 91–93) but also the Yogācāra claim that cognition is reflexively aware (svasaṃvitti/
svasaṃvedana) (MAV ad MA 91). Following Śāntarakṣita, Kamalaśīla and Haribhadra like-
wise adopt the Dignāga-Dharmakīrti tradition of logic and epistemology while also taking a
conciliatory approach to Yogācāra. Below, we will return to the question of how best to under-
stand Śāntarakṣita’s Yogācāra-Madhyamaka synthesis.
Both of Śāntarakṣita’s main treatises, the Compendium and the Ornament, exemplify his
wide-ranging and systematic critical engagement with the philosophical views of competing
Buddhist and non-Buddhist schools of thought. The Compendium – over 3,600 stanzas organ-
ized in twenty-six chapters – in some ways resembles a critical doxographical survey of the
eighth-century Indian religio-philosophical landscape, yet such a description does not adequately
reflect its dialogical structure or its in-depth engagement with these competing systems.10 The
first twenty-three chapters of the treatise examine and ultimately reject a succession of cosmogo-
nical theories, ontological categories, semantic theories, epistemological theories, and candidate
464
Śāntaraks∙ita
sources of knowledge advanced by Nyāya, Vaiśeṣika, Mīmāṃsā, Sāṃkhya, Jaina, Vedānta, and
Lokāyata/Cārvāka traditions, as well as Buddhist traditions such as the Vātsīputrīya.11 The final
three final chapters that constitute nearly the second half of the treatise are largely aimed at
Mīmāṃsakas, first rejecting their claim that the Vedas lack a human author (apauruṣeya), next
critically examining their theory that veridical cognition is self-certified (svataḥprāmāṇya), and
finally concluding with an argument in support of the possibility of omniscience. Given the
breakdown of the text, the Compendium would seem primarily concerned with competing theo-
ries of the Nyāya and Mīmāṃsā, and Śāntarakṣita makes important contributions in formulating
Buddhist responses to the Naiyāyika philosopher Uddyotakara (fl. c. 600), as well as the Pūrva-
Mīmāṃsaka Kumārila Bhaṭṭa (fl. c. 660), both of whom were influential critics of Dignāga.
The Ornament similarly rejects a range of positions from competing Buddhist and non-
Buddhist systems, though it is more metaphysical in its focus than the Compendium, with
roughly two-thirds of the ninety-seven stanzas devoted to the neither-one-nor-many argument
(ekānekaviyogahetu), which sets out to demonstrate that nothing possesses an intrinsic nature
(svabhāva). While the argument takes aim at entities advanced by many of the same non-
Buddhist traditions addressed in the Compendium, the Ornament is concerned foremost with
competing Buddhist views.
In presenting and rejecting the views he addresses in both the Compendium and the Orna-
ment, Śāntarakṣita utilizes a dialectical/pedagogical device of provisionally adopting what
he deems to be successively more rational positions in order to reject less rational ones. Sara
McClintock has influentially described this method whereby Śāntarakṣita argues from progres-
sively shifting perspectives as a “sliding scale of analysis.”12 To illustrate, in the Ornament,
Śāntarakṣita adopts the Sautrāntika representationalist theory of perception in order to reject
the Vaibhāṣika direct realist theory of perception. He then assumes the Yogācāra idealist posi-
tion on which mental representations have no external referents in order to reject Sautrāntika
representationalism. Finally, he uses the Madhyamaka account of the nonexistence of funda-
mentally real cognition to reject Yogācāra theories of the mind and mental content.
Although the Compendium includes several allusions to the superiority of the Madhyamaka
perspective,13 it might be read as culminating in the Yogācāra perspective. By contrast, the
Ornament, in which Śāntarakṣita presents his definitive account of Madhyamaka, devotes
more critical attention to Yogācāra than to any other competing system. Yet it is in this same
text that Śāntarakṣita presents his provisional endorsement of Yogācāra idealism on the level
of conventional truth. The following sections will take up Śāntarakṣita’s contributions to the
Madhyamaka theory of two truths in the Ornament.
Ultimate Truth and the Neither-One-Nor-Many Argument

The Madhyamaka central commitment, or ultimate truth, is the emptiness of intrinsic nature
(svabhāvaśūnyatā), which might be glossed as the universal negation of ontological inde-
pendence. In other words, according to Mādhyamikas, nothing lays claim to ontological self-
sufficiency, which is commonly identified as a necessary condition for fundamentality and
substancehood. The Madhyamaka view thus might be described as a kind of thoroughgoing
anti-foundationalism as well as a form of substance nihilism. But if there are no ontologically
independent or fundamental entities, then whatever there is depends for both its nature and its
existence on something else.
In his Ornament, Śāntarakṣita makes a major contribution to the Madhyamaka canon of
arguments for emptiness with his presentation of the neither-one-nor-many argument. Although
he expands on his predecessor Śrīgupta’s more condensed formulation of the argument in the
465
Allison Aitken
Introduction to Reality, it is Śāntarakṣita’s influential Ornament that popularizes the argument

in both Buddhist India and Tibet. In his Illumination of the Middle Way (Madhyamakāloka),
Kamalaśīla presents the neither-one-nor-many argument among a set of five Madhyamaka
arguments for emptiness, which subsequently became known in Tibet as the “five great argu-
ments” for emptiness (gtan tshigs chen po lnga).
Nāgārjuna articulates an early precursor to the neither-one-nor-many argument in his Pre-
cious Garland (Ratnāvalī), stating:
Something is not a unity if it has multiple loci. There is nothing that lacks multiple
loci. In the absence of any unity, neither is there a multiplicity.
(Precious Garland 1.71)
Here, Nāgārjuna argues that whatever is divisible into multiple discrete spatial or temporal loci
does not count as a true unity. And everything, he claims, is so divisible. Just as each bit of matter –
regardless of how minute – has a right side and a left side, a top and a bottom, and so on; likewise,
each moment of time – no matter how brief – has a beginning, a middle, and an end. Otherwise,
the existence of the spatially and temporally extended ordinary objects that populate our world –
like computers, kangaroos, and cognitions – would be impossible. After all, the thought goes, how
could fundamental building blocks that lack spatial/temporal extension ever yield anything that
has spatial/temporal extension? And since a plurality presupposes unities as its basic constituents,
if there is nothing that is truly one, then neither is there anything that is truly many.
Śrīgupta’s expanded formulation of the neither-one-nor-many argument makes explicit
the implication that nothing has an intrinsic nature, while also formalizing the argument and
defending its soundness according to the Dignāga-Dharmakīrti tradition of logic and episte-
mology. Closely following Śrīgupta’s Introduction to Reality (TA 1), Śāntarakṣita articulates
the central inference of the neither-one-nor-many argument in the opening stanza of the Orna-
ment as follows:
In reality, everything that is theorized by our own and other schools of thought is
without intrinsic nature, due to lacking an intrinsic nature that is either one or many,
like a reflection.
(MA 1)
The argument poses a destructive dilemma, which says: if anything has an intrinsic nature,
then it is either one or many. Śāntarakṣita, in effect, argues that nothing can satisfy either dis-
junct of the consequent and therefore, by modus tollens, that nothing can satisfy the anteced-
ent. Upon analysis, nothing possesses an intrinsic nature.
Three features of the disjunctive property pair <one or many> are critical for the argument
to go through. First, the terms translated as “one” and “many” here (eka and aneka in Sanskrit)
are perhaps more precisely rendered as “unity” and “non-unity,” reflecting the fact that they
are a mutually exclusive and contradictory pair, conforming to the conceptual, logical, and
grammatical structure F and not-F. As Śāntarakṣita makes clear, if anything had an intrinsic
nature, then on pain of violating the law of excluded middle, it would have to either be a unity
or non-unity:
Aside from unity and not-unity, an object’s having some other classification is impos-
sible, since it is established that these two properties are mutually exclusive.
(MA 62)
466
Śāntaraks∙ita
Second, the operator, or qualifier, “in reality” (tattvataḥ) in the statement of the central infer-
ence clarifies that the target here is a true unity and a true multiplicity. This should be con-
trasted with a merely conventional status, like the unity of an aggregate such as a flock of
sheep or a heap of sand. And, as indicated by Nāgārjuna, a true unity is defined as a mereologi-
cal simple, that is, something that lacks proper parts, where something has proper parts just in
case it is either physically or conceptually divisible. While physical divisibility is more or less
straightforward, we can understand that x is conceptually divisible in the mereological sense
just in case there are conceptually isolatable proper parts ys that compose x, such that x is the
sum of the ys. A true multiplicity, then, is something that has proper parts, the most basic of
which are themselves true unities.
Finally, a third feature of this property pair is that, unlike most contradictories, unity and
non-unity share not only a conceptual priority relation but also a metaphysical priority rela-
tion: the existence of a non-unity presupposes the existence of some unities. As Nāgārjuna
pointed out in his Precious Garland, a plurality requires singular things as its building blocks.
Śāntarakṣita explains,
Thus, a “multiplicity” is defined as a composite of unities. If no unity exists, neither

does a multiplicity, just like if no trees exist, neither does a forest.
(MAV ad MA 61)
But if a multiplicity depends for its existence on some unities just like a forest does on some
trees, then a multiplicity is not a candidate for ontologically independent being after all. As
it turns out, then, true unity is a necessary criterion for ontological independence. And this
should not be so surprising: just like ontological independence, true unity is commonly cited
as a necessary condition for fundamentality as well as for substantial reality. The neither-one-
nor-many argument thus reduces to a rejection of true unities, which is to say a rejection of
mereological simples.
The Neither-One-Nor-Many Argument Against Material Simples

Śāntarakṣita’s rejection of material simples closely follows Vasubandhu’s (c. fourth–fifth cen-
tury) anti-atomist argument in his Yogācāra work, the Twenty Verses (Viṃśikā 11–15). This
section of his neither-one-nor-many argument also features in the “Examination of External
Objects” (Bahirarthaparīkṣā) chapter of his Compendium, wherein he assumes the Yogācāra
perspective. Śāntarakṣita targets three kinds of views about how atoms aggregate to constitute
composites, which recur in debates of this kind in pre-modern Indian philosophy:
i. Each atom conjoins with surrounding atoms.

ii. Atoms have interceding space between them.
iii. Atoms are spatially continuous, neither conjoining with surrounding atoms nor having
interceding space between them.14
To each of these views, Śāntarakṣita, in effect, poses the following dilemma: If matter is con-
stituted by fundamental, simple particles, then those particles either face surrounding particles
at one and the same locus or at spatially differentiable loci. If, on the one hand, fundamental
particles did not have spatially differentiable loci at which to face neighboring particles, and
were thus spatially unextended, then they could not compose an extended composite. If, on
the other hand, fundamental particles did have spatially differentiable loci at which to face
467
Allison Aitken
surrounding particles (e.g., a right side, a left side, etc.), then they would have spatially dis-
crete parts, which means that they would be composites themselves and could not be funda-
mental unities after all. As Śāntarakṣita argues in both his Ornament and Compendium,
Whether atoms are (i) conjoined, (ii) located at a distance from one another, or (iii)
located continuously without interceding space, if the very same part of the central
atom in a composite which is facing one atom were also imagined to be facing another
atom, then the aggregation of atoms composing mountains, etc., would not be feasible.
(TS 1989–90 = MA 11–12)
If instead it were accepted that a different part of the central atom faced another
atom, then how indeed could an atom like that, i.e., with distinct parts facing different
atoms, be truly unitary?
(TS 1991 = MA 13)
He concludes that there is no account on which matter could be founded in simple particles.
And given the metaphysical priority of true unities to a true multiplicity, in the absence of
material simples, a material multitude is also precluded.
The Neither-One-Nor-Many Argument Against Mental Simples

While the Compendium restricts the subject of the neither-one-nor-many argument to external
objects, the Ornament grants the argument a universal scope of application. Having rejected
the true unity of any extramental entities within the first fifteen stanzas of the Ornament,
Śāntarakṣita devotes stanzas 16–60 to rejecting the true unity of the mind, addressing a variety
of Buddhist and non-Buddhist accounts of the mind and mental content, with the argument
culminating in a sixteen-stanza section targeting Yogācāra theories. Śāntarakṣita introduces
this section, remarking:
Even though the Yogācāra view has merit, we shall consider whether such mental
entities are to be accepted as real or as satisfactory only when not analyzed.
(MA 45)15
The succeeding argument turns on an analysis of the relation between the mind and mental
content qua cognition ( jñāna) and mental representations (ākāra).
Śāntarakṣita targets two families of views from the Yogācāra tradition on the ontological
status of mental representations:
i. Representational realism (*satyākāravāda): representations are real in the same way as

cognition is taken to be.16
ii. Representational antirealism (*alīkākāravāda): representations are unreal figments.17
It is important to keep in mind that for Yogācārins, who reject mind-independent material
objects, a representation does not actually represent any extramental entity but is simply the
intentional object of a cognition. Thus, the question of the ontological status of representations
concerns not the represented content (like a desk or a dragon) but rather the representation
itself as a feature of the mind.18 Representational realism is commonly associated with a second
claim which says that cognition is necessarily and intrinsically endowed with representations
468
Śāntaraks∙ita
(sākāra), while representational antirealism is commonly associated with the claim that cogni-
tion is not necessarily endowed with representations, and that invariably veridical enlightened
cognition lacks representations (nirākāra).19
In addressing the first view, on which cognition is intrinsically and necessarily endowed
with real representations, Śāntarakṣita takes it that, according to this theory, a cognition and
its representation are non-distinct, constituting a single subject. Supposing that representa-
tions and cognition share a strict identity relation, his reasoning here turns on a version of the
law of noncontradiction according to which contradictory properties cannot be predicated of
the same subject. Śāntarakṣita observes that a moment of cognition seems to be indisputably
simple, and yet the content of cognition looks obviously complex; in any given moment, ordi-
nary experience presents us with a multiplicity of data – a white patch here, a blue patch there,
and so on. And, indeed, perhaps the most intuitive representational realist view (the so-called
citrādvaita, or “variegated nonduality” theory) says that unitary cognition is non-distinct from
its multifaceted representation. But, Śāntarakṣita argues, given the law of noncontradiction, if
cognition and representations are non-distinct, it follows that:
i. since a representation is manifold, so too is cognition, or else

ii. since cognition is truly unitary, so too is its representation.
Śāntarakṣita lays out this argument from dilemma as follows:
It is difficult to deny that: (i) due to being non-distinct from real representations, cog-
nition must accord with the nature of representations and therefore be manifold. Or
else, (ii) due to being non-distinct from unitary cognition, representations would have
to be unitary in accord with the nature of cognition. On account of having contradic-
tory properties, ultimately, representations and cognition would have to be distinct.
(MAV ad MA 46)
Śāntarakṣita takes up the second horn of the dilemma first, arguing that if a representation were
truly unitary in accord with cognition, then absurd consequences would follow. For instance,
since a simple representation could not be analytically divisible into phenomenal proper parts, we
would be unable to conceptually isolate different aspects of our phenomenal field (like the right
side and left side of this page). Furthermore, in the absence of phenomenal parts, it could never be
the case that one element of our experience was in motion while another was at rest (MA 47–48).
On the other hand, Śāntarakṣita reasons, the alternative that cognition is manifold in accord
with its complex representation is susceptible to the same kind of argument that he leveled
against material atomism: just as an extended material object could not be constituted by unex-
tended material simples, a phenomenally extended representation (like the one you may have
of this page) could not be composed of phenomenally unextended building blocks (MA 49).20
And given the metaphysical priority of unity to multiplicity, if there are no simple phenomenal
parts, neither can there be a true multiplicity of them. The parts of cognition, then, could not
exist in numerical parity with representational parts, since there can be no determinate number
of them to which cognition might correspond. He thus concludes that cognition and a real
representation could be neither truly one nor truly many.
Śāntarakṣita next turns to the representational antirealist view on which cognition is not
actually endowed with real representations, which only seem to appear to cognition due to an
error (MA 52). This view, Śāntarakṣita argues, is incapable of accounting for ordinary experi-
ence, for how could we perceive anything if no percept exists (MA 53–54)? Indeed, he insists
469
Allison Aitken
that there can be no cognition at all in the absence of an intentional object, since cognition is
intentional by its very nature; to cognize is to have a cognition of something (MA 55). Moreo-
ver, Śāntarakṣita argues, an unreal representation could stand in no relation whatsoever with
cognition, whether that be an identity relation or a causal relation (MA 57). If a representation
stood in an identity relation with cognition, then given the law of noncontradiction, either:
i. since cognition is real, the representation too would be real, or else

ii. since the representation is unreal, cognition too would be unreal (MAV ad MA 57).
Neither alternative is admissible for the representational antirealist. Furthermore, an unreal

representation is no more capable of standing in a causal relation with cognition than an iden-
tity relation, since if a representation were caused, then it would be real, but if it had no cause,
then there could be no explanation for its appearing with spatiotemporal determinacy or con-
sistency (MA 58). Unable to get the semblance of a defeasible account of cognition up and
running on this view, Śāntarakṣita does not even bother to apply the neither-one-nor-many
analysis to the representational antirealist position.
Having dismissed both the representational realist and antirealist views as untenable,
Śāntarakṣita takes himself to have demonstrated that there is no account in which a mental
entity could exist as a true unity or a true multitude, and thus concludes that nothing – whether
material or mental – lays claim to ontologically independent being. Instead, the only kind
of unity and being that exist are conventional and dependent. To flesh out precisely what
Śāntarakṣita means by this, let’s turn to the second of the two truths.
Conventional Truth and Yogācāra Ontology

Upon concluding his neither-one-nor-many argument for the ultimate truth, Śāntarakṣita pre-
sents his account of conventional truth to clarify that the rejection of ontological independence
does not entail an unmitigated nihilism. The term satya translated here as “truth” has a seman-
tic range that is also inclusive of “reality,” and Śāntarakṣita’s account of the conventional is
a theory of truth as well as an ontological theory. Yet Mādhyamikas, like Śāntarakṣita, affirm
only an ultimate truth and not an ultimate reality qua ontological status. In fact, the ultimate
truth as the emptiness of intrinsic nature might be interpreted as the claim that nothing is ulti-
mately real.21 There is thus only one ontological status: conventional reality.
According to Śāntarakṣita, whatever is conventionally real (i) has the capacity for causal effi-
cacy (arthakriyāśakti/arthakriyāsamartha), (ii) is dependently originated ( pratītyasamutpanna),
and (iii) satisfies our ordinary notions of unity and being only when not subjected to analysis into
its final nature (avicāraramaṇīya/avicāramanohara) (MA 64). Conventional truths, then, are
pragmatically efficacious claims that concern conventionally real things and which may be veri-
fied by our epistemic instruments of perception and inference. With this account, Śāntarakṣita
once again follows his predecessor Śrīgupta, who presents the earliest extant formulation of this
threefold criterion,22 which was subsequently adopted by Jñānagarbha, Kamalaśīla, Haribhadra,
the later Bhāviveka (c. eighth century), Atiśa (982–1054), and others.23 The first criterion for
conventional reality – having the capacity for causal efficacy – is a repurposing of Dharmakīrti’s
criterion for ultimately reality.24 Though an apparent subversion of Dharmakīrti’s intent, this
criterion represents yet another Dharmakīrtian influence on this branch of the Madhyamaka
tradition. We will return to the third criterion in treating the role of analysis in Śāntarakṣita’s
account of conventional truth below, but it is with the second criterion, being dependently origi-
nated, that Śāntarakṣita incorporates Yogācāra ontology into his system.
470
Śāntaraks∙ita
The claim that whatever is conventionally real comes into being in dependence on other
things goes back to Nāgārjuna.25 The ontological dependence relation implicated here is inclu-
sive of mereological dependence, mind-dependence, as well as causal dependence, such that
every conventionally real thing: (i) comes into being in dependence upon its parts and those
parts upon their own parts, ad indefinitum, (ii) is individuated as a conventional unity in
dependence upon mental designation, and (iii) is a product of causes and conditions, each
of which is in turn a product of its own causes and conditions, ad indefinitum. But, marking
a significant departure from prior Madhyamaka accounts, Śāntarakṣita identifies all things
involved in causal relations as mental in nature, thereby aligning his presentation of conven-
tional reality with Yogācāra ontology and its central commitment that everything consists in
cognition alone (vijñaptimātra):26
Whatever exists as cause and effect is, in fact, merely cognition ( jñāna). Whatever is
established by cognition itself exists in cognition.
(MA 91)
In order to understand how Śāntarakṣita understands Yogācāra ontology to map onto con-
ventional reality, it is necessary to first pin down what precisely he means by affirming the
Yogācāra commitment that everything is merely mental (cittamātra). Here are some possible
interpretations:
i. A phenomenological claim on which the only things relevant to our experience are mental
ii. A kind of skepticism which says that we cannot know whether or not there exist any extra-
mental entities
iii. An epistemological idealism which says that all objects of knowledge are determined by,
or dependent on, the mind and the structure of thought
iv. An immaterialism, or metaphysical idealism, on which there are no material things, and
the only kinds of things that exist are mental
The strongest claim, (iv) immaterialism, is an eliminative idealism insofar as it effectively

eliminates, or precludes the existence of, extramental things. The former three options are
varieties of non-eliminative idealism insofar as they grant some kind of primacy to the mental
but leave open the possibility that extramental things exist.
Some have argued for a version of the (i) phenomenological reading of Śāntarakṣita’s
Yogācāra by pointing out that he seems to identify conventionally real things with appear-
ances and thus exclusively with what lies within the domain of experience.27 Indeed, in a rather
customary Madhyamaka move following an argument for emptiness, Śāntarakṣita insists that
the conclusion of his neither-one-nor-many argument does not entail the denial of appear-
ances (MA 78ab). To do so would be tantamount to an implausible thoroughgoing nihilism.
But Śāntarakṣita’s nondenial of appearances should not be read as an anti-metaphysical, phe-
nomenological turn; nor does it serve to restrict of the scope of knowledge to the domain of
appearances, indicative of skepticism along the lines of view (ii). After all, he rejects as irra-
tional and untenable both direct realist and representationalist theories of perception, which
suppose mind-independent external objects to be the direct and indirect objects of perception,
respectively. And he does not simply argue that such objects are unknowable or irrelevant to
our experience. Rather, he insists that the existence of external objects founded in atoms is
incoherent and thus metaphysically impossible. And just as Śāntarakṣita’s arguments rejecting
a substantial self and a creator god, for example, are not intended to leave the back door open to
471
Allison Aitken
the existence of such entities lying beyond the scope of our experience, ordinary cognition, or
the reach of human knowledge, presumably the same is true of his rejection of material objects.
That Śāntarakṣita’s Yogācāra is best read not as (iii) an epistemological idealism (in which all
objects of knowledge are determined by, or dependent on, the mind) but as an eliminative (iv)
metaphysical idealism is supported by his commentary on MA 91, where he states in no uncertain
terms that external material objects do not exist and what we ordinarily take to be external objects
simply are one’s own mind. This agrees with the Compendium chapter on the “Examination of
External Objects,” wherein Śāntarakṣita assumes the Yogācāra perspective. Here, he lays out two
main lines of reasoning against external objects, the first an epistemological argument against the
possibility of having knowledge of the existence of external objects founded in atoms and the
second an argument against the metaphysical possibility of such objects.28 He transitions from
the epistemological to the metaphysical argument with reference to the discerning epistemic
agent ( prekṣāvat), once again gesturing to the prominent place of rational analysis in his system:
Be as it may that atoms are not established by any source of knowledge, there may
nonetheless be doubt. But how should a discerning person come to have certainty
about their nonexistence?
(TS 1988)
Śāntarakṣita aims to pull the rug out from under the external world by rejecting the existence
of its purported foundations: material simples.29 That Śāntarakṣita takes the Yogācāra mind-
only thesis to negate the existence (and not merely the epistemic accessibility) of real external
objects is reiterated in his Yogācāra formulation of the neither-one-nor-many argument:
Thus, it is appropriate for discerning individuals to ascertain that atoms are non-
existent, due to being empty of an intrinsic nature that is either one or many, like a
lotus in the sky.
(TS 1996)
According to Śāntarakṣita’s Yogācāra, atoms – the purported building blocks of material

objects – are no more real than lotuses growing in midair.
Śāntaraksita’s Madhyamaka-Yogācāra Synthesis: An Instrumentalist

Approach to the Ultimate
With Śāntarakṣita’s characterization of Yogācāra ontology in place, let us turn now to the
question of how precisely to understand his Yogācāra-Madhyamaka synthesis. That is, does
Śāntarakṣita propose a genuine and coherent unification of these two systems, or does he
have an instrumentalist story in mind? Given his identification of the conventional truth with
Yogācāra ontology, this question is tied up with how best to understand the relation between
the two truths, that is, between metaphysical idealism as the conventional truth and the univer-
sal negation of ontological independence as the ultimate truth. Are these two truths compatible
or contradictory?
One Madhyamaka story about the relation between the two truths sees them as contradic-
tory in as much as (i) conventional truths are true according to the non-veridical beliefs and
linguistic-conceptual norms of ordinary folks whose understanding is obscured by metaphysi-
cal ignorance, while (ii) the ultimate truth reflects the veridical cognition of an ideal epistemic
agent. But this is not Śāntarakṣita’s story. The radically unintuitive claim that all things are
472
Śāntaraks∙ita
merely mental in nature is plainly difficult to square with the commonly accepted view of
the world.30 The conventional truth of metaphysical idealism can hardly be “read off ” our
linguistic-conceptual norms. For Śāntarakṣita, the general consensus is not a guide to what is
conventionally true or real. Instead, he insists that a discerning person should seek a correct
understanding of conventional reality by employing analysis; rationality is king as a guide to
both the conventional and the ultimate.
Another Madhyamaka story about the two truths sees them as perfectly consistent and even
synonymous. On this view, the ultimate truth as the universal negation of ontological inde-
pendence is the obverse (viz. negative equivalent) of the conventional truth as the universal
affirmation of dependent origination. Yet Śāntarakṣita does not see the two truths as wholly
compatible either. This is evident from a comparative analysis of his Yogācāra and Madhy-
amaka versions of the neither-one-nor-many argument in the Compendium and Ornament. Not
only are the subjects of the two arguments different (the Yogācāra iteration takes up atoms and
the Madhyamaka argument concerns all things), but the predicate is also different, yielding
distinct inference warranting entailment relations (vyāpti):31
Yogācāra entailment relation: Whatever is neither-one-nor-many does not exist.

Madhyamaka entailment relation: Whatever is neither-one-nor-many lacks an intrin-
sic nature.
How can Śāntarakṣita consistently maintain that this same neither-one-nor-many reason estab-
lishes the non-existence of atoms in his Yogācāra iteration of the argument and the absence
of an intrinsic nature of all things in his Madhyamaka formulation? If, on the one hand, the
Yogācāra entailment relation holds, then his Madhyamaka neither-one-nor-many argument –
which applies this same reason to an unrestricted domain – commits him to a thoroughgo-
ing nihilism. If, on the other hand, the Yogācāra entailment relation does not obtain, then
Śāntarakṣita advances a fallacious argument in support of his account of conventional truth.
Śāntarakṣita must surely reject the first alternative; nihilism is not an option. He thus looks
committed to the second horn of the dilemma, that his argument in support of his view of
conventional truth is fallacious. Yet this dilemma stands only if Śāntarakṣita intends to hold
the Yogācāra and Madhyamaka perspectives simultaneously.32 But, as he points out, once one
has realized the Madhyamaka ultimate truth that all things are equally devoid of ontologi-
cal independence, one necessarily relinquishes the Yogācāra ontology that grants a privileged
position to the mental:
Those whose intellectual capacity is not slight and particularly those who are highly
industrious will, upon analyzing whether the mind has a unitary or manifold nature,
ultimately perceive no such entity. Thus, in reality the mind-only view is not accepted.
(MAV ad MA 92)
But why bother reasoning our way to metaphysical idealism if it is not ultimately true? Accord-
ing to Śāntarakṣita’s sliding scale of analysis, there are better and worse conventional truths,
with more rational stories supplanting the less rational. But there is no definitive convention-
ally true story. In setting up his characterization of cause and effect as mental in nature in MA
91, Śāntarakṣita states:
Whoever accepts the conventional reality of those things that stand in causal rela-
tions should analyze what those accepted conventional things are in order to respond
473
Allison Aitken
to fallacious arguments: Are they merely in the nature of the mind and mental con-
stituents or is their nature also extramental?
He goes on to explain that among Mādhyamikas, there are two opinions on the matter. The
first camp, exemplified by Bhāviveka, accepts conventionally real material and mental entities
alike. He introduces the second view – that cause and effect are merely mental – as simply “the
opinion of others.” Significantly, he does not reject the first opinion. He does, however, explain
the pedagogical utility of the latter. Śāntarakṣita thus sets out an instrumentalist account of
the relation between the two truths on which Yogācāra metaphysical idealism is provisionally
accepted as conventionally true as an expedient means to arrive at an understanding of the
Madhyamaka ultimate truth.
Śāntarakṣita’s conventional truth is not determinate but contextual, as indicated by the third
criterion for conventional truth that it does not withstand analysis. Importantly, Śāntarakṣita
does not follow Candrakīrti (Introduction to the Middle Way, Madhyamakāvatāra 6.35) or
Jñānagarbha (SDV 21) in articulating this criterion as the normative claim that one ought
not to analyze conventional truths. This would be incompatible with Śāntarakṣita’s progres-
sive path, which the discerning person traverses precisely by analyzing conventional truths.
Instead, following Śrīgupta, Śāntarakṣita simply claims that conventional truths satisfy when
not analyzed; a conventionally real thing satisfies our notions of reality, independence, and
unity when its ultimate nature is not subjected to analysis. We may arrive at progressively
more rational conventional truths through analysis, but there is no final conventionally true
theory; metaphysical inquiry into the nature of things has no termination point.
Still, if there is no determinate conventional truth, what makes one conventionally true theory
more rational than another? Śāntarakṣita implies that since nothing withstands analysis, a theory
that posits fewer ontologically independent entities is more rational than one that posits more inso-
far as it is closer to the ultimate truth. In both his Compendium and Ornament, Śāntarakṣita uses
analysis to gradually eliminate ontological categories, with the Yogācāra sparse ontology of “mind-
only” being, as it were, the last man standing. But at the end of the day, this category too does not
withstand analysis. The primary utility of Yogācāra isn’t in what it affirms but in what it denies.
From an idealist, rather than a dualist view of conventional reality, it simply takes fewer
steps to arrive at an understanding of the ultimate truth, which undermines the fundamentality
of anything – whether material or mental. Śāntarakṣita introduces his Yogācāra-Madhyamaka
synthesis, stating:
Based on the perspective of the mind-only system (cittamātra), one should under-
stand that there are no real external objects. Based on this Madhyamaka system, one
should understand that the mind too is utterly selfless.
(MA 92)
Those who hold the reins of rationality while riding the chariot of these two systems
will thereby achieve the state of a genuine proponent of the Mahāyāna.
(MA 93)
Śāntarakṣita recommends approaching the ultimate truth via the Yogācāra view, but it is just
that – an approach to the ultimate, not a definitive or unrevisable claim about the final nature
of things. Thus, on this picture, one cannot definitively claim that the two truths are either
compatible or contradictory. The conventional truth is a moving target on shifting sands. That
Yogācāra idealism is just one among other instructively efficacious conventional stories by
474
Śāntaraks∙ita
way of which one might arrive at the ultimate is supported by the following remark from
Kamalaśīla’s Illumination of the Middle Way:
Thus, one who is unable to instantaneously realize the fact that all things without
exception lack an intrinsic nature should, by temporarily relying on the mind-only
system, proceed in stages, beginning with understanding that external objects lack an
intrinsic nature.
(D 3887, 157a)
In this same instrumentalist spirit, Śāntarakṣita (MAV ad MA 70) cites Bhāviveka’s famed
metaphor of conventional truth as a ladder to the ultimate truth:
Without the ladder of conventional truth, it would not be possible for the learned to
ascend to the pinnacle of the palace of reality.
(Verses on the Heart of the Middle Way, Madhyamakahṛdayakārikā 3.12)
In Śāntarakṣita’s presentation of the two truths, the Yogācāra idealist ontology – though not the
exclusive, determinate, or preeminent conventional truth – is nonetheless a highly efficacious
penultimate steppingstone to understanding the Madhyamaka ultimate truth.
Conclusion
Śāntarakṣita remained a Madhyamaka authority to be reckoned with well into the final period
of Buddhism in India. His direct students, Kamalaśīla and Haribhadra, adopted and devel-
oped his Madhyamaka-Yogācāra synthesis, and he was recognized as a principal source on
the Madhyamaka neither-one-nor-many argument throughout the succeeding centuries by
authors such as Dharmamitra (fl. ca. 800), Jitāri (late tenth century), Bodhibhadra (fl. c. 1000),
Prajñākaramati (ca. 950–1030), Atiśa, Prajñāmokṣa (ca. eleventh century), Abhayākaragupta
(late eleventh–early twelfth century), and so on. Later prominent Yogācāra philosophers includ-
ing Ratnākaraśānti (ca. eleventh century) and Jñānaśrīmitra (ca. eleventh century) also cited
and responded to Śāntarakṣita’s neither-one-nor-many argument, indicating that they consid-
ered his attack on Yogācāra theories of the mind and mental content a serious enough threat to
necessitate critical engagement several hundred years later. In Tibet, Śāntarakṣita’s Ornament
together with Jñānagarbha’s Differentiation of the Two Truths and Kamalaśīla’s Illumination
of the Middle Way came to be known as the major works of the so-called “three Mādhyamikas
of the East” (dbu ma shar gsum), with commentaries composed on the Ornament by such
philosophically and temporally diverse luminaries as Chapa Chökyi Senggé (Phya pa Chos
kyi seng ge, 1109–1169), Tsongkhapa Lobsang Drakpa (1357–1419), and Jamgön Ju Mipham
Gyatso (1846–1912). Śāntarakṣita’s emphasis on the role of rationality and a progressive path,
as well as his synthesis of Madhyamaka with the Dignāga-Dharmakīrti tradition of logic and
epistemology, left a lasting and definite impact on Indian and Tibetan Buddhist philosophy.33
Notes
1 See Frauwallner 1961, 141–43.
2 See, for example, the Records of the Ba Clan (sba/dba’ bzhed) for one of the earliest sources (Wangdu
and Diemberger 2000). For a much later account from a compilation of sources, see Butön Rinchen
Drup’s (bu ston rin chen grub, 1290–1364) History of Buddhism (chos ’byung gsung rab rin po che’i
475
Allison Aitken
gter mdzod), translated in Obermiller (1932, 187–92), and Gö Lotsawa Zhönnu Pel’s (’gos lo tsā ba
gzhon nu dpal, 1392–1481) Blue Annals (deb gter sngon po), translated in Roerich (1949, 41–44). For
a recent compilation of sources on the life and work of Śāntarakṣita, see Eltschinger 2019.
3 Portraits of the members of this lineage were painted on the walls of Samyé; see, for instance, Ober-
miller 1932, 190; Roerich 1949, 34.
4 Tibetan doxographies commonly present Śāntarakṣita as an exemplar of the so-called *Yogācāra-
Svātantrika-Madhyamaka school of thought and Bhāviveka as the representative of *Sautrāntika-
Svātantrika-Madhyamaka, while Buddhapālita and Candrakīrti are standardly cited as paradigmatic
proponents of *Prāsaṅgika-Madhyamaka. “Yogācāra” in Śāntarakṣita’s doxographical designa-
tion indicates his synthesis of Yogācāra ontology into his account of conventional truth, while the
*Svātantrika label signifies (in part) his style of argumentation which, following Bhāviveka, deploys
independent inferential arguments (svatantrānumāna) to establish his theses rather than exclusively
utilizing reductio ad absurdum arguments ( prasaṅga) to undermine the theses of his opponents. On
the Tibetan doxographical assignments of Śāntarakṣita, see Seyfort Ruegg 1981, 87–100; Blumenthal
2004, 41–470, and for a critical analysis of this doxographical assignment, see McClintock 2003. For
discussions on the historical development of the *Svātantrika and *Prāsaṅgika categories, see Drey-
fus and McClintock 2003 and Seyfort Ruegg 2006.
5 Śāntarakṣita’s authorship of this work has been called into question by Tsongkhapa Lobsang Drakpa
(Tsong kha pa Blo bzang grags pa, 1357–1419) and Tāranātha (1575–1634) owing to the fact that
Kamalaśīla appears to argue against the author of the SDVP in the TSP; see Eckel 1987, 27–31; Blu-
menthal 2004, 29; Seyfort Ruegg 1981, 68 n. 224.
6 The attribution of Tattvasiddhi to Śāntarakṣita has been called into question by Steinkellner (1999, 356–57).
7 In addition to the texts already mentioned, the other texts attributed to Śāntarakṣita in the Tengyur
are: two short praises, the Praise of the Eight Tathāgatas (*Aṣṭatathāgatastotra) and Praise of the
Bhagavan: Song of Śrī Vajradhara (*Śrīvajradharasaṃgītibhagavatstotra) together with auto-
commentary (ṭīkā); two tantric works, Five Great Instructions on Kurukulla Arisen from Hevajra
(Hevajrodbhavakurukullāyāḥ Pañcamahopadeśa) and Ritual for the Extensive Recitation of the Pre-
vious Aspirations of the Seven Tathāgatas Collected from Sūtras (Saptatathāgatapūrvapraṇidhāna-
viśeṣavistārakalpavacanavidhi-sūtrāntasaṃkṣepa); and a commentary (vṛtti) on Candragomin’s
(seventh century) Twenty Verses on the Bodhisattva Vows (Bodhisattvasaṃvaravṃśaka). In his Com-
mentary on the Ornament, Śāntarakṣita appears to allude to another text he authored titled Ascertain-
ment of the Ultimate (*Paramārthaviniścaya) (Ichigō 1985, 330). As Eltschinger (2019, 384–85)
points out, Śāntarakṣita also seems to allude to this work in TS 2083.
8 See McClintock 2010, 58–62, 2013; Tillemans 2011, 153–54, 2016, 143–44 on Kamalaśīla’s elabora-
tion on this concept, particularly in the context of discerning the correct understanding of conven-
tional truth; on the term, prekṣāvat, see also Eltschinger 2007, 137–50, 2014, 195 n. 17, 219–34.
9 On Śrīgupta’s Introduction to Reality (Tattvāvatāra = TA) and accompanying autocommentary (-vṛtti
= TAV), see Aitken (Forthcoming).
10 For a short synopsis of the Compendium, see Seyfort Ruegg 1981, 89–90, and for a compilation of sec-
ondary scholarship on this work, see Steinkellner and Much 1995, 56–63; Eltschinger 2019, 385–86.
11 Śāntarakṣita presents a summary of the topics of the text in the first six stanzas of the Compendium,
with the main points of the first twenty-three chapters glossed as modifying dependent origination
and those of the final three chapters modifying the Buddha who taught dependent origination. On the
two-part structure of the Compendium, see McClintock 2010, 97–98.
12 On the application of this framework for describing Śāntarakṣita’s method, see McClintock 2003,
2010, 85–91; on this same concept applied to the work of Dharmakīrti, see Dreyfus 1997 and Dunne
2004, 53–79. Blumenthal (2004, 43, 44, 46) articulates this same approach of Śāntarakṣita’s in terms
of a “graded ascent of philosophical views,” “multiple levels or stages of provisionality,” and an
“ascent through provisional views.”
13 McClintock (2003, 68–76) points to TS 1916–17 as instances where Śāntarakṣita acknowledges the superi-
ority of Madhyamaka. Śāntarakṣita additionally nods to the Madhyamaka tradition in the framing of the text,
mirroring Nāgārjuna’s opening to his Fundamental Verses on the Middle Way (Mūlamadhyamakakārikā)
in laying out the purpose and structure of the Compendium in its opening stanzas.
14 (i) Vaiśeṣikas defend the conjoined view, (ii) Vaibhāṣikas such as Saṅghabhadra (c. fourth–fifth
century) defend the interceding space view, and (iii) Vasubandhu defends the spatially continuous
view in the Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya) ad 1.43d2, where
476
Śāntaraks∙ita
he attributes the position to the Kaśmīri Vaibhāṣikas such as Vasumitra (c. second century CE);
Śubhagupta (c. 720–780) also defends this view in his Proof of External Objects (Bāhyārthasiddhi)
50, 52, 53, 56, although Śāntarakṣita cites Śubhagupta in connection to the second view.
15 In his autocommentary, Śāntarakṣita goes on to praise Yogācāra for its theory being confirmed and known
by a source of knowledge ( pramāṇa), for serving as a corrective for mistaken views, for its rejection of
material simples, and for its agreement with scriptural sources (MAV ad MA 45; Ichigō 1985, 124).
16 MA 46–51; see Dharmakīrti’s Explanation of the Sources of Knowledge (Pramāṇavārttika) 3.209–22
and Śrīgupta’s TA 4.
17 MA 52–60; see TA 5–6, TAV transitional stanzas (antaraśloka) 1–4.
18 In Cartesian terms, the question here concerns the formal reality of thoughts themselves rather than
the objective reality of whatever might be represented in thought.
19 *Satyākāravāda (rnam pa bden par smra ba) and *Alīkākāravāda (rnam pa brdzun par smra ba) (lit-
erally “theory/proponent of real representations” and “theory/proponent of unreal representations”)
are not attested in extant Indic doxographies, where we instead find the Sākāravāda-Nirākāravāda
distinction. While these labels were imposed onto diverse sets of thinkers in contriving subschools
of Yogācāra, they are nevertheless useful for clarifying the structure of Śāntarakṣita’s argument
and the dialectical lay of the land as he understood it. One should be careful to distinguish the use
of the terms sākāra vs. nirākāra in the Yogācāra context from the use of this same pair of terms to
designate, respectively, representationalist vs. direct realist accounts of ordinary perception among
realists about external objects.
20 To the contrary, Berkeley (Principles in Works vol. 2, 98) and Hume (Treatise 1.2.4), for instance,
both argue for theories of a minima sensibilia, a kind of phenomenal atomism on which a perception
is reducible to indivisible, unextended simples.
21 This account can be traced back to Nāgārjuna, who argues that nothing lays claim to the Abhidharma
ontological category of ultimate reality ( paramārthasat), or substantially reality (dravyasat).
22 TA 11; see Aitken (Forthcoming) and (2021a) for an interpretation of Śrīgupta’s version of this three-
fold criterion. For a comparison of Śrīgupta and Śāntarakṣita on the two truths, see Aitken (2021b).
23 See, for example, Kamalaśīla’s MAP ad MA 64, Haribhadra’s Illuminating the Ornament of Realiza-
tion (Abhisamayālaṃkārālokā) (Wogihara 1932–35, 594.18–25), Bhāviveka’s Compendium on the
Meaning of the Middle Way (Madhyamakārthasaṃgraha) 9–11 and Jeweled Lamp for the Middle
Way (Madhyamakaratnapradīpa) 1.4, and Atiśa’s Introduction to the Two Truths (Satyadvayāvatāra)
4. Jñānagarbha also sets out versions of these three criteria (SDV 8, 12, and 21).
24 Explanation of the Sources of Knowledge 3.3ab and Essence of Reasoning (Nyāyabindu) 1.15.
25 Śāntarakṣita cites Nāgārjuna’s Fundamental Verses on the Middle Way 24.18 in support of this
criterion.
26 Importantly, Śāntarakṣita does not regard this as an innovation, citing Nāgārjuna’s Sixty Stanzas on
Reasoning (Yuktiṣaṣṭika) 21 and 34 (among other sources) in support of this presentation.
27 For example, Garfield (2016) argues that Jamgön Ju Mipham Gyatso (’Jam mgon ’Ju Mi pham rgya
mtsho, 1846–1912) reads Śāntarakṣita’s Yogācāra phenomenologically.
28 See Saccone’s 2018 edition, translation, and analysis of this chapter. It is this second, metaphysical
argument that follows Vasubandhu’s previously discussed anti-atomist argument from his Twenty
Verses 11–15.
29 It is worth pointing out that Śāntarakṣita’s rejection of atoms does not in principle preclude the pos-
sibility of conventionally real external objects that are not founded in atoms, though he does not
explicitly consider such a scenario, presumably owing to the fact that all external world realists in his
intellectual milieu were atomists.
30 To the contrary, Berkeley famously makes a case for subjective idealism as part of his project to
restore commonsense.
31 This difference is also reflected in the different examples cited in the two inferences (in which the
entailment relation is instantiated), with the Madhyamaka example being a reflection, which lacks an
intrinsic nature, and the Yogācāra example being a lotus growing in the sky, which is a nonexistent
thing.
32 Since atoms are stipulated by their proponents as partless, and therefore true unities by definition (TS
1992ab), the rejection of the unity of atoms would necessarily preclude their existence. The problem
here lies in the sweeping nihilistic implications of the generalized Yogācāra entailment relation when
applied to the Madhyamaka all-inclusive subject.
477
Allison Aitken
33 Although Śāntarakṣita’s conciliatory approach to Yogācāra was rejected by many later Tibetan
Mādhyamikas in favor of Candrakīrti’s account of conventional truth, his synthesis of Madhyamaka
with Dignāga and Dharmakīrti’s logic and epistemology remains influential in Tibetan Madhyamaka
cutting across traditions.
Bibliography
Aitken, Allison. Forthcoming. Introduction to Reality: Śrīgupta’s Tattvāvatāravṛtti. Harvard Oriental
Series. Cambridge, MA: Harvard University Press.
———. 2021a. “No Unity, No Problem: Madhyamaka Metaphysical Indefinitism.” Philosophers’
Imprint 21 (31): 1–24.
———. 2021b. “The Truth About Śrīgupta’s Two Truths: Longchen Rabjampa’s ‘Lower Svātantrikas’ and
the Making of a New Philosophical School.” Journal of South Asian Intellectual History 3 (2): 185–225.
Blumenthal, James. 2004. The Ornament of the Middle Way: A Study of the Madhyamaka Thought of
Śāntarakṣita. Ithaca: Snow Lion Publications.
———. 2009a. “Dynamic and Syncretic Dimensions to Śāntarakṣita’s Presentation of the Two Truths.”
Asian Philosophy 19 (1): 51–62.
———. 2009b. “The ‘Neither-One-Nor-Many’ Argument of Śāntarakṣita: A Classical Buddhist Argu-
ment on the Ontological Status of Phenomena.” In Buddhist Philosophy: Essential Writings, edited by
William Edelglass and Jay L. Garfield, 46–60. New York: Oxford University Press.
———. 2015. “Two Topics Concerning Consciousness in Śāntarakṣita’s Yogācāra-Madhyamaka Syncre-
tism.” In Madhyamaka and Yogācāra: Allies or Rivals?, edited by Jay L. Garfield and Jan Westerhoff,
242–51. New York: Oxford University Press.
Dreyfus, Georges. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpretations.
Albany: State University of New York Press.
Dreyfus, Georges, and Sara McClintock. 2003. The Svātantrika-Prāsaṅgika Distinction: What Differ-
ence Does a Difference Make? Boston, MA: Wisdom Publications.
Eckel, Malcolm David. 1987. Jñānagarbha on the Two Truths: An Eighth Century Handbook of Madhy-
amaka Philosophy. Albany, NY: State University of New York Press.
Eltschinger, Vincent. 2007. “On 7th and 8th Century Buddhist Accounts of Human Action, Practical
Rationality and Soteriology.” In Pramāṇakīrtiḥ: Papers Dedicated to Ernst Steinkellner on the Occa-
sion of His 70th Birthday, Part 1, edited by Birgit Kellner, Horst Lasic, M. T. Much, and Helmut
Tauscher, 135–62. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
———. 2019. “Śāntarakṣita.” In Brill’s Encyclopedia of Buddhism, Volume II: Lives, edited by Jonathan
A. Silk, Richard Bowring, Vincent Eltschinger, and Michael Radich, 383–90. Leiden and Boston: Brill.
Frauwallner, Erich. 1961. “Landmarks in the History of Indian Logic.” Wiener Zeitschrift für die Kunde
Süd- und Ostasiens 5: 125–48.
Garfield, Jay L. 2016. “Cittamātra as Conventional Truth from Śāntarakṣita to Mipham.” Journal of Bud-
dhist Philosophy 2: 263–80.
Ichigō, Masamichi, ed. 1985. Madhyamakālaṃkāra of Śāntarakṣita with His Own Commentary or Vṛtti
and with the Subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido.
———. 1989. “Śāntarakṣita’s Madhyamakālaṃkāra.” In Studies in the Literature of the Great Vehi-
cle: Three Mahāyāna Buddhist Texts, edited by Luis O. Goméz and Jonathan A. Silk, 141–240. Ann
Arbor: Center for South and Southeast Asian Studies, University of Michigan.
———. 2016. “Śāntarakṣita on Satyadvaya.” Journal of Indian and Tibetan Studies 20: 166–94.
Jha, Ganganatha. 1937–1939. The Tattvasaṅgraha of Śāntarakṣita with the Commentary of Kamalaśīla.
2 vols. Baroda: Oriental Institute.
Kapstein, Matthew. 2001. Reason’s Traces: Identity and Interpretation in Indian and Tibetan Buddhist
Thought. Boston: Wisdom Publications.
———. 2009. “Śāntarakṣita’s Tattvasaṃgraha. A Buddhist Critique of the Nyāya View of the Self.” In
Buddhist Philosophy. Essential Readings, edited by William Edelglass and Jay L. Garfield, 320–33.
Keira, Ryusei. 2004. Mādhyamika and Epistemology: A Study of Kamalaśīla’s Method for Proving the
Voidness of All Dharmas: Introduction, Annotated Translations and Tibetan Texts of Selected Sec-
tions of the Second Chapter of the Madhyamakāloka. Wiener Studien zur Tibetologie und Buddhis-
muskunde, Heft 59. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
478
Śāntaraks∙ita
McClintock, Sara L. 2003. “The Role of the ‘Given’ in the Classification of Śāntarakṣita and Kamalaśīla
as Svātantrika-Mādhyamikas.” In The Svātantrika-Prāsaṅgika Distinction: What Difference Does a
Difference Make? edited by Georges B. J. Dreyfus and Sara L. McClintock, 125–71. Boston: Wisdom
Publications.
———. 2010. Omniscience and the Rhetoric of Reason: Śāntarakṣita and Kamalaśīla on Rationality,
Argumentation, and Religious Authority. Boston: Wisdom Publications.
———. 2013. “Kamalaśīla and Śāntarakṣita on Scripture and Reason – The Limits and Extent of ‘Practical
Rationality’ in the Tattvasaṃgraha and Pañjikā.” In Scriptural Authority, Reason, and Action: Proceedings
of a Panel at the 14th World Sanskrit Conference, Kyoto, September 1st-5th 2009, edited by Vincent Eltsch-
inger and Helmut Krasser, 209–38. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Obermiller, Eugene. 1932. The History of Buddhism in India and Tibet by Bu Ston: Part II. Heidelberg:
Harrassowitz.
Roerich, George. 1949. The Blue Annals, Part I. Calcutta: Royal Asiatic Society of Bengal.
Saccone, Margherita Serena. 2018. On the Nature of Things: A Buddhist Debate on Cognitions and Their
Object. Vienna: Arbeitskreis für tibetische und buddhistische Studien Universität Wien.
Sāṅkṛtyāyana, Rāhula. 1935/1936. “Dharmakīrti’s Vādanyāya. With the Commentary of Śāntarakṣita.”
Journal of the Bihar and Orissa Research Society: 21–22.
———. 2006. “The Svātantrika-Prāsaṅgika Distinction in the History of Madhyamaka Thought.” Indo-
Iranian Journal 49 (3/4): 319–46.
Sferra, Francesco. 2004. “Śāntarakṣita: Vātsīputrīyaparikalpitātmaparīkṣa, Disamina del se immaginato
dai seguaci di Vātsīputra (Tattvasaṃgraha, 336–49 con il commento [Pañjikā] di Kamalaśīla).” In
La Rivelazione del Buddha, vol. II, edited by Raniero Gnoli, 1155–74. Milan: Arnoldo Mondadori.
Shastri, Dwarka Das. 1968. Tattvasaṅgraha of Ācārya Shāntarakṣita with the Commentary Pañjikā of
Shri Kamalashīla. 2 vols. Varanasi: Bauddha Bharati.
Steinkellner, Ernst. 1999. “Yogic Cognition, Tantric Goal, and Other Methodological Applications
of Dharmakīrti’s kāryānumāna Theorem.” In Dharmakīrti’s Thought and Its Impact on Indian and
Tibetan Philosophy, Proceedings of the Third International Dharmakīrti Conference, Hiroshima,
November 4–6, 1997, edited by Shoryu Katsura, 349–62. Vienna: Austrian Academy of Sciences Press.
2001. “Is the Ultimate Cognition of the Yogin Conceptual or Non-Conceptual? Part I: A Critical Edition of
the Tantric Tattvasiddhi, Final Section.” In Le Parole e i marmi: Studi in onore di Raniero Gnoli nel suo
70° compleanno, edited by Raffaele Torella, 835–52. Rome: Istituto Italiano per l’Africa e l’Oriente.
———. 2014. The Edition of Śāntarakṣita’s Vādanyāyaṭīkā Collated with the Kundeling Manuscript.
Vienna: Arbeitskreis für tibetische und buddhistische Studien Universität Wien.
Steinkellner, Ernst, and M. T. Much. 1995. Texte der erkennt-nistheoretischen Schule des Buddhismus: sys-
tematische Übersicht über die buddhistische Sanskrit-Literatur II. Göttingen: Vandenhoeck and Ruprecht.
Tillemans, Tom T. F. 1982. “The ‘Neither One Nor Many’ Argument for śūnyatā and Its Tibetan Interpre-
tations: Background Information and Source Materials.” Études de Lettres (University of Lausanne)
3: 103–28.
———. 1983. “The ‘Neither One Nor Many’ Argument for śūnyatā and Its Tibetan Interpretations.” In
Contributions on Tibetan and Buddhist Religion and Philosophy, edited by Ernst Steinkellner and
Helmut Tauscher, 305–20. Vienna: Arbeitskreis für Tibetische and Buddhistische Studien.
———. 1984. “Two Tibetan Texts on the ‘Neither One Nor Many’ Argument for Śūnyatā.” Journal of
Indian Philosophy 12: 357–88.
———. 2011. “How Far Can a Mādhyamika Reform Conventional Truth? Dismal Relativism, Fiction-
alism, Easy-Easy Truth, and the Alternatives.” In Moonshadows: Conventional Truth in Buddhist
Philosophy, edited by the Cowherds, 151–66. New York: Oxford University Press.
———. 2016. How do Mādhyamikas Think? And Other Essays on the Buddhist Philosophy of the Mid-
dle. Somerville, MA: Wisdom Publications.
Wangdu, Pasang, and Hildegard Diemberger. 2000. dBa’ bzhed: The Royal Narrative concerning the
bringing of the Buddha’s Doctrine to Tibet. Vienna: Austrian Academy of Sciences Press.
Westerhoff, Jan. 2015. “Nāgārjuna’s Yogācāra.” In Madhyamaka and Yogācāra: Allies or Rivals? edited
by Jay L. Garfield and Jan Westerhoff, 165–83. New York: Oxford University Press.
Wogihara, Unrai, ed. 1932–1935. Abhisamayālaṃkārāloka Prajñāpāramitāvyākhyā (Commentary on
Aṣṭasāhasrikā-Prajñāpāramitā) by Haribhadra Together with the Text Commented On. Tokyo: The
Toyo Bunko.
479
29
KAMALAŚĪLA
Mādhyamika Champion of Magical Reason
Sara McClintock
Kamalaśīla’s Tantalizing Traces

Unlike the majority of figures in this volume, the monastic scholar Kamalaśīla (c. 740–795)
leaves tantalizing historical traces of his embodied existence beyond the many treatises cred-
ited to his name and the usual hagiographies. These traces mainly take the form of later Tibetan
historiographical writings documenting what has come to be known as either the Great Debate
at Samyé or the Council of Lhasa (c. 792–794). This legendary intellectual contest of unknown
duration features Kamalaśīla and Heshang Moheyan (Héshang Móhēyǎn), a Chinese rival
monk representing the Chan Buddhist tradition, both of whom are said to have been invited to
Tibet by the powerful king Trisong Detsen (Khri srong lde btsan, c. 742–797). The earliest and
most important of the Tibetan works documenting this event – the historicity of which remains
an open question – is the Testament of Ba (dBa’ bzhed), an account of the arrival of Buddhism
in Tibet which survives in diverse forms (Wangdu and Diemberger 2000) and upon which
later Tibetan historiographical writings appear to rely (Tucci 1958, 10). Other early sources
preserved at Dunhuang in both Tibetan and Chinese likewise hint at Kamalaśīla’s existence
without, however, naming him – though one text, at least, the Ratification of the True Principle
of the Mahāyāna Teaching of Sudden Awakening (Dunwu dacheng zhengli jue) of Wang Xi,
refers to the Indian opponent of Moheyan as a “brahman monk” (Demiéville 1952; Seyfort
Ruegg 2013), a moniker that fits what we learn of Kamalaśīla from hagiographers. Thus,
while much remains uncertain, all sources point to a conflict between Indian proponents of a
so-called gradual path and Chinese advocates for a practice allowing for a sudden or “subitist”
breakthrough to enlightenment. Not surprisingly, there is no agreement about who actually
won the debate, assuming that it actually took place or in any way resembled the traditional
accounts. While Chinese documents from Dunhuang identify the Chinese side as the winner
(Eltschinger and Marks 2019, 273), Tibetan Buddhist historiographers are unanimous in nam-
ing Kamalaśīla the victor, thus ensuring that Tibetan Buddhism would hence forward rely
primarily on Indian (and not Chinese) sources for the transmission of Buddhism to Tibet. With
this narrative as a backdrop, Kamalaśīla becomes something of a hero of the Tibetan Buddhist
philosophical tradition, especially in its monastic and scholastic dimensions.
While the facts may be murky, however, there is still quite a bit in the traditional accounts
that finds support in Kamalaśīla’s oeuvre. First, Kamalaśīla’s commitment to gradualism is a
480 DOI: 10.4324/9781351030908-42

Kamalaśīla
persistent theme in his writings, as we shall soon see. Second, Kamalaśīla’s faith and expertise
in reasoned debate and analysis as a means to resolve doubts is well attested in his numer-
ous philosophical treatises, especially his epistemological works but extending to his Madh
yamaka compositions as well. And third, his three interrelated tracts, Stages of Cultivation
(Bhāvanākrama), take frequent and direct aim at a form of subitism later interpreters identify
with the position of the Chinese monk Moheyan. Indeed, it is widely supposed that these
three texts on the methods and goals of meditation were composed in Tibet specifically in
preparation for the Great Debate, and there is speculation that this fact explains their lack of
renown among later Buddhists in India. Sadly, their absence from the scene in India, along
with the absence as well of Kamalaśīla’s most important Madhyamaka text, Light of the Mid-
dle (Madhyamakāloka), may also corroborate another part of Kamalaśīla’s legend: several
sources tell of Kamalaśīla’s demise a few years after the famous debate at the hands of assas-
sins who are said to have “squeezed” his kidneys, thus preventing the monk from returning
to India. Whether these murderers were sent by a jealous Moheyan or by non-Buddhist sup-
porters of Tibetan indigenous traditions opposed to the importation of Buddhism to Tibet is a
matter on which even the Tibetan sources do not agree (Keira 2004, 2).
What is clear, however, is that Kamalaśīla and his teacher Śāntarakṣita (c. 725–788), who
is widely credited with having established Samyé Monastery where he ordained the first seven
Buddhist monks in Tibet, have both emerged as cultural heroes for Tibetan Buddhists. Many,
like Tenzin Gyatso, the current Dalai Lama, see them as key representatives of what the he calls
“the Nālandā tradition” in reference to the great North Indian monastery and center of learning
where both Indian pandits are said to have taught and to have served as abbot. Our knowledge of
Kamalaśīla’s life in India and role at Nālandā, however, is even more sketchy than is our under-
standing of his activities in Tibet. He is clearly associated with Śāntarakṣita through his writings –
having composed two important commentaries on Śāntarakṣita’s chief works – and through the
historical legends maintaining that he was called to Tibet upon Śāntarakṣita’s deathbed advice
to the Tibetan king. He is also connected, though more remotely, with Śāntarakṣita’s own pur-
ported teacher, Jñānagarbha. All three are thought to have hailed from Bengal, with later Tibetans
sometimes speaking of them as a group using variants on the phrase the “three Svātantrika-
Mādhyamikas from the east” (dbu ma rang rgyud gyi shar gsum) (Tauscher 1999, 387).
We shall return to the problematic doxographical classification of Svātantrika-Mādhyamika
in what follows. For now, we note simply that all that remains beyond these hints about
Kamalaśīla’s historical personage are the surviving works attributed to him and the later recep-
tion of these works in Tibet. Ironically, despite Tenzin Gyatso’s lauding of the Nālandā tradition
and the foundational role of these Indian thinkers in the establishment of Buddhist monastic
scholasticism in Tibet, Śāntarakṣita and Kamalaśīla’s lineage appears not to have remained
strong in the Land of Snows, where the interruption of their oral lineage combined with the later
ascendancy of the so-called Prāsaṅgika-Madhyamaka of Candrakīrti starting in the eleventh
century seems to have lessened their influence. This may be particularly true for their epis-
temological writings, concerning which the fifteenth-century Sakya scholar Gorampa Sönam
Sengé (Go ram pa bSod nams seng ge, 1429–1489) said, “Although the abbot Śāntarakṣita
and the supreme scholar Kamalaśīla appear to have founded an oral tradition (bshad-sgrol),
nowadays, apart from the mere book, there does not appear a continuity of its teaching” (van
der Kuijp 1983, 4–5). This statement intriguingly suggests something more about how figures
like Śāntarakṣita and Kamalaśīla function as authors. That is, in addition to composing treatises,
their philosophical activity is understood to require an embodied dialogical dimension. Reading
the mere book ( yig cha tsam) without an authoritative guide empowered to transmit its contents
is unlikely to produce the insights toward which these philosophers would like us to progress.
481
Sara McClintock
We are, nevertheless, products of our own time, and this emboldens us to try to under-
stand these works even in the absence of an unbroken transmission lineage. A review of the
writings attributed to Kamalaśīla reveals him to be an extraordinary scholar, whose depth of
learning and fluency of expression yield a rare combination among philosophers: a thinker
who can communicate dense, elliptical arguments with a refreshing mix of clarity and preci-
sion. As an author, he has the further virtue of offering frequent, explicit signposts concerning
his purposes, presuppositions, and audience. These signposts serve as invaluable instructions
to readers while simultaneously reflecting his understanding of the philosophical enterprise:
by attending to the signposts, we understand that Kamalaśīla sees the philosophical project
as inherently embedded in social and linguistic contexts, thus rendering its value entirely
at the conventional (saṃvṛti) or transactional (vyāvahārika) level of reality. This is so even
though philosophy itself is also an important tool on the path to the realization of the ultimate
( pāramārthika) reality or the truth of the highest meaning ( paramārthasatya). We shall delve
into this topic in greater depth subsequently, as it is crucial to understanding Kamalaśīla’s
character as an author committed to a syncretic and gradualist approach.
Stages of Cultivation and Light of the Middle

First, however, to get a better picture of Kamalaśīla as an author, we need to give a fuller account
of his oeuvre. We have already encountered Kamalaśīla’s two most famous independent trea-
tises: the three interrelated texts known as Stages of Cultivation and the extensive late treatise
on Madhyamaka, Light of the Middle. As previously indicated, these works are widely consid-
ered to have been composed in Tibet (Keira 2004, 7), and this fact is understood to account for
the near total lack of their citation among Indian Buddhists. In the case of Light of the Middle,
we do find parallel passages in the twelfth-century Indian scholar Abhayākaragupta’s Orna-
ment of the Sage’s Thought (Munimatālaṃkāra), lending credence to the report that when Atiśa
Dīpaṃkaraśrījñāna (c. 982–1054) visited Samyé in the eleventh century, he found the Sanskrit
manuscript of Light of the Middle, copied it, and had it sent back to India (Keira 2004, 8). As for
Stages of Cultivation, there is some doubt about the first of the three texts in the series, since we
find numerous parallel passages in the contemporary Indian scholar Haribhadra’s seminal Light
for the Ornament of Realization (Abhisamayālaṃkārāloka); we likewise find parallels in the
Instructions on the Stages of Cultivation of the Perfection Vehicle (Pha rol tu phyin pa’i theg
pa’i sgom pa’i rim pa’i man ngag; *Pāramitāyānabhāvanākramopadeśa), a work preserved in
Tibetan attributed to the relatively unknown Indian author Jñānakīrti (c. 900 CE) and the San-
skrit title for which is unattested (Taniguchi 1992). So it is also possible that Kamalaśīla wrote
the work in India or that he had copies sent back, even as the text also appears to address some
of the core issues at stake in the legendary Great Debate. Such situations again raise the intrigu-
ing problem of the nature of authorship in this period, causing us to reflect that our attributions
of authors to particular texts are indeed conventional designations arising from myriad factors,
most if not all of which cannot be ascertained with any true certainty.
In any case, although both Stages of Cultivation and Light of the Middle are widely revered
in Tibet, neither received much direct commentary there. According to the fourteenth Dalai
Lama, Tenzin Gyatso, even though Tsongkhapa, the fifteenth-century progenitor of the
Gelugpa school of Tibetan Buddhism, quotes extensively from Stages of Cultivation in his
own Great Stages of the Path (lam rim chen mo), the transmission lineage for Kamalaśīla’s trio
of meditation texts was “fairly rare” in central and southwestern Tibet (Dalai Lama 2001, 23).
The Dalai Lama relates how he himself had to go to some effort to obtain the teachings, thus
reinforcing the idea that notwithstanding Kamalaśīla’s reputation among Tibetans, the study of
482
Kamalaśīla
his texts had been largely eclipsed. As for Light of the Middle, the only known commentaries
in Tibetan are a recently discovered treatise by Chapa Chökyi Sengé (Phya pa chos kyi seng
ge, 1109–1169) and a late partial treatment by the Mongolian scholar Ngawang Tendar (Ngag
dbang bstan dar, 1835–1915).1 Although Stages of Cultivation has received some sustained
attention (Adam 2002; Schmid 2020), perhaps in part due to the fact that two of the three texts
survive in Sanskrit (Tucci 1958, 1971), Light of the Middle is extant only in Tibetan and is still
relatively unknown. Ryusei Keira’s excellent Mādhyamika and Epistemology (2004), a study
and translation of Kamalaśīla’s arguments for emptiness in Light of the Middle, remains the
only book-length treatment of this work in English. Still, Keira treats only a relatively small
part of this long and masterful text. Philosophically, both Stages of Cultivation and Light of
the Middle are extraordinarily rich. Here we touch only on a few of the most salient contribu-
tions. Later, after we have reviewed Kamalaśīla’s oeuvre more thoroughly, we shall consider
the thorny question of Kamalaśīla’s doxographical affiliation. Here, however, we shall attempt
to encounter his philosophy on its own terms.
The texts making up Stages of Cultivation contain considerable overlap in their themes, and
it is not clear why Kamalaśīla composed three different works.2 For our purposes, we can iso-
late four distinct themes running through the three works that also characterize Kamalaśīla’s
philosophical positions more generally. These themes are (1) that the bodhisattva path is a
gradual affair with distinct stages (krama) involving the perfection and integration of method
(upāya) and wisdom ( prajñā); (2) that conceptual philosophical activity, including rational
analysis (vicāra) and examination ( pratyavekṣā), is integral to the bodhisattva path resulting
in buddhahood; (3) that the ultimate nature of reality is to be empty (śūnya) of any essential
nature (svabhāva), such that the final realization of that emptiness (śūnyatā) in the perfec-
tion of wisdom ( prajñāpāramitā) is necessarily nondual (advaya) and nonconceptual (nirvi-
kalpa) in nature; and (4) that the root of buddhahood is compassion, which when combined
with insight into reality (bhūtārtha) results in a state characterized as non-located nirvāṇa
(apratiṣṭhitanirvāṇa). This last state is that of the omniscient buddhas, located in neither
saṃsāra nor nirvāṇa and thus able to serve the needs of suffering beings by teaching them
while remaining undefiled by their unpurified ignorance.3
Kamalaśīla’s gradualism is on full display here, as he insists on the necessity of first
using rational analysis to verify the emptiness – known otherwise as the naturelessness
(niḥsvabhāvatā) – of all things before attaining direct nonconceptual realization of that real-
ity in yogic perception ( yogipratyakṣa).4 But how does one verify naturelessness? While
Kamalaśīla touches on this problem in the Stages of Cultivation, it is only in his specifi-
cally Madhyamaka works, especially Light of the Middle but also his commentary on his
teacher Śāntarakṣita’s well-known Ornament of the Middle (Madhyamakālaṃkāra), that he
provides a developed set of answers to this question. In these texts, Kamalaśīla sets out the
proof statements ( prayoga) that establish emptiness as the incontrovertible ultimate truth
( paramārthasatya) concerning all entities. Developing arguments inherited from the larger
Madhyamaka tradition and expressed in the language of the Buddhist epistemological tradi-
tion, he details in Light of the Middle five key proof statements for naturelessness.5 The five
proof statements – each of which is referenced by way of its reason (hetu) establishing empti-
ness – are as follows: (1) the “diamond splinter” (vajrakaṇa) reason, which refutes that a thing
is produced either by itself, by another, by both itself and another, or without a cause; (2) the
reason that refutes the production of either existent or nonexistent things; (3) the reason that
refutes that one can produce many, many can produce one, many can produce many, or one
can produce one; (4) the reason by virtue of the “dependent arising” ( pratītyasamutpāda) of all
things; and (5) the reason of having “neither one nor many” (ekānekaviyoga) essential nature.
483
Sara McClintock
Light of the Middle is structured around two basic divisions: an earlier segment consisting
of a series of 83 so-called prior positions ( pūrvapakṣa) raising objections to Kamalaśīla’s
Madhyamaka system and a later segment consisting of the so-called subsequent positions
(uttarapakṣa) rebutting these objections. The second segment is traditionally understood to be
organized into four chapters: a chapter on the proof of emptiness via scripture, a chapter on the
proof of emptiness by reasoning, a chapter on the two truths, and a chapter establishing that
there is ultimately only one single vehicle (ekayāna) or path to awakening. Not surprisingly,
Kamalaśīla addresses the objections with extreme rigor, harnessing the full arsenal of argu-
mentation techniques from the Buddhist epistemological tradition to beat back his Buddhist
opponents, some of whom have been identified as generic Vaibhāṣikas and Yogācāras, others
tentatively as Dharmakīrti’s early commentators Devendrabuddhi and Śākyabuddhi, one as
Dharmapāla, and one even as Dharmakīrti himself.6
Noteworthy here and elsewhere is Kamalaśīla’s embrace of flagship theories and positions
associated with the Yogācāra stream of Buddhist thought in India. For example, in Light of
the Middle, we see Kamalaśīla accept the Yogācāra theory of the three natures (trisvabhāva)
and their attendant three kinds of naturelessness, but we see him doing so in such a way as to
interpret all these in light of the Madhyamaka theory of the two truths (satyadvaya). As is well
known, the three natures are a Yogācāra scheme to explain our experience of reality consist-
ing of the imagined nature ( parikalpitasvabhāva), the dependent nature ( paratantrasvabhāva),
and the perfected nature ( pariniṣpannasvabhāva). Each of these natures is also said to have
its own distinct kind of naturelessness: character naturelessness (lakṣaṇa-niḥsvabhāvatā),
production naturelessness ( utpatti-niḥsvabhāvatā), and ultimate naturelessness ( paramārtha-
niḥsvabhāvatā). Kamalaśīla accepts all these categories explicitly. Yet his final position differs
from that of the typical Yogācāra thinker in that he wants to emphasize the ultimate emptiness of
all three natures and not just the lack of identity in the case of the imagined nature and the lack of
production in the case of the dependent nature.7 To take just one example, Kamalaśīla states that
“the dependent nature is an entity that is accepted so long as it is not analyzed, that is just as it
appears, and that is dependently arisen, like an illusion.”8 By this statement, Kamalaśīla alludes
to key characteristics of the conventional advanced by his direct predecessors in Madhyamaka
thinking, Jñānagarbha and Śāntarakṣita. He shows that in his view, the Yogācāra category of the
dependent nature can be accepted, but that status is nevertheless thoroughly conventional even
if it lacks conceptuality. The conventional nature of appearances, even without the presence of
language and concepts, is a hallmark of Jñānagarbha and Śāntarakṣita’s Madhyamaka thought.9
The Extensive Commentary on Śāntaraksita’s Compendium of Reality

Beyond Stages of Cultivation and Light of the Middle, Kamalaśīla is best known for his mas-
sive Extensive Commentary (Pañjikā) on his teacher Śāntarakṣita’s Compendium of Realities
(Tattvasaṃgraha). Unlike the previous two works, Kamalaśīla most certainly composed this
one in India, before he went to Tibet. The Extensive Commentary, along with the Compen-
dium, which are rarely, if ever, read apart from one another, are staggering in their compre-
hensiveness, with Śāntarakṣita’s root text coming in at more than 3,600 verses in length and
Kamalaśīla’s prose commentary at over 1,000 pages in modern printed editions. Both works are
famed not only for their sophisticated argumentation utilizing the epistemological and logical
theories of Vasubandhu and Dharmakīrti but also for their extensive engagement with and cita-
tion of a wide range of Buddhist and non-Buddhist authors representing a plethora of contem-
porary Indian philosophical positions. Since both Śāntarakṣita and Kamalaśīla quote opponents
directly and often quite faithfully, the works have become valuable resources for the study not
484
Kamalaśīla
only of Indian Buddhism but also of non-Buddhist Indian philosophical traditions whose own
root texts have been lost. Kamalaśīla’s unusual practice of directly naming opponents adds
extra value. We are extremely fortunate that two medieval manuscripts were preserved in Jain
libraries. While the two complete modern editions (Krishnamacharya 1926; Shastri 1968) and
the single complete English translation (Jha 1937/1939) all leave significant room for improve-
ment, their publication has been an enormous boon to the field of Indian philosophical studies.10
Some controversy has emerged over how best to read the texts, with a key question emerg-
ing regarding how to imagine the original audience for the works. The Compendium of Reali-
ties itself is divided into 26 chapters or “investigations” ( parīkṣā) covering major topics in
Indian metaphysics, epistemology, and argumentation, and, as is not uncommon during this
period, the investigations mostly proceed through the presentation of an opponent’s views,
the so-called prior position, followed by section designed to refute those views, the so-called
subsequent position. Often within a single investigation we find multiple variants of the prior
position, and in some cases, the investigation is itself further subdivided to address variant
views. For example, the seventh chapter, the “Investigation of the Self ” (the ātmaparīkṣā),
includes six subsections treating the theories of the Self as propounded by proponents of the
Nyāya, Mīmāṃsā, Sāṃkhya, Digambara Jaina, Advaita Vedānta, and Buddhist Vātsīputrīya
traditions. In each case, Kamalaśīla identifies specific thinkers and also identifies schools
of thought by name. His Extensive Commentary is frequently offered as the site of the first
appearance of the school name Advaita in a surviving Sanskrit composition.
The question of the audience is important not only in terms of getting the intellectual history
right, but, perhaps even more so, it is also an important question philosophically. Most likely we
will never settle whether the Compendium and its Extensive Commentary were written primarily
to win over opponents, to prepare Buddhist scholars for oral debate, or as part of the spiritual train-
ing of Buddhists who are “inclined toward wisdom” ( prajñānusārin), as Kamalaśīla puts in his
opening essay to the Extensive Commentary (Eltschinger 2014; McClintock 2013). Most likely,
all three of these purposes come into play (Marks 2019). What is beyond doubt, however, is that
within these works, we see the exemplification of a complex theory of the transactional nature of
reason itself according to which all human reasoning, whether undertaken for oneself or another,
must start with the lived realities, the embodied experiences, in short, the shared world (loka), of
the participants in the debate. To seek an uncontaminated starting point free from the distortions of
primordial ignorance would be sheer folly. Yet to eschew what appears in common on the grounds
that it is a distortion would be to abandon the commitment to rational inquiry for the sake of the
reduction or elimination of the suffering of sentient beings. Kamalaśīla shows a strong commit-
ment to reason throughout the Extensive Commentary, robustly defending the epistemological
instruments ( pramāṇa) as developed by Dharmakīrti and his earliest commentators, Devendrab-
uddhi and Śākyabuddhi. But he shows an equally strong commitment to working within the world
of ordinary experience, in one place pointing out that the elaborate philosophical theories of infer-
ence of those who espouse formal reasoning (nyāyavādin) are really no different from the day-to-
day forms of reasoning that are undertaken by ordinary persons in the world (McClintock 2019).
His argument seems designed to save philosophical inquiry from irrelevance not by elevating it
above conventional truths and realities but by confirming its embeddedness with them.
Another hermeneutical conundrum concerns the doxographical classification of the
Extensive Commentary. While it is generally counted as a work of epistemology (tshad
ma = pramāṇa), Tibetan commentators have nonetheless held that it, like the epistemologi-
cal works of Dignāga and Dharmakīrti, presents a Yogācāra position for its final view. This
assessment is made despite the fact that Kamalaśīla clearly indicates much of the argumenta-
tion to be offered at the so-called Sautrāntika level of analysis. At this point in the scholastic
485
Sara McClintock
development of Buddhist philosophy, the term Sautrāntika generally refers to a system in which
objects of knowledge external to the mind are held to exist, even if they are also understood to
be fleeting and infinitesimal. Yet despite this preponderance of Sautrāntika arguments, we also
find Kamalaśīla designating certain passages as being from the perspective of the Yogācāra –
which he likewise terms Vijñānavāda (“doctrine of cognition”), Vijñaptimātra (“representation
only”), or Cittamātra (“mind-only”) – a position in which no objects of knowledge exter-
nal to the mind are held to exist and which Kamalaśīla clearly judges to be superior to the
Sautrāntika. At the same time, a close reading of the commentary also reveals subtle refer-
ences to Madhyamaka (McClintock 2010, 2019), a position to which the work might be said
to gesture as its logical culmination. Whatever the final classification, there is no question that
the “sliding scale of analysis” at play in the work is yet another manifestation of Kamalaśīla’s
commitment to a vision of rational inquiry as thoroughly conditioned and context dependent.11
An important example of a Yogācāra doctrine defended in the Extensive Commentary
is reflexive awareness (svasaṃvedana, svasaṃvitti). This is the idea that awareness has the
nature of illumination ( prabhāsvara) and for this reason is always inherently self-aware.12
The doctrine of reflexive awareness is a hallmark of Dignāga and Dharmakīrti’s epistemology,
and it is not surprising to see it turn up here as well. For Dharmakīrti, reflexive awareness is
the ultimate trustworthy awareness ( pāramārthika-pramāṇa), which makes sense in Yogācāra
terms since a mind knowing the real must by definition be knowing the only thing that is real,
mind itself.13 Yet this kind of knowing is also radically nondual since knower and known are
ultimately one and the same. Kamalaśīla asserts that this reflexive awareness, which amounts
in the end to a vision of reality (tattvadarśana), is the primary form of the mind.14 Other forms
of trustworthy awareness in which the mind appears to know an object external to itself are
accepted in merely conventional (saṃvṛti) or transactional (sāṃvyavahārika) terms.15 In keep-
ing with Kamalaśīla’s gradualism and Madhyamaka leanings, however, we must remember
that the acceptance of reflexive awareness as the ultimate trustworthy awareness, as well as
the Yogācāra tenet that everything is mind only, are themselves valid only in conventional
terms.16 Alternatively, we can take a page from Jay L. Garfield (2016) and recognize that for
Kamalaśīla, as for Śāntarakṣita, the Cittamātra position obtains phenomenologically, while the
Madhyamaka position obtains on an ontological level.17
Other Works on Madhyamaka and Epistemology

in Kamalaśīla’s Oeuvre
So far we have considered five of Kamalaśīla’s most famous works: the three texts that all bear
the same name and are collectively called Stages of Cultivation; the single long, independent
treatise that sets out to establish the Madhyamaka view of emptiness, Light of the Middle; and
the detailed, polemical work covering a host of philosophical topics, the Extensive Commen-
tary on Śāntarakṣita’s Compendium of Realities. Of these, the Stages of Cultivation and the
Light of the Middle are both classified as belonging to the Madhyamaka (dbu ma) section, and
the Extensive Commentary is contained in the Epistemology (tshad ma = pramāṇa) section
of the Treatises, or Tengyur (bstan ’gyur) portion of the Tibetan canon. Kamalaśīla is credited
with many more texts than these, up to 30 in the Derge edition of the Tengyur. Unfortunately,
current scholarship does not allow us to make a clear determination as to whether all these
texts should be attributed to a single hand.
A number of these texts, however, can be confidently attributed to the same Kamalaśīla.
These include several more works in the Madhyamaka section, including the Exten-
sive Commentary on the Ornament of the Middle (Madhyamakālaṃkārapañjikā), a
486
Kamalaśīla
substantive prose subcommentary on his teacher Śāntarakṣita’s well-known Ornament of the

Middle (Madhyamakālaṃkāra). Also in this section are the two shorter works Light of Real-
ity (Tattvāloka) and Demonstration of the Naturelessness of All Things (Sarvadharmaniḥsva-
bhāvasiddhi), both of which again utilize the tools of Buddhist epistemology to demonstrate
the correctness of the doctrine of emptiness. Finally, in the Madhyamaka section, we find two
short works highlighting the centrality of meditation and spiritual practice, the Introduction to
the Cultivation of Yoga (Yogabhāvanāvatāra) and the Cultivation of the Mind of Awakening
(Bodhicittabhāvanā). These again point to Kamalaśīla’s project of integrating Madhyamaka’s
nonfoundational insights into a gradual path of yogic cultivation that insists on the necessity of
adhering to conventionally correct ethical and epistemological procedures along the way. The
Cultivation of the Mind of Awakening discusses a passage from the Secret Communion Tantra
(Guhyasamājatantra), a point which raises an important question about Kamalaśīla’s famili-
arity and involvement with the Buddhist esoteric stream and prompts Marks and Eltschinger
(2019, 274) to note that “any pronouncement concerning this work’s authenticity seems prema-
ture.” More broadly, one might say that any pronouncement concerning the engagement or lack
thereof with tantric activities on the part of Kamalaśīla and his Indian confrères is premature.
There remains one more work by Kamalaśīla in the Epistemology section of the canon,
namely the Summary of Prior Positions in the Drop of Reason (Nyāyabindupūrvapakṣa
saṃkṣipta). This interesting tract, which sadly does not survive in Sanskrit, seeks to illumine
Dharmakīrti’s basic guide to reasoning, the Drop of Reason (Nyāyabindu), not through direct
commentary but rather through providing an account of the opponents to whom Kamalaśīla
understands Dharmakīrti’s treatise to respond. Marks (2019, 263) wonders about the purpose
of this little-studied work. Was it “an assignment given to a close disciple, a young teacher’s
lecture notes, the work of an overeager student, or perhaps even preliminary research toward
the Pañjikā [i.e., the Extensive Commentary]?” While we may not be able to answer such
questions, Kamalaśīla’s interest in the dialogical nature of reasoning is again evident here.
For Kamalaśīla, it appears, stating the rules for constructing a proper inference, for example,
makes sense only in the context of an actual argument with someone who holds a different
position. Reasoning, in other words, is always a thoroughly contextual and conditioned affair.
The Summary of Prior Positions in the Drop of Reason is also notable for including a short
exposition on a famous scriptural quotation that appears in both the Compendium and in its
Extensive Commentary. The quotation epitomizes Śāntarakṣita and Kamalaśīla’s attitude toward
reason and religious authority and is therefore of great interest for understanding their perspec-
tive on philosophy as a whole. Presented as spoken by the Buddha himself, the verse reads:
Just as wise persons accept gold, having first tested it through heating it, cutting it,
and rubbing it on a touchstone, so too, O monks, should you accept my words only
after testing them, and not out of respect for me.18
Kamalaśīla’s Summary of Prior Positions makes clear that the three tests referenced here are
metaphors for three kinds of examinations that wise persons make in relation to three kinds of
epistemic objects.19 First, there are objects that can be known directly ( pratyakṣa) through per-
ception. This direct knowing is similar to testing gold by heating or melting it. Next are objects
that can be reliably known only indirectly ( parokṣa) through inference. This indirect knowing
is similar to testing gold by cutting it. Finally, there are objects that are epistemically remote
(atyantaparokṣa) or unknowable for ordinary persons and which can therefore only be known
through the testimony of an awakened authority like the Buddha. This extremely indirect kind
of knowing is similar to testing gold by rubbing it on a touchstone. As Kamalaśīla makes
487
Sara McClintock
clear, following Dharmakīrti, Buddhists should rely primarily on perception ( pratyakṣa) and
inference (anumāna), specifically inference that functions through the force of a real entity
(vastubalapravṛttānumāna), to obtain valid knowledge. Only in cases where it is truly neces-
sary for one’s spiritual progress to have knowledge of some supersensible matter should wise
people rely on scriptural testimony through a mechanism known as scripturally based infer-
ence (āgamāśritānumāna), a kind of pseudo-inference whose justification is entirely cast in
pragmatic terms.20 It is not hard to see why the current Dalai Lama is fond of citing this verse,
given his self-proclaimed commitment to the methods of modern science.
Sūtra Commentaries
An aspect of Kamalaśīla’s contribution to Indian Buddhist philosophy that is often overlooked
(and where there remains considerable room for future research) can be found in his substan-
tive Mahāyāna sūtra commentaries preserved in the Tibetan canon. The lack of attention is
not so surprising, in part because these texts do not survive in Sanskrit but just as importantly
because the philosophical contribution of sūtra literature is only recently gaining the atten-
tion it deserves. One difficulty in reading and interpreting the sūtras is their dialogical nature.
Unlike other philosophical works in the Buddhist tradition, including the verse and prose trea-
tises we have considered thus far, sūtras (and their Pāli counterparts, suttas) are presented not
as works penned by an individual author but rather as the recordings of oral teachings given
by the Buddha or his representative to some particular audience or addressee. As a genre,
therefore, sūtras require a different set of hermeneutical tools to read and interpret than do the
systematic treatises, or śāstras, that we more readily recognize as forms of philosophy. The
Tibetan canon credits Kamalaśīla with five sūtra commentaries. These can be useful for refin-
ing our understanding of Kamalaśīla as a philosopher, although space allows only a brief hint
of how this may be so.
We begin with Kamalaśīla’s Commentary on the Rice Seedling Sūtra (Śālistambaṭīkā).21 This
sūtra is an extended meditation on a pithy, enigmatic teaching – itself also termed a sūtra within
the larger text – that the Buddha is said to have uttered upon spying a young rice seedling. The
teaching goes: “Who sees dependent arising sees the Dharma. Who sees the Dharma sees the
Buddha.” In the Rice Seedling Sūtra, we encounter this curious teaching through the report of
the monk Śāriputra, who further claims that after uttering the sūtra, the Buddha remains silent.
Śāriputra reports the teaching to the bodhisattva Maitreya, who subsequently gives a detailed
explication, including many core elements of the technicalities of dependent arising. This expli-
cation makes up the bulk of the larger sūtra. But despite his thoroughness, Maitreya refrains
from commenting on the Buddha’s silence, a fact that Richard Nance (2012, 126) helpfully
points out. Kamalaśīla, on the other hand, does remark on it.22 As Nance unpacks this passage
(126–128), we learn that Kamalaśīla understands the Buddha’s silence as fulfilling two distinct
functions keyed to the Buddha’s two audiences within the sūtra: learned monks, like Śāriputra,
and bodhisattvas, like Maitreya. For the monks, the silence is designed to destroy pride, since
the monks will be forced to recognize that they do not understand this enigmatic teaching. For
the bodhisattvas, the silence conveys the teaching’s profundity. Thus, in this short commentarial
passage, we encounter in a different way from what we find in his epistemological treatises
Kamalaśīla’s central idea that meaning is always determined at least in part by the audience to
which is it directed. This sociolinguistic aspect of language runs through Kamalaśīla’s work as
a philosopher, insofar as he understands all philosophy to be transactionally produced.
The Commentary on the Dhāraṇī of Entering the Nonconceptual (Avikalpapraveśadhāraṇīṭīkā,
known also as the Nirvikalpapraveśadhāraṇīṭīkā)23 is important in relation to another topic in
488
Kamalaśīla
Kamalaśīla’s philosophical thought: namely the issue of the relationship between conceptual
(savikalpa) and nonconceptual (nirvikalpa) knowledge or awareness ( jñāna) and how a practi-
tioner moves from the former to the latter. In the first of his three Stages of Cultivation, Kamalaśīla
cites a passage from the Dhāraṇī for Entering the Nonconceptual that counsels “non-mentation”
(amanasikāra) as a method for turning away from signs such as form and so on.24 But lest one
imagine that this practice of non-mentation in any way invalidates or contradicts conceptual
cognition and rational analysis, Kamalaśīla is quick to point out, in the words of David Seyfort
Ruegg, “that what is here intended by the term amanasikāra is not simple absence of mentation
(manasikārābhāvamātra) but, rather, that non-objectifying or non-apprehension which belongs to
him who analyses through discriminative knowledge.”25 What does this mean? In his Commentary
on the Dhāraṇī for Entering the Nonconceptual Kamalaśīla lays out two explanations for how to
construe the scriptural quote.26 First, non-mentation may be understood as the defining characteristic
(lakṣaṇa) of the proper examination of the real (bhūtapratyavekṣā). In this view, proper examination
of the real is the necessary antidote to our primordial, unexamined, misguided, and habitual disposi-
tion to conceptualize signs in appearances. That conceptualization is mentation (manasikāra), and
the proper examination of the real is its opposite, non-mentation or non-perception (anupalambha).
Alternatively, non-mentation may be understood to be the result of proper examination of the real,
which by way of metonymy is then called non-mentation. What really happens is that through the
proper examination of the real, which still operates under the influence of primordial ignorance
(avidyā), objectification of objects is removed and one enters the space of signlessness (one of three
doorways to freedom, along with emptiness and wishlessness). Even though the proper examination
of the real has a conceptual nature, it operates in such a way as to ultimately undermine and disman-
tle its objects of knowledge, much as the fire sticks used to start a fire are destroyed in the ensuing
blaze – an image found in the Questions of Kāśyapa (Kāśyapaparivarta) and other scriptures and
in various Madhyamaka treatises, including Kamalaśīla’s third of the three Stages of Cultivation.27
Kamalaśīla’s remaining three sūtra commentaries all fall in the Perfection of Wisdom
section of the Tibetan canon. These are the Commentary on the Perfection of Wisdom in 700
Lines (Saptaśatikāprajñāpāramitāṭīkā), the Commentary on the Diamond Cutter Perfection
of Wisdom (Prajñāpāramitāvajracchedikāṭīkā), and the Commentary on the Heart Sūtra Per-
fection of Wisdom (Prajñāpāramitāhṛdayanāmaṭīkā). Here, as in his other sūtra commentar-
ies, Kamalaśīla employs the techniques recommended by his predecessor Vasubandhu in his
Logic of Explication (Vyākhyāyukti).28 Scattered references to and treatments of these com-
mentaries may be found in contemporary scholarship in English, but for the most part, these
texts remain only partially studied and understood.29 M. David Eckel helpfully highlights that
Kamalaśīla’s primary aim in his Heart Sūtra commentary was to emphasize the necessity of
inference (anumāna) as the means to generate certainty (niścaya) about the ultimate truth – a
position perhaps surprising given the Heart Sūtra’s repeated apophatic discourse but entirely
in keeping with Kamalaśīla’s epistemological stance and commitment to gradualism.30
Tantra Commentaries and Other Writings

As with other authors from this era, there exists a sizeable collection of works (nine or more)
attributed to Kamalaśīla in the Tantra (rgyud) or Tantra Commentary (rgyud ’grel) sections of the
Tibetan canon. The general current consensus now is that these tantric texts are unlikely to be the
product of the same Kamalaśīla whose philosophical works we have been discussing, even though,
as we have noted, Kamalaśīla does comment on one tantric verse in a text attributed to him in the
Madhyamaka section of the Tibetan canon.31 Here again we come up against a difficult problem
about authorship: we have an idea about who Kamalaśīla must be based on not only the knowledge
489
Sara McClintock
we receive from our own reading of the texts that we ascribe to him but also on our reception of
hagiography, commentary, and contemporary scholarship. This image and the power it exerts over
us is what Michel Foucault (2003) has called the “author function,” a set of conditions through
which the idea we construct of an author serves to help explain the ideas we find in the texts we
attribute to that same author. The author function prompts us to seek and find unity in the author’s
oeuvre and also provides a key by which differences can be resolved. Similarly, the author function
allows us to recognize a certain style or voice across genres of texts attributed to a given author. In
the case of Kamalaśīla, the author function with which we currently operate colors how we imagine
him and tends to make the ascription of tantric commentaries feel dubious. Yet archeological evi-
dence at Nālandā, as well as Tibetan records, suggest that tantric practice was widespread among
Indian monastics of that period. Given that tantra was then an esoteric practice, is it surprising that
allusions to it are missing from non-tantric works? Of course, other means to determine authenticity
can be applied,32 but sadly the state of our field does not allow us to say more at this point.
The Epistles (spring yig) section of the Tibetan canon contains two works attributed to
Kamalaśīla. The first, Lamp for the Generation of Faith (Śraddhotpādapradīpa), draws on prior
Indian teachings about the preciousness of human life and the importance of developing the altru-
istic mind of awakening (bodhicitta). In this, it sets forth many of the ethical and religious themes
that would later be developed in the Tibetan tradition known as Stages of the Path (lam rim).
Although it is in the Epistles section, this text does not mention any particular addressee. In contrast,
the second work in this section, the Teaching on the Varieties of Suffering (Duḥkhaviśeṣanirdeśa),
does mention an addressee, who intriguingly appears likely to have been a woman.33 The fact that
the colophon contains no mention of a translator further suggests that Kamalaśīla may have writ-
ten the work directly in Tibetan and in response to the questions of a Tibetan disciple. Whatever
the case, the work lays out in poetic detail eight forms of suffering – starting with the cycle of
rebirth (saṃsāra) and moving through birth, old age, sickness-and-death, seeking the nonexistent,
seeking what is unobtainable, meeting an enemy, and separation from friends.34 These are then
equated with the five psychophysical aggregates (upadāna-skandha) that both constitute and fuel
the cycle of rebirth itself. Kamalaśīla punctuates the verses describing the varieties of suffering –
which he describes as a cesspool (’dam rdzab) – using iterative laments modeled after the first
interjection: “There is no way to bear the suffering of saṃsāra here in this world!” Suffering, of
course, is the first of the Buddha’s four noble truths. That there is no way to bear this suffering
leaves only one option: to seek to obtain the freedom of nirvāṇa.
Another text ascribed to Kamalaśīla is Remembering the Words of the Fifty Verses for Nov-
ices (lnga bcu pa’i tshig gi brjed byang, *Śramaṇapañcāśatkārikāpādābhismaraṇa), a com-
mentary on Nāgārjuna’s training manual for novice monks in verse. Here we find Kamalaśīla
in full monastic register, providing summaries of some of the most important narratives sur-
rounding the root downfall ( pārājikā) rules for monks as drawn from the Mūlasarvāstivāda
Vinaya lineage of monastic discipline. Although the text has not appeared in English transla-
tion, it continues to be read and studied among Tibetan monastics to this day. The lack of a
recorded Sanskrit title suggests that the text may have been composed in Tibet, a possibility
that makes sense given what must have been a need for pedagogical materials for the newly
budding monastic community at Samyé during the period of Kamalaśīla’s reputed stay.
Doxographical Classification
Many articles about Kamalaśīla begin with his doxographical affiliation as this has been
assigned by later Tibetans and continued by contemporary scholars. Such an approach paints
Kamalaśīla as a representative of a school and not as an individual, and this is why we have
490
Kamalaśīla
avoided starting with such classifications. Until now, we have spoken only in terms of school
names that Kamalaśīla himself would have recognized: Madhyamaka, Yogācāra, Sautrāntika,
and so on. We’ve also spoken about the Buddhist epistemological tradition, a stream of Indian
Buddhist thinking that Kamalaśīla likely viewed not so much in terms of doctrinal distinctions
but rather in terms of a commitment to particular interscholastic forms of rational inquiry.
Indeed, in his Extensive Commentary on the Compendium of Realities, Kamalaśīla praises
and identifies with the designation of Nyāyavādin, the “propounder of formal reasoning.”35
It is thus abundantly clear that Kamalaśīla’s commitment to reason and rational inquiry is a
through line in nearly all of his writings, as we have seen repeatedly previously.
Kamalaśīla’s commitment to the procedures of formal reasoning is also in large part behind
what was eventually to become the most prevalent doxographic label attached to his name, that
of the so-called Yogācāra-Svātantrika-Mādhyamika (rnal ’byor spyod pa’i dbu ma rang rgyud
pa). This label has its own complicated history, combining as it does two separate doxographi-
cal divisions devised by Tibetans.36 The first can likely be credited to the eighth-century Tibetan
scholar Yeshé Dé (Ye shes sde), who spoke of two kinds of Mādhyamikas: the Sautrāntika-
Mādhyamika (mdo sde pa’i dbu ma pa), who accepts the existence of objects of knowledge
external to the mind, and the Yogācāra-Mādhyamika (rnal ’byor spyod pa’i dbu ma pa), who
does not. Both Śāntarakṣita and Kamalaśīla fall into the later camp, while Bhāviveka is seen to
represent the former. A second doxographical division was promulgated later, from the eleventh
century onward, originating perhaps with the translator Patshab Nyimadrak (Pa tshab nyi ma
grags, b. 1055), namely the division between the so-called Svātantrika-Mādhyamika (rang rgyud
pa’i dbu ma pa) and the Prāsaṅgika-Mādhyamika (thal ’gyur ba’i dbu ma pa). At base, this divi-
sion depends on whether the Mādhyamika in question accepts the use of so-called autonomous
proof statements (svatantra-prayoga = rang rgyud kyi sbyor ba). On this view, Śāntarakṣita and
Kamalaśīla, as enthusiastic adopters of Dignāga and Dharmakīrti’s epistemological apparatus
with its fulsome reliance on proof statements, must be classified as Svātantrika-Mādhyamikas.
Others, such as the seventh-century Mādhyamika Candrakīrti, reject the use of the technical tools
of Buddhist epistemology as inappropriate for a Mādhyamika. Such thinkers are then classified
by later Tibetans as Prāsaṅgika-Mādhyamikas in a gesture toward their preference for arguments
relying on the urging of unwanted consequences ( prasaṅga = thal ’gyur) on their opponents.
A detailed examination of these later Tibetan doxographical schemas is beyond the scope of
this chapter. Yet because these doxographical systems have exerted a strong influence on con-
temporary scholars, it is necessary to have a basic understanding of the issues at hand. Key in
all of this is the problem of natures (svabhāva), since the inferential reasoning of the Buddhist
epistemologists depends on identifying a natural relation (svabhāva-pratibandha) between the
elements in an inference (i.e., between the evidence and that which it proves). Those Tibetans
who maintain, as nearly all Tibetans eventually do, that the so-called Prāsaṅgika-Mādhyamika
is superior to the so-called Svātantrika-Mādhyamika do so on the basis that the Svātantrika
puts inappropriate stock in natures, a move held to be antithetical to Madhyamaka thought.
Also key is that such natures would operate “autonomously” (rang rgyud kyis) or “by their
own nature” (rang bzhin gyis). But such an idea is far less likely to have been acceptable to
Kamalaśīla. As a Mādhyamika, Kamalaśīla does not accept that natures are established auton-
omously or intrinsically. His commitment to natures is pragmatic and operates fully on the
conventional or transactional level – which, we must remember, includes even nonconceptual
objects of awareness such as the objects known in direct perception (see McClintock 2019).
The distinction, therefore, is dubious, though not without philosophical interest.
Recently, Kevin Vose (2020) has argued that the development of the Svātantrika-Prāsaṅgika
distinction should indeed be traced to Patshab Nyimadrak, specifically to his close reading
491
Sara McClintock
of Śāntarakṣita and Kamalaśīla’s epistemological writings. Vose explains how for Patshab,
Śāntarakṣita and Kamalaśīla’s reliance on autonomously established natures qua appearances
“makes ultimate conventions in the world” (707). In the conclusion of his article, Vose con-
cedes that the development of the distinction “may be unfair” and involves “several interpre-
tive leaps” that rely on the “stretching of terminology” (745). The upshot, then, is that the
label of Svātantrika-Mādhyamika is probably best avoided when discussing Śāntarakṣita and
Kamalaśīla except insofar as it offers an excellent opportunity to further investigate and refine
our understanding of how these thinkers square their commitment to the structures of reason
with their equally vehement commitment to a Madhyamaka vision of naturelessness.
How to resolve these two elements of Kamalaśīla’s thought, the rational side which relies
on natures to undertake a proper examination of the real and the anti-foundationalist side
which deconstructs natures at every turn? Gradualism is one part of the answer, as we have
seen. Kamalaśīla asks us to utilize the natures that appear to us in order to examine them and
discover their ultimate unreality. In doing so, we are gradually weaned from grasping at objects
external to the mind (at the Sautrāntika level of analysis), until eventually we are weaned as
well from grasping the mind itself (at the Yogācāra level of analysis). Finally, the philosopher
is freed when the Madhyamaka nonconceptual vision of naturelessness dawns. But arriving
there takes some work. For Kamalaśīla, it is necessary to engage with natures, despite their
lack of ultimate reality, in order to obtain freedom from seeing them as ultimately real. Walk-
ing away from natures without doing the work is not an option. Likewise, it is equally neces-
sary to engage with natures if we are to be of benefit to others. Again, this is true even though
natures may be ultimately unreal. In Light of the Middle,37 Kamalaśīla states that “even though
an illusion is false, it is relied upon when there is a purpose, as in the case of magicians.”
Comparing buddhas and bodhisattvas to magicians, Kamalaśīla insists that these compassion-
ate ones “do not turn away from anything whatsoever that is of use to others.” Natures, which
form the basis for investigation, are ultimately unreal. Yet like other unreal appearances, they
arise in a way that is dependent on causes and conditions. Thus, “just as illusions and so forth
arise in dependence on the collection of their own causes, likewise the ‘illusion’ which is the
wisdom of the yogis and so forth, which is acceptable as long as it is not analyzed, also arises.”
Even the pinnacle of philosophical knowledge, wisdom itself, is an illusion. But it is a kind
of necessary illusion, especially for those motivated by compassion. Reason is thus a form of
magic for this Mādhyamika champion of reason. Although an illusion, it nevertheless serves a
great purpose – namely to convey sentient beings to freedom from the grips of illusion.
Abbreviations
BhK Bhāvanākrama, Stages of Cultivation
MAV Madhyamakālaṃkāravṛtti, Ornament of the Middle
MAP Madhyamakālaṃkārapañjikā, Extensive Commentary on Ornament of the Middle
MĀ Madhyamakāloka, Light of the Middle
NBPS Nyāyabindupūrvapakṣasaṃkṣipta, Summary of Prior Positions in the Drop of Reason
TS Tattvasaṃgraha, Compendium of Realities
TSP Tattvasaṃgrahapañjikā, Extensive Commentary on Compendium of Realities
Notes
1 Chapa’s commentary, entitled Extensive Explanation Commentary on Light of the Middle (dBu
ma snang ba’i ’grel pa rgya cher bshad pa or sNang bshad), is published in vol. 6 of the Kadam
492
Kamalaśīla
Collected Works (bKa’ gdams gsung ’bum) and is currently being edited and translated by Pascale
Hugon and Jongbok Yi. It is also discussed briefly in Hugon 2015. As for the partial commentary by
Ngawang Tendar, entitled Notes on Light of the Middle (dBu ma snang ba’i brjed tho), although it
covers only the prior positions ( pūrvapakṣa), it is nonetheless quite useful. See also Ye, Li, and Kano
2013, 37–38 for discussion of a partial Sanskrit manuscript of an apparent commentary on the MĀ.
2 Seyfort Ruegg 1981, 96–99 summarizes the contents of all three texts. Arnold 2017 provides a philo-
sophical summary.
3 Kellner 2020 nicely summarizes these interlocking themes in the BhK texts.
4 Funayama 2011 discusses Kamalaśīla’s theory of yogic perception.
5 Keira 2004, 10–13 n. 32 locates the five arguments in the MĀ and discusses other lists of proofs in
Indian texts as well as the reception of Kamalaśīla’s list in Tibet.
6 Ichigō 1992, 199–202 reviews various identifications made especially by Japanese scholars while
also providing a synopsis of the prior positions according to the commentary of Ngawang Tendar.
Keira 2004 includes English translations of the subsequent positions answering to prior positions
4–14 and 33–42 according to Ngawang Tendar’s numbering.
7 Keira 2009 translates relevant passages from MĀ.
8 See Keira 2009, 14–15 for translation and analysis of the larger passage.
9 See Eckel 1987b for a translation of Jñānagarbha’s Distinction Between the Two Truths
(Satyadvayavibhaṅgavrtti) with excerpts from Śāntarakṣita’s commentary in the endnotes.
10 Note that Krishnamacharya mistakenly believed a verse to have been elided from the manuscript, so
his verse numbers are off by one digit starting at verse 527. We follow Shastri’s numbering throughout.
11 On the topic of the sliding scale of analysis, see my treatment in McClintock 2010, 85–91, in which
I also reference the discussion in Dreyfus 1997, 83–105; Dunne 2004, 53–79. Kellner 2011 critiques
the idea.
12 See TSP ad TS 3434.
13 Dharmakīrti invokes the term pāramārthika-pramāṇa at Pramāṇaviniścaya I.58. For a recent in-depth
exploration of Dharmakīrti’s theory of perception, including reflexive awareness, see Yiannopoulos 2020.
14 See TSP ad TS 3435–3436. See also TS 2000ff. with Kamalaśīla’s commentary.
15 See TSP ad TS 2980–2981.
16 See, for example, Kamalaśīla’s commentary on Śāntarakṣita’s MAV 16 (Ichigō 1985, 71), where he
indicates that the characterization of reflexive awareness as independently luminous (gsal ba gzhan la
mi bltos pa) is performed according to “the conventional path” (tha snyad kyi lam). For Kamalaśīla’s
understanding of Yogācāra philosophy as both crucial and provisional on the path of meditation, see
Arnold 2017.
17 The idea is that the Yogācāra philosophical stream is not to be read as arguing for a strict idealism
in which nothing but mind is real; rather, it is best understood as providing tools for the analysis of
phenomenal experience. These tools reveal that phenomenologically speaking – that is, in terms of
how we experience the world – there can be no ultimately real division between subject and object.
Instead, the mind, whose nature is to be illuminating, is aware of itself in an endless variety of forms.
But this observation concerns our experience alone and says nothing about what does and does not
exist in the world. Such ontological questions are best left to the Madhyamaka philosophical stream,
which recognizes all things, whether mental or physical, to be empty of essential nature.
18 TS 3587. Kamalaśīla also cites this verse in his commentary on the opening verses of the TS and
includes it as well in his NBPS 1.2.1. See also TSP ad TS 3343.
19 See Eltschinger 2013, 99–101; McClintock 2010, 61.
20 See Keira 2009; McClintock 2013 for Kamalaśīla’s understanding of scripturally based inference. For
Dharmakīrti’s position, see Tillemans 1999.
21 The sūtra and Kamalaśīla’s commentary have been critically edited, studied, and translated in Schoe-
ning 1995.
22 Nance notes that the two audiences may overlap in some instances.
23 Kamalaśīla’s commentary has been partially translated in Robertson 2006.
24 BhK I, Tucci 1958, 212. For a study of amanasikāra, including its reception in Tibet, see Higgins
2006.
25 Seyfort Ruegg 1989, 94. See Tucci 1958, 212.
26 The explanation that follows relies on Seyfort Ruegg 1989, n. 179.
27 Tucci 1971, 20. See also Kellner 2020, 52.
28 See Schoening 1995, 138–43. See also Nance 2012, 100–22 and 129–52.
493
Sara McClintock
29 Tucci 1956, 129–71 contains a summary of the Commentary on the Diamond Cutter Perfection of Wisdom.
30 Eckel 1987a, 76.
31 Eltschinger and Marks (2019, 274) speculate these works may have been authored by the tantric
master Padampa Sangyé (Pha dam pa Sangs rgyas, d. 1117).
32 Steinkellner (1999, 355–60) exemplifies the kind of required close reading when he analyzes the
tantric text Establishing Reality (Tattvasiddhi) attributed to Kamalaśīla’s teacher Śāntarakṣita. Stein-
kellner concludes that while the text does show evidence of training in the Buddhist epistemological
tradition, the author “commits at least one major logical blunder,” thus effectively nullifying the
possibility of the text being authored by the Śāntarakṣita we know. Steinkellner does not discuss,
however, the possibility of scribal error or other factors that could explain this blunder. While there
may be reasons to doubt the attributions of authorship of tantric texts to famous Indian philosophers,
there are reasons to keep this question open for future research.
33 Dietz 1984, 75–76 provides a summary of the work and discusses the identity of the addressee, whose
name is given as Lho Zamo Tshangpé Yang (Lho za mo tshangs pa’i dbyang) and variants thereof.
Dietz 1984, 340–57 also includes an edition of the Tibetan text and an annotated German translation.
34 This list of eight forms of suffering differs from lists found in other texts. Dietz 1984, 343 n. 3 notes
the unusual way Kamalaśīla considers death to fall within the category of sickness.
35 See McClintock 2019 for details.
36 See Mimaki 1983 for details.
37 Citations are from the Derge edition, vol. sa, D 222b–223b.
Bibliography
Adam, Martin T. 2002. “Meditation and the Concept of Insight in Kamalaśīla’s Bhāvanākramas.” PhD
diss., McGill University.
Arnold, Dan. 2017. “Pushing Idealism Beyond Its Limits: The Place of Philosophy in Kamalaśīla’s Steps
of Cultivation.” In The Oxford Handbook of Indian Philosophy, edited by Jonardon Ganeri, 379–99.
The Dalai Lama. 2001. Stages of Meditation. Root text by Kamalashila. Translated by Venerable Geshe Lob-
sang Jordhen, Losang Choephel Ganchenpa, and Jeremy Russell. Ithaca, NY: Snow Lion Publications.
The Dalai Lama, Khöntön Peljor Lhündrup, and José Ignacio Cabezón. 2011. Meditation on the Nature
of Mind. Somerville, MA: Wisdom Publications.
Demiéville, Paul. 1952. Le Concile de Lhasa. Une controverse sur le quiétisme entre bouddhistes de
l’Inde et de la Chine au VIIIe siècle de l’ère chrétienne. Paris: Presses Universitaires de France.
Reprint Paris: Collège de France, Institut des Hautes Études Chinoises, 2006.
Dietz, Siglinde. 1984. Die Buddhistische Briefliteratur Indiens: Nach dem tibetischen Tanjur heraus-
gegeben, übersetzt und erläutert. Wiesbaden: Otto Harrassowitz.
Dreyfus, Georges J. B. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpreta-
Dreyfus, Georges J. B., and Sara L. McClintock, eds. 2003. The Svātantrika-Prāsaṅgika Distinction:
What Difference Does a Difference Make? Boston: Wisdom Publications.
Dunne, John D. 2004. Foundations of Dharmakīrti’s Philosophy. Somerville, MA: Wisdom Publications.
Eckel, Malcolm David. 1987a. “Indian Commentaries on the Heart Sūtra: The Politics of Interpretation.”
Journal of the International Association of Buddhist Studies 10 (2): 69–79.
———. 1987b. Jñānagarbha’s Commentary on the Distinction Between the Two Truths: An Eighth-
Century Handbook of Madhyamaka Philosophy. Albany: State University of New York Press.
Eltschinger, Vincent. 2013. “Turning Hermeneutics into Apologetics: Reasoning and Rationality under
Changing Historical Circumstances.” In Scriptural Authority, Reason and Action. Proceedings of a
Panel at the 14th World Sanskrit Conference, Kyoto, September 1st – 5th 2009, edited by Vincent Eltsch-
inger and Helmut Krasser, 71–145. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
———. 2014. Buddhist Epistemology as Apologetics: Sources on the History, Self-Understanding and
Dogmatic Foundations of Late Indian Buddhist Philosophy. Vienna: Verlag der Österreichischen
Akademie der Wissenschaften.
Eltschinger, Vincent, and James Marks. 2019. “Kamalaśīla.” In Brill’s Encyclopedia of Buddhism. Vol II:
Lives, edited by Jonathan A. Silk et al., 273–78. Leiden and Boston: Brill.
Foucault, Michel. 2003. “What Is an Author?” In The Essential Foucault: Selections from the Essential Works
of Foucault 1954–1984, edited by Paul Rabinow and Nikolas Rose, 377–91. New York: The New Press.
494
Kamalaśīla
Funayama, Toru. 2011. “Kamalaśīla’s View on Yogic Perception and the Bodhisattva Path.” In Religion
and Logic in Buddhist Philosophical Analysis. Proceedings of the Fourth International Dharmakīrti
Conference. Vienna, August 23–27, 2005, edited by Helmut Krasser, Horst Lasic, Eli Franco, and
Birgit Kellner, 99–111. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
Garfield, Jay L. 2016. “Cittamātra as Conventional Truth from Śāntarakṣita to Mipham.” Journal of Bud-
dhist Philosophy 2: 263–80.
Higgins, David. 2006. “On the Development of the Non-Mentation (amanasikāra) Doctrine in Indo-
Tibetan Buddhism.” Journal of the International Association of Buddhist Studies 29 (2): 255–303.
Hugon, Pascale. 2015. “Proving Emptiness: The Epistemological Background for the ‘Neither One Nor
Many’ Argument and the Nature of Its Probandum in Phya pa Chos kyis senge ge’s Works.” Journal
of Buddhist Philosophy 1: 58–94.
Ichigō, Masamichi. 1985. Madhyamakālaṃkāra of Śāntarakṣita with His Own Commentary or Vṛtti and
with the Subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido.
———. 1992. “On the Dbu ma Snaṅ ba’i Brjed Tho.” Asiatische Studien/Études Asiatiques 46: 195–211.
Jha, Ganganatha, trans. 1937/1939. The Tattvasaṅgraha of Śāntarakṣita with the Commentary of
Kamalaśīla. Baroda: Oriental Institute.
Kajiyama, Yuichi. 1978. “Later Mādhyamikas on Epistemology and Meditation.” In Mahāyāna Buddhist
Meditation: Theory and Practice, edited by Minoru Kiyota, 114–43. Honolulu: University of Hawai’i
Press.
Voidness of All Dharmas. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität
Wien.
———. 2009. “The Description of Niḥsvabhāvatā and its Intentional Meaning: Kamalaśīla’s Solution
for the Doctrinal Conflict Between Mādhyamika and Yogācāra.” Acta Tibetica et Buddhica 2: 1–24.
Kellner, Birgit. 2011. “Dharmakīrti’s Criticism of External Realism and the Sliding Scale of Analy-
sis.” In Religion and Logic in Buddhist Philosophical Analysis. Proceedings of the Fourth Inter-
national Dharmakīrti Conference. Vienna, August 23–27, 2005, edited by Helmut Krasser, Horst
Lasic, Eli Franco, and Birgit Kellner, 291–98. Vienna: Verlag der Österreichischen Akademie der
Wissenschaften.
———. 2020. “Using Concepts to Eliminate Conceptualization: Kamalaśīla on Non-Conceptual Gnosis
(nirvikalpajñāna).” Journal of the International Association of Buddhist Studies 43: 39–80.
Krishnamacharya, Embar, trans. 1926. Tattvasaṅgraha of Śāntarakṣita: With the Commentary of
Kamalaśīla. 2 vols. Gaekwad’s Oriental Series, no. 30–31. Baroda: Oriental Institute.
Marks, James Michael. 2019. “Playfighting: Encountering Aviddhakarṇa and Bhāvivikta in Śāntarakṣita’s
Tattvasaṃgraha and Kamalaśīla’s Pañjikā.” PhD diss., University of California, Berkeley.
Meinert, Carmen. 2003. “Structural Analysis of the Bsam gtan mig sgron: A Comparison of the Fourfold
Correct Practice in the Āryāvikalpapraveśanāmadhāraṇī and the Contents of the Four Main Chapters
of the Bsam gtan mig sgron.” Journal of the International Association of Buddhist Studies 26 (1):
175–95.
McClintock, Sara. 2010. Omniscience and the Rhetoric of Reason: Śāntarakṣita and Kamalaśīla on
Rationality, Argumentation, and Religious Authority. Boston: Wisdom Publications.
———. 2013. “Kamalaśīla and Śāntarakṣita on Scripture and Reason: The Limits and Extent of ‘Practi-
cal Rationality’ in the Tattvasaṃgraha and Pañjikā.” In Scriptural Authority, Reason and Action. Pro-
ceedings of a Panel at the 14th World Sanskrit Conference, Kyoto, September 1st – 5th 2009, edited
by Vincent Eltschinger and Helmut Krasser, 209–37. Vienna: Verlag der Österreichischen Akademie
der Wissenschaften.
———. 2019. “How to Do Things with Natures: A Madhyamaka Approach to Arguments and Appear-
ances.” Journal of the International Association of Buddhist Studies 42: 405–447.
Mimaki, Katsumi. 1983. “The Blo gsal grub mtha’ and the Mādhyamkika Classification in Tibetan Grub
mtha’ Literature.” In Contributions on Tibetan and Buddhist Religion and Philosophy, edited by Ernst
Steinkellner and Helmut Tauscher, 161–67. Vienna: Arbeitskreis für Tibetische und Buddhistische
Studien Universität Wien.
Moriyama, Seitetsu. 1985. “An Annotated Translation of Kamalaśīla’s Sarvadharmaniḥsvabhāvasuddhi
Part IV.” Bukkyo Daigaku Daigakuin kenkyū kiyō 69: 36–86.
Nance, Richard F. 2012. Speaking for Buddhas: Scriptural Commentary in Indian Buddhism. New York:
495
Sara McClintock
Robertson, Raymond E. 2006. A Study of the “Dharmadharmatāvibhaṅga” (Volume 1): The Root
Text and its Scriptural Source (the Avikalpapraveśadhāraṇī), with Excerpts from Kamalaśīla’s
Avikalpapraveśadhāraṇīṭīkā. Beijing: China-Tibet Publishing House.
Schmid, Karl. 2020. “Kamalaśīla’s Theory of Vipaśyanā: An Exposition and Philosophical Defense.”
PhD diss., Emory University.
Schoening, Jeffrey D. 1995. The Śālistamba Sūtra and its Indian Commentaries. 2 vols. Vienna: Arbeit-
skreis für Tibetische und Buddhistische Studien Universität Wien.
———. 1989. Buddha-Nature, Mind and the Problem of Gradualism in a Comparative Perspective:
On the Transmission and Reception of Buddhism in India and Tibet. London: School of Oriental and
African Studies.
———. 2013. “On the Tibetan Historiography and Doxography of the ‘Great Debate of Samyé’.” In The
Tibetan History Reader, edited by Gray Tuttle and Kurtis R. Schaeffer, 108–20. New York: Columbia
University Press.
Shastri, Dwarika Das. 1968. Tattvasaṅgraha of Ācārya Shātarakṣita: With the Commentary ‘Pañjikā’ of
Shrī Kamalashīla, 2 vols. Varanasi: Bauddha Bharati.
Steinkellner, Ernst. 1999. “Yogic Cognition, Tantric Goal, and Other Methodological Applications of
Dharmakīrti’s Kāryānumāna Theorem.” In Dharmakīrti’s Thought and its Impact on Indian and
Tibetan Philosophy: Proceedings of the Third International Dharmakīrti Conference, Hiroshima,
November 4–6, 1997, edited by Shōryū Katsura, 349–62. Wien: Verlag der Österreichischen Akad-
emie der Wissenchaften.
Taniguchi, Fujio. 1992. “Quotations from the First Bhāvanākrama of Kamalaśīla Found in Some Indian
Texts.” In Tibetan Studies: Proceedings of the 5th Seminar of the International Association for
Tibetan Studies, Narita 1989, edited by Ihara Shōren and Yamaguchi Zuihō, 303–7. Tokyo: Naritasan
Shinshoji.
Tauscher, Helmut. 1999. “Phya pa chos kyi seng ge’s Opinion on prasaṅga in his dBu ma’i shar gsum gyi
stong thun.” In Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy: Proceed-
ings of the Third International Dharmakīrti Conference, Hiroshima, November 4–6, 1997, edited by
Shōryū Katsura, 383–93. Wien: Verlag der Österreichischen Akademie der Wissenchaften.
Tillemans, Tom J. F. 1999. “How Much of a Proof Is Scripturally Based Inference (āgamāśritānumāna)?”
In Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy: Proceedings of the Third
International Dharmakīrti Conference, Hiroshima, November 4–6, 1997, edited by Shōryū Katsura,
395–404. Wien: Verlag der Österreichischen Akademie der Wissenchaften.
Tucci, Giuseppe. 1956. Minor Buddhist Texts. Part I. Rome: Istituto Italiano per il Medio ed Estremo
Oriente.
———. 1958. Minor Buddhist Texts. Part II. First Bhāvanākrama of Kamalaśīla: Sanskrit and Tibetan
Texts with Introduction and English Summary. Rome: Istituto Italiano per il Medio ed Estremo
Oriente.
———. 1971. Minor Buddhist Texts. Part III. Third Bhāvanākrama. Rome: Istituto Italiano per il Medio
ed Estremo Oriente.
Ueyama, Daishun, Kenneth W. Eastman, and Jeffrey L. Broughton. 1983. “The Avikalapa-praveśa-
dhāraṇī: The Dharani of Entering Non-Discrimination.” Bulletin of Institute of Buddhist Cultural
Studies 22: 32–42.
van der Kuijp, Leonard. 1983. Contributions to the Development of Tibetan Buddhist Epistemology:
From the Eleventh to the Thirteenth Century. Wiesbaden: Franz Steiner Verlag.
Vose, Kevin. 2020. “When Did Svatantra Inference Gain Its Autonomy? Śāntarakṣita and Kamalaśīla as
Sources for a Tibetan Distinction.” Journal of Indian Philosophy 48: 703–50.
Wangdu, Pasang, and Hildegard Diemberger, trans. and ed. 2000. dBa’ bzhed: The Royal Narrative Con-
cerning the Bringing of the Buddha’s Doctrine to Tibet. Vienna: Verlag der Österreichischen Akad-
Ye, Shaoyong, Xuezhu Li, and Kazuo Kano. 2013. “Further Folios from the Set of Miscellaneous Texts
in Śāradā Palm-leaves from Zha lu Ri phug.” China Tibetology 1: 30–48.
Yiannopoulos, Alexander. 2020. “The Structure of Dharmakīrti’s Philosophy: A Study of Object-Cognition
in the Perception Chapter ( pratyakṣapariccheda) of the Pramāṇasamuccaya, the Pramāṇavārttika,
and Their Earliest Commentaries.” PhD diss., Emory University.
496
30
HARIBHADRA
The Voice of Perfect Wisdom
Introduction
Haribhadra (c. eighth century CE) is renowned as one of the foremost specialists in the lite
rature of Perfect Wisdom ( prajñā-pāramitā) in late Indian Buddhism. In Haribhadra, we meet
two of the most impressive qualities of Buddhist philosophers: exegetical expertise and specu-
lative ingenuity. As an exegete, he stays faithful to the ideal of the Buddhist intellectual who
is able both to master the understanding of authoritative texts (āgama) and to defend and
establish their claims through reasoning ( yukti). At the same time, although what is left of
him consists essentially in commentarial work, his writings showcase a creativity that elevates
Haribhadra as a standard for the Perfect Wisdom tradition. Like so many other groundbreak-
ing philosophers, his contributions forced subsequent authors to take a position either for or
against him. He was not to be ignored by either late Indian or Tibetan authors, especially as
with the passage of time, his authority grew to become almost undisputed in Tibet.
Haribhadra’s philosophical contribution might best be summarized as an attempt to pro-
vide the metaphysical and ethical foundations of the Buddhist path (mārga). The exegeti-
cal tradition with which he is most closely associated, that of the Ornament of Realizations
(Abhisamayālaṃkāra, henceforth Ornament), is dedicated to unveiling the “vast” dimension
of the sūtras of Perfect Wisdom ( prajñā-pāramitā-sūtras).1 This is said to be in contrast to
the “profound” dimension on which well-known Mādhyamika philosophers like Nāgārjuna,
Āryadeva, and Bhāviveka famously focus. While the “profound” dimension concerns the
metaphysical conception of the nature of things qualified as empty (śūnya) or insubstantial
(niḥsvabhāva), the “vast” dimension is about the path or process of transformation that leads an
individual from their ordinary human condition to the perfect state of buddhahood. Haribhadra
can thus be identified as a “philosopher of the path,” with the proviso that this implies much
more than just ethical or pragmatic concerns, since the path is an all-encompassing concept
that involves in addition metaphysical, cognitive, emotional, physical, and affective aspects.
Biographical and Bibliographical Context

Given the obscurity surrounding many figures of classical Indian philosophy, we are lucky to
have some rather specific information regarding Haribhadra’s life.2 Mano (1972) confirms the
DOI: 10.4324/9781351030908-43 497

scholarly consensus by estimating his dates as ca. 735–795. The historical context surround-
ing Haribhadra’s life is marked by the flourishing period of the Pāla dynasty, especially under
the kingship of Dharmapāla (not to be confused with the scholar of the same name), which
represents a golden age in terms of military expansion and economic prosperity. Such favora-
ble conditions allowed for strong patronage of Buddhist monasteries and communities. The
great academic centers of Vikramaśīla, Odantapurī, and Somapur, which attracted so many of
the famous later Buddhist philosophers featured in this volume, were established during the
eighth to ninth centuries with the support of Pāla rulers.
Haribhadra’s intellectual profile can be inferred from a few clues (Harter 2019a). As the
colophon of his long and complex commentary indicates, he wrote it at the monastery of
Trikaṭuka, considered to be situated in Bengal (Sanderson 2009, 88–89). This suggests that he
was probably an “academic monk,” and one can imagine that he was renowned for his mastery
of the literature of Perfect Wisdom given the massive commentary he composed and, given the
fact that his commentaries often refer to the opinions of unnamed interpreters, that he evolved
in a learned and scholastic environment, debating and disagreeing with other scholars. It is
unclear how close he was to the intense scholastic activity of the big university-monasteries
mentioned previously. He is not remembered to have been part of the “teaching staff ” there,
but his works cite such well-known scholars as Śubhagupta, Jñānagarbha, and Śāntarakṣita,
which indicates that he was well aware of the debates of the time and even took a position on
some of the issues discussed then.
The tradition does not remember him as anything other than a specialist of Perfect Wisdom
literature, which sets him apart from other late Indian Buddhist philosophers who composed
treatises in different branches of Buddhist philosophy. As argued in the following, the reason
might be that he saw Perfect Wisdom literature as encompassing all aspects of Buddhist phi-
losophy and doctrine, which is why he did not feel compelled to write independent treatises
on epistemology ( pramāṇa) or Abhidharma. Perfect Wisdom literature can be said to present
two forms during his time: a sūtra form and a śāstra form. A sūtra is considered by the tradition
to be a teaching uttered or “authorized” by the Buddha. Many sūtras of Perfect Wisdom were
circulating at his time of varying lengths: in 25,000 verses, in 18,000 verses, in 8,000 verses,
or even in one syllable (Conze 1973, 1975, 1978). A śāstra is a treatise providing a theory to
be implemented practically (Pollock 1985), ranging from elephant training to medicine and
philosophy. The Ornament is one such treatise. Dedicated to the topic of Perfect Wisdom, it
emerged around the fourth century CE. Haribhadra is the first known commentator to ascribe
its authorship to Maitreya. The treatise established the possibility for a theory of the path based
on the sūtras of Perfect Wisdom by providing an interpretation of these sūtras that supposedly
uncovers their intention to present the whole Buddhist path with its stages, its prerequisites,
the virtues and qualities necessary to its cultivation and the faults that slow it down, and the
metaphysics that lays at its foundation (Harter 2015, 4–10). The Ornament is not a manual of
the path that lists practical advice about how to practice the Buddhist path. Instead, it takes
the form of an abstract treatise that reflects on what constitutes the path and what does not and
on the theoretical principles the path requires. This is why it is organized into eight chapters,
which do not mirror the stages of the path in a chronological order: it begins at the end, so to
speak, with a chapter on the omniscience of the buddha, before backtracking to the omnisci-
ence of lower forms of realization (the bodhisattva and the hearer).
Haribhadra’s bibliography displays his expertise in both sūtra and śāstra literature. He
remains famous for being the editor of the Sūtra on Perfect Wisdom in Twenty-Five Thousand
Verses (Pañcaviṃśatisāhasrikā, D 3790, P 5188), which was later transmitted only in Tibet
(Conze 1978, 34–39; Makransky 1997, 132). He also wrote three commentaries. The least
498
Haribhadra
well known of his commentaries concerns another important sūtra from the Perfect Wisdom
literature, the Versified Summary of the Precious Qualities (Ratnaguṇasaṃcayagāthā), a sort
of précis of the texts and ideas of the Perfect Wisdom literature (Yuyama 1976). He also
wrote two commentaries on the Ornament: his short Explanation on the Treatise ‘The Versified
Ornament of Realizations’ (Abhisamayālaṃkārakārikāśāstravivṛti, Amano 2000; Naughton
1989; hereafter Explanation) and his much longer commentary entitled The Light Illuminating
the Ornament of Realizations: A Commentary on Perfect Wisdom (Abhisamayālaṃkārālokā
prajñāpāramitāvyakhyā, Wogihara 1932–1935; Sparham 2006–2012; hereafter Light).
All three of these works focus on the same subject, Perfect Wisdom, albeit with different
emphases. The commentary on the Versified Summary of the Precious Qualities focuses on
this sūtra while blending in references to the structure of the Ornament in eight chapters and
seventy topics and reading the sūtra through that lens. The Light balances its focus on both the
Sūtra on Perfect Wisdom in Eight-Thousand Verses and the Ornament. It refers to both texts
constantly, offering a gloss on every word. Finally, the short Explanation focuses exclusively
on the Ornament (except for a few quotations from sūtras devoid of commentary), becoming
the first autonomous commentary on that text. With this commentary, Haribhadra elevated the
Ornament to a new status, that of a theoretical investigation into the concept of the path. These
three commentaries thus seem to make a statement that the meaning of Perfect Wisdom is not
enclosed in one text or even one form of text (sūtra or treatise) but rather resonates in different
corners of Buddhist literature, even when it is not recognized.
Haribhadra and the Voice of Perfect Wisdom

Haribhadra’s works suggest that his main philosophical concern can be identified with an
attempt to recover the voice of Perfect Wisdom. Speaking of a voice is not only a poetic meta-
phor: Haribhadra explains that a sūtra is a statement (vacana) and needs to be analyzed as such
(Wogihara 1932–1935, 2). It also implies that the voice of Perfect Wisdom be articulated with
Haribhadra’s own voice, which raises an important issue: is there a space for Haribhadra’s
philosophical individuality?
To address this issue, we should not dodge the philological question regarding the authorship
of the three commentaries ascribed to Haribhadra. To this question, the answer can be unequiv-
ocal: there is absolutely no reason to doubt the attributions of all three works to the same author.
The two commentaries on the Ornament are very close in terms of content and style, and entire
passages are reproduced from one to the other. As for the Commentary on the Versified Sum-
mary of the Precious Qualities, it showcases the same features as the two other commentaries:
blended references to the sūtras of Perfect Wisdom and the Ornament, usage of “excursus” to
expand on chosen subjects, and the exegetical reflex of adding “in reality” to explain away the
seeming nihilism that texts from the Perfect Wisdom literature can be accused of falling into.3
The philological question thus leads us to an interesting literary remark: there is a similar tone
that resonates in all of Haribhadra’s commentaries, which echoes the voice of Perfect Wisdom
by defending it from nihilist interpretations. Haribhadra does so by reevaluating the status of
relative truth (saṃvr․ ti-satya) and expanding the significance of Perfect Wisdom beyond a mere
defense of emptiness by providing philosophical essays on various subjects.
Removing doubts regarding the authorship of all three works to Haribhadra certainly helps
sketching Haribhadra’s philosophical profile, in contrast, for instance, to the uncertainty sur-
rounding Nāgārjuna’s authorship of so many texts. It nonetheless would be simplistic to think
that the issue is resolved. Indeed, Haribhadra is and presents himself merely as a commentator.
He never wrote independent works. His voice is thus constantly intertwined with the voice he
499
is trying to serve. Does a commentator have an individual voice? This question is not unique
to the Indian tradition. The vast majority of philosophical traditions are exegetical traditions,
at least until the nineteenth or twentieth century (although, arguably, exegesis is still largely
practiced by today’s philosophers, if in a subtler form). Philosophers have long thought that
the purpose of philosophical work is not to display originality in the way that, for instance, a
poet might want to. Truth and clarity are supposed to trump any other preoccupation. Hegel
himself stated that the task of the philosopher was nothing else than to be “the secretary of the
universal Spirit,” which tells us something important about such philosophical perspective:
what matters is less originality than authenticity as the most faithful way to serve truth, or the
Spirit, or Perfect Wisdom, and it is the particular way with which a philosopher renders that
service that marks their individuality.4
What is, then, this Perfect Wisdom of which Haribhadra could be said to be the secretary?
Haribhadra recognizes the polysemy of the term and articulates its different meanings to give
a comprehensive and rational account of Perfect Wisdom. He writes:
Perfect Wisdom is the perfection of wisdom, which is defined as the discernment of

phenomena. The principal Perfect Wisdom is the blessed buddha, the nondual know
ledge that is similar to illusion. But, in conformity with its attainment, the secondary
Perfect Wisdom is the text, which is a collection of words and sentences, and the
path, defined by the different stages like vision.5
Because the Buddhist tradition speaks of the sūtras of Perfect Wisdom, the term Perfect Wis-
dom is generally understood as designating a specific corpus of texts. But Haribhadra argues
that this is Perfect Wisdom only in a secondary sense; that is, it is derived from the original
sense of Perfect Wisdom. He defines it as “the discernment of dharmas” (dharma-pravicaya),
which is a long-standing definition of wisdom in the Buddhist tradition: wisdom or knowledge
is the capacity of telling apart, identifying, (etymologically) analyzing the different factors or
constituents (dharmas) that make up reality and our experience of it. For Haribhadra, this ana-
lytical power not only results in an ontological classification between different kinds of con-
stituents of reality but also concerns the different “modes of existence” of these factors, that
is, the ultimate truth ( paramārtha-satya) and the relative truth (saṃvr․ ti-satya).6 Thus Perfect
Wisdom is nothing other than the capacity of seeing reality as it is ( yathābhūta) in its nature
and its multiplicity, freed from the delusions of false ideas such as self and other substances
(svabhāva). More generally, Haribhadra’s commentaries extend the meaning of Perfect Wis-
dom so as to comprise any kind of knowledge, including those he considers inferior, like the
knowledge of non-Buddhists and of other Buddhist schools and groups (śrāvakas and pratyeka-
buddhas). Hence, the term Perfect Wisdom arguably includes both the objective aspect of truth
(what reality is) and its cognitive counterpart (the apprehension of reality), standing as the
ultimate standard of truth undifferentiated into subject and object.
By extending the meaning in this way, Haribhadra accomplishes what we can call a “de-
textualization” of Perfect Wisdom: Perfect Wisdom is first and foremost the actuality of wis-
dom, before it is even considered a text.7 In the previous quotation, he describes this wisdom
in two ways: he calls it “buddha” and “nondual knowledge that is similar to illusion.” It is not
absolutely clear what he means by “buddha” in this passage because of the complexity of his
theory regarding what a buddha is,8 but it is clear that he does not simply mean the historical
figure born sometime in India around the sixth to fifth centuries BCE and known as Śākyamuni
Gautama. He means rather the actual embodiment of wisdom, including its embodiment in
a human life. The second description suggests also that “buddha” is primarily taken as an
500
Haribhadra
epistemic term, indicating a form of knowing cognition – the actual event of understanding
rather than the abstract disposition of knowledge. The expression “nondual knowledge that is
similar to illusion” (māyopamaṃ jñānam advayam) summarizes well the metaphysical posi-
tion Haribhadra holds, which we will explore in the following.
But in a secondary sense, Haribhadra argues, “Perfect Wisdom” can also signify texts that
explain this primary, actual wisdom, or the path that leads to it, since both texts and path are
called “Perfect Wisdom” only in reference to the first, primary sense of the phrase. Engineer-
ing is not a textbook on engineering but rather the knowledge used to accomplish bridge
construction or road design; yet this primary sense of engineering gives its name to textbooks
through which we can learn it. In the same way, Haribhadra has an instrumentalist vision of
texts and the path: they are both teleologically oriented toward the realization of Perfect Wis-
dom, that is, the attainment of the summum bonum that buddhahood or nondual knowledge
are understood to be. Thus philosophy’s main accomplishment is outside its texts and even
outside the “philosophical practices” that are designed to make us wise. Rather, philosophy’s
main accomplishment is the result consisting in wisdom. In this sense, there is much affinity
between Haribhadra’s thought and pre-modern forms of philosophy in Greece, Rome, or the
Islamic, Jewish, and Latin Middle Ages.9
Even though Perfect Wisdom does not have primarily a textual nature, it does express
itself in many texts, in sūtras and in treatises. By relating the multiplicity of sūtras of Perfect
Wisdom to a common non-textual source, Haribhadra can provide a commentary on Perfect
Wisdom mixing in references to the Sūtra in Eight-Thousand Verses as well as other sūtras
and to the Ornament. This plasticity in the expression of Perfect Wisdom suggests that Hari
bhadra’s conception of a “text” differs from the way modern scholars think about texts.10 It is
an ideal text expressing the essence of Perfect Wisdom he is aiming at in his commentaries
rather than any specific historical text. It is within this space between the ideal text and the
actual texts that Haribhadra’s creativity as an individual commentator can best be seen: his
comments sometimes seem rather loosely related to the words he is explaining, and he allows
himself to make many digressions or excursuses, which function as mini philosophical essays
throughout his Light. In other words, he is not a literalist and is not blindly subservient to the
written texts he has in front of his eyes.
This broad consideration of what constitutes the “text” of Perfect Wisdom is also the basis
for his criticism of those who preceded him in commenting the Ornament. In the opening
verses of his Explanation, Haribhadra emphasizes the comprehensiveness of Perfect Wisdom
and how the Ornament is a key to the understanding of all sūtras of Perfect Wisdom, not
only one of them, as Vimuktisena (ca. fifth to sixth centuries) thought when he explained the
Ornament as exclusively referring to the sūtra in 25,000 verses. Others showed partiality in
their perspective, like Vasubandhu, who seems to be accused of explaining the text from his
Yogācāra bias. In affirming his difference, Haribhadra found his own individual voice as a
commentator who echoed, distinctly, that of Perfect Wisdom with its wide compass. Key to
Haribhadra’s philosophical perspective is his commitment to such comprehensiveness and the
multiplicity of approaches, which reflects his conception of the totality of the Buddhist path.
The Philosopher of the Path as the Great Synthesizer

Haribhadra can rightfully be seen as one of the greatest Buddhist philosophers of the path,
that is, the process of transformation that leads an individual to the summum bonum of bud-
dhahood. His commentaries give an account of what constitutes this process, the qualities to
develop and faults to get rid of, the possibility of different trajectories and stages, and so on.
501
This account is not really descriptive, as if he were to report his own experience or that of
others. It rather provides a theory of the path: given a certain conception of buddhahood, argu-
ments can be elaborated about what counts as the path and what does not, what can be deemed
the best path, what is required for specific betterment, and so on. This is why his commenta
ries are not directly practical, since their purpose is not to provide a recipe for awakening but
instead to reflect about what the path is (Dreyfus 2003, 172–76).
Many Buddhist philosophers have focused on issues of metaphysics, understood as a dis-
course about the general nature of things. But once we know, for instance, that reality is insub-
stantial, that it is devoid of any substance or essence, as the Mādhyamika philosophers like
Nāgārjuna or Āryadeva argue, what do we do about it? And in what sense do we know that
this is the case? The reflection on the status and the different kinds of knowledge is a long-
standing approach in the Buddhist tradition. Haribhadra situates himself in a line of thought
that privileges forms of knowledge that are “nonconceptual” (avikalpita) in contrast with
forms of knowledge that are “conceptual.” What is wrong with a concept? Haribhadra quotes
Dharmakīrti’s definition of conceptualization:
One talks about conceptualization because something appears that is not a thing, and
it is a disturbance because it is discordant.11
The idea is that a concept or conceptualization (vikalpa) is something that is not a particular
thing but rather a general entity or universal. General entities have very different characte
ristics than do particulars: they are not restricted in place and time, they do not “appear” to a
subject (at least in the same sense as things can appear in perception), and they are the result of
processes of determination and abstraction. In that sense, concepts are not (particular) things,
are not real (because reality is impermanent and insubstantial), and are some sort of fictions
resulting from a judgment subsuming multiple instances under a unity – hence they are “dis-
cordant” in never exactly corresponding to particulars. Nevertheless, these fictions can allow
us to understand things and reality if we don’t get mesmerized by them such that we come to
believe that they have more ontological weight than they actually do. This is a very important
aspect of Haribhadra’s (and other Buddhist thinkers’) thought that cuts through the heart of his
theory of the path: fictions and even errors can help us to understand and to progress toward
the summum bonum as long as we can see through them, as we shall see in the following.
Such definition already shows that epistemological developments of Dignāga and
Dharmakīrti are presupposed and put in use in his theory of the path. These epistemologi-
cal and ontological considerations have epistemic consequences for the path. If a conceptual
cognition does not conform to reality but can still help to understand it, the process of cogni-
tive progress can reproduce this passage from conceptual to nonconceptual understanding.
This is the paradigm he uses as a representation of the path:
The science of all aspects, which is the full knowledge of all aspects such as the produc-
tion of the thought of awakening, is stated first for the bodhisattva longing for buddha-
hood by indicating that it is the fruit [...] Then in order to have mastery over the three
kinds of omniscience that have been acquired like that, again one realizes in every way
the awakening of all aspects by cultivating the three kinds of omniscience, which con-
sists in taking together all the sorts of gnoses of aspects, paths, and elements.12
In the first sentence of this typically technical passage, Haribhadra indicates that the first
chapter of the Ornament is about the science of all possible aspects of reality – in other
502
Haribhadra
words, the highest and broadest knowledge possible, that of a buddha. It is presented first,
he writes, because it is the end of the path, and the bodhisattva, the ideal practitioner of the
Buddhist path, needs to know the destination to which the path leads in order to be willing
to engage in it. The next two chapters present two lower kinds of knowledge, which have all
paths and all elements as their objects, respectively, and belong to bodhisattvas and hearers.
Reading about these three kinds of omniscience might be enough to “acquire” ( prāpta) them
(one can understand, through arguments, what reality is and, broadly, what ethical practices
are warranted), but this is not enough to “realize” (adhigamya) them. These two cognitive
and soteriological states of acquiring and realizing represent the two ends of the path, and in
other passages, Haribhadra clearly identifies the first moment of knowledge as a conceptual
one and the second as a nonconceptual one. The path is this passage from a certain inferior
kind of knowledge about reality in all its dimensions (i.e., elements, paths, and aspects) to
a superior form of knowledge, a knowledge that is transformative because it is actual and
no longer merely abstract. The passage from the first to the second form of knowledge is
ascribed to cultivation (bhāvanā), that is, a form of engagement that is not only cognitive,
but also practical, emotional, bodily, and so forth, that consists in habituating the practi-
tioner to a certain understanding so that it is appropriated, implemented, and becomes truly
the way she sees reality – hence the quotation indicating that all aspects, paths, and elements
need to be “taken together.”
This conception raises questions about the conditions for transformation, specifically about
the epistemological characteristics a cognition must bear to be able to be transformative. In
more familiar Western philosophical terms, the question is how understanding can determine
the will. How is the transformation of an individual based on a philosophical approach pos-
sible? Greek philosophers, for instance, considered progress toward the good life an essential
aspect of what philosophy is about, but they said little about how to proceed with it. The con-
cept of the path implies that a theoretical proposal cannot by itself trigger the change necessary
for the summum bonum to be attained. In other words, listening to the Buddha does not make
you become a buddha. There is much work to be done. An active process of transformation is
necessary that would guarantee the passage from the first stage of knowledge to the second.
This is what the Ornament, and by implication Haribhadra’s commentaries, are about.
Haribhadra, following a lead from Vimuktisena, addresses the question of the subject matter
of Perfect Wisdom in the Ornament at the beginning of his two commentaries. He says there
are three options concerning this topic: it can be about the fundamental elements of reality
(vastu) – a sort of ontological and metaphysical categorization of what reality is made of. But,
he argues, this would just reproduce what has already been done by Abhidharma treatises.
Or the Ornament could present all the “remedies” ( pratipakṣa) – a sort of ethical approach
that addresses the practices and ideas necessary for eliminating all obstacles that prevent us
from attaining the summum bonum such as negative inclinations and affects. Such a treatise,
Haribhadra explains, would be incomplete: to present “remedies,” you have to account for
what they are remedies of. Finally, it could be about the totality of aspects (ākāra), that is, the
totality of objects of knowledge. But Haribhadra wonders why this would differ from the first
proposal: aspects should be related to the fundamental elements of reality. He finally makes his
own proposal of a comprehensive and updated conception of the path as the subject matter of
the Ornament and Perfect Wisdom that would comprehend all three options:
It is the entire unmistaken path, comprehending the realizations of hearers, indepen

dent buddhas, bodhisattvas, and buddhas – thus including the three kinds of things
mentioned above and avoiding the faults of each option – organized in the stages of
503
the eight realizations starting with the science of all aspects, and leading to goods,
whether they are success in existence or the final good.13
This gives us a helpful glimpse into Haribhadra’s wholistic understanding of the path: the path is
a total transformation, whether it is cognitive, dispositional, emotional, ethical, and even physi-
cal, which is why its theory needs to comprehend all aspects, all paths, and all elements of real-
ity. This is why Haribhadra’s commentary displays such comprehensiveness in its approach as
well: within a Mādhyamika framework, he incorporates Abhidharma, epistemology ( pramāṇa),
and Yogācāra elements, discussing positions to either appropriate them or refute them but also
to articulate them along the dynamic arch of the path (Makransky 1997; Harter 2015).
Philosophical Excurses
As noted, Haribhadra’s long commentary is a very rich text not only because it manages to pro-
vide a word-by-word exegesis of the long Eight-Thousand Sūtra on Perfect Wisdom together
with the Ornament but also because it uses certain passages from these texts as opportunities
to launch sustained philosophical developments – a procedure that turned into a norm in Tibet,
where commentaries of the Ornament basically became encyclopedias of the Buddhist path in
which commentators would address any issue they deemed interesting. These “digressions” or
“excursuses” ( prasaṅga), as he calls them, are at times closely tied to the text of the sūtra and
are at other times seemingly completely autonomous and quite long. Among many possible
examples, we examine two here.
The Path as Beneficial Illusion

Haribhadra advocates unfailingly for the Mādhyamika philosophy of universal emptiness: noth-
ing has a substance or inherent essence (svabhāva), and yet things appear as if they do. Such a
position is potentially threatening to the whole Buddhist path, since a nihilist interpretation is
always possible: if nothing has any substance, what is the point of cultivating virtues, eliminating
anger, developing compassion? How would a path with its different stages and practices be
consistent? This is the reason one of Haribhadra’s main concerns in his commentaries is to
assert the reality of the relative truth (saṃvṛti-satya), which is the ontological dimension where
things appear so that they can be differentiated and identified – a necessary ontological dimen-
sion for any language and action – in contrast with the ultimate truth, which rejects any real
existence of substances or essences. The danger of a nihilist reading of the sūtras of Perfect
Wisdom was always looming, since these texts use negations heavily to reject any substantial
conception of reality. They even go as far as negating the existence of the Buddha, of wisdom,
of the path, of practitioners, of bodhisattvas, of virtues, and so on. One of Haribhadra’s signa-
tures is to add as a gloss “in reality” (tattvena) to restrict such negations to the dimension of
ultimate truth and to preserve some sort of ontological status for the relative truth. This strategy
is essential for a philosopher of the path, without which the path would lose all significance.
In one of his excurses, Haribhadra addresses the question of the status of these appearing
realities, with particular reference to the appearance of beings (Wogihara 1932–1935, 159–61;
translation in Sparham, vol. 2, 107–10). Often sūtras state that things are “similar to illusions” or
“analogous to illusions” (māyopama). In one particular passage, the gods, who are often charac-
ters in the stories or interlocutors of the Buddha, wonder if there is a difference between saying
that beings are similar to illusions or that they are just plainly illusions – the question possibly
pointing at the seeming similarity of the Buddhist position to the Vedāntin idea of the world
504
Haribhadra
being an illusion (māyā). If beings are just similar to illusions, it might mean that they are not
really illusions and that behind an illusory veil we can find what they really are. That is a slip-
pery slope for Mādhyamikas. In a nutshell, Haribhadra explains that, even if there is no onto-
logical difference between beings and illusions, there is still a phenomenological one. Indeed,
beings are illusions: the Buddhist tradition is known for its rejection of the existence of a self
as some sort of a core that would give beings their identities and unities. There is no such inhe
rent core or soul but rather fleeting bodily, sensuous, conscious, and unconscious processes that
function together for some time and to which we attribute identities. Such “consistent function-
ing” gives the illusion of identity, which is a conceptual construction (vikalpa) subsuming under
an imaginary unity multiple different instances in time and place. And so, Haribhadra argues,
there is no ontological difference between beings and illusions; otherwise we would have to
believe that there is something real “behind” the illusion of individual beings, and we would
be back to the same substantialist assumption that the Mādhyamikas fight so hard to uncover.
Using the analogy of a magic show, Haribhadra explains that what appears during the show
is not real in the sense that it is a pure illusion fabricated by a performer who knows how to use
tricks to conjure up images appearing to spectators. And yet, during the magic show, there are
still things that appear to the spectators, and these appearances make a difference between what
can actually appear and have a relative reality (due to their being produced by the mechanism
of the tricks; in other words, due to causal efficiency, the mark of relative truth for Haribhadra
following Dharmakīrti) and what cannot appear because it is just purely nonexistent, as illus-
trated by the well-known examples of the son of a barren woman or the horns of a rabbit, which
are outside any possible causal production. This double dimension of reality, Haribhadra con-
tinues, does not suffer exceptions: even the Buddha or nirvāṇa, he states in a very Nāgārjunian
fashion, are, ontologically speaking, illusions because they are not ultimately real. No single
phenomenon can be excused from emptiness; otherwise emptiness would be only partial. But
that does not mean that suffering in cyclic existence or being liberated from it are phenomeno-
logically the same. Phenomenologically speaking, they are said to be like illusions because
they can still be differentiated one from another and because they still function in ways that
really do allow the practitioner to abandon ignorance and to cultivate Perfect Wisdom.
A second “excursus” (Wogihara 1932–1935, 205; translation in Sparham, vol. 2, 152) could
be usefully connected with this one to get a fuller theory of illusion, one which allows Hari
bhadra to argue that the liberating process of the path can make use of illusions. In this pas-
sage, Haribhadra wonders whether the perfect knowledge of a buddha has any intentional
object at all: if all things are insubstantial according to Mādhyamika metaphysics, wouldn’t
a consistent epistemological position amount to an absence of any intentional object, since
intentional objects always entail a certain nature or substance? This is a common objection
raised by non-Buddhist thinkers in India, namely that the Buddhist summum bonum resembles
more that of a stone than of an accomplished human being. And after all, the sūtras do repeat
that the final cognitive state is “without intentional object” (anālambana).
Haribhadra responds by making a distinction between error (bhrānti) and aspect (ākāra).
Using the example of the dream, he explains that the objects seen in a dream are indeed ulti-
mately nonexistent (they disappear once we wake up), and yet these objects, which are mental
aspects seen by the person who is dreaming, follow certain causal processes: they provoke fear
or desire in that person, they can be acted upon, and so on. In the same way, in the waking state,
many cognitive states like perception arise with various clear and distinct objects, which do
appear as having unity (although there is none) and as being external (although they always
appear through our cognitions and we have no way to step outside of them to verify whether
they reflect what is actually outside). These aspects of our cognitions follow causal processes
505
that explain their various appearances and modifications, but these aspects themselves are not
“false” once we understand how they are the results of causes and conditions. Only our naïve
judgment that their appearance indicates ultimate reality is problematic. This is why Haribhadra
concludes:
Even when error stops, it is impossible for aspects to stop since there is no connection
between them.14
Spinoza argued in a similar fashion when he pointed out that even when we understand that the
appearance of the sun as being only 200 feet away from us is false and that it is in truth more
than 600 diameters of the earth away from us, the false appearance of it being so close does not
cease. Why? Because it is created by a complex of causes and conditions (our situation in the
universe, the laws of reflection of light, the particular condition of our eyes, etc.) that results in
such appearance. The appearance is true when we can understand its causal genealogy, but the
judgment we superimpose on it – that the sun is indeed 200 feet away – is erroneous.
Such argument is more radical than it seems because it means that all appearances can be
transfigured into elements of truth if we know how to consider them: when believed naïvely,
they can fool us, but when we can understand how they come to be, they open up opportuni-
ties for understanding and liberation from error. And thus, even though they can be deemed
ultimately false, when reconsidered within the web of the causal process, they can turn out to
be useful for the progress on the path. A fascinating example Haribhadra takes is how dedi-
cation, that is, the mental disposition and action to dedicate the benefits of an action to the
wellbeing of all beings (a core practice in Mahāyāna Buddhism consisting in orienting oneself
to constantly act for the benefit of others instead of pursuing one’s selfish goals), can be used
to progress on the path even though it is devoid of any reality.15 To dedicate an object or an
action to all beings, one has to use a concept of that object; that is, one has to superimpose
onto past, present, and future instances of that object (which can never coexist) a unity that
is a pure fabrication. And yet this illusion can be beneficial and contribute to the progress on
the path (by cultivating compassion and lessening self-promotion, for instance), as long as the
practitioner is aware of the manipulation. It is when this dedication (and everything entailed
by it: the practitioner, its addressees, etc.) is understood as being “similar to an illusion” that
a gap between the naïve realist apprehension and the nihilist diversion from ethical action can
open up for the possibility of knowledge and progress toward liberation.
Haribhadra the Metaphysician: The Great Discussion

on the Ultimate Nature of Things
The longest excursus in which Haribhadra indulges is an inquiry into the nature of reality or
suchness, which he frames as an inquiry into the nature of buddhahood. For Haribhadra, bud-
dhology (understood as the inquiry into what a buddha is) is never separate from metaphysics,
and there is a metaphysical reason for that: buddhahood is not a personal condition but rather
nothing else than the actual knowledge of reality. This knowledge is nondual in the sense
that it does not consider reality an object projected in front of (ob-jectum) a subject (as if the
subject could exclude itself from reality). Such a framework of subject–object is rejected as
a distortion of reality. Haribhadra, in a somewhat similar fashion to Neoplatonist thinkers,
maintains a radical or idealized conception of knowledge free from the duality of subject and
object, even though he still accepts a range of inferior kinds of knowledge that maintain a
dualist form and can be more or less attuned to it. The long excursus of chapter four (Wogihara
506
Haribhadra
1932–1935, 624–42; translation in Sparham, vol. 3, 238–71) deserves a much longer study,
but we can at least focus on the result of the inquiry. After asserting the absence of a difference
between what a buddha is and reality, Haribhadra wonders what account he can give of what is
real and whether it has a single or a multiple nature, a line of inquiry common to many think-
ers of the time.16
His inquiry takes the shape of a review of hypotheses that he identifies with specific philo-
sophical positions. He first refutes the position that ascribes reality to objects of knowledge
( jñeya), especially the materialist-atomistic position that takes atoms as constituting reality.
Next he turns toward the position that considers cognitive instances ( jñāna) the only existent
reality, which he attributes to Yogācāra thinkers. Haribhadra seems to accept the idea that
such mental events can be described as nondual, impermanent, and endowed with aspects.
This last characteristic indicates that our cognitive instances display appearances of distinct
objects as well as appearances of subjects, including our own consciousness as aware of inten-
tional objects. The rest of the discussion concerns the ontological status of these aspects, and
once again he frames the discussion in terms of two sub-positions in the Yogācāra perspective
(although he never actually frames them as being “schools”). The first is the position that takes
aspects to be real (sākārajñāna), and the second is the position that takes aspects to be false
(nirākārajñāna). He rejects the first position as lending too much credit to aspects. Haribhadra
posits that the mind needs to be thought of as a capacity of being a unity, holding a potentially
infinite number of changing aspects. Such multiplicity in unity can only be noncontradictory if
those aspects are unreal, since a real unity would be contradictory with a real multiplicity.17 He
criticizes the second option as lending not enough credit to aspects when they are dismissed
as false, because such a perspective misses their causal and relative natures. Thus cognitive
instances are no more ultimately existent than objects of knowledge but no more illusory than
other relative phenomena.
This extended, detailed, and subtle argument can be understood as a summary and reas-
sessment of his commitment to insubstantiality, reformulated in terms of the kind of ideal
knowledge (that of a buddha) to which the practitioner of the path should aspire. Haribhadra’s
final characterization of such knowledge is that of “a mental continuum that is merely nondual
consciousness, which is similar to an illusion.”18 The awareness that knows reality as it is is
itself impermanent; that is, it is a continuum or series ( prabandha) of mental events arising
due to a complex causal process. This claim therefore means that even the knowledge of a
buddha has a relative nature being made up by a causal process – quite a bold statement for
some Buddhist thinkers. This mental continuum is not a pure (subjective) idealism à la Fichte
or Berkeley but a nondual reality, since objects never appear outside of a conscious event, and
awareness is itself not separated from its objects qua mental appearances. But these should not
be given an ultimate ontological status either, since the whole structure of nondual awareness
is itself “similar to illusion,” an expression, as we have already discussed, that tries to escape
the pitfalls of realist and nihilist interpretations.
Haribhadra’s position thus avoids the tempting characterization in terms of idealism, since
awareness or consciousness is not given a privileged metaphysical status. It is just as insubstan-
tial as other phenomena. He should be understood in his own terms: as a Mādhyamika thinker
who strives to maintain a balance between ultimate and relative truths all the way to his inter-
pretation of final truth or buddhahood. Finally, his style of metaphysics does not adopt a purely
theoretical approach, as the common contemporary conception of metaphysics would want. It is
a metaphysics closer to Neoplatonist or a range of medieval (Latin, Arabic, or Jewish) metaphy
sical thoughts, which see metaphysics as a practical process of gradual insight bringing about
cognitive, spiritual, ethical, and emotional progress. He stresses at the end of this passage how
507
this final position should be cultivated through the well-known Buddhist process of the three
kinds of understanding: understanding through study (śrutamayī), which is purely receptive;
understanding through reflection (cintāmayī), which uses discursive analysis to deconstruct
substantial assumptions and arrive at an insight into the nature of “illusion-like” reality; and
finally understanding through habituation (bhāvanāmayī), which appropriates in one’s mental
continuum the results of the two previous understandings. Metaphysics is thus more than a theo-
retical scholastic exercise: it is what fuels the very process of the path, and it has direct practical
implications.
Conclusion
Haribhadra’s thought shows a quite remarkable breadth and force, which explains why he
exerted a powerful influence on late Indian and Tibetan Buddhist thinkers (the Ornament was
never translated in China before the twentieth century; hence he was completely unknown to
pre-modern Chinese and Japanese thinkers). Indian commentators saw his work as a game-
changer in the exegetical tradition of the Ornament, which explains why his small commentary
was in turn the object of subcommentaries by Dharmamitra (ninth century) and Dharmakīrtiśrī
(ninth to tenth century). But he was also hotly contested by important late Indian figures like
Ratnākaraśānti (ca. 970–1045 CE) due his pro-Mādhyamika interpretation and his stance on
multiple topics of the Buddhist path. Tibetan commentators of the Ornament almost univer-
sally considered him an authority, although rare exceptions like Dol po pa (1292–1361) went
as far as reviling him as an inferior commentator with a narrow understanding of Mādhyamika
philosophy.
As argued previously, Haribhadra’s work offers a double perspective worth exploring: a
comprehensive view of Buddhist philosophy and especially of the central concept of the path,
which can be extracted from an overall consideration of his various compositions, and more
focused reflections on multiple subjects ranging from hyper-specialized questions on the path
to general metaphysical and Mādhyamika topics. His thought is complex since it is formulated
within the confines of commentaries of one of the most technical texts of Buddhist philoso-
phy, and it is not always easy to distinguish between what is truly original and what is an
appropriation of the exegetical tradition. It also represents a challenge for philosophers, as it
seems to vacillate between what could be designated at times as philosophy and at other times
as theology, depending on the nature and the extent of the assumptions in play. But arguably,
the history of theology is also a part of the history of philosophy in the West, and it would be
problematic to exclude it, even for the very understanding of the evolution of philosophical
thought. Just as problematic is the neglect of Haribhadra’s works, which represent a line of
thought that has marked late Indian Buddhist philosophy decisively in shaping a comprehen-
sive and systematic account of the Buddhist path.
Notes
1 The qualifications “vast” and “deep” are Tibetan interpretative tools and are useful heuristic means
for presenting the aims of this literature. See Harter 2019b.
2 For details, see Seyfort Ruegg 1981, 101–2; Makransky 1997, 213–14; Sparham 2006, xv; Apple
2008, 29–30; Harter 2015, 11–18; Harter 2018, 204–5.
3 In addition, during the history of Buddhist philosophy, no scholar seems to be have questioned the
authorship of these three works, apart from Ngog blo ldan shes rab, who doubted that the Commen-
tary on the Versified Summary of the Precious Qualities was Haribhadra’s (Obermiller 1937, 5).
508
Haribhadra
4 The Hegelian statement seems to be an apocrypha coming from the Hegelian tradition, since no such
statement can be found in Hegel’s writings.
5 My translation. Wogihara 1932–1935, 23; Sparham 2006, 199.
6 On the topic, see chapters on Nāgārjuna, Āryadeva, Śāntarakṣita, Candrakīrti in this volume.
7 The move is not unique to Haribhadra, although it is certainly quite forceful in this passage. For
another similar “de-textualization” strategy, see the beginning of Vasubandhu’s Commentary on the
Treasury of Abhidharma (Abhidharmakośa-bhāṣya) and his definition of abhidharma.
8 This is known as the theory of the buddha-bodies (kāya), which was harshly debated by commentators
of the Ornament. See Makransky 1997 for a comprehensive overview of the Indian and Tibetan debates.
9 Some might talk of a common conception of philosophy as a way of life as in Pierre Hadot’s works,
although differences should not be underestimated. See Hadot 1995; Kapstein 2013; Harter 2018.
10 Maria Heim develops a similar idea about Buddhaghosa (Heim 2018, 3–4).
11 Wogihara 1932–1935, 158: vikalpo ’vastu-nirbhāsād visaṃvādād upaplava. The full definition can
be found in PVin 1.33ab.
12 Wogihara 1932–1935, 16.
13 Ibid., 4.
14 Ibid., 205.
15 See the whole passage in Wogihara 1932–1935, 332–49; translation in Sparham, vol. 2, 232–40.
16 See, for example, the chapter on Śāntarakṣita and Kamalaśīla in this volume.
17 See also Wogihara 1932–1935, 534; Sparham, vol. 2, 176–78.
18 Wogihara 1932–1935, 641.3: māyopāmādvayavijñānamātraprabandha.
Bibliography
Amano, Koei H. 2000. Abhisamayālaṃkāra-kārikā-śāstra-vivṛti: Haribhadra’s commentary on the
Abhisamayālaṃkāra-kārikā-śāstra edited for the first time from a Sanskrit manuscript. Kyoto: Heira-
kuji Shoten.
Apple, James. 2008. Stairway to Nirvāṇa: A Study of the Twenty Saṃghas Based on the Works of Tsong
kha pa. Albany: State University of New York Press.
Conze, Edward. 1973. The Perfection of Wisdom in Eight Thousand Lines and Its Verse Summary. Boli-
nas: Four Seasons Foundation. Reprint, San Francisco: City Lights, 2006.
———. 1975. The Large Sutra on Perfect Wisdom. Berkeley: University of California Press.
———. 1978. The Prajñāpāramitā Literature. 2nd ed. Tokyo: The Reiyukai.
Dreyfus, Georges. 2003. The Sound of Two Hands Clapping: The Education of a Tibetan Buddhist Monk.
Berkeley, Los Angeles and London: University of California Press.
Hadot, Pierre. 1995. Philosophy as a Way of Life. Malden and Oxford: Blackwell Publishing.
Harter, Pierre-Julien. 2015. “Buddhas in the Making: Path, Perfectibility, and Gnosis in the
Abhisamayālaṃkāra Literature.” PhD diss., University of Chicago.
———. 2018. “Spiritual Exercises and the Buddhist Path: An Exercise in Thinking with and against
Hadot.” In Buddhist Spiritual Practices: Thinking with Pierre Hadot on Buddhism, Philosophy,
and the Path, edited by David V. Fiordalis and Luis O. Gómez, 147–79. Berkeley, CA: Mangalam
Research Center for Buddhist Languages.
———. 2019a. “Haribhadra,” In Brill’s Encyclopedia of Buddhism. Volume II: Lives, 204–8. Leiden and
Boston: Brill.
———. 2019b. “From Exegesis to Hermeneutics: Rongtön and the Hidden Meaning (sbas don) of Per-
fect Wisdom.” In Reasons and Lives in Buddhist Traditions: Studies in Honor of Matthew Kapstein,
edited by Dan Arnold, Cécile Ducher, and Pierre-Julien Harter, 379–91. Somerville, MA: Wisdom
Publications.
Heim, Maria. 2018. Voice of the Buddha: Buddhaghosa on the Immeasurable Words. Oxford: Oxford
University Press.
Kapstein, Matthew. 2013. “Stoics and Bodhisattvas Spiritual Exercise and Faith in Two Philosophical Tradi-
tions.” In Philosophy as a Way of Life: Ancient and Modern. Essays in Honor of Pierre Hadot, edited by
Michael Chase, Stephen R. L. Clark, and Michael McGhee, 99–115. Malden and Oxford: Wiley Blackwell.
Makransky, John. 1997. Buddhahood Embodied: Sources of Controversy in India and Tibet. Albany:
Mano, Ryūkai. 1972. Genkan shōgonron no kenkyū. Tokyo: Sankibōshoten.
509
Naughton, Alexander. 1989. “The Buddhist Path to Omniscience.” PhD diss., University of Wisconsin,
Madison.
Obermiller, Eugene. 1937. Prajñā Pāramitā-Ratna-guṇa-saṃcaya-gāthā: Sanscrit and Tibetan Text.
Bibliotheca Buddhica, 29. Leningrad: Akad. Nauk.
Pollock, Sheldon. 1985. “The Theory of Practice and the Practice of Theory in Indian Intellectual His-
tory.” Journal of the American Oriental Society 105 (3): 499–519.
Sanderson, Anderson. 2009. “The Śaiva Age – The Rise and Dominance of Śaivism during the Early
Medieval Period.” In Genesis and Development of Tantrism, edited by Shingo Einoo, 41–349. Tokyo:
Institute of Oriental Culture, University of Tokyo.
Sparham, Gareth, trans. 2006–2012. Abhisamayālaṃkāra with Vṛtti and Ālokā, Vṛtti by Vimuktisena,
Ālokā by Haribhadra, vols. 1–4. Fremont: Jain Publishing Company.
Wogihara, Unrai. 1932–1935. Abhisamayālaṃkār’ālokā Prajñāpāramitāvyākhyā, Commentary on the
Aṣṭasāhasrikā-Prajñāpāramitā by Haribhadra together with the text commented on. Tokyo: The
Toyo Bunko.
Yuyama, Akira. 1976. Prajñā-pāramitā-ratna-guṇa-saṃcaya-gāthā: Sanskrit Recension A, Edited with
an Introduction, Bibliographical Notes, and a Tibetan Version from Tunhuang by Akira Yuyama.
Cambridge and New York: Cambridge University Press.
510
31
ŚĀNTIDEVA
Virtue on the Empty Path of the Bodhisattva
Stephen Harris
Śāntideva was an eighth-century CE Mahāyāna Indian Buddhist monk. He was the author of
two surviving texts, the Introduction to the Practice of Awakening (Bodhicaryāvatāra; here-
after BCA) and the Compendium of Trainings (Śikṣāsamuccaya; hereafter ŚS). Śāntideva’s
philosophical influence is most significant in the field of moral philosophy, with each text
making substantial contributions to the study of virtue, well-being and the relation between
metaphysics and ethics. The ninth chapter of the BCA also presents an influential exposition
and defense of the core Madhyamaka metaphysical tenet of emptiness (śūnyatā).
Life and Major Works

Very little is known about Śāntideva’s life. He is generally accepted to have lived between the
late seventh and mid-eighth century CE. Traditional Tibetan biographies claim that he spent
much of his career at the Indian monastic university of Nālandā. Both the BCA and the ŚS
are training manuals for the development of the bodhisattva, the being who takes a vow to
achieve full buddhahood in order to liberate all sentient beings. The influence of the shorter
text, the BCA, is attested to by the existence of numerous Indian and Tibetan commentaries
(Williams 1998, 4–5). Recently, it has undergone a resurgence of popularity, with a number
of influential contemporary practitioners of Tibetan Buddhism, including the fourteenth Dalai
Lama and Pema Chödrön, writing commentaries on it (Gyatso 1994; Chödrön 2007). The last
several decades have also seen an increase in academic attention to the BCA; it is often char-
acterized as the most significant text for the study of Indian Buddhist ethics (Goodman 2009;
Garfield 2012). By contrast, until recently, the ŚS received relatively little academic attention
and almost no philosophical study. A new translation of the ŚS by Charles Goodman (2016),
as well as several recent important studies (Mahoney 2002; Clayton 2006; Harrison 2007; Lele
2007; Mrozik 2007), have increased its academic visibility.
The received version of the BCA is composed of ten chapters that total 913 verses. This
recension has been the subject of all extant commentaries and has been translated into Eng-
lish recently a number of times (Śāntideva 1997, 2006, 2008). A shorter version of the BCA,
composed of nine chapters and 701.5 verses, is preserved in Tibetan translation but has not yet
been published (Saito 1993). Most contemporary scholarship takes the longer text as its object
of study, as do I in this chapter.
DOI: 10.4324/9781351030908-44 511

Stephen Harris
The primary purpose of both the BCA and the ŚS is to instruct the reader in developing
the virtuous qualities of the bodhisattva. These include the six perfections ( pāramitās) of
generosity (dāna), ethical restraint (śīla), patience (kṣānti), energetic effort (vīrya), concentra-
tion (samādhi) and wisdom ( prajñā), as well as closely associated virtues such as compas-
sion (karuṇā), mindfulness (smṛti) and introspection (saṃprajanya). A significant difference
between the two texts is that the BCA is presented in Śāntideva’s own voice and is organized
around the six perfections ( pāramitās). By contrast, the ŚS is largely composed of quota-
tions from Mahāyāna Buddhist sutras. It presents its insights through a frame of twenty-seven
verses written by Śāntideva, articulating the central training points of the bodhisattva path,
which are interspersed throughout the text’s body. Each verse is further developed with occa-
sional commentary by Śāntideva and through quotations from Mahāyāna sutras which consti-
tute most of the text’s length.
As Susanne Mrozik argues, the ŚS should not be looked on merely as an anthology of
Mahāyāna sutras; Śāntideva’s organization of the sutra quotations, as well as his brief verse
and prose commentary, allow a distinctive authorial voice to emerge (2007, 10–12). Neverthe-
less, we cannot assume that every line of the sometimes lengthy sutra quotations expresses a
view which Śāntideva would fully endorse. For this introduction, therefore, I focus primarily
on the shorter and more influential BCA. Nevertheless, the ŚS treats certain topics in much
greater depth than the BCA, making it invaluable in understanding his thought. I draw upon
it in my explanation of Śāntideva’s presentation of generosity (dāna), a topic which receives
much less attention in the BCA than it does in the ŚS.
In this chapter, I focus on four philosophically significant topics treated in the BCA and
the ŚS: bodhicitta and the bodhisattva vow; the perfection of generosity; the perfection of
patience; and the perfection of wisdom, with emphasis on its role in the development of the
other virtues. In terms of this last topic, I give particular attention to an influential argument in
the BCA’s eighth chapter, in which Śāntideva links acceptance of the nonexistence of a self to
altruistic commitment. In addition, one of the most philosophically stimulating claims made in
the BCA is that adopting the apparently demanding bodhisattva path benefits the bodhisattva.
Therefore, in this chapter, I also explore how altruism and self-interest converge as a result of
Śāntideva’s development of the perfections.1 This in no way exhausts Śāntideva’s philosophi-
cal contributions; nevertheless, exploring these themes will serve as an introduction to much
of what is most distinctive about his ethical thought.2
As we examine these positions, however, we should keep in mind that the BCA and ŚS are not
merely theoretical texts meant to provide a conceptual understanding of virtue and moral motiva-
tion. Rather, both are meditation manuals whose purpose is to aid in the development of virtue
for those engaged on the bodhisattva path. Most of the characterizations of virtue presented in the
texts, as well as the arguments Śāntideva provides to defend his ethical positions, also function
as meditations designed to develop virtue and moral commitment. This is the case, for instance,
in Śāntideva’s argument linking the metaphysical position of selflessness (ānatman) with a com-
mitment to impartial benevolence, which I will consider below. It also explains Śāntideva’s
predilection for providing numerous distinguishable meditations and arguments for developing
a given virtue. Philosophical understanding plays an important role in moral development for
Śāntideva, but theoretical knowledge for its own sake is not his primary aim.
Bodhicitta and the Bodhisattva’s Vow

A central concept in both texts is the bodhisattva, the being (sattva) who vows to achieve the
full awakening (bodhi) of a buddha in order to most effectively work for the benefit of sentient
512
Śāntideva
beings. The motivation to attain this goal is bodhicitta, the mind (citta) dedicated to full awak-
ening (bodhi). The first four chapters of the BCA are focused on bodhicitta: the first praises
its virtuous qualities; the second and third present a ritual in which it is strengthened and the
bodhisattva vow is taken; the fourth focuses on maintaining it. These early chapters stress the
dangerously unstable state we are in prior to entering the bodhisattva path. Śāntideva intro-
duces this theme in a well-known pair of verses at the beginning of the first chapter.
Just as lightning illuminates the darkness of a cloudy night

for an instant, in the same way, by the power of the Buddha,
occasionally people’s minds are momentarily inclined toward merit.
Thus, virtue is perpetually ever so feeble, while the power of

vice is great and extremely dreadful. If bodhicitta did not exist,
what other virtue would overcome it?
(BCA 1:5–6, translation modified)3
The verses make two striking and philosophically significant points. First, the minds of those
outside the bodhisattva path are depraved, unable to focus on virtuous development with any con-
sistency. Vice ( pāpa) here indicates the mental afflictions (kleśas), in particular craving (lobha),
anger (krodha) and delusion (moha). Our dispositions (anuśaya) for their arising are so strong
that consistent virtuous activity is impossible. Second, bodhicitta itself is the best, or perhaps the
only, solution to our predicament. This is surprising, since developing bodhicitta and undertak-
ing the bodhisattva path requires remaining in saṃsāra for eons, during which time one per-
fects generosity by sacrificing one’s body (BCA 7:20–22) and willingly taking painful rebirths
(BCA 8:107–108) to work for the benefit of beings. Yet this claim, of the self-beneficial power of
bodhicitta and the virtues of the bodhisattva, is a frequent and repeated theme within the BCA:
Those who long to overcome the abundant miseries of mundane existence,

those who wish to dispel the adversities of sentient beings,
and those who yearn to experience a myriad of joys should never forsake bodhicitta.
(BCA 1:8, translation altered)
The second line gives the expected advice that those dedicated to eliminating the suffering of
sentient beings should adopt the bodhisattva path. Developing the perfections, like generosity,
patience and so on, will enable the bodhisattva to effectively intervene for their aid. Surpris-
ingly, the verse also claims that those pursuing their own well-being, both in terms of libera-
tion from saṃsāra and ordinary saṃsāric happiness, should also develop bodhicitta. Claims
like this are repeated forcefully in the eighth chapter, which emphasizes the self-beneficial
aspects of compassion:
All those who are unhappy in the world are

so as a result of their desire for their own
happiness. All those who are happy in the world are so
as a result of their desire for the happiness of others.
(BCA 8:129)
The central concepts which make Śāntideva’s claimed convergence between altruism and self-
interest intelligible are the mental afflictions (kleśas) and a set of virtuous mental states called
513
Stephen Harris
the perfections ( pāramitās). The most important mental afflictions, for Śāntideva and many
Mahāyāna authors, are craving (lobha), anger (krodha) and delusion (moha). For Śāntideva’s
Madhyamaka school, delusion refers most centrally to the superimposition of intrinsic existence
(svabhāva), an essence which constitutes non-relational identity, upon impermanent and depend-
ently originated phenomena. Attachment and anger are the affective states of attraction and repul-
sion, which are generated as a result of this superimposition. Śāntideva also recognizes a number
of other mental afflictions – including laziness (ālasya), jealousy (īrṣyā) and pride (māna) –
which contaminate our experience. The mental afflictions are the deep causes of our suffering,
inciting harmful actions and generating negative karma, which leads to unfortunate rebirths.
The perfections ( pāramitās) are six virtuous mental states which many Mahāyāna
authors, including Śāntideva, present as the central qualities developed by the bodhisattva.
They are generosity (dāna), ethical restraint (śīla), patience (kṣānti), energetic effort (vīrya),
concentration (dhyāna) and wisdom ( prajñā). Their primary role is to eliminate the mental
afflictions (kleśas). Energetic effort, for instance, acts as the antidote to laziness by providing
the bodhisattva with the motivational force to progress along the path and, at its conclusion,
to tirelessly work for the benefit of sentient beings. The perfections, then, have two inter-
related functions. First, they eliminate the mental afflictions and therefore remove all deep
sources of suffering for their possessor. This is why progressing on the bodhisattva path,
which is largely constituted by their development, is in the individual’s own benefit. Second,
they enable the bodhisattva to work to liberate sentient beings from suffering. Developing
the perfections, therefore, is the most effective way to pursue one’s own well-being, as well
as the well-being of others.
These early chapters also present a ceremony called “The Ritual of Supreme Worship
(anuttara-pūjā),” which constitutes the entirety of BCA chapters two and three. In this cer-
emony, the bodhisattva makes offerings to the buddhas and bodhisattvas and confesses past
wrongdoings as a way of preparing to take the bodhisattva vow. Powerful expressions of com-
passion are made directly before the vow is taken.
May I be a protector for those who are without protectors, a

guide for travelers, and a boat, a bridge, and a ship for those
who wish to cross over.
May I be a lamp for those who seek light, a bed for those
who seek rest, and may I be a servant for all beings who
desire a servant.
(BCA 3:17–18)
The Bodhisattva Vow follows shortly afterwards:
Just as the Sugatas of old adopted bodhicitta

and just as they properly conformed to the practice of the
bodhisattvas,
So I myself shall generate bodhicitta for the

sake of the world; and so I myself shall properly engage in
those practices.
(BCA 3:22–23, translation altered)
514
Śāntideva
The incorporation of ritual activity in a philosophical text may at first seem incongruous,
but it illustrates that Śāntideva’s primary commitment in the BCA is to providing practices
which will develop virtue and moral commitment in practitioners on the bodhisattva path. For
Śāntideva, ritual activity complements clear intellectual understanding by developing a feeling
of joyous confidence ( prasāda) in the efficacy of Buddhist trainings. Moreover, both the ritual
and the taking of the vow itself embed the bodhisattva in an infinitely long lineage of buddhas
and bodhisattvas who have previously traveled on the bodhisattva path. Both thereby aid the
bodhisattva in adopting a particular kind of conventional identity, grounded in the central pro-
ject of liberating sentient beings from suffering. Philosophical argumentation and ritual activ-
ity are each kinds of cultivation (bhāvanā) by which to develop the virtues of bodhisattvahood.
The bodhisattva vow, expressing the bodhisattva’s infinite commitment to remain in saṃsāra
forever to work for the benefit of beings, also indicates a creative tension running throughout
the entire text. With his promise, the bodhisattva binds all his future lives into a shared project
of liberating the world from suffering. Nevertheless, Śāntideva later explicitly acknowledges
that future lives are not numerically identical to the current one (BCA 8:98), presumably given
their lack of psychological and physical continuity (Harris 2018). Moreover, the bodhisattva
himself, and all those he liberates, are empty of intrinsic existence. The bodhisattva’s infinite
commitment, therefore, is open to charges on multiple fronts that it is inconsistent with Bud-
dhist commitments to selflessness (anātman) and impermanence (anitya). Śāntideva confronts
this tension in the ninth chapter, which is dedicated to developing wisdom ( prajñā).
[Qualm:] If no sentient being exists, for whom is there

compassion?
[Mādhyamika:] For one who is imagined through delusion,

which is accepted for the sake of the task.
[Qualm:] If there is no sentient being, whose is the task?
[Mādhyamika:] True. The effort, too, is due to delusion.

Nevertheless, in order to alleviate suffering, delusion with
regard to one’s task is not averted.
(BCA 9:75–76)
Here, Śāntideva’s opponent objects that the metaphysical selflessness of the bodhisattva and
those he liberates makes his task incomprehensible. Śāntideva’s response is to claim that delusion
itself can be embraced to reestablish the existence of persons and provide a provisional ground for
compassion. The term translated as delusion is moha, a synonym for avidya, the superimposition
of intrinsic existence (svabhāva) upon empty (śūnya) phenomena. It is striking to find Śāntideva
invoking it as having at least provisional benefit for the bodhisattva’s progression along the path.
These verses suggest that Śāntideva is less interested in resolving the tension between emp-
tiness, impermanence and the bodhisattva’s multi-life commitment than he is in manipulating
both the metaphysical fact of emptiness, and our habitual tendencies towards reification and
superimposition, for liberative benefit. Here, Śāntideva invokes delusional self-existence to pro-
vide a basis for compassion to arise. Elsewhere, he offers meditations upon the nonexistence of the
self (anātman) to dissolve selfishness (BCA 8:101–103) and eliminate anger (BCA 6:22–41).
We also find him turn anger against itself (BCA 8:41) and invoke terror of death to provide
515
Stephen Harris
motivational energy to enter into the bodhisattva path (BCA 2:28–66) and develop the perfec-
tion of energetic effort (BCA 7:4–13) (Harris 2017). I return to Śāntideva’s strategic employ-
ment of the realization of selflessness when I discuss the perfection of wisdom subsequently.
In the next two sections, which examine Śāntideva’s development of generosity and patience,
I explore more systematically another of Śāntideva’s overarching themes in the BCA, the con-
vergence of altruistic commitment and self-interest as one progresses along the bodhisattva path.
Generosity (dāna)
Although generosity is not given its own chapter in the BCA, Śāntideva characterizes it in
several important verses dispersed throughout the text. A much more extensive treatment is
provided in the ŚS, which I also draw upon in this section. For Śāntideva, generosity is a
mental state which fully renounces all items in one’s possessions; it does not primarily refer to
the physical distribution of goods. This is stated most clearly in a verse in the BCA, in which
Śāntideva replies to a hypothetical opponent who claims that the existence of poverty implies
no previous buddhas or bodhisattvas perfected generosity.
If the perfection of generosity makes the world free of

poverty, how is it possible that the Protectors of the past
acquired it, when the world is still impoverished today?
The perfection of generosity is interpreted simply as a state

of mind due to the intention of giving away everything,
together with the fruits of that, to all people.
(BCA 5:9–10)
Śāntideva’s response is to explain that the perfection of generosity is a mental quality rather
than an act of giving which reduces poverty. It is the state of mind which abandons (tyāga)
everything to sentient beings. Giving, therefore, is initially and primarily the elimination of
craving. Moreover, the perfection of generosity is total; it is the willingness to give everything
to others. In the ŚS, the totality of generosity is indicated by listing three kinds of objects
which the bodhisattva renounces: his possessions, his body and mental states and his karmic
merit (Goodman 2016, 20; see also Mahoney 2002; Clayton 2006). Together, these items con-
stitute the entire experience of the bodhisattva, all of which is offered to sentient beings.
Another distinctive feature of Śāntideva’s characterization of generosity is its self-benefi-
cial nature. This is stated forcefully in the BCA, where generosity is equated with liberation.
Abandoning everything is nirvāṇa, and my mind seeks

nirvāṇa. If I must abandon it, it is
better that I give it to sentient beings.
(BCA 3:11, translation modified)
This striking verse presents a three-fold equation between abandoning (tyāga), giving (dāna)
and nirvāṇa. Linking abandoning with nirvāṇa is relatively natural; nirvāṇa is the cessation of
the mental afflictions (kleśas), and these arise in relation to possessions, as a result of attach-
ment (tṛṣṇā/rāga) to them, aversion ( pratigha) to losing them, jealously (īrṣyā) of another
person possessing them and so on. A practitioner first abandons possessions by becoming a
monastic, then abandons mental afflictions through meditational cultivation (bhāvanā) and
516
Śāntideva
ultimately abandons the false sense of self. Abandoning, in all these senses, is in the deepest
interest of the individual; its perfection is individual liberation.
Remarkably, in 3:11, Śāntideva extends this equivalence between abandoning and nirvāṇa
to the first of the Mahāyāna perfections, generosity (dāna). Giving is, first and foremost, aban-
doning. One can abandon physical possessions by giving them to others and diminish crav-
ing in so doing (Lele 2007). More significantly, as we saw previously, perfect generosity is
itself the mental attitude which abandons craving towards possessions, even when they remain
within our control. Since giving for Śāntideva is total, encompassing all mental and physical
experience, perfect generosity is equivalent to the complete abandonment of craving, which is
nirvāṇa. Giving is, therefore, a self-protecting and beneficial virtue.
This characterization of the perfection of generosity as the elimination of craving is devel-
oped more extensively in the ŚS, where one of Śāntideva’s favorite Mahāyāna sutras to quote is
the Questions of Ugra (Ugraparipṛcchā-sūtra). This text is ideal for guidance about generosity,
since many of its instructions are directed towards household bodhisattvas who remain legal
owners of property even while dedicating themselves to awakening. Such a situation provides
the ideal opportunity to consider what attitude a bodhisattva should take towards possessions.
For thus, leading merchant, a household bodhisattva should not develop a sense of
grasping or a sense of “mine” toward any property, nor clinging, nor addiction, nor
the propensity to craving.
(Goodman 2016, 23, translation slightly altered)
Notice that the household bodhisattva is not instructed to physically abandon or give away his
possessions but instead eliminates grasping ( parigraha) and his sense of ownership (mamat-
vam) towards property. A quotation from the Questions of Nārāyaṇa sutra repeats this advice
but also brings into view the connection between generosity as the elimination of grasping,
with giving in its ordinary sense of transferring items to those in need:
Don’t appropriate anything if you can’t arouse the thought of giving it away. Don’t
own any property if you can’t arouse the thought of giving it up. . . . A bodhisattva
great being should arouse the following thought: “I have given away and offered
this, my body, to all sentient beings; how much more so any external property! If any
sentient being needs anything for any reason, whatever it might be, if it exists, I will
give it to that being. I will give my hand to whoever needs a hand. I will give my foot
to whoever needs a foot. I will give my eye to whoever needs an eye. I will give my
flesh to whoever needs flesh. I will give my blood to whoever needs blood. . . . What
need is there even to speak of external property, such as money, grain, gold, silver,
gems, jewelry.
(Goodman 2016, 25)
The bodhisattva is instructed to refrain from acquiring material goods until they are able to
relate to them as if they are not the owner and therefore can give them up without hesitation
or regret. As we saw in the last quotation, this is accomplished through abandonment (tyāga)
of any sense of grasping ( parigraha) or ownership (mamatvam) towards possessions. In the
current quotation, these insights are also applied to giving in its ordinary sense of the transfer
of property to those in need. Based upon our prior abandonment of any sense of ownership, we
can effortlessly give items away. Giving, therefore, emerges as a two-step process: first, mental
appropriation (upādāta) and all forms of craving are eradicated, and second, physical property
517
Stephen Harris
is transferred. What is essential to understand about Śāntideva’s account of giving is that it is

the first, rather than the second, of these steps that is the perfection of generosity. Certainly, the
bodhisattva wants to benefit others through giving away possessions, but for this to be deeply
effective, perfect giving must have already taken place, privately, by mentally turning over
everything to sentient beings in general. Perfect gifts are made to nobody in particular, and
this is why it is most natural to think of them initially and primarily as acts of abandonment.
In Śāntideva’s characterization of generosity, we find several distinctive features which
reoccur in many of his descriptions of the virtues. First, generosity, and indeed all the perfec-
tions, are mental states. Śāntideva interprets the perfection of ethical discipline as the elimi-
nation of the attitude which wishes to harm others (BCA 5:11) and defines the perfection of
patience as the mind that is free of anger (BCA 5:12). The other perfections, of energetic effort,
concentration and wisdom, are all naturally understood as mental. In his bodhisattva manu-
als, Śāntideva’s primary concern is to develop the state of the bodhisattva’s mind; only once
mental training is complete can systematic effort be made to intervene in the lives of others.
Second, Śāntideva’s account of generosity is structured to be both self-beneficial and altru-
istic. Generosity benefits oneself, because it is the abandonment of craving, which is one of
the deep sources of suffering. Once generosity as a mental attitude is perfected, however, it
is then possible for the bodhisattva to intervene in the world skillfully by gifting possessions.
It is this two-fold structure of self-benefit and altruism, repeated in his development of many
of the perfections, which explains how Śāntideva can characterize the apparently demanding
bodhisattva path as being conducive to one’s own well-being. We will find this same structure
again in his characterization of patience, to which we now turn.
Patience (ksānti)
The sixth chapter of the BCA is devoted to developing two forms of patience (kṣānti). Early
in the chapter, Śāntideva offers meditations to develop patience in the form of the ability to
endure great pain without mental distress (BCA 6:12–21). The remainder of the chapter pre-
sents meditations which allow the bodhisattva to prevent anger from arising when others harm
us.4 To understand Śāntideva’s characterization of these two forms of patience, it helps to keep
in mind the relation between the perfections ( pāramitās) and the mental afflictions (kleśas).
It is the mental afflictions which both cause one’s own suffering and prevent the bodhisattva
from working effectively for the well-being of others. Likewise, it is the perfections which
eliminate the mental afflictions and in so doing protect the bodhisattva’s well-being and allow
him to work for others’ welfare.
In developing meditations to develop both forms of patience, Śāntideva recognizes that
certain events, such as physical pain and abuse from others, reliably trigger the arising of the
mental affliction of anger (krodha/dveṣa). This stimulates further mental afflictions, creates
negative karma, causes us to harm others in retaliation and simply hurts. In these ways, anger
increases our own suffering and prevents us from working for the well-being of others. Situ-
ations in themselves, however, no matter how severe, do not inevitably cause us to react with
anger. Chapter six, therefore, offers meditations to change our attitude towards painful and
unpleasant situations as a way of preventing anger from arising (Bommarito 2014). Under-
standing patience as the mental factor which protects against anger by changing our perspec-
tive on difficult situations shows that the two kinds of patience are subtypes of a single mental
state. Physical pain and abuse by sentient beings are reliable triggers of anger, and patience
intervenes by altering our attitude towards these situations so that anger no longer arises.
518
Śāntideva
Patience as Enduring Pain

Śāntideva uses two methods to develop the kind of patience which allows one to endure physi-
cal pain without distress. Here he combines the strategy of altering our perspective on painful
situations with the recognition that over time we become habituated to, and suffer less from,
physical pain. Ultimately these are not separate strategies, however; becoming habituated to
a painful sensation without adverse mental response is itself a way of taking an equanimous
attitude towards the sensation. This type of patience is vital for the bodhisattva’s development.
The bodhisattva path, especially at the later stages, requires endurance of vast amounts of pain
as the bodhisattva willingly gives up his body and life to benefit sentient beings (BCA 7:20–22)
and takes rebirth in hell realms to work for the beings there (BCA 8:107–108). Moreover,
Śāntideva has claimed that the bodhisattva path is conducive to one’s own well-being. Patience
in its form of enduring pain, therefore, must be developed to an extremely high degree, which
will allow the bodhisattva to endure any grade of physical pain without mental distress.
In the verses on becoming habituated to pain, Śāntideva points out that we already fre-
quently experience unpleasant physical sensations without psychological distress.
There is nothing whatsoever that remains difficult as one

gets used to it. Thus, through habituation with slight
pain, even great pain becomes bearable.
Do you not consider the pain of bugs, gadflies, and

mosquitoes, of thirst and hunger, and the irritation of a
serious rash and the like as insignificant?
Cold, heat, rain, wind, traveling, illness, captivity, and

beatings should not induce a sense of fragility. Otherwise,
the distress becomes greater.
(BCA 6:14–16)
Verse 6:14 begins by suggesting that over time we may become used to painful sensations, and
supports this position in 6:15, by reference to irritations like bug bites and skin rashes which
many of us endure with ease. Śāntideva then suggests developing these same attitudes to more
serious hardships, since adding mental distress to physical pain only increases overall suffering.
The second part of the technique for developing the ability to endure pain is to alter one’s
perspective on it, recognizing that its badness is minor or even that the experience of pain has
positive elements. This is dependent on the recognition that the mental affliction of anger,
rather than physical pain, is the real cause of suffering. One of the ways Śāntideva encourages
us to change our perspective on painful physical sensations is by considering its relation to
our overall goals (Bommarito 2014). He sometimes employs everyday examples outside of
religious life to illustrate this strategy, as in this example from chapter four.
If fishermen, outcasts, farmers, and others, whose minds are

fixed merely on their own livelihoods, withstand the adversities
of cold and heat, then why do I not endure for the sake
of the well-being of the world?
(BCA 4:40)
519
Stephen Harris
This verse combines the two strategies that Śāntideva recommends to develop the ability to
endure pain. First, fishermen and other kinds of workers experience discomfort related to
their occupations on a daily basis and slowly become accustomed to them. More importantly,
however, these discomforts are willingly endured because the workers are intent on a valued
goal, here the earning of their livelihood. This changes their perspective on pain, since it is
seen as a necessary condition of pursuing the valued item. Of course, the bodhisattva’s pain
will eventually be much greater, when he offers his life for others, but the goal of liberating all
sentient beings is much greater as well, a fact which Śāntideva reminds his listener of in the
last line of the verse.
Śāntideva repeats this point in chapter six, but also suggests the possibility that patience can
eliminate all adverse reaction to pain:
Some, seeing their own blood, show extraordinary valor,

while some faint even at the sight of others’ blood.
That comes from mental fortitude or from timidity.

Therefore, one should become invincible to suffering, and
surmount pain.
Not even in suffering should a wise person disrupt his

mental serenity, for the battle is with the mental afflictions;
and in battle pain is easily obtained.
(BCA 6:17–19)
There are several important points being made in these verses. First, Śāntideva again empha-
sizes that contemplating the value of a goal will alter our attitude towards physical pain which
must be endured in its pursuit. If we like, we can substitute contemporary examples like the
running of a marathon for the warriors’ endurance of battlefield suffering. Second, verse 6:18
indicates that it is our mental attitudes of fortitude (dṛḍhatva) or timidity (kātaratva) which
determine our ability or inability to withstand pain. Over time, these mental responses can be
trained, so that we can endure physically painful situations with tranquility. Verses 6:18 and
6:19 also indicate the radicalness of Śāntideva’s position. With the proper training and perse-
verance, we can become invincible to any grade of physical pain.
The approach so far has been to emphasize that pain is insignificant, at least in relation to
the goal of buddhahood. In the following verse, Śāntideva takes this strategy a step further,
claiming that pain itself may have positive aspects.
Suffering has another quality since

arrogance diminishes because of despair,
and one feels compassion for beings in the
cycle of existence, fear of sin, and a
yearning for the Jina.
(BCA 6:21)
Feeling pain has great value in relation to the bodhisattva’s goals, since it aids in developing
compassion, by helping us to empathize with others who have experienced similar kinds of
discomfort. Likewise, it stimulates effort (vīrya) by motivating us to deeply commit to the
bodhisattva path, which will eliminate all suffering.
520
Śāntideva
The results of these meditations is a reevaluation of physical pain, which is recognized

as relatively innocuous and in some cases can be conducive to progress on the bodhisattva’s
path. This does not eliminate the physical sensations of pain but alters our perspective on them
so that we are less averse to experiencing them. Therefore, anger, which is the real source of
disvalue, does not arise. It is worth emphasizing that Śāntideva also instructs the bodhisat-
tva to preserve their bodies in ordinary circumstances as instruments for progressing on the
bodhisattva’s path (BCA 5:70). Although eventually the bodhisattva must be able to endure
great pain without distress, this should not be interpreted as a senseless masochism in which
pain is sought out for its own sake.
Patience Towards Others’ Wrongdoings

Śāntideva also employs the strategy of altering our perspective on unpleasant situations when
he presents meditations designed to develop patience towards those who harm us. In these
meditations, Śāntideva argues that anger is always irrational, both in that it conflicts with our
wish to be happy and that it arises from mistaken views about reality (Lele 2007). These verses
should be read both as arguments for the badness of anger and as meditations by which to dis-
solve it. Early in the chapter, Śāntideva emphasizes the mental discomfort which anger causes.
The mind does not find peace, nor does it enjoy pleasure
and joy, nor does it find sleep or fortitude when the thorn
of hatred dwells in the heart.
(BCA 6:3)
The irrationality of anger is also argued for by pointing out that it is always a needless addi-
tional suffering, since resolutions of difficult situations can be sought without anger (BCA
6:10), and that it causes social unrest (BCA 6:4–5). Meditations like these do not depend
on accepting Buddhist presuppositions. Other meditations, however, rely on the efficacy of
karma, such as Śāntideva’s claim that anger can cause one to be reborn in hell:
If one is unable to endure even this slight suffering of the

present, then why does one not ward off anger which is the
cause of pain in hell?
Thus, solely due to anger I have brought myself into hells

thousands of times, and I have not brought about benefit for
myself or others.
(BCA 6:73–74)
These verses illustrate a reoccurring feature of the BCA. Śāntideva often presents basic psy-
chological or social observations as a way of beginning to develop virtue and dissolve mental
afflictions. These initial insights are then magnified through the incorporation of Buddhist cos-
mological and karmic presuppositions. Anger should be abandoned because it keeps us awake
at night (BCA 6:3) and disrupts social relations (BCA 6:4–5) but even more so because it will
cause us to be reborn in hell (BCA 6:73–74).
When working with Śāntideva’s texts philosophically and interculturally, therefore, it can
be helpful to recognize that arguments depending on cosmological and karmic presupposi-
tions will often be expanding on a central insight which may also be presented in a more
521
Stephen Harris
interculturally available fashion. Nevertheless, we lose something vital if we do not pay atten-
tion to the karmic and cosmological aspects of Śāntideva’s work. They illustrate the vastness
of his project and the extraordinary degree to which he develops the virtues. Generosity and
patience are total; the bodhisattva gives away everything, including his life and body, and
endures any amount of physical pain without discomfort. Moreover, he develops compassion
so strongly that he remains in saṃsāra forever.
We have seen that Śāntideva draws attention to positive aspects of physical pain as a way of
developing the bodhisattva’s ability to endure it. He employs the same strategy in developing
patience towards sentient beings’ wrongdoing:
For a supplicant is not a hindrance to generosity at the time

of almsgiving; and when a person who bestows an ordination
arrives, he is not called a hindrance to the ordination.
Beggars are easy to meet in the world, but malefactors are

difficult to find, for no one will wrong me when I do no
wrong.
Therefore, since my adversary assists me in my

bodhisattva way of life, I should long for him
like a treasure discovered in the house and
acquired without effort.
(BCA 6:105–107)
The verses point out that the bodhisattva has committed to the extraordinarily valuable goal of
developing the perfections and achieving full buddhahood for the benefit of all. This requires
developing patience, since patience is one of the perfections, and this in turn requires ene-
mies whose actions give the bodhisattva the opportunity to develop it. Therefore, the attitude
towards such persons should be gratitude, rather than anger.
All the meditations surveyed thus far can be understood as parts of Śāntideva’s extended
argument for the irrationality of anger. Perhaps the most obvious defense of this claim, how-
ever, comes in a set of verses in which Śāntideva emphasizes the incompatibility between
anger and the dependently originated nature of phenomena.
I am not angered at bile and the like even though they cause
great suffering. Why be angry at sentient beings, who are
also provoked to anger by conditions?
(BCA 6:22)
Śāntideva’s argument continues for several verses, and there has been scholarly disagreement
as to how we should understand it. The general point is that accepting that beings are subject
to innumerable causes and conditions entails accepting as well that our enemies are influenced
by numerous factors beyond their control. Recognizing this will dissolve our anger towards
them, since they are not deeply responsible for their actions. Some contemporary commen-
tators have understood Śāntideva in these verses to be endorsing a kind of incompatibilism
in which moral responsibility is determined to be incompatible with dependent origination
(Goodman 2002). A weaker reading is to see Śāntideva as merely drawing attention to the
relatively powerless position of our enemies who are dominated by the mental afflictions. Both
522
Śāntideva
readings entail that anger towards our enemies is an irrational response. This in turn suggests
that Śāntideva, like the Stoics, held at least a weak cognitivism, in claiming that emotions like
anger are highly responsive to our judgments. Moreover, it is a distinctive position of Bud-
dhists and Stoics alike that the emotions are highly malleable and with long-term training can
wholly be brought under our control (McRae 2015).
There is, however, an obvious contrast between Śāntideva’s position and authors like Aris-
totle who claim that anger is natural and in the right circumstances is an essential component
of a flourishing life. Likewise, some philosophers have emphasized that circumstances of
injustice may require anger as an appropriate moral response (Nussbaum 1994, cf. Vernezze
2008). By contrast, Śāntideva claims that anger is always harmful and seems to have little
concern for justice. This is largely because anger causes suffering, and like virtually all Indian
Buddhist ethicists, Śāntideva places the removal of suffering at the center of his moral project.
It suggests as well that at least a certain strand of deontic thought will have little traction with
Śāntideva. If moral indignation increases the overall suffering of a particular situation, then it
must be resisted. Nevertheless, this should not be mistaken for a moral passivity for the suf-
fering of those in need. Rather, it demonstrates another of Śāntideva’s central commitments,
impartial benevolence, manifested in an equal-minded concern for all, where this includes
both the victims of aggression and the aggressors who will suffer horribly from the negative
consequences of their anger.
Wisdom ( prajñā) and the Relation Between Metaphysics and Ethics

Chapter nine of the BCA, dedicated to developing wisdom ( prajñā), provides an extended
defense of the core Madhyamaka metaphysical tenet of emptiness (śūnyatā). Although
Śāntideva does not explicitly define emptiness in the text, he appears to accept the standard
Madhyamaka position that wisdom is the realization that all phenomena are empty (śūnya) of
intrinsic existence (svabhāva). This means that entities do not bear their own identity conditions
but exist only in dependence upon their causes and conditions and upon conceptual imputation.
Most of the chapter presents a series of objections brought by philosophical opponents, includ-
ing Abhidharma Buddhists (BCA 9:38–53), Yogācāra Buddhists (BCA 9:17–35) and other phil-
osophical schools such as Sāṃkhya (9:60–67) and Nyāya (9:68–69). The chapter is a valuable
contribution to the earlier expositions and defenses of Madhyamaka metaphysics by authors
such as Nāgārjuna and Candrakīrti. Nevertheless, Śāntideva’s arguments largely support posi-
tions already established by these earlier authors, and I am not able to survey them here.
What is perhaps more distinctive about Śāntideva’s approach is how he integrates various
forms of wisdom, including the realization of emptiness (śūnyatā), selflessness (anātman)
and dependent origination ( pratītyasamutpāda), into his development of the other perfections
( pāramitās). We saw one example of this in the last section, in which he develops patience by
arguing that dependent origination entails the irrationality of anger. A similar argument occurs
in the eighth chapter, in which Śāntideva appeals to the core Buddhist tenet of selflessness
(anātman) in arguing that prioritizing one’s own well-being is irrational. This argument has
attracted considerable contemporary philosophical interest. It opens by drawing attention to
the similarity of all beings in their desire to be happy and free from suffering.
One should first earnestly meditate on the equality of oneself

and others in this way: “All equally experience suffering and
happiness, and I must protect them as I do myself.”
(BCA 8:90)
523
Stephen Harris
When happiness is equally dear to others and

myself, then what is so special about me that I strive after
happiness for myself alone?
(BCA 8:95)
When fear and suffering are equally abhorrent

to others and myself, then what is so special about
me that I protect myself but not others?
(BCA 8:96)
As with the passages in the patience chapter, these verses should be understood both as med-
itations designed to increase our compassion and as premises in Śāntideva’s argument for
impartial benevolence. They act as meditations, since reflecting on the fact that others’ wish
for happiness and freedom from suffering stimulates compassion for them. But Śāntideva also
challenges his egoistic opponent to provide a distinction (viśeṣa) which justifies giving greater
priority to his own well-being. Since the wish for freedom from suffering is the same for all,
one’s own desire to be free of it cannot provide the needed justification.
The remainder of the argument consists of a series of potential distinctions offered by the
egoist to justify his self-interest, each of which is in turn dismissed by Śāntideva. The most
powerful of these reasons for egoistic behavior is that this particular pain belongs to me.
Śāntideva dismisses this response in a trio of verses which conclude his argument.
The continuum of consciousness, like a series, and the

aggregation of constituents, like an army and such, are
unreal. Since one who experiences suffering does not exist,
to whom will that suffering belong?
All sufferings are without an owner,

without exception. They should be
warded off simply because they are suffering.
Why is any restriction made in this case?
Why should suffering be prevented? Because everyone

agrees. If it must be warded off, then all of it must be warded
off; and if not, then this goes for oneself as it does for
everyone else.
(BCA 8:101–103, translation altered)
These verses respond to an implied objection that I am justified in giving my own well-being
greater emphasis because it is mine. Śāntideva responds by pointing out that since the self
does not exist (8:101), suffering does not belong to anyone (8:102). Rather, it is only the
impersonal badness of suffering which should motivate its removal. Verse 103 responds to
an implied follow-up objection. Since there are no persons, the opponent wonders, why is it
worth eliminating anyone’s suffering? Śāntideva responds by pointing to the robust universal
agreement that suffering is bad. Arguments must end somewhere, and the self-evident bad-
ness of suffering is a sufficient reason for committing to its removal. The argument concludes
by claiming that since no justifying distinction has been offered, we should commit to remov-
ing the suffering of all.
524
Śāntideva
The argument has generated a great deal of philosophical attention, with some commenta-
tors being critical of its success (Williams 1998; Harris 2011; Westerhoff 2015) and others
defending it (Pettit 1999; Siderits 2000). One of the objections frequently leveled against
Śāntideva is that he has not taken seriously enough the possibility that eliminating the belief in
a self makes plausible the nihilistic possibility that we no longer need to care about anyone’s
pain (Harris 2011). I cannot survey these contemporary treatments here. We can, however,
notice two features of Śāntideva’s argument that reinforce repeating themes in his text. First,
it illustrates the close connection between meditation and reasoning in the BCA; even if the
argument fails to convince his interlocuter, meditating on the nonexistence of the self, and con-
templating others’ suffering, will respectively lessen selfishness and strengthen compassion.
Second, we find in these verses once more the creative tension running through the text
between the conventional and ultimate levels of reality. At the conventional level, we meditate
on the pain of others and generate compassion as a way to motivate ourselves to progress along
the bodhisattva path. At the ultimate level, we recognize that selves do not exist and use this
insight to dissolve selfishness. But dipping down to the level of ultimate reality threatens the
coherence of the entire path, since it casts doubt on the existence of those whom we vow to
save. We have already seen that Śāntideva’s solution to this tension is not to resolve it but to
float between ontological levels as a way of stimulating compassion and dissolving craving
and aversion (Todd 2013). What is most distinctive in his approach to wisdom, therefore, may
not be the long defense of it in chapter nine but his manipulation and employment of it, along
with even delusion itself, to develop the perfections and help the bodhisattva make progress
along the bodhisattva path.
Conclusion
In both India and Tibet, Śāntideva’s presentations of the bodhisattva path have exerted great
influence on the development of Mahāyāna Buddhism. Although the BCA is far from exhaus-
tive in its treatment of Mahāyāna doctrine, its organization, and most importantly the viva-
cious style with which Śāntideva presents core Mahāyāna insights, have made the text an icon
of Buddhist literature. The ŚS has had somewhat less historical influence but is now being
recognized as a valuable complement to the BCA for understanding Śāntideva’s thought.
The last twenty years have seen the academic study of Buddhist ethics develop into a signif-
icant subfield of Buddhist studies. The BCA has become one of the key texts of this developing
movement. In this introduction, I have emphasized some of the text’s core contributions to
moral philosophy. This includes Śāntideva’s consideration of the relation between metaphysi-
cal positions, in particular not-self, and the normative conclusion of committing to impartial
benevolence. Just as striking is his attitude to wisdom itself, which he integrates into medita-
tions designed to develop the perfections. His account of generosity, best understood through
consultation of both BCA and ŚS, is striking in its interpretation of giving as initially and
primarily a mental attitude of non-clinging. Likewise, his development of patience is remark-
able in its radicalness in that it aims to eliminate all anger and enable the bodhisattva to endure
any amount of physical pain without mental discomfort. Finally, the bodhisattva vow itself is
extraordinary in its infinite magnitude but also in the tensions it contains, in that bodhisattvas
without selves make multi-lifetime vows to liberate empty beings. All of these topics also
factor into a reoccurring theme of the text as a whole: Śāntideva’s claim that a perfect and
limitless commitment to altruism is the best way to eliminate the mental afflictions which are
the source of all suffering. For Śāntideva, altruism and self-interest converge perfectly as one
progresses along the bodhisattva path.
525
Stephen Harris
Notes
1 For a partly contrasting view, see Edelglass 2017.
2 A natural question to ask is whether Śāntideva’s ethical commitments are best captured by one or
another of the ethical theories developed in Western ethics, such as consequentialism and virtue eth-
ics. Śāntideva’s emphasis on the development of virtue throughout his texts does not of itself settle the
issue, given that most ethical theories accept the importance of virtue, with consequentialists empha-
sizing its ability to maximize good outcomes and deontologists theorizing virtue’s role in enabling
keeping one’s commitments. Śāntideva’s deepest goals in the texts are the promotion of a particular
consequence, the ending of suffering, as well as the development of the virtues of the bodhisattva. It
is less clear whether he takes consequences or virtue as the deeper theoretical commitment in terms
of theory of right action, however, or whether this was even an interesting question for him. The most
systematic treatment of this issue is provided by Goodman 2009. Goodman takes Śāntideva to be a
consequentialist whose central ethical commitment is to promote the development of virtue. Good-
man’s interpretation is insightful but also controversial. See Barnhart 2012; Davis 2013; Harris 2015
for responses. See Edelglass 2013 for an accessible introduction to similarities and contrasts between
Buddhist and Western moral thought.
3 All translations of the BCA are from Śāntideva 1997. All translations of the ŚS are from Goodman
2016.
4 In the ŚS, Śāntideva describes these as the patience which endures pain (duṣkhādhivāsanakṣānti) and
the patience towards others wrongdoing ( parāpakāramarṣanakṣānti), respectively (Goodman 2016,
178). Śāntideva also distinguishes a third kind of patience, which arises from reflecting on dharma
(dharmanidhyānakṣānti). However, I agree with Lele (2007, 120) that, at least in the BCA, dharmic
patience is treated as a variety of the patience towards others’ wrongdoing. Therefore, I do not distin-
guish them here.
Bibliography
Bommarito, Nicolas. 2014. “Patience and Perspective.” Philosophy East and West 64 (2): 269–86.
Barnhart, Michael. 2012. “Theory and Comparison in the Discussion of Buddhist Ethics.” Philosophy
East and West 62 (1): 16–43.
Chödrön, Pema. 2007. No Time to Lose: A Timely Guide to the Way of the Bodhisattva. Translated by the
Padmakara Translation Group. Boston: Shambhala.
Clayton, Barbra. 2006. Moral Theory in Śāntideva’s Śikṣāsamuccaya: Cultivating the Fruits of Virtue.
London and New York: Routledge.
Davis, Gordon. 2013. “Traces of Consequentialism and Non-Consequentialism in Bodhisattva Ethics.”
Philosophy East and West 63 (2): 275–305.
Edelglass, William. 2013. “Buddhist Ethics and Western Moral Philosophy.” In A Companion to Bud-
dhist Philosophy, edited by Steven M. Emmanuel, 476–90. West Sussex: Wiley-Blackwell.
———. 2017. “Buddhism, Happiness and the Science of Meditation.” In Meditation, Buddhism and
Science, edited by David L. McMahan and Erik Braun, 62–63. New York: Oxford University Press.
Garfield, Jay L. 2012. “Mindfulness and Ethics: Attention, Virtue and Perfection.” Thai International
Journal of Buddhist Studies 3: 1–24.
Goodman, Charles. 2002. “Resentment and Reality: Buddhism on Moral Responsibility.” American
Philosophical Quarterly 39 (4): 359–72.
———. 2009. Consequences of Compassion: An Interpretation and Defense of Buddhist Ethics. Oxford:
———. 2016. The Training Anthology of Śāntideva: A Translation of the Śikṣāsamuccaya. Oxford:
Gyatso, Tenzin. 1994. A Flash of Lightning in the Dark of Night: A Guide to the Bodhisattva’s Way of
Life. Translated by the Padmakara Translation Group. Boston: Shambhala.
Harris, Stephen. 2011. “Does Anātman Rationally Entail Altruism? On Bodhicaryāvatāra 8:101–103.”
Journal of Buddhist Ethics 18: 92–123.
———. 2015. “On the Classification of Śāntideva’s Ethics in the Bodhicaryāvatāra.” Philosophy East
and West 65 (1): 249–75.
———. 2017. “The Skillful Handling of Poison: Bodhicitta and the Kleśas in Śāntideva’s
Bodhicaryāvatāra.” Journal of Indian Philosophy 45: 331–48.
526
Śāntideva
———. 2018. “Promising Across Lives to Save Non-Existent Beings: Identity, Rebirth and the Bodhisat-
tva’s Vow.” Philosophy East and West 68 (2): 386–407.
Harrison, Paul. 2007. “The Case of the Vanishing Poet: New Light on Śāntideva and the Śikṣā-
Samuccaya.” In Indica et Tibetica: Festschrift für Michael Hahn, edited by Konrad. Klaus and Jens-
Uwe Hartmann, 215–48. Zum 65. Vienna: Geburtstag von Freunden und Schülern überreicht.
Lele, Amod Jayant. 2007. “Ethical Revaluation in the Thought of Śāntideva.” PhD diss., Harvard
University.
Mahoney, Richard. 2002. “Of the Progresse of the Bodhisattvamārga in the Śikṣāsamuccaya.” M. A.
Thesis, University of Canterbury.
McRae, Emily. 2015. “Buddhist Therapies of the Emotions and the Psychology of Moral Improvement.”
History of Philosophy Quarterly 32 (2): 101–22.
Mrozik, Susanne. 2007. Virtuous Bodies: The Physical Dimensions of Morality in Buddhist Ethics.
Nussbaum, Martha. 1994. The Therapy of Desire. Princeton: Princeton University Press.
Pettit, John. 1999. “Paul Williams: Altruism and Reality: Studies in the Philosophy of the
Bodhicaryāvatāra.” Journal of Buddhist Ethics 6: 120–37.
Saito, Akira. 1993. “A Study of Akṣayamati (=Śāntideva)’s Bodhisattvacaryāvatāra as Found in the
Tibetan Manuscripts from Tun-huang.” Report of the Grant-in-Aid for Scientific Research. Miye
University, Japan.
Śāntideva. 1997. A Guide to the Bodhisattva Way of Life (Bodhicaryāvatāra). Translated by Vesna Wallace
and Alan Wallace. Ithaca, NY: Snow Lion Publications.
———. 2006. The Way of the Bodhisattva. Translated by the Padmakara Translation Group. Boston:
Shambhala.
———. 2008. The Bodhicaryāvatāra. Translated by Kate Crosby and Andrew Skilton. Oxford: Oxford
University Press.
Siderits, Mark. 2000. “The Reality of Altruism: Reconstructing Śāntideva.” Philosophy East & West 50 (3):
412–24.
Todd, Warren. 2013. The Ethics of Śaṅkara and Śāntideva: A Selfless Response to an Illusory World.
London and New York: Routledge.
Vernezze, Peter. 2008. “Moderation or the Middle Way: Two Approaches to Anger.” Philosophy East and
West 58 (1): 2–16.
Westerhoff, Jan. 2015. “The Connection Between Ontology and Ethics in Madhyamaka Thought.” In
Moonpaths: Ethics and Emptiness, edited by the Cowherds. New York: Oxford University Press.
Williams, Paul. 1998. Altruism and Reality: Studies in the Philosophy of the Bodhicaryāvatāra. Surrey,
UK: Curzon Press.
527
32
PRAJÑĀKARAGUPTA
Buddhist Epistemology as the Path to the
Wisdom of Non-Duality
Shinya Moriyama
Introduction
In the Buddhist epistemological tradition that was founded by Dignāga (ca. 480–540) and
developed by Dharmakīrti (ca. 600–660), a variety of topics concerning the means of valid
cognition ( pramāṇa) were explored extensively, mainly through commentaries on the works of
these two authors. Through word-by-word glossing, paraphrasing, and analyzing each stanza
and each sentence, commentators clarified the intentions behind Dignāga and Dharmakīrti’s
subtle arguments. Besides this work of clarification, commentators have sometimes inserted
their original philosophical ideas in between the lines. In this manner, Buddhist epistemology
developed through extending and updating the founders’ thought in accordance with each new
stage of Indian philosophy.
Among these commentators, Prajñākaragupta (ca. mid-eighth century–ninth century)
stands out due to his extensive commentary on Dharmakīrti’s Commentary on Epistemol-
ogy (Pramāṇavārttika), The Ornament of the Commentary on Epistemology (Pramāṇavārtti
kālaṅkāra).1 Prajñākaragupta’s commentary covers the chapters on the establishment of the
means of valid cognition (chapter II), perception (chapter III), and inference for others (chapter
IV) of the Commentary on Epistemology and does not address the chapter on inference for one-
self (chapter I), on which Dharmakīrti composed his own commentary. Before Prajñākaragupta,
Devendrabuddhi (ca. 630–690) had already composed a commentary ( pañjikā) on Dharmakīrti’s
work. But Prajñākaragupta’s commentary was particularly influential on later Buddhist and
Indian philosophy. This influence is seen in Ravigupta (ca. first half of ninth century), Jayanta
(the Buddhist, ca. tenth century), Jñānaśrīmitra (ca. 980–1040), Yamāri (ca. 1000–1060), and
the anonymous author of the Secret of Logic (Tarkarahasya). Prajñākaragupta’s philosophy
also received severe criticisms from Nyāya and Jaina philosophers such as Bhāsarvajña (ca.
first half of tenth century) and Vidyānandin (ca. second half of tenth century).
We have little biographical information on Prajñākaragupta. Although the Tibetan monk-
scholar Tāranātha (1575–1634) writes that Prajñākaragupta was a lay devotee and lived at
the time of King Mahāpāla, no clear evidence for the description exists (Schiefner 1869,
230). According to Theodore Stcherbatsky, Dharmakīrti’s followers can be classified into
three groups: the “philological school” represented by Devendrabuddhi and Śākyabuddhi, the
“philosophical school” represented by Dharmottara, and the “religious school” represented by
528 DOI: 10.4324/9781351030908-45

Prajñākaragupta
Prajñākaragupta. Stcherbatsky’s classification of “religious school” is motivated by the fact

that Prajñākaragupta’s commentary puts stress on the explanation of religious issues discussed
in the second chapter of Dharmakīrti’s Commentary on Epistemology (Stcherbatsky 1930–32,
39–47). It is also interesting to note that Prajñākaragupta’s school is distinguished from Dhar-
mottara’s school, because we know that Prajñākaragupta criticized several times an opponent’s
view that can most probably be attributed to Dharmottara and that the two philosophers stand
in opposition concerning interpretations of Dharmakīrti’s philosophy, specifically in terms of
the ontological status of mental image in cognition. In the lineage of Prajñākaragupta, three
followers’ names are well known: Ravigupta, Jayanta, and Yamāri, the latter two of whom
wrote commentaries on the Ornament of the Commentary on Epistemology.
In this chapter, I will introduce Prajñākaragupta’s most prominent arguments related to his
ontology, epistemology, and philosophy of religion. First, his ontology presents an original view
on causation known as “backward causation.” According to this view, a future entity is taken
to play the role of a cause of a present entity, against our commonsense view that cause pre-
cedes effect. Second, his epistemology is determined by the definition of existence for which he
argues. While Dharmakīrti’s definition of the “purposeful action/causal efficacy” is well known,
Prajñākaragupta several times refers to an idealistic view that acknowledges the dependence of
the existence of all entities on the mind. Third, with regard to philosophy of religion, his argu-
ments in defense of the Buddha’s omniscience are analyzed from two viewpoints, namely the
process for its attainment and the proof of it. Finally, Prajñākaragupta’s philosophy is summa-
rized as a path to the wisdom of non-duality from its opposite, dualistic view of reality.
Backward causation (bhāvikāran avāda)

Before Prajñākaragupta, Dharmakīrti established a Buddhist theory of causation based on
the idea that everything arises and perishes at each moment. He argues both that there is a
continuous chain of causes and effects and that several causes co-operate simultaneously in
producing a single effect. For instance, a seed, which, strictly speaking, continuously arises
moment to moment, produces a sprout when sunshine, earth, water, and so on cooperate.
Since one observes that a certain entity in appropriate conditions produces another entity,
and the entity does not arise when the former entity is lacking, one infers the former from the
latter, namely a cause from its effect, for example, fire from smoke. However, Dharmakīrti
rejected the reverse inference of an effect from its cause, because there is always the pos-
sibility of impediments between a cause and its effect. Regarding the reverse inference,
Dharmakīrti just argued that one can infer the fitness or possibility ( yogyatā) of an effect’s
arising when one observes an entity being surrounded by conditions appropriate to its pro-
ducing a possible effect. He states:
When the fitness of an effect’s arising is inferred from a cause in appropriate complex
of conditions, the fitness is called the “essential property” (svabhāva), because it
does not depend on anything else.
(PV I.7, Steinkellner 1991, 1999, 2013, Teil I, 17)
In Dharmakīrti’s system of logic, there are two types of valid reason that can result in a proof.
The first type is an essential property of an entity (svabhāvahetu, e.g., being produced for
proving impermanence), and the other is an effect (kāryahetu, e.g., smoke for proving fire).
Whereas the former is established in proving a property from another property, both of which
refer to the same entity, the latter is used in the inference of an entity, which is a cause, from
529
Shinya Moriyama
another entity, which is its effect. Based on this distinction, Dharmakīrti maintains that the
reverse inference of an effect from its cause can be described as a kind of inference based on
an essential property as logical reason; that is, it is the appropriateness or completeness of a
cause that proves another essential property, that is, the fitness for the production of its effect.
The two properties (i.e., the cause’s completeness and the fitness for its effect’s arising) indi-
cate one and the same state of a cause that necessarily produces its effect. However, he never
argued the reverse inference by relying on an effect as logical reason, even though, as Stein-
kellner (1999) has pointed out, Dharmakīrti probably would have been interested in explaining
the future attainment of Buddhahood or rebirth within his system of logic.
It is Prajñākaragupta who goes one step further and claims that one can infer a future life
from a present life by using the reverse inference. However, it should be noted that in this case,
he does not accept any new type of valid reason such as a cause as logical reason (kāraṇahetu).
Instead, by using the words “cause” and “effect” in different manners, he provides a peculiar
explanation to interpret the present life as an “effect” resulting from its future “cause.” In
doing so, the case is still categorized under the inference using effect as a reason (kāryahetu).
The idea is based on the theory of future cause (bhāvikāraṇavāda), or so-called“backward
causation,” in which a future entity plays the role of causing a present entity, against our com-
monsense view that cause precedes effect (Franco 2007, 2015).
How, then, is the backward causation justified? On this point, he resorts to an argumentative
device well known in Indian philosophy (Cardona 1967–68; Inami 1999) called “following co-
presence and co-absence” (anvayavyatirekānuvidhāyitva) as the grounds for defining “cause,”
which can be paraphrased by a famous formulation of the Buddhist doctrine of dependent
origination ( pratītyasamutpāda): “When A is present, B is also present,” and “When A is
absent, B is also absent.” Thus, if a particular entity a is established as such a type of A, a is
said to be the “cause” of b that belongs to type B. In our usual understanding of causality, one
might assume a temporal relation between A and B; yet, the formulation does not imply any
temporal order, at least not for Prajñākaragupta, who states:
If causation is determined merely from the relation in which an entity’s occurrence is

brought forth by another entity’s occurrence (tadbhāvabhāvitā), what does it contra-
dict? What is the use of the temporal sequence ( pūrvaparabhāva) here?
(PVA 69.11, v. 440)
To be more precise, the formulation “If A, then B” is described in Sanskrit with the terms of a
locative case ending (sati). Other commentators, such as Devendrabuddhi and Śākyabuddhi,
interpret this locative as denoting a temporal order where the cause precedes its effect.
Prajñākaragupta disagrees. Once the necessary connection of two types, A and B, is established,
there is no longer any distinction between inferences of “cause to effect” and “effect to cause,”
because in both cases, it is equally true that what is to be proved does not exist at the present,
regardless of whether it is a past entity or a future one. That is to say, in as much as past and
future entities are here considered as conceptual constructions, in the technical sense that they
are not actual but fabricated by the mind, both inferences have the same value. If one acknowl-
edges the validity of inference of a past entity that has already perished, similarly, one should
also accept the validity of inference of a future entity that has not yet come into existence.
Moreover, a folk belief largely shared at that time in India provides an important back-
ground to Prajñākaragupta’s idea. For instance, one’s present pleasant feeling is said to be
influenced by one’s future happiness. Likewise, when one perceives an omen of death (ariṣṭa)
in a person’s body, one believes that it is caused by the person’s future death. Those examples
530
Prajñākaragupta
run counter to our belief that the past determines the present, that there are cases in which we
also believe that the future determines the present (PVA 67.30, v. 435, 68.29–31; Franco 2007).
Although we know little about its exact impact on later Indian philosophy, there are at least
two examples of the legacy of Prajñākaragupta’s argument. The first is Bhāsarvajña, who pointed
out that Prajñākaragupta’s backward causation deviates from Dharmakīrti’s view: while the lat-
ter argues that the necessary relationship (avinābhāva) between the two is ascertained only after
one’s empirical determination of causal relations,2 the former maintains that the causal relation
is a priori determined as the necessary connection (Moriyama 1998). The second can be found
in a short work by Jitāri (ca. 940–1000), Treatise on Future Cause (Bhāvikāraṇavāda), which
has been very recently found and studied by Eli Franco (Franco 2015). In the treatise, Jitāri
makes an argument based on backward causation while referring to Prajñākaragupta’s view.
Existence defined by perception

Another innovative idea of Prajñākaragupta concerns the definition of existence (sattā/sattvam).
While Dharmakīrti’s definition of existence as “purposeful action/causal efficacy” (arthakriyā)
is well known (Katsura 1984; Dunne 2004, 272–78), Prajñākaragupta refers several times to
another definition, “existence is perception” or “existence is what is perceptible” (upalambhaḥ
sattā).” This resonates with George Berkeley’s famous phrase, Esse est percipi. With this
famous phrase, the eighteenth-century Irish philosopher aimed at criticizing the commonsense
view that presupposes the mind-independent existence of objects. According to Berkeley, the
idea is ultimately grounded on an eternal spirit’s perception that acts even while others do
not perceive an object, that eternal spirit being none other than God. Similarly, the eighth
century Buddhist philosopher aimed to establish an idealistic view that acknowledges the
mind-dependent existence of all entities. The interesting point of comparison between the two
authors is that, although Prajñākaragupta refers to the theory of existence in several contexts,
in one place (PVA 112.1–6), he argues for it by resorting to the Buddha’s omniscience regard-
ing all entities in the past, present, and future. This showcases an interesting affinity between
idealism and the argument for the religious authority in both traditions of Christianity and
Buddhism (Moriyama 2015, Forthcoming).
This idealistic view has its roots in Dharmakīrti’s discussion of the concept of “non-
perception” (anupalabdhi) as a logical reason. According to his account, the absence of a
perceptible object in a certain place is never directly perceived but inferred, and the inference
is based on the “non-perception of an object that is perceptible under appropriate conditions
(upalabdhilakṣaṇaprāpta)” (Kellner 2003). In this case, there must be a necessary connection
between the logical reason and the target to be proved, that is, between the non-perception and
the nonexistence of a perceptible object. Dharmakīrti explains the necessary connection by its
contraposition:
That is, existence is nothing but either perceptibility (upalabdhi) that is characterized
by the real entity’s causal capability, or perception (upalabdhi), namely, a cognition’s
occurrence based on the entity’s causal capability.
(PVSV 4.9–11, Kellner 2003, 125–28; Steinkellner 2013, Teil I, 11)
The Sanskrit term upalabdhi is grammatically analyzed as a derivative from the verb “to per-
ceive” (upa-LABH), and the noun “perception” is analyzed from the perspective of both the
object (karman) and the agent (kartṛ). Dharmakīrti plays on both possibilities, first on focusing
on the perceptibility of the object, and second on focusing on the occurrence of the perceiving
531
Shinya Moriyama
cognition. In the first use of upalabdhi, which stands for a causal account of perception, a cog-
nition occurs by the causal capability of its object, which has arisen in the moment preceding
the perceptive cognition, which implies that there must be a temporal gap between an object’s
existence and that which is cognized through its perception. However, the second use of upa
labdhi suggests that an object’s existence is equated with the occurrence of its perception. In
the two definitions of existence, some commentators like Śākyabuddhi takes the first alterna-
tive as standard and explains the second one as a metaphorical usage of the term upalabdhi
(PVṬ, Vol. Je, 13b3; Steinkellner 2013, II, 50 n. 54).
In contrast, Prajñākaragupta rejects the causal account of perception and claims that the
object and its cognition are essentially not different.3 He explains as follows:
If the term upalabdhi indicates the state of being perceived, the state is exactly the essence
of an object. Alternatively, even though the term upalabdhi indicates the perception in
the sense of agency, it means that the existence of a perceptible object is conceptually
pervaded by its “perception.” And if x is not separable from y, one does not understand
that x and y are different, because the difference is understood only in separable ones.
Dharmakīrti said therefore that existence is perception because the latter does not deviate
from the former. Moreover, the perception of objects does not occur due to perception
that is different from the object and that occurs at the same time of the object, but rather
due to “perception” characterized by “being perceived,” which is entailed in the object
(tadanupraviṣṭopalabhyamānatā). Therefore, existence is nothing but perception.
(PVA 633.4–8, Moriyama, forthcoming)
Compared to Dharmakīrti’s original argument, Prajñākaragupta’s interpretation emphasizes the

inseparability between an object’s existence and its perception by denying the idea that it is pos-
sible to separate the two as cause and effect. For him, there is no reason to suppose a gap between
an object’s existence and its perception. This is because, he argues, the term “object” really
refers to something that is already in the state of being perceived. That is, a mental image of the
object necessarily appears in the perceiver’s cognition. In fact, his epistemological position is
often called the position that accepts cognition endowed with mental images (sākārajñānavāda),
which summarizes well in epistemological terms his claim that things are mind dependent.4
This Indian version of “esse est percipi” is further discussed in Prajñākaragupta’s com-
mentary on verse III.53d (meyaṃ tv ekaṃ svalakṣaṇam) of Dharmakīrti’s Commentary on
Epistemology, where Dharmakīrti argues that the two kinds of objects, that is, particulars
(svalakṣaṇa) and universals (sāmānyalakṣaṇa), are ultimately reducible to the particular as
the object that is capable of purposeful action (arthakriyā). For instance, an aggregate of atoms
that forms the so-called “pot” is defined by the action of “holding water,” which is expected
when one perceives or infers the pot.5 That is, the existence of a particular entity is deter-
mined through one’s act toward the entity after its perception or inference. Since the immedi-
ate object of perception and inference is causally connected to the particular that fulfills the
purpose expected by the cognizer, Dharmakīrti claims that cognizable objects are reducible to
the particular. To this, Prajñākaragupta proposes another, non-causal account that maintains
the equality between an object’s existence and its perception at the ultimate level:
The existence defined by perception (upalambha) is indeed the ultimate existence.

The ultimate existence of entities is expressed in this way: “Existence is nothing but
perception.”
(PVA 213.22, v. 230)
532
Prajñākaragupta
Here we should pay attention to the term “ultimate,” which he uses in other contexts for quali-
fying the state of self-awareness or reflexive awareness (svasaṃvedana), in which a cognition
not only cognizes an external object but is also reflexively aware of its cognitive state. To be
more precise, every cognition arises holding two different aspects, namely the aspect of an
object and that of its own grasping cognition. Thus, by focusing on the internal structure of
cognition that is endowed with those two aspects, Prajñākaragupta even goes a step farther
to argue that ultimately all cognitions are reducible to self-awareness. As he states elsewhere
(PVA 31.22–23, Ono 1999, 83.2): “The means of valid cognition is only self-awareness as the
single type of perception” (svasaṃvedanam evaikaṃ pratyakṣaṃ pramāṇam). For him, ulti-
mately, there is only self-awareness, distinct from all other cognitions that are established in
everyday activity (vyavahāra). According to Prajñākaragupta’s understanding, Dharmakīrti’s
account works well in the analysis of those ordinary cognitions, in which an object’s existence
should be defined by purposeful action, without introducing the analysis from the viewpoint
of self-awareness. In contrast, at the ultimate level, an object’s existence is equated with the
appearance of its mental image that is being perceived reflexively in the cognition. This view-
point can be identified as Prajñākaragupta’s idealist approach to the definition of existence.
In this connection, it is also noteworthy that a similar idea to distinguish the two levels of
existence emerges again in the epistemological works of Jñānaśrīmitra, a late tenth- to early
eleventh-century Buddhist philosopher. He distinguishes the existence defined by illumination
( prakāśa) from that defined by purposeful action or causal efficacy. While the latter presup-
poses the position that acknowledges external objects, the former is linked with Yogācāra
idealism (SSŚ 398.24–399.2; cf. CAP 131.13–25). Of course, because the terminology of
“illumination” differs from “perception,” it is still questionable whether Jñānaśrīmitra’s view
is inspired by Prajñākaragupta’s. Nevertheless, we should not overlook the structural similar-
ity between the two arguments.
The Buddha’s omniscience as the ultimate means of valid cognition

The major contribution of Prajñākaragupta in the field of philosophy of religion is related to
his arguments about the Buddha’s omniscience. To explain the point, we should briefly look at
the religious philosophical background of Dharmakīrti’s epistemology, which is specifically
discussed in the second chapter of the Commentary on Epistemology (Franco 1997). In the
period from the late sixth to seventh century, a controversy arose regarding religious autho
rity between Mīmāṃsā thinkers, who accept the eternal Vedic scripture as their authority;
Nyāya-Vaiśeṣika thinkers, who accept a god (Īśvara) as their authority; and Buddhist think-
ers, who consider the Buddha the ultimate authority. Kumārila (ca. late sixth century–early
seventh century), a representative philosopher of the Mīmāṃsā school, argued against the
Buddha’s omniscience. Dharmakīrti, avoiding a direct reply to the criticism, demonstrated
the Buddha’s authority based on his practical knowledge on the primary purpose of human
beings ( pradhānapuruṣārtha), namely the cessation of the cause of all sufferings and the
means for its attainment, that is, the Four Noble Truths. Since the Buddha does not mislead
about (avisaṃvādin) the path that he himself has already experienced (svadṛṣṭamārga), and
because he presents the Four Noble Truths which are not yet known (ajñātārthaprakāśa) to
other non-Buddhist teachers, he is called the “means of valid cognition” whereby people are
led to the right cognition on the final purpose of life (Franco 1997, chapters 1 & 2; Krasser
2001; Steinkellner 2003; Moriyama 2014, chap. 1).
However, later Buddhist philosophers, including Śāntarakṣita, Kamalaśīla, and Prajñā
karagupta, selected a different way to defend the Buddha’s omniscience from the Mīmāṃsā
533
Shinya Moriyama
criticisms (McClintock 2010; Moriyama 2014). In the case of Prajñākaragupta, his commen-
tary on the second chapter of the Commentary on Epistemology provides several interesting
arguments for establishing the Buddha’s omniscience as a reliable source for radically inac-
cessible matters (atyantaparokṣa), including past and future lives ( paraloka) and the connec-
tion between karmic causes and their results, and so on, which remain hidden for us since our
conventional means of valid cognition do not reach what lies beyond our senses. He remarks:
Therefore our ignorance is removed only by a treatise that has been proclaimed by an
omniscient being, but not by anything else. Thus, in this manner, only the words of an
omniscient being are a means of valid cognition. Hence, ultimately ( paramārthatas),
only the cognition of an omniscient being, but not another, is a means of valid cogni-
tion. This is the ultimate meaning of Dharmakīrti’s statement of PV II 5b, “a treatise
is for the removal of ignorance.”
(PVA 29.26–27, Ono 1999, 77.16–18; Moriyama 2014, 21)
Identifying the Buddha’s omniscience and his teaching as the ultimate means of valid cogni-
tion, Prajñākaragupta distinguishes it from the conventional (sāṃvyavahārika) ones, namely
our ordinary perception and inference. However, it should be noted that even such omnis-
cience is classified as “conventional” in comparison to non-dual cognition (advayajñāna),
because the omniscience presupposes entities that are established in our everyday activity
based on the division of cause and effect (Iwata 2004; Moriyama 2014).
The Mīmāṃsā opponents point out the impossibility for a human being, such as the Buddha,
to attain omniscience because of the limitation of human ability. Dharmakīrti already argued
that unlike physical ability like jumping, which has the limitation of its training, mental abili-
ties can be developed infinitely through training. For instance, by repeatedly training oneself to
arise compassion, one can develop an infinite compassion to the point of extending to all living
beings. Following Dharmakīrti, Prajñākaragupta further argues that the same is possible for
the Buddha’s meditation for the attainment of omniscience. Since he has compassion, namely
the wish for the salvation of all living beings (PVA 53.15–17), he is motivated to undertake
the practice for omniscience. Prajñākaragupta describes this kind of practice as the Buddha’s
making various kinds of inferences for knowing past and future entities that relate to his own
mental stream as well as others’ ones. Following the chain of cause and effect, he infers what
has happened and what will happen and finally reaches to the direct intuition of every sin-
gle phenomena. Therefore, Prajñākaragupta calls one side of the Buddha’s omniscience the
“inference of all aspects of entities” (sarvākārānumāna) and another side the immediate cogni-
tion (sākṣātkaraṇa) of everything. This would imply that the notion of omniscience breaks the
border between perception and inference as defined by Dignāga. However, since the Sanskrit
term pratyakṣa (“perception”) etymologically (i.e., by vyutpattinimitta) indicates a “cognition
that occurs in accordance with ( prati) each sense faculty (akṣa),” it should be distinguished
from inference. To solve the issue, Prajñākaragupta replies that the term akṣa here indicates the
mental faculty (manas) as the cause of the immediate cognition, and hence, the term pratyakṣa
covers the “inference of all aspects of entities,” a cognition based on the mental faculty as
well. Just as one is able to perceive fire from smoke spontaneously after the absolute repetition
(atyantābhyāsa) of inference of fire from smoke, the Buddha is also able to perceive everything
spontaneously after the absolute repetition of inferences of various kinds of cause and effect. In
this manner, the process of how the Buddha has attained omniscience is explained as a combina-
tion of inference and perception (Moriyama 2014, Chap. 3).
534
Prajñākaragupta
The Mīmāṃsā opponents also point out the difficulty concerning the proof of an omnisci-
ent being. Kumārila already raised the objection that it is impossible in principle to prove that
somebody else is an omniscient being unless the person who judges that person is, herself,
an omniscient being. In this regard, Prajñākaragupta emphasizes the importance to verify the
Buddha’s teaching even partially, to the extent that one is able to examine it (śakyavivecana).
Commenting on the first half of verse II.280 of the Commentary on Epistemology, where
the Buddha’s knowledge, which is considered true, lasting, and endowed with all qualities
(tattvasthirāśeṣaviśeṣajñāna), is said to be justified by its capacity to “protect living beings,”
namely the fact that the teaching of the Four Noble Truths can liberate them from suffer-
ing, Prajñākaragupta shows how it is possible to prove the Buddha’s omniscience from his
teaching (PVA 164.28–12, Wakahara 1985; Moriyama 2014, Chap. 4). Suppose we listen to a
teacher’s statement, “All things are perishing at every moment,” one aspect of the Truth of suf-
fering. If we learn Dharmakīrti’s proof of momentariness based on the reason of “existence”
as defined by causal efficacy (sattvānumāna), we are able to examine it and conclude that the
proposition is true (Steinkellner 1968). This leads us to another assumption – that the teacher’s
knowledge of the proposition is either transmitted from another teacher’s teaching or ascer-
tained by himself. Even in the first case, another teacher’s knowledge is assumed to be either
originated from a further different teacher’s teaching or ascertained by himself. In this manner,
it is inevitable that one needs to finally assume an initial teacher who immediately perceived
everything as momentarily perishing and ascertained the proposition established universally.
Since the statement, “All things are perishing at every moment,” is a universal proposition,
the teacher who initially held it is required to possess the knowledge that the proposition is
established without any exception. If there were even one unseen entity for the teacher, the
possibility would remain that the proposition is inconclusive. Therefore, in order to explain
that the proposition is universally established, there is no other way than to assume that the
teacher, the Buddha, is omniscient. In other words, the validity of inference of momentariness,
which is verifiable even by our logical thinking, guarantees the high probability of the initial
teacher’s omniscience, without which the universal validity of the proposition would not be
grounded on all entities of this world. In addition, by applying the same way to other proposi-
tions the Buddha taught, it becomes more certain that he is omniscient, and as a consequence,
we may have confidence in the Buddha’s words with respect to radically inaccessible matters.
Conclusion: the path toward the wisdom of non-duality

Before concluding this chapter, let us turn to the topic of the wisdom of non-duality that is at
the core of Prajñākaragupta’s Buddhist philosophy. He first mentions this special cognition in
his commentary on Dharmakīrti’s second definition of the means of valid cognition, as that
which “illuminates the object that is not yet known to us (ajñātārtha),” and states that this
special cognition is the ultimate means of valid cognition. Commenting on the “object that is
not yet known to us,” Prajñākaragupta interprets it as meaning the nature of non-duality and
states that a means of valid cognition is nothing but that which illuminates non-duality. For
him, Dharmakīrti’s system of epistemology is set as a program that leads people from their
lower level of knowledge to the highest level of the wisdom. Starting from the position that
presupposes the existence of external objects independent from the mind, one gradually attains
higher understandings as one’s analysis becomes more and more subtle. For instance, in the
view that cognitions are endowed with images that correspond to external objects, further
investigation reveals a thorny issue, since there can be no numerical correspondence between
535
Shinya Moriyama
images and external objects. This is because the former are essentially non-divisible, while the
latter are divisible into many atoms. Thus, for solving the problem, Dharmakīrti claims in the
third chapter of his Commentary on Epistemology that cognition is empty owing to the neither-
one-nor-many argument (209–219) and that cognition has the nature of variety-in-non-duality
(citrādvaita) (220–222). As Inami (2011) has clarified in detail, Prajñākaragupta goes one step
further and provides his own view of this section: “There can never exist any distinction in the
ultimate sense, so that the distinctions between singular and multiple, and between variegated
and unvariegated, are considered untenable” (Inami 2011, 178).
Prajñākaragupta’s fundamental position on the non-duality of cognition can thus shed more
light on how his own views on causality, ontology, and religious philosophy are articulated. For
instance, his peculiar view of backward causation entails the view that the temporal sequence
is fictional. This claim can now be understood as implying that time is seen by Prajñākaragupta
from the perspective of non-duality, which abolishes the temporal differences of the past,
present, and future. Likewise, the definition of existence as perception is also based on the
doctrine of non-duality that denies the subject-object dichotomy. Finally, it is also remarkable
that the Buddha’s role loses its authoritative position from the point of view of non-duality,
since no distinction between self and other or between teacher and disciples remains. In other
words, the Buddha’s role, in particular his omniscience, occupies its place only in a world of
duality that is conceptually constructed by ordinary people who sees themselves as separated
from and inferior to the Buddha.
Overall, Prajñākaragupta’s philosophy seeks to lead from the conventional reality, based
on duality, to the ultimate reality. As one’s attachments to the world of duality are gradually
abandoned through investigation of the Buddha’s teaching, a new horizon of non-dual cogni-
tion opens little by little. The goal that Prajñākaragupta finally aimed to reach is the state of
non-duality, which is called the “highest vehicle” (advayaṃ yānam uttamam, PVA 32.15, Ono
1999, 84.14f.).
At the very end of The Ornament of the Commentary on Epistemology, Prajñākaragupta
writes, “Village girls are longing for ornaments to make the body elegant due to their desire.
I composed the Ornament (Alaṅkāra) of Dharmakīrti’s Commentary on Epistemology
(Vārttika), but it is not out of my arrogance” (PVA 648.9–10, v. 622). His motivation as a com-
mentator is thus based on the simple wish to clarify Dharmakīrti’s condensed arguments rather
than to present his own view. However, his commentary includes in fact numerous original
ideas that introduce novel perspectives and are different from Dharmakīrti’s account. Although
Prajñākaragupta will say that they are the natural consequences that can be drawn from the
depth of his predecessor’s ideas, in our eyes, they are good examples of the fascinating devel-
opment of Buddhist epistemology through the “Commentary” tradition.
Abbreviations
CAP Citrādvaitaprakāśavāda, ed. by A. Thakur, 1975
PV I Pramāṇavārttika, the first chapter (Svārthānumāna), ed. by R. Gnoli, 1960
PV II Pramāṇavārttika, the second chapter (Pramāṇasiddhi), ed. by Y. Miyasaka, 1971/72,
and T. Vetter, 1990
PV III Pramāṇavārttika, the third chapter (Pratyakṣa), ed. by H. Tosaki, 1979/1985
PVA Pramāṇavārttikālaṅkāra, ed. by R. Sāṅkṛtyāyana, 1953
PVSV Pramāṇavārttikasvavṛtti, ed. by R. Gnoli, 1960
PVṬ Pramāṇavārttikaṭīkā, sDe dge Tibetan Tripiṭaka, vol. Je.
SSŚ Sākārasiddhiśāstra, ed. by A. Thakur, 1987.
536
Prajñākaragupta
Notes
1 For bibliographical information on PVA, see Ono 1999, xi–xxv; Steinkellner and Much 1995, 74f. On
Prajñākaragupta’s date, see Franco 2019. Although Sāṅkṛtyāyana’s edition of PVA is still useful, correc-
tions are necessary for a better edition by relying on photocopies of manuscripts in Watanabe 1998 and
Tibetan materials, including Ravigupta’s commentary on PV and sub-commentaries by Jayanta and Yamāri.
Very recently, a Sanskrit manuscript of Yamāri’s subcommentary has been found and is now being edited
by Eli Franco and his research group at the University of Leipzig. Apart from PVA, the Tibetan Tripiṭaka
contains a treatise called the Sahopalambhaniyamasiddhi (sDe dge no. 4255/Peking no. 5753) attributed to
Prajñākaragupta, though Iwata (1991, 255f.) questions whether it was actually composed by him.
2 Dharmakīrti argues in several different places that causal relation is determined by perception and non-
perception. On the empirical process of determination, there was a controversy among his followers.
See Kajiyama 1963.
3 In this connection, we know that there were two different opinions about the interpretation of the logi-
cal reason of non-perception (anupalabdhihetu): while a party represented by Dharmottara accepts
the non-perception of the property that pervades another property (vyāpakānupalabdhi), which pre-
supposes only a conceptual relation between the existence of perceptible objects and their percep-
tions, another party represented by Prajñākaragupta accepts the non-perception of intrinsic nature
(svabhāvānupalabdhi), which presupposes the real identity between existence and perception. See
Kellner 1997; Tani 2000, 186–95.
4 Concerning the later controversy around a Yogācāra position that cognition is endowed with images
(sākāra) and another position that cognition is not endowed with images (nirākāra), Jñānaśrīmitra
evaluates that Prajñākaragupta completed the former position, which was originally intended by
Dharmakīrti. See SSŚ 367.18–20; Oki 1975, 94.
5 For the verse and its interpretation, see Franco and Notake 2014, 136–39. For Prajñākaragupta’s inter-
pretation of svalakṣaṇa and its related concept pravartaka, see Kobayashi 2011; McCrea 2011; Miyo
2014. On Prajñākaragupta’s theory of sensory perception, see Yokoyama 2018.
Bibliography
Cardona, George. 1967–1968. “Anvaya and Vyatireka in Indian Grammar.” Adyar Library Bulletin
31–32: 313–52.
Franco, Eli. 1997. Dharmakīrti on Compassion and Rebirth. Vienna: Arbeitskreis für Tibetische und
Buddhistische Studien Universität Wien.
———. 2007. “Prajñākaragupta on pratītyasamutpāda and reverse causation.” In Pramāṇakīrtiḥ: Papers
dedicated to Ernst Steinkellner on the occasion of his 70th birthday, Part 1, edited by Birgit Kellner,
Helmut Krasser, Horst Lasic, Michael T. Much, and Helmut Tauscher, 163–85. Vienna: Arbeitskreis
für Tibetische und Buddhistische Studien Universität Wien.
———. 2015. “Jitāri on Backward Causation (bhāvikāraṇavāda).” In Buddhist Meditative Praxis: Tra-
ditional Teachings & Modern Application, edited by K. L. Dhammajoti, 81–116. Hong Kong: Centre
of Buddhist Studies University of Hong Kong.
———. 2019. “Prajñākaragupta.” In Brill’s Encyclopedia of Buddhism, Volume Two, edited by Jonathan
A. Silk et al. Leiden: Brill.
Franco, Eli, and Miyako Notake. 2014. Dharmakīrti on the Duality of the Object. Pramāṇavārttika III
1–63. Zürich and Berlin: LIT Verlag.
Inami, Masahiro. 1999. “On the Determination of Causality.” In Dharmakīrti’s Thought and Its Impact
on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakīrti Conference,
Hiroshima, November 4–6, 1997, edited by Shoryu Katsura, 131–54. Vienna: Verlag der Österreichis-
chen Akademie der Wissenschaften.
———. 2011. “Nondual Cognition.” In Religion and Logic in Buddhist Philosophical Analysis. Proceed-
ings of the Fourth International Dharmakīrti Conference, Vienna, August 23–27, 2005, edited by
Helmut Krasser et al., 177–96. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
Iwata, Takashi. 1991. Sahopalambhaniyama. Stuttgart: Franz Steiner Verlag.
———. 2004. “Prajñākaragupta’s Proof of the Buddha’s Authority.” In Three Mountains and Seven
Rivers: Prof. Musashi Tachikawa’s Felicitation Volume, edited by Shoun Hino and Toshihiro Wada,
355–74. New Delhi: Motilal Banarsidass Publishers.
537
Shinya Moriyama
Kajiyama, Yuichi. 1963. “Trikapañcakacintā: Development of the Buddhist Theory on the Determination
of Causality.” Miscellanea Indologica Kiotiensia 4–5: 1–15. [Reprint in Y. Kajiyama, Studies in Bud-
dhist Philosophy (Selected Papers), edited by Katsumi Mimaki et al., 475–89. Kyoto: Rinsen Book].
Katsura, Shoryu. 1984. “Dharmakīrti’s Theory of Truth.” Journal of Indian Philosophy 12: 215–35.
Kellner, Birgit. 1997. “Non-Cognition (anupalabdhi) – Perception or Inference? The Views of Dharmot-
tara and Jñānaśrīmitra.” Tetsugaku 49: 121–34.
———. 2003. “Integrating Negative Knowledge into Pramāṇa Theory: The Development of the
Dṛśyānupalabdhi in Dharmakīrti’s Earlier Works.” Journal of Indian Philosophy 31: 121–59.
Kobayashi, Hisayasu. 2011. “Prajñākaragupta’s Interpretation of Svalakṣaṇa.” Indogaku Bukkyōgaku
Kenkyū 59 (3): 182–87.
Krasser, Helmut. 2001. “On Dharmakīrti’s Understanding of Pramāṇabhūta and His Definition of
Pramāṇa.” Wiener Zeitschrift für die Kunde Südasiens 45: 173–99.
McClintock, Sara. 2010. Omniscience and the Rhetoric of Reason. Śāntarakṣita and Kamalaśīla on
Rationality, Argumentation, and Religious Authority. Boston: Wisdom Publications.
McCrea, Lawrence. 2011. “Prajñākaragupta on the pramāṇas and their objects.” In Religion and Logic in
Buddhist Philosophical Analysis. Proceedings of the Fourth International Dharmakīrti Conference,
Vienna, August 23–27, 2005, edited by Helmut Krasser et al., 319–28. Vienna: Verlag der Österrei-
chischen Akademie der Wissenschaften.
Miyo, Mai. 2014. “Controversy Between Dharmottara and Prajñākaragupta Regarding Pravartaka.”
Indogaku Bukkyōgaku Kenkyū 62 (3): 223–28.
Moriyama, Shinya. 1998. “Raise no ronshō ni miru Prajñākaragupta no mirai genin setsu.” [“Prajñāka
ragupta’s Theory of Future Cause in the Proof of Future Lives.”] Indo Tetsugaku Bukkyōgaku Kenkyū
5: 44–57.
———. 2014. Omniscience and Religious Authority. A Study on Prajñākaragupta’s Pramāṇa
vārttikālaṅkārabhāṣya ad Pramāṇavārttika II 8–10 and 29–33. Zürich and Berlin: LIT Verlag.
———. 2015. “Prajñākaragupta no chikaku-sonzai setsu” [“Prajñākaragupta’s Theory of Existence
Defined by Perception.”] Indo Tetsugaku Bukkyōgaku Kenkyū 22: 175–90.
———. Forthcoming. “Prajñākaragupta on sattvam upalabdhir eva or an Indian version of “esse est per-
cipi.” In Gedenkschrift für Helmut Krasser, edited by Vincent Eltschinger, Chizuko Yoshimizu, et al.
Oki, Kazufumi. 1975. “Citrādvaita riron no tenkai – Prajñākaragupta no ronjutsu” [“A Development of
the Theory of Citrādvaita: An argument of Prajñākaragupta.”] Tōkai Bukkyō 20: 94–81.
Ono, Motoi. 1994. “Prajñākaragupta ni yoru Dharmakīrti no pramāṇa no teigi no kaishaku –
Prajñākaragupta no shinriron” [“Prajñākaragupta’s Interpretation of Dharmakīrti’s Definition of Valid
Cognition – Prajñākaragupta’s Theory of Truth.”] Indogaku Bukkyōgaku Kenkyū 42 (2): 885–75.
———. 1999. Prajñākaraguptas Erklärung der Definition gültiger Erkenntnis. Vienna: Verlag der Öster-
reichischen Akademie der Wissenschaften.
Patil, Parimal. 2009. Against a Hindu God. Buddhist Philosophy of Religion in India. New York: Colum-
Schiefner, Anton. 1869. Tāranātha’s Geschichite des Buddhismus in Iniden. St. Petersburg: Commis-
sionäre der Kaiserlichen Akademie der Wissenschaften.
Stcherbatsky, Theodore. 1930/32. Buddhist Logic, 2 vols. Leningrad: The USSR Academy of Sciences.
Steinkellner, Ernst. 1968. “Die Entwicklung des kṣaṇikatvānumānam bei Dharmakīrti.” Wiener Zeitschrift
für die Kunde Süd- und Ostasiens 12–13: 361–77.
———. 1991. “Dharmakīrti on the Inference of Effect (kārya).” In Papers in Honour of Prof. Dr. Ji
Xianlin on the Occasion of His 80th Birthday (II), edited by Li Zheng et al., 711–36. Jiangxi: Jiangxi
Renmin Chubanshe.
———. 1999. “Yogic Cognition, Tantric Goal, and Other Methodological Applications of Dharmakīrti’s
Kāryānumāna Theorem.” In Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy.
Proceedings of the Third International Dharmakīrti Conference, Hiroshima, November 4–6, 1997,
edited by Shoryu Katsura, 349–62. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.
———. 2003. “Once More on Circles.” Journal of Indian Philosophy 31: 323–41.
———. 2013. Dharmakīrtis frühe Logik. Annotierte Übersetzung der logischen Teile von Pramāṇavārttika
1 mit der Vṛtti, 2 vols. Tokyo: The International Institute for Buddhist Studies.
Steinkellner, Ernst, and Michael T. Much. 1995. Texte der erkenntnistheoretischen Schule des Buddhis-
mus. Göttingen: Vandenoeck & Ruprecht.
Tani, Tadashi. 2000. Setsunametsu no kenkyū. [A Study on Momentariness.] Tokyo: Shunjūsha.
538
Prajñākaragupta
Tosaki, Hiromasa. 1979/1985. Bukkyō ninshikiron no kenkyū [A Study on Buddhist Epistemology], 2 vols.
Tokyo: Daitō shuppansha.
Vetter, Tilmann. 1990. Der Buddha und Seine Lehre in Dharmakīrtis Pramāṇavārttika. Vienna: Arbeit-
skreis für Tibetische und Buddhistische Studien Universität Wien.
Wakahara, Yusho. 1985. “Āgama no kachi to zenchisha no sonzai shōmei.” [“Authority of Scripture and
the Proof of the Omniscient Being.”] Bukkōgaku Kenkyū 41: 52–78.
Watanabe, Shigeaki, ed. 1998. Sanskrit Manuscripts of Prajñākaragupta’s Pramāṇavārttikabhāṣyam,
Facsimile Edition. Patna and Narita: Bihar Research Institute/Naritasan Institute for Buddhist Studies.
Yokoyama, Akito. 2018. “Prajñākaragupta’s Theory of Sense-Perception (indriyapratyakṣa): With a
Focus on the Criticism of the Nyāya-Vaiśeṣika School.” Journal of Indian and Buddhist Studies 66 (3):
148–52.
539
PART 7
Late-Period Commentators
(Tenth–Twelfth Century)
LATE-PERIOD COMMENTATORS
The tenth through twelfth centuries represent the start of the waning of the classical period of
Buddhist philosophy in India. Yet this period is also marked by tremendous creativity, often
in ways that modern scholars are only beginning to appreciate and understand, with Buddhist
philosophers propounding new types of syntheses both in the debate halls of the great monastic
universities of North India and in scholarly treatises. Tantric practice and thought are increas-
ingly influential, while the epistemological theories of Dignāga and Dharmakīrti continue to
hold considerable sway. Madhyamaka and Yogācāra continue to vie for dominance, even as
thinkers like Ratnākaraśānti attempt to knit them together, sometimes with the aid of tantric
or Buddha-nature theories. Scholars like Jñānaśrīmitra and Ratnakīrti continue to argue with
non-Buddhist thinkers, extending conversations on topics including philosophy of language
and the warrants for knowledge ( pramāṇas) such as perception and inference, even as practi-
tioners like Saraha (see Part 2) offer an alternative path rejecting such scholastic approaches.
Some philosophers travel outside of India, bringing new ideas to other regions and greatly
influencing the course of Buddhist thought, as with Atiśa’s promotion of Candrakīrti in Central
Tibet. In turn, Tibetans and others visit the monastic universities, seeking knowledge to bring
back to their homelands, where Buddhist philosophy would continue to develop long past the
time of its demise in India.
According to Tibetan sources, Jitāri (tenth century) was associated with tantric traditions
and clearly defended Madhyamaka in his doxographical writings. However, as Junjie Chu
shows in his chapter, two recently discovered manuscripts of a treatise attributed to Jitāri, Top-
ics of Debate, complicate our understanding of Jitāri’s own philosophical position. Topics of
Debate makes possible a more comprehensive understanding of Jitāri’s philosophy than previ-
ously available texts because of the breadth of its scope: in seventeen chapters, he focuses on
seventeen significant topics of debate among Buddhist, Brahmanical, and Jain philosophers.
These chapters typically begin with an argument for Jitāri’s own position followed by refuta-
tions of objections and other positions. Chu gives an overview of each of these chapters and
concludes that while Jitāri may have been influenced by Madhyamaka thought, he is better
understood as a Yogācāra thinker in the tradition of Dharmakīrti.
Jñānaśrīmitra (tenth century) belongs to a group of scholars from the great Buddhist uni-
versity of Vikramaśīla. He also wrote from within the logical and epistemological tradition
derived from Dignāga and Dharmakīrti. Jñānaśrīmitra analyzes cognitions into “appearance”
DOI: 10.4324/9781351030908-47 543

Late-Period Commentators
( pratibhāsa) and “determination” (adhyavasāya). Any conceptual or nonconceptual aware-

ness, according to Jñānaśrīmitra, involves the appearance of an image in consciousness.
“Determination” is the process whereby a phenomenal image is given as presenting something
other than itself; this is, Jñānaśrīmitra argues, conceptual awareness. But we are mistaken in
taking a cognition that includes an image of something beyond itself as actually giving us an
external object. Rather, as Lawrence McCrea makes clear in his chapter, for Jñānaśrīmitra,
appearances in consciousness are just that – appearances in consciousness – for there is
nothing that exists outside of consciousness. Jñānaśrīmitra defends an idealist position that
there is only one undifferentiated mental image, what he calls “variegated nonduality” (citra-
advaita). He deploys the distinction between appearance and determination to offer creative
solutions to a variety of problems in ontology and metaphysics, philosophy of language, epis-
temology, logic, and philosophy of religion, even as he grounds himself as an exegete of the
Dharmakīrtian tradition.
Jñānaśrīmitra’s work was influential for much Buddhist philosophy in India and was also
engaged by prominent non-Buddhist Indian philosophers for centuries. Ratnakīrti (eleventh
century), his student from Vikramaśīla, concisely summarizes, defends, and elaborates on the
views of his teacher. Like Jñānaśrīmitra, Ratnakīrti works within the broader Dharmakīrtian
tradition of logic and epistemology and addresses a range of issues in Buddhist philosophy. In
his chapter, Patrick McAllister explores Ratnakīrti’s ontology with a particular focus on two
themes. The first is Ratnakīrti’s position that for anything to be real, it must be momentary,
because if something were not momentary, it could not be a cause. The second is Ratnakīrti’s
ontological idealism that nothing exists outside of the mind. As McAllister shows, Ratnakīrti
devoted considerable attention to making sense of cognition and the Buddhist path from a
perspective in which our everyday experience of enduring and external objects is thoroughly
erroneous.
The Mahāyāna view that ordinary persons perceive conventional reality and take it to be
ultimately real, while a buddha does not perceive conventional reality, raises the problem of
how a buddha can communicate with and help ordinary persons. According to Ratnākaraśānti
(eleventh century), grasping subjects and the objects they grasp are both imagined. They are
representational forms that arise in the consciousness of an ordinary person, but a buddha is
precisely free from these erroneous representations and perceives reality as it ultimately is.
As Gregory Max Seton shows, Ratnākaraśānti makes the controversial argument that, moti-
vated by compassion, a buddha purposefully retains a degree of cognitive error, which enables
perception of the external objects and the grasping subject even as a buddha simultaneously
knows these representational forms are erroneous. Seton focuses on the account of conscious-
ness and the theory of error Ratnākaraśānti employs as he attempts to develop an ontology
and representationalist epistemology that would constitute the correct interpretation of both
Nāgārjuna’s Madhyamaka and Asaṅga’s Yogācāra.
We can see another kind of syncretic project in the work of Atiśa (982–1054), who inte-
grates Candrakīrti’s and Bhāviveka’s methodological approaches to Nāgārjuna’s Middle Way
philosophy, as well as teachings on the bodhisattva path and esoteric tantric practices. While
Atiśa primarily follows Candrakīrti’s reductio ad absurdum method of demonstrating contra-
dictions in the views of others, like Bhāviveka, he employs reasoning as an aid to cultivate
insight, even as he eschews the approach to valid means of cognition ( pramāṇa) developed
by Dignāga and Dharmakīrti. The result, as James B. Apple suggests, is a form of nominal-
ism, according to which cognitions and objects are mere appearances and not ultimately real
because they are mental imputations. Even the mind, itself dependently originated, is also a
nominal designation, an appearance that cannot be established as inherently existing. Atiśa
544
calls his approach the “Great Middle Way,” integrating a multiplicity of teachings and prac-
tices to support progress on the path to a nondual awakening, ultimately letting go of even
nonconceptual wisdom.
Abhayākaragupta (twelfth century) was active at both Nālandā and Vikramaśīla, the two
most important centers of Buddhist learning in North India. His writings typically consist
of passages from earlier Buddhist texts, without naming the author or text, compiled into a
coherent whole. Despite what might be regarded as a lack of originality, his ordering of ideas
proved influential in both India and Tibet. In his chapter, Kazuo Kanō explores five themes
in Abhayākaragupta’s texts: the scriptural authority of Mahāyāna texts, the two truths, the
two truths and an evaluation of the mind-only doctrine, the illusory but beneficial mind, and
Buddha-nature. According to Abhayākaragupta, the doctrine of the Two Truths is the founda-
tional Buddhist view: all other Buddhist doctrines need to be understood in the framework of
the Two Truths. Abhayākaragupta applies this framework to the mind and the achievements
of a bodhisattva, which are both beneficial even if they are illusory because they can result in
ordinary beings being freed from suffering. Thus, he concludes, even if something exists only
conventionally, if it is beneficial, it should be accepted in some form as true.
545
33
JITĀRI
A Later Buddhist Master of Debate1
Junjie Chu
Jitāri2 is a renowned Buddhist philosopher who exercised a strong influence on the later
period of Indian Buddhist philosophy. He is regarded by Durvekamiśra (c. 970–1030), an
important commentator in his period, as his teacher or guru (Dharmottarapradīpa 149.24,
Hetubinduṭīkāloka 411.2); the initial formal reasoning of his “proof of the omniscience of
the Buddha,” as well as the conclusion of his “proof of cognitive invalidity of the Veda,”
are quoted by Ratnakīrti (Ratnakīrti-Nibandhāvaliḥ 31,13–17; 101,16–17), another influential
philosopher in the eleventh century. Further, in the Tibetan tradition, Jitāri was said to be the
teacher of Atiśa (cf. Roerich 1949, 243), a key person in the second transmission of Buddhism
to Tibet in the eleventh century.
Unfortunately, we know very little about his life due to the lack of reliable and precise
chronological materials. His date can be fixed only conjecturally on the basis of Tibetan his-
toriographical sources. According to studies of these materials by modern scholars (Vidyab-
husana 1921, 337f.; Iyengar 1952, viii; Tucci 1956, 249–54, 1971, 249ff.; Shirasaki 1981b,
345–42), the information concerning Jitāri’s date can be summarized as follows: Jitāri’s father
lived at the court of King Sanātana of Varendra, a vassal of the Pāla kings. It is said that Jitāri
received the title of Paṇḍịta from King Mahāpāla who, according to some sources, reigned
until 940 CE. As mentioned, it is also said in the Tibetan sources that Jitāri was the teacher of
Atiśa when the latter was a youth. Atiśa was born in 980 and arrived in Tibet in 1042. Com-
paring these dates, Jitāri is thought to have flourished in the second half of the tenth century
(c. 940–1000).
The early seventeenth-century Tibetan scholar Tāranātha’s History of Indian Buddhism
(rGya gar chos ’byung, GGCB) includes a passage describing Jitāri’s life and conveying the
following two significant points:
(1) Jitāri is closely connected with the tantric practice of Mañjughoṣa, a tantric form of Mañjuśrī.
It is said that Jitāri’s father sNying po’i zhabs (Garbhapāda) performed the tantric initiation
of Guhyasamāja for the king Sanātana. For this reason, when Jitāri was sent to a school at
the age seven, he was persecuted by other Brahman boys, and was told: “Being a Buddhist
tantric practitioner your father gave the low caste (śūdra) queen a higher status and, while
worshipping, he mixes the low and high castes without discrimination.” To subdue them,
Jitāri’s father decided to bestow the tantric empowerment of Mañjughoṣa on him and let him
DOI: 10.4324/9781351030908-48 547

Junjie Chu
practice. Tāranātha sums up Jitāri’s achievement of tantric practice: “He learned whatever
his father had mastered – the Guhyasamāja, Cakrasaṃvara, Hevajra, etc. He also attended
many other teachers. Uniquely, he was able to listen to all doctrines from Śrī Mañjughoṣa
himself ” (cf. GGCB 233,6‒234,17, 234,9–11; Chimpa and Chattopadhyaya 1970, 290–92).
(2) With regard to Jitāri’s Mahāyāna scholarship, he is affiliated with Śāntideva’s Madhyam-
aka teaching tradition. Tāranātha concludes: “He taught a lot on various doctrines and his
fame spread widely. He composed short commentaries on the Compendium of Training
(Śikṣāsamuccaya), Engaging in Bodhisattva Conduct (Bodhicaryāvatāra), the Essence of
Space Sūtra (Ākāśagarbhasūtra), etc. He composed about one hundred treatises on many
sūtras and tantras” (GGCB 235,2–235,4; cf. Chimpa and Chattopadhyaya 1970, 292).
Sadly, we have no other historical materials to prove the reliability of this description. However,
the first point – that is, Jitāri’s association with Mañjughoṣa – is reflected in several introductory
verses in his own works.3 For example, the verse of praise (maṅgalācaraṇa) at the beginning of
his Topics of Debate (Vādasthānāni), which is also included in the eleventh-century anthology
of court poetry known as the Jewel Treasury of Elegant Sayings (Subhāṣitaratnakośa), reads:
May he who consecrates his worshiper into the kingdom of his law with anointing
liquid, golden red as saffron, poured from the golden amphora, his foot, wherein his
lovely toes are ceremonial buds: may Mañjuśrī watch over you for your happiness
and good.
(Translation in Ingalls 1965, 67; Sanskrit in
Kosambi and Gokhale 1957, 6, 9–12.)
Concerning the second point – that is, his relation to the Madhyamaka tradition – to my
knowledge, clear support can be found only in his Verses on the Classification of Buddhist
Systems and its commentary (Sugatamatavibhaṅgakārikā and Sugatamatavibhaṅgabhāṣya;
the Madhyamaka section is translated in Shirasaki 1986a, 26ff.). This is a doxographic
work which follows the pattern of Āryadeva’s Compendium on the Essence of Knowledge
(Jñānasārasamuccaya, JSS), explaining the four Buddhist philosophical systems, namely the
Sarvāstivāda, Sautrāntika, Yogācāra, and Madhyamaka. As seen previously, Tāranātha men-
tions that Jitāri wrote commentaries on Śāntideva’s Compendium of Training and his Engaging
in Bodhisattva Conduct. According to Ruegg, both Āryadeva and Śāntideva are related to the
religio-philosophical movement of the synthesis of Madhyamaka and Vajravāya (Ruegg 1981,
106); that is to say, they are the Madhyamaka philosophers who are influenced by tantric or
Vajrayāna theory and practice. In Verses on the Classification of Buddhist Systems verses 7–8
(D 8a2–3), which are almost identical to JSS verses 27–28,4 it is said clearly that the aware-
ness (vijñāna) postulated in the Yogācāra system is not accepted by the Mādhyamika as the
ultimate reality because it is devoid of either a single or a multiple essence (ekānekasvabhāva),
just like a sky-lotus, and the latter accepts only the reality that is free from the four extremes
of existence, nonexistence, both, and neither (Verses on the Classification of Buddhist Systems
verses 8a3–4; JSS 26b7–27b1; cf. Mimaki 1976, 189). An extended discussion on this topic
can be found in Commentary on the Verses on the Classification of Buddhist Systems (46b5ff ).
Based on Verses on the Classification of Buddhist Systems and its commentary, there seems
to be no problem with regarding Jitāri as a Mādhyamika. However, disagreements arise with
regard to Jitāri’s opinion about the cognitive image of objects (ākāra). In his commentary on
Verses on the Classification of Buddhist Systems verse 7, Jitāri criticizes Dharmottara’s posi-
tion. The controversy is based on their different interpretations of the famous Dharmakīrtian
548
Jitāri
argument that a cognition and its object are not different because they are necessarily per-
ceived together (sahopalambhaniyama). Matsumoto concludes that “Jitāri should be regarded
as a Mādhyamika who advocates Satyākāravāda,” that is, the theory that cognition bears the
true image of an object (Matsumoto 1981, 966). Shirasaki has a different opinion: he considers
Jitāri a Mādhyamika who advocates Nirākāravāda, the theory that cognition does not bear the
image of its object (cf. Shirasaki 1986a, 14ff.).5
It is remarkable that the previously mentioned studies are based mainly on Jitāri’s doxo-
graphic works and in some cases are influenced by later Tibetan doxographic works as well.
However, more convincing conclusions could be made by comparing these doxographic trea-
tises to Jitāri’s own philosophical works.
Of Jitāri’s philosophical works, although a considerable number were brought to Tibet, only
a few were translated into Tibetan. The following works are included in the Tibetan canon:
1. Ascertainment of Property and the Property-Bearer (Dharmadharmiviniścaya = Chos dang

chos can gtan la dbab pa, D4262; cf. the synopsis in Kyuma 2003). In the initial verse, the
topics of the treatise are summarized. The distinction between property and the property-
bearer is viewed from five vantage points: (1) the natures, (2) the different aspects, (3) the cat-
egorical divisions, (4) the particular determinations, and (5) the etymological interpretations.
2. Logic Introduced to Unlearned Persons (Bālāvatāratarka = Byis pa ’jug pa’i rtog ge zhes
bya ba, D4263; critically edited in Shirasaki 1981a, 32–52). This is a treatise on episte-
mology for beginners which follows Dharmakīrti’s Ascertainment of the Means of Valid
Cognition (Pramāṇaviniścaya) and Drop of Reason (Nyāyabindu). It is divided into three
chapters: perception ( pratyakṣa), inference for oneself (svārthānumāna), and inference
for others ( parārthānumāna).
3. Instruction on the True Nature of a Reason (Hetutattvopadeśa = gTan tshigs kyi de kho na
nyid bstan pa, D4261; the Sanskrit edition = Tucci 1956, 261–74; Sanskrit-Tibetan and
Japanese indices are available in Miyasaka 1964). As Tucci points out in his introduc-
tion, in this treatise, Jitāri follows closely the Entry to Reason (Nyāyapraveśa), a work of
Śaṅkarasvāmin, the disciple of Dignāga. It starts with the introductory verse of the Entry
to Reason and it follows the content and structure of the latter very closely, even repeating
verbatim many sentences from the latter. This shows that Dignāga’s logic had not been com-
pletely suppressed by that of Dharmakīrti even as late as the tenth or early eleventh century.
It is worth mentioning that, concerning the content, text structure, and argumentative style,
these works are very different from his Topics of Debate, to which we now turn.
As has been reported elsewhere,6 the present writer now has access to two newfound manu-
scripts (hereafter named “A” and “B”) of a philosophical work attributed to Jitāri that had
been brought to Tibet. The work contains twenty-one chapters, of which six have previously
been edited and published by other scholars; the remaining fifteen are completely new.7 At the
beginning of the work, after the salutation to the Buddha and the verse of praise already cited
above, Jitāri prefaces his work in the initial verse as follows:
In compliance with the wish of friends, putting my shyness aside, some topics of
debate (vādasthānāni) [between Buddhists, Brahmans, and Jainas] are written here
according to my ability, according to my recollection.
It would thus seem that Topics of Debate was the title of the collection as a whole. However,
titles of philosophical works in Sanskrit do not usually appear in plural form, and the term
549
Junjie Chu
might also have been used merely as a description of the content of the work.8 Since there is
no general colophon at the end of the work, no certainty on this matter can be arrived at. Thus,
for lack of anything better, Vādasthānāni or Topics of Debate will be used as the title of the
work.9 Although no general colophon is present, the end of each chapter does include a colo-
phon containing the name of the author and the title of the treatise. Needless to say, colophons
cannot always be relied on to convey the original title of a work. In the case under discussion,
it is clear that they were not written by Jitāri himself, whose name appears in honorific forms
such as jitāripāda and mahāpaṇḍitaśrījitāripāda (always in the plural), which he was highly
unlikely to have used to refer to himself. Furthermore, if we take titles appearing in the colo-
phons as the titles of each chapter, there would be two cases where a title was doubly used for
two different texts; this suggests also that the colophons were added by different scribes later
on. In addition to the titles appearing in the colophons, however, a list of the chapters is written
on the cover page of manuscript A. Unfortunately, because of poor legibility, only seventeen
titles can be clearly identified. In comparison with the titles that appear in the colophons, those
contained in the list have one obvious advantage: they describe the contents or the focus of
each chapter more precisely, and there is no repetition of titles. For these reasons, in what fol-
lows, I use the list to identify each chapter, supplementing the illegible part of the list with the
titles found in the colophons.
Topics of Debate is now considered a single work containing multiple chapters, since Jitāri
gives them a common title by composing the initial verse for all of them. However, the chapter
divisions were not made beforehand – as a matter of fact, these chapters are independent in
content from one another, exhibiting a variety of topics or perspectives. It is clear that they
also circulated singly and in various combinations. Although Jitāri gives regular indications
of the changes of subject matter with an introductory sentence usually placed at the beginning
each chapter – something like “Here, at the beginning, the theory of the generic property is first
refuted” (A1b2) and “Now, the nonexistence of the Self is proved” (A11b3) – he does not seem
to make any reference to any formal division of his work. All this strengthens the hypothesis
that we are indeed dealing here with a set of single compositions which have been collected
together later for convenience.
In Topics of Debate, Jitāri addresses a wide range of topics, most of which have been
subjects of long-lasting debates among Buddhist, Brahmanical, and Jaina philosophers, with
occasional discussions originating from inside the Buddhist system. He generally starts his
discussions with a statement of formal reasoning ( prayoga), which forms the main argument,
or the main topic, of the chapter. This is usually followed by the refutation of various objec-
tions that claim that the reason Jitāri uses in his reasoning is invalid in one of three usually
recognized ways: by being “unestablished” (asiddha), “inconclusive” (anaikāntika), or “con-
tradictory” (viruddha). In what follows, I list twenty of the twenty-one chapters10 according to
the order they appear in manuscript A. Each title is accompanied by a short description of the
main topic of the work, based chiefly on the available formal reasoning at the beginning of the
chapter. Because the critical edition of the most of these chapters is not yet complete, and due
also to space limitations, a more precise or comprehensive description is not possible.
Here is a list of the contents of Jitāri’s Topics of Debate:
1. Negation of the Existence of Universals (Jātyādiniṣedha; A1b1–5b6, B1b1–6b3 = Jātinirākṛti

in Tucci 1971, 251–54; Iyengar 1952, 72–80; Bühnemann 1985, 30–38; cf. Shirasaki
1983). The generic property is unreal because it cannot be said to be different from or
identical to the individual. All things that cannot be said to be different from or identical
to the individual are not real, just like a sky-lotus.
550
Jitāri
2. Refutation of the Existence of Universals (Sāmānyanirākṛti; A6a1–8a5, B6b3–10a1). The

universal cannot be treated as existent in its related particulars, because it is not thus
perceived. When one thing is not perceived to be existent in another thing, even when the
conditions for its perception are fulfilled, this thing cannot be treated as existent in that
thing, just like a horn on the head of a horse.
3. Refutation of the Existence of God (Īśvaranirākaraṇa; A8a5–11b3 = Bühnemann 1985,
39–43;11 cf. Shirasaki 1995). The refutation is centered around the formal reasoning used
by the opponent to prove the existence of God: All products are perceived to be insepa-
rable from their producer, just like a pot. And products in the world such as a body, and
so on, have also an creator, namely God. The refutation comes to the conclusion that the
world’s diversity is caused by karma and does not have a single universal producer.
4. Proof of the Nonexistence of the Self (Nairātmyasiddhi; A11b3–12a3, B10a1–10b3 = Büh-
nemann 1985, 29). A living body cannot possess a Self, because it does not have any
connection with a Self. Anything that is not connected with another thing cannot possess
that thing through the connection, just like the Himalaya mountains with the Malaya
mountains.
5. On the Cognitive Invalidity of the Veda (Vedāprāmāṇya; A12a3–14b4 = Vedāprāmāṇyasiddhi
in Bühnemann 1985, 23–26; quoted in its entirety in Ratnakīrti-Nibandhāvaliḥ 99,16–
101,17, as identified by Bühnemann 1985, 13; cf. Eltschinger 2003). The Vedic verbal
testimonies are not valid cognitions concerning external objects, because they have no
connection with them. Anything that has no connection with another thing cannot be a
valid cognition for that thing, just like an ass cannot be the valid cognition for fire.
6. Proof of the Existence of Mere Cognitive Representation (Vijñaptimātratāsasiddhi;
A14b4–20a6, B49a1–55b2, cf. Chu 2020). In contrast to his Verses on the Classifica-
tion of Buddhist Systems and its Commentary, in this chapter, Jitāri aims to prove the
Yogācāra theory that all things that become manifest or appear in cognition are only cogni-
tion’s self-manifestation. He does so by refuting the theory of existence of external object
(bahirarthavāda), which is divided into two branches: one accepts that cognition possesses
an image of the external object, and the other does not accept that cognition possesses
an image of the external object. The main target of his refutation is the second one, that
is, the theory that cognition does not possess an image (nirākārajñānavāda). His formal
reasoning runs: Anything that becomes manifest in cognition is cognition itself, just as the
conceptual construction of a blue thing. And a visible thing, and so on, becomes manifest
in cognition; therefore, this is an instance of a cognition with the image of objects. In the
context of countering the attack that the reason is “inconclusive,” Jitāri refutes the oppo-
nent’s thesis that the manifestation of the external object in cognition is caused by cogni-
tion. He does so by using an argument in the form of “four alternative propositions,” that
is, negating all four logically possible propositions with regard to the relationship between
the manifestation and the object – difference, non-difference, both, and neither – with the
conclusion that the manifestation of the object is merely the cognition’s self-manifestation.
And thus, the reason, that is, “becoming manifest,” cannot be inconclusive, since the mani-
festation never occurs in any other way, and the thesis that is ascertained through the reason
free from all fallacies, each of which is proved not to be existent, should be accepted.
7. Negation of the Naiyāyika’s Concept of the Whole Consisting in Parts of a Gross Object
(Avayaviniṣedha; A20a6–24b6, B41a1–46b1). All seen things like a pot cannot be regarded
as being single, because they are involved with contradictory properties; whatever is
involved with contradictory properties cannot be a single thing, just like a variegated cloth,
and so on.
551
Junjie Chu
8. Proof of the Theory of Exclusion (Apohasiddhi; A24b6–32b1, B32a1–40b3). Every

expression (vācaka) in verbal form such as “pot” refers to a superimposed single excluded
thing, because every expression takes a superimposed single excluded thing as its sphere
of reference, just like a nonexistent verbal form.
9. On the Non-Perception of the Pervader (Vyāpakānupalambha; A32b1–46a1). At the
beginning, Jitāri says that he composes the treatise for the purpose of examining the
proper nature of the means of cognition that invalidates the occurrence of the reason
“being existent” in the opposite of what is to be proved (viparyayabādhakapramāṇa),
which is applied in the proof of the Buddhist doctrine of the momentariness.12 Accord-
ing to Jitāri, this should not refer to the perception of the contrary of the pervader
(vyāpakaviruddhopalambha); rather, it should be understood as the non-perception of the
pervader (vyāpakānupalambha), which is in line with Dharmakīrti’s intention.
10. Negating the Superiority of Brahmanhood (Brāhmaṇyaniṣedha; A46a1–57b4). Jitāri
questions the opponent who assumes to be a twice-born (i.e., a Brahman or other member
of the first three classes) about his high self-esteem, which is claimed to be connected
to the property not in common with the Śūdra, and so on. Jitāri negates the existence of
such a distinctive property; he lists all possible interpretations of this property (i.e., the
birth, the activity, the family, or the special capacity), then he negates each of them. For
example, he negates the special birth with the formal reasoning: That which is not known
through a means of valid cognition cannot be treated as existent, just like a sky-lotus. The
birth uncommon to Śūdra, and so on, is not known through the means of valid cognition
in those who assume to be twice-born. He draws the conclusion that the designations of
the four traditional classes (varṇa) of Brāhmaṇa, Kṣatriya, Vaiśya, or Śūdra are only dif-
ferent in names, which are devoid of connection to any real substance, and one should not
adhere to them merely due to the merely nominal difference.
11. Examination of the Theory of Non-Momentariness (Akṣaṇikavādavicāra; A57b4–62b2,
B18a – 22b). The proponents of non-momentariness are divided into two groups accord-
ing to whether they regard things as permanent or not. The first group holds that the per-
manent things exist without a cause, like ether; they can never and in no way come to a
cessation of their proper nature. The second group maintains that things exist with a cause
and have an impermanent nature, so they come necessarily to destruction at any time in
any manner, but they are not momentary, because when they would cease to attain unin-
terruptedly their proper nature, everything would thus become momentary. These two are
refuted in detail.
12. Proof of the Omniscience of the Buddha (Sarvajñasiddhi; A62b2–64a4, B11b –
117b3 = Bühnemann 1985, 27–28). The Buddha’s speech, like “all produced things are
momentary,” is a valid cognition that is reliable and has an ascertained objects because
every speech that is non-belying and has an ascertained object presupposes directly or
indirectly the knowledge that apprehends intuitively its object, like the expression “fire
burns.” Note that here, “omniscient” refers to “knowing all that is useful” but not “know-
ing everything whatever” (cf. McClintock 2010, 133 and n. 338; Moriyama 2011, 2014,
84–86). In this chapter, Jitāri proves that the Buddha’s teachings which are useful for
liberation are reliable.
13. On the Theory of Backward Causation (Bhāvikāraṇavāda; A64a4–69b3, Franco 2015,
90–99). Whatever has a positive and negative concomitance followed by another thing is
the cause of that thing, just like fire is the cause of smoke. And every future effect has nec-
essarily positive and negative concomitances followed by its cause, so this future effect
is the cause of that cause. Here, Jitāri emphasizes that the determining factor of a causal
552
Jitāri
relationship consists in the positive and negative concomitance, not other factors, such as
the temporal sequence of their manifestation.
14. On the Theory of the Generic Property or Universal (Jātivāda; A69b3–70b3). The
existence of the generic property cannot be proved, because it is not capable of pro-
ducing an effect. Everything that is not capable of producing an effect cannot be
proved to be existent, even by taking existence as its basis (sattāśrayatvena), just like
a sky-lotus.
15. Proof of the Human Authorship of the Vedic Injunction (Śrutikartṛsiddhi; A70b3–77b4,
B26b – 131b3). Vedic statements, such as “a person who wants to go to heaven should
scarify the fire-offer,” are human products. Every verbal statement is a human product,
just like a verbal statement of a passerby.
16. On the Invalidity of Vedic Verbal Testimony (Śabdāprāmāṇya; A77b5–85b4). Verbal tes-
timonies can be a valid cognition only because they lead to attain the object of the state-
ment, not due to being related to the mere intention of the speakers . . . (the remaining part
of the sentence is illegible).
17. Destruction of the Causal Complex (Sāmagrībhaṅga; A85b5–87b1). The opponent holds
that object and its cognition have a subject-object relationship on account of their being
commonly dependent on the same causal complex (ekasāmagrī), that is, the sense facul-
ties. This is refuted on the basis of the nonexistence of such a complex. According to
Jitāri, a thing that is whole is called “complex,” but the sense faculty cannot be the whole
cause, because this can neither be proved by perception, since it is supersensory, nor by
inference, since inference is preconditioned by perception.
18. Introduction to Momentariness (Kṣaṇabaṅgopanyāsa; A87b1–90a5). All things, such as
visible things, visual senses, and so on, are momentary because they are existent; when a
thing is not momentary, it is not existent, just like in the case of a sky-lotus.
19. Proof of Momentariness (Kṣaṇabhaṅgasiddhi; A90b1–93b4). The opinions of Arcaṭa
and Dharmottara are reported, and they are labeled advocates of the “internal pervasion”
(antarvyāptivādin) and the “external pervasion” (bahirvyāptivādin), respectively. These
are the two theories about the nature of the pervasion of the reason “existence” by the
property to be proved “momentariness.” In the case of the “internal pervasion,” the perva-
sion is apprehended in the subject of the inference itself without reference to an external
case, so no example is necessary, whereas in the case of the “external pervasion,” the
pervasion is apprehended outside of the subject of inference with reference to an external
case, that is, something other than the subject that is usually indicated with an example.13
Arcaṭa’s reasoning runs: Whatever is existent is momentary, and a visible thing, visual
sense, and its cognition, and so on, are existent; therefore they are momentary. In this
reasoning, the pervasion is apprehended within its subject “every existent thing,” so no
other example is given. Furthermore, according to this theory, the invalidity of the reason
is excluded by the means of valid cognition, which invalidates the occurrence of the rea-
son in the opposite to the thesis to be proved, as it is formulated as follows: A thing that
does not have the capacity to produce an effect in a successive or non-successive way is
not capable of producing an effect at all, just like a sky-lotus. And a non-momentary thing
is not capable of producing an effect in a successive or non-successive way, so it is not
capable of producing any effect and thus is not existent, too. This is because the nature
of being existent is equal to the nature of being capable of producing an effect, and this
capacity is pervaded by successiveness or simultaneousness in producing an effect. On
the other hand, Dharmottara’s reasoning is as follows: Whatever is existent is momentary,
just like a pot. And these pots, and so on, are existent, so they are momentary. Here, in
553
Junjie Chu
contrast to Arcaṭa’s reasoning, an example is given. But the necessity of the example in
this reasoning is questioned; a detailed discussion is provided.
20. Refutation of Non-Absolutism (Digambaramataparīkṣā; A93b4–97b5, B23a – 125b3 =
Anekāntavādanirāsa in Iyengar 1952, 80–85; cf. Shirasaki 1974; Toshiaki 1978). In the
Jaina system, it is maintained that the substance (dravya) keeps its intrinsic nature, while
the modes ( paryāya) of that substance can be different, just like gold keeps its essential
nature, while its modes can be various gold ornaments. Jitāri refutes this by pointing out
the incompatibility between a single substance and multiple modes.
In all these chapters, Jitāri defends the most important Buddhist doctrines and refutes vari-
ous non-Buddhist theories by following Dharmakīrtian epistemologico-logical theoretical
patterns and argumentative methods. Although Jitāri does not seem to offer many innovative
ideas or clearly independent theories, the Topics of Debate should still occupy an important
place in the history of post-Dharmakīrtian Yogācāra philosophical literature, as it summarizes
or discusses from a new perspective the most important topics that later Indian Buddhists
were debating with various non-Buddhist systems. The main target of his criticisms should be
the Mīmāṃsaka, and he mentions Kumārila by name many times. There are also other cases
where he has as his opponents different Buddhist interpreters of certain theories. This enriches
significantly our knowledge about later development of the Yogācāra system and its interac-
tion with various non-Buddhist or Buddhist opponents.
With regard to Jitāri’s philosophical position in the Topics of Debate, in my view, he is a
Yogācāra who follows Dharmakīrti’s system, although he may be influenced by the Madhyamaka
system to some extent. In many chapters, his Yogācāra background is quite clear. For instance, in
chapter 6, Proof of the Existence of Mere Cognitive Representation, he proves that all manifesta-
tions of external objects are merely the self-manifestation of cognition. So, in this chapter, it is
more natural to say that he establishes the fundamental theory of the Vijñānavāda or the Yogācāra
system that the mere cognitive representation (vijñapti) or awareness (vijñāna) exists rather than
that he negates it from the Madhyamaka point of view. That task is reserved for his Verses on
the Classification of Buddhist Systems and its commentary, where he says unmistakably that the
“awareness” regarded in the Yogācāra system as the ultimate reality is not accepted from the
Madhyamaka point of view. As for his opinion about the cognitive image of an object, it is also
clear that he maintains the theory that awareness has the image of the object (sākāravāda) on the
basis of the Vijñānavāda or the Yogācāra system rather than supporting the theory that cogni-
tion does not have the image of object (nirākāra) from the Madhyamaka point of view, as he is
interpreted to hold in his Verses on the Classification of Buddhist Systems and its Commentary.
In any case, his position in the Topics of Debate is different from that in the Verses on
the Classification of Buddhist Systems and its Commentary. According to my present knowl-
edge, in the former, no statement can be found in agreement with those in the latter. Indeed,
the argument “being devoid of the single or multiple essence” used by Śāntarakṣita in his
Madhyamakālaṃkāra 1, which is often regarded as the indication of author’s Madhyamaka
position, is not found to be used in the Topics of Debate, and “four alternatives” is often used
to negate various non-Buddhist theories, but it is not found to be used against the Yogācāra. If
we do not assume that there are two philosophers who bear the same name, we have to con-
sider the possibility that Jitāri articulates his different philosophical positions in his two works.
In his philosophical work Topics of Debate, he refutes various mainly non-Buddhist theories
from the Yogācāra point of view, and in his doxographical work Verses on the Classification
of Buddhist Systems and its Commentary, he refutes Yogācāra theory from the Madhyamaka
position. In this reading, we can say that he speaks differently to different audiences. Another
554
Jitāri
possibility, however, is that these two works were composed in different periods and that in
between these two periods, his thought underwent a radical change from the one philosophical
perspective to the other. Since the edition and the philosophico-historical studies on the Topics
of Debate are still in the early stage, no decisive conclusion can be offered.
Notes
1 I am indebted to Dr. Sara L. McClintock, whose valuable comments and suggestions improved my
English and made the presentation more clear and precise.
2 Regarding the Sanskrit form of his name, although Tibetan sources suggest something like Jetāri, on
the basis of colophons in the published texts as well as in the newfound manuscripts, the correct form
must be Jitāri, with dGra las rnam rgyal or dGra las rgyal ba as the Tibetan translation.
3 Additionally, a large number of tantric works are attributed to him (cf. Shirasaki 1981b, 336–28).
However, these are beyond the scope of the present study.
4 The Sanskrit verses of JSS 27 and 28 are quoted in Mokṣākaragupta’s Language of Logic
(Tarkabhāṣā; 70,4–5) and in Prajñākaramati’s Commentary on Engaging in Bodhisattva Conduct
(Bodhicaryāvatārapañjikā, 174,11–12) respectively. See Mimaki 1976, 188.
5 Another related topic discussed is how Jitāri regards Dharmakīrti’s thought. According to Ruegg 1981,
100; Shirasaki 1986b, Jitāri regards Dharmakīrti as a Mādhyamika master. This conclusion gives rise
to an interesting discussion in Steinkellner 1990, where the author points out that there is not sufficient
evidence to prove Dharmakīrti a Mādhyamika, and the source for Jitāri’s view is Śāntarakṣita’s Com-
mentary on the Ornament of the Middle (Madhyamakālaṃkāravṛtti; Steinkellner 1990, 82f.).
6 See Chu and Franco 2016. In the following, some discussions related to the manuscripts are based on
that article, with some information updated.
7 According to Kellner et al., there is a third manuscript of Jitāri’s Proof of the Existence of Mere
Cognitive Representation (Vijñaptimātratāsiddhi; see no. 6), and there is also a copy of the Delibera-
tion on the Rule of the Necessary Co-Perception of a Cognition and Its Object (Sahopalambhaniy-
amasamarthana) which is also ascribed to Jitāri. See “Preface” in Kellner et al. 2020, xiv. However,
these are not available to the present author.
8 See Dignāga’s initial verse in Pramāṇasamuccaya 1.1ab: “for the purpose of establishing the valid
means of cognition, from my earlier ideas scattered in various treatises, here a single collection (sam-
uccaya) is composed.”
9 Note that in Iyengar 1952, “Vādasthānāni” is also used as the title of Jitāri’s work which contains
only two sections: Jātinirākṛti and Anekāntavādanirāsa.
10 The last chapter, Negation of Ungrammatical Language (Apaśabdanirākṛti; A112b2–7 = B47a1–
48b5), is not included because the legibility in both manuscripts is very poor, and for the time being,
I am unable to offer an understandable transliteration.
11 In the manuscript Bühnemann uses, the title is missing; a tentative title “*Īśvaravādimataparīkṣā” is
suggested, see Bühnemann 1985, 19.
12 For the technical term and the related topics, see Steinkellner 1991, 314ff.; Shiga 2011; Sakai 2019,
2020.
13 For more detailed explanations of these two terms, “internal pervasion” and the “external pervasion,”
see Mookerjee 1935, 398–400; Kajiyama 1958, 362.
Bibliography
Āryadeva. Tibetan translation of Jñānasārasamuccaya (Āryadeva), see D 3851=P 5251.
Bühnemann, Gudrun, ed. 1985. Jitāri: Kleine Texte beschrieben und ediert. 2nd expanded ed. Wien:
Arbeitskreis für tibetische und buddhistische Studien, Universität Wien.
Chimpa, Lama, and Chattopadhyaya, Alaka, trans. 1970. Taranatha’s History of Buddhism in India. New
Delhi: Motilal Banarsidass.
Chu, Junjie. 2020. “Jitāri’s Vijñaptimātratāsiddhi.” In Reverberations of Dharmakīrti’s Philosophy, Pro-
ceedings of the Fifth International Dharmakīrti Conference Heidelberg, August 26 to 30, 2014, edited
by Birgit Kellner, Patrick McAllister, Horst Lasic, and Sara McClintock, 1–19. Vienna: Austrian
Academy of Sciences Press.
555
Junjie Chu
Chu, Junjie, and Eli Franco. 2016. “Rare Manuscripts of Works by Jitāri.” In Sanskrit Manuscripts in
China II, edited by Xuezhu Li, Horst Lasic, Xuezhu Li, and Horst Lasic, 15–48. Beijing: China Tibe-
tology Publishing House.
Daisetz, T. Suzuki, ed. 1957. The Tibetan Tripitaka, Peking Edition – Kept in the Library of the Otani
University, Kyoto. – Reprinted under the Supervision of the Otani University. Tokyo-Kyoto: Tibetan
Tripitaka Research Institute.
Eltschinger, Vincent. 2003. “La Vedāprāmāṇyasiddhi de Jitāri. Introduction et traduction.” Journal Asia-
tique 291 (1–2): 137–72.
Franco, Eli. 2015. “Jitāri on Backward Causation (bhāvikāraṇavāda).” In Buddhist Meditative Praxis:
Traditional Teachings and Modern Application, edited by K. L. Dhammajoti, 81–116. Hong Kong:
Centre of Buddhist Studies.
Ingalls, Daniel H. H., ed. 1965. An Anthology of Sanskrit Court Poetry, Vidyākara’s “Subhāṣitaratnakośa.”
Cambridge, MA: Harvard University Press.
Iyengar, H. R., ed. 1952. Tarkabhāṣa and Vādasthāna of Mokṣākaragupta and Jitāripāda. Mysore: Hin-
dusthan Press.
Jitāri. Tibetan translation of Sugatamatavibhaṅga (Jitāri), D3990)=P 5461.
———. Tibetan translation of Sugatamatavibhaṅgakārikā (Jitāri), see D3899=P 5296.
Kajiyama, Yuichi. 1958. “On the Theory of the Intrinsic Determination of Universal Concomitance in
Buddhist Logic.” Journal of Indian and Buddhist Studies 7 (2): 360–64.
Kellner, Birgit, Patrick McAllister, Horst Lasic, and Sara McClintock, eds. 2020. Reverberations of
Dharmakīrti’s Philosophy, Proceedings of the Fifth International Dharmakīrti Conference Heidel-
berg, August 26 to 30, 2014. Vienna: Austrian Academy of Sciences Press.
Kosambi, D. D., and V. V. Gokhale, eds. 1957. The Subhāṣitaratnakośa, Compiled by Vidyākara. Cam-
bridge, MA: Harvard University Press.
Kyuma, Taiken, ed. 2003. “Jitari ni kiserareru Dharmadharmiviniscaya ni tsuite [On Dharmadharmivi-
niscaya ascribed to Jitāri].” Indogaku Bukkyōgaku Kenkyū 51 (2): 900–4.
Malvania, Dalsukhbhai, ed. 1955. Paṇḍita Durveka Miśra’s Dharmottarapradīpa, Being a Subcommen-
tary on Dharmottara’s Nyāyabinduṭīkā, a Commentary on Dharmakīrti’s Nyāyabindu. Patna: Kashi
Prasad Jayaswal Research Institute.
Matsumoto, Shiro. 1981. “On the Philosophical Positions of Dharmottara and Jitāri.” Journal of Indian
and Buddhist Studies/Indogaku Bukyōgaku Kenkyū 29 (2): 966–69.
McClintock, Sara L. 2010. Omniscience and the Rhetoric of Reason: Śāntarakṣita and Kamalaśīla on
Rationality, Argumentation, and Religious Authority. Somerville, MA: Wisdom Publications.
Mimaki, Katsumi. 1976. La Réfutation bouddhique de la permanence des choses (Sthirasiddhidūṣaṇa) et
la preuve de la momentanéité des choses (Kṣaṇabhaṅgasiddhi), 2 vols. Paris: Institut de civilisation
indienne.
Miyasaka, Yūsho. 1964. “A Sanskrit-Tibetan and Japanese Index to the Hetutattvopadeśa of Jitāri: Com-
paring with the Chinese of Buddhist Logical Terms.” Mikkyō Bunka 68: 31–57.
Mookerjee, Satkari. 1935. The Buddhist Philosophy of Universal Flux: An Exposition of the Philosophy
of Critical Realism as Expounded by the School of Dignaga. Calcutta: University of Calcutta.
Moriyama, Shinya. 2011.“Pramāṇapariśuddhasakalatattvajña, sarvajña and sarvasarvajña.” In Reli-
gion and Logic in Buddhist Philosophical Analysis, Proceedings of the Fourth International Dharma-
kirti Conference Vienna, August 23–27, 2005, edited by Helmut Krasser, Horst Lasic, Eli Franco, and
Birgit Kellner, 329–40. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
———. 2014. Omniscience and Religious Authority. A Study on Prajñākaragupta’s
Pramāṇavārttikālaṅkārabhāṣya ad Pramāṇavārttika II 8–10 and 29–33. Zürich and Berlin: LIT
Verlag.
Roerich, George N., trans. 1949. The Blue Annals (Parts 1–2). Calcutta: Royal Asiatic Society of Bengal.
Ruegg, David Seyford. 1981. The Literature of the Madhyamaka School of Philosophy in India. Wies-
baden: Harrassowitz.
Sakai, Masamichi. 2019. “Antarvyāpti and Antarvyāptivāditva: Arcaṭa on the Buddhist Inference of
Momentariness and Durvekamiśra’s Comparison of Him with Dharmottara.” Journal of Indological
Studies 30/31: 99–128.
———. 2020. “Dharmottara on the viparyaye bādhakapramāṇa and trairūpya in Dharmakīrti’s
sattvānumāna.” In Reverberations of Dharmakīrti’s Philosophy, Proceedings of the Fifth Interna-
tional Dharmakīrti Conference Heidelberg, August 26 to 30, 2014, edited by Birgit Kellner, Patrick
McAllister, Horst Lasic, and Sara McClintock, 375–89. Vienna: Austrian Academy of Sciences Press.
556
Jitāri
Sanghavi, Sukhlalji, and Muni Shri Jinavijayaji, eds. 1949. Hetubinduṭīkā of Bhaṭṭa Arcaṭa with the Sub-
Commentary Entitled Āloka of Durveka Miśra. Baroda: Oriental Institute.
Shiga, Kiyokuni. 2011. “antarvyāpti and bahirvyāpti re-examined.” In Religion and Logic in Buddhist
Philosophical Analysis, Proceedings of the Fourth International Dharmakirti Conference Vienna,
August 23–7, 2005, edited by Helmut Krasser, Horst Lasic, Eli Franco, and Birgit Kellner, 423–36.
Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Shirasaki, Kenjō. 1974. “Jitāri no Anekāntavāda hinhan [The Anekāntavāda of Jitāri].” Indogaku
Bukkyōgaku Kenkyū 22 (2): 999–1003.
———. 1981a. “The Bālāvatāratarka.” Indo Gakuhō 3: 23–52.
———. 1981b. “Jitāri, hito to shisō [Jitāri: Life and Thought].” In Sōden no kenkyū: Kimura Takeo kyōjū
koki kinen, edited by Takeo Kimura, 320–48. Kyoto: Nagata Bunshōdō.
———. 1983. “Jitāri no ‘Fuhen-jitsuzai-ron-hihan [Japanese translation of Jitāri’s Jātinirākr. ti]’.”
Bukkyōshi Kenkyū 26: 1–19.
———. 1986a. “Sugatamatavibhaṅgabhāṣya dai yon shō chūganha no kyō gi wayaku [A Japanese
Translation of the Sugatamatavibhaṅgabhāṣya of Jitāri: Chap. IV on Madhyamaka Doctrine].” Nanto
Bukkyō 55: 1–104.
———. 1986b. “Dharmakīrti wa chūganronsha de aru [Dharmakīrti Is a Mādhyamika].” Bukkyō Ronsō
30: 110–14.
———. 1995. “Jitāri no Īśvaravādimataparīkṣā [The Īśvaravādimataparīkṣā of Jitāri].” Kōbe Joshi
Daigaku Bungakabu Kiyō 28 (2): 105–24.
Steinkellner, Ernst. 1990. “Is Dharmakīrti a Madhyamika?” In Earliest Buddhism and Madhyamaka,
edited by David Seyfort Ruegg and Lambert Schmithausen, 72–90. Leiden: E. J. Brill.
———. 1991. “On the Logic of Svabhāvahetu In Dharmakīrti’s Vādanyāya.” In Studies in the Buddhist
Epistemological Tradition: Proceedings of the Second International Dharmakirti Conference, June,
11–16, 1989, 311–24. Wien: Verlag Der Österreichischen Akademie Der Wissenschaften.
———. 2005. “Dignāga’s Pramāṇasamuccaya, Chapter 1, a Hypothetical Reconstruction of the Sanskrit
Text with the Help of the Two Tibetan Translations on the Basis of the Hitherto Unknown Sanskrit
Fragments and the Linguistic Materials Gained from Jinendrabuddhi’s Ṭīkā.” www.oeaw.ac.at/ias/
Mat/dignaga\_PS\_1.pdf.
Takasaki, Jikidō, Yamaguchi Zuihō, and Hakamaya Noriaki, eds. 1981–1984. Sde dge Tibetan Tripiṭaka
Bstan ’gyur, preserved at the Faculty of Letters, University of Tokyo. Vol. 1–20, Inmyōbu (Tshad –
ma). Tokyo: Sekai Seiten Kankō Kyōkai.
Tāranātha. rGya gar chos ’byung dgos ’dod kun ’byung, spyi’i ang 175 deb ang TBE167001. ’Dzam thang
par ma la gzhi bcol ba yin. Ser byes rig mdzod kyis zhu sgrig byas. www.serajeyrigzodchenmo.org/
Thakur, Anantalal, ed. 1987. Ratnakīrtinibandhāvaliḥ (Buddhist Nyāya Works of Ratnakīrit). Patna:
K. P. Jayaswal Research Institute.
Toshiaki, Tamaru. 1978. “Jitāri no Anekātnavādanirāsa [On Jitāri’s Anekātnavādanirāsa].” Bukkhōgaku
Kenyū 34: 22–40.
Tucci, Giuseppe, ed. 1956. “Hetutattvopadeśa of Jitāri and Tarkasopnāna of Vidyākaranśānti.” In Minor
Buddhist Texts, 247–309. Rome: Istituto italiano per il Medio ed Estremo Oriente.
———. 1971. “The Jātinirākṛti of Jitāri.” In Opera Minora, 249–54. Part I. Roma: G. Bardi. Reprint
of “The Jātinirākṛti of Jitāri.” Annals of the Bhandharkar Oriental Research Institute (Poona) (11)
(1930): 54–58.
Vaidya, Paraśurāma Lakṣmaṇa, ed. 1960. Bodhicaryāvatāra of Śāntideva with the Commentary Pañjikā
of Prajñākaramati. Darbhanga: Mithila Institute.
Vidyabhusana, Satish Chandra. 1921. A History of Indian Logic: Ancient, Mediaeval and Modern
Schools. Calcutta: Calcutta University.
557
34
JÑĀNAŚRĪMITRA
Variegated Non-Duality
Lawrence McCrea
Jñānaśrīmitra (c. 1025 CE) was among the most accomplished and celebrated philosophers
active during the last great flowering of Buddhist intellectual culture in India. He wrote exten-
sively on a variety of philosophical topics in support of positions he took to be grounded in
the logical and epistemological tradition descending from Dignāga and Dharmakīrti. While
his philosophical writings do not appear to have survived in India, he and his works were
still known and quoted by non-Buddhist Indian philosophers centuries later, even after Bud-
dhism had largely disappeared as a serious intellectual force in India; his views are quoted
and discussed, for example, by Udayana (eleventh century), Mādhava (fourteenth century),
and Śaṅkaramiśra (fifteenth century). He is listed as one of the six famous Dvārapaṇḍitas
(“Gateway Scholars”) of the great Buddhist university of Vikramaśīla, and it is clear that he
was a major figure in the educational and public disputational culture of that institution. Many
of the more important Buddhist intellectuals of the succeeding generation declare themselves,
or are said by others, to have been pupils of Jñānaśrīmitra, including Atiśa, Advayavajra,
Ratnakīrti, and Yamāri (on Atiśa, see Tārānātha 1990, 302; on Advayavajra, see Tatz 1987,
698; on Ratnakīrti, see Thakur 1957, 17–18).
Jñānaśrīmitra’s twelve extant Sanskrit philosophical works (all preserved in a single manu-
script and none existing in Tibetan translation) are all topical monographs, or prakaraṇas,
rather than full commentaries on the works of Dharmakīrti or other predecessors in the Bud-
dhist epistemological tradition. Many of these philosophical essays cover topics that had
long become standard among the Buddhist epistemologists: the momentariness of all objects
(kṣaṇabhaṅga),1 the exclusionary or negative nature of linguistic and conceptual content
(apoha),2 the nature of and mode of determining the pervasion relationship that undergirds
inference (vyāpti),3 the inferential establishment of absences through non-apprehension (anu-
palabdhi),4 the refutation of the existence of God (īśvara-dūṣaṇa), and so on. Though the
range of topics he covers is largely standard, there are some distinctive issues that preoccupy
him and some features of his methods and his conceptual inventory set him apart. Building on
the insights of the earlier Dharmakīrtian commentator Prajñākaragupta, Jñānaśrīmitra articu-
lates a consistently idealist reading of Buddhist epistemology, deriving from this a remarkably
uniform methodology in which nearly all philosophical problems are resolved through an
analysis of cognitions into two contrasting aspects: “appearance” ( pratibhāsa) and “deter-
mination” (adhyavasāya). All awareness, both conceptual and nonconceptual, consists in the
558 DOI: 10.4324/9781351030908-49

Jñānaśrīmitra
appearance of a certain image (ākāra) in our consciousness. Conceptual awareness is that

which, through the process of determination, takes one or more of the cognitions containing
these phenomenal images to be intentional – that is, to contain or present images of something
other than themselves. All such intentional determinations are erroneous – nothing exists apart
from the mental images themselves – but all worldly activity and belief, and all language,
relies inescapably upon these ultimately false determinative projections. Considered indepen-
dently of these false projections, what appears in our consciousness (and therefore what really
exists) is nothing but a single, undifferentiated mental image, which Jñānaśrīmitra, following
Prajñākaragupta, terms “variegated non-duality” (citra-advaita).
“Appearance” and “Determination”

Earlier philosophy in the Dignāgan/Dharmakīrtian tradition had largely conducted its philosophical
analysis, and its critique of the rival realist analyses of the Brāhmaṇical Nyāya and Mīmāṃsā tradi-
tions, in terms of two sets of binary oppositions: between the two authoritative means of awareness
( pramāṇa), namely perception ( pratyakṣa) and inference (anumāna), and between two types of
knowable objects, namely self-characterized particulars (sva-lakṣaṇas) and commonly character-
ized universals (sāmānya-lakṣaṇas). The former are understood as unique individuals, absolutely
dissimilar and sharing no common features whatsoever. The latter comprise all purported common
properties (including natural kinds such as “cow” as well as property-universals such as “blue”)
which are thought to unite these individuals into groups or classes, but, at least for Dharmakīrti and
his followers, these common properties are purely mental constructions with no ultimate objective
reality. A key tenet of the Buddhist epistemologists is that there is a one-to-one correspondence
between the two pramāṇas and the objects they cognize, perception apprehending only particulars
and inference only conceptually constructed universals. What is distinctive about Jñānaśrīmitra’s
reworking of Dharmakīrtian epistemology and ontology is that these two linked binaries are over-
shadowed by, and in fact radically reconceived in the light of, the appearance/determination binary.
Jñānaśrīmitra’s basic model of the role of appearance and determination within the opera-
tion of the pramāṇas is derived not from Prajñākaragupta, whom he most typically looks to
as his model and inspiration, but from Dharmottara, with whom he has a more complex and
often contestatory relationship. It was Dharmottara who first clearly asserted that each of the
two pramāṇas has not just one, but two objects – one “grasped,” that is, directly apprehended,
and the other “determined.” Jñānaśrīmitra adopts this dual object model and makes it basic
to his entire epistemology.5 One of the clearest statements of his basic position on this issue
can be found in his Investigation of Pervasion (Vyāpticarcā), his analysis of the pervasion
relationship that underlies inference and of the manner in which knowledge of this pervasion
relationship is acquired. A large portion of the work is devoted to addressing specific criti-
cisms of Dharmakīrtian inference theory offered by rival philosophical traditions. One of the
more important interlocutors here is the Nyāya philosopher Trilocana, author of a (lost) work
entitled The Flower Cluster of Reasoning (Nyāyamañjarī), with whom Jñānaśrīmitra engages
frequently in his works. Taking the basic Dignāgan position to be that perception has only
particulars for its objects, Trilocana argues that this will make inference altogether impossible.
Inference depends on prior awareness of an invariant relationship between the inferential rea-
son and the thing to be proven by it (e.g., in the standard example, between the smoke that one
sees rising from a mountain and the fire whose presence there one infers from it). Dharmakīrti
and his followers hold that this relationship is learned by a combination of perception (of, e.g.,
fire occurring together with smoke) and non-apprehension (e.g., not observing smoke occur-
ring where fire is absent).6 But, if one can perceive only particulars, and all inference depends
559
Lawrence McCrea
on prior awareness of a link between two universals, there can be no pramāṇa, no valid source
of knowledge, that will allow one to know such a relationship. Jñānaśrīmitra summarizes Tri-
locana’s argument as follows:
But Trilocana raises the following objection: Since both perception and non-
apprehension have only particulars as their objects, how can one come to know
the connection between two universals by means of them? If you say that what we
come to know is the connection of “what is excluded from non-smoke” with “what
is excluded from non-fire,” we may still ask by what means can this connection be
known? It cannot be known by perception, since this has only particulars for its
objects. Nor can it be known by inference, since inference itself depends upon first
knowing this connection.7
Jñānaśrīmitra’s response to this objection proceeds by challenging Trilocana’s account of the

Buddhist theory of perception. It is not in fact the Buddhist position that perception has only
particulars as its objects; universals too are cognized via perception, but not in the same way
particulars are. It is here that Dharmottara’s account comes into play. Dharmottara argued
that all episodes of valid awareness, both perceptual and inferential, have two objects, one
“grasped” and one “determined.” When Dignāga and Dharmakīrti insist that perception has
only particulars as its object, what they really mean is that perception has only a particular as
its grasped object.8 Adopting this model derived from Dharmottara, Jñānaśrīmitra argues that
perception is able to apprehend universals and pervasion relationships between them as the
second, determined type of object. He seeks to undercut Trilocana’s argument by examining
the Buddhist epistemologists’ claim that “Perception has the specific individual as its object.”
There are two ways to understand the predicate of this claim: that perception has only the spe-
cific individual as its object (such that connection with any other kind of object is excluded) or
that perception does indeed have the specific individual as its object (such that non-connection
with the specific individual is excluded, without ruling out connection with other objects as
well). Jñānaśrīmitra argues that Trilocana has misconstrued the Buddhist’s claim by taking it
to entail the first sort of exclusion rather than the second:
To the objection of Trilocana, we reply as follows: Does the claim that “Perception
has the specific individual as its object” mean that it also has the specific individual as
its object, or that it only has the specific individual as its object? On the first view, that
there is exclusion of non-connection [of perception with a specific individual] there
is nothing to prevent perception having a universal as its object. How then is there no
opportunity for perception and non-apprehension of the two universals [linked in the
inference warranting relationship]? When you say that there is “exclusion of connec-
tion with anything other than a universal,” do you mean that this is so according to our
own established conclusion, or that it is really so? Not the first, since, according to us,
both pramāṇas have both particulars and universals as their objects, since these objects
are divided into those that are “grasped” and those that are “determined.” Whatever
appears in any particular awareness is what is “grasped.” But that thing toward which
one acts is “determined.” This being the case, in perception it is a particular that is
grasped, and a universal that is determined. For inference, on the other hand, the case
is just the reverse. Now, when one perceives something that serves as a means of
action, and desires the pragmatic effect to be achieved by that means, even though that
perception apprehends only a single moment, it does have a universal as its object,
560
Jñānaśrīmitra
insofar as it bears on the object-continuum. But, in a situation in which one learns

the pervasion relationship, one perceives only a single individual, but this perception
too has a universal for its object, since it takes as its object all particulars of the same
class. When one has the thought “I ascertain by perception that smoke is born from
fire,” one is aware of a determination arising from this perception that incorporates
all individual instances of smoke, excluding whatever is different from smoke. This is
just like cases where one sees merely visible form, but, because one determines that
one has seen, for example, a pot, which is a specific collection of visible, gustatory,
olfactory, and tactile properties, the perception has that pot as its object.9
Here Jñānaśrīmitra explicitly adopts Dharmottara’s two-object model of both perception

and inference as his own. Jñānaśrīmitra applies this theory specifically to the problem raised
by Trilocana’s critique: How can one apprehend, through perception alone, the pervasion-
relationship between, for example, smoke and fire? Because all perception has a second, deter-
mined and universal, object in addition to the momentary particular that it directly grasps,
perception is able to, and indeed necessarily does, make us aware of universals through deter-
mination. That any given perceptual awareness grasps only a momentary particular poses no
problem, as determination creates a second, universal object toward which one directs one’s
subsequent activity. This is as much true when we perceive a purportedly individual object
such as a single pot (which is not a true individual, according to the Buddhist epistemolo-
gists, but already a universal constructed by grouping individual sensibilia into an imagined
whole), as when we determine class-universals based on a finite set of observations – such as
that “smoke (in general) is born from fire (in general).” Trilocana’s misreading of the Buddhist
position, as Jñānaśrīmitra understands it, hinges on a misunderstanding of what exactly a “uni-
versal” is on the Buddhist account. As Jñānaśrīmitra makes clear, a universal is not a distinct
kind of entity – inaccessible to perception but required for inference. Rather, it is merely one
possible way of conceptualizing the particulars that perception grasps:
There is no other thing called a “universal.” Rather, it is the particulars themselves

that are called a “universal” when their differences from one another are not sepa-
rately discerned. When those differences are separately discerned, each can be desig-
nated by its own term, “particular.” So, even if all the particulars distinct from what is
other do not appear in a perceptual cognition, and hence cannot be its “object” in the
sense of being grasped by it, nevertheless it cannot be denied that they are its object
in the sense of being determined by it.10
The Buddhist view of universals, as typically presented by both themselves and their oppo-
nents, is that they are mentally constructed and hence wholly fictitious objects, “non-things”
(a-vastu). But Jñānaśrīmitra insists that this is not correct. A “universal” is simply a set of par-
ticulars conceptualized as a class rather than as individually distinct entities. Hence perception
can, through the process of determination, allow us to be aware of universals and the universal
pervasion relations between them.
Exclusion (apoha)
Some of Jñānaśrīmitra’s most original and influential writing revolves around his analysis of
the Buddhist theory of “exclusion” (apoha), the view first advanced by Dignāga that the con-
tent of all linguistic and conceptual awareness takes the form of a negation, excluding the
561
Lawrence McCrea
conceptualized item from some domain without ascribing to it any objectively real feature or
property. Jñānaśrīmitra devoted a full monograph to the topic,11 in which he not only surveys and
refutes the most important criticisms of the theory offered by anti-Buddhist philosophers but ana-
lyzes and critiques a variety of internal Buddhist positions on the question. This long essay proved
quite influential, and, along with the briefer summary of the same arguments by Jñānaśrīmitra’s
pupil Ratnakīrti in his own Proof of Exclusion (Apohasiddhi), redefined the Buddhist position on
this central issue. Jñānaśrīmitra’s discussion in particular drew a detailed critique from Udayana
in his Determination of the Essence of the Self (Ātmatattvaviveka) and thereby established itself
as a touchstone for later critiques from Nyāya and other Brāhmaṇical systems.
Here, as in so many other areas, Jñānaśrīmitra’s distinctive and innovative treatment of
this standard Buddhist topic turns on a careful discrimination of the roles of appearance and
determination in linguistic and cognitive processes. By his time, Buddhists had been arguing
among themselves for centuries about the proper understanding of the nature of the exclusion
said to define the content of linguistic and conceptual awareness. They asked: Should we take
the content of a conceptual awareness to be (1) a pure negation (which exists only internally,
as a mental image), (2) an external object qualified by such a negation, or (3) some hybrid
of the two? (All three of these positions – sometimes dubbed “negativist,” “positivist,” and
“synthesist” – have been advanced by authors within the Buddhist epistemological tradition
at one time or another; see Mookherjee 1993, 131–33; Katsura 1986, 171.) Jñānaśrīmitra was
the pioneering and most prominent advocate of the “synthesist” position – that the “object”
of conceptual awareness necessarily involves both a positive and a negative component. And,
significantly and typically, he links these two aspects to the two basic cognitive modes of
appearance and determination. He summarizes his own view of exclusion in a single verse, on
which his entire Monograph on Exclusion (Apohaprakaraṇa) is constructed as a commentary:
First of all, it is the external object that is primarily expressed by words. This being
the case, exclusion is understood as a qualifier of that. One of these – the external
object – is conditionally adopted as an object due to determination; the other – the
exclusion – is conditionally adopted as an object due to appearance. But really, noth-
ing at all is expressed.
While the traditional position of the Buddhist epistemologists is that linguistic awareness has
only a negation or “exclusion” as its referent (so that the term “cow” signifies “what is not a
non-cow”), Jñānaśrīmitra stresses, as many of his predecessors did not, that conceptual con-
tent must have a positive component and that this component is, at least phenomenologically,
the predominant element in such awareness, exclusion serving only as a necessary subordinate
element. It is this that accounts for our common intuitions about linguistic reference. Even the
exclusion theorist acknowledges that when we use the expression “There is a cow,” or recog-
nize an object before us as “a cow,” we typically feel ourselves to be speaking of and recog-
nizing a physical object external to our own language usage and thought. This is the “external
object” that is “primarily expressed by words.” But this purported external object is not what
appears in a conceptual awareness episode; rather, it is what is determined by it. Through
analyzing the components of inferential/verbal awareness in terms of the appearance/deter-
mination binary, Jñānaśrīmitra is able to explain how such awareness functions on a conven-
tional level without admitting the ultimate reality of any verbal referent or conceptual object
at all. Given the need to reconcile two seemingly contradictory positions earlier taken by the
Buddhist epistemologists – that words have exclusion as their referent and that words have
no referents at all – Jñānaśrīmitra ultimately concludes that the question is perspectival: one
562
Jñānaśrīmitra
may legitimately speak of different things being the referents of words from the perspective
of appearance, of determination, or of ultimate truth. He closes out his discussion as follows:
So, when one is asked, “How is exclusion expressed by a word?” we answer, “As an
element of that [external object that is primarily expressed],” the meaning of which
is as described. If the question is, “Why are the mental image, or the particular, or
the contingent features not expressed?” these questions are dispensed with in order
by saying, “This is because of the absence of determination, the absence of appear-
ance, and the absence of both.” But if the question is “What is it that is expressed by
words?” then, having set out these options (1) on the basis of appearance, (2) on the
basis of determination, or (3) really, the answers are, in order, (1) “the image that is
excluded from what is other, that resides in conceptual awareness”; or (2) “the particu-
lar that is excluded from what is other”; or (3) “nothing.” This has already been said.
Therefore, establishing the position that words and inferential reasons have exclusions
as their objects is for the sake of making it known that all properties are inexpressible.
For Jñānaśrīmitra, all conceptual awareness presupposes the existence of objects which are
both apparent to us and available to us as objects of action (even if that action is only verbal
or mental). But ultimately there can be no object that meets both of these criteria. Cognitive
images appear in our awareness, but these, being momentary, are not amenable to deliberate
action; determination constructs the imagined external particulars toward which all our activ-
ity (even verbal and mental activity) is directed, even though these determined particulars do
not actually appear in our awareness. But there exists no one thing that can both appear and be
determined. “Exclusion” must be accepted as an element that qualifies the (fictitious) external
particular determined by a conceptual awareness but cannot be a self-sufficient object of such
awareness. Hence he defends the Dignāgan theory of exclusion but only as a partial account of
awareness on the conventional level. Ultimately, language has no object at all, neither a posi-
tive entity nor an exclusion.
Conditionally Adopted Positions

A key feature of Jñānaśrīmitra’s treatment of exclusion, as can be seen from the previous sum-
mary, is his relegation of many of even the most characteristic and prestigious positions of the
Buddhist epistemological tradition – such as the claim that the objects of linguistic or conceptual
awareness are exclusions (apoha) – to the realm of conventional, rather than ultimate, truth. For
him, conceptual awareness really has no object at all, universal or particular, internal or exter-
nal. It is only conventionally that may speak of anything as the “object” of such an awareness.
Why then did Buddhist thinkers beginning with Dignāga insist that exclusion is the object of
inferential/verbal awareness if this is not really the case? To explain this, Jñānaśrīmitra develops
a sophisticated theory of “conditionally adopted positions” (vyavasthā), that is, positions that
one may legitimately adopt in a philosophic argument even while knowing them to be untrue:
By relying on a little bit of the truth, a certain conditionally adopted position is, for
a specific purpose, constructed by us, in one way, even though the actual state of
affairs is different, just as in examples such as the “self ” or the “arising of a thing.”
For “arising” can be a property only of an existing object qualified by a prior absence.
By relying on a little bit of the truth, namely, the prior absence, we conditionally
adopt the position that “there is arising of a nonexistent thing” in order to foreclose
563
Lawrence McCrea
any worries about the doctrine that effects preexist in their causes. Or, by relying on
the conceptual construction of a single continuum, we conventionally say: “Who else
will experience the result of an action done by this very person?” in order to frustrate
the deceptive view that there is the passing away of what has been done and the onset
of what has not been done.
He then applies this theory specifically to his interpretation of the theory of exclusion:
Here too, the idea that linguistic expression takes a positive entity as its object is just
the same [in that it, too, is a conditionally adopted position]. Here we conditionally
adopt the position that exclusion, even though it is really just a necessarily attendant
awareness, is the object of conceptual awareness, in order to set aside any suspicion
that we accept the position pushed by our opponents that it is only the positive entity
that is really expressed. And therefore, we don’t talk in terms of just the positive
entity [when describing the semantic value of a word]. But when someone pushes
the position that “exclusion alone is the primary meaning of a word,” then we put
forth the positive entity as well. As stated, “First of all, it is the external object that is
primarily expressed by words.” But in perception, because there is no disagreement
of this sort, it is proper that one should not conditionally adopt this position.12
It turns out that for Jñānaśrīmitra, both perceptual and inferential/verbal awareness necessarily
have both a positive and a negative component. What differentiates perceptual from inferen-
tial/verbal awareness is not that one has a positive and the other a negative “object” but that
the rhetorical context in which the Buddhist theory is advanced justifies the claim of a negative
object in the second case but not the first. In the case of conceptual and verbal awareness, the
advancing of the exclusion-based account, although untrue, is justified by the need to counter
the realist contention that the positive entity alone is expressed. And it is the theory of condi-
tionally adopted positions that provides this justification.
Through this explanation of the “conditionally adopted position,” Jñānaśrīmitra clearly
theorizes, perhaps for the first time and certainly the most perspicaciously, a process that had
been underway for some time. Almost since the first layer of commentary on Dharmakīrti’s
works, philosophers in the Buddhist epistemological tradition had been engaging in a mostly
tacit program of rational reconstruction, adjusting, expanding, or reformulating Dharmakīrti’s
key concepts and arguments to confront new objections and to maximize the coherence and
explanatory power of the system as a whole and often adopting stances that would likely have
been anathema to Dharmakīrti himself. (Dharmottara’s dual-object model of the pramāṇas
is a good case in point.) What is most striking about Jñānaśrīmitra’s treatment is the open-
ness with which he acknowledges and elucidates this process, openly dismissing the overt
positions of Dignāga and Dharmakīrti while simultaneously justifying them as “white lies,”
ostensibly deliberate misrepresentations rendered virtuous through the pedagogical or rhetori-
cal purposes they serve.13
Lower- and Higher-Order Conventional Reality

By thus relegating key elements of the established doctrine of the Dignāgan and Dharmakīrtian
tradition to the realm of conventional rather than ultimate truth, Jñānaśrīmitra is in part doing
what commentators often do, updating the stances of their forbears to meet new philosophi-
cal and dialectical needs. But, as we have seen already in his development of the idea of
564
Jñānaśrīmitra
conditionally adopted positions, he is also unusually open and self-conscious in making this
move, and, when he questions the ultimate truth of the overtly stated views of the philosophers
whose tradition he ostensibly upholds, he seems to call into question the ultimate value of their
having advanced these views in the first place. By linking the legitimacy of conventional asser-
tion to the ability to articulate a legitimate purpose for the assertion of things not strictly and
ultimately true, Jñānaśrīmitra raises overtly the question of how we can meaningfully assess
the value or usefulness of philosophical positions as distinct from their accurate representation
of what is really the case. If all the claims of philosophers, being verbal and thus necessarily
conceptual, are thereby in the final analysis misrepresentative of ultimate reality, how then are
they to be weighed against one another? If the “truths” presented by the Buddhist epistemolo-
gists are merely conventionally rather than ultimately true, why should they be at all preferred
to the views of rival Brāhmaṇical philosophers or, for that matter, to the philosophically untu-
tored beliefs of ordinary people? In response to such objections, Jñānaśrīmitra finds it neces-
sary to develop a more sophisticated account of conventional truth. As far back as the time
of Nāgārjuna, Buddhist philosophers had routinely spoken of “the two truths,” conventional
(saṃvṛti) and ultimate ( paramārtha); many things that may legitimately taken as “true” for
everyday practical purposes – such as the existence of perduring extra-mental objects – must
be rejected as ultimately false when subjected to philosophical analysis. But Jñānaśrīmitra goes
beyond this binary division, arguing that different conventional truths must be seen as ranked
in a continuum ranging from lower (adhara) to higher (uttara), with the latter being systemati-
cally preferred to the former, even though both necessarily fall short of ultimate truth.
This issue is discussed most fully in the opening section of Jñānaśrīmitra’s Treatise on
Moment by Moment Destruction (Kṣaṇabhaṅgādhyāya). As a committed idealist, Jñānaśrīmitra
is confronted with a basic problem at the outset of his extensive proof of momentariness.
What is the point of proving, and what does it even mean to prove, that mind-external objects
must exist only as series of causally related moments if one will only go on to argue that no
such objects exist at all? Jñānaśrīmitra takes the position that the long-held Buddhist tenet of
momentariness must be defended as a higher-order convention, in preference to the lower-
order convention upholding the existence of temporally persistent objects, and it is in con-
nection with this argument that he offers his most fully developed theorization of higher- and
lower-order conventional truths:
Even though the causal capacity of objects is only conventionally real, it must necessar-
ily be accepted if this investigation is to have any validity. And, because of this, postula-
tion of purportedly stable entities such as “universals” and “inherence”14 should not be
accepted, in so far as they are falsified by the proof of moment-by-moment destruction
adduced from this causal capacity of objects, provided one accepts that objects have
pragmatic efficacy. (If, on the other hand one does not accept that objects have pragmatic
efficacy, then there would have to be conventional acceptance of reality for something
incapable even of producing an awareness of itself; and this should be considered to
be an even lower-order convention.) Therefore, given that progressively higher-order
(uttara) conventions are conditionally established (vyavasthāpyamāna) as “ultimately
real” in comparison with progressively lower-order conventions, the conventional belief
in pragmatic effectiveness, having no convention higher than it, must be accepted as
“ultimately true,” and whatever exists with this capacity of pragmatic effectiveness is
called “ultimately real.” This is just like what happens when a child thinks there is a
face in the surface of a mirror, and, in contrast to this belief, one says to the child as an
ultimate truth that “What is seen there is not a real face,” even though, in contrast with
565
Lawrence McCrea
this belief, there is a further “ultimate truth” that the rays of light from one’s eyes, turned
back upon oneself by striking the mirror, actually apprehend one’s own face. Thus when-
ever successively later conceptual constructions prove more powerful than those that
precede them, we say that each earlier one is “conventional,” and each one that follows
after is “ultimately true,” up until the point when pragmatic effectiveness is achieved.
The belief that external objects exist only momentarily is not in the final analysis any more
true than the view that they exist as perduring over time. Both views are ultimately false, as
a philosophically educated view will hold that no such objects can possibly exist in the first
place. Nevertheless, the view that such objects exist as series of causally related moments
is to be preferred as a higher-order convention, being “stronger” than the lower-order con-
vention it displaces. We are not told what exactly the criteria of “strength” in this context
are, except that they are to be explained in terms of “pragmatic effectiveness” (arthakriyā).
One conventional view is ranked as stronger and therefore “higher” than another based on
its usefulness for some purpose, not because it more accurately approaches or approximates
ultimate truth. In light of Jñānaśrīmitra’s theory of conditionally adopted positions discussed
previously, presumably such purposes include rhetorical effectiveness in philosophically
refuting the views of others as well as soteriologically assisting the holders of such beliefs in
the quest for enlightenment.
Jñānaśrīmitra’s “With-Image” Idealism and the Struggle with

Ratnākaraśānti
One of the central preoccupations of Jñānaśrīmitra’s philosophy concerns an entirely intra-
Buddhist debate regarding the ultimate ontological status of the mental contents or images
(ākāra) that appear in our awarenesses. Two of his surviving works, together comprising more
than a third of his extant philosophical oeuvre, are devoted to this topic: he wrote a A Treatise
Proving That Cognitions Contain Real Images (Sākārasiddhiśāstra), an extended prose piece
devoted programmatically to this topic, as well as a shorter (but still lengthy) verse summary
of the same argument, his Brief Summary of the Argument That Cognitions Contain Real
Images (Sākārasaṃgrahasūtra). Jñānaśrīmitra begins his Treatise Proving That Cognitions
Contain Real Images with a general statement of his idealist ontology, which holds that all that
really exists is a single multiform (citra) cognitive image:
It has been established that this entire triple world is nothing but consciousness.
Here in this work we explain the greater or lesser strength of internal divisions
within this consciousness.
To explain: By the steps of refuting mind-external objects and the like,
This world is, first of all, not external, since in itself it has a form apparent to con-
sciousness, and due to the necessary co-apprehension [of supposed objects with the
cognitions that apprehend them]. No other thing can possibly be established through
reasoning, and, even if it were possible, it would not have the characteristic of being
apprehensible. Therefore, all this world is mere consciousness.
The world is one variegated sensory image, comprising colors such as gold,
sounds such as the third note of a musical scale, tastes such as sweet, odors such as
the pleasantly fragrant, tactile sensations such as soft, and mental sensations such as
pleasure and displeasure. Thus there is the following scriptural passage, “Indeed, O
Monks, these three worlds are mind only.”15
566
Jñānaśrīmitra
Jñānaśrīmitra, following Prajñākaragupta and the idealist Yogācāra tradition more broadly,
takes the basic position that all that exists is mere consciousness, with no corresponding extra-
mental objects. The very fact that we are aware of objects at all is sufficient to establish that
these “objects” are mental. To be apparent to awareness is precisely the mark of the mental. This
is often summed up by the principle of “co-apprehension” (sahopalambha) first articulated by
Dharmakīrti that “there is no distinction between ‘blue’ and the awareness of blue, since they are
necessarily apprehended together” (Pramāṇaviniścaya 1.54ab); one can never be aware of any
purported “object” such as blue as distinct from one’s awareness of it, and hence there can be
no evidence for the existence of such an object. But this basic position, by Jñānaśrīmitra’s time
widely (if not universally) shared by Buddhist intellectuals, gives rise to a further intra-Buddhist
debate about the nature and epistemic status of the purported internal contents of our awareness.
Are the images that appear in our awareness themselves ultimately real, or are they themselves
fictions or illusions ultimately to be displaced? Jñānaśrīmitra argues forcefully that these images
are not, and in fact cannot, be displaced or overturned in any way, and are ultimately real (are,
in fact, all that is ultimately real). In making this case, he summarizes what he takes to be the
arguments of the two principal Buddhist opponents of this view, the Mādhyamikas and the “No
Image” (nirākāra) Yogācāras. He begins by briefly presenting the Mādhyamika position:
On this point, some people [i.e., the Mādhyamikas] say: Let us agree that sensory pro
perties such as white, and internally sensed properties such as pleasure have the nature
of pure consciousness. Even so, they cannot escape the emptiness of nature characteris-
tic of all things. If you seek to establish that external objects cannot exist by introducing
an argument based on the impossibility of their being either unitary or multiple, there
is no reason why this argument would recoil from properties such as brightness even if
they bear a purely mental nature [since they would still be inexplicable as either unitary
or multiple]. Therefore, it is better to say that all properties [whether mental or extra-
mental] are just empty, as is set forth in the Buddhist scriptures in many places, and
also in Dharmakīrti’s Critique of the Sources of Valid Awareness (Pramāṇavārttika),
where it is said that, “In whatever way one considers the objects of cognition, they col-
lapse in just that way” (Pramāṇavārttika, chapter on perception, 209). The view that all
is consciousness is merely a conventional and not an ultimate truth.16
Having thus briefly surveyed the Mādhyamikas’ argument against the ultimate reality of aware-
ness’s phenomenal content, Jñānaśrīmitra then turns to introduce and confront his principle
adversary, the “Without Image” Idealist (Nirākāra-yogācāra) school as represented by his con-
temporary and rival Ratnākaraśānti. He begins by summarizing this view, which maintains that,
while all the purported objects of our ordinary awareness are ultimately “blocked” (bādhita), that
is, overturned and displaced by other cognitions and hence to be rejected as false (alīka), “illumi-
nation itself ” ( prakāśamātra) remains as the constant and ineradicable nature of all awareness,
which can never be falsified or overturned and hence must be accepted as ultimately real:
Others [i.e., Ratnākaraśānti] resolve this in the following manner: Let us accept that
the forms of awareness are false, in that they are blocked. But the illumination itself
is true, and accepted by our tradition, because an awareness consisting of this illu-
mination is perceptual, since it cannot possibly be blocked. “Illumination” itself is
the innate form of an illuminating awareness; it cannot be introduced into it through
some confusion, such that the awareness of it could be an error. Specific phenomenal
content such as “blue,” because it is a form different from the awareness itself, could
567
Lawrence McCrea
be introduced through a confusion, and hence the awareness of it could be an error.

So, there is an opportunity for the blocking of phenomenal content such as blue, but
not of illumination. Negation should always be understood to be of that which is false
or conceptually constructed.17
Jñānaśrīmitra’s refutation begins by attacking the notion of the “illumination” which the
Nirākāra theorist imagines to remain unfalsified when all specific awareness-contents have
been falsified. If in all cognitions what appears to us are just specific cognitive images (ākāra),
how can we ever come to be aware of this supposed “illumination” as distinct from the images
that invariably appear along with it?
Regarding this view the following must be considered: What exactly is this “illumi-
nation” as distinct from phenomenal content such as blue? If there is no distinction,
then how can you avert the fear of illumination being blocked? In any case, there is
no valid authority to believe that it exists as something distinct. Perception cannot be
the authority, since it is impossible in the awareness of blue to apprehend anything
other than the perceived blue. Nor can inference be the authority since one can never
become aware of any potential inferential sign’s having an inference-warranting con-
nection with something that has never previously been perceived. Furthermore, what
could be more of an embarrassment than to claim that the “illumination” existing in
one’s own stream of cognitions is known through inference, which only apprehends
things beyond the range of one’s perception?18
If the existence of “illumination” as distinct from the specific phenomenal content of any
awareness cannot be established by either perception or inference (the only two valid means of
knowledge accepted by the Buddhist epistemologists), then there is no way one could ever be
aware of this “illumination” as distinct from the image presented by the awareness. And, in the
absence of any evidence for illumination as a distinct element, there is no way one can argue
that the cognitive image is blocked or overturned while the illumination remains somehow
unfalsified. An awareness and its image are ultimately inseparable, and there is no way the one
can be falsified without the other.
This line of argument, if successful, effectively undermines Ratnākaraśānti’s position but does
not by itself establish Jñānaśrīmitra’s countervailing view, that all the images that appear in our
awareness remain unfalsified and are therefore to be accepted as ultimately real. As a central pillar
of his case for the ultimate reality of mental content, Jñānaśrīmitra seeks to undermine the entire
notion that the presentational content of any cognition can ever be “blocked” or falsified in any
way – an argument principally directed against Ratnākaraśānti but which effectively undermines the
Mādhyamika opponent as well, as it makes it ultimately impossible to falsify the presentational con-
tent of any awareness. Jñānaśrīmitra’s argument turns, once again, on analyzing the process of falsi-
fication through the familiar rubric of appearance and determination. The falsification or “blocking”
of one cognition by another necessarily requires the determination of the second cognition as being
a cognition which apprehends the first cognition as “false,” which, Jñānaśrīmitra argues, can occur
only at the level of conventional reality. This argument is presented most forcefully in the follow-
ing passage from the fourth chapter of the Treatise Proving That Cognitions Contain Real Images:
But how can there be any falsity (alīkatā) of the nature of an experiential awareness?
You may say, “Because another awareness blocks it.” But what is it that is “blocked,”
and by what? You may argue that, “There is blocking of the whole collection of
568
Jñānaśrīmitra
variegated mental images, by the non-apprehension of a factor that would necessarily

pervade them.” But this is protecting the cow after you’ve already sold it:
A relation of blocker and blocked has been defended in Prajñākaragupta’s Com-

mentary on Dharmakīrti’s Critique of the Sources of Valid Awareness, but only on
the conventional level, since a contradiction has been established on the theory of
“appearance of the non-existent,”19 which would pervade the relation of blocker
and blocked.20
The idea that one cognition could block another depends on the belief that a second cognition
could somehow reveal that the seeming content of the first was neither a real external object
or the cognition’s own real form but something totally non-existent. But this can be accepted
only on the level of conventional truth, because it is ultimately not possible for a cognition to
reveal anything other than itself and therefore for one cognition to reveal the falsity of another.
Jñānaśrīmitra then proceeds to flesh out this claim of impossibility by asking precisely how
one cognition could block the content of another:
Furthermore, an awareness cannot block itself, since its consciousness of its own form
would overpower any possible blocking. But, if one awareness were to be blocked by
another, would that other awareness be one with the same object, or a different object?
And, would it be simultaneous or non-simultaneous with the awareness that it blocks?
Supposing it has the same object, it would then be incoherent to say that this aware-
ness has the object of the first cognition as its object, but nevertheless blocks that very
object. But how could the object of the first cognition be blocked by a cognition with
a different object, since all the second cognition does is to present its own object? And
why shouldn’t the second cognition be blocked by the first? Furthermore, one cogni-
tion could not be blocked by another occurring simultaneously: If multiple objects,
whether similar or dissimilar, were to appear at the same time, how could one know
which is blocked by which? But, if they occur at different times, this is all the more rea-
son why there is no occasion for one to block the other. This would lead to impossibili-
ties such as the ancient king Mahāsaṃmata being in conflict with the much later king
Śrīharṣa.21 You may argue that what is meant by “blocking” is that one cognition makes
known that another has no object, and that there is no contradiction in this occurring
between two cognitions occurring at different times. But if the image to be blocked did
in fact appear in the prior cognition, then it surely did exist at that time; how then could
it be “blocked”? If, on the other hand, it did not appear in the initial cognition, then it
simply does not exist on its own accord – what need is there of any further negation
by the later blocking cognition? And, in any case, the so-called “blocking” cognition
could not be said to have the same object as the earlier cognition.22
No cognition can coherently be said to block itself, but no other cognition, whether simultane-
ous or successive, could block it either. For this to occur, it would be necessary for the two
cognitions to be “of ” the same thing but to present it differently, and this Jñānaśrīmitra shows
to be impossible. It is only conventionally that one could imagine any cognition to represent
or misrepresent anything other than itself:
You might argue that there is prohibition of what is superimposed.23 But we have already
explained that superimposition can be established only on the level of conventional
569
Lawrence McCrea
reality, and hence the blocking of it also cannot go beyond the conventional. You may
argue as follows: Since “error” is simply the arising of a successive cognition in reli-
ance on a specific kind of dependent co-origination, in just the same way “blocking”
is just the arising of an awareness of the form “This is like this, and not otherwise.”
If so, then it is established that there is really no blocking at all. A given awareness
arises through taking on a certain image, but nothing is either established or blocked
by it, just as an awareness of sweetness arising immediately after an awareness of blue
neither establishes nor blocks that prior awareness. If you say that there is a difference,
since in this case there is blocking of the other through determination, we respond
as follows: To say that this blocking is “established only through determination” is
precisely to say that it is conventional. Therefore the meaning of “blocking” cannot be
that it makes known that another awareness lacks an objective foundation.24
Seen simply in terms of their phenomenal appearance as cognitions, one following the other
in sequence, there is no way a second cognition could make known the unreality or falsity of
the image appearing in the first. Instead, this can happen, in Jñānaśrīmitra’s account, only on
the level of conventional reality. “Determination” consists of taking the phenomenal image
that appears in one’s awareness to be an image “of ” something else and is necessarily a form
of conceptual awareness and, as such, ultimately erroneous. It is only by falsely projecting
the image that constitutes an awareness as being the image of something outside that aware-
ness that one can take one awareness to falsify another. At the ultimate level, the level of
appearance at which no awareness reveals anything beyond itself, blocking is impossible and
unimaginable.
Throughout this entire argument, we see again how the appearance/determination binary
grounds Jñānaśrīmitra’s approach to the most important philosophical problems he deals with.
He shows at great length that any theory of the “blocking” of one cognition by another entails
taking the second, blocking cognition to be a cognition of something other than itself, which
necessarily depends upon the process of determination. What appears in any cognition is noth-
ing more than the image (ākāra) that appears in it; taking it to cognize anything other than
this image is precisely what Jñānaśrīmitra understands determination to consist of. But all
determination is conceptual and therefore a form of error. Hence the appearance-content of all
awareness remains unfalsified and must be accepted as ultimately real.
Conclusion
Through his thoroughgoing rereading of the Dignāgan/Dharmakīrtian tradition through the
lens of the binary appearance/determination rubric and his theoretically self-conscious deve
lopment and deployment of the “conditionally adopted position” as his key hermeneutic tool,
Jñānaśrīmitra is able to effectively position himself both as the culmination of the half millen-
nium of Buddhist epistemological thought that precedes him and as a creative and often pro-
vocatively innovative thinker in his own right. His work offers an excellent example of the ways
in which Indian philosophers operating within broadly scholastic intellectual traditions could
satisfy the sometimes conflicting imperatives of exegetic faithfulness and original and creative
philosophizing. His rich and still very much understudied oeuvre stands as one of the monu-
ments of late first-millennium Indian philosophy and one of the most sophisticated attempts to
synthesize the diverse mass of Buddhist epistemological theory into a coherent system.
570
Jñānaśrīmitra
Notes
1 See his Treatise on Momentary Destruction (Kṣaṇabhaṅgādhyāya), partly edited and translated in
Kyuma 2005.
2 See his Monograph on Exclusion (Apohasiddhi), edited and translated in McCrea and Patil 2010.
3 See his Investigation of Pervasion (Vyāpticarcā), edited and translated in Lasic 2000.
4 See his Secret of Non-Apprehension (Anupalabdhirahasya), edited and translated in Kellner 2007.
5 On the history of the theory of determination within the Dharmakīrtian tradition and Dharmottara’s
transformative role in it, see McCrea and Patil 2010, 16–20; McCrea and Patil 2006, 318–31; Dreyfus
1997, 354–64.
6 The precise nature and number of such prior observations and non-observations required to
acquire knowledge of a pervasion relationship was itself a matter of controversy, and Jñānaśrīmitra
devoted another short work to this problem, his Proof of the Relation Between Cause and Effect
(Kāryakāraṇabhāvasiddhi, in Thakur 1959, 317–22).
7 Thakur 1959, 161.
8 See Dharmottara, Nyāyabinduṭīkā 1.12; for further discussion of Dharmottara’s account and
Jñānaśrīmitra’s deployment of it, see also McCrea and Patil 2006.
9 Thakur 1959, 166.
10 Thakur 1959, 166.
11 McCrea and Patil 2010.
12 Thakur 1959, 204–5.
13 On the conditionally adopted position as “white lie,” see Patil 2007.
14 “Universals” (sāmānya) are held by the Brāhmaṇical realists to be eternal unchanging class-proper-
ties or essences manifested in individual members of a class such as “cow”; “inherence” (samavāya)
is the relation by which these universals inhere in individual class-members.
15 Thakur 1959, 367. The original scriptural source of the quotation is unknown, but was earlier quoted
by Vasubandhu (Introduction to Viṃśikā verse 1) and several other authors following him.
16 Thakur 1959, 367–68.
17 Ibid., 368. As Kajiyama has shown, this is a summary of Ratnākaraśānti’s presentation of his own posi-
tion in his Instruction in the Perfection of Wisdom (Prajñāpāramitopadeśa). See Kajiyama 1965, 36–37.
18 Thakur 1959, 368.
19 One of the classical theories of error first laid out by Maṇḍanamiśra in his Vibhramaviveka, the doc-
trine holds that what appears in an erroneous awareness is something wholly unreal (asat). The Bud-
dhist epistemologists reject this doctrine, and Prajñakāragupta argues against it in his commentary
on Pramāṇavārttika, Pratyakṣa 331, as part of his overall rejection of the idea that any cognition can
ultimately make known anything other than itself (Prajñākaragupta 1943, 356).
20 Thakur 1959, 437.
21 Mahāsaṃmata is a legendary king, said to have lived long before the time of the historical Buddha.
“Śrīharṣa” refers to Harṣavardhana, who ruled a large North Indian kingdom centered in Kanauj from
607–647 CE.
22 Thakur 1959, 437. If, as supposed here, the content to be blocked did not appear in the initial cogni-
tion in the first place, there is no way the later cognition that supposedly overturns it can be seen as
bearing on the same object – hence the content to be overturned was never affirmed in the first place,
and the notion of “blocking” is rendered incoherent.
23 That is to say, the image the first cognition was thought to “grasp” was never really part of that cog-
nition but was erroneously superimposed on it. The second “blocking” cognition therefore does not
overturn the real nature of the prior one but merely a misidentification wrongly imposed on it.
24 Thakur 1959, 437.
Works Cited
Dreyfus, Georges B. J. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpreta-
Kajiyama, Yuichi. 1965. “Controversy Between the Sākāra- and Nirākāra-vādins of the Yogācāra
School – Some Materials.” Indogaku Bukkyōgaku Kenkyu 14 (1): 26–37.
571
Lawrence McCrea
Katsura, Shoryu. 1986. “Jnānaśrīmitra on Apoha.” In Buddhist Logic and Epistemology: Studies in the
Buddhist Analysis of Inference and Language, edited by Bimal K. Matilal and Robert D. Evans,
171–84. Dordrecht: D. Reidel.
Kellner, Birgit. 2007. Jñānaśrimitra’s Anupalabdhirahasya and Sarvaśabdābhāvacarcā: A Critical Edi-
tion with a Survey of his Anupalabdhi Theory. Wien: Arbeitskreis für tibetische und buddhistische
Studien, Universität Wien.
Kyuma, Taiken. 2005. Sein und Wirklichkeit in der Augerblicklichkeitslehre Jñānaśrīmitra’s Ksanab-
hangadhyaya I. Paksadharmatādhikāra: Sanskrittext und Ubersetzung. Wien: Arbeitskreis für tibet-
ische und buddhistische Studien, Universität Wien.
Lasic, Horst. 2000. Jñānaśrīmitras Vyāpticarcā: Sanskrittext, Übersetzung, Analyse. Wien: Arbeitskreis
für tibetische und buddhistische Studien, Universität Wien.
McCrea, Lawrence J., and Parimal G. Patil. 2006. “Traditionalism and Innovation: Philosophy, Exege-
sis, and Intellectual History in Jnānaśrīmitra’s Apohaprakaraṇa.” Journal of Indian Philosophy 34:
303–66.
———. 2010. Buddhist Philosophy of Language in India: Jnānaśrīmitra on Exclusion. New York:
Mookherjee, Satkari. 1993. The Buddhist Philosophy of Universal Flux: An Exposition of the Philosophy
of Critical Realism as Expounded by the School of Dignāga. New Delhi: Motilal Banarsidass.
Patil, Parimal G. 2007. “Dharmakīrti’s White Lie: Philosophy, Pedagogy, and Truth in Late Indian Bud-
dhism.” In Pramānakīrtih: Papers Dedicated to Ernst Steinkellner on the Occasion of His Seventieth
Birthday, Part 2, edited by Birgit Kellner, Helmut Krasser, Michael T. Much, and Helmut Tauscher,
597–619. Wien: Wiener Studien zur Tibetologie und Buddhismuskunde.
Prajñākaragupta. 1943. Pramāṇavārttikabhāṣyam. Edited by Rahul Sankrityayana. Patna: Kashi Prasad
Jayaswal Research Institute.
Tārānātha. 1990. History of Buddhism in India. Translated by Lama Chimpa and Alaka Chattopadhyaya.
New Delhi: Motilal Banarsidass.
Tatz, Marm. 1987. “Life of the Siddha Philosopher Maitrīgupta.” Journal of the American Oriental Soci-
ety 107 (4): 695–711.
Thakur, Anant Lal, ed. 1957. Ratnakīrtinibandhāvali. Patna: Kashi Prasad Jayaswal Research Institute.
———. 1959. Jñānaśrīmitranibandhāvali. Patna: Kashi Prasad Jayaswal Research Institute.
572
35
RATNAKĪRTI
Aligning Everyday Experience with
Momentariness and Idealism
Patrick McAllister
The Scholar and His Philosophical Works

Ratnakīrti was a Buddhist scholar (mahāpaṇḍita) active in the monastery of Vikramaśīla in
today’s Bihar during the first half of the eleventh century CE. Today, Ratnakīrti is best known
for his philosophical works. All of these works were composed in Sanskrit, and most of them
can best be characterized as concise and analytic summaries of the much longer and more com-
plex works composed by his teacher, Jñānaśrīmitra. In their texts, Jñānaśrīmitra and Ratnakīrti
present a highly developed formulation of the main tenets of Dharmakīrti’s system. Their works
provide a coherent and detailed analysis of our conventional understanding of everyday real-
ity and argue that this conventional understanding is an erroneous fiction based on ignorance.
Given that our conventional understanding is based on ignorance, Jñānaśrīmitra and Ratnakīrti
then theorize the nature of liberation and how it can be attained under such conditions.
There are several recurring topics in Ratnakīrti’s philosophical oeuvre. In A Proof of an
Omniscient Being,1 he discusses the possibility of a buddha’s omniscience. His arguments about
omniscience are also applied to the refutation of god-like beings and their characteristics in A
Refutation of the Proof of God.2 His considerations of universal momentariness and of the way in
which everyday activity seemingly engages with permanent objects were developed into a theory
of existence that serves as the ontological basis for all his works. These considerations are found
in A Positive Proof of Momentary Cessation, A Negative Proof of Momentary Cessation, A Refu-
tation of the Proof of Enduring Things, A Proof of Exclusion, and A Theory of Multifaceted yet
Non-Dual Appearance.3 The main subject of this last work and A Refutation of Other Continua4
is the philosophy of mind. In these two works, Ratnakīrti develops a typical Buddhist form of ide-
alism, which results from his analysis of the mind as always having what he calls a multifaceted,
yet non-dual appearance and his proof that we cannot know other minds. Writing in the tradition
of the logico-epistemological school of Indian Buddhist philosophy, all his treatises showcase
his mastery of logic. He explicitly discusses technical problems of epistemology and logic in A
Treatise on the Existence of Other Means of Valid Cognition5 and in An Assessment of Pervasion.6
The following sketches will focus on his most prominent ontological theories. Both of
these, momentariness and the non-existence of anything but the mind, fly in the face of every-
day experience. Ratnakīrti, like his predecessors, was pushed to account for how the engage-
ment with enduring, extra-mental objects could possibly be an illusion.
DOI: 10.4324/9781351030908-50 573

Patrick McAllister
Momentariness and the Definition of Reality

Ratnakīrti’s two Proofs of Momentariness seek to establish the truth of the following infer-
ence:7 To exist is to be momentary; for example, a pot exists and is therefore momentary. And
these things under discussion do exist.8 To be valid, this kind of inference (a parārthānumāna;
see Kajiyama 1998, 86) requires certain elements: First, a relation between a reason – here,
“to exist” – and the probandum – here “to be momentary” – is stated to hold. This relation,
called “pervasion” (vyāpti), implies that the reason cannot be true of some thing without the
probandum being true of the same thing. Second, an example of such a relation is provided,
which should be acceptable to both the proponent and the opponent. In the last step, the reason
is stated to apply to the case under consideration. Ratnakīrti specifies that the reason used in
this inference is one that consists in the proper nature of a thing (svabhāvahetu).9 We should
therefore understand that, just like an oak is a tree by its very nature of being an oak, things that
exist are momentary by their very nature; in other words, to exist is to cease or perish momen-
tarily. Both of Ratnakīrti’s proofs of momentariness take up the validity of this inference. A
Positive Proof of Momentary Cessation first discusses how to define “existence” and to what
extent it qualifies the objects that the inference is about and then asks whether there are – and
whether we can know about – cases in which things are momentary by their very nature of
being existent.10 Ratnakīrti begins this discussion as follows:
One cannot declare that a reason shows an unperceived object without first elimi-
nating the doubts about whether the reason might be a pseudo-reason. And pseudo-
reasons are of three kinds: unestablished, contradictory, and ambiguous reasons.11
The core of the argument of A Positive Proof of Momentary Cessation consists in showing
that the reason for inferring momentariness, “existence,” is not one of these three so-called
pseudo-reasons (unestablished, contradictory, and ambiguous, a differentiation that long pre-
ceded Ratnakīrti). Broadly speaking, an unestablished reason is one that is not actually present
in the object that the inference is about (it may be true that an oak is a tree, for instance, but if
the object of the inference is not actually an oak, the inference does not apply). A contradictory
reason is one that can never have joint presence with the property that should be established
(so there must be at least one instance in which existence and momentariness, or “being an
oak” and “being a tree,” occur together). An ambiguous reason is one that occurs both together
with and apart from the inferred property and as such is not a sure indicator of the latter prop-
erty (some existent things would be momentary, others not, like some oaks would be trees,
others not). In other words, we can think of Ratnakīrti’s analysis of the three potential faults
of the reason “existence” as investigating (1) the relevant definition of the property of exis
tence and whether the things under discussion have this property, (2) the possibility that these
existent things are momentary, and (3) the necessity that these existent things be momentary.
Ratnakīrti’s investigation into whether “existence” might be an unestablished reason
focuses on the analysis of existence. He first states that there is a generally acceptable defini-
tion of existence: the state of producing an effect (arthakriyākāritva). This definition of exis
tence had been accepted within Buddhism at least since Dharmakīrti. According to Ratnakīrti,
this definition can serve as common ground for himself and his opponents: on the one hand, it
is the notion of existence implicitly presupposed by anyone who engages in everyday activi-
ties, and on the other, it is only this kind of existence that can be generally established by per-
ception and inference, the two means of knowledge ( pramāṇa) acceptable to both Buddhists
and virtually all of their opponents (see Kajiyama 1998, 30–38). Other definitions had been
574
Ratnakīrti
proposed by competing philosophical systems, yet, Ratnakīrti argues, since such definitions
meet these two criteria only by resort to causality, it would be useless to discuss them in detail.
Existence is thus analyzed in causal terms: to exist is to produce an effect. The next ques-
tion to examine is how this characteristic is linked to momentariness. That is, can things that
produce an effect be momentary? If this cannot be affirmed, then the purported reason “exis
tence” would be a pseudo-reason of the second type, a contradictory one. If it were impossible
for momentary things to produce an effect, then to say that something exists would preclude
that thing’s momentariness.
At this point, Ratnakīrti has to consider another logical peculiarity of the proof of momen-
tariness. In the usual inference of its type, “This is a tree because it is an oak,” there are three
groups of things that have well-defined relations to each other: the thing currently under consi
deration, all or at least some other things for which it has been established that they are oaks and
trees, and everything that is not an oak (though perhaps a tree). In the tradition of logical analy-
sis that Ratnakīrti subscribes to, an inference generates knowledge only when the property used
as a reason (“being an oak,” “existing”) is “pervaded by” the property to be established (“being
a tree,” “being momentary”). This relation of pervasion can usefully be compared to sets: the
instances of oaks are a proper subset of the instances of trees, and the set of trees contains other
types of trees apart from oaks. If this is true, then the reason “being an oak” is proof of “being
a tree”; in other words, “being an oak” is pervaded by (or subsumed under) “being a tree.” This
approach generates two problems for the inference of momentariness. First, the set of existing
things and the set of momentary things are, according to the claim of the inference, the same:
there are no momentary things that do not exist, and no things exist that are not momentary.
Second, there are no (existing) things that do not belong to this set. In other words, there is
nothing that could be adduced as an example for being both existent and momentary. Apart
from being a formal requirement of inferences in Ratnakīrti’s tradition, the question is larger: if
there were no other corroborating instances, how could we be sure the inference is true? Since
the inference is about all existing things, every “similar instance” – everything for which it is
established that it is momentary, independently of this inference – would be included in the
group of things that the inference is about and would, for that reason, not yet have been proven
to be momentary (or the present inference would be unnecessary). Conversely, it is impossible
to present anything that does not belong to this group and yet illustrate the inference through
another case, since by definition such a thing would not exist.
Ratnakīrti’s formulation of the inference of momentariness simply states “as in the case
of a pot” as an example of an instance of momentariness in which the reason “existence” is
present. His opponent asks how the pot could possibly be a similar instance: since the subject
of the inference comprehends all things, the pot included, and the inference aims at proving its
momentariness, the pot thus cannot be used as the example case for which momentariness is
supposed to be already established. Hence the opponent claims that there is no way to establish
the pot’s momentariness: perception does not establish it (we experience temporally extended
things); nor does the inference from existence, as that inference is currently under scrutiny, and
its employment for its own proof would lead to an infinite regress, and there is no other infer-
ence that can establish momentariness (and if there were, the present inference from existence
would be useless).
Ratnakīrti’s answer focuses first on the notion of the capability to produce an effect.12 For
example, a pot which contains water is the cause of the water’s remaining in a certain place;
consequently, it is evident that the pot is presently capable of carrying water. But what about
past or future times in which the pot was or will be? If capability is determined through
observed effect, then Ratnakīrti claims there are only three options: either the pot produces
575
Patrick McAllister
the very same effect of carrying water at those other times, it produces another effect, or it
produces no effect at all. The first option can be dismissed without a long discussion – if true,
it would then follow that the pot is now producing an effect that has already been produced,
a contradiction in terms. The pot’s effect on the water yesterday, its spatially and temporally
determined containment, is not being produced now. Even if the pot broke now, it would not
spill the water yesterday. Thus the pot today or tomorrow cannot be considered capable of
holding the water yesterday.
The second option deserves closer consideration. Both Ratnakīrti and the opponent agree
that one can coherently maintain that the pot is capable of producing the effect it is currently
producing. But should one say that it is capable or incapable of the future or past effects? Evi-
dently, for Ratnakīrti, if the pot produces the present effect because it is capable of it, and if it
were currently capable also of producing the future effects, there would be no reason it should
produce only the present effect but not the future ones in the present moment. Therefore, the
pot must be considered currently incapable of producing the effects that one thinks it produced
before or will produce later. However, to think that the same pot is both capable and incapable
of the same thing (i.e., the future effect) would require assigning it contradictory properties. It
would be more plausible to assume that the thing changes from something that is not capable
to something that is.
The third option, that the pot stops producing effects altogether, is incoherent. As the present
pot producing a present effect shows, capability is temporally directly adjacent to the actual
production of that effect. But if a thing we call “capable” were at some moment not causing
anything at all, then this criterion would be contradicted and we should not call it “capable.”
The only possibility that Ratnakīrti considers coherent is that the pot changes its state from
not being capable of a certain effect to being capable of that effect. Implied in this notion of
capability is that when something is capable of an effect, it must produce it – if it did not, there
would be no reason to call it capable of it. The very idea of potential causation is thus rejected.
Ratnakīrti formalizes and refines this argument with a logical tool that allows for infer-
ences that have no examples. This tool consists of two parts: the formulation of an “unwanted
consequence” ( prasaṅga) and the formulation of a “contraposition of the unwanted conse-
quence” ( prasaṅgaviparyaya). The “unwanted consequence” is an inference showing that a
certain assumption – usually one held by the opponent – implies an untenable commitment.
The “contraposition” is an inference that, based on the negation of that untenable commitment,
proves the opposite of the previous assumption. Ratnakīrti’s formulation of the unwanted con-
sequence is as follows:
That which can be treated coherently as producing something at some time must
produce that thing at that time. Like the last group of causes in a causal line must
produce its specific effect. And, according to you, this pot is something that is cohe
rently treated as producing past or future effects, both at the time of producing the
present effects as well as at the time after all effects have been produced. So there is
an unwanted consequence based on a reason that is a thing’s nature.13
This reasoning is based on the opponent’s assumption that the pot stays the same throughout its
usage, be it yesterday, today, or tomorrow. It can thus correctly be thought of, talked about, and
used as the same single thing that produces effects in the past and in the future. But according
to Ratnakīrti, if it really were always capable of those effects, it would always have to produce
them in every moment, even after it has already produced them. There would be continuous
causation of all effects. This is, obviously, not something that the opponent could accept.
576
Ratnakīrti
The argument seems to be liable to an obvious objection: ordinarily, we agree that the
capability for an effect does not mean that this capability has to be exercised. A grain is, after
all, commonly considered to be capable of producing a shoot, but it does not do so unless it is
in the soil in the right conditions. Likewise, a present pot is commonly held to be capable of
carrying tomorrow’s water.
Ratnakīrti’s answer to this objection relies on the notions of causality and capability that
were described previously: a thing cannot truly be called capable of producing a certain effect
until the moment when it is actually producing that effect. In the moments leading up to that
moment of production, the thing should be characterized only as “the cause of the cause”
(kāraṇakāraṇatva) of the intended effect but not yet the cause of that effect. In the case of
a grain, the property of “generating a shoot” is applicable to a grain only in that state where
it actually is producing a shoot. Before that, it is gradually attaining – in combination with
other factors like humidity, temperature, soil, and so on – a state in which it can properly be
described as capable of producing a shoot. For Ratnakīrti, then, capability is synonymous with
“being a cause.” Capability must always be exercised immediately, since there is no coherent
explanation of why it should be exercised at one time rather than another.
This line of argumentation is complemented with the following “contraposition of the
unwanted consequence”:
That which at some time does not produce something cannot at that time coherently
be treated as being capable of producing something. Like millet, insofar as it does
not produce a rice sprout, cannot be treated coherently as being capable of producing
a rice sprout. And this pot, at the time of producing the present effect and at the time
after all its effects have been produced, does not produce a past or future effect. So
the logical reason of not apprehending a pervading element separates the capable
moment from the incapable moment.14
This reasoning aims to prove that when something does not produce a certain effect, it cannot
be treated as capable of producing that effect. It is the contraposition of the previous reasoning
in that it takes as its premise the negation of what was there the consequence and deduces the
negation of this premise in the previous “unwanted consequence.” The absurd consequence
would be that if it were correct to say that a pot were capable of producing past and future
effects, the pot would have to produce all of these effects now. The contraposition concludes
that since the pot does not produce these effects now, it is not correct to say that it is capable
of producing these effects.
This reasoning is intended to establish the corroborating example for the general inference
of momentary cessation. These two supporting inferences jointly establish that the example in
the inference of momentariness – the pot – is not always in the same state with regard to the
production of an effect. And since the states of not being capable and being capable constitute
contradictory states, the thing to which these states belong cannot be said to remain the same
throughout such changes in states.
A Negative Proof of Momentary Cessation complements A Positive Proof of Momentary
Cessation by investigating whether there are dissimilar instances, cases where momentari-
ness and existence are both absent. Since it is difficult, using the definition of existence as
causal efficiency, to prove that some things lack the property of existence, the actual purpose
of this second work is to show that the opposite is impossible. That is, non-momentariness
must be shown to be incompatible with existence defined as cause and effect both within
the model where a cause precedes its effect and within the model wherein cause and effect
577
Patrick McAllister
coexist. If that can be shown to be the case, then the only conclusion is to accept that exis
tence implies momentariness.15
Concepts and Activity Without a World

Ratnakīrti develops his epistemological framework in A Proof of Exclusion and A Theory of Multi
faceted yet Non-Dual Appearance, as well as in additional discussions contained in A Refutation
of Other Continua. This framework implies the ontological background established and explained
previously in his texts on momentariness. It is useful to treat these texts in accordance with the
degree to which the existence of external objects is tolerated in them: in A Proof of Exclusion,
Ratnakīrti examines how everyday engagement with extra-mental, middle-sized objects is pos-
sible. In A Theory of Multifaceted yet Non-Dual Appearance and A Refutation of Other Continua,
he proposes an analysis of the mind that results in a strong form of idealism and solipsism.
The professed aim of A Proof of Exclusion is to establish the object or referent (artha)
of words.16 By “referent of words,” as the discussion soon shows, Ratnakīrti actually means
universals – in the broad sense of any commonness that the discontinuous and unrepeatable
particulars proved in A Positive Proof of Momentary Cessation and A Negative Proof of Momen-
tary Cessation are thought to share. These “universals” are the objects of any concept-driven
activity, be it linguistic, mental, or physical. Any everyday activity that engages with tempo-
rally extended or similar objects has this “referent of words” as its object: to say “An apple is
on the table,” to think a thought that can be expressed by those sounds, or to purposefully act
in a way that is coherent with an apple being on the table are all activities that are engaged with
“universals” in this sense. The apple that one thinks of as the same from the time it is first seen
on the table to the time it is picked up to be eaten consists, according to the analysis of momen-
tariness, of a causal continuum of apple-moments; “this apple” is already an abstraction of the
individual moments, a result of what Ratnakīrti calls determination (adhyavasāya), the judg-
ment that something (a sequence of apple-moments) is something else (a continuous apple).
A continuous apple is closer in its nature to a universal (like cow-hood) than to a particular:
both the continuous apple and cow-hood must be temporally extended entities.
We have seen that, for Ratnakīrti, to be a real thing is necessarily to be momentary, discon-
tinuous, and unrepeatable. Any notion that posits an object as either temporally extended, con-
tinuous, or repeatable does not accord with reality and is therefore erroneous. Nevertheless,
such a notion can be practically true – reliable in that it enables activity that meets an agent’s
expectations by directing her towards obtaining or avoiding desired or undesired effects of
particulars. Concepts are “true enough” for someone to engage with particulars in everyday
activity, but they do not let a person know those particulars in a stricter sense.
What, then, is the nature of universals or temporally extended objects that are “practically
true”? The answer is given in the theory of exclusion.17 This theory was first formulated by
Dignāga around the fifth century and remained a point of much contention between Bud-
dhists and rival groups for at least seven centuries. Its central postulation is that things are
deemed similar to each other – be it in terms of outward appearance, causal effects, genus,
or otherwise – because, in reality, they share the same difference from, or “exclusion from,”
other things. Though this exclusion is often introduced as the referent of words, it is in fact
the object of any conceptual cognition. An enduring middle-sized object like the apple on the
table would thus be analyzed not as a substance in which certain slowly changing properties
inhere but rather as something that is defined only through a set of distinctions from other
things: momentary phases that are causally linked in a certain way so as to be judged a con-
tinuum. The phases constituting such a sequence can be judged to be the same in some way
578
Ratnakīrti
(say, as distinguished from what are not apples or from what is in a different place); any phase
in such a continuum can cause, in someone already used to the idea of dealing with temporally
extended objects, the notion of a slowly changing object in the same place.
Ratnakīrti, like many authors in the tradition of Buddhist epistemology who have broached
this subject, admits frankly that this analysis is very different from what everyday experience
looks like to most people: moving about in the world, one does have a strong feeling that there
are middle-sized, continuous objects. His justification of this gap relies on two arguments: (1)
an ontological one that shows that, apart from exclusion, no other object of words and con-
cepts is actually possible and (2) an epistemological one that explains how these appearances
can cohere with the world as it really is.
Ratnakīrti characterizes his opponents as proposing one of three kinds of objects as possible
referents of words: either a particular, a true universal (like cow-hood) that may or may not be
connected to a particular, or the form of awareness. This “form of awareness” (buddhyākāra)
is the technical term for the object that any given awareness is directly aware of and with
which it is identical. For the present discussion, it is best to think of it as the internal object
of cognition. For Ratnakīrti, this form of awareness is a real thing like a particular, but not a
mind-external, one.
A pure particular, unqualified by a genus, and the form of awareness are discarded as viable
candidates for the referent of a word without much discussion: speakers do not use the word
“cow” for only one particular cow, be it a particular in the strict sense of the theory of momen-
tariness or even in the loose sense of a single cow, and the form of awareness that a person has
when hearing or saying a word is not what a speaker usually refers to. Both of these objects are
useless as the referent of words: though they can be referred to, neither the one particular cow
nor the imagined cow are things that can satisfy a central and observable function of the word
“cow” – that it refers to all cows. Hence, A Proof of Exclusion consists mostly of refutations
that a universal or genus is the referent of a word, either directly or as connected to – but still
separable from – particulars.
Ratnakīrti’s own definition of the word referent is this: “The word referent is a positive ele-
ment qualified by other-exclusion.”18 With this formulation, A Proof of Exclusion effectively
characterizes the referent of a word as something that has the structure of a property and its
substrate (dharma and dharmin); the property in this case is other-exclusion (apoha), and that
which is qualified by it is a “positive element.” Ratnakīrti tells us that this term, “positive ele-
ment,” refers to two things: an external particular19 and the form that a particular awareness
has. This internal form is also a particular, but not an extra-mental, one. While the external
particular does not appear in conceptual cognition, whose objects are exclusions, both items
are factors in any conceptual cognition and, as such, constitute two aspects of the object of
conceptual cognition. Yet it is only the external particular that is considered the referent of a
word (though, as we shall see, not directly but rather through its quality of other-exclusion).
This follows from the fact that any predication that is performed in a conceptual cognition
must refer to an external object.
One of Ratnakīrti’s examples concerns the property “existence”: to attribute existence to
the form of a particular awareness is redundant, since a present form must exist. To negate the
present form as existent is incoherent because, again, predications that are performed in a con-
ceptual cognition must refer to an external object. Predicating existence or its absence about
external things is, however, useful and not redundant.
The underlying reason for this argument is that the internal form is the object of a spe-
cial kind of perception, that of awareness by itself (Ratnakīrti and his tradition call this self-
awareness, svasaṃvedana). Insofar as a thing is known through perception, it is fully known:
579
Patrick McAllister
at least since Dharmakīrti, Buddhist epistemologists agreed that the distinction of properties
and substances was not real and that perception shows its object as it really is; this means that
any perception must show its object fully, with all the parts that might later be analyzed as
properties or substrates. Ratnakīrti’s point is that this form appearing in awareness must, by
definition, be known in all its aspects. Predications about it would thus be redundant: anything
true that could be said about it would already be known, and anything false about it would be
contradictory.
A conceptual cognition does not apprehend an external thing through such a direct per-
ception, by definition. Rather, the external thing is only indicated, and not shown, through
determination. Determination is a judgment which, caused by the specific nature of the mind-
internal particular form, itself causes activity directed at an object that is construed as external
and enduring. Its truth consists in enabling everyday activity according to an agent’s expecta-
tions, even though the temporally extended object that the agent takes to be acting towards it
does not exist. The two “positive elements” are keyed to two modes of awareness: the form of
awareness is known through self-awareness, in a perceptual mode, and is thus known in all its
particular characteristics; the external particular is known indirectly through a determination
that, based on the particular form of awareness, instigates activity, the agent of which believes
to be acting towards a temporally stable, unitary object. This fictional object might, if the cog-
nition is “true enough,” align with an actual, external particular.
Ratnakīrti’s definition of a word’s object attributes the property of other-exclusion to these
two positive elements. This property is likewise analyzed as internal and external since it
qualifies both positive elements. In the case of the external element, the exclusion from others
is said to result from the specific set of causes that generate the particular: through these, the
particular is “restricted” to a particular causal capacity that makes it useful (or not) for a par-
ticular aim. For Ratnakīrti, it is clearly coherent to say that two particulars can share an “exclu-
sion from others” and that they have nothing real in common.20 This allows him to maintain,
on the one hand, that two particulars are really completely different from each other and, on the
other, that their treatment as having something in common is successful. For Ratnakīrti, this
position is tenable because an “exclusion from others” is a form of absence or negation but not
a substantially real universal (as it is for all his opponents). The property of “being a horse” can
be absent from two particulars without those two particulars therefore sharing a substantially
real property. They can thus be held to be excluded from horses without commitment to a real,
shared property. What his opponents generally take to be the effect of universals – that certain
lumps of matter share a common property like “being a cow” – is analyzed in terms of sets
of differences. To be a cow is to be qualified by a certain set of differences from other things
that are not cows. Insofar as a cow is an animal, it will share these differences with those other
animals (usually, they are distinguished from horses). But to all these differences common to
cows and other animals, there must be differences specific to cows (notably, a dewlap, horn,
tail, and so on). Each cow, of course, has its further individual differences from all other cows.
And, if analyzed carefully, each moment of an individual cow’s existence has its own set of
differences from all other moments in that causal chain. This theory of universals is thus fully
compatible with the theory of momentariness.
Ratnakīrti does not even pretend that this is “how things appear” in conceptual cognition.
All these sets of differences, which should be infinitely many, do not appear in cognition at all.
Ratnakīrti’s solution to this is a rather surprising one: the form of awareness, the mind-internal
positive element described previously, is what appears, yet there is always a simultaneous
grasping of the differences from other things. This apprehension of the differences is equated
by Ratnakīrti to a capacity (śakti), modeled on the stock example of the grasping of absence:
580
Ratnakīrti
as the perception of an empty stretch of floor is capable of generating infinite ascertainments

(all of the form “X is not here”), so the internal positive element, shown by self-awareness,
contains the ability to form such ascertainments. This means that to be a competent user of
the word “cow” is to be able to, upon hearing the word “cow,” act (in speech or otherwise)
according to that set of differences that the word “cow” is accepted to refer to. Such a person
could say both “This is not a horse” and “This is an animal,” or, to use Ratnakīrti’s example of
hearing the command “Tie up the cow!”, tie up a cow but not a horse. The emphasis here is on
the possibility because not every difference has to come to awareness every time that the word
is heard. In other words, the form that an awareness takes at any particular time is the basis for
a certain set of possible predications of differences.
This also suggests that the attribution of “other-exclusion” to the “positive element”
must be understood ontologically not only in the case of the external particular but also
in the case of the form of awareness, where it could also be understood as concerning
the content of the cognition: the external particular and the form of awareness are both
particulars in that they are caused by a specific set of causes and for that reason also have
particular causal capacities. What appears to cognition whenever the word “cow” is heard
(or used) need not be the same every time;21 the forms of these awareness-events just need
to be similar in the sense that they cause a person to act in accordance with the same set of
differences. If there is such an act, it will have been caused by an awareness whose causal
capacity is defined by the causal capacity of the awareness that arose on hearing (or using)
the word “cow.” Hearing a statement such as “Tie up the cow!” twice will, in a listener who
understands the statement, twice create an awareness whose form is such that it delimits
the activities compatible with the command to tying up the cow; the listener would also, in
all likelihood, say that she has received the same command, understood it the same way,
and did the same thing in accordance with it. But none of these instances of sameness are
real. Instead, two spoken sentences caused two awareness-events that, due to the particular
characteristics of the awareness at that time (i.e., previous experiences that influence the
person’s understanding of that command, along with a deep habituation to erroneous beliefs
that there are enduring, extra-mental objects), resulted in two actions that one might judge
to be the same (i.e., whose observation would lead to other awareness-events that share
similar sets of differences).
The last important feature of Ratnakīrti’s theory of exclusion is that he lays great stress on
the simultaneity of the appearance of the positive element and its quality, the exclusion from
others. It is not possible to say that one of them is the primary and the other the secondary
element in a conceptual cognition, in terms of the sequence in which one becomes aware of
them. Since Ratnakīrti does not accept a real difference between properties and the things they
qualify, they are also not known apart from each other: to perceive something is to perceive
both what can be analyzed as the substrate and what can be analyzed as its properties in the
exact same moment. This holds true for the substrate that is the form of cognition and its
ephemeral quality, other-exclusion, since the substrate is known through a perception of the
type self-awareness. For the external thing with its quality, there is no discussion of the simul-
taneous perception of the property and its bearer because it is not directly perceived through
any conceptual cognition. Ratnakīrti illustrates this with the word indīvara, which is a name
for blue water lilies that cannot be analyzed as a compound of two parts, one for “blue” and
one for “water lily”: someone who knows what the word means does not become aware of
two different things in sequence, the color blue and a water lily (as could perhaps be the case
if one heard the phrase “blue water lily”); rather, there is a simultaneous awareness of both the
color blue and a water lily. It is the same in the case of the word “cow,” which refers to what is
581
Patrick McAllister
different from non-cows: hearing it, a competent user of that word will have a form of cogni-
tion that is qualified by the according exclusion of non-cows. As stated before, this quality is a
capacity of the cognition, not an actual presentation of everything that is not a cow.
The theory of other-exclusion is thus a device to bridge the apparent gap between the realm
of particulars and that of conceptual cognitions, which appear to deal with universals, unreal
insofar as they are devoid of causal ability. This theory reinterprets these universals as the dif-
ferences or exclusion from others. They are the only general things acceptable to Ratnakīrti
because they are not real things in the causal sense. In the final analysis of our interaction with
everyday objects, Ratnakīrti would conclude that there is no gap because the conceptual can
be fully reduced to particulars as well: the infinitely many sequences of causes and effects that
constitute the extra-mental world and the infinite causal sequence that constitutes awareness.
Any lasting or repeatable phenomenon is unreal and ephemeral, figuring in any cognition only
as a horizon at which activity may be directed.
The Transition to Idealism

This justification of the difference between how everyday objects purportedly appear and how
they are reasonably analyzed allows Ratnakīrti to make two important further points: first, he
can both explain and deconstruct everyday experiences of extra-mental objects in terms of
awareness-events that are erroneously directed “outward,” facilitating a move towards ideal-
ism; second, he can tie the Buddhist theory of nescience and the possibility of liberation to the
very core of everyday experience.
Further details of this strategy can be culled from A Theory of Multifaceted yet Non-Dual
Appearance. This text is a highly condensed and focused treatise based on Jñānaśrīmitra’s
works that discuss the teaching that awareness has a real form (sākāravāda) and whose central
concerns – to show that awareness must always have a form; what that form is; and what this
means for the analysis of cognition, reality, and liberation – are not yet well understood.22
There is no doubt, however, that, at least for A Theory of Multifaceted yet Non-Dual Appea
rance, there is a clear link between the following tenets: cognition always has a form of a
well-defined kind; determination (the primary result and expression of the nescience of unen-
lightened beings) works without dependence on any extra-mental, real entities; and liberation
and its opposite, saṃsāra, can be understood in terms of determination.
A Theory of Multifaceted yet Non-Dual Appearance discusses these matters in a tightly
defined context: Ratnakīrti assumes that any and all extra-mental objects have been proven
to not exist; all theories that awareness does not have or show a form at all (or that its form is
somehow ephemeral or unreal) have been refuted; and the whole world has been established
as only the appearance of a form to awareness, like what appears in a dream.23 A rather specific
form of idealism is thereby presupposed, in which external objects are like objects in a dream,
devoid of extra-mental existence but real – causally efficacious – in the context of the dream
or insofar as they are cognitions.
The main thesis that Ratnakīrti seeks to establish in A Theory of Multifaceted yet Non-Dual
Appearance is that awareness has a single yet variegated form.24 Typically, this point is pre-
sented as an inference:
What appears, that is one. Like the form of blue occurring amidst other forms. And
this collection of various forms, such as white, the sound “ga,” sweet, fragrant, soft,
pleasure and its opposite, etc., appears. This is a proof using an essential nature as a
reason.25
582
Ratnakīrti
This inference is formally valid: to appear (the reason) is to be one (the probandum). The
example supplied is an apprehension of the color blue, as blue never appears by itself but
instead appears always together with other “forms” (shapes, sounds, etc.). The reason is then
said to apply to a variegated form of awareness, all aspects of which appear together at the
same time and therefore must be a unit. The discussions focus largely on the unity the form of
cognition possesses: according to Ratnakīrti, everyone agrees that something appears and that
it is something manifold that appears – after all, everyone experiences a world of variegated
objects. But Ratnakīrti claims that what appears to be manifold is actually a single, unitary
thing. In being aware of something blue, one is not aware only of blue: blue appears amidst a
group of other appearances (both external ones like sounds, smells, shapes, as well as pleasure
and other factors attributable to the subject of the experience). Yet in the end, the whole world,
including its temporal and spatial distinctions, is nothing but a single, variegated form that is
of the nature of awareness (see Kajiyama 1965, 12).
The opponent objects to Ratnakīrti’s inference that the example (the form of blue) is actu-
ally not one with the variegated yet unitary form of awareness that is under discussion in the
inference since the two forms, the example and the subject of the inference, appear at different
times in different places and look different. But if they were identical, the example would not
be admissible since the example in an inference needs to be distinct from the case that the
inference is about.26 Ratnakīrti’s answer is that the counter instance – a form that would not
be unitary but less or more than one – is known to not appear. How is this known, the oppo-
nent immediately asks, since in order to know this counter instance, it would certainly have
to appear to cognition? The trap is obvious: if it appears and is not one, the main inference is
shown to be based on a wrong relation; if it does not appear, then how can we ascertain that it
is not one? Ratnakīrti responds that the opponent cannot possibly mean that it should appear
directly: insisting on that requirement would make most inferences impossible, since certainty
that the reason is never present where the property to be proven is absent is a precondition for
inference (e.g., the certainty of the absence of oak where there is no tree), but how could this
co-absence ever be certain if it were required that all cases where the property to be proven is
absent must appear to awareness? The counter instances are presented, rather, through deter-
mination. In this case, the actual counter instances need not appear but are cognized indirectly,
as a group of cases for all of which this co-absence is true; in other words, all counter instances
are known by a unifying general feature, the absence of the reason and the probandum. The
inference that lets one know this group in the present case is of the same type as the one used
to establish the momentariness of the example in the inference to general momentariness.27
Ratnakīrti’s discussion of determination presents itself as an explanation of how a counter
instance can be known. The larger issue, however, is how determination makes it possible to
apprehend something other than cognition without this other object having to be really differ-
ent from cognition. Awareness is always one, and therefore anything that appears in it would
be both real and identical with it, insofar as it appears in it. Yet insofar as what appears in
awareness is determined – that is to say, produces activity that is directed at something that
is not awareness – one may speak of external objects. This is how one can meaningfully act
upon things with an ontologically problematic status, that is, past and future things, impossible
things, and things whose existence is doubtful. They are real as appearances to awareness, but
they are not real as they are acted towards.
The main inference of A Theory of Multifaceted yet Non-Dual Appearance allows Ratnakīrti
to give a clever example of this. The inference holds only if it can be shown that what is not
of a single form – that is, any counter instance – cannot appear. To establish this, however,
the counter instance must be known in some way. This is the puzzle that must be solved by
583
Patrick McAllister
determination: it should give cognitive access to something that is not directly present and that
does not have to exist in the way that it is accessed, even though it must exist in some way so
that it can be accessed. The basis of Ratnakīrti’s solution was already described in the treatment
of other-exclusion previously: the form of awareness directly appears and has the capacity to
cause activity that conforms to its properties (the other-exclusions that qualify it). Speaking
from the position of radical idealism in A Theory of Multifaceted yet Non-Dual Appearance,
Ratnakīrti adds to this analysis that the form of awareness is not really caused by external things
at all but only by awareness itself, given that external things do not exist and hence cannot cause
anything. Likewise, determination may direct activity outwards, and one of those activities can
be acting towards an object in conformance with the form of awareness. But determination nei-
ther is caused in any way by an external thing, nor does it show an external thing, nor does it let
anyone really obtain an external thing. All it truly does is give rise to the next momentary phase
of awareness, continuing a process that only the achievement of liberation can end. Liberation
is defined accordingly: determination is saṃsāra; its destruction liberation (mokṣa).28 Determi-
nation is thus not only a central function within Ratnakīrti’s epistemology but also a function of
the nescience that makes living beings stay within the cycle of existence. Determination’s main
operations, externalizing and generalizing objects, are what makes everyday activity possible
but also what makes everyday activity an expression of the suffering of unenlightened beings.
To end determination is thus, for Ratnakīrti, to attain liberation.
Notes
1 “A Proof of an Omniscient Being (Sarvajñasiddhi)” (Thakur 1975, 1–31).
2 “A Refutation of the Proof of God (Īśvarasādhanadūṣaṇa)” (Thakur 1975, 32–57; see Patil 2009).
3 “A Positive Proof of Momentary Cessation (Kṣaṇabhaṅgasiddhiḥ – anvayātmikā)” (Thakur 1975,
67–82); “A Negative Proof of Momentary Cessation (Kṣaṇabhaṅgasiddhiḥ – vyatirekātmikā)”
(Thakur 1975, 83–95); “A Refutation of the Proof of Enduring Things (Sthirasiddhidūṣaṇa)” (Thakur
1975, 112–28); “A Proof of Exclusion (Apohasiddhi)” (McAllister 2020, 47–82); “A Theory of Mul-
tifaceted yet Non-Dual Appearance (Citrādvaitaprakāśavāda)” (Thakur 1975, 129–44).
4 “A Refutation of Other Continua (Santānāntaradūṣaṇa)” (Thakur 1975, 145–49).
5 “A Treatise on the Existence of Other Means of Valid Cognition (Pramāṇāntarbhāvaprakaraṇa)”
(Thakur 1975, 96–105).
6 “An Assessment of Pervasion (Vyāptinirṇaya)” (Lasic 2000).
7 Mimaki (1976, 9–45) has shown how Ratnakīrti recasts the historical discussions of the Buddhist theory of
momentariness as a logically ordered whole. He is shown to have separated the discussions in the Exami-
nation of Enduring Entities (Sthirabhāvaparīkṣā) and the Examination of the Relation of Action and Result
(Karmaphalasambandhaparīkṣā) of Śāntarakṣita’s “A Summary of What Is the Case (Tattvasaṅgraha)”
(Krishnamacharya 1926), an encyclopedic work of Sanskrit classical philosophy, into three works: A Posi-
tive Proof of Momentary Cessation, A Negative Proof of Momentary Cessation, and A Refutation of the
Proof of Enduring Things. It is reasonable to assume that the characteristics described by Mimaki (1976,
9–45) would apply also to Jñānaśrīmitra’s “Section on Momentary Cessation (Kṣaṇabhaṅgādhyāya)”
(Thakur 1987, 1–159) so that Ratnakīrti’s creativity would lie, as usual, in the reorganisation and com-
pression of his teacher’s material rather than in the masterful analysis of the history of the arguments for
momentariness. For a very useful tabular overview of the correspondences, see Mimaki 1976, 45.
8 A Positive Proof of Momentary Cessation 67.
9 See A Negative Proof of Momentary Cessation 83.
10 This question marks the main difference to the second text, A Negative Proof of Momentary Cessa-
tion, which discusses whether there are (and whether we can know) instances in which existence is
present but not momentariness.
11 See A Positive Proof of Momentary Cessation 67.
12 This paragraph and the next are essentially a paraphrase of A Positive Proof of Momentary Cessation
68, though the illustration of the pot containing water has been added.
13 See A Positive Proof of Momentary Cessation 68. See Mimaki 1976, 56 for an analysis of this inference.
584
Ratnakīrti
14 See A Positive Proof of Momentary Cessation 69.

15 This is a paraphrase of Ratnakīrti’s closing statement in A Negative Proof of Momentary Cessation
95.
16 A Proof of Exclusion, 47: “Exclusion is declared as the referent of words.”
17 See Patil 2003; Patil 2009; McAllister 2020 for the following paragraphs.
18 A Proof of Exclusion, 49.
19 That the “external positive element” is a particular is the present author’s interpretation (see McAl-
lister 2020, 235–58). Patil (2009, 281–83) understands Ratnakīrti to be talking about an external, but
general, object, a continuum as a whole in the case of everyday objects.
20 Compare also Dharmakīrti’s famous example of the medical herbs in “Inference for Oneself (The
First Chapter in the Additional Explanations to Dignāga’s Compendium of Means of Valid Cogni-
tion)” in Gnoli (1960, 74), translated by Eltschinger et al. (2018, 85) as: “Or else, to give another
example ( yathā), one observes that in spite of their diversity, certain plants and not others are capable,
whether individually or collectively, of alleviating fever, etc.”
21 But Ratnakīrti does clearly assume that a sort of generalized image-and-speech-sound appears in
hearing words (A Proof of Exclusion, 60). But insofar as this is always a particular form of awareness,
it cannot ever be truly the same as another.
22 See Moriyama 2011, 2012 for a translation into Japanese (which the present author does not read) and
the introduction to Yūichi Kajiyama 1965 for a general characterization of the form and history of
the idealism defended by Ratnakīrti. For Jñānaśrīmitra, see “A Systematic Exposition of the Proof of
Cognition Having a Form (Sākārasiddhiśāstra)” (Thakur 1987, 367–513) and “Collected Statements
on Cognition Having a Form (Sākārasaṃgrahasūtra)” (Thakur 1987, 515–78). Ratnakīrti’s A Theory
of Multifaceted yet Non-Dual Appearance also includes material from Jñānaśrīmitra’s “Treatise on
Exclusion (Apohaprakaraṇa)” (Thakur 1987, 201–32; trans. in McCrea and Patil 2010), attesting to
the close connection of the topics.
23 See Kellner and Taber 2014 for a general characterization of this form of idealism. Ratnakīrti bases
himself on Dharmakīrti’s sahopalambhaniyama argument, which, on the principle that two things
that always occur together are identical, aims to establish that cognition and its object are not different
from each other. Ratnakīrti quotes Dharmakīrti’s “First Chapter in the Ascertainment of the Means
of Valid Cognition (Pramāṇaviniścayaḥ)” verse 54ab (Vetter 1966) at A Theory of Multifaceted yet
Non-Dual Appearance 129.
24 To say that awareness “has a form” is to say that awareness shows that form or that one is aware of the
form. The sākāra accepts that awareness is self-illuminating and holds that awareness really assumes
the form of its object.
25 A Theory of Multifaceted yet Non-Dual Appearance 129. See Kataoka 2017; McAllister 2017a for
investigations of the sections in which determination is discussed. “The sound ‘ga’ ” is the third note
of a common solmization in Indian classical music.
26 For the background of this position (technically called bahirvyāptivāda), see Bhattacharya 1986.
27 See Kajiyama 1998, 116 ff., n. 310 for a succinct description of such an inference.
28 See A Theory of Multifaceted yet Non-Dual Appearance 137, 9–16.
Bibliography
Bhattacharya, Kamaleswar. 1986. “Some Thoughts on Antarvyāpti, Bahirvyāpti, and Trairūpya.” In Bud-
dhist Logic and Epistemology. Studies in the Buddhist Analysis of Inference and Language, edited by
Bimal Krisha Matilal and Robert D. Evans, 89–105. Dordrecht: Reidel.
Eltschinger, Vincent, John Taber, Michael Torsten Much, et al. 2018. Dharmakīrti’s Theory of Exclusion
(apoha). Part 1, On Concealing: An Annotated Translation of Pramāṇavārttikasvavṛtti 24, 16–45, 20
(Pramāṇavārttika 1.40–91). Tokyo: The International Institute for Buddhist Studies.
Gnoli, Raniero, ed. 1960. The Pramāṇavārttikam of Dharmakīrti: The First Chapter with the Autocom-
mentary; Text and Critical Notes. Rome: Istituto Italiano per il Medio ed Estremo Oriente.
Kajiyama, Yūichi. 1965. “Buddhist Solipsism – A Free Translation of Ratnakīrti’s Saṃtānāntaradūṣaṇa.”
Journal of Indian and Buddhist Studies (Indo-Gaku Bukkyō-Gaku Kenkyū) 13 (1): 420–35 (9–24).
———. 1998. An Introduction to Buddhist Philosophy: An Annotated Translation of the Tarkabhāṣā of
Mokṣākaragupta. Reprint with corrections in the author’s hand. Vienna: Arbeitskreis für Tibetische
und Buddhistische Studien Universität Wien.
585
Patrick McAllister
Kataoka, Kei. 2017. “Dharmottara’s Notion of āropita: Superimposed or Fabricated?” In Reading Bhaṭṭa
Jayanta on Buddhist Nominalism, edited by Patrick McAllister, 217–50. Vienna: Austrian Academy of
Sciences.
Kellner, Birgit, and John Taber. 2014. “Studies in Yogācāra-Vijñānavāda Idealism I: The Interpretation of
Vasubandhu’s Viṃśikā.” Asiatische Studien/Études Asiatiques 68 (3): 709–56.
Krishnamacharya, Embar, ed. 1926. Tattvasaṅgraha of Śāntarakṣita With the Commentary of Kamalaśīla,
2 vols. Baroda: Central Library.
Lasic, Horst. 2000. Ratnakīrtis Vyāptinirṇaya: Sanskrittext, Übersetzung, Analyse. Vienna: Arbeitskreis
für Tibetische und Buddhistische Studien Universität Wien.
Matilal, Bimal Krishna, and Robert D. Evans, eds. 1986. Buddhist Logic and Epistemology. Studies in the
Buddhist Analysis of Inference and Language. Dordrecht: Reidel.
McAllister, Patrick. 2017a. “Competing Theories of Conceptual Cognition. Dharmottara and Trilocana
vs. Dharmakīrti?” In Reading Bhaṭṭa Jayanta on Buddhist Nominalism, edited by Patrick McAllister,
291–321. Vienna: Austrian Academy of Sciences.
———, ed. 2017b. Reading Bhaṭṭa Jayanta on Buddhist Nominalism. Vienna: Austrian Academy of
Sciences.
———. 2020. Ratnakīrti’s Proof of Exclusion. Vienna: Austrian Academy of Sciences.
McCrea, Lawrence J., and Parimal G. Patil. 2010. Buddhist Philosophy of Language in India:
Jñānaśrīmitra on Exclusion. New York: Columbia University Press.
Mimaki, Katsumi. 1976. La réfutation bouddhique de la permanence des choses (sthirasiddhidūṣaṇa) et
la preuve de la momentanéité des choses (kṣaṇabhaṅgasiddhi). Paris: Institut de Civilisation Indienne.
Moriyama, Shinya. 2011. “An Annotated Japanese Translation of Ratnakīrti’s Citrādvaitaprakāśavāda
(1).” South Asian Classical Studies 6: 51–92.
———. 2012. “An Annotated Japanese Translation of Ratnakīrti’s Citrādvaitaprakāśavāda (2).” South
Asian Classical Studies 7: 135–67.
Patil, Parimal G. 2003. “On What It Is That Buddhists Think About – apoha in the Ratnakīrti-
Nibandhāvali.” Journal of Indian Philosophy 31 (1–3): 229–56.
———. 2009. Against a Hindu God: Buddhist Philosophy of Religion in India. New York: Columbia
University Press.
Thakur, Anantalal, ed. 1975. Ratnakīrti-nibandhāvaliḥ: Buddhist Nyāya Works of Ratnakīrti. 2nd ed.
Patna: Kashi Prasad Jayaswal Research Institute.
———. 1987. Jñānaśrīmitranibandhāvali: Buddhist Philosophical Works of Jñānaśrīmitra. 2nd ed.
Patna: Kashi Prasad Jayaswal Research Institute.
Vetter, Tilmann. 1966. Dharmakīrti’s Pramāṇaviniścayaḥ 1. Kapitel: Pratyakṣam. Einleitung, Text
der tibetischen Übersetzung, Sanskritfragmente, deutsche Übersetzung. Vienna: Austrian Academy of
Sciences.
586
36
RATNĀKARAŚĀNTI
The Illumination of False Forms
Gregory Max Seton
Introduction
Renowned as the “Omniscient One of the Degenerate Age” (Sankrtyayana 1935, 35 n. 4),
Ratnākaraśānti (c. 970–1045 CE) was considered one of the foremost Indian Buddhist scholars
and tantric adepts (siddha) of his time (Guide, 231a1; Chimpa and Chattopadhyaya 1990, 295,
313; Roerich 1988, 206).1 But his signature contribution to philosophy was his unique onto-
epistemological answer to the Mahāyāna buddhological problem of how a buddha, who per-
ceives the ultimate reality, can communicate with and bring benefit to ordinary persons, who
perceive merely conventional reality. Since Ratnākaraśānti’s answer to this question relies on
his buddhological assertion that a buddha retains cognitive error (bhrānti), it implicitly con-
tradicts the foundational Buddhist premise of a buddha’s infallibility and cognitive perfection
and, not surprisingly, was considered controversial and criticized by most subsequent scholars.
But his tantric commentaries, into which he often seamlessly wove this philosophical view-
point, appear to have been widely appreciated for their ability to address the many pressing
contemporary soteriological questions, such as why and how an obviously fallible human
being might attain buddhahood within a single lifetime through antinomian tantric practices.
This chapter will focus on explaining Ratnākaraśānti’s philosophical position, which claims
to be the correct interpretation of both the Madhyamaka philosophy founded by Nāgārjuna and
the Yogācāra philosophy explained by Asaṅga and Maitreya (Guide, 231a3). By claiming this,
Ratnākaraśānti distinguishes his own position from the positions of prior Buddhist scholars.
Among these, he explicitly calls Dharmottara and Śubhagupta materialists (bahirārthavādin)
and refutes them (Proving Mere Representation, 309a2–3). He also refutes the Yogācāra view-
point of Prajñākaragupta as nonsensical (Proving Mere Representation, 309a5–7; Instructions,
143a6 – b2, 148b1–49a5; Nondual Variegated, 137) and the “pseudo-Mādhyamika” (dbu ma ltar
snang ba) Candrakīrtian viewpoint as nihilistic materialism (Guide, 226b4; Instructions, 143b4).
Despite his finding more common ground with the epistemological position of Kamalaśīla,
Ratnākaraśānti rebuts its refutation of the ultimate reality of mind (Instructions, 151b1–6).2
Before explaining Ratnākaraśānti’s philosophical intervention, we will examine the central
Mahāyāna problem of how an awakened buddha’s immanent awareness can know what an
ordinary confused person perceives.
DOI: 10.4324/9781351030908-51 587

Gregory Max Seton
The Immanent Awareness Problem and Ratnākaraśānti’s

Intervention
According to the Mahāyāna, ordinary persons needlessly suffer in cyclic existence, because,
just as in a dream, their sleeplike ignorance causes them to cling to a dreamlike personal iden-
tity that experiences a constant state of existential dis-ease (saṃskāraduḥkhatā). However, a
buddha (lit. Awakened One) overcomes this ignorance and suffering by directly perceiving the
ultimate reality of all dreamlike phenomena perceived by ordinary persons. After awakening,
a buddha comes to have three so-called bodies, which are ultimately not distinct but are merely
the transcendent and immanent aspects of buddhahood perceived as different bodies through
the subjectivity of unawakened ordinary persons, partially awakened beings (bodhisattva), and
fully awakened ones (buddha) (Thurman 2004, 93–97; Mahāyāna Scriptures, ix.56–66). First,
a buddha’s consciousness is transformed into a nonphysical “natural body” (svābhāvikakāya)
– also called the true body (dharmakāya) – which embodies the true nature of a buddha that
she only knows. Due to a buddha’s prior aspirations to help others awaken from their dream-
like states, the transcendent natural body manifests in the perception of partially awakened
bodhisattvas as an immanent body of enjoyment (sāṃbhogikakāya) and in the perception of
unawakened ordinary persons as an immanent body of emanation (nairmāṇikakāya).
Despite agreeing that buddhas manifest three bodies to help awaken the two other types of
beings, the two main Mahāyāna traditions, Madhyamaka and Yogācāra, debated what it is like
for a buddha to fully overcome ignorance and still know the reality of deluded persons in order
to communicate with them and lead them out of their own dreamlike cyclic existence. On the
one hand, the Mādhyamika skeptics, who claimed to refute all ontological reals, essentially
sidestepped the epistemological question by suggesting that what a buddha is truly like is
beyond all conceptual proliferations (niṣprapañca) and that his so-called bodies only appear
from others’ perspectives to be communicating with them. On the other hand, the monistic
Yogācāra schools – who held consciousness to be the only ontologically real thing – offered
nuanced epistemological explanations of how the immanent awareness of a buddha could per-
ceive simultaneously the ontological reality accessible only to buddhas and the conventional
reality perceived by ordinary deluded persons.
According to Yogācāra ontology, while ordinary persons may feel as though they experi-
ence things external to consciousness, those things cannot be established as anything other than
mind. That is to say, ordinary persons may think that they are seeing various colored things that
exist somehow in the world. But their consciousness (vijñāna) – that is, the intentional “being
conscious of ” something – is merely illuminating and knowing perceptual content (ālambana)
in “representational forms” (ākāra) that are made of a mental substance rather than material sub-
stance external to mind. Although dualists might argue that external material things can be infer-
entially established as producing those representational forms, the Yogācāra scholars respond,
in effect, that the putative correspondence between external phenomena and their mental repre-
sentations could never be established, since consciousness, which has no way to access anything
other than its own representational forms, could never know anything outside itself or verify the
correspondence between its own mental representations and the supposedly external referents of
those representations (e.g., Twenty Verses 3, 30). Furthermore, since the representational forms
must be mental in substance and, at the very least, shaped by concepts and linguistic structures,
they “represent” only the content of mind’s latent impressions produced by prior mental activi-
ties (karma) that take on particular forms due to ignorance, like appearances in a dream.
Although all Yogācāra scholars agree that the external things imagined by ordinary persons
are ontologically grounded in nothing but consciousness, the two main camps of Yogācāra
588
Ratnākaraśānti
epistemology disagree on what that consciousness is ultimately like, especially once its basis
undergoes a fundamental transformation (āśrayaparāvṛtti) upon awakening from that dream.
Among the two camps, the “Form-bound” (sākāra) scholars claim that consciousness is inten-
tional by definition and must be ultimately “bound to” representational forms.3 Their rivals, the
“Form-free” (nirākāra) scholars, argue that the consciousness of an awakened being is intentional
but in a way that is different from ordinary beings and ultimately “free” from representational
forms. On the surface, the “Form-bound” and “Form-free” disagreement may seem to revolve only
around the ontological existence or nonexistence of representational forms at the level of a buddha.
But, in fact, their epistemologies entail fundamentally divergent understandings of the nature of
representational forms, and these are reflected in their rival theories of consciousness and error.
For Form-bound scholars such as Dharmapāla (sixth century), the fundamental cognitive
error in an ordinary person’s consciousness leads to a conceptual misinterpretation of the rep-
resentational forms, which are viewed as ontologically distinct from consciousness. But a bud-
dha’s consciousness, which eliminates this cognitive error, sees the very same representational
forms that ordinary beings see, except it apprehends them in a more accurate way without
superimposing externality and other conceptual proliferations. Hence, a buddha’s conscious-
ness is intentional, like that of an ordinary person. But a buddha’s consciousness, rather than
misinterpreting the nature of the representational forms, is reflexively aware that its intentional
objects are nothing but mind.
For Form-free scholars such as Sthiramati (sixth century), the fundamental cognitive error
is a drifting (viplava) of consciousness into the false and nonexistent representational form of
a grasping subject (grāhaka) and its grasped objects (grāhya). The characteristics of a grasp-
ing subject and its grasped objects are shot through with conceptualization and only imagined
( parikalpita) to be real. But since the term “representational form” refers to a moment of
consciousness in the mode of grasping momentary content (ālambana), its referent is only the
grasping subject, not the grasped object. In other words, the representational form of a grasping
subject is simply a false appearance (vitathapratibhāsa) of real consciousness, but a grasped
object, which is not a representational form, has nothing true or real about it (Middle-Beyond-
Extremes, 16). In any event, when the fundamental cognitive error – on which the imagined
nature of the grasping subject and grasped object are dependent ( paratantra) – dissolves, then
both subject and object are eliminated and consciousness’s own inherent nature is realized
( pariniṣpanna) to be free of subject-object structuring (Middle-Beyond-Extremes, 84–85, 143;
Instructions, 138a4 – b7).4 Hence, unlike the Form-bound notion, the Form-free scholars hold
that for a buddha at the ultimate level, consciousness is free of subject-object duality and is
thus intentional only in the sense that the sheer existence of all phenomena manifests directly
within a buddha’s own mind (Textual Commentary ad 3627, 6–8; Funayama 2007, 194–96).5
Skeptical critics of these two Yogācāra onto-epistemologies, like Kamalaśīla, pointed out
many weaknesses and incoherencies in their explanations of what consciousness is ultimately
like and how a buddha’s mind could be aware of the nature and variety of limitless objects and
still communicate with ordinary people perceiving different particular conventional realities
(Textual Commentary ad 535, ad 3627). In brief, they argue that the Form-bound epistemology
is coherent, but its ontology fails to fully account for the transcendent qualities of a buddha’s
natural body, whereas the Form-free ontology is clear, but its epistemology fails to account for
the immanence of a buddha’s bodies of enjoyment and emanation. Thus, as an intervention,
Ratnākaraśānti’s somewhat hybrid position, which agrees with and responds to Kamalaśīla’s
critiques, can be seen as accepting the basic Form-free ontology but making some limited con-
cessions to Form-bound epistemology so as to explain the mechanism through which a buddha
interacts with ordinary persons.
589
Gregory Max Seton
Ratnākaraśānti’s philosophical intervention, contra Sthiramati, explains the grasped object

and a grasping subject as two types of representational forms that appear in consciousness and
argues that a buddha’s reflexive awareness is ultimately free of these. But Ratnākaraśānti also
asserts that – by deliberately retaining a small amount of cognitive error – a buddha’s imma-
nent awareness perceives the two representational forms, which she reflexively knows to be
unreal and false, so that she can keep an epistemic connection to ordinary persons. He states:
After the representational forms dissolve into a transcendent (lokottara) awareness,

the transcendent awareness itself arises as an all-pervasive buddha free of appear-
ances, free of the two representational forms, and free of conceptual proliferations.
Even though her pure immanent (laukika) awareness has representational forms,
those forms are discerned to be false and unreal, and, in this sense, that awareness is
said to be free of representational forms.
(Guide, 226a7)
In Ratnākaraśānti’s position, when a buddha realizes the ontological reality of the dream, she
does not exit to some awakened world beyond the dream where she knows the mere exist-
ence of things directly. Rather, her transcendent awareness is inseparable from an immanent
awareness that compassionately manifests within the dream world, like a lucid dreamer who
interacts with the dream-forms of ordinary persons in the shared dream-space of conventional
reality. Thus, for Ratnākaraśānti, since a buddha chooses to participate in the dream and knows
the dream-forms to be false and unreal, she can still be understood to be ultimately “free.”
Although Ratnākaraśānti himself characterizes his unique viewpoint as a sub-position
within the Form-free (nirākāra) school – where he incidentally also places Kamalaśīla’s
Mādhyamika position – Ratnakīrti (eleventh century) and others classify Ratnākaraśānti’s posi-
tion as “False-form” (alīkākāra) and differentiate it from the paradigmatic Form-free position
associated with Sthiramati and others (Guide, 226b2; Nondual Variegated, 122). Since many
modern and traditional scholars incorrectly conflate these two distinct positions and use the
names Form-free and False-form interchangeably, it must be emphasized here that this chapter
follows Ratnakīrti in using the terms to refer to very different viewpoints (Kajiyama 1965,
426–25).6 Now that these two names have been distinguished, the following two main sections
will focus on the two main pillars of his onto-epistemological argument – namely his theory
of the fundamental nature of consciousness and his implied theory of error – which aimed to
bridge some of the gaps between the rival ontologies of Yogācāra monists and Mādhyamika
skeptics and between the two divergent representationalist epistemologies.
Ratnākaraśānti’s Ontological Account of Consciousness

When arguing that all phenomena are nothing but consciousness, Yogācāra scholars often
compare consciousness to a lamp that illuminates itself while illuminating other things. For
Form-bound scholars, the two functions of consciousness – namely its “illumination” that
allows phenomena to appear and its “awareness” that knows them – are like two equal sides
of the same coin. Hence, the analogy of a lamp also means that awareness knows itself reflex-
ively as it knows other things. But Ratnākaraśānti’s ontological argument treats illumination
( prakāśa) as a nature of consciousness more fundamental than even its awareness, because –
insofar as a lamp illuminates itself while illuminating other things – illumination is the very
condition of reflexivity that makes awareness of anything possible (Instructions,145b3). In
other words, Ratnākaraśānti presents illumination as the substance (dravya) of consciousness
590
Ratnākaraśānti
and awareness as its natural function (Instructions,143a5, 145b3, 150a2). By grounding mind’s
knowing quality within its illuminating quality, Ratnākaraśānti suggests that the lamp-like
reflexivity of illumination underlies both an ordinary person’s intentional consciousness and
a buddha’s simultaneously transcendent and immanent awareness. In this way, he sets up his
explanations both of how illumination remains unchanged when an ordinary person becomes
awakened and of how buddhas can still communicate with ordinary beings.
To explain his ontology, Ratnākaraśānti establishes what it means that all phenomena are
only consciousness in four main steps. In the first step, Ratnākaraśānti argues that phenomena
are ontologically grounded in awareness due to an identity (tādātmya) relation between what
we might call illumination qua phenomena and illumination qua awareness. Second, he argues
that if awareness grounds phenomena and if awareness is aware of phenomena, then aware-
ness is ultimately reflexively aware of itself. Third, he argues that the identity relation between
real illumination and unreal representational forms is not contradictory because the former is
the “bare” nature, while the latter is something “superimposed” upon that. Finally, he explains
how illumination, as an ontological substance, manifests, due to cognitive error, either in the
three types of intentional consciousness of ordinary persons or in the three bodies of a buddha.
Each of these five points will be expanded upon in the four following subsections.
Phenomena Are Ontologically Grounded in Awareness

Ratnākaraśānti’s basic ontological argument is that illumination is the only real substance,
but it can take on unreal, adventitious representational forms due to cognitive error. To begin
with, he proves that all phenomena have nothing but consciousness as their inherent nature,
and since this means they lack the nature of a grasping subject or grasped object, they them-
selves do not have their own inherent nature (Instructions, 145a6). The upshot of his argument
is that phenomena may appear to be different from awareness, but through direct perception
( pratyakṣa) and inference (anumāna), phenomena can be proven to be simply representational
forms made out of the substance of illumination whose nature is clear and indiscernible from
awareness. For Ratnākaraśānti, the fact that phenomena are just illumination can be established
first through direct perception because we directly experience illumination in the phenomena
shining forth in our minds. That is to say, just like a film spectator is aware of the light within
film images, an ordinary person is always subconsciously aware that phenomena are nothing
but illumination in various representational forms because they are shining forth, manifest-
ing, or appearing in awareness and are neither “dim” ( jaḍa) nor “inaccessible” ( parokṣa) to it
(Instructions, 145a6). For this reason, when an ordinary person says “I see a blue form,” it is
identical to saying “I see the illumination of a blue form.” Here, Ratnākaraśānti explains the
blueness of the form as appearing due to mind’s latent impressions; he explains illumination’s
nature as being clear – that is, lacking its own color, shape, form, and so on – because, if it
were opaque, nothing would be able to appear in it and no other phenomena could ever be
established through it. For Ratnākaraśānti, on the one hand, this clear illumination cannot be
known unless it is the substance of awareness itself. On the other hand, an intentional object
is by definition accessible and connected to awareness only through illumination that is clear,
like a nonmaterial mirror that illuminates objects. That is to say, any intentional object that
appears in that mirrorlike clear nature, such as a blue color patch and so on, must share a con-
nection (sambandha) of identity (tādātmya) with that illumination (Instructions, 153a4).7 So,
if one were to say, “I see the illumination of a blue thing,” this would be the same as saying
“Illumination qua awareness is taking on the representational form of grasping the illumina-
tion in the blue representational form grasped.” In other words, the representational form of
591
Gregory Max Seton
a grasped blue object and the representational form of a grasping subject are both made out
of illumination. In this way, Ratnākaraśānti argues, on the basis of direct perception and infe
rence, that all phenomena have awareness as their nature because illumination is established
as the substance out of which both phenomena and awareness are made.
Awareness Is Ultimately Reflexively Aware of Itself

Ratnākaraśānti next defends the implication that if awareness is the ontological ground of
phenomena and if awareness is aware of phenomena, then awareness must be ultimately only
reflexively aware of itself. To this end, he puts forth a two-part proof that makes use of the
ambiguous and multivalent Sanskrit instrumental case to insinuate simultaneous concomi-
tance, causality, and identity. First, Ratnākaraśānti argues that, if an intentional object cannot
be known except through (vinā) illumination, then that intentional object is known as or by
that illumination (Instructions, D153a4). His point here seems to be that, for instance, a blind
dog who knows a woman only through her scent knows her as or by her scent. Hence, there is
no contradiction in saying that one can be reflexively aware of intentional objects as illumina-
tion, because it is the intentional object’s illumination alone that is known through illumination
qua awareness. Second, he argues that, if an intentional object is something to be known by or
as illumination, then that illumination must be awareness (Instructions, D153a5). Since only
awareness can be known through reflexive awareness (svasaṃvedya), the awareness of illu-
mination qua phenomena cannot be anything other than an identity connection with it. Thus,
he concludes that the nature of pure consciousness is proven to be the reflexive awareness of
illumination, and hence, the false and unreal representational forms of objects grasped by con-
sciousness are also refuted as being external to the mind (Instructions, D153a6).
The “Bare” and “Superimposed” Natures Are Not Contradictory

Next, Ratnākaraśānti defends his argument against a possible objection – that is, that an iden-
tity connection between a real thing and unreal thing is self-contradicting – by distinguishing
the representational forms as a “superimposed” (āropita) false nature and their illumination
qua awareness as their “bare” (anāropita, lit. “unsuperimposed”) true nature. According to
Ratnākaraśānti, it is meaningful to speak of an identity between the unreality of the representa-
tional forms and the reality of their illumination, because “unreal” and “real” – insofar as they
are based on (niṣṭha) different natures that are, respectively, superimposed and bare – can be
considered contradictory types of properties that refer to a single locus (vastu) without negat-
ing one another (Instructions, 149a3 – a6; Textual Commentary ad 3623–24). To explain this
special type of identity connection, we may compare it to Gibbard’s (1975, 187) notion of so-
called “contingent identity” between an indestructible form-free lump of clay and a destroy-
able clay statue made from it.8 Gibbard’s idea seems to be that even though a form-free lump
of clay and the statues or pots fashioned out of it can be considered identical, their identity is
contingent, in the sense that the statues or pots can be destroyed by the hands, but the lump of
clay – which is “form-free” in the sense of being substance without a particular form – is not
destroyed by that action. Likewise, Ratnākaraśānti’s notion of sheer illumination being “bare”
is comparable to the indestructible lump of clay, and his notion of representational forms being
“superimposed” is comparable to the destroyable clay statues or pots fashioned out of it. In
this way, Ratnākaraśānti asserts a special type of identity relationship between reflexively
aware illumination and its false representational forms. By holding this real substance to be
a momentary material cause in line with mainstream Buddhist doctrine, Ratnākaraśānti shifts
592
Ratnākaraśānti
the focus away from the function of awareness toward the substance of illumination in order
to avoid the prior criticisms of Form-free scholars who refute any possible causal relationship
between a real awareness and any unreal intentional objects. Furthermore, Ratnākaraśānti’s
argument sets up his subsequent assertion that – by retaining a small amount of cognitive
error – a buddha’s nondual illumination is an immaterial substance that has the moment-by-
moment power (śakti) to be aware reflexively either of its own transcendent inherent nature
alone, for example, when it is like a lump of clay, or of its immanent inherent nature in the
false representational forms themselves, for example, when it is like a clay statue.
Illumination’s Reflexivity as Intentional Consciousness

Next Ratnākaraśānti connects the perception of buddhas to ordinary persons epistemically. To
do so, he explains illumination’s inherent reflexivity as having two nondual aspects – namely a
cognizing ( jñāna) and cognized aspect ( jñeya) – that appear, due to a fundamental cognitive
error (bhrānti), either as the quasi-intentional immanent awareness of a buddha or as the inten-
tional consciousness of an ordinary person. In the case of a buddha, the quasi-intentional imma-
nent awareness arises due to cognitive error as a cognizing aspect that recognizes the inherent
nature (svarūpa) of its cognized aspect as being identical to its own inherent nature (Excellent,
15; Seton 2016, 96). But in the case of an ordinary person, since intentional consciousness has
conceptual proliferations in addition to the fundamental cognitive error, it does not recognize the
identical nature of these two nondual aspects and instead superimposes upon them the imagined
characteristics of a grasped object and grasping subject, even though these are unreal (asat). That
is to say, in the ordinary person’s imagination, the cognized and cognizing aspects appear to be
discrete representational forms of a grasped object and grasping subject despite their not being
real discrete substances (avastu) (Excellent, 15; Seton 2016, 93–95). In this way Ratnākaraśānti
demonstrates that an ordinary person’s consciousness is not perceiving the bare inherent nature
of the cognized or cognizing aspects of his or her reflexively aware illumination, simply because
he or she is fixated upon the unreal content in the non-discrete representational forms superim-
posed upon the cognized and cognizing aspects due to their conceptual proliferations. That said,
to compare how illumination manifests in an ordinary person’s intentional consciousness and
in the buddha’s quasi-intentional immanent awareness subsequently, it will be helpful first to
describe the way that illumination is the substance of the three streams of an ordinary person’s
intentional consciousness – that is, the “container consciousness” (ālayavijñāna), the “afflicted
ego mind” (kliṣṭamanas), and the “actively aware consciousness” ( pravṛttivijñāna) – which
arise with their own false and unreal representational forms of a grasper and grasped moment
to moment.
To begin with, illumination is the substance of the unselfconscious “container conscious-
ness,” which is like the sleeping mind’s consciousness in which and out of which a dream world
appears (Instructions, 139a2). This stream of consciousness is called a “container” because
its primary function is to contain the latent impressions of past actions (karma) that shape our
character, dispositions, and tendencies and to project them onto our mindscreen, like the reels
holding the film that is projected frame by frame in the mental movie of our dream.9 Within this
nonconscious awareness, illumination’s natural reflexivity – with its cognizing and cognized
aspects – manifest as the animate beings and inanimate world within a dream (Instructions,
139a4–5; Five Aggregates, 91–92). This container consciousness is intentional because it oper-
ates passively in the background of all experience – even when the other two streams of con-
sciousness cease temporarily during deep sleep, fainting, meditative cessation, and so on – as a
subconscious awareness that registers moment by moment what occurs in the world, mind, and
593
Gregory Max Seton
body (Instructions, 141a1; Five Aggregates, 93).10 It also serves as the ground for the continu-
ity of consciousness when mind alternates between its own unconsciousness and the conscious
mental states – the actively aware consciousness and afflicted ego mind – which arise out of it.
Second, illumination is the substance of the “afflicted ego mind,” a consciousness that arises
from the container consciousness like a wave arising from the ocean. The ego-mind manifests
as a self that is “afflicted” because it is painfully engrossed in self-delusion (ātmamoha),
self-imaging (ātmadṛṣṭi), self-attachment (ātmasneha), and self-conceit (ātmamāna). The
ego-mind is the subconscious sense of a subject grasping the container consciousness as an
unseen internal personal identity with particular characteristics gleaned from the various rep-
resentational forms (Five Aggregates, 108; Instructions, 139a1), like emotionally identifying
with the experiencer of the dream images as a “me” with certain characteristics, such as body
size, personal history, and so on. In this way, the ego-mind influences the actively aware con-
sciousness to become engrossed in particular aspects of the dream as though they were real
and related to “me” (Instructions, 139a5).
Third, illumination is the substance of six types of “actively aware consciousness”
( pravṛttivijñāna), namely the mental consciousness and the five types of sense consciousness.
Like further waves, the mental consciousness arises from the container consciousness moment
by moment, and the five types of sense consciousness arise in sequence or simultaneously
based on conditions (Instructions, 141a1–3). Each of the five sense perceptions is a grasping
of different grasped intentional objects produced by latent impressions emerging from the
container consciousness, like perceiving various images in a dream (Five Aggregates, 92;
Instructions, 138b7). Since the mental consciousness is purely conceptual, it is a conceptual
grasping of its intentional objects (artha) as possessing the characteristics of cowness or blue-
ness common to all particular manifestations (vyakti) of cows or blue color patches. Hence,
its grasped perceptual content “cow” or “blue” appears in representational forms that are false
or nonveridical due to being connected with universalizing linguistic structures that cogni-
tively exclude non-blue and non-cow.11 In modern neuroscientific terms, we might say that our
mental images of “cow” or “blue” are purely our human brain’s predictive interpretations of
light and that they do not correspond to any blue color or cow that exists in external reality.
Thus, the illumination that appears as the representational forms is real, but their presenta-
tion ( prakhyānam) as a “cow” or as “blue” is merely a mentally constructed determination
(adhyavasāya) of unreal intentional objects (Instructions, 149a3-a6), like determinations of a
person’s character in a dream.
Now, in order to realize the bare nature of illumination that forms the substance of the
superimposed streams of consciousness, reflexive awareness must be established by valid cog-
nition in a nondual way – that is, free of the subject-object structuring of ordinary intentional
consciousness – through the following four stages of phenomenological and ontological self-
reflection practice ( yogabhūmi).12 In the first stage, one must reduce all appearances to a fixed
set of categories consisting of what is grasping and grasped. In the second stage, one must
examine the grasping subject and the grasped object as a purely mental experience. In the
third stage, one must identify illumination’s emptiness of forms as the single common charac-
teristic (sanimittā) of these forms and then reflect upon that emptiness until one discovers the
transcendent awareness that is free of subject-object structuring. Although this transcendent
awareness is free of intentional content (nirālambana), it is not free of all content, insofar as
it is reflexively aware of its own suchness being empty of forms (Instructions, 149b7).13 This
discovery at the end of the third stage marks the point where an ordinary person first enters
the so-called path of seeing (darśanamārga). Here, if she did not previously do accompanying
594
Ratnākaraśānti
Mahāyāna practices to cultivate unconditional compassion, she would become a first-phase

“arhat” – that is, someone who permanently remains in a discrete body of sheer illumination
never to be reborn in further physical bodies. As such, she remains unable to benefit others in
the way that a buddha can. If she did previously do accompanying Mahāyāna practices, she
would at this time become a first level (bhūmi) “noble bodhisattva.”
In the fourth stage, one must practice resting in reflexive awareness of illumination’s
emptiness qua freedom from any identifiable characteristics (nirnimittā) (Instructions,
149b7) on the so-called path of cultivation, where a noble bodhisattva alternates between
her transcendent awareness – which involves resting in nondual reflexive awareness with-
out representational forms during meditative states – and its after-effect, a pure immanent
awareness (śuddhalaukikajñāna) during post-meditative states in order to perfect her ability
to rest in the nondual reflexive awareness even while interacting with the world (Instructions,
159a4). According to Ratnākaraśānti, the post-meditative awareness itself is called “imma-
nent” because it has strayed (’khrul ba nyid kyis; bhrāntatvena) from sheer illumination into
the world of illusory appearances, and yet “pure” because it is able to discern the suchness
(tathatā) of these illusions (Guide, 226a7). Finally, at the end of the fourth stage, one wakes
up as a buddha with a nondual reflexive awareness in which the distinction between the medi-
tative “transcendent” awareness and the post-meditative “immanent” awareness disappears.
According to Ratnākaraśānti, once a buddha awakens, the very basis of the three streams
of intentional consciousness is permanently transformed (āśrayaparāvṛtti), and the bare
cognizing aspect of reflexive awareness simply recognizes the bare cognized aspect of its
own illumination that was always present in the superimposed representational forms of a
grasping subject and grasped subject. In the Form-free Yogācāra notion of this awakening,
a buddha’s reflexive awareness is called nondual because its cognizing and cognized aspects
never again take on representational forms. That is to say, the transformed consciousness
will never again be intentional, because in a miraculous way it penetrates all phenomena
(sarvadharmavibhutvalābhatas) of ordinary persons without a subject-object structure (Thirty
Verses Commentary ad 30). However, in Ratnākaraśānti’s False-form position, such a non-
dual reflexive awareness free of representational forms can only result in the state of an arhat
(Sky-like, 231). For Ratnākaraśānti, what distinguishes a buddha’s nondual reflexive aware-
ness from that of an arhat is its ability to retain cognitive error – which provides a mechanism
for interacting with the world of ordinary persons through representational forms – without
being subject to its negative effects, like a lucid dreamer who is not at all subject to the dream.
In order to distinguish the buddha’s nondual reflexive awareness from a bodhisattva’s pure
immanent awareness, Ratnākaraśānti qualifies the amount of error retained as being “small” to
suggest that it produces representational forms so translucent that their empty nature is effort-
lessly obvious to a buddha at all times – unlike to bodhisattvas who, in effect, must return to
meditative states in order to recharge their awareness of illumination.
Hence, a buddha’s nondual awareness is completely “nondual” in two senses. First, her
reflexive awareness is free of the ignorant subject-object structuring of an ordinary person’s
intentional consciousness. Second, she is able to simultaneously and completely recognize her
own inherent nature in those representational forms, as if her dream images were so translucent
she could not avoid seeing their illumination. Thus, in Ratnākaraśānti’s False-form position, a
buddha’s nondual reflexive awareness is not just an ordinary person’s inferential knowledge of
the reflexivity of awareness, not just an arhat’s direct perception of sheer illumination free of
forms after physical death, not just a bodhisattva’s transcendent awareness of sheer illumina-
tion during meditation with its aftereffect of immanent awareness during post-meditation, but
595
Gregory Max Seton
rather a transcendent and immanent, quasi-intentional consciousness of representational forms

that never loses sight of sheer illumination being the only real thing.
Ratnākaraśānti’s Implied Theory of Error

Ratnākaraśānti speaks often of error without ever providing a systematic theory of error.
Hence, this section will have to tease out his unique perspective from his various assertions
about the three bodies of a buddha and how they function in the world. But before delving into
these assertions, we must first address the fact that we have been translating the term bhrānti
as “cognitive error,” but the term has a wider range of meaning in Ratnākaraśānti’s writings.
First, at the level of ordinary persons, Ratnākaraśānti uses the term bhrānti to refer not only to
a “cognitive error” – which causes ordinary persons to perceive a dualistic illusion of repre-
sentatational forms whose nature is real illumination – but also to a “conceptual error,” which
is the conceptual proliferations about a false and unreal grasping subject and grasped object.14
Second, unlike any other prior Buddhist scholar, Ratnākaraśānti claims that a buddha, out of
compassion, retains a small amount of bhrānti so that she can engage with representational
forms and still accurately know the content of those illusory forms to be false and unreal. By
specifying only a “small amount,” Ratnākaraśānti implies that a buddha retains only the most
basic cognitive error that gives rise to representational forms but not the conceptual error
tied to ignorance that leads to suffering. Thus, beyond the “cognitive” versus “conceptual”
distinction, Ratnākaraśānti’s notion of bhrānti might not best be translated as “error” because
the buddha is not deceived by retaining it, and her own reflexive awareness remains non-
erroneous and “free” of representational forms. That is to say, in Ratnākaraśānti’s system, the
word bhrānti – a noun formed from the Sanskrit verbal root √bhram, whose range of meanings
includes “to wander,” “to err,” or “to spread” – might be better translated as a more neutral
“spreading” or “diffusion” of illumination which can be erroneous or not.
Whatever English translation is generally adopted, though, it should be emphasized that
Ratnākaraśānti’s idea of a buddha’s bhrānti is that it enables her transcendent awareness to
appear in an intersubjective world of representational forms as an accurate, non-erroneous,
immanent awareness – rather than causing her somehow to be lost in ignorance and fallible.
In this way, unlike the Form-free notion of bhrānti as an all-or-nothing, purely accidental error
that is eliminated upon awakening as a buddha, Ratnākaraśānti treats bhrānti more like a neutral
efficient cause that makes the substance of sheer illumination – which exists moment to moment
in a stream ( pravṛtti) – appear as representational forms, like a crystal through which a ray of
clear illumination is diffused in order to produce visible colored forms that others can perceive.
That said, in Ratnākaraśānti’s system, when sheer illumination is diffused by bhrānti, it is
not just perceived by others, but rather it illuminates both itself and others in a way that it can
consciously interact with other beings in an intersubjective world. In Ratnākāraśānti’s non-
solipsistic universe, distinct ordinary persons are lost in their own bubbles of self-illuminating
awareness distorted by their cognitive and conceptual error into various representational forms
of grasping subjects and grasped objects. But persons in different groups share common illu-
sions that they grasp differently due to their different conceptual proliferations, like different
audience members in the same theater grasping different aspects of the same unreal movie
images on screen. So, the difference between a buddha’s and an ordinary person’s represen-
tational forms is that a buddha has no conceptual proliferations and recognizes the cognized
and cognizing aspects of diffused illumination to be her own projections, whereas an ordinary
person has conceptual proliferations and misperceives the cognized and cognizing aspects as
an imaginary grasped and grasper.
596
Ratnākaraśānti
As for how a buddha can function in a false and unreal world in order to awaken other
beings, Ratnākaraśānti explains the three bodies of a buddha as being both real and illusory
forms of illumination free of conceptual proliferations. To begin with, a buddha’s natural body
consists in real illumination reflexively aware of its own emptiness. However, as a natural out-
come (niṣyandaphala) of the natural body being the pure source of qualities (dharmadhātu),
it automatically diffuses itself through the prism of bhrānti to manifest a light body of enjoy-
ment, which is its own mental construct (vikalpa). This diffusion of illumination is reflexively
aware of itself but also able to be perceived directly in various forms by beings who are almost
awake (bodhisattva) and indirectly, as an emanation body, by ordinary persons, who are still
asleep. Hence, even though Ratnākaraśānti does not explain what representational forms are
like for a buddha, his notion of bhrānti enabling her both to see the illusory world and to be
seen in it might explain why Ratnākaraśānti, unlike Sthiramati, holds both the grasper and
grasped to be representational forms consisting of illumination. Whatever the case, though,
he makes clear that an awakened person’s natural body does not just embody its own free-
dom (vimuktikakāya) – like that of an arhat – but embodies and shines forth the qualities
(dharmakāya) that can participate in different illusory dimensions, like an awakened person
entering into a lucid dream in order to project herself into others’ dream worlds.
In this regard, Ratnākaraśānti also makes clear that the three bodies ultimately consist in
the same real illumination but he treats the transcendent aspect of a buddha as the natural
body – that is, a sunlike orb of reflexively aware illumination – and her immanent aspect as
the sunray-like enjoyment body shining forth illusory representational forms to communicate
with beings. High-level bodhisattvas, close to awakening themselves, directly perceive and
enjoy the real illumination of that illusory enjoyment body in translucent representational
forms, like semi-lucid dreamers with sunlight on their faces noticing the brightening of all
their dream images.15 But those ordinary persons, who are light sleepers, perceive that same
illumination indirectly as opaque representational forms of physical and verbal content, like
ordinary dreamers imagining the sunlight on their faces to be bright dream lightbulbs shining
in their dream eyes (Guide, 228b3). In both cases, it is bhrānti that allows a buddha to shine
forth her own illusory bodies with a cognized and cognizing aspect as representational forms
that appear to others like a grasping subject aware of what is grasped. In this way, a buddha’s
bhrānti is instrumental in allowing the substance of a buddha’s reflexively aware illumination
to take on the form of a real enjoyment body at the conventional level of bodhisattvas and
to appear as various emanation bodies in the conventional dream world of ordinary persons,
without a buddha actually entering the dreams herself.
Therefore, in Ratnākaraśānti’s implied theory of error, bhrānti is more or less a diffusion
that causes the cognizing and cognized aspects of sheer illumination to appear distinct, like
visual floaters seeming to be objects outside an eye, whereas the conceptual proliferations cre-
ate erroneous content. By retaining only a small amount of bhrānti, a buddha’s illumination is
diffused just enough to produce representational forms but not enough to produce erroneous
conceptual proliferations that would cloud the cognizing and cognized aspects of a buddha’s
reflexive awareness. In this way, bhrānti is central to explaining the transcendent and immanent
aspects of a buddha’s natural body that has the power (śakti) to communicate accurately with
ordinary beings without being subject to the same ignorance and suffering that they experience.
Conclusion
For Ratnākaraśānti, the Form-bound ontology cannot explain the transcendent qualities of a
buddha’s natural body; whereas the Form-free epistemology cannot explain the immanence of
597
Gregory Max Seton
a buddha’s bodies of enjoyment and emanation. For this reason, he argues for a position that can
be seen as agreeing with Form-free ontology but making some limited concessions to Form-
bound epistemology with regard to a buddha’s interaction with ordinary persons. To counter
those who follow Kamalaśīla’s idea of a true dream but refute illumination, Ratnākaraśānti
asserts that a buddha must realize illumination – that is, the substance of intentional conscious-
ness – through reflexively aware valid cognition but, out of compassion for others, retain a
small amount of bhrānti in order to produce physical bodies able to communicate with ordi-
nary persons who perceive only false and unreal forms, and to teach them the path to realize
the nature of their consciousness as illumination.16 By arguing that seemingly contradictory
properties – such as real and illusory, superimposed and bare, and so on – can in fact coincide
within a single locus, he establishes a type of contingent identity relation between illumina-
tion and forms. Based on this type of identity, we are meant to understand why these physical
bodies of buddhas can have both real and illusory properties and how their nondual reflexive
awareness can be intentionally conscious of representational forms while reflexively aware
of their illumination. In effect, Ratnākaraśānti’s ontological assertion of sheer illumination
ultimately free of representational forms (nirākāra) does not indicate that a buddha does not
have any forms, but simply that a buddha’s awareness has the ability to be “free” of forms.
Furthermore, a buddha engages in cognitive diffusion that is free of the conceptual error repre-
sented by conceptual proliferations. Hence, although Ratnākaraśānti classifies his own view-
point simply as “Form-free” and although many traditional and modern scholars conflate the
names False-form (rnam brdzun) and Form-Free (rnam med) and use them interchangeably,
we can call his position “False-form insofar as it asserts the illumination of false forms.” Thus,
in this way, Ratnākaraśānti’s assertions truly distinguish his position from all other scholars
within the history of Mahāyāna Buddhist philosophy.
Notes
1 As described in the Guide’s colophon, ostensibly written by Śāntibhadra, likely a student of his.
2 See McClintock’s (2013, 4) discussion of Kamalaśīla’s agreement with the “third” position in Textual
Commentary 3626–27.
3 Here, the term “bound” does not signify this position’s self-description or any particular terminology
but rather a metaphorical interpretation of the word sa- (“with”) in sākāra (viz. “with representational
form”). Different sākāra positions explain consciousness as bound in different ways, such as indivis-
ibly (citrādvaita), and so on. See Textual Commentary, 3627, where a buddha’s momentary aware-
ness is described in several different ways, for example, as arising in conjunction (upagraheṇotpatti)
with representational forms.
4 Here and subsequently, where my conclusions are based upon my own reading of Ratnākaraśānti’s
works extant in Sanskrit or Tibetan, which have not been published in English translation, I will cite
an abbreviation of a translated English title, but the subsequent folio numbers refer to the Sanskrit or
Tibetan editions, while “a” or “b” refers to front or back sides of the folios.
5 I have interpreted the phrases sattāmātreṇa and sattām anubhavati differently from Funayama 2007,
194 n. 35.
6 Kajiyama (1989) explicitly conflates the two names nirākāra (“Form-free”; Tib. rnam med) and
*alīkākāra (“False-form”; rnam brdzun).” For more on classifying the position, see Seton 2016,
chapters 2–4 and Luo 2018.
7 It is widely accepted among Buddhist scholars (e.g., Textual Commentary ad 3459) that there are
only two types of connection between things, either an identity connection (tādātmya), meaning they
share the same nature, or a causal connection (tadutpatti), meaning one arises from the other. Other
epistemological idealists hold that awareness and its intentional object share a causal connection.
8 Tomlinson (2018, 4) discusses the Gibbard’s clay example in a slightly different way.
9 In modern terms, one might say that the unconscious processes during resting and digesting all rely
on the identification of some sort of intentional objects, but active awareness is just unaware of them.
598
Ratnākaraśānti
10 Here, meditative cessation is a state into which meditators enter into with an intention to return to
consciousness after a given period of time; that is, the basic continuity of consciousness is not lost
during unconsciousness.
11 There are many Buddhist theories of exclusion. The basic idea is that the predictive mechanism of the
mind pragmatically develops the abstract universal “cow” together with learned linguistic conven-
tions – that is, the word “cow” – and it does so not by establishing all the common characteristics of
“cow-ness” but by excluding the main characteristics of a non-cow. Based on this abstract universal,
a conceptual cognition can then determine those unique characteristics (which are not non-cow char-
acteristics) as belonging to particular instances of cows.
12 For a brief comparison of phenomenology in Buddhist and non-Buddhist contexts, see Tomlinson
2019, 20–26.
13 See, for example, Pearl Garland 97.
14 Ratnākaraśānti’s virtual silence on conventional “perceptual errors” – like the optical illusions of a
rotating firebrand or cataracts disease – shows his general disinterest in epistemological explanations
of how to act accurately in the world.
15 For Ratnākaraśānti, mental constructs exist qua substance. Cf. Guide, 228b3 or Pearl-Garland, i.37–38.
16 For more on the true dream, see McClintock 2014, 10.
Bibliography
Chan, Ngan Che. 2007. “A Study of Yogācāra Theory of the Ten Causes.” PhD diss., University of
Hong Kong.
Chimpa, Lama, and Alaka Chattopadhyaya. 1990. Taranatha’s History of Buddhism in India. Edited by
Debiprasad Chattopadhyaya. Reprint edition. New Delhi: Motilal Banarsidass.
Dowman, Keith. 1986. Masters of Mahamudra: Songs and Histories of the Eighty-Four Buddhist Sid-
dhas. Albany: State University of New York Press.
Excellent Commentary = Rinpoche, Samdhong and Vrajavallabh Dwivedi, eds. 1992. Mahāmāyātantra
with Guṇavatī by Ratnākaraśānti. Rare Buddhist Text Series 10. Project edition. Sarnath, Varanasi:
Central Institute of Higher Tibetan Studies.
Five Aggregates Explanation = Kramer, Jowita, ed. 2013. Sthiramati’s Pañcaskandhakavibhāṣā:
Critical Edition (Part 1). Vienna: China Tibetology Publishing House, Austrian Academy of Sciences
Press.
Funayama, Toru. 2007. “Kamalaśīla’s Distinction Between the Two Sub-Schools of Yogācāra. A Provi-
sional Survey.” In Pramāṇakīrtih. Papers Dedicated to Ernst Steinkellner on the Occasion of His 70th
Birthday, Part 1, edited by Birgit Kellner, Horst Lasic, M. T. Much, and Helmut Tauscher, 187–202.
Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
Gibbard, Allan. 1975. “Contingent Identity.” Journal of Philosophic Logic 4 (2): 187–221.
Guide to Adornment of the Middle Way = Rin-chen-’byung-gnas-zhi-ba (Ratnākaraśānti). dBu ma rgyan
gyi man ngag (*Madhyamālaṃkāropadeśa, MAu). Translated by Śāntibhadra & Śākya-’od. Revised
by Amogha & ‘O ru. Tōh. 4085.
Instructions on Transcendent Wisdom = Ratnākaraśānti. Shes rab kyi pha rol tu phyin pa’i man ngag
(Prajñāpāramitopadeśa, PPu). Translated by Zhi-ba-bzang po (Śāntibhadra) and’Gos [Khug pa]-lhas-
btsas. Tōh. 4079.
Innate Sādhana = Isaacson, Harunaga. 2001. “Ratnākaraśānti’s Hevajrasahajasadyoga (Studies in
Ratnākaraśānti’s Tantric Works I).” In Le Parole E I Marmi: Studi in Onore di Raniero Gnoli Nel Suo
70 Compleanno, edited by Raffaele Torella, 457–87. Roma: Istituto per L’Africa e L’Oriente.
Isaacson, Harunaga, and Francesco Sferra, eds. 2014. The Sekanirdeśa of Maitreyanātha (Advayavajra)
with the Sekanirde-śapañjikā of Rāmapāla: Critical Edition of the Sanskrit and Tibetan Texts with
English Translation and Reproductions of the MSS. Manuscripta Buddhica 2. Napoli: Università degli
Studi di Napoli “L’Orientale.”
Kajiyama, Yūichi. 1989. Studies in Buddhist Philosophy: Selected Papers. Edited by Katsumi Mimaki
et al. Kyoto: Rinsen Book Co., LTD. [Reprint of Kajiyama, Yūichi. 1965. “Controversy between the
sākāra- and nirākāra-vādins of the yogācāra school – some materials.” Journal of Indian and Buddhist
Studies 14 (1): 418–29].
Kano, Kazuo. 2016. Buddha-Nature and Emptiness: rNgog Blo-ldan-shes-rab and A Transmission of
the Ratnagotravibhāga from India to Tibet. Vienna: Arbeitskreis für Tibetische und Buddhistische
Studien, Universität Wien.
599
Gregory Max Seton
Luo, Hong. 2018. “Is Ratnākaraśānti a gZhan stong pa?” Journal of Indian Philosophy 46: 577–619.
Mahāyāna Scriptural Adornment = Bagchi, S., ed. 1970. Mahāyānasūtrālaṅkāra of Asaṅga. Bud-
dhist Sanskrit Texts, no. 13. Darbhanga: Mithila Institute. Funahashi, Nagoya, ed. 1985.
Mahāyānasūtrālaṃkāra Chapter I, II, II, IX, X. Tokyo: Kokushokankokai.
Makransky, John J. 1998. Buddhahood Embodied: Sources of Controversy in India and Tibet. Albany:
McClintock, Sara. 2014. “Kamalaśīla on the Nature of Phenomenal Content (ākāra) in Cognition: A Close
Reading of TSP ad TS 3626 and Related Passages.” Journal of Indian Philosophy 42 (2/3): 327–37.
Middle-Beyond-Extremes Commentary = Bhattacharya, Vidhushekhara and Giuseppe Tucci, eds. 1932.
Madhyāntavibhāgasūtrabhāṣyaṭīkā of Sthiramati: Being a sub-commentary on Vasubandhu’s Bhāṣya
on the Madhyāntavibhāgasūtra of Maitreyanātha (Part I). London: Luzac & Co.
Nondual Variegated Illumination = Thakur, Anantalal, ed. 1957. “Citrādvaitaprakāśa.” In
Ratnakīrtinibandhāvalī. Patna: Kashiprasad Jayaswal Research Institute.
Pearl-garland Commentary = Tripāṭhī, Rāmaśaṅkara and Ṭhākurasena Negī, ed. 2001. Hevajratantram:
Ratnākaraśāntiviracita-Hevajrapañjikā-muktāvalī-saṃvalitam. Vārāṇasī: Kendrīya Uccạ Tibbatī
Śikṣāsaṃsthānam. Cf. Muktāvalī, I.viii (unpublished), ed. by Harunaga Isaacson.
Proving Mere Representation = Ratnākaraśānti. Vijñaptimātratāsiddhi. rNam par rig pa tsam nyid du
grub pa zhes bya ba. Tōh. 4259, D, tshad ma, zhe 306b4–309b3.
Roerich, George de, and Gendün Chöphel. 1988 (1949). The Blue Annals by Gö Lotsawa. Translated by
George de Roerich and Gendün Chöphel. New Delhi: Motilal Banarsidass. Reprint of Calcutta, Royal
Asiatic Society of Bengal, in two volumes.
Quintessential Commentary = Ācārya Ratnākaraśānti. 1979. Sāratamā: A Pañjikā on the Aṣṭasāhasrikā
Prajñāpāramitā Sūtra. Edited by Padmanabh S. Jaini. Calcutta: Kashi Prasad Jayaswal Research
Institute.
Schmithausen, Lambert. 2005. On the Problem of the External World in the Ch’eng wei shih lun. Studia
Philolgica Buddhica (Occasional Paper Series XII). Tokyo: International Institute for Buddhist Stud-
ies of the ICPBS.
Seton, Gregory Max. 2016. “Defining Wisdom: Ratnākaraśānti’s Sāratamā.” DPhil diss., University of
Oxford.
———. 2018. “Ratnākaraśānti.” In Brill’s Encyclopedia of Buddhism, Volume 2: Lives, edited by Jonathan
A. Silk. Leiden: Brill.
———. 2019. “Ratnākaraśānti.” The Treasury of Lives. https://treasuryoflives.org/biographies/view/
Ratnākaraśānti/P00EGS1016642.
Sky-like Commentary = Upādhyāya, Jagannāth, ed. 1983. “Khasama-nāma-ṭīkā.” In Śramaṇavidyā,
edited by Gokul Chandra Jain. Varanasi: Sampurnanand Sasnkrit Vishvavidyalaya.
Ten levels Scripture = Rahder, J., ed. 1926. Daśabhūmikasūtra et Bodhisattvabhūmi: Chapitres Vihāra et
Bhūmi. Paris: Paul Geuthner Louvain, J. B. Istas.
Textual Commentary on the Compendium of Reality = Krishnamacharya, Embar, ed. 1926. Tattvasaṃgraha
of Śāntarakṣita with the Commentary of Kamalaśīla. Gaekwad’s Oriental Series, no. 30–31. Baroda:
Central Library.
Thirty Verses Commentary = Buescher, Harmut, ed. 2007. Sthiramati’s Triṃśikāvijñaptibhāṣya. Critical
Editions of the Sanskrit Text and its Tibetan Translation. Vienna: Verlag der Östereichischen Akademie
der Wissenschaften.
Thurman, Robert. 2004. The Universal Vehicle Discourse Literature: Mahāyānasūtrālaṃkāra by
Maitreyanātha/Āryāsaṅga together with its Commentary (Bhāṣya) by Vasubandhu. Translated by L.
Jamspal, R. Clark, J. Wilson, L. Zwilling, M. Sweet, and R. Thurman. New York: American Institute
of Buddhist Studies at Columbia University.
Tomlinson, Davey. 2018. “Ratnākaraśānti’s Metaphysics of Difference: Limiting the Scope of the Neither-
One-Nor-Many Argument.” Delivered in the Buddhist Studies Unit at the Conference of the American
Academy Religion, Denver, November 17–20, 2018 (unpublished).
———. 2019. “Buddhahood and Philosophy of Mind: Ratnākaraśānti, Jñānaśrīmitra, and the Debate
over Mental Content (ākāra).” PhD diss., University of Chicago.
Twenty Verses = Silk, Jonathan A., ed. and trans. 2018. Materials Toward the Study of Vasubandhu’s
Viṃśikā (I): Sanskrit and Tibetan Critical Editions of the Verses and Autocommentary; An English
Translation and Annotations. Harvard Oriental Series, no. 81. Cambridge: Department of South Asian
Studies, Harvard University.
600
37
ATIŚA
The Great Middle Way of Mere Appearance
James B. Apple
Introduction
Atiśa, also known as Dīpaṃkaraśrījñāna (982–1054), is famous as a master from the ancient
Indian Buddhist land of Bengal and for his journeys in Indonesia and Nepal. In the last thirteen
years of his life, he became one of the most influential Indian Buddhist masters ever to set foot
in Tibet. His Mahāyāna Buddhist teachings, encompassing instructions on entering the way
of bodhisattvas up through highly advanced practices of the esoteric secret way of mantras,
came to influence all subsequent traditions of Buddhism in Tibet. Atiśa followed a lineage of
Indian Buddhist Madhyamaka (Middle Way) that was based on the teachings of Nāgārjuna
and Candrakīrti and influenced by his direct teachers, including Bodhibhadra (fl. c. 1000), a
scholar-monk at Nālandā; the contemplative-monk Vidyākokila (fl. c. 1000); and the tantric
yogi Avadhūtipa (fl. c. 1000). Atiśa’s Madhyamaka lineage was upheld and actively taught in
Tibet among communities of his Kadampa (bka’ gdams pa) followers until it was superseded
by forms of Madhyamaka infused with epistemology that developed at the famous Tibetan
monastery of Sangphu Neuthok (gsang phu ne’u thog) in the twelfth century (Apple 2019a).
Atiśa’s lineage of the Middle Way was contemplative in nature and did not utilize episte-
mological warrants ( pramāṇa) to realize ultimate reality. Still, although not advocating the
employment of epistemological warrants for Buddhist meditative realization, Atiśa did utilize
reasoning ( yukti) for the direct cognition of insight. He grounded his use of reasoning from
scriptural authoritative texts attributed to the Buddha that demonstrate a continuity in the use
of reasoning within Buddhist tradition. He denied a substance-based ontology while preserv-
ing a Buddhist soteriological program that accepts causality in the efficacy of conventional
practices of moral discipline and meditation. Atiśa offers the modern reader a rare glimpse
into an integrated Indian Buddhist Middle Way philosophy composed of a nominalism of
“mere appearances” (snang ba tsam) in which both objects and cognitions are dependently
designated and are therefore mere imputations ( prajñāptimatra). Atiśa advocates a mentalist
theory of Madhyamaka in which the mind, as mere appearance, is not at all established and is a
mere nominal designation free from the extremes of existence and nonexistence. In its overall
orientation, Atiśa’s Middle Way outlines a program of meditative cultivation that results in a
nondual awakening whereby all conceptual thought has been eliminated and not even noncon-
ceptual wisdom exists in Buddhahood.
DOI: 10.4324/9781351030908-52 601

James B. Apple
Historical Context
Atiśa lived at a unique juncture in the history of India and Tibet. He resided in India dur-
ing a revival of the East Indian Pāla Dynasties (760–1142) of Bihar and Bengal. During this
time, Bengal was an “international” region with trade routes on land from Assam and Burma
passing through ancient Buddhist pilgrimage holy sites of Magadha, including Vajrāsana, the
“Diamond Seat,” where Śākyamuni Buddha attained awakening, located in present-day Bodh
Gayā, India. The international environment of northeast India also included maritime routes
connecting the Bay of Bengal to harbors in South and Southeast Asia (Acri 2016). The Pālas
ruled over the northeastern lands of Bihar, West and North Bengal, and explicitly claimed to be
Buddhist in their inscriptions: they utilized the Dharma wheel of the Buddha’s teachings on the
top of inscribed copperplates and, in the colophons of manuscripts they sponsored, described
themselves as “entirely devoted to the Buddha” (Sanderson 2009). Furthermore, the Pālas
supported a perpetual endowment of donations for the Buddhist monastic community and
established a network of major monasteries, including the prestigious centers of Vikramaśīla
and Somapura. Atiśa, the Princely-Lord, came of age, studied, meditated, and taught during
this buoyant time of maritime Buddhist Asia.
The Pāla dynastic period was permeated with the religious practices of esoteric Buddhism,
or Secret Mantra practices. This form of Buddhism, based on groups of texts called tantras,
was primarily disseminated from master to disciple. Institutional monastic esoterism and the
esoterism of siddhas, accomplished adepts on the margins of society, dominated the Buddhist
culture of the time. By the late tenth and early eleventh century, during the lifetime of Atiśa,
Buddhist monastic communities and siddha culture had undergone almost two centuries of
blending and accommodation. The accommodation of esoteric Buddhism into monastic com-
munities was influenced, in part, by a highly competitive environment between Buddhists and
non-Buddhists seeking patronage and economic support. Esoteric Buddhists lived in a cultural
context, where Buddhists who did not support esoteric practices, as well as non-Buddhist
esoteric practitioners including the diverse groups of Śaivas (followers of Śiva), contested for
prestige and authority. These groups took part in a whole range of ascetic and ritual practices.
While residing in India before his journey to Tibet, Atiśa served as a monastic official at
monasteries in Bodh Gayā and Vikramaśīla. One way to understand the form and content of
the Mahāyāna Buddhism practiced and taught by Atiśa is to compare Atiśa with his esteemed
colleagues at Vikramaśīla. Most of Atiśa’s colleagues there followed various forms of Yogācāra
thought and worked with the philosophical traditions of epistemology based on the texts of the
seventh-century Buddhist thinker Dharmakīrti. His colleagues at Vikramaśīla were also holders
of diverse monastic ordination lineages: Atiśa’s colleague Jñānaśrīmitra, a prominent scholar who
initially upheld Saindhava śrāvaka views, advocated a form of Yogācāra thought that asserts that
cognitive images (ākāra) are real. Jñānaśrīmitra also formulated a sophisticated version of the
exclusion (apoha) theory of language based on complex epistemological arguments (McCrea and
Patil 2010). Atiśa’s senior colleague and teacher Ratnākaraśānti was ordained in the Sarvāstivāda
school at Odantapurī. Ratnākaraśānti composed numerous works on subjects such as valid cogni-
tion, the Perfection of Wisdom Sūtra (Prajñāpāramitā Sūtra), Yogācāra, and Tantra, as well as
Buddhist verse metrics (cansaḥśāstra) and riddles (Isaacson 2013). In terms of his viewpoint,
Ratnākaraśānti articulated a middle way based on Yogācāra principles that incorporated the the-
ory of the three natures (trisvabhāva) with an emphasis on self-awareness (svasaṃvedana) as
equivalent to luminosity ( prakāśa) (Apple 2018a). Atiśa’s junior colleague Ratnakīrti followed a
nondual consciousness (citrādvaitavāda) understanding of Yogācāra and articulated sophisticated
arguments against the Nyāya school’s God-like being called “Īśvara” (Patil 2009). In contrast
602
Atiśa
to all of his known Vikramaśīla colleagues, Atiśa followed a lineage of Madhyamaka based on
Nāgārjuna and Candrakīrti that was transmitted through his teachers Avadhūtipa, Bodhibhadra,
and Vidyākokila, with whom he studied in his youth. Such differences between Atiśa and his col-
leagues may have been a contributing factor in Atiśa’s decision to leave Vikramaśīla for Tibet.
At that time, the kings in West Tibet were seeking to rejuvenate Buddhism in order to rep-
licate the order, ethical principles, and stability that Mahāyāna Buddhist ideals had brought to
the Tibetan Empire during the seventh to ninth centuries. In seeking out Atiśa as the rejuvenator
of Mahāyāna Buddhist ideals, the Tibetans likely initially considered him a descendant of the
great bodhisattva-scholar Śāntarakṣita, who was also from Bengal. Śāntarakṣita was responsi-
ble for establishing the first ordained monks in Tibet and contributing to the translation of texts
and the construction of Tibet’s first monastery of Samyé (Blumenthal 2004). Although Atiśa
and Śāntarakṣita may have been from the same region, Atiśa lived in an era of Indian Buddhism
different than Śāntarakṣita’s: while Śāntarakṣita was a Mūlasarvāstivāda monk who blended
Yogācāra and Madhyamaka philosophies, Atiśa upheld the Mahāsāṃghika ordination lineage
and was a follower of the Madhyamaka tradition of Candrakīrti. Atiśa was also a lineage holder
of a number of Yoginī tantras, such as the Laghuśaṃvaratantra and Hevajratantra, which were
only introduced into Tibetan culture in the late tenth century by Rinchen Zangpo (958–1055).
Tibetans at this time were not fully familiar with these areas of monastic discipline, Madhy-
amaka philosophy, and esoteric Buddhism that Atiśa brought with him to Tibet.
Writings
Throughout his life in India, Indonesia, Nepal, and Tibet, Atiśa composed over 100 works. Recent
publication of his collected writings in Tibetan organize his compositions into those related to
view, conduct, union of view and conduct, and Secret Mantra practices. Atiśa focused his writings
and teachings on essential practices and integrative meditative cultivations leading to the nondual
realization of full buddhahood (Miyazaki 2007a). Such essential practices consist of teachings that
Atiśa distilled from the vast corpus of Buddhist scriptural and commentarial literature that was
preserved at the monasteries of Vikramaśīla, Nālandā, and Somapura. He composed several works
focusing on essential practice, including the Essential Summary (Garbhasaṃgraha), the Essence
Clearly Explained (Hṛdayanikṣepa), Lamp for the Summary of Conduct (Caryāsaṃgrahapradīpa),
and Bodhisattva’s Jewel Garland (Bodhisattvamaṇyāvalī). In these works, Atiśa emphasized culti-
vation of the awakening mind, moral conduct, and practices for generating merit.
The Open Basket of Jewels (Ratnakaraṇḍodghāṭamadhyamakopadeśa; Miyazaki 2007b;
Apple 2019a, 63–113) is one of the primary works among Atiśa’s collected writings and perhaps
the most extensive of his extant writings composed in India. Within this work, Atiśa cites sūtras
and tantras that provide one of the most thorough overviews on the awakening mind (bodhic-
itta) in Indian Buddhist literature. Atiśa’s Lamp for the Path to Awakening (Bodhipathapradīpa;
Apple 2019b) is his most well-known work. The text contains sixty-eight verses outlining the
integration of three forms of discipline, including the vows of the monastic disciplinary code
( prātimokṣa), bodhisattva precepts, and the precepts of the Secret Mantra Vehicle. Mahāyāna
and Vajrayāna practices and cultivations are also discussed. Written in response to questions
from King Jangchup Ö in West Tibet around 1042, Atiśa’s Lamp became “one of the most influ-
ential of Indian texts received by Tibetans” and was “the model for mainstream Tibetan monas-
tic Buddhists for the next nine hundred years” (Davidson 1995, 293). The Song with a Vision
for the Realm of Reality (Dharmadhātudarśanagīti; Blo bzang rdo rje rab gling 1999; Apple
2019b), which elucidates the dharmadhātu, or realm of reality, provides a summation of Bud-
dhist and non-Buddhist philosophical views on reality in a little over 140 verses. Atiśa discusses
603
James B. Apple
the realm of reality in the first twenty-five verses, fourteen of which are based on the work of
Nāgārjuna. Atiśa then summarizes Buddhist views (vv. 26–96), beginning with Madhyamaka
and descending to Vaibhāṣika tenets, follows this with non-Buddhist philosophical tenets (vv.
97–142), and then provides concluding verses (vv. 143–147) (Apple 2019b).
Atiśa’s Madhyamaka thought has traditionally been understood based on his Entry
to the Two Realities (Satyadvayāvatāra), Special Instructions on the Middle Way
(Madhyamakopadeśa), and Commentary on the Difficult Points in the Lamp for the Path to
Awakening (Bodhimārgapradīpapañjikā, D. 3948; Sherburne 2000). Entry to the Two Realities
and Special Instructions on the Middle Way are considered by traditional Gelukpa historians
the two foremost textual teachings (gzhung) on the view (lta ba) within Atiśa’s works (Apple
2019a, 270). Entry to the Two Realities succinctly lays out in twenty-eight verses a general
exposition on the two realities. Atiśa composed this work between 1012 and 1025 while resid-
ing in Sumatra and studying under Serlingpa. Serlingpa inquired about Atiśa’s philosophical
views in a letter, and Atiśa composed this set of verses as a response. Atiśa wrote Entry to
the Two Realities as an attempt to transform Serlingpa’s philosophical view from a Yogācāra
position to a Madhyamaka one. This introductory text on Madhyamaka presents Atiśa’s under-
standing based on the synthesis of a number of previous Indian Madhyamaka thinkers. Special
Instructions on the Middle Way is Atiśa’s advice for self-transformation through the practice
of Madhyamaka, given in Lhasa at the request of Ngok Lekpai Sherap. The brief text provides
instructions for how Mādhyamikas meditate, cultivating the three wisdoms of learning, reflec-
tion, and meditation within the context of meditative equipoise and postmeditative wisdom
construed through the purviews of conventional and ultimate reality.
A General Explanation of, and Framework for Understanding, the Two Realities (Bden
gnyis spyi bshad dang/bden gnyis’jog tshul; hereafter A General Explanation) is a late elev-
enth-century Indo-Tibetan Madhyamaka text that records oral teachings attributed to Atiśa on
the two realities (satyadvaya). The text furnishes an exposition of the Middle Way thought of
Nāgārjuna based on an exegesis of conventional reality and ultimate reality within the frame-
work of Mahāyāna path structures found in texts attributed to Maitreyanātha (Apple 2019a,
171–266). A General Explanation provides an early account of points of Madhyamaka exegesis
that would go on to become polemical points of debate in later decades and centuries in Tibetan
Buddhist thought. Atiśa discusses such points – whether Mādhyamikas have a thesis, an infer-
ence that is known to others (gzhan la grags pa’i rjes dpag), the object that is negated (dgag
bya), the negation of self-characteristics (rang gi mtshan nyid ≈ svalakṣaṇa), and the notion that
two things are “a single nature but different conceptual isolates” (ngo gcig ldog pa tha dad) –
in a late eleventh-century Indo-Tibetan Buddhist historical context.1 The text emphasizes that
correct conventional realities are indicated through nonimplicative negations (med dgag) and
that things are mere appearances that are transactually designated without being established.
Atiśa’s Middle Way Thought

Atiśa was an adherent to Madhyamaka thought and practice with a basis in the works of
Nāgārjuna. For Atiśa, Nāgārjuna was a towering figure of Buddhist culture who not only had
great insight and realization but also sparked innovations in other areas such as politics and
medicine. Atiśa cites and comments on a number of texts attributed to Nāgārjuna – such as the
Commentary on the Awakening Mind (Bodhicittavivaraṇa), Twenty Verses on the Great Vehi-
cle (Mahāyānaviṃśīkā), Verses on the Heart of Dependent Arising (Pratītyasamutpādahṛdaya
kārikā), and Stages of Meditation (Bhāvanākrama) – that he considered vital to understanding
the Middle Way. These texts were not yet fully translated into Tibetan by the eleventh century
604
Atiśa
and are often not included in the exegesis of Madhyamaka by later traditional Tibetan scholars
or by modern scholars. As such, Atiśa followed a commentary attributed to Nāgārjuna, No
Fear from Anywhere (Akutobhayā), for his interpretation of verses found in the Fundamental
Verses of the Middle Way (Mūlamadhyamakakārikā). Atiśa understood Nāgārjuna’s devotional
praises, such as the Hymn to the Realm of Reality (Dharmadhātustava), as complementary
to his works emphasizing reasoning, such as the Sixty Stanzas on Reasoning (Yuktiṣaṣṭikā).
Nāgārjuna was thus the principal master of Atiśa’s Madhyamaka in this holistic and inclusive
way of interpretation. Atiśa also considered Nāgārjuna to have lived for 600 years and to have
written esoteric Buddhist works. Moreover, Atiśa described his teachers’ visionary encounters
with Nāgārjuna as part of an ongoing revelatory lineage of Madhyamaka. Atiśa also consid-
ered Candrakīrti had lived for 400 years in India and been a direct disciple of Nāgārjuna.
Atiśa’s Madhyamaka lineage represented a contemplative tradition that emphasized a cultiva-
tion of the resolution for awakening, the development of compassion, and the realization of empti-
ness leading to the inconceivable state of buddhahood. Although reasoning does have its place in
Atiśa’s system, he advocated a faith-based Madhyamaka that valued the instructions of the spiritual
teacher and held predictions, prayers, and meditative cultivation in high regard. Atiśa’s Middle Way
synthesized the teachings of Bhāviveka and Candrakīrti, built on the basis of Nāgārjuna’s teach-
ing. In both India and Tibet, Atiśa pedagogically utilized the works of Bhāviveka – particularly his
Blaze of Reasoning (Tarkajvālā) and Jewel Lamp of the Middle Way (Madhyamakaratnapradīpa)
– as an introduction to Madhyamaka and then taught advanced students Candrakīrti’s system as
exemplified in the Introduction to the Middle Way (Madhyamakāvatāra). Atiśa primarily synthe-
sized the teachings of Bhāviveka and Candrakīrti by bringing together compatible elements of their
teachings for soteriological efficacy in progression on the path. He emphasized the commonality
of Bhāviveka and Candrakīrti’s teachings, as both pertain to practices within conventional reality
to reach the goal of realizing ultimate reality while downplaying any minor points of philosophi-
cal difference between the two thinkers. Atiśa identified this commonality between Bhāviveka
and Candrakīrti as “the stairway of correct conventional reality” ( yang dag kun rdzob kyi skas),
providing him with an avenue for advocating conventional practices, such as moral discipline and
meditation, as the dependently arisen means that lead toward the goal of realizing ultimate reality.
Atiśa’s Madhyamaka thought and practice underlie his understanding of advanced bodhisat-
tva practices within both the Perfection and Secret Mantra Vehicles. In Mahāyāna forms of
Buddhism, the term “vehicle” ( yāna) is a metaphor for the way or path that a bodhisattva
takes to achieve Buddhahood. The exoteric way is the vehicle of the Perfections ( pāramitā)
in which a bodhisattva practices perfections, such as morality, patience, and wisdom, as out-
lined in Mahāyāna scriptures (i.e., sūtras). The esoteric way is the vehicle of Secret Mantra in
which a bodhisattva receives a consecration from a qualified preceptor in secret to utilize man-
tras, maṇḍalas, and mudrās to quickly achieve Buddhahood as outlined in Mahāyāna ritual
texts (i.e., tantras). Atiśa called his understanding of Madhyamaka, or the Middle Way, “Great
Madhyamaka.” Atiśa’s classification of Great Madhyamaka, in brief, represents his effort to
differentiate the meaning of the Perfection of Wisdom as taught by Nāgārjuna from its meaning
as taught by the Yogācāra scholar Asaṅga. Atiśa consistently upheld a Madhyamaka view over
and against Yogācāra ontological tenets in all of his known works. Along these lines, Atiśa
rejected the utilization of epistemological cognitions for the realization of emptiness. As Atiśa
states in his Entry to the Two Realities (v. 10),
Direct perception and inference are the two valid cognitions ( pramāṇas) accepted
by Buddhists. The deluded whose vision is narrow say that emptiness is understood
by these two.
605
James B. Apple
For Atiśa, the science of epistemology and logic was a secular science common to Buddhist
and Indian non-Buddhist schools. Cognitions derived from epistemological warrants were,
for Atiśa, only utilized at the level of conventional reality to refute opponents. According
to Atiśa, valid cognitions are only conventional and are not in the domain of ultimate real-
ity, nor are they able to realize ultimate reality. Along these lines, Atiśa also disparaged the
practice of debate in his A Brief Treatise on the Method of Practicing the Great Vehicle Path
(Mahāyānapathasādhana-varṇasaṃgraha; Sherburne 2000, 452–53):
Neglecting one’s hard-to-tame mental stream while practicing argument in order to

learn debate, or engaging in the explanation of the teaching in every moment of the
day and night for worldly things such as fame and so forth, life quickly passes on
without purpose and one degenerates from the supreme path.
However, although Atiśa did not advocate the practice of debate, this does not mean that he
rejected the use of reasoning ( yukti) – for Atiśa, reasoning is an essential method that leads
to the direct realization of insight. Reasoning thus signifies an “internal” Buddhist form of
critical analysis that is different from the “external” epistemological devices used to defend
Buddhist Dharma and defeat non-Buddhist opponents.
A General Explanation specifies that the object of negation of reasoning is a conceived
object based on conceptualization that imputes things as either existent or nonexistent. The
object negated by reasoning consists of conceptual thought that imputes an object as existing
with its own character. According to Atiśa, things appear to be substantially existent but, under
analysis through reasoning, are realized to subsist without essence. Atiśa refers to appearances
without substance as mere appearances. However, Atiśa stresses in this work that unestablished
mere appearances are not refuted by reasoning. Rather, appearances are overturned through
antidotes cultivated while practicing the path, particularly the path of vision (darśanamārga)
and the path of meditation (bhāvanāmārga). The vision of gnosis cultivated in Atiśa’s Middle
Way meditation instructions, in brief, dissipates the cognition of appearances as real. A Gen-
eral Explanation offers an early distinction between objects negated by an antidote (gnyen po’i
dgag bya) while implementing the path and objects negated by reasoning (rigs pa’i dgag bya)
when searching out the inherent existence of something (Apple 2019a, 207–8).
Atiśa states “four great reasons” (gtan tshigs chen po bzhi) for the cultivation of insight in
meditation in his Commentary on the Difficult Points in the Lamp for the Path to Awakening
(Bodhimārgapradīpapañjikā) and also in A General Explanation. The “four great reasons”
proving emptiness for Atiśa are: (1) the reason refuting production according to the tetra-
lemma (mu bzhi skye ’gog gi gtan tshigs; catuṣkoṭyupādapratiṣedhahetu), (2) the “diamond-
splinters” reason (rdo rje gzegs ma’i gtan tshigs; vajrakaṇahetu), (3) the reason of being
neither one nor many (gcig dang du ma bral ba’i gtan tshigs; ekānekaviyogahetu), and (4) the
reason consisting in dependent origination (rten ’brel gyi gtan tshigs; pratītyasamutpādahetu).
Atiśa’s four reasons constitute a system of analytical procedures to prove the absence of
intrinsic essence in all things. For Atiśa, the (4) reason consisting in dependent arising
is based on the argument that because things arise due to causes and conditions that are
mutually related, they therefore lack an intrinsic essence. Atiśa claims that the reasoning of
dependent arising (rten ’grel gyi rigs pa) contains all four reasonings within it. The (1) reason
refuting production according to the tetralemma, or four points, is an analytical procedure
which demonstrates that things lack intrinsic essence because (i) something already intrin-
sically existent cannot dependently arise, and, likewise, (ii) something nonexistent cannot
arise. Things also cannot arise (iii) through being existent and nonexistent at the same time,
606
Atiśa
nor (iv) through being neither existent nor nonexistent. The (2) diamond-splinters reason
(rdo rje gzegs ma), metaphorically like a diamond that crushes the rock of substantial views,
is an analytical procedure that examines the possibility of intrinsic essence from the point of
view of the cause. The (3) reason of being neither one nor many, which examines both the
cause and effect in determining the possibility of intrinsic nature, argues against things hav-
ing intrinsic identity or intrinsic plurality in proving the lack of intrinsic essence in things.
Atiśa’s system of positing four reasons for proving emptiness is different from
Kamalaśīla’s, which includes five reasons (Keira 2004, 10–13). Atiśa leaves out the reason
refuting the production of existent and nonexistent things ( yod med skye ’gog gi gtan tshigs,
*sadasadutpādapratiṣedhahetu) that is discussed by earlier Mādhyamikas. A General Expla-
nation clarifies that these reasons are based on the reasoning of dependent arising and that all
four reasons are accepted as consequences that nonimplicatively negate the intrinsic existence of
things but do not negate the mere appearance of causes and effects. The mere appearances that
arise from causes and effects are overturned through antidotes cultivated while practicing the path.
Based on statements in the Kadampa Collection on the Two Realities and A General Expla-
nation, Atiśa followed a lineage of Madhyamaka that advocated the use of consequences that
exposed contradictions and employed other-acknowledged inferences (Apple 2019a, 39). In
fact, A General Explanation contains the earliest extant mention of the four types of con-
sequence used by Mādhyamikas. The four types of consequence are (1) consequences that
compose contradictions (’gal ba brjod pa’i thal ’gyur, *virodhacodanāprasaṅga), (2) the
inference that is known to others (gzhan la grags pa’i rjes dpag), (3) the evidence that is not
established due to the equivalence with what is being established (bsgrub bya dang mtshungs
pa’i ma grub pa), and (4) the equivalence of the reason. Atiśa explains these four as follows:
The text of the Ācārya Nāgārjuna states (1) consequences that expose contradictions
(gal ba brjod pa’i thal ’gyur); (3) a pseudo-sign similar to what is to be proven; (4)
equivalence (mgo bsgre ba): “If you accept in this way, because the reason is not
different, you must accept this as well.” These are bound to the opponent. (2) Infer-
ence that is known to others (gzhan la grags pa’i rjes dpag) in which one states:
“If you yourself accept in this way, your own understanding is contradictory with
this conclusion.” The property of the subject and the entailment are bound with the
opponent and are established by their acceptance. Even though these arguments by
consequence are proclaimed through four reasons, they are not different than being
included within the reason of dependent-arising.
(Apple 2019a, 204)
Atiśa did not accept that the Madhyamaka posits a thesis (dam ’cha’ ba ≈ pratijñā), nor that a
Madhyamaka concedes to the principle of common establishment. The early Kadampa com-
mentaries provide evidence that Atiśa’s early followers accepted his position on Madhyamaka
reasoning of not adhering to the principle of common establishment (ubhayasiddhatva), which
became in later Tibetan scholarship the contentious issue of “commonly appearing subjects”
(chos can mthun snang ba) (Apple 2019a, 39).
The principle of common establishment, traceable to statements made by Dignāga (fifth to
sixth century), was that the reason and subject must be established for both parties in a debate.
That is, for both parties in a debate, there must be something that similarly appears (mthun
snang ba) in order for logical debate and discussion to take place. Some Mādhyamikas in India
accepted this principle; others did not. Śāntarakṣita, Kamalaśīla, and Jñānagarbha accepted it
at least on the conventional level (Keira 2004). Candrakīrti, on the other hand, appears to be
607
James B. Apple
one of the few scholars within Indian Madhyamaka traditions not to have followed the rule of
common establishment. Atiśa followed this lineage tradition (Apple 2019a, 38).
Atiśa compares the reasoning process to two sticks, which, after rubbing together and
generating a fire, burn up and become nonexistent. Although Atiśa does not state his textual
source, he draws this example from the Chapter on Kāśyapa (Kāśyapaparivarta), which is
cited in the Jewel Lamp of the Middle Way. Kamalaśīla also cites this sūtra in his Stages of
Meditation (Bhāvanākrama) and Commentary on the Dhāraṇī “Entering into Non-Conceptu-
ality” (Avikalpapraveśadhāraṇīṭīkā) as an example to illustrate that although the analysis of
reality is indeed the nature of conceptual thought, it will nevertheless be consumed by the fire
of correct wisdom produced by it (Apple 2018c). In his Open Basket of Jewels, Atiśa states
that “the wisdom of individual analysis (so sor rtog pa’i shes rab) itself turns into clear light.”
Atiśa’s Indian disciple Prajñāmukti is even clearer in his Commentary on the Special Instruc-
tions (Madhyamakopadeśavṛtti; Apple 2019a, 285–86):
That the very wisdom which individually discriminates is not established either
means that analytical cognition negates itself. Since wisdom is a particularity of an
entity, when an entity is not established, the very wisdom itself is also not established,
just like when a tree is not established, the wood and so forth are negated. As it is
said, “In this regard, a fire which burns fuel, having burned its fuel does not remain.”
These passages indicate that, for Atiśa and his followers, reasoning is a conventional process
that dissolves itself when seeking to establish the existence of an object. Analytical reasoning
which dissolves itself is preparatory for wisdom – more specifically, for non-conceptual gnosis
(nirvikalpajñāna). The texts suggest a difference between discernment ( prajñā) at the level
of learning and the reflection, reasoning (rigs pa’i shes rab), and non-conceptual realization
that constitute gnosis ( jñāna). The numerous reasonings which appear in Atiśa’s works and
his Kadampa commentaries are meditations that lead the reader through merelogical forms of
analysis to dissolve the conceptual thought that reifies things and their relations.
Atiśa did not differentiate Madhyamaka into separate schools. Instead, he utilized the texts of var-
ious Madhyamaka authors for pedagogical and soteriological purposes. Atiśa introduced his students
to the thought and practice of Madhyamaka with the works of Bhāviveka. To support more advanced
levels of understanding, Atiśa closely followed the works of Nāgārjuna, Candrakīrti, and Śāntideva.
According to Atiśa’s Madhyamaka, most prominently in the General Explanation (Apple
2019b, 191–266), all appearances are based on ignorance and composed of ignorance. Appear-
ances are not nonexistent but rather false, erroneous, and mistaken. The synonyms of the con-
ventions of mere appearance in Atiśa’s nominalistic Madhyamaka are mere name, mere word,
mere convention, and mere imputation. For Atiśa, the subjective perceiver, the appearance of
the cognition, and the appearance of the mind, which is like an illusion, are only imputedly
established. Thus, the cognizer, what is cognized, and the phenomenal awareness that occurs
in such a conjunction of ephemeral conditions lack any substantive ontological status, either
conventionally or ultimately, in Atiśa’s Madhyamaka.
A key Madhyamaka classification that structures Atiśa’s thought and practice is the bifurca-
tion of all things in the universe into the categories of two realities: ultimate reality and conven-
tional reality. As Atiśa states in his Song with a Vision for the Realm of Reality, ultimate reality
is a purified appearance of nonappearance like the center of space, known through individually
intuited knowledge. For Atiśa, the “ultimate” is a conventional expression, also embodied by
the phrase “realm of reality” (dharmadhātu), consisting of selfless nonappearances that are
realized with nonconceptual pristine awareness ( jñāna). The realm of reality, for Atiśa, is a
608
Atiśa
naturally pure object of realization that is the ever-present real state of things that may be cog-
nized in meditative equipoise but is not completely actualized until full buddhahood.
Conventional realities are appearances that arise due to causes and conditions. Though all
conventional realities are false and deceiving, they are not nonexistent. Rather, they are mere
appearances, subject to the principle of cause and effect imputed through dependent arising. In
Atiśa’s system, conventional realities are classified either as mistaken or correct when viewed
from different perspectives in relation to the cognitive understanding of ordinary individuals
or the realizations of those who have reached the path of vision. The three ways of identifying
(’jogs lugs) mistaken and correct conventional realities are explained based on Candrakīrti’s
Introduction to the Middle Way. Conventional realities are dependently designated in relation to
the perspectives of (1) the worldly (lo ka pa), (2) philosophical tenets (grub mtha’), and (3) yogic
awareness (rnal ’byor pa’i blo). A General Explanation posits correct and mistaken conventional
realities based on the nature of dependent arising in relation to yogic awareness. This understand-
ing of the dependent arising of conventional reality in correlation with its states of awareness
accords with what Wangchuk (2009) has called “the relativity theory of the purity and valid-
ity of perception” in Madhyamaka works such as Candrakīrti’s Introduction to the Middle Way
(6.27) and Śāntideva’s Introduction to the Practice of Awakening (Bodhicaryāvatāra) (9.3–4a).
In the words of A General Explanation, “things are not higher or lower; awarenesses are higher
or lower” (Apple 2019a, 180–81). Atiśa accepted the distinction between correct and incorrect
conventional realities even though he considered conventional realities false and unreal. This dis-
tinction is structured within a framework of shifting perspectives as one progresses along the path
in accord with Bhāviveka’s Blaze of Reasoning commentary on the Heart of the Middle Way and
Candrakīrti’s Introduction to the Middle Way (Apple 2016). As outlined in the following para-
graphs, Atiśa refers to this progress on the path as the “stairway of correct conventional reality.”
Mistaken conventional realities are appearances of ignorance that impute impermanent and
empty things as either existent or nonexistent. Mistakes are impermanent and cause suffering,
and they are also deceptive and false. Mistaken appearances are like the hair that is perceived
by someone suffering from eye disease. A General Explanation states that its acceptance of
mistaken conventional realities is similar to how True Aspectarians (*Satyākāravadins) posit
aspects, or cognitive images (ākāra). This process is comparable to Candrakīrti, who, as
pointed out by MacDonald (2009, 151), skillfully adapts the Sautrāntika theory of cognition
on the conventional level to justify his own views. Atiśa’s position on the status of external
objects is similar to the presentation of “internal” Madhyamaka (nang gi dbu ma) in the Jewel
Lamp of the Middle Way (D, 280a3–81a3; Del Toso 2014). However, Atiśa also states in his A
General Explanation – in correlation with Nāgārjuna accepting appearances as mind in works
such as the Stages of Meditation and in terms of the mind as mere appearance, a perspective
equal to correct conventional reality – that all sentient beings are one single continuum (rgyud
gcig; 712.1–713.5). All sentient beings are considered one continuum, for even though they
have differences of karmic conditions, they share an undifferentiated self-nature (rang gi ngo
bo la tha dad med pa) that is free from the two extremes of intrinsic essence. Atiśa also asserts
consciousness to be one group (rnam shes tshogs gcig, *ekavijñānakāya) or one single flow,
as opposed to the six distinct types of consciousnesses (ṣaḍvijñānakāya) found among main-
stream Buddhist schools (Apple 2019a, 213, 399–400). Positing one group entails that one
single consciousness moves to individual objects in dependence upon individual sense organs
rather than six distinct active sensory consciousnesses that arise based on the contact of a
sensory faculty with a sense object. In this way, the General Explanation presents a mentalist
theory of Madhyamaka in which the mind, as mere appearance, is not at all established and is
a mere nominal designation free from the extremes of existence and nonexistence.
609
James B. Apple
Correct conventional reality for the yogi, according to Atiśa, occurs only after completing
the path of vision, a stage on the path when emptiness is directly cognized in a non-conceptual
manner. Correct conventional realities are considered nondeceptive, nonerroneous, and trust-
worthy in that, from this perspective, appearances are realized to be unproduced like an illusion,
and objects are cognized as essenceless entities. Although correct, they are conventional due to
arising through causes and conditions and are thus considered illusions of pristine awareness.
Correct conventional realities are nonerroneous illusions and are imputations conducive to puri-
fication, since the causes of purification have nondeceptive individual results. This accords
with appearances of “purified worldly knowledge” mentioned by Bhāviveka in his Heart of the
Middle Way and his Blaze of Reasoning. Atiśa refers to correct conventional reality, following
Bhāviveka, as the stairway of correct conventional reality ( yang dag kun rdzob rnams kyi skas
≈ tathyasaṃvṛtisopānam). Atiśa’s understanding of correct conventional reality offered him an
avenue for upholding the conventional practices of moral virtue that eventually lead to realizing
ultimate reality, practices that some critics argued were undermined by Madhyamaka analysis.
Based on these factors of Madhyamaka thought and practice, Atiśa outlines a program
of bodhisattva training that integrates the cultivation of wisdom and compassion with the
development of an awakening mind at both the ultimate and conventional levels. The ulti-
mate awakening mind – the birthless, luminous, nonconceptual realm of reality equated with
emptiness – is cultivated during meditative equipoise. The conventional awakening mind is
cultivated during postmeditative practices. In this program of training, the two levels of the
awakening mind are integrated and stabilized, having the essence of emptiness and compas-
sion. A snapshot of instructions for this integrated cultivation is found in the Open Basket of
Jewels (Apple 2019a). Atiśa states,
Regarding the training, first the mind did not come from anywhere and will not go
anywhere at the end. The mind does not abide anywhere and is without color and
without shape. The mind does not arise from the beginning nor does it cease at the
end. The mind is empty of inherent existence and has the nature of clear light. One
should recall this again and again.
On the other hand, one should stabilize through accustoming the awakening mind
to love and compassion. One should completely purify the mind and stand firm,
being continuously mindful of each moment of thought with mindfulness, awareness,
thoughtfulness, and conscientiousness.
The Open Basket of Jewels then discusses how the integrated awakening mind is protected
and increased while the bodhisattva advances through the ten stages as outlined by the Sūtra
of the Ten Stages (Daśabhūmikasūtra). In the course of the training, the bodhisattva alternates
between cognizing the nonconceptual space-like realm of reality in meditative stabilization
and then viewing things as illusions in the postmeditative state. The alternation, however,
ceases at the stage of buddhahood.
According to Atiśa, based on numerous citations from the hymns of Nāgārjuna, at the stage
of buddhahood, the purified realm of reality directly and constantly fuses with the Dharma body
(dharmakāya), the ultimate cosmic reality of buddhahood, without any mental element or gnosis
existing at all. Atiśa emphatically advocated in his works that buddhas are completely fused with
ultimate reality in a nondualistic fashion whereby all conceptual thought has been eliminated
and not even nonconceptual wisdom exists. Throughout his writings, Atiśa references Mahāyāna
sūtras; tantras; and the works of Nāgārjuna, Āryadeva, Candrakīrti, and Śāntideva as authorita-
tive texts for this standpoint. In his Open Basket of Jewels, he provides an extended discussion,
610
Atiśa
based on the works of Nāgārjuna, of the point that buddhahood does not have any mind or men-
tal factors. Related to Atiśa’s understanding of buddhahood as bereft of any mental qualities is
his position that a buddha does not have a continuum of wisdom. This issue is specifically related
to the interpretation of a partial stanza in Śāntideva’s Introduction to the Practice of Awakening
(9.15ab) that states, “upon conditions having their continuum cut, an illusion does not arise even
conventionally.” Atiśa directly repeats this portion of Śāntideva’s stanza in his Entry to the Two
Realities (vv. 23cd). An early Kadampa commentary to this verse, likely following an oral tradi-
tion of Atiśa’s teachings, explains that appearances occur due to various causes and conditions,
and if the continuance of the conditions for such appearances is interrupted, then appearances
will no longer arise even conventionally. For Atiśa and his Kadampa followers, all appearances
are due to ignorance and are composed of ignorance. Therefore, when the conditions for any type
of appearance are exhausted, including wisdom or gnosis ( jnāna), then such appearances will no
longer occur. In addition to this discussion within the Entry to the Two Realities, Atiśa directly
addressed the status of gnosis at the level of a buddha in his Commentary on the Introduction to
the Practice of Awakening (Bodhisattvacāryāvatārabhāṣya), a summary of Śāntideva’s Intro-
duction to the Practice of Awakening. Several sections of the text are composed in the form of
a dialogue with the bodhisattva Mañjuśrī to address difficult points of exegesis. Atiśa discusses
the gnosis at the level of a buddha in a section describing the vision of liberating gnosis (rnam
par grol ba’i ye shes mthong ba ≈ vimuktijñānadarśana) at the level of a buddha. Atiśa’s discus-
sion in this work clearly demonstrates that he did not accept that a buddha would have a con-
tinuum of wisdom. On this point, Atiśa’s understanding is congruent with a strict Madhyamaka
understanding of the nature of buddhahood found in the works of Nāgārjuna, Candrakīrti, and
Śāntideva (Apple 2019a, 43–44). This understanding of buddhahood would not be followed by
a majority of Tibetan scholars, particularly after the time of Chapa Chökyi Sengé (1109–1169).
In his works, Atiśa outlined an undifferentiated Madhyamaka tradition. In Open Basket of
Jewels, he emphasizes that his teachings focus on the lineage of Nāgārjuna and defends “the
Mādhyamika followers of Nāgārjuna” as having no faults. His Commentary to the Lamp for the Path
to Awakening documents a tradition that passed from Nāgārjuna through Āryadeva, Candrakīrti,
Bhāviveka, and Śāntideva down to Bodhibhadra. This work also briefly describes his vision of
Madhyamaka as Great Madhyamaka (dbu ma chen po) in its section on insight (shes rab). Atiśa’s
General Explanation claims that Great Madhyamaka represents the definitive understanding of
Nāgārjuna’s thought. Atiśa’s epithet of Great Madhyamaka, in brief, signifies his understanding
of Nāgārjuna’s Middle Way as a realization free of any extremes or conceptual proliferations,
a direct vision of the non-duality of appearances and emptiness. In his own writings, therefore,
Atiśa did not distinguish between individual followers of Nāgārjuna, that is, Mādhyamikas (Apple
2019a, 59–60). Madhyamaka, for Atiśa, was what Ruegg (1981, 30, 57, 59) has labeled “pure”
Madhyamaka, a designation that signifies a Middle Way tradition that is not differentiated based
on features of individual doctrines upheld in the works of Indian Buddhist authors.
Stages of the Path

Atiśa’s greatest contribution and legacy was his synthesis and reintegration of the teachings
of sūtras and tantras into a coherent system of Buddhist practice. Atiśa’s well-known Lamp
for the Path to Awakening and his important but lesser-known Stages for the Path to Awak-
ening (Byang chub lam gyi rim pa ≈ Bodhipathakrama; hereafter Stages) are long-standing
testaments to his innovative and dynamic system. Atiśa began to formulate teachings that
integrate sūtra and mantra in his early writings on the esoteric Buddhist deity Cakrasaṃvara.
His Analysis of Realization (Abhisamayavibhaṅga) articulates the “stages of the path of the
611
James B. Apple
essential meaning,” The stages of the path were initially formulated by Atiśa based on these
advanced stages of esoteric Buddhist teachings.
Atiśa further developed his stages of the path system under the urging of Jangchup Ö while
he was in West Tibet. Atiśa’s Lamp was the result of Jangchup Ö’s patronage and inquiries,
and the treatise became a model for subsequent generations of Tibetan scholars in their under-
standing and practice of the Buddhist path. The Lamp was an imperially sanctioned public
teaching (tshogs chos) that influenced Tibetan Buddhists for over nine centuries. The Stages,
on the other hand, written at the behest of Dromtönpa, was a secret teaching (lkog chos) that
was transmitted among early followers of Atiśa until the time of Pakmo Drupa (1110–1170).
Although the Stages has not been widely studied (or even recognized) by later traditional
and modern scholars of Tibetan Buddhism, its impact was just as great as the Lamp: the Stages
contains within it all the major topics found in later stages of the path literature that the Lamp
does not mention. The Stages clearly outlines the practices and realizations of the individu-
als of small and middling capacity, including such topics as taking refuge, meditations on
death and impermanence, the elimination of wrongdoing, and recollecting the sufferings of
cyclic existence. Although Atiśa’s Stages is primarily devoted to the ethics of karmic cause
and effect, in the context of instruction on pointing out a nonconceptual direct vision of the
emptiness of one’s own mind, it also develops an account of the Middle Way between the
extremes of nihilism and permanence:
All dharmas are the mind, the mind itself is free from all extremes. The multiple vari-
ous causes and effects of virtue and wrongdoing is unceasing, definitely free from the
extreme of nihilism. Since whatever appears of the cause and effect of the round of
rebirth and nirvāṇa is the nature of one’s own mind, which is not at all established, it
is definitely free from the extreme of permanence. Emptiness indivisible with cause
and effect is the nature of one’s own mind, free from the proliferations of extremes,
the Great Middle Way.
(Apple 2019b)
These two stanzas situate Atiśa’s Great Middle Way between the extremes of nihilism and
permanence based on the nonduality of emptiness and cause and effect. They occur in the
instructions on tranquility and insight and represent the type of instruction given to disciples
of advanced spiritual capacity. Later Kagyüpa scholars such as Gampopa (sgom po pa bsod
nams rin chen, 1079–1153) and Pakmo Drupa ( phag mo gru pa rdo rje rgyal po, 1110–1170)
described them as Mahāmudrā, or Great Seal, teachings (Apple 2019a, 52–53). The topics
found in the Stages and the sequence in which they are presented impacted all subsequent
stages of the path literature in Tibet.
Conclusion
Atiśa was an Indian Buddhist itinerant teacher whose instruction was influenced by historical
conditions of time, place, and patronage. Atiśa brought with him to Tibet an active Madhyam-
aka lineage, teaching the works of Nāgārjuna based on those of Candrakirti. This Madhyamaka
lineage was contemplative in nature and based on faith, compassion, and resolutions to attain
a miraculous form of buddhahood rather than on formal logical proof, linguistic semantics, or
metaphysical speculation. Atiśa’s Madhyamaka lineage was actively taught and followed for
at least fifty years after he came to Tibet. His teachings were disseminated before the rise of the
early Kadampa monastery of Sangphu and its debating traditions that, particularly beginning
612
Atiśa
in the twelfth century, emphasized the merger of Madhyamaka with epistemology and came to
shape all later forms of Tibetan Buddhist scholasticism.
Note
1 The intellectual history of most of these topics in Buddhist thought in general, and specifically in
Indian Mādhyamika traditions, have yet to be adequately studied. Ruegg (2000, 2010) has devoted
considerable attention to the topic of whether a Mādhyamika has a thesis. For discussion of the
other topics in later post–fourteenth-century Tibetan Buddhist understandings, see Cozort (1998) and
Ruegg (2002).
Bibliography
Acri, Andrea, ed. 2016. Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts,
Icons. Singapore: ISEAS – Yusof Ishak Institute.
Apple, James B. 2010. “Atiśa’s Open Basket of Jewels: A Middle Way Vision in Late Phase Indian
Vajrayāna.” The Indian International Journal of Buddhist Studies 11: 117–98.
———. 2013. “An Early Tibetan Commentary on Atiśa’s Satyadvayāvatāra.” Journal of Indian Philoso-
phy 41: 263–329.
———. 2015. “A Study and Translation of Atiśa’s Madhyamakopadeśa with Indian and Tibetan Com-
mentaries.” Acta Tibetica et Buddhica 7: 1–82.
———. 2016. “An Early Bka’-gdams-pa Madhyamaka Work Attributed to Atiśa Dīpaṃkaraśrījñāna.”
———. 2017. “Atiśa’s Teachings on Mahāmudrā.” The Indian International Journal of Buddhist Studies 18: 1–42.
———. 2018a. “Atiśa and Ratnākaraśānti as Philosophical Opponents with Attention to Yuktiṣaṣṭikā,
Verse 34.” Critical Review for Buddhist Studies 23: 9–38.
———. 2018b. “Khu lo tsā ba’s Treatise: Distinguishing the Svātantrika/*Prāsaṅgika Difference in Early
Twelfth Century Tibet.” Journal of Indian Philosophy 46: 935–98.
———. 2018c. “The Spiritual Exercises of the Middle Way: Reading Atiśa’s Madhyamakopadeśa with
Hadot.” In Hadot and Buddhist Thought, edited by Luis Gómez and David Fiordalis, 105–44. Berke-
ley: Mangalam Research Center.
———. 2019a. Jewels of the Middle Way: The Madhyamaka Legacy of Atiśa and His Early Tibetan Fol-
lowers. Somerville, MA: Wisdom Publications.
———. 2019b. Atiśa Dīpaṃkara: Illuminator of the Awakened Mind. Boulder, CO: Shambhala.
Atiśa Dīpaṃkaraśrījñāna. Bodhisattva’s Jewel Garland. Bodhisattvamaṇyāvalī. Byang chub sems dpa’i
nor bu’i phreng ba. Tôh. no. 3951 (also, no. 4471). Dergé Tanjur, vol. khi, ff. 294b7–296a1. Tibetan
translation by the author and Tshul khrims rgyal ba.
———. Commentary on the Difficult Points in the Lamp for the Path to Awakening.
Bodhimārgapradīpapanjikā. Byang chub lam gyi sgron ma’i dka’ ’grel. Tôh. no 3948. Dergé Tanjur
vol. khi, ff. 241a4–293a4. Tibetan translation by Atiśa and Tshul khrims rgyal ba. Translated and
Tibetan edition, Sherburne 2000.
———. Entry to the Two Realities. Satyadvayāvatāra. Bden pa gnyis la ’jug pa. Tôh. no. 3902 (also, no.
4467). Dergé Tanjur, vol. a, ff. 72a2–73a7. Tibetan translation by the author and Rgya Brtson ’grus seng ge.
———. Lamp for the Path to Awakening. Bodhipathapradīpa. Byang chub lam gyi sgron ma. Tôh. no.
3947. Dergé Tanjur, vol. khi, ff. 238a6–241a4. Tibetan translation by the author and Dge ba’i blo gros.
———. Lamp for the Summary of Conduct. Caryāsaṃgrahapradīpa. Spyod pa bsdus pa’i sgron ma.
Tôh. no. 3960 (also, no. 4466). Dergé Tanjur, vol. khi, ff. 312b2–313a7. Tibetan translation by the
author and Tshul khrims rgyal ba.
———. Open Basket of Jewels. Ratnakaraṇḍodghāṭa-nāma-madhyamakopadeśa. Dbu ma’i man ngag
rin po che’i za ma tog kha phye ba. Tôh. no. 3930. Dergé Tanjur, vol. ki, ff. 96b1–116b7. Tibetan
translation by the author, Rgya Brtson ’grus seng ge, and Tshul khrims rgyal ba.
———. The Song with a Vision for the Realm of Reality. Dharmadhātudarśanagīti. Chos kyi dbyings lta
ba’i glu. Tôh. no. 2314 (also, 4475). Dergé Tanjur, vol. zhi, ff. 254b7–260b5. Tibetan translation by
the author and Tshul khrims rgyal ba. See also Blo bzang rdo rje rab gling 1999, 66–214.
———. “Stages of the Path to Awakening. Bodhipathakrama. Byang chub lam gyi rim pa.” In Byang
chub lam gyi rim pa, 21–63.7. Leh, Ladakh: Thupten Tsering, 1973. TBRC W1KG506. Also in Bka’
613
James B. Apple
gdams gsung ’bum phyogs bsgrigs glegs bam go gcig, 21–202. Dpal brtsegs bod yig dpe rnying zhib
’jug khang, 2015.
Blo bzang rdo rje rab gling. 1999. Five Treatises of Ācārya Dīpaṃkaraśrījñāna. Sarnath, Varanasi: Cen-
tral Institute of Higher Tibetan Studies.
Blumenthal, James. 2004. The Ornament of the Middle Way: A Study of the Madhyamaka Thought of
Śāntarakṣita. Ithaca, NY: Snow Lion Publications.
Cozort, Daniel. 1998. Unique Tenets of the Middle Way Consequence School. Ithaca, NY: Snow Lion
Publications.
Davidson, Ronald M. 1995. “Atiśa’s A Lamp for the Path to Awakening.” In Buddhism in Practice, edited
by Donald S. Lopez Jr, 290–301. Princeton, NJ: Princeton University Press.
Del Toso, Krishna. 2014. “Some Problems concerning Textual Reuses in the Madhyamakaratnapra-dīpa,
with a Discussion of the Quotation from Saraha’s Dohākośagīti.” Journal of Indian Philosophy 43:
511–57.
Hahn, Michael. 2016. Vṛttamālāstuti of Jñānaśrāmitra with Śākyarakṣita’s Vṛttamālā(stuti)vivṛti. New
Delhi: Aditya Prakashan.
Isaacson, Harunaga. 2013. “Yogācāra and Vajrayāna according to Ratnākaraśānti.” In The Foundation
for Yoga Practitioners: The Buddhist Yogācāra bhūmi Treatise and Its Adaptation in India, East Asia,
and Tibet, edited by Ulrich Timme Kragh, 1036–51. Cambridge, MA: Harvard University Press.
Voidness of All Dharmas: Introduction, Annotated Translations, and Tibetan Texts of Selected Sec-
tions of the Second Chapter of the Madhyamakāloka. Wien: Arbeitskreis für tibetische und buddhis-
tische Studien, Universität Wien.
MacDonald, Anne Elizabeth. 2009. “Knowing Nothing: Candrakīrti and Yogic Perception.” In Yogic
Perception, Meditation and Altered States of Consciousness, edited by Eli Franco and Dagmar Eigner,
133–69.
McCrea, Lawrence J., and Parimal G. Patil. 2010. Buddhist Philosophy of Language in India:
Jñānaśrīmitra on Exclusion. New York: Columbia University Press.
Miyazaki, Izumi. 2007a. “Atiśa (Dīpaṃkaraśrījñāna): His Philosophy, Practice and Its Sources.” Memoirs
of the Research Department of the Toyo Bunko 65: 61–89.
———. 2007b. “Annotated Tibetan Text and Japanese Translation of the Ratnakarandodghaṭa-nāma-
madhyamakopadeśa of Atiśa.” Memoirs of the Faculty of Letters, Kyoto University 46: 1–126.
https://repository.kulib.kyoto-u.ac.jp/dspace/handle/2433/73130?mode=full.
Patil, Parimal G. 2009. Against a Hindu God: Buddhist Philosophy of Religion in India. New York:
Ruegg, David Seyfort. 1981. The Literature of the Madhyamaka School of Philosophy in India. History
of Indian Literature series. Wiesbaden: Harrassowitz.
———. 2000. Studies in Indian and Tibetan Madhyamaka thought. Wien: Arbeitskreis für tibetische und
buddhistische Studien, Universität Wien.
———. 2002. Two Prolegomena to Madhyamaka Philosophy: Candrakīrti’s Prasannāpāda madhyamaka-
vṛttiḥ on Madhyamakakārikā I.1 and Tsoṅ kha pa blo bzaṅ grags pa/rgyal tshab dar ma rin chen’s
dka’ gnad/gnas brgyad kyi zin bris. Studies in Indian and Tibetan Madhyamaka Thought. Wien: Arbe-
itskreis für tibetische und buddhistische Studien, Universität Wien.
———. 2010. The Buddhist Philosophy of the Middle: Essays on Indian and Tibetan Madhyamaka.
Boston, MA: Wisdom Publications.
Sanderson, Alexis. 2009. “The Śaiva Age.” In Genesis and Development of Tantrism, edited by Shingo
Einoo, 41–349. Tokyo: University of Tokyo.
Sherburne, Richard. 1983. A Lamp for the Path and Commentary of Atīśa. The Wisdom of Tibet
Series 5. London: Allen & Unwin.
———. 2000. The Complete Works of Atīśa Śrī Dīpaṁkara Jñāna: Jo-Bo-Rje: The Lamp for the Path
and Commentary, Together with the Newly Translated Twenty-Five Key Texts (Tibetan and English
Texts). New Delhi: Aditya Prakashan.
Wangchuk, Dorji. 2009. “A Relativity Theory of the Purity and Validity of Perception in Indo-Tibetan
Buddhism.” In Yogic Perception, Meditation, and Altered States of Consciousness, edited by Eli
Franco and Dagmar Eigner, 215–37. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
614
38
ABHAYĀKARAGUPTA
A Last Great Pandita
Kazuo Kano
Life
Abhayākaragupta (twelfth century CE) was one of the most influential scholars of the great
monastic university Vikramaśīla.1 Born, according to a Tibetan historical account, into a
Brāhmaṇa family in eastern India, in his youth, he received the instruction from a young
yoginī to go to Magadha, where he became a Buddhist. Returning to Bengal, he took novice
ordination and studied non-tantric teachings. From there he moved on to tantric teachings,
which he studied under Ratnākaragupta, whose name is embedded in the opening verse of his
Instruction on the Procedure for Self-Empowerment (Svādhiṣṭhānakramopadeśa). Eventually
he established a sufficiently high reputation to gain the support of a Pāla king, very likely
Rāmapāla. According to the Tibetan author Tāranātha, Abhayākaragupta served as a mkhan
po2 of both Vikramaśīla and Nālandā monasteries.
Many Tibetan scholars visited Abhayākaragupta in Vikramaśīla and Nālandā in order to
study under him, and he supported their translations of a large number of Sanskrit Buddhist
treatises into Tibetan.3 In Tibet he later came to be regarded as, as Seyfort Ruegg (1981, 115)
states, “a representative of Śāntarakṣita’s Yogācāra-Svātantrika-Madhyamaka school.” In the
following, we shall see that he replicates large sections of Kamalaśīla’s Illuminator of the Mid-
dle (Madhyamakāloka), thus lending weight to this doxographical attribution.
Writings
Some twenty-six works are currently attributed to Abhayākaragupta (see Mori 2009, 2–3; Sinclair
2011; Luo 2020). Among them, twenty-five are preserved in Tibetan, and thirteen may be found
in the Sanskrit original. They include texts which are ascribed to Abhayākaragupta but whose
authorship is yet to be confirmed (e.g., no. 17 subsequently). Four of the works are classified as
non-tantric and the other twenty-two as tantric works. His non-tantric works are the following:
1. Adornment of the Sage’s Thought (Munimatālaṃkāra) (Skt., Tib.)

2. Flower Cluster of the Middle Way (Madhyamakamañjarī) (Skt.)
3. Moonlight on Crucial Points (Marmakaumudī) (Tib.)
4. Manual for Taking the Bodhisattva Vow (Bodhisattvasaṃvaragrahaṇavidhi) (Tib.)
DOI: 10.4324/9781351030908-53 615

Kazuo Kano
No. 1 (completed in 1108/1109) is his main non-tantric work. Encyclopedic in scope, it is one of the
most extensive sources for determining his philosophical position, a topic to which we shall return
in the following. No. 2 expounds upon Madhyamaka philosophy, seeking to establish the absence of
an intrinsic nature among all phenomena and refuting other philosophical positions. This text is pre-
served only in Sanskrit, and an edition is currently being prepared by Hong Luo. No. 3 is an extensive
commentary on the Adornment of Realization (Abhisamayālaṃkāra) on the basis of the Perfection of
Wisdom in Eight Thousand Lines (Aṣṭasāhasrikā Prajñāpāramitā). No. 4 is a ritual manual for taking
the bodhisattva vow and contains passages parallel to the Adornment of the Sage’s Thought.
Among his 22 tantric works, the following three are commentarial literature:
5. Flower Cluster of Transmissions (Āmnāyamañjarī) (Skt., Tib.)

6. Guide by the Fearless One (Abhayapaddhati) (Skt., Tib.)
7. Moonlight on the Five Procedures (Kramakaumudī) (Skt., Tib.)
No. 5 (completed in 1115/1116) is his largest work. Styled as a commentary on the Tantra of Emer-
gence from Sampuṭa (Saṃpuṭodbhavatantra), it is in fact an encyclopedic work covering a wide vari-
ety of topics over the entire field of tantric Buddhism, including a number of philosophical subjects,
such as a discussion of differences between non-tantric and tantric Mahāyāna (see Tomabechi and
Kano 2008). No. 6 is his commentary on the Tantra of the Skull of the Buddha (Buddhakapālatantra)
and No. 7 on the Five Procedures (Pañcakrama), which is a tantric meditation manual stemming
from the Guhyasamājatantra (an edition of which is currently being prepared by Toru Tomabechi).
The following three, called collectively the “Garland Trilogyˮ ( phreng ba skor gsum), are
systematic compendiums of ritual or meditation manuals dealing with various tantric tradi-
tions of Indian Buddhism:
8. Garland of Vajras (Vajrāvalī) (Skt., Tib.)

9. Garland of Completion Yoga (Niṣpannayogāvalī) (Skt., Tib.)
10. Flower Cluster of Fire Oblation (Jyotirmañjarī) (Skt., Tib.)
No. 8 is the key part of the trilogy, while the other two (Nos. 9 and 10) are called, by
Abhayākaragupta himself, “supplements” ( parikara) to it. No. 8 treats some fifty ritual manuals,
focusing on describing “maṇḍalas to be drawnˮ (lekhyamaṇḍala) for the purpose of consecrating
( pratiṣṭhā and abhiṣeka) either objects of worship or people. The Garland of Vajras had a great
impact on Tibetan tantric tradition, including serving as the basis for the production of a set of
scroll paintings (thang ka) depicting maṇḍalas at Ngor monastery, inspiring Tsong kha pa’s Great
Gradual Paths of Esoteric Teachings (sNgags rim), and so on. No. 9 details twenty-six particular
kinds of “maṇḍalas for meditationˮ (bhāvyamaṇḍala), describing visualizations of each deity
within them. No. 10 is a tantric manual for the fire-offering ritual (homa, see Mori 2011).
There are two works each on the Cakrasaṃvara cycle and Kālacakra cycle:
11. A Guide to the Realization of Śrīcakrasaṃvara (Śrīsaṃvarābhisamayopāyikā) (Skt., Tib.)

12. Instruction on the Procedure for Self-Empowerment (Svādhiṣṭhānakramopadeśa) (Skt., Tib.)
13. Key Points of the Kālacakra (Kālacakroddāna) (Tib.)
14. An Entrance Gate to the Kālacakra (Kālacakrāvatāra) (Skt., Tib.)
No. 11 is a meditation manual based on Lūyīpāda’s Realization of Cakrasaṃvara

(Cakrasaṃvarābhisamaya) dealing with the 62-deity maṇḍala of Cakrasaṃvara (the Sanskrit orig-
inal of which was recently found by Péter-Dániel Szántó), while No. 12 deals with the visualization
616
Abhayākaragupta
leading to one’s self-transformation into the two-armed Cakrasaṃvara. Nos. 13 and 14 are short
works on the Kālacakratantra, with parts of No.13 having to do with astrology (gza’ skar).
The following four are miscellaneous tantric works:
15. Flower Cluster of Instruction (Upadeśamañjarī) (Tib.)

16. Guide to Awakening (Bodhipaddhati) (Tib.)
17. Short Treatise on Consecration (Abhiṣekaprakaraṇa) (Tib.)
18. Flower Cluster of Transgressions within the Vajrayāna (Vajrayānāpattimañjarī) (Tib.)
No. 15 argues several different positions regarding tantric realizations of the generation stage
(utpattikrama) and the stage of the arisen (utpannakrama) including those of Ratnākaraśānti
(alluded to by the phrase “a certain master [ācārya] of Pūrvadeśaˮ), Nāgabodhi, Ānandagarbha,
Jñānapāda, Indrabodhi, and Kambalāmbara. It sometimes discusses philosophical topics,
such as an argument on the ontological status of visualized maṇḍala-deities, the applicabil-
ity of tantric visualization to the Three Nature theory of Yogācāra, and different positions
on the classification of the bodies of a buddha. Its opening verse may contain an allusion
to Abhayākaragupta (mi ’jigs mgos phyag ’tshal), and the work contains a number of refer-
ences to his other works (e.g., Garland of Vajras). These facts suggest that this is a work by
Abhayākaragupta, though the attribution is not certain.
No. 16 deals with a series of worshipful mantra recitations. No. 17 is a treatise on consecra-
tion rituals (abhiṣeka). No. 18 is a brief list of fourteen kinds of fundamental transgressions
(mūlāpatti) and eight kinds of coarse but less grave transgressions (sthūlāpatti) in tantric practice.
Finally, there are eight further minor works of tantric ritual manuals focusing on specific
tantric deities or themes:
19. Accumulation of Rituals of Mahākāla (Mahākālakarmasaṃbhāra) (Tib.)

20. Ritual Procedure for Rejoicing the Lord Kākayoni (Nāthakākayonitarpaṇavidhikrama) (Tib.)
21. Perfection of Raktayamāntaka (Raktayamāntakaniṣpanna) (Tib.)
22. Meditation Guide of Jñānaḍākinī (Jñānaḍākinīsādhana) (Tib.)
23. Meditation Guide of Ucchuṣmajambhala (Ucchuṣmajambhalasādhana) (Tib.)
24. Fire Oblation of Vajramahākāla for Harming Enemies (Vajramahākālābhicārahoma) (Tib.)
25. Ritual Manual for Rain (Varṣāpaṇavidhi) (Skt.)
26. Ritual Manual for the Tantric Feast (Gaṇacakravidhi) (Tib.)
These all are ritual or meditation manuals dealing with the worship of various deities or for the
fulfillment of mundane wishes.
Writing Strategy
One of Abhayākaragupta’s concerns appears to have been to avoid establishing a new tradition
of his own, and a key strategy towards this end was to fill his text with a patchwork of pas-
sages written by other authors. He frequently draws upon such passages without mentioning
the titles or authors of these sources. In this regard, Abhayākaragupta may seem to be more a
compiler than an author, although such a judgment risks also overlooking his contribution. In
any case, he usually reproduces the original text almost verbatim, not drastically changing the
original wording or word order, and he only exceptionally omits words. Thanks to this mosaic-
like patterning, we can recover a number of passages from rare Buddhist works in the original
language when those Sanskrit texts have yet to be found.
617
Kazuo Kano
For instance, in his Adornment of the Sage’s Thought, passages from Kamalaśīla’s Illu-
minator of the Middle (Madhyamakāloka) and Candrakīrti’s Treatise of the Five Aggregates
(Pañcaskandhaka), both of which are so far available only in Tibetan translations, have been
recovered (see Kano and Li 2020). In his Flower Cluster of Transmissions, we likewise find
passages from works of Ratnākaraśānti, Kamalanātha, Bhavabhaṭṭa, Nāgabodhi, and so forth
(Isaacson and Sferra 2019, 251). These discoveries alone make his works valuable. But they
are also interesting in terms of their presentation of doctrine.
Locating Abhayākaragupta’s Doctrinal Position

Abhayākaragupta leans philosophically towards Madhyamaka thinkers, especially
Kamalaśīla, but also Nāgārjuna, Śāntarakṣita, and Candrakīrti, among others. With
regard to his understanding of the Adornment of Realization tradition, he mainly fol-
lows Āryavimuktisena, while when dealing with tantric issues, he frequently reuses pas-
sages from Nāgabodhi, Jñānapāda, Ratnākaraśānti, Kamalanātha, and others, including
his master Ratnākaragupta (Isaacson and Sferra 2019, 251). His own impact is seen, for
instance, in works by Daśabalaśrīmitra and Ratnarakṣita and also within the Tibetan tra-
ditions. Many of his works are yet to be studied, and future research can be expected to
clarify his doctrinal position more precisely.
The works that center on philosophical topics are the Adornment of the Sage’s Thought,
Moonlight on Crucial Points, Flower Cluster of the Middle Way, and a number of tantric trea-
tises, such as the Flower Cluster of Transmissions and Flower Cluster of Instruction.
The Adornment of the Sage’s Thought is the most comprehensive and systematic among his
non-tantric works. One might call it Abhayākaragupta’s encyclopedic overview of the entire
system of non-tantric Buddhist doctrines and practices.4 Abhayākaragupta himself states his
purpose in composing the work in the opening verses:
The highest thought of the great sages (i.e., buddhas), namely, the own nature of
this body (mūrti, i.e., dharmakāya), was revealed by the excellent noble ones (i.e.,
Nāgārjuna, Maitreya, etc.). But worldly people, who are confounded by darkness, do
not understand this thought – therefore, the succession ( paraṃparā) of light shed on
this thought (i.e., in the four chapters of this work).
According to this third verse of the opening, Abhayākaragupta composed this work in order to
guide ignorant persons who do not comprehend the buddhas’ thought (i.e., the “sage’s intent”
or munimata) as expounded by the noble ones Nāgārjuna, Maitreya, Āryavimuktisena, and
others. He defines this work as:
The light from (i.e., the chapters of) this Adornment of the Sage’s Thought, which
consists of the jewel of qualities, shows the indescribably correct view.
The word “lightˮ or “illuminatorˮ or “elucidationˮ (āloka) in this and previous verses indicates
the four chapters of the work, each of whose titles ends with the word āloka. The chapters are:
Chapter 1: The Light of the Resolve to Become a Buddha (bodhicittāloka)

Chapter 2: The Light of Cultivating the Resolve to Become a Buddha (bodhicittabhāvanāloka)
Chapter 3: The Light of the Eight Realizations (ạṣṭābhisamayāloka)
Chapter 4: The Light of the Resultant Qualities (guṇāloka)
618
Abhayākaragupta
With this as his framework, Abhayākaragupta undertakes a comprehensive, systematic treat-

ment of Mahāyāna Buddhism, elucidating its doctrinal foundations in chapter 1, its soterio-
logical path in chapters 2 and 3, and the fruits of its path in chapter 4.
Chapter 1 delves in depth into the resolve to become a buddha (bodhicitta) by describing sig-
nificant related doctrinal and practice-oriented topics of non-tantric Buddhism: the significance of
compassion as a starting point of Buddhist practices; the ritual involved in taking a vow to become
a Buddhist; the Six Perfections, especially the perfection of moral conduct; the cosmologies of
sentient beings and the inanimate world; the Abhidharmic classifications of all phenomena into
the five aggregates ( pañcaskandha), and so on; the four truths of the noble ones (caturāryasatya);
the three natures of Yogācāra (trisvabhāva); the two truths (satyadvaya); and the single vehicle
(ekayāna). Chapter 2 elucidates the cultivation of the twenty-two kinds of the resolve to become a
buddha that are taught in the Adornment of Realization (following Jñānakīrti’s Commentary on the
Entrance to the Reality (Tattvārthāvatāravyākhyā) [D 3709]; see Isoda 1998). Chapter 3 discusses
the eight kinds of realization at the core of the Adornment of Realization. Chapter 4 deals with quali-
ties resulting from the cultivation of the resolve to become a buddha. Of the four chapters, the first
provides the richest insight into Abhayākaragupta’s central philosophical leanings, the two truths
being by his own confession one of the most crucial doctrinal points for him (see subsequently).
Selected Topics Pertinent to Abhayākaragupta’s Philosophical

Position
In the following, we shall review selected topics of Abhayākaragupta’s philosophical position,
mostly focusing on the Adornment of the Sage’s Thought. These topics shall be:
(1) The scriptural authority of Mahāyāna texts

(2) The two truths
(3) The two truths and an evaluation of the mind-only doctrine
(4) The illusory but beneficial mind
(5) Buddha-nature
The Scriptural Authority of Mahāyāna Texts

In the middle of the last chapter of the Adornment of the Sage’s Thought, Abhayākaragupta
inserts a discussion of an extra argument with the expression “now, enough of excessive argu-
mentation.” This passage (Skt. Ms. fols. 148r1–150v2) appears in the context of the eighth
of the ten great aspirations (mahāpraṇidhāna), that is, the “application of one’s single will to
enter the Great Vehicle on account of not being desirous of other Vehicles,” one of the qualities
in the first stage of a bodhisattva. To describe these qualities, he borrows, verbatim, sentences
from Vasubandhu’s Commentary on the Sūtra of the Ten Stages (Kano 2019). The passage in
question, which is probably in Abhayākaragupta’s own words, is inserted in between a series
of these borrowed sentences. There, he refutes the claim to scriptural authority on the part of
the rival Vehicle he designates as the Śrāvakayāna (Vehicle of the Disciples) and establishes
the Mahāyāna’s superiority to this Vehicle. A summary of the argument and extractions of key
points follows (note that the following is not a translation but rather a synopsis):
Abhayākaragupta maintains: The Mahāyāna is great because of seven points of

greatness,5 but still there are those who do not believe in the Mahāyāna. How can the
Mahāyāna not be the Buddha’s words?
619
Kazuo Kano
An opponent replies: This is because the Mahāyāna is external to the traditional

Three Baskets, i.e., Sūtra, Vinaya, and Abhidharma, and it does not meet the definition
of the Buddha’s words: “What fits into the Sūtra Piṭaka, what is shown in the Vinaya
Piṭaka, and what does not go against the reality (dharmatā) – that is the Buddha’s word.”
Abhayākaragupta then insists that the Mahāyāna does meet this definition.
The opponent retorts: The Mahāyāna is not what is proclaimed as the Buddha’s
words in the fixed Three Baskets of the eighteen schools.6
Abhayākaragupta’s response: If the Three Baskets of the eighteen schools are the
Buddha’s words, why do they contain texts that mutually contradict one another? Such
texts cannot be the Buddha’s words. Furthermore, they cannot even be the words of past
arhats, for arhats, who have abandoned the defilements, do not compose such mutually
contradictory texts. They must have been composed by ordinary beings who have not yet
abandoned their remaining traces of ignorance and doubt. Therefore, the sacred scrip-
tures of the Mahāyāna cannot be included in such recent texts made by ordinary people.
Even assuming that there were ancient Śrāvakayāna scriptures that had been lost (i.e.,
even though someone insists that the teaching of Mahāyāna scriptures correspond to
lost Śrāvakayāna scriptures), Mahāyāna scriptures could not be included among them,
for there is a huge gap between the Śrāvakayāna and the Mahāyāna. If the Mahāyāna
could fit into or be included within the Hīnayāna (i.e., Śrāvakayāna), the very essence
of the Mahāyāna would be lost. It is like the great ocean, which does not lose its qual-
ity of being a receptacle (āśraya) of all kinds of water, even though it cannot fit into a
small hole (i.e., it does not matter that the Mahāyāna scriptures are not included in the
Śrāvakayāna scriptures, and this is not a fault of the Mahāyāna, but rather a virtue).
Abhayākaragupta goes on to ask how the bright, variegated Mahāyāna teach-
ings can be useful for the blind followers of the Śrāvakayāna, who are the deserved
objects of compassion? (See Kano forthcoming)
So much for a summary of the passage. In the history of Indian Buddhism, the scriptural author-
ity of the Mahāyāna has been discussed by several authors, including Nāgārjuna, Vasubandhu,
Īśvara, Bhāviveka, Haribhadra, and Prajñākaramati.7 Abhayākaragupta’s argument fundamen-
tally follows the pattern already seen in the works of these previous authors, who pointed out
contradictory statements in the Śrāvakayāna scriptures. Abhayākaragupta himself seems to know
these scriptures well, quoting a number of them from the different schools elsewhere in this work.
On the other hand, the following two points do not seem to be found in similar discussions
by previous authors and thus probably represent his own position: (1) the Śrāvakayāna scrip-
tures are not a depository of the Buddha’s words and not even one of past arhats’ words; rather,
they contain merely the words of ordinary beings, and (2) the Mahāyāna scriptures do not need
to be included in the Śrāvakayāna scriptures, given the superiority of the former to the latter.
This rather harsh attitude towards the followers of the Śrāvakayāna likely reflects the fact that
Śrāvakayāna monks were still active and influential at the time of Abhayākaragupta.
The Two Truths

In one of his closing verses of the Adornment of the Sage’s Thought, Abhayākaragupta evalu-
ates the two truths as being the most crucial doctrine:
Here (i.e., in this work), in sum, the life (i.e., crucial point) of the pair of truths (i.e., the
two truths) has been presented by me from the perspective of the three methods (i.e., the
620
Abhayākaragupta
śrāvaka, pāramitā, and mantra approaches). If there is misunderstanding, please forgive

me for it. As for the two truths, those who are enjoying the bhūmis are the authority.8
“Those who are enjoying the bhūmisˮ mean bodhisattvas practicing the ten stages. In the
following two passages from chapter 1, he explains that the Two Truths encompass all other
Buddhist doctrinal points. In his introductory statement of this Two Truths passage, he states:
These very same five aggregates, twelve āyatanas, and eighteen elements are nothing
but the four truths, that is, the truths of suffering, origination, cessation, and the path;
the three natures, that is, the natures of conceptualization, dependence, and perfec-
tion; the two truths, that is, the conventional truth and the ultimate truth; and finally
the single, nondual ultimate truth, that which is again beyond all proliferation; that
which is precisely the Mahāyāna, Perfection of Wisdom ( prajñāpāramitā), and the
ultimate resolve to become a buddha (bodhicitta).9
In the closing part of this same passage, Abhayākaragupta quotes the Sūtra on the Descent into
Laṅkā (Laṅkāvatārasūtra) 6.5 in order to re-emphasize that these doctrinal topics encompass
the whole of the Mahāyāna teaching and that all of them are in turn included in the two truths:
On the other hand, a summary division of them (i.e., reality and cognition of it) was
taught in the Ārya-Laṅkāvatārasūtra:
“The Five dharmas, the Three Natures, the Eight Cognitions, the Two Selfless-
nesses – these ought to encompass the whole of the Mahāyāna.ˮ
Here, the Five dharmas are name, manifestation, conceptualization, reality, and
the true gnosis. Among these, name, manifestation, and conceptualization are men-
tioned when specifying the conventional truth, and reality and the true gnosis are
mentioned, both literally and figuratively, when specifying the ultimate truth.
The Three Natures are those of conceptualization, dependence, and perfection.
The Eight Cognitions are the storehouse cognition (ālayavijñāna), the defiled mind,
and the cognitions of the six sense faculties, i.e., those of eyes and the like. The
Two Selflessnesses are the selflessness of persons and the selflessness of dharmas.
Among these the conceptual nature, dependent nature, and the Eight Cognitions are
mentioned when specifying the conventional truth. The perfect nature and the Two
Selflessnesses are mentioned when indicating the ultimate truth.10
In sum, Abhayākaragupta subsumes the major doctrinal points under the two truths in the fol-
lowing manner:
Instruction of the Instruction of the

Doctrinal Topics
Conventional Truth Ultimate Truth
Five Dharmas name, manifestation, and reality and the true gnosis
conceptualization
Three Natures the nature of conceptualization and the the nature of perfection
nature of dependence
Eight Cognitions all eight cognitions —
Two Selflessnesses — the two Selflessnesses
621
Kazuo Kano
As an adherent of Madhyamaka, he naturally takes all Eight Cognitions (including the

ālayavijñāna) as falling under the conventional truth. The Eight Cognitions are a part of Yogācāra
theory. They are ālayavijñāna (storehouse consciousness; the substratum of the world, which
is nothing but the mind), kliṣṭaṃ manas (defiled mind; the mind which grasps ālayavijñāna as
“Iˮ), and six kinds of pravṛttivijñāna (forthcoming form of the mind, such as eye-cognition).
The Two Truths and the Mind-Only Doctrine

Abhayākaragupta continues his discussion with an evaluation of the mind-only doctrine
(the mind-only doctrine is a part of Yogācāra doctrine that teaches all phenomena as nothing
but a product of the mind).11 The central question is whether the mind is to be accepted as
merely conventionally existent, and he comes down on the side of that being the case. His
exposition of the mind-only doctrine is found in the same section of the Adornment of the
Sage’s Thought that deals with the two truths, most of it sticking closely to Kamalaśīla’s
train of thought in the Illuminator of the Middle (along with very frequent unacknowledged
passages from Kamalaśīla).
The relevant portion of text can be summarized as follows: Abhayākaragupta views the mind-
only doctrine as the entrance gate to understanding that all phenomena are characterized by their
absence of an intrinsic nature, whence he goes on to explain that externally perceived objects
are only conventionally existent, to disallow any real existence to the mind and external objects,
to deny the existence of a mind that has no mental image, to establish the usefulness of the
mind-only doctrine for attracting ordinary beings, and to refer to a position (most likely that of
Śāntarakṣita) that accepts the conventional existence of the mind but refuses to accord it to exter-
nal objects. He concludes that the adherents of the mind-only doctrine are realists (vastuvādin)
who teach a pseudo ( prativarṇikā) version of the Perfection of Wisdom ( prajñāpāramitā).12
Although he introduces a position that can be identified with that of Śāntarakṣita, who
accepts the conventional existence of the mind but denies even that to external objects,
Abhayākaragupta leaves open his final judgement on the matter, saying:
Thus it does not matter whether external objects, too, exist conventionally, or just
the state of mind-only; there is absolutely no Buddhist teaching at all that assuredly
teaches any real intrinsic nature of phenomena.13
For him, the truth of any final judgement on the ontological status of conventional forms of
existences is less crucial than the pragmatic effect, in other words, the benefit, it has. What
is more important for him, that is, is how the Mind-only doctrine works for the sake of those
exposed to it. In line with Kamalaśīla’s Illuminator of the Middle (D 3887, 159r6–7), he states:
And the mind-only doctrine is not just talk for deluding other people, in as much as
the Buddha’s instruction in general brings forth fruitful results, by bringing benefit
and bliss to those undergoing training. Also, there is no delusiveness in the Fortunate
Ones (i.e., buddhas), since only something that benefits others is the truth. And it was
taught by the master Nāgārjuna: “The truth is something that benefits others, whereas
delusion is the opposite, for it does not bring forth any benefit to others.ˮ14
Although Abhayākaragupta takes the mind-only doctrine as a provisional teaching, he insists

on the significance of this doctrine’s efficacy in benefiting others, and thus, for that very
622
Abhayākaragupta
reason, considers it a truth. He continues to solidify this argument in a context that deals with
the beneficial aspect of the mind, again following Kamalaśīla, as laid out in the next section.
The Illusory but Beneficial Mind

Let us overview the series of arguments concerning the beneficial aspects of the illusory mind,
which follow the passage mentioned just previously. The passages in question are again, in
large part, silent textual borrowings from Kamalaśīla’s Illuminator of the Middle.15
An opponent first criticizes Abhayākaragupta’s statement that the selflessness of all phe-
nomena is directly perceived through yogic perception ( yogipratyakṣa), in as much as selfless-
ness, a state of nonexistence, should be imperceptible; that is, nonexistence cannot give rise to
any perception (cf. Kano and Li 2021, 9; Ms. fol. 64r2–3).
In reply, Abhayākaragupta answers that even nonexistence (for instance, a rabbit’s horn)
can generate a perception and that yogins (i.e., perceivers of reality) do not see the world as
really existent, but they do see its lack of a self. He then quotes the definition of direct percep-
tion from Dharmakīrti’s Commentary on the Means of Knowledge (Pramāṇavārttika): “Direct
perception is free from conceptualization and from delusion” (kalpanāpoḍham abhrāntaṃ
pratyakṣam). And he states that “even if there is no object, a perception does arise very vividly
through the power of cultivation” (cf. Kano and Li 2021, 11; Ms. fol. 64v4) and that this per-
ception is free from conceptualization. He thus refutes the opponent’s objection.
He then explains the three kinds of direct perception (namely those of ordinary beings,
yogins, and buddhas), which are free from delusion (abhrānta) with regard, respectively,
to verbal expressions (vyavahārataḥ), the selflessness of persons, and every aspect of real-
ity (cf. Kano and Li 2021, 12–13, Ms. fol. 65r1–4). With regard to the illusory nature of all
phenomena, he states that “an ultimate nature or the like superimposed by ordinary people
is invalidated by means of inference (or perception). Its being invalidated establishes that all
phenomena are ultimately devoid of an intrinsic natureˮ (65r4). Scriptures are quoted in sup-
port of this (cf. fol. 65r4–v2).
Regarding inference, he mentions two proof statements ( prayoga) and refers to the Flower
Cluster of the Middle Way for further references to the relevant arguments:
And regarding both the reason “being dependently arisen on the part of all phenomenaˮ
( pratītyasamutpannatva) and the reason “being devoid of the nature of oneness and
manifoldness on the part of all phenomenaˮ (ekānekasvabhāvarahitatva), there are
namely, (proof statement 1) “What is dependently arisen is, from the viewpoint of
ultimate reality, empty by nature, just as an illusion is. And all dharmas are depend-
ently arisen”; and (proof statement 2) “What lacks the intrinsic nature of oneness and
manifoldness lacks, from the viewpoint of ultimate reality, an intrinsic nature, just
as a reflection does. And all dharmas lack, from the viewpoint of ultimate reality, an
intrinsic nature of both oneness and manifoldness”. The demonstration of these two
proof statements is not laid out here, for I fear that those who do not know logical
procedures will not listen to it, given their lack of understanding. However, it was
explained by me in a summarized manner – unsurpassably so – in the Flower Cluster
of the Middle Way. Therefore, the wise may reflect on it there.16
The opponent points out an undesirable consequence: If the real existence of all phenomena
is refuted by inferring their selflessness, then what about direct perception, which should be
623
Kazuo Kano
the foundation for this very inference? That is, given that the basis of the inference is to be
abandoned, the inference, too, should be abandoned (cf. Ms. fol. 65v3–4).
Abhayākaragupta accepts this consequence but judges it no problem, because he who has
the ‘fire’ of analytic gnosis (i.e., inference) burns the ‘fuel,’ that is, all these illusory phenom-
ena, and then, having doused the ‘fire’ of analysis, realizes the excellent emancipation (nirvṛti)
which has the cessation of conceptualization as its characteristic (cf. Kano and Li 2021, 15–16;
Ms. fol. 65v4–66r1).
The opponent then points out a further undesirable consequence: Given that even the
mind is devoid of an intrinsic nature on the ultimate level, the gnosis which realizes this very
absence would have to be abandoned (as being merely an illusion), and so there would be no
gnosis among yogins and no gnosis among buddhas, and finally, “there would be the destruc-
tion of the entire system” of Buddhist teaching.
Abhayākaragupta replies by pointing out that although the mind does not exist and is illu-
sory on the ultimate level, it should not be abandoned. If it were both illusory and not benefi-
cial, only then should it be abandoned. But the mind is illusory but indeed has the qualities
that benefit oneself and others and so should not be abandoned. It is like a buddha’s emanation
body (nirmāṇakāya), which is illusory on the ultimate level but beneficial for everyone. In
order to nail down this point, he states:
It is indeed not the case that existence alone is reason for accepting the mind, nor is
it the case that nonexistence is reason for abandoning it.
In this regard, he provides examples of the stances of śrāvakas, yogins, and buddhas stances
regarding the acceptance of nonexistence:
Śrāvakas: That is to say, although Śrāvakas, in embracing the idea of truly existing entities,
abandon the first two truths (i.e., the truth of suffering and that of origination), they
accept the third (i.e., the truth of cessation), which is not an entity.
Yogins: And yogins (i.e., bodhisattvas), whose minds are filled with compassion, create
various emanations – elephants and the like – for the sake of beneficial activities
for other beings even though the latter are illusory.
Buddhas: Likewise, it is precisely admitted, though they have already abandoned all error,
that the Fortunate Ones – who have mastered the means of bringing forth infinite
benefit and bliss for others by depending on great compassion, who have totally
dispelled the darkness of delusion by means of the most eminent light of wisdom,
and who wish benefit for others – will remain until the end of sentient beings, of
their karma, of their defilements, and of space, by virtue of the power of aspiration
in their past lives.17
Abhayākaragupta then adds the following comment, in which he makes reference to the pan-
theon of tantric buddhas such as Saṃvara and others, who continue to be active in the world.
Even though in Jambudvīpa there are no disciples of the Buddha in Śākyamuni’s

form (i.e., because he had gone long before), the Fortunate One remains even today
in the form of Saṃvara and so on, manifests to bodhisattvas who have accumu-
lated wholesome accumulations over many kalpas, and who are capable of teach-
ing the profound new method (i.e., the tantric path), and he still teaches them his
teachings.18
624
Abhayākaragupta
Arguing the possibility of saving infinite beings, he continues:
Objection: Saving all beings is impossible because they are limitless.

Answer: Precisely for this reason, buddhas’ compassionate thoughts toward them are limit-
less. Otherwise their activity would not accord with their affirmation (i.e., their commit-
ment to save all beings), for they promise release from suffering for all beings at the very
moment when their resolve to attain enlightenment first arises.19
He then argues that bodhisattvas’ gnosis, which is illusory but beneficial, stems from
their accumulations of merit and wisdom. Likewise, the non-abiding form of nirvāṇa
(apratiṣṭhānanirvāṇa) stems from a buddha’s compassion and wisdom. Given their compas-
sion, buddhas do not completely enter nirvāṇa, and given their wisdom, they are not defiled in
saṃsāra (cf. Ms. fol. 66v3–67r1).
He goes on to explain that bodhisattvas bring forth great fruit because – for instance, in the
case of the perfection of offering (dānapāramitā) – they understand the absence of any intrinsic
nature within offered materials, such understanding removing any possible attachment to them
and being supported by the realization of emptiness in a meditative state (cf. Ms. fol. 67r1–v2).
From there he moves on, in the last portion of chapter 1, to discuss establishing the truth
of the single vehicle.
Thus a summary of the text portion, which represents one of the highlights of this work’s
philosophical discourse. As with other sections of the work, in large part, Abhayākaragupta
arranges and contextualizes Kamalaśīla’s passages so as to fit into his own flow of thought and
occasionally interjects statements of his own into them.
Among his own foci are topics that may motivate readers to take up the practice, such as the
beneficial aspects of the mind, rather than proofs of selflessness or disproving wrong views.
These more polemical tasks are, according to him, the concerns of another work, the Flower
Cluster of the Middle Way. His choice of passages serves the specific purposes of the Adorn-
ment of the Sage’s Thought: chapter 1 deals with the doctrinal foundation for the series of cul-
tivation practices dealt with from chapter 2 onward. He homes in on passages from Kamalaśīla
that motivate readers to undertake such practices.
Buddha-Nature
Another doctrine which propels one towards practice is that of Buddha-nature. Here I shall
just briefly summarize Abhayākaragupta’s understanding of Buddha-nature (for the details,
see Kano 2016, 108–23).
To show that all sentient beings, whatever vehicle they may initially adopt, can attain the
Mahāyāna awakening in line with the single-vehicle (ekayāna) theory, Abhayākaragupta states in
chapter 1 of Adornment of the Sage’s Thought that the Buddha taught that “every sentient being
has Buddha-nature (tathāgatagarbha)”; that is, all sentient beings are able to reach the state of
perfect awakening. The term tathāgata, the first member of the compound tathāgatagarbha, refers
to the sphere/element of reality (dharmadhātu), namely the natural luminosity which pervades
all sentient beings in the form of the twofold selflessness (cf. Ms. fol. 69r4–5). Abhayākaragupta
here takes tathāgatagarbha in the sense of dharmadhātugarbha (a term also found in the
Śrīmālāsūtra), and this dharmadhātu is characterized by the selflessness of persons and phenom-
ena ( pudgaladharmanairātmyalakṣaṇasya dharmadhātoḥ). These words are strikingly simi-
lar to Kamalaśīla’s statement in his Madhyamakāloka, where again the single-vehicle theory is
expounded (D 3887, 242b4–7).
625
Kazuo Kano
In chapter 3 of Adornment of the Sage’s Thought, Abhayākaragupta states (cf. Ms. fol. 85v1):
“The dharmadhātu has precisely as its characteristic all dharmas’ absence of an own-being.”
He then roughly synonymizes the dharmadhātu with tathāgatagarbha and gotra (Ms. fol.
85v2: idam eva dharmatātmakaṃ gotraṃ tathāgatagarbha ity uktaṃ). Further on, he poses a
question and answers it:
Question: Now, how can the dharmadhātu, which is without an own-being, be a base
( pratiṣṭhā) that gives rise to the resolve to become a buddha (cittotpāda)?
Reply: This is just like space (ākāśa), which is without an own-being, being a base for the
rays of the moon and sun, and for the removal of darkness as well.
(Ms. fol. 86v1)
Abhayākaragupta rehashes this passage from Ratnākaraśānti’s Quintessence (Sāratamā)

(D 30a4–5), with the slight modification of arūpin (gzugs med pa) into niḥsvabhāva. He claims
that just as space or the sky (ākāśa), which lacks an intrinsic nature (niḥsvabhāva), functions
as a base for removing darkness, so too the dharmadhātu, with no nature of its own either,
serves as a base for cittotpāda.20
Abhayākaragupta does not accept the dharmadhātu (or Buddha-nature) as a cause
which engenders something ( jananahetu), since it has been categorized as unconditioned
(asaṃskṛta), something which cannot produce anything. Instead, he takes it as a supporting
cause ( pratiṣṭhāhetu). In this way, Buddha-nature can function as a cause, even if it lacks a
nature of its own for engendering something.
In the Moonlight on Crucial Points (D 3805, 49b4), in the commentary on the Adornment
of Realization I.37–39, we find a similar idea, Buddha-nature (synonymized with prakṛtistha-
gotra and dharmadhātu, again characterized by the absence of an intrinsic nature within all
phenomena) being explained as a support or cause of bodhicitta (Kano 2016, 120).
In chapter 1 of his tantric work, the Flower Cluster of Transmissions, Abhayākaragupta states:
“If the minds of these sentient beings were not luminous by nature, they, like charcoal, would
not become immaculate even when the obscuration in them is washed out” (Skt. Ms. 19r3,
Tib. D 7b4–5). He then quotes Mahāyānasūtrālaṃkāra IX.37 and concludes that the mind of
every sentient being is luminous by nature, tathatā pervading everything without exception.
A similar idea is found in his Pañcakramatātparyapañjikā Kramakaumudī on Pañcakrama 2.1.
When explaining the word kulam, “buddha-family,ˮ in Emergence from Sampuṭa I-i, again
in Flower Cluster of Transmissions chapter 1, Abhayākaragupta has occasion to repeat that
every sentient being has Buddha-nature. That Buddha-nature lacks any intrinsic nature is espe-
cially suggestive of the doctrinal position of Mādhyamikas, which he espouses:
Here, indeed, everything has the mind as its foremost element. And sentient beings
characterized by the mind continuum have a tathāgata within/have Buddha-nature
(tathāgatagarbhāḥ) – have as their nature a state of absence of any intrinsic
nature (naiḥsvābhāvyātma-).21
This is a passage borrowed from Ratnākaraśānti’s Commentary on the Manual of Maṇḍala-

Practice of the Guhyasamājatantra (Guhyasamājamaṇḍalavidhiṭīkā),22 where the phrase
“state of absence of any intrinsic natureˮ is not found, for Ratnākaraśānti does accept the real
existence of the mind in the form of “illuminationˮ ( prakāśa), which has an intrinsic nature.
The bold-faced words (i.e., naiḥsvābhāvyātma-) were inserted by Abhayākaragupta in order to
show that everything is just illusory in accordance with his Madhyamaka position.
626
Abhayākaragupta
In sum, Abhayākaragupta utilizes the Buddha-nature doctrine to establish the single-vehi-

cle doctrine, which teaches that all sentient beings can and will attain Mahāyāna awakening.
All along, he faithfully follows Kamalaśīla’s position. His interpretation of Buddha-nature
as being devoid of any intrinsic nature (niḥsvabhāva) is also paralleled in Kamalaśīla’s
Madhyamakāloka. His explanation of the dharmadhātu’s causal potency is drawn from
Ratnākaraśānti’s Sāratamā. He sometimes borrows passages from Ratnākaraśānti’s works that
deal with Buddha-nature and related topics and carefully modifies the borrowed passages to
show Buddha-nature as lacking an intrinsic nature.
Abhayākaragupta’s discussion of Buddha-nature had an impact on later Indian masters,
including Daśabalaśrīmitra, who utilizes the discussion of the three states of tathatā; Ratnarakṣita,
who further delves into the association of hetutantra with Buddha-nature; and Jayānanda, who
probably makes use of the single-vehicle discussion in the Adornment of the Sage’s Thought.
The role played by Abhayākaragupta, then, was to pick up on his predecessors’ discussions of
Buddha-nature, to “Madhyamakanizeˮ them, and to pass them on so altered to later generations.
His analysis of Buddha-nature accords with that of the mind (see previously), in which he
stresses the beneficial aspect of the mind (i.e., for the sake of others), even though the mind
itself is illusory from the viewpoint of the ultimate. Buddha-nature, too, is illusory (i.e., is
devoid of an intrinsic nature) at the ultimate level, but it is beneficial for guiding others to the
Buddhist path and for motivating them to undertake its practices (Kano 2016, 32–33).
Notes
1 Colophons of three of his works provide three dates within his lifetime, those of the Abhayapaddhati,
Munimatālaṃkāra, and Āmnāyamañjarī mentioning the dates of completion as the twenty-fifth year
(1103/1104), the thirtieth year (1108/1109), and the thirty-seventh year (1115/1116) of Rāmapāla’s
enthronement, respectively. The dPag bsam ljon bzang of Sum pa mkhan po gives the dates of his
birth and death as corresponding to 1064 and 1125. Regarding them, Isaacson and Sferra (2019, 249)
state that “these dates need not be given particular authority, but they are consistent with those given
in the three works mentioned previously for their completion.ˮ Isaacson and Sferra (2019, 249–51)
and Luo (2020) offer the most recent overviews of Abhayākaragupta’s life and works.
2 The term mkhan po in this passage is frequently translated as “abbot.” An exact equivalent to it, how-
ever, is not attested in the India materials; that is, there is no evidence that Buddhist monasteries in
India had what could be called mkhan pos. The only references to such personages occur in histories
of Indian Buddhism authored by Tibetans. See Kano 2017.
3 Erb (1997, 28) estimates approximately 135 works.
4 The lone Sanskrit manuscript consists of 202 palm leaves; its Tibetan translation, Tōhoku no. 3903
in the Derge Tanjur, covers some 220 folia. Hirofumi Isoda has studied the Munimatālaṃkāra inten-
sively on the basis of the Tibetan translation and has published editions of parts of it. Recently, Kazuo
Kano and Xuezhu Li have started to edit the Sanskrit original and already have published several
portions of the text.
5 See Mahāyānasūtrālaṃkārabhāṣya on chapter 19, verses 59–60.
6 See Haribhadra, Abhisamayālaṃkārālokā 401.25–402.24.
7 See Nāgārjuna’s Ratnāvalī, Mahāyānasūtrālaṃkāra, Vasubandhu’s Vyākhyāyukti, Īśvara’s
Abhidharmadīpa, Bhāviveka’s Madhyamakahṛdaya and its autocommentary (Tarkajvālā), Haribhadra’s
Abhisamayālaṃkārālokā, and Prajñākaramati’s Bodhicaryāvatārapañjikā. For details of these refer-
ences, see Fujita 2011, 115–16. Abhayākaragupta extracts selected passages on the scriptures of the vari-
ous schools from Bhāviveka’s Tarkajvālā (D 3856, 170a6–180a2; I owe this to Vincent Tournier) and a
passage on the definition of the Buddha’s words from an unknown work called skyabs gsum gyi brtag pa
attributed to Vasubandhu (cf. D 3811, 270b4–271a1) in the Adornment of the Sage’s Thought (Skt. Ms.
fol. 8v2–9r4 and 148v5–r1, respectively).
8 The third verse in closing verses.
9 Skt. Ms. 58r5–v1, Tib. 138r1–2.
10 Kano and Li 2014, 20–21; Ms. 69v1–3; D 150b4–151a1.
627
Kazuo Kano
11 Some citations from the Adornment of the Sage’s Intention in sections “The Two Truths and the Mind-
Only Doctrineˮ and “The Illusory but Beneficial Mindˮ have also been translated from the Tibetan by
Kapstein 2001.
12 These topics are treated in the Sanskrit manuscript, fol. 62v5–64r2. For the Sanskrit text and a Japa-
nese translation, see Kano and Li 2018.
13 See Kano and Li 2018, 13; Ms. fol. 63v4.
14 See Kano and Li 2018, 42; Ms. fol. 63v2–3.
15 See Isoda 1993. The summary in this section is based on the Sanskrit manuscript fol. 64r2–67v2.
16 See Kano and Li 2021, 14–15; Ms. fol. 65v2–3.
17 See Kano and Li 2021, 17; Ms. fol. 66r5–v1.
20 To illustrate Buddha-nature, RGV I.52–63 compares it to ākāśa as being itself a foundation which
does not admit of any subdivision.
21 Skt. Ms. fol. 54r5–v1; D 1198, 7b3–6; cf. Saṃpuṭa 1.1.
22 Maṇḍalavidhiṭīkā, D 1871, 80r6–v1 on Guhyasamājamaṇḍalavidhi v. 94.
Bibliography
Erb, Felix. 1997. Śūnyatāsaptativṛtti: Candrakīrtis Kommentar zu den „Siebzig Versen über die Leerheit
“des Nāgārjuna [Kārikās 1–14]. Tibetan and Indo-Tibetan Studies, no. 6. Stuttgart: Franz Steiner
Verlag.
Fujita, Shōdo. 2011. “Daijōbussetsuron no ichi danmen – Daijō sōgon kyōron no shiten kara 大乗仏説
論の一断面―『大乗荘厳経論』の視点から.ˮ Sirīzu daijō bukkyō 1 Daijō bukkyō towa nanika シ
リーズ大乗仏教1 大乗仏教とは何か edited by Jikidō Takasaki et al., 113–50. Tokyo: Shunjūsha.
Isaacson, Harunaga, and Francesco Sferra. 2019. “Indian Tantric Authors: Overview.” In Brill’s Ency-
clopedia of Buddhism, Vol. 2: Lives, edited by Jonathan Silk et al., 228–60. Leiden and Boston: Brill.
Isoda, Hirofumi. 1993. “Abhayākaragupta to Madhyamakāloka AbhayākaraguptaとMadhyamakāloka.ˮ
In Indo bukkyōgaku kenkyū: Miyasaka Yūshō hakase koki kinen ronshū jō インド学密教学研究宮
坂宥勝博士古稀記念論文集上, 501–16. Kyoto: Hōzōkan.
———. 1998. “Haramitsu no jissen – Abhayākaragupta Munimatālaṃkāra dai ni shō 波羅蜜の実践 –
Abhayākaragupta『Munimatālaṃkāra』第二章 –. ˮ In Indo mikkyō no keisei to tenkai: Matsunaga
Yūkei kokikinen ronshū インド密教の形成と展開：松長有慶古稀記念論集, 287–304. Kyoto:
Hōzōkan.
Kano, Kazuo. 2016. Buddha-Nature and Emptiness: rNgog Blo-ldan-shes-rab and a Transmission of the
Ratnagotravibhāga from India to Tibet. Wiener Studien zur Tibetologie und Buddhismuskunde, no.
91. Wien: Arbeitskreis für Tibetische und Buddhistische Studien, Universität Wien.
———. 2017. “Vikramaśīla ji no rokkenmon wo meguru shiryō to sono mondaitenヴィクラマシー
ラ寺の六賢門をめぐる史料とその問題点.ˮ Indogaku bukkyōgaku kenkyū 印度学仏教学研究 65
(2): 108–14.
———. 2019. “Seshin saku juujikyōron no bonbun itsubun 世親作十地経論の梵文佚文.ˮ Indogaku
bukkyōgaku kenkyū 印度学仏教学研究 67 (2): 116–20.
———forthcoming. “Abhayākaragupta no daijō bussetsuron—Muniishusōgonron daiyonshō kara—
アバヤーカラグプタの大乗仏説論－牟尼意趣荘厳論第4章から－.ˮ Mikkyō bunka 密教文化
247/248.
Kano, Kazuo, and Li Xuezhu. 2014. “Bonbun muniishusōgon dai isshō matsubi bubun no kōtei to
wayaku—Chūgankōmyō ichijō ronshō dan no bonbun danpen no kaishū 梵文『牟尼意趣荘厳』第
１章末尾部分の校訂と和訳―『中観光明』一乗論証段の梵文断片の回収―.ˮ Mikkyō bunka
密教文化 232: 7–42.
———. 2018. “Bonbun muniishusōgon dai isshō (fol. 61r5–64r2)—Chūgankōmyō sezoku to gonzetsu
oyobi yuishinsetsu hihan kasho itsubun—梵文校訂『牟尼意趣荘厳』第一章（fol. 61r5–64r2）―
『中観光明』世俗と言説および唯心説批判箇所佚文―.ˮ Mikkyō bunka 密教文化 241: 31–56.
———. 2020. “A Survey of Passages from Rare Buddhist Works Found in the Munimatālaṃkāra.ˮ In
Sanskrit Manuscripts in China III: Proceedings of a Panel at the 2016 Beijing Seminar on Tibetan
Studies, edited by Birgit Kellner, Jowita Kramer, and Xuezhu Li. Beijing and Vienna: China Tibetol-
ogy Publishing House/Institute for the Cultural and Intellectual History of Asia.
628
Abhayākaragupta
———. 2021. “Bonbun muniishusōgon dai isshō (fol. 64r2–67v2)—Chūgankōmyō itsubun gyōja no
chokkanchi to mujishōronshō—梵文校訂『牟尼意趣荘厳』第一章（fol. 64r2–67v2）―『中観
光明』佚文・行者の直観知と無自性論証―.ˮ Mikkyō bunka 密教文化 246: 5–39.
Kapstein, Matthew T. 2001. “Abhayākaragupta on the Two Truths.ˮ In Reason’s Traces: Identity and
Interpretation in Indian and Tibetan Buddhist Thought, 393–415. Boston: Wisdom Publications.
Luo, Hong. 2010. Abhayākaragupta’s Abhayapaddhati, Chapters 9 to 14. Critically edited and translated
by Luo Hong, with a preface by Harunaga Isaacson and Alexis Sanderson. Sanskrit Texts from the
Tibetan Autonomous Region, no. 14. Beijing/Hamburg: China Tibetology Publishing House, Univer-
sity of Hamburg.
———. 2018. “The Opening and Concluding Verses of Abhayākaragupta’s Madhyamakamañjarī.ˮ
China Tibetology 2 (September): 15–23.
———. 2020. “A First Investigation of Abhayākaragupta’s Madhyamakamañjarī.ˮ Indogaku chibetto-
gaku kenkyū 24: 57–75.
Mori, Masahide. 2009. Vajrāvalī of Abhayākaragupta: Edition of Sanskrit and Tibetan Versions. 2 vols.
Tring: The Institute of Buddhist Studies.
———. 2011. Indo mikkyō no girei sekai インド密教の儀礼世界. Kyoto: Sekaisisōsha.
Seyfort Ruegg, David. 1977. “The gotra, ekayāna and tathāgatagarbha Theories of the Prajñāpāramitā
according to Dharmamitra and Abhayākaragupta.ˮ In Prajñāpāramitā and Related Systems: Studies
in Honor of E. Conze, edited by Lewis Lancaster and Luis O. Gomez, 283–312. Berkeley: Berkeley
Buddhist Studies Series.
———. 1981. The Literature of the Madhyamaka School of Philosophy in India, Vol. 7, Fasc. 1 of a His-
tory of Indian Literature. Wiesbaden: Otto Harrassowitz.
Sinclair, Iain. 2011. “Review of Vajrāvalī of Abhayākaragupta: Edition of Sanskrit and Tibetan Versions
by Masahide Mori.” Saṃbhāṣā: Nagoya Studies in Indian Culture and Buddhism 29: 94–100.
Tomabechi, Toru, and Kazuo Kano. 2008. “A Critical Edition and Translation of a Text Fragment from
Abhayākaragupta’s Āmnāyamañjarī: Göttingen, Cod.ms.sanscr.259b.ˮ Tantric Studies 1: 22–44.
629
PART 8
Modern Philosophers
MODERN PHILOSOPHERS
Indian Buddhist philosophy is typically associated with early figures such as Nāgārjuna, Vasu-
bandhu, Dignāga, and their premodern commentators. But in the nineteenth and twentieth
centuries, a number of Indian scholars and social reformers took up Buddhist thought in new
and intellectually creative ways. Using the language of “Buddhist Revival” – and thus empha-
sizing what they described as India’s glorious Buddhist past – they reformulated Buddhism as
a modern way of life appropriate for democratic, egalitarian societies. Figures such as Iyothee
Thass, Rahula Sankrityayana, Dharmananda Damodar Kosambi, P. Lakshmi Narasu, and B.
R. Ambedkar found in Buddhist traditions a universal religion, characterized by reason and
morality, that they simultaneously proposed as a national religion for India. Responding to the
pressing issues of the time – such as colonialism and independence; the formation of a new
republic; and religious, class, and political conflicts – they sought resources and inspiration
in Buddhist traditions for developing a non-sectarian morality; advocating for freedom of
thought; and critically rethinking and reconstructing society to be more equal in terms of caste,
gender, economics, and politics. The philosophers whose work is explored in this section thus
display two of the many ways in which Buddhist philosophy has been adapted to address the
challenges of modernity.
In the first chapter of this section, William Edelglass focuses on three intertwined threads
of the philosophy of religion, Buddhism as a socially engaged practice, and Buddhist political
philosophy in the writings of Bhimrao Ramji Ambedkar (1891–1956), the most prominent
of the modern Indian Buddhist figures. Ambedkar was born into an outcaste community, the
Mahars, and went on to earn multiple graduate degrees at both Columbia and the London
School of Economics, eventually becoming the first minister of law in the newly independ-
ent India and chair of the Drafting Committee of the Indian Constitution. While he is pri-
marily known as the chief architect of the constitution, Ambedkar was also a prolific author
who worked in disciplines across the social sciences and humanities, including philosophy.
According to Ambedkar’s analysis, the inequality that permeated Indian society was due in
part to the dominant forms of Indian religion; he argued for the annihilation of caste, as well as
the abandonment of the religious framework that justified caste hierarchy. However, Ambed-
kar believed religion was still important, as it could provide the basis for a sacred morality
and responsibility to others that would bind social groups together in a democratic society.
Ambedkar devoted the last decade of his life to developing a Buddhist philosophy and practice
DOI: 10.4324/9781351030908-55 633

Modern Philosophers
that would explicitly center the experience of marginalized communities and lead to a more
free, just, and equitable society.
Jay Garfield, in the final chapter, presents an overview of the philosophy of the Dalai Lama
XIV (the Ven. Tenzin Gyatso, 1935–). Characterizing the Dalai Lama XIV as a modern Indian
philosopher may not be an obvious choice. However, the Dalai Lama XIV has lived his entire
adult life in India, his own philosophical work is rooted in Indian Buddhist traditions as taken
up in Tibet, and for decades he has been in dialogue with Indian philosophical and religious
thinkers. The Dalai Lama XIV’s philosophical writings include commentaries on Indian Bud-
dhist texts, treatises in moral philosophy seeking to develop a universal ethic, reflections on
science and Buddhist metaphysics, philosophy of religion and interfaith dialogues, and works
devoted to nonviolence and political thought. As Garfield shows, these texts offer a distinctive
synthesis of classical Indian and Tibetan Buddhist philosophy with contemporary Indian and
Western thought and make the Dalai Lama XIV one of the most significant and influential
contemporary Indian philosophers.
634
39
B. R. AMBEDKAR
Justice, Religion, and Buddhist Political
Philosophy
William Edelglass
Introduction
Bhimrao Ramji Ambedkar (1891–1956) served as the first minister of law of the newly independent
India and is widely known for his work as chair of the Drafting Committee of the Indian Constitu-
tion.1 Ambedkar is also known as an ex-untouchable, a Dalit2 who founded and led numerous pro-
jects to promote social justice and political rights for subordinated communities, including schools,
publications, and political parties. In addition, he organized and led protest movements demanding
universal access to public water and an end to prohibitions against intercaste marriage and dining. For
many years, he worked as a barrister to defend the rights of low-caste Indians and negotiated with the
British and the Congress Party on their behalf. Ambedkar is also famous for his extensive education.
After completing his B.A. in Economics and Politics at Elphinstone College at Bombay University,
he was offered a scholarship to pursue graduate work at Columbia University, where he studied phi-
losophy, history, sociology, and anthropology, eventually earning two M.A. degrees and then a Ph.D.
in economics. In 1916, he went to London, where he studied law, was admitted to the bar at Gray’s
Inn, and continued his research on Indian economic history at the London School of Economics,
receiving a M.Sc. and D.Sc. in economics. Following his formal education, Ambedkar continued to
read widely and deeply in a range of disciplines, which informed his work as a professor, dean, barris-
ter, politician, publisher, activist, community leader, delegate to conferences with the British leading
up to independence, leader of the People’s Education Society with its colleges, theorist of caste and
society, and religious reformer. Ambedkar’s prodigious learning is reflected in his many writings and
speeches: his writings and speeches in English were collected and published posthumously in seven-
teen volumes that together surpass 14,000 pages (under the title Dr. Babasaheb Ambedkar Writings
and Speeches, hereafter cited as BAWS); in addition, there is a vast corpus of his writings in Marathi.
Ambedkar’s academic study and much of his writing focused on economics, sociology,
history, political science, and religious studies. His work was also informed by a deep engage-
ment with philosophy.3 During his time at Columbia, Ambedkar took three seminars with the
American philosopher John Dewey, with whom he felt a close connection, and he developed a
lifelong interest in American pragmatism. According to Arun Mukherjee,
So deeply embedded is Dewey’s thought in Ambedkar’s consciousness that quite

often his words flow through Ambedkar’s discourse without quotation marks. . . .
DOI: 10.4324/9781351030908-56 635

William Edelglass
Ambedkar not only borrowed concepts and ideas from Dewey, his methodological
approach and way of argumentation also show Dewey’s influence.
(2009, 347–48)
While in New York to receive an honorary degree from Columbia, shortly after Dewey’s death
in June 1952, Ambedkar wrote in a letter to his wife: “I owe my whole intellectual life to him”
(Stroud 2018, 64). In addition to Dewey, Ambedkar was particularly influenced by Rousseau,
Marx, and Bergson (Kumar 2015). He read widely in British, Greek, French, and German phi-
losophy, as well as American pragmatism (Rodrigues 2017, 101). He also read and responded to
the philosophical works of his prominent contemporaries in India. As a young student in India,
Ambedkar had been denied access to Sanskrit and the study of the Vedas. Later, though, he
devoted considerable attention to these Brahmanical texts, the Indian epics, and classical Indian
philosophy, especially Sāṃkhya, which he critically engaged as the philosophical foundation
of Hinduism. Ambedkar regarded philosophy and critical thought as necessary for addressing
social injustice and drew on a wide variety of thinkers in his analyses of Indian society. Thus,
in a single sentence, he can recommend both communist and Catholic texts, for example, argu-
ing in a speech at the 1943 All India Trade Union Worker’s Study Camp that everyone in the
working class should study Rousseau’s Social Contract, Marx’s Communist Manifesto, Pope
Leo XIII’s Encyclical on the Conditions of Labor, and Mill’s On Liberty (BAWS vol. 10, 110).
Ambedkar regarded contemporary Hinduism – and the Brahmanical traditions which
gave it authority – as a primary cause of social conflict and injustice in India. Moreover, he
was deeply critical of the conflation of religion and nationalism that permeated the anticolo-
nial struggle and obscured the profound inequality and social suffering fostered by India’s
caste society. Because of these views and his leading role in drafting the Indian Constitu-
tion, Ambedkar is often associated with secular constitutionalism. But his constitutionalism
was informed by a conception of political faith. This political faith includes both the trust
between citizens that is a necessary condition for democratic politics and a shared faith in a
transcendent moral good that is ultimately, according to Ambedkar, the only viable basis for
protecting the most vulnerable and marginalized in any society. While this shared faith need
not be explicitly religious, over time, Ambedkar did argue for a revolutionary reconstruction
of Indian society specifically grounded in religious transformation. He came to believe that
while change in the political and legal sphere was necessary, such change was insufficient for
the radical social transformation required to dismantle the caste structure of graded inequality
that privileged some and subordinated many. In particular, after years of reading in compara-
tive religion, he argued that Buddhism was uniquely suited to support a moral order in which
social conflict and injustice would be diminished. Drawing on classical Buddhist texts and
contemporary scholarship on Buddhism, as well as the work of Buddhist activists in India
and European philosophers and social scientists, Ambedkar developed a Buddhist philosophy
that sought to reconstruct Buddhism in ways that would address the suffering of marginalized
communities to create a more just society.
Ambedkar’s engagement with Buddhism culminated in two major events in twentieth-
century Indian Buddhism. First, on October 14, 1956, Ambedkar publicly converted to Bud-
dhism by taking refuge in the Three Jewels, accepting the Five Precepts, and making twenty-two
vows, including the vow to “endeavor to establish equality” (Ambedkar 1989, vol. 17, 529–
31). The second event was the publication, shortly after his death in December 1956, of The
Buddha and His Dhamma (hereafter BHD). In BHD and other late works, Ambedkar devel-
ops what he calls a “neo-Buddhism,” what many Ambedkarite Buddhists call “Navayāna,” or
“new vehicle,” a socially engaged Buddhism that can play a role in reconstructing the world.
636
B. R. Ambedkar
At the same time, he was developing a Buddhist political philosophy informed by the ideals
of liberty, equality, and fraternity. According to Ambedkar, only in a constitutional democracy
can liberty, equality, and fraternity all be present: liberty without limits undermines equal-
ity; valorizing equality without limits undermines liberty. Only a social order grounded in
fraternity, or the sympathies of shared ways of life, which Ambedkar comes to call maitrī (the
Sanskrit term for benevolence, one of the four divine abodes [brahmavihāras] cultivated in
Buddhist practice [mettā in Pali]), can sustain the conditions for a society that embodies lib-
erty, equality, and fraternity – or, as Ambedkar characterizes it, a just society.
Today, like some of the other figures in this volume, Ambedkar – widely referred to as
“Babasaheb” by Ambedkarite Buddhists – is venerated as a bodhisattva or even a buddha.
Ambedkar’s Philosophy of Religion

In his posthumously published Philosophy of Hinduism, Ambedkar argues that the discipline
of comparative religion has demonstrated the arbitrariness of European hierarchies of religions.
According to these hierarchies, Western “revealed religions” were superior to other religious
traditions. Comparative religion thus contributes to a move beyond Eurocentric frameworks,
making space for non-Western traditions to be recognized as equally valid religions. In the twen-
tieth century, this recognition of religious relativism was part of a larger anticolonial rethinking
of culture. At the same time, even as he recognizes the anticolonial power of religious relativ-
ism and the “capricious distinction between true and false religions based on purely arbitrary
and a priori considerations,” Ambedkar believes that comparative religion has resulted in some
misconceptions. “The most harmful one,” according to Ambedkar, is the idea “that all religions
are equally good and that there is no necessity of discriminating between them” (BAWS vol. 3,
69). In contrast, Ambedkar writes at the beginning of the book, philosophy of religion is both
descriptive and normative. It is descriptive to the degree that it addresses the teachings of a par-
ticular religion; “in so far as it involves the use of critical reason for passing judgment on those
teachings,” however, “the Philosophy of Religion becomes a normative science” (BAWS vol. 3,
5). Ambedkar argues that analyzing and evaluating religions is necessary because like any other
institution, particular religions can have beneficial or harmful impacts on individuals, groups,
and societies.
In defending the role of religion in reconstructing society, Ambedkar was informed by a
wide array of Indian and Western thinkers. These include Indian religious reformers such as
M. G. Ranade, R. G. Bhandarkar, Gopal Hari Deshmukh, Gopal Ganesh Agarkar, and espe-
cially P. Lakshmi Narasu, Mahatma Jotirao Phule, and Dharmananda Kosambi. While there
were many differences between these thinkers and Ambedkar was not in complete agreement
with any of them, he shared their view that a critical analysis and ultimately a transformation
of religion and caste would be more effective than focusing solely on transforming politi-
cal structures. He also agreed with them that religion should focus on human social rela-
tions rather than being concerned primarily with God, the soul, and life-after-death. This, he
believed, was a necessary element in the construction of a more socially just and equal India.
Ambedkar also engaged Western scholars of religion. He read Marx with care but makes a
clear departure from the Marxist view of religion as justifying inequality. Like Weber, Ambed-
kar emphasizes the ways in which religion influences society; unlike Weber or Marx, he did not
believe that religion would eventually wither away. Rather, as with Durkheim, he believed that
religion provided a foundation for the moral order and was therefore necessary for social and
political life (Omvedt 2004b). Ambedkar’s understanding of religion was deeply influenced
by Durkheim’s The Elementary Forms of Religious Life, in which Durkheim proposes a kind
637
William Edelglass
of Rousseauean “civil religion,” grounded in a universal reason and morality that could bind
communities together in solidarity. It was also informed by Dewey, who saw in civil religion
the possibilities of forging a common foundation for sustaining and supporting democratic life
and individual rights. Finally, as Aishwary Kumar has shown, Ambedkar’s mature understand-
ing of religion was deeply influenced by Bergson’s The Two Sources of Morality and Religion
(Kumar 2015), in which Bergson distinguishes between two forms of morality and religion.
One is closed and static, with rigid rules that function to support the basic needs of a commu-
nity. A closed morality and religion excludes other communities, so communities with a closed
morality and religion are always preparing for conflict with other societies or even war. In con-
trast, Bergson suggests, an open morality and religion is oriented towards progress and creativ-
ity. Open morality and religion is not exclusive to one’s own community; it is universal and
therefore includes other communities. It is oriented toward peace and solidarity instead of war.
There is an echo of Bergson’s characterization of two forms of morality and religion in the distinc-
tion between religion grounded in principles and religion based on rules and ritual that is at the heart
of Ambedkar’s philosophy of religion. In his 1936 text, Annihilation of Caste, Ambedkar writes:
A religious act may not be a correct act but must at least be a responsible act. To per-
mit for this responsibility, religion must mainly be a matter of principles only. It can-
not be a matter of rules. The moment it degenerates into rules it ceases to be religion
as it kills responsibility which is the essence of a truly religious act.
(BAWS vol. 1, 75)
His discussion of nonviolence provides a good example of what he means when he argues that
a responsible act must be a matter of principle. According to Ambedkar, the Buddha articulated
nonviolence as a way of life and a matter of principle, not a rule. The Buddha did not simply
prohibit violence but rather prohibited violence only when it was unnecessary. Thus the prohi-
bition against violence was not a rule to be followed blindly: “He enunciated it as a matter of
principle or way of life. . . . A principle leaves you freedom to act. A rule does not” (BHD, 183).
Even though the Buddha was in principle against violence, “he was also in favor of justice and
where justice required he permitted the use of force” (BHD, 184). A truly religious act, then, is
a responsible act because it requires freedom of thought, the freedom to discern if a principle
is applicable in a particular context. And because contexts are always changing, a true religion
would be one that abandoned any claim that its ideas, texts, or rituals were infallible. Instead, a
true religion would always be open to rethinking its doctrines and practices based on principles.
True religion, then, provides a framework for thinking for ourselves and attunes us to the moral
dimension of our relations with others, helping us discern what we ought to do in any particular
context. In contrast, a religion of rules – in which thoughts and actions are determined – under-
mines freedom of thought, blinds us to the needs and value of others, and creates an excessively
powerful priestly class through its emphasis on ritual (BHD, 47).
“The word ‘religion’,” Ambedkar writes, “is an indefinite word with no fixed meaning”
(BHD, 167). There are, he suggests, very different forms of religion, and they need to be dis-
tinguished. For Ambedkar, true religion is moral, social, and rational and therefore belongs
to the public sphere. This puts him at odds with much of European Enlightenment thinking,
which characterizes religion as a private and personal affair (BHD, 168). In Enlightenment
thought, the very distinction between religion and the secular was taken to be a distinction
between the private, personal realm and a realm of public reason in which autonomous agents
could argue based on principles. By situating religion in a transcendent realm that was comple-
mentary to but outside the realm of public reason, Enlightenment thinking allowed traditional
638
B. R. Ambedkar
religious belief and practice to be preserved as a personal matter alongside the public require-
ments of rationality. In Ambedkar’s view, however, a religion which is said to “transcend
reason” is nothing more than a religion of rules and rituals, governed by superstition, in which
moral responsibility and free thinking based on principles are weakened.
In contrast to religion as private and personal, Ambedkar describes true religion, what he
often refers to as “Dhamma,” as social. It is social because fundamentally it is about social
relations and morality: “Morality is Dhamma and Dhamma is Morality,” Ambedkar claims
(BHD, 172). Dhamma is not about performing rituals or a relationship with God; it is about
a moral relationship with others: “one man, if he is alone, does not need Dhamma. But when
there are two men living in relation to each other, they must find a place for Dhamma whether
they like it or not” (BHD, 168). Dhamma, or true religion, is about the moral order; it is about
responsibility to others. By yoking religion with responsibility and responsibility with prin-
ciples, Ambedkar positions religion as also about justice and universality. Thus, according to
Ambedkar, religion is not a private matter and must be evaluated in the public sphere.
In “Buddha and Future of His Religion,” published in the journal of the Maha Bodhi Soci-
ety in May, 1950, Ambedkar articulates four requirements for a suitable contemporary reli-
gion. It must (1) provide a morality to govern society; (2) be consistent with science; (3)
recognize and affirm liberty, equality, and fraternity; and (4) not regard poverty as ennobling.
In other words, it must be what Bergson would call an open religion, one which aims toward
peace between social groups. Or, in Ambedkar’s terminology, it would be a religion of the
principle, a religion of responsibility, a religion of morality. “So far as I know,” Ambedkar
writes, “the only religion which satisfies all these tests is Buddhism” (BAWS vol. 17.2, 105).
Ambedkar’s Philosophy of Hinduism

Ambedkar was born in 1891 in Mhow, Maharashtra, the fourteenth child of a military school-
teacher who belonged to the Mahar jāti.4 Mahars, the largest outcaste community in Maharash-
tra, were traditionally responsible for defending villages and apprehending criminals. Some,
like Ambedkar’s father, were able to find employment in the British army. Thus, in the hierar-
chy of jātis outside the varṇa system, the Mahars were above some other groups. But they also
had other “defiling” responsibilities, such as caring for corpses. Thus, while they were often
regarded as inherently impure, that impurity was magnified by their polluting occupations.
Their perceived impurity was regarded as contagious, and as outcastes, they were marginal-
ized in social, economic, and political life. Stigmatized by his Mahar identity, Ambedkar’s
existence was conditioned by the violence of this marginalization. At school he was unable to
sit at a desk, drink from the water fountain, use the chalkboard, or even be close to his class-
mates; teachers would not touch his papers. Despite his accomplishments in New York and
London, upon his return to India, he was still on occasion refused housing, for example, as
Mahars were often denied access to the common goods of public life, such as drinking water,
transportation, and accommodation, as well as education, and sometimes even public roads.5
The countless indignities and constant humiliation, exclusion, and discrimination resulted in
Ambedkar’s experience of himself as outside Indian social life: “I am not a part of the whole;
I am a part apart” (BAWS vol. 2, 261). Or, as Ambedkar famously told Gandhi in 1931:
“I have no homeland” (BAWS vol. 17, part 1, 53).
Ambedkar drew on his wide-ranging studies of history, economics, culture, society, phi-
losophy, and religion when analyzing Indian society, trying to understand and end the vio-
lence and social injustice that conditioned so much of his own lived experience. According
to Ambedkar’s analysis, the violence of untouchability and the unequal distribution of rights
639
William Edelglass
and privileges are inherent to the varṇa system. Ambedkar characterizes this social structure
as a form of “graded inequality” (BHD, 56). Men in the lowest groups in the hierarchy, for
example, were denied the privilege of bearing arms or accessing education, the very resources
necessary for liberation from the oppression of the caste system. Women of any caste were
also denied education; thus, like lower-caste men, they were unable to follow the Brahmanic/
Hindu ideal stages of life, which begin with the stage of the student. The superior power and
authority of some groups over others, then, was not a mere side effect of Brahmanism/Hindu-
ism, Ambedkar argues, but its essence: “The inequality preached by Brahmins was its official
doctrine. . . . It was opposed to equality” (BHD, 57–58). According to Ambedkar’s analysis,
the graded inequality at the heart of the caste structure does not result in social harmony, free-
dom, or morality. Instead, it leads to “an ascending scale of hatred and a descending scale of
contempt,” as the lower castes hate those who are above and the higher castes have contempt
for those who are below in the social hierarchy (BHD, 58). Caste thus damages those in both
the lower and the higher ranks of the social order.
Ambedkar recognizes several forces that maintain this system. First, the structure of graded
inequality ensures that most people, no matter how much they resent the rights and privi-
leges of those who occupy superior positions in the hierarchy, do not want to lose their own
rights, privileges, and power over those in subordinate positions. Second, regardless of the
violence and suffering inherent to the caste system, many Hindus believe that it is a “divinely
ordained social order” grounded in sacred Brahmanical texts. While caste hierarchy is a fea-
ture of Indian society across all religious communities – including Muslim ones, which also
sometimes contain so-called “untouchable” jātis not justified by a logic of karma – Ambedkar
emphasizes the way in which caste is legitimized by the Brahmanical understanding of karma,
a form of theodicy that explains one’s status at birth as determined by actions in previous lives.
Caste hierarchy is also codified in the Laws of Manu (Manusmṛti), which Ambedkar regards as
the most authoritative Brahmanical legal text, with its horrific punishments for those outside
the varṇa system who overstep the limits of their social condition.
Already in Annihilation of Caste, Ambedkar argues that it is not that people are somehow
wrong because they observe caste. Rather, they are simply being religious, and their religion
asks them to observe caste, with all its exclusions. Thus, it is not people but their religion
which is wrong (BAWS vol. 1, 68). Because Ambedkar believes caste to be at the heart of Hin-
duism, he argues that Hinduism cannot be reformed; it has to be rejected. Perhaps Ambedkar
may have already thought as much when he publicly burned a copy of the Laws of Manu in
1927. Indian elites in the independence movement in the late nineteenth and early twentieth
centuries often conflated Hinduism with nationalism. For Ambedkar, though, freedom required
not just independence from the British but freedom from Hinduism itself, a religion of rules
with its scriptures held to be infallible, justifying an oppressive and violent social order that
set communities against each other. Ambedkar sought to show that the Indian social order, an
order widely accepted as a natural given, was arbitrary and unjust and ought to be annihilated.
The Buddha and His Dhamma: Ambedkar’s Buddhism

In contrast to his analysis of Hinduism as a religion of rules, Ambedkar presents Buddhism
as a religion of principles committed to freedom of thought. “If a modern man who knows
science must have a religion,” he writes in the preface to BHD, “the only religion he can have
is the Religion of the Buddha” (BHD, xxv). Buddhism, in this account, is scientific because
the Buddha makes no use of miracles to teach the Dhamma (BHD, 235). In contrast to other
religions grounded in claims of divine authority or supernatural revelation, the Buddha teaches
640
B. R. Ambedkar
through rational argument, the views of one person that could and should be challenged:
“He never claimed infallibility for his message. . . . It was based on universal human experi-
ence of life in the world. He said that it was open to anyone to question it, test it, and find what
truth it contained” (BHD, 121). The Buddha, in Ambedkar’s account, emphasizes wisdom and
rationality “because he did not wish to leave any room for superstition” (BHD, 168). Seeing
Buddhism as a religion of principles, Ambedkar regards responsibility as the heart of Bud-
dhism: “morality,” in Buddhism, “has been given the place of God” (BHD, 132). Ambedkar
sees this in the Buddha’s emphasis on the Five Precepts, the Eightfold Path, and the Ten Per-
fections, which he regards as “the basis of his religion.” According to Ambedkar, this is what
distinguishes Buddhism from other religions (BHD, 152).
Ambedkar’s Buddhism, as Gitanjali Surendran shows, was preceded by numerous Indian
thinkers who argued that Buddhism was universal, rational, and ethical and an appropriate
civic religion for India (Surendran 2013).6 Ambedkar develops these ideas with a particular
emphasis on Buddhism as especially responsive to the lives of the poor and marginalized.
He rejects the idea of a stable canon understood as the true teaching of the Buddha. Instead,
Ambedkar explicitly articulates three hermeneutic principles that justify and guide his inter-
pretation, allowing him to reject doctrines he believed to be introduced by monks or Brah-
mins that are in tension with the rational, democratic, and moral teachings of the Buddha.
Because the Buddha “was nothing if not rational, if not logical,” Ambedkar argues, we can
accept what is logical and rational as the word of the Buddha. Ambedkar’s second hermeneu-
tic principle is that “anything attributed to the Buddha which did not relate to man’s welfare
cannot be accepted to be the word of the Buddha.” Finally, he insists, we need to pay attention
to the distinction between those views which the Buddha articulated with certainty and those
about which the Buddha only expressed his views tentatively (BHD, 184–85; Queen, 1996).
These hermeneutic principles, especially rationality and social benefit, are apparent
throughout BHD. In his introduction, Ambedkar presents what he views as four problems in
the traditional accounts of the life and teachings of the Buddha. They are problems because, in
his view, they do not appear to be rational or of social benefit. The first concerns the Buddha’s
motivation to renounce the world and pursue a spiritual path. According to tradition, the Bud-
dha’s renunciation comes about because, while on a chariot ride outside the palace at the age of
twenty-nine, for the first time he saw someone who was sick, someone else who was old, and a
corpse. As Ambedkar notes, the idea that as a grown man he would not have encountered sick-
ness, ageing, and death is absurd (BHD, xxix). The story that Ambedkar tells instead is one in
which Siddarth Gautama belonged to the minority anti-war view in the Sakya council, which
had resolved to wage war against the neighboring Koliyas. To avoid a social boycott against
his family or the confiscation of their lands, Siddarth offers to leave his home and renounce
the world (BHD, 17–23).
When Gautama hears that the war between the Sakyas and the Koliyas is over, he wonders
if he should return home. On further reflection, he realizes that the problem of war – which
originally motivated his leaving home – is part of a much larger problem of conflict that can
be seen between and within nations, families, and communities. But while the conflict between
nations, for example, is intermittent, “the conflict between classes is constant and perpetual.
It is this which is the root of all sorrow and suffering in the world.” Recognizing that the issue
that originally led to his departure from home is the much more pervasive problem of social
conflict, he determines to explore for himself whether there is a solution to the “root of all sor-
row and suffering” (BHD, 41).
This analysis of the root of suffering in social conflict suggests a more politically engaged
approach to the cause of suffering than we find in the canonical account of the Four Noble
641
William Edelglass
Truths, which Ambedkar raises as the second problem. He suggests that the Four Noble
Truths – that suffering is pervasive, that this suffering arises with craving, that to let go of
suffering is to let go of craving, and the Eightfold Path to letting go of craving – were not
taught by the Buddha.7 Instead, it was monks who later introduced what he regards as the pes-
simistic teaching of the Four Noble Truths, focusing on the individual’s path rather than the
social conditions of the community that, according to Ambedkar, result in the conflict that was
the Buddha’s primary concern (BHD, xxx). It is pessimistic, in Ambedkar’s view, because it
does not seem to offer a way to address the immediate sufferings caused by poverty and social
injustice. As he argues later in the text, it is not simply greed and delusion that cause suffer-
ing: “man’s misery is the result of man’s inequity to man” (BHD, 152). Ambedkar repeatedly
emphasizes that it is the social and material conditions of our lives that cause suffering and that
“the Buddha’s conception of Dukkha is material” (BHD, 265). This is why, as Ambedkar fre-
quently insists, the Buddha found poverty and destitution abhorrent and never claimed, as we
hear in Jesus’s famous Sermon on the Mount, that the poor and the downtrodden are blessed.
Instead, he encourages the moral acquisition of adequate wealth and justice in this world. Any
account that obscures or downplays the role that material and social conditions play in human
suffering – for example, by suggesting that the marginalized can achieve freedom from their
suffering through individual practice – distracts us from the actual causes of suffering and
leaves them in place.
Ambedkar’s critique of the canonical view of freedom from suffering does not mean that
individual practice is not important. Indeed, he argues that Buddhism teaches us that instead
of striving for “some imaginary heaven,” to remove our misery, “each one must learn to be
righteous in his conduct in relation to others, and thereby make the earth the kingdom of
righteousness.” For Ambedkar, this emphasis on moral cultivation is what distinguishes Bud-
dhism from other religions, particularly through the Five Precepts, the Eightfold Path, and the
Ten Perfections. These are, in Ambedkar’s words, “the basis” of Buddhism, “because they
constitute a way of life which alone can make man righteous.” This “way of life” is a form of
cultivation that begins with training the mind in order to cultivate and train our dispositions – it
requires courage to stand up and do what is right (BHD, 152–53). Morality, then, is the very
heart of Buddhism. We are called to cultivate and train our dispositions and virtues, but we are
not responsible for the suffering caused by our social conditions.
Ambedkar’s insistence on not holding individuals responsible for the suffering that arises
from their social conditions is also reflected in his approach to the third problem he proposes to
address: how to understand karma (BHD, xxxi). One conception of karma suggests that social
conditions are determined by actions in previous lives. According to this view, which Ambed-
kar associates with Brahmanism/Hinduism, there is no innocent suffering. The corollary of
this view is that the soul is reborn and can therefore suffer the consequences of past actions in
future births (BHD, 178–82). Ambedkar distinguishes this view from what he characterizes as
the Buddha’s view of karma. According to Buddhist doctrine, because there is no substantial
self, there is no being who is reborn who could suffer the consequences of actions in a past
life (BHD, 174).8 This does not mean that Ambedkar rejects karma or denies the importance
of karma for the Buddha. Indeed, he argues, “to believe that karma is the instrument of moral
order is dhamma.” For Ambedkar, the law of karma means that human actions create the moral
order. Skillful actions contribute to a better moral order, and unskillful actions lead to a worse
moral order. Whether there is a good or bad moral order, resulting in less or more conflict and
human suffering, is up to human, not divine, beings (BHD, 131–32).
The fourth problem that Ambedkar poses in his introduction is the Buddha’s understand-
ing of the proper role of monastics. “Was the object to create a perfect man?” Ambedkar
642
B. R. Ambedkar
asks, “Or was his object to create a social servant devoting his life to service of the people
and being their friend, guide, and philosopher?” (BHD, xxxi). Quoting especially from the
Dhammapada, Ambedkar notes that a Buddhist monastic – he uses the Pāli term Bhik-
khu – should cultivate virtues of temperance and wisdom. A Bhikkhu is disciplined, has
faith, integrates meditation and knowledge, lives a life of purity, and practices restraint.
Still, Ambedkar is very clear: the Bhikkhu is not an ascetic. Indeed, he regards asceticism
as pointless: “There is no escape from the world even for an ascetic.” With Marx’s famous
thesis in mind – “the philosophers have only interpreted the world, in various ways; the
point is to change it” – Ambedkar continues, “what is necessary is to change the world and
to make it better.” Rather than a futile devotion to escaping suffering as an individual, the
Buddha realizes that “it was his duty to return to the world and serve it, and not sit silent
as the personification of inactive impassivity” (BHD, 65). And the Bhikkhu is not like the
Brahmin, who is born a Brahmin and whose primary responsibility is to perform rituals.
Instead, the Bhikkhu – without caste, family, or property – is primarily interested in spread-
ing and cultivating the dhamma through teaching and conversion, thereby increasing virtue
and improving the moral order. The role of the Bhikkhu is to serve the larger community
(BHD, 226–40). The Buddha’s teaching was not oriented primarily toward the monk; it was
“principally for the laity” (BHD, 239).
In BHD, impermanence and emptiness – often regarded as the two most significant met-
aphysical concepts in Buddhism – are also interpreted in an ethical framework. Analyzing
Ambedkar’s employment of emptiness, Aishwary Kumar shows how he “imbues the word
with an ethical substance” (Kumar 2018, 187). It is ethical because, according to Ambedkar, it
describes our creaturely finitude – our bodily vulnerability, our need for sustenance and physi-
cal comfort – which he regards as the basis for the “sentiment of fraternity.” Kumar points to a
passage in The Philosophy of Hinduism where Ambedkar contrasts rational exhortations to be
kind to others because we are all children of God with his own focus on how fraternity arises.
He argues that the “condition for the growth of this sentiment of fraternity lies in sharing in the
vital processes of life. It is sharing in the joys and sorrows of birth, death, marriage and food.
Those who participate in these come to feel as brothers” (BAWS vol. 3, 64). For Ambedkar,
our creaturely finitude – named by the Buddhist terms “emptiness” and “impermanence” – is
the basis of morality because it leads us to gather and share our lives with others and culti-
vate sympathy and friendship. It is also the basis of justice, which Ambedkar, like Aristotle,
believes to be dependent on fraternity (Kumar 2015, 294).
Ambedkar’s Buddhist Political Philosophy

Ambedkar was the chief architect of the Indian constitution, which granted equal political
rights to all adult citizens. He believed equal rights enshrined in the law were necessary for
justice. Ambedkar was committed to the values of a modern, liberal democracy, values he
worked hard to write into the constitution. He affirmed the citizen virtues and individual rights
that enable shared democratic life. And yet, while “the idea of making a gift of fundamental
rights to every individual is no doubt very laudable,” Ambedkar writes, “the question is how to
make them effective.” Ambedkar notes that it is widely held that rights are safeguarded as soon
as they are written into law. But, he argues, experience shows otherwise: “rights are protected
not by law but by the social and moral conscience of society.” If the majority of a community
opposes these rights, they cannot be guaranteed by law, legislators, or judges. Of what use,
he asks in 1943, are fundamental rights to Blacks in the United States or to Untouchables in
India? (BAWS vol. 1, 222) Any democratic constitution, even while affirming noble ideals,
643
William Edelglass
can function to suppress the lives of minorities and the poor. Therefore, to really achieve the
kind of equality necessary for justice and democratic life, one needs to pay attention to and
work at the level of social relations. Ambedkar clearly states this view in a speech on All-India
Radio in October 1954: “Indians today are governed by two different ideologies. The political
ideal set out in the Preamble to the Constitution,” a preamble written by Ambedkar himself,
“affirms a life of Liberty, Equality, and Fraternity.” The problem, though, is that “their social
ideal embodied in their religion denies them” this life (BAWS vol. 17.3, 503).
Earlier in that same speech, Ambedkar insists that he has not borrowed his political phi-
losophy – at the heart of which are democracy and the ideals of liberty, equality, and frater-
nity – from French Enlightenment thought, as listeners might expect. Rather, he says, “My
philosophy has roots in religion and not in political science. I have derived them from the
teachings of my Master, the Buddha” (BAWS vol. 17.3, 503). We may be tempted to dismiss
this claim, as Ambedkar had been invested in the political ideals of liberty, equality, and frater-
nity long before his turn to Buddhism. But, in the end, it is through Buddhism that he frames
his mature understanding of how liberty, equality, and fraternity are related and how they can
sustain a democratic way of life.
According to Ambedkar’s account, the Buddha valued liberty and equality. However, liberty
and equality are in tension. Liberty without limits, in Ambedkar’s interpretation of the Buddha,
destroys equality, as we see today with the rise of massive inequality in the wake of deregula-
tion. And equality without limits destroys liberty, as Ambedkar argues in his reflections on
Marxism. While we may appreciate and support the goal of creating equality that motivated the
Russian Revolution, he argues, “society cannot afford to sacrifice fraternity or liberty. Equality
will be of no value without fraternity or liberty” (BAWS vol. 3, 462). But there are limits to
what a constitution or the law can do to achieve equality or fraternity. In Ambedkar’s presenta-
tion of the Buddha’s views, “law had a place only as a safeguard against the breaches of liberty
and equality” (BAWS vol. 17.3, 503). The law, or the constitution, could help when liberty or
equality were transgressed here and there. But, quoting Edmund Burke, he observes that while
the law may hold an individual perpetrator accountable, it is powerless against a multitude who
choose to transgress it (BAWS vol. 1, 222). According to Ambedkar, the Buddha also “did not
believe that law can be a guarantee for breaches of liberty or equality” (BAWS vol. 17.3, 503).
As he writes in “Buddha and Future of His Religion,” “In all societies, law plays a very small
part. It is intended to keep the minority within the range of social discipline.” The majority is
constrained not by law but by morality. Thus, while the law and the constitution are important,
it is morality that is necessary to actually achieve equality. “Religion in the sense of morality,
must therefore, remain the governing principle in every society” (BAWS vol. 17.2, 104).
This form of religion as morality, Ambedkar’s Buddhism, is often interpreted as a secular,
modernist project primarily serving a political purpose.9 However, alongside Ambedkar’s dis-
course of religion as instrumental, secular, and thoroughly rational, he also presents Buddhism
as a religion of principles and responsibility that calls for an ethical response beyond rules and
social custom. Thus, Ambedkar’s Buddhism makes space for an alternative to the dominant
Indian forms of modernity. As Debjani Ganguly argues, in making this space, and thus making
Indian modernity less coercive, Ambedkar’s modernity “dares to speak the non-sociological,
non-secular language of transcendence” (Ganguly 2002, 344). For even as Ambedkar’s Bud-
dhism constitutes a kind of critique of Brahmanism/Hinduism – and, one could argue, of much
previous and existing Buddhism as well – it is also, as Ajay Skaria has argued, a critique of
“secularism, and that criticism is articulated moreover as a religion” (Skaria 2015, 451). This
critique gestures toward a responsibility to others, an attentiveness to the vulnerability and
need of embodied life that transcends rules, social custom, and natural relations and thus also
644
B. R. Ambedkar
instrumental rationality. As Aishwary Kumar notes, for Ambedkar, secularism as a political

program or social condition does not offer a responsible and just alternative to religious fanati-
cism and the oppression of religious minorities. According to Kumar’s reading of Ambedkar’s
critique,
at [secularism’s] empty center is not the truthful equality of faiths but an unraveling
project to manage the insurrectionary force of the outnumbered minority, held in its
place by a peculiarly modern pact between constitutional restraint in the political
realm and fanatical populism in the social.
(2018, 156)
Thus, while Ambedkar is deeply committed to liberal democracy, he is critical of a purely

secular modernity for, as he claims in BHD, “the only way to put a stop to conflict is to have
common rules of morality which are sacred to all.” Without a sacred morality, without a moral-
ity grounded in a force that is more than secular, “the privileged remain privileged. . . . This
means that there can be liberty for some, but not for all. This means that there can be equality
for a few, but none for the majority.” The remedy, Ambedkar argues, is fraternity, which he
sometimes refers to as maitrī, the Buddhist term for benevolence, care, friendship, and kind-
ness to others. Maitrī, or fraternity, he argues, “is another name for morality. This is why the
Buddha preached that Dhamma is morality; and as Dhamma is sacred, so is morality” (BHD,
173). Because law is secular, Ambedkar suggests, there can be social norms that transgress it
with impunity. “Law is secular, which anybody may break,” Ambedkar insists, but “fraternity
or religion is sacred which everybody must respect” (BAWS vol. 17.3, 503). There is a neces-
sary power to religious ideals that, Ambedkar believes, is beyond anything a purely secular
ideal can offer. This is why, as Kumar argues, “it is impossible to ignore the religious founda-
tion of his political thought and practice” (2015, 32).
True religion, then, is a religion that supports fraternity, that inculcates a way of life that
is attentive to the vulnerability and need of others and responsible to ideals of justice. It is a
religion that therefore creates the conditions for equality and freedom. This is why Ambedkar
argues, sometimes quoting Burke, that “religion is necessary for a free society” (BAWS vol. 1,
76; BAWS vol. 3, 442). By a “free society,” Ambedkar means democracy. But,
a democracy is more than a form of Government. It is primarily a mode of associated

living. The roots of Democracy are to be searched in the social relationship, in the
terms of associated life between the people who form a society.
(BAWS vol. 17.3, 519)
How would such a society be described? It is one characterized by cooperation, solidarity,

mutuality of sympathy, common purpose, a sense of belonging to a common community.
Brahmanism/Hinduism, in Ambedkar’s view, with its caste system, obstructs the very social
foundations necessary for a democracy of liberty, equality, and fraternity. The Buddha, on the
other hand, Ambedkar writes, “was born a democrat and he died a democrat . . . he was a thor-
ough equalitarian” (BAWS vol. 3, 451–52). And the Buddha’s teachings support democracy
because Dhamma, in his view, “breaks down barriers between man and man” (BHD, 161).
Democracy requires a society in which the minority is not excluded. A parliamentary
democracy may enshrine individual political rights in a constitution and ensure freedom of
contract, but such a democracy could also maintain social and economic inequality. Instead,
for Ambedkar, citizens in a democracy should have the freedom and power to make society
645
William Edelglass
more just. Contrasting his own definition with Lincoln’s famous line – that democracy is “a
government of the people, by the people, and for the people” – Ambedkar defines democracy
as “a form and a method of government whereby revolutionary changes in the economic and
social life of the people are brought about without bloodshed” (BAWS vol. 17.3, 475).10
Conclusion
During a conference in October 1935 at Yeola, located in the Nashik district of Maharashtra,
Ambedkar famously declared that while he may have been born a Hindu, “I solemnly assure
you that I will not die a Hindu” (BAWS vol. 17.3, 95). According to Ambedkar, freedom from
untouchability required the annihilation of caste, and because caste was at the heart of Brahman-
ism/Hinduism, it meant an emancipatory political choice of leaving the religion into which he had
been born.11 On October 14, 1956 (a day associated with Emperor Aśoka’s conversion to Bud-
dhism), in Nagpur (a city associated with Buddhist history), Ambedkar converted to Buddhism.
His conversion was witnessed by almost half a million Mahars, who, that weekend, also converted
to Buddhism. Within a short time, the vast majority of Mahars had converted to Buddhism, and
today there are millions of Indians who think of themselves as “Ambedkarite Buddhists.” Statues
of Ambedkar can be seen in the private homes of many Dalits and in public places across India.
Despite his success in the conversion of millions of Dalits, critics have pointed out that the
universal religion and morality that Ambedkar sought to achieve has not been realized: while
Ambedkar’s goal was a universal maitrī that could overcome divisive communal identities,
Neo-Buddhism today is very much associated with Dalits and thereby seen by some as replicat-
ing casteist thinking (Fuchs 2001, 265). While Ambedkar never argued that all forms of Bud-
dhism, or even any actually existing form of Buddhism, was a true religion of principles that
could sustain a moral social order, recent Buddhist atrocities and genocide against Rohinga in
Myanmar and against Tamils in Sri Lanka, the expulsion of so many Nepali-speaking Hindus
in the name of a nationalist Buddhist society in Bhutan, the violent ideology of Imperial Zen
in Japan, or any number of other historical instances of Buddhist violence against outgroups
should give us pause in thinking that mass conversions to Buddhism would somehow create
a shared sympathy across social groups that results in equality, liberty, and justice. In the end,
then, we are still left with questions about shared life in democratic societies: How do we culti-
vate fraternity when we belong to different social groups? How do we live in a pluralistic soci-
ety that is not held together by one established religion? Or, if shared religious commitments
are necessary to cultivate fraternity, how do we avoid the problem of religious nationalism?
Although Ambedkar’s project of conversion may not have led to the all-encompassing
social transformation he envisioned, it did contribute to the increased power and voice of Dal-
its, often manifest by protests in parks across India named after Ambedkar. Furthermore, while
we may be left with some of the same political questions with which he struggled, Ambedkar’s
analysis of the limits of the law and the need for cultivating fraternity in pluralistic democra-
cies is still relevant today. He is a capacious thinker who explores the intersecting dynamics
of subordination based on gender, class, caste, and race.12 And while his social context was
very much Indian society, he was a cosmopolitan thinker. As Ambedkar wrote in a 1946 letter
to W. E. B. Du Bois, “There is so much similarity between the position of the Untouchables
in India and of the position of the Negroes in America that the study of the latter is not only
natural but necessary.”13 Ambedkar’s analysis of the limits of democratic constitutions and the
law for achieving social equality and the rights of minorities drew on other social contexts and
is applicable well beyond Indian caste society.
646
B. R. Ambedkar
For Buddhists in particular, in India and elsewhere, Ambedkar provides a model of a

socially engaged Buddhism, rooted in tradition and addressing the needs of contemporary
people. His is a model of Buddhism as a way of life that works toward a more just society and
recognizes – invoking and revising Marx – that “the function of Religion is to reconstruct the
world” (BAWS vol. 3, 442). This is why the slogan most closely associated with Ambedkar is
taken from the labor movements in the United States and England: “educate, agitate, organ-
ize!” Regardless of whether we are persuaded by Ambedkar’s Buddhism, or his understanding
of the role it can play in social life, he challenges us to think and practice ways of life that can
nourish sympathy and benevolence across social differences.
Notes
1 I am grateful to Giulia Davis, Sarah Fleming, Richard Henning, Aishwary Kumar, Sara McClintock,
Meg Mott, Valerian Rodrigues, Andy Rotman, Eva Seligman, and Sonia Sikka for helpful comments
on a previous draft.
2 Ambedkar generally uses the term “untouchable” in his speeches and writings; he rarely uses “Dalit.”
In the late nineteenth century, “Dalit” – from the Sanskrit root dal in the sense of “broken” – came to
refer to groups excluded from the four Brahminic castes. While social reformers such as Savitribai
and Jotirao Phule popularized the term in the nineteenth century, it was embraced by ex-untouchables
across India in the twentieth century to affirm a positive cultural and political identity even as it simul-
taneously names the history and present of marginalization. This positive identity enacts a refusal of
the condition of untouchability, with its subordination and exclusions, and thus contests the power
of upper-caste and upper-class Indians across Indian society. In recent decades, some have favored
the term “Bahujan,” or “Dalit-Bahujan.” “Bahujan” is intended to build coalitions between the most
marginalized groups, such as Dalits and Adivasis, and the many other people in contemporary Indian
political, juridical, and social life, including peasants, religious minorities, and others who find them-
selves subordinated to high-caste power.
3 As Valerian Rodrigues has argued, philosophical thinking permeates Ambedkar’s work, which
addresses issues in metaphysics, philosophy of the person, epistemology, ethics, social and political
thought, and philosophy of religion (Rodrigues 2017, 101).
4 A jāti is an endogamous group that traditionally defined social roles, occupations, identity, and sta-
tus in Indian society. There are thousands of regional jātis in India in contrast to the four transre-
gional varṇas in the Brahminic/Hindu social order: brahmins (priests), kshatriyas (warriors), vaishyas
(merchants), and shudras (servants). Under the British, jātis were associated with particular varṇas,
strengthening the varṇa system. Some jātis were regarded as outside the varṇa system and thus
were outcastes and referred to, during Ambedkar’s time, as “untouchable.” Today, many scholars use
“caste” to describe jātis and “class” to describe varṇas.
5 For more on Ambedkar’s life, see biographies by Jaffrelot (2005) and Omvedt (2004a) as well as his
own notes for an unfinished autobiography, “Waiting for a Visa,” BAWS vol. 12, 661–91
6 This approach to Buddhism was also developed elsewhere in South and Southeast Asia, where Bud-
dhism was formulated as a rational religion with a universal morality and therefore appropriate for
modernity as part of a larger anticolonial movement (Edelglass 2020).
7 Ambedkar was not alone among twentieth-century Buddhist scholars and reformers either in arguing
that the four Noble Truths were a later development or giving Buddhism a more explicitly engaged and
political interpretation. Christopher Queen, for example, has shown how Ambedkar underlined his copy
of Hermann Oldenberg’s The Buddha: His Life, His Doctrine, His Order, which also makes the argu-
ment that the four Noble Truths are not found in the earliest stages of Buddhism (Queen 2004, 140–41).
8 While the details of his argument may be different from classical Buddhist discussions, Ambed-
kar’s rejection of Brahmanical attempts to “naturalize” social status is consistent with much Indian
Buddhist thought. See, for example, Vincent Eltschinger’s work on Buddhist critiques of realist
interpretations of caste and social status from early canonical texts through Āryadeva, Vasubandhu,
Dharmakīrti, Candrakīrti, Kumārila, and Prajñākaragupta (Eltschinger 2012).
9 For critiques of these readings, see Viswanathan 1996; Ganguly 2002; Gannon 2011.
10 After I wrote this chapter, Valerian Rodrigues kindly shared with me part of his forthcoming book
on Ambedkar’s political philosophy. Rodrigues’s work, especially his chapter on “Religion and the
647
William Edelglass
Modern Public,” addresses much of what I have discussed here in greater detail and with nuance and
sophisticated analysis.
11 Religious conversion as a means for those outside the varṇa system to achieve social and political
emancipation was not unique to Ambedkar; there had been such conversions to Christianity, Sikhism,
Islam, and Buddhism in the nineteenth and early twentieth centuries. Ambedkar himself considered
these other traditions as possibilities before committing himself to Buddhism.
12 “Ambedkar’s view that the creators of the Vedic literature should be viewed as a people rather than as
a race was far ahead of his times. This view is winning approval in modern Indology” (Sharma 2005,
862).
13 www.saada.org/item/20140415-3544. Accessed January 15, 2019.
Bibliography
Ambedkar, B. R. 1982–2003. Dr. Babasaheb Ambedkar Writings and Speeches (BAWS), vols. 1–17.
Bombay: Government of Maharashtra, Department of Education.
———. 2011. The Buddha and His Dhamma: A Critical Edition. Edited, introduced, and annotated by
Aakash Singh Rathore and Ajay Verma. New Delhi: Oxford University Press.
Blackburn, Anne M. 1993. “Religion, Kinship and Buddhism: Ambedkar’s Vision of a Moral Commu-
nity.” The Journal of the International Association of Buddhist Studies 16 (1): 1–23.
Edelglass, William. 2020. “Aspiration, Conviction, and Serene Joy: Faith and Reason in Indian Buddhist
Literature on the Path.” In Asian Philosophies and the Idea of Religion, edited by Sonia Sikka and
Ashwani Kumar Peetush, 13–35. London: Routledge.
Eltschinger, Vincent. 2012. Caste and Buddhist Philosophy: Continuity of Some Buddhist Arguments
Against the Realist Interpretation of Social Denominations. Translated by Raynald Prévèreau in col-
laboration with the author. Delhi: Motilal Banarsidass Publishers.
Fiske, Adele, and Christoph Emmrich. 2004. “The Use of Buddhist Scriptures in B. R. Ambedkar’s The
Buddha and His Dhamma.” In Reconstructing the World: B. R. Ambedkar and Buddhism in India,
edited by Surendra Jondhale and Johannes Beltz, 97–119. New Delhi: Oxford University Press.
Fuchs, Martin. 2001. “A Religion for Civil Society? Ambedkar’s Buddhism, the Dalit Issue and the
Imagination of Emergent Possibilities.” In Charisma and Canon: Essays on the Religious History of
the Indian Subcontinent, edited by Vasudha Dalmia, Angelika Malinar, and Martin Christof, 250–73.
New Delhi: Oxford University Press.
———. 2004. “Buddhism and Dalitness: Dilemmas of Religious Emancipation.” In Reconstructing the
World: B. R. Ambedkar and Buddhism in India, edited by Surendra Jondhale and Johannes Beltz,
283–300. New Delhi: Oxford University Press.
Ganguly, Debjani. 2002. “History’s Implosions: A Benjaminian Reading of Ambedkar.” Journal of Nar-
rative Theory 32 (3): 326–47.
Gannon, Shane P. 2011. “Conversion as a Thematic Site: Academic Representations of Ambedkar’s Bud-
dhist Turn.” Method and Theory in the Study of Religion 23 (1): 1–28.
Gokhale, Pradeep P. 2004. “Universal Consequentialism: A Note on B.R. Ambedkar’s Reconstruction
of Buddhism with Special Reference to Religion, Morality, and Spirituality.” In Reconstructing the
World: B. R. Ambedkar and Buddhism in India, edited by Surendra Jondhale and Johannes Beltz,
Jaffrelot, Christophe. 2005. Dr. Ambedkar and Untouchability: Fighting the Indian Caste System. New York:
Kumar, Aishwary. 2015. Radical Equality: Ambedkar, Gandhi, and the Risk of Democracy. Stanford:
Stanford University Press.
———. 2018. “In the Void of Faith: Sunnyata, Sovereignty, Minority.” In Tolerance, Secularization and
Democratic Politics in South Asia, edited by Humeira Iqtidar and Tanika Sarkar, 156–90. Cambridge:
Mukherjee, Arun P. 2009. “B. R. Ambedkar, John Dewey, and the Meaning of Democracy.” New Literary
History 40 (2): 345–70.
Omvedt, Gail. 2004a. Ambedkar: Towards an Enlightened India. Haryana: Penguin Books.
———. 2004b. “Confronting Brahmanic Hinduism: Dr. Ambedkar’s Sociology of Religion and Indian
Society.” In Reconstructing the World: B. R. Ambedkar and Buddhism in India, edited by Surendra
Jondhale and Johannes Beltz, 49–62. New Delhi: Oxford University Press.
648
B. R. Ambedkar
Queen, Christopher. 1996. “Dr. Ambedkar and the Hermeneutics of Buddhist Liberation.” In Engaged
Buddhism: Buddhist Liberation Movements in Asia, edited by Christopher S. Queen and Sallie B.
King, 45–71. Albany: State University of New York Press.
———. 2004. “Ambedkar’s Dhamma: Source and Method in the Construction of Engaged Buddhism.”
In Reconstructing the World: B. R. Ambedkar and Buddhism in India, edited by Surendra Jondhale
and Johannes Beltz, 132–50. New Delhi: Oxford University Press.
Rodrigues, Valerian. 2017. “Ambedkar as a Political Philosopher.” Economic & Political Weekly 52 (15):
101–7.
Sharma, Arvind. 2005. “Dr. B. R. Ambedkar on the Aryan Invasion and the Emergence of the Caste
System in India.” Journal of the American Academy of Religion 73 (3): 843–70.
Skaria, Ajay. 2015. “Ambedkar, Marx and the Buddhist Question.” Journal of South Asian Studies 38
(3): 450–65.
Stroud, Scott. 2018. “Creative Democracy, Communication, and the Uncharted Sources of Ambedkar’s
Deweyan Pragmatism.” Education and Culture 34 (1): 61–80.
Sumant, Yashwant. 2004. “Situating Religion in Ambedkar’s Political Discourse.” In Reconstructing
the World: B. R. Ambedkar and Buddhism in India, edited by Surendra Jondhale and Johannes Beltz,
Surendran, Gitanjali. 2013. “ ‘The Indian Discovery of Buddhism’: Buddhist Revival in India, c. 1890–
1956.” PhD diss., Harvard University.
Viswanathan, Gauri. 1996. “Religious Conversion and the Politics of Dissent.” In Conversion
to Modernities: The Globalization of Christianity, edited by Peter van der Veer, 89–114. New York:
Routledge.
649
40
THE DALAI LAMA XIV
A Modern Indian Philosopher
Jay L. Garfield
Introduction
It may sound odd to some ears to represent the present Dalai Lama (the Ven. Tenzin Gyatso,
1935–) as an Indian philosopher. After all, he was born and educated in Tibet and is associated
primarily with Tibetan Buddhist philosophy. But there are several good reasons to read him as
an Indian philosopher, and these reasons also serve as a good introduction to his philosophi-
cal career. First, Tibetan Buddhist philosophy does represent a clear continuation of an Indian
Buddhist tradition, and most of the reference points even for contemporary Tibetan philoso-
phy – and particularly for the work of the Dalai Lama XIV – are Indian. That would hardly
be sufficient, though; in that case, every Tibetan philosopher would count as Indian, and the
distinctive character of Tibetan philosophy would be elided.
But when we add several other reasons, the case becomes more compelling. The second
reason is this: the Dalai Lama XIV has spent his entire adult life (and professional philosophi-
cal career) in India, after fleeing Tibet in 1959 amidst an uprising against Chinese occupation.
And that residency in India has hardly been one of isolation within the Tibetan community;
he has been in constant dialogue with Indian philosophical and religious figures and with the
wider world. Third, as a contemporary Indian philosopher, one of his most important touch-
stones beyond the classical Indian Buddhist tradition has been Mohandas K. Gandhi, one of
the most important public Indian philosophers of the twentieth century.
Finally, the approach to philosophy taken by the Dalai Lama throughout his extensive cor-
pus is utterly continuous with that of modern Indian philosophers. That is, he works self-con-
sciously in engagement with a classical Indian tradition but also in self-conscious dialogue with
contemporary philosophy derived from non-classical, non-Indian traditions, addressing philo-
sophical problems by drawing freely from both (see Bhushan and Garfield 2017). The fact that
this long-term resident of India works self-consciously in the context of recent Indian philoso-
phy and uses a methodology consonant with that of his compatriots more than justifies reading
the Dalai Lama XIV as an Indian philosopher. Indeed, the Dalai Lama XIV himself agrees with
this characterization, writing in Beyond Religion: Ethics for the Entire World (2011):
Sometimes I describe myself as a modern-day messenger of ancient Indian thought.

Two of the most important ideas I share whenever I travel – the principles of
650 DOI: 10.4324/9781351030908-57

The Dalai Lama XIV
nonviolence and interreligious harmony – are both drawn from ancient Indian her-
itage. Though I am of course a Tibetan, I also consider myself to be, in a sense, a
son of India. Since childhood my mind has been nourished by the classics of Indian
thought. . . . And since early adulthood my body, too, has been nourished by Indian
fare: rice and dal.
(11)
The Dalai Lama XIV’s corpus is vast, comprising about 50 books and countless transcribed
lectures, interviews, and teachings. It would be impossible to survey this entire body of work in
a single chapter. Instead, I will discuss the outlines of his philosophical project and views and
some of the respects in which he is philosophically distinctive. We can divide the Dalai Lama
XIV’s published work into several broad (overlapping) genres. First, he has written substantial
commentarial work on Indian Buddhist literature. Second, he has written systematic ethical
treatises aimed at a global audience. Third, he has written on metaphysics and science. Fourth,
he has written on the philosophy of religion, particularly on interfaith dialogue. Fifth, he has
written religious homilies aimed sometimes at Buddhist audiences and sometimes at more gen-
eral audiences, designed to help readers to lead happier lives. Finally, he has written autobio-
graphical texts. Here, I will address the work comprising the first four of these rough categories.
General Philosophical Framework

The overarching philosophical framework within which the Dalai Lama XIV works represents
an intersection between several philosophical traditions. First and foremost, he is a follower
of the Prāsaṅgika Madhyamaka tradition of Buddhism, a tradition grounded in Candrakīrti’s
(seventh century CE) reading of Nāgārjuna’s (ca. second century CE) Fundamental Verses on
the Middle Way (Mūlamadhyamakakārikā). Philosophers in this tradition take seriously the
ultimate emptiness of all phenomena of any intrinsic nature in virtue of their interdependence.
But at the same time, they take seriously the conventional reality of these interdependent phe-
nomena. Since this is a Mahāyāna tradition, Madhyamaka moral philosophers ground ethical
thought in the bodhisattva ideal – that is, in a commitment to attain awakening, conditioned by
the cultivation of karuṇā, an attitude of universal care, for the benefit of all sentient beings. On
the ethical side, the principal texts to which the Dalai Lama XIV refers are Śāntideva’s (eighth-
century) How to Lead an Awakened Life (Bodhicāryāvatāra) and Kamalaśīla’s (eighth-cen-
tury) Stages of Meditation (Bhāvanākrama).
The Dalai Lama XIV’s understanding of this tradition is mediated in several respects (even
prior to the interaction with the other strains of thought that combine in his philosophical out-
look). First, he reads this tradition through the extensive Tibetan commentarial tradition. Most
specifically, he is influenced enormously by the commentaries and treatises of his own lineage,
the dGe lugs (Geluk) school founded by rJe Tsongkhapa (1357–1419), a lineage that, more than
most, emphasizes a robust realism about conventional reality and the possibility of genuine
knowledge of conventional truth (Cowherds 2011; Jinpa 2019; Thakchöe 2007; Yakherds 2021).
That influence is rendered less parochial, however, in virtue of the Dalai Lama XIV’s
commitment to an ecumenical approach to Buddhist philosophy following the nineteenth- to
twentieth-century Tibetan ris med, or non-sectarian movement, and he has received instruc-
tion from many scholars belonging to lineages other than the dGe lugs tradition, often com-
menting on texts from the perspective of these other lineages. Moreover, the commitment to
the Prāsaṅgika Madhyamaka school is mitigated by the Dalai Lama XIV’s deep interest in
651
Jay L. Garfield
and sympathy with other strands of Buddhist thought, including that of the Theravāda and
Yogācāra traditions. His own Madhyamaka thought often incorporates ideas from these other
schools, something unusual in Tibetan scholarship but reflecting the kind of cosmopolitan
approach to philosophy characteristic of Indian philosophers.
This cosmopolitan sensibility also informs the Dalai Lama XIV’s philosophy of religion,
in particular his religious pluralism. I have already noted his intra-Buddhist ecumenism deriv-
ing from the Tibetan non-sectarian movement, a tendency amplified by his constant dialogue
with Buddhists of other traditions. But like Gandhi, and Ramakrishna before him, the Dalai
Lama XIV has been a constant advocate for interfaith dialogue and for pluralism in religious
practice and commitment. Given the historically missionary nature of Buddhism (a tendency
still very much in evidence), this is a remarkable posture. Not only has the Dalai Lama XIV
urged that all religions share common core commitments and that there is no single religious
view appropriate to all religions and cultures, but he has actually offered commentary on the
Christian gospels. This pluralist ecumenism is another respect in which his modernism and
cosmopolitanism are very much in evidence.
The second line of influence to note is that from Mohandas K. Gandhi, in particular Gan-
dhi’s emphasis on ahiṃsā, or avoidance of harm. Although the Dalai Lama XIV relies princi-
pally on Buddhist teachings, while Gandhi draws inspiration from a combination of Hindu and
Western pacifist teachings (Bhushan and Garfield 2017), their views converge on this point,
as well as in the conviction that ahiṃsā is first and foremost an inner state, or commitment
demanding self-cultivation, and only secondarily an external recusal from violent action. The
Dalai Lama XIV draws heavily on Gandhi’s theorization of non-harm both in his own ethical
theory and in his approach to the political campaign for greater Tibetan autonomy within the
People’s Republic of China (see especially Beyond Religion, 9).
Gandhi’s political thought, as many have noted, has a complex relationship to modernity.
On the one hand, his theological outlook and his hostility towards technology, urbanization,
and central government – ideas inspired by Thoreau, Ruskin, and Tolstoy – contribute to a
trenchant critique of modernity, most explicitly in Hind Swaraj. On the other hand, Gandhi
draws heavily on a discourse of rights and rule of law that depends on ideals of European
modernity, arguing for Indian independence, for instance, on the grounds that colonial rule
violates fundamental rights. The Dalai Lama XIV is more friendly to modern science, tech-
nology, and government but follows Gandhi on this latter route, developing a political and
ethical theory that takes rights seriously in a way unprecedented in Buddhist thought. This has
allowed him to become a potent actor on the world political stage and has allowed him to bring
Buddhist ethical reflection into dialogue with global modernity.
The Dalai Lama XIV’s modernism is also evident in his robust engagement with science and
in his scientific realism. He has a long-standing interest in physics, in the biological sciences,
and in cognitive science in particular. In keeping with the realistic spirit of the Geluk under-
standing of conventional reality, he has taken modern science to provide the best understanding
of the nature of that reality and has always insisted that where science and Buddhist doctrine
conflict, science wins. (Whether he always follows that dictum faithfully is a matter of some
controversy, but I leave that aside here.) Through a series of dialogues among senior Tibetan
scholars and scientists in which he is a regular participant, organized by the Mind and Life
Foundation, and in a series of influential books inspired by those dialogues, the Dalai Lama
XIV has argued that Buddhist philosophy entails a respect for science and that modern science
confirms many of the most central tenets of Buddhist philosophy. He has also been instrumen-
tal in revising monastic curricula, incorporating extensive education in modern science.
652
The Dalai Lama XIV
Buddhist Philosophy
Not surprisingly, a large portion of the Dalai Lama XIV’s corpus is devoted specifically to top-
ics in Buddhist philosophy. The majority of that consists in commentarial exposition of Indian
Madhyamaka texts; the remainder are independent treatises addressed predominately to popu-
lar audiences. The Dalai Lama XIV has commented on sūtras, texts addressing the graduated
path to awakening, and tantric texts. These commentaries include: The Essence of the Heart
Sūtra (2005); The Path to Enlightenment (1994a); Tantra in Tibet (2007); A Flash of Lightning
in the Dark of Night (1994b); Sleeping, Dreaming, and Dying (1997); Dzogchen: The Heart
Essence of the Great Perfection (2000); For the Benefit of All Beings: A Commentary on
The Way of the Bodhisattva (2009); and Practicing Wisdom: The Perfection of Shantideva’s
Bodhisattva Way (2014). In most of these texts, his approach eschews traditional line-by-line
commentary as the primary mode of explanation (although there is some of this). Instead,
he favors the reconstruction of a broader philosophical picture in which fundamental meta-
physical and epistemological ideas are first defended and then used as premises on the basis
of which an account of cultivation and of the good life, or of salutary meditative practices, is
defended. These commentaries are very much addressed to a Buddhist audience interested in
advice regarding spiritual practice or illumination of difficult texts in Buddhist philosophy.
The Dalai Lama XIV has returned to How to Lead an Awakened Life – a text he often
teaches in public – many times. In his discussion of this text, unlike some others, the Dalai
Lama XIV is self-consciously addressing both a Buddhist and a non-Buddhist audience, treat-
ing How to Lead an Awakened Life both as a cultivation text within the Buddhist tradition and
as a secular text, defending an approach to ethical cultivation and conduct for everyone from
premises that require no religious commitment. His books on this text include Transcendent
Wisdom (1988), A Flash of Lightning in the Dark of Night (1994b), Practicing Wisdom (2004),
A Profound Mind: Cultivating Wisdom in Everyday Life (2011), For the Benefit of All Beings:
A Commentary on The Way of the Bodhisattva (2009), Practicing Wisdom: The Perfection of
Shantideva’s Bodhisattva Way (2014), and Perfecting Patience (2018). While some of these
are traditional verse-by-verse commentaries on the text, others are more discursive, focusing
on a single chapter or set of verses, explaining their meaning in the text, their context in Madh
yamaka ethical thought generally, and their importance in moral cultivation and daily life. In
each case, the Dalai Lama XIV works to ground ethical ideas in metaphysical and psychologi-
cal claims about the nature of the person, following the lead of Śāntideva and his predecessor
Āryadeva (third century CE). And in each case, the principal link to which he draws atten-
tion is that between interdependence ( pratītyasamutpāda), the fundamentally social nature of
human beings, and care (karuṇā), arguing that a true understanding of the nature of reality and
of human life spontaneously gives rise to an attitude of care.
Religious Pluralism
One might expect the leader of a major branch of a traditionally missionary religious move-
ment to be a religious monist, to believe that the religious tradition that he leads is the one true
religion. This is indeed the case regarding most religious leaders of this stature, reflecting the
truth-claims religions make. There have been notable exceptions to this stance in the history
of Indian philosophy: Jainism is pluralistic, and the nineteenth- to twentieth-century Vedānta
sage Ramakrishna was a pluralist as well. The Dalai Lama XIV follows in this Indian religious
tradition, advocating not only tolerance of other religious beliefs but also genuine pluralism
653
Jay L. Garfield
regarding them. He has argued that each religion can legitimately be regarded as true by its fol-
lowers, that all religions have a common ethical core, that the metaphysical differences among
religions are of no great import, and that different religions are best for different individuals,
just as different medicines are appropriate to different patients and yet all aim at a cure.
The Dalai Lama XIV has brought this pluralist spirit to interfaith dialogue, including nota-
ble dialogues with the Catholic monk Thomas Merton, with a group of Jewish leaders, and
with the Anglican Archbishop Desmond Tutu. He has also defended and expressed these views
in several notable books. In The Good Heart: A Buddhist Perspective on the Teachings of
Jesus (1996), a volume arising from a teaching residency in a Benedictine monastery in Eng-
land, the Dalai Lama XIV offers commentary on the Christian Gospels. In Beyond Religion:
Ethics for a Whole World, he argues that the values advanced by religions are equally well –
or even more effectively – advanced in a secular context, radicalizing his vision of religious
pluralism to include irreligion as an option he also endorses.
In The Good Heart, the Dalai Lama XIV introduces his project by saying
I believe the purpose of all the major religious traditions is not to construct big tem-
ples on the outside, but to create temples of goodness and compassion inside, in our
hearts. Every major religion has the potential to create this. The greater our aware-
ness is regarding the value and effectiveness of other religious traditions, then the
deeper will be our respect and reverence toward other religions. This is the proper
way for us to promote genuine compassion and a spirit of harmony among the reli-
gions of the world.
(38–39)
He reinforces this strong pluralism a few pages later:
People often experience feelings of exclusivity in their religious beliefs – a feeling

that one’s own path is the only true path – which can create a sense of apprehension
about connecting with others of different faiths. I believe the best way to counter that
force is to experience the value of one’s own path through a meditative life, which
will enable one to see the value and preciousness of other traditions.
(41)
This does not amount to a universalism, according to which all religions are taken to be basi-
cally the same, albeit articulated in different ways, but a genuine pluralism, respecting the real
differences among religions, as the Dalai Lama XIV immediately emphasizes:
I have always felt that we should have different religious traditions because human
beings possess so many different mental dispositions: one religion simply cannot
satisfy the needs of such a variety of people. If we try to unify the faiths of the world
into one religion, we will also lose many of the qualities and richnesses of each par-
ticular tradition.
(Ibid.)
He makes the same point in the introduction to his exposition of the Heart of Wisdom Sūtra
(2005, 9–13) in a discussion replete with genuine appreciation for the benefits of views foreign
to Buddhism. He concludes that discussion with an admonition against proselytizing and in
654
The Dalai Lama XIV
favor of maintaining faith in one’s own tradition (13–14). He then characterizes productive
interfaith dialogue as follows:
At the beginning of such a dialogue, it’s important that all the participants fully rec-
ognize not only the many areas of convergence between each other’s faith traditions,
but more crucially, that they recognize and respect the differences between the tradi-
tions. Furthermore, we should look at the specific causes and conditions that gave
rise to the differences between the traditions. . . . Then, having clarified the differ-
ences and the origins, we look at the religions from a different perspective: becoming
aware of how different religious philosophies and practices can give rise to similar
results. By entering into interfaith dialogue in this way, we develop genuine respect
and admiration for each other’s religious traditions.
(14–15)
The Dalai Lama XIV argues that the benefits from interfaith dialogue go well beyond the
achievement of greater understanding of and respect for other religious traditions; he argues
that traditions can actually benefit from lessons learned from other faiths:
Though I don’t recommend that a person abandon his or her native religion, I believe
that a follower of one tradition can certainly incorporate into his or her own spiritual
practice certain methods for spiritual transformation found in other traditions. For
example, some of my Christian friends, while remaining deeply committed to their
own tradition, incorporate ancient Indian methods for cultivating single-pointedness
of mind through meditative concentration. . . . These devout Christians, while remain-
ing deeply committed to their own spiritual tradition, embrace aspects and methods
from other teachings. This, I think is beneficial to them and wise.
Buddhists can incorporate elements of the Christian tradition into their practice – for
instance, the tradition of community service. In the Christian tradition, monks and
nuns have a long history of social work. . . . In providing service to the greater human
community through social work, Buddhism lags far behind Christianity.
(17–18)
In The Compassionate Life, he devotes an entire chapter (chapter 4) to an exploration of reli-

gious pluralism, developing these ideas in greater detail, and arguing directly for a serious
religious pluralism that embraces difference to the benefit of all traditions.
Let us now consider one example of the Dalai Lama XIV’s approach to a Christian scrip-
ture, drawn from his commentary on the beatitudes in the Sermon on the Mount. Here we see
both his insistence on recognizing difference but also his genuine interest both in preserv-
ing that difference and in utilizing it to gain deeper insight into commonality. After pointing
out that both Buddhist and Christian philosophers see the problem of suffering as structuring
human life and that both affirm a close connection between the moral valence of one’s actions
and the nature of their consequences for one’s future life, he notes:
One of the most difficult concepts involved here, especially for Buddhists, is the
concept of a divine being, God. Of course, one can approach this concept in terms
of something which is inexpressible, something which is beyond language and
655
Jay L. Garfield
conceptuality. But one must admit that, at the theoretical level, the conceptions of
God and Creation are a point of departure between Buddhists and Christians.
(55)
The Dalai Lama XIV then uses this “point of departure” to explore the motivations for the-
ism and a theory of universal creation by an omniscient, omnipotent, omnibenevolent god in
Christianity and the motivations for rejecting such an account in Buddhism, concluding that
there are good arguments on each side. The discussion is a model of judicious pluralism in the
philosophy of religion. He returns to the question of theism later in the text, explicitly ground-
ing this pluralism in Buddhist philosophy:
All [of the Buddha’s various teachings] are aimed toward sentient beings’ diverse
mental dispositions, needs, and spiritual inclinations. And when I understand the truth
of this, I am able to truly appreciate the richness and value of other traditions, because
it enables me to extend the same principle of diversity to other traditions as well . . .
Within Christianity there seem to be a number of diverse interpretations or under-
standings of the concept of God. . . . [Q]ualities such as compassion can also be
attributed to [a] divine ground of being. Now if we are to understand God in such
terms – as an ultimate ground of being – then it becomes possible to draw parallels
with certain elements in Buddhist thought and practice . . .
We should also be careful not to reduce everything to a set of common terms so
that at the end of the day we have nothing left . . . that is distinct about our specific
traditions. . . . For example, if one were to try hard to draw parallels between Bud-
dhism and the idea of the Trinity, the first thing that might come to mind would be
the idea of the three kayas, the doctrine of the three embodiments of the Buddha:
dharmakāya, sambhogakāya and nirmāṇakāya. But . . . we should not push these
lines of comparison too far.
(72–73)
We could say much more. But it should be clear that the Dalai Lama XIV is approaching the
philosophy of religion through a sophisticated, cosmopolitan pluralism grounded in Buddhist
philosophy, that is, as a true modernist.
Ethics
The Dalai Lama XIV has written a good deal on ethics, and indeed this may be the area in
which he has made the most substantive – as well as the most public – contributions to philoso-
phy. It is also an area in which his modernism and in which his connections to Western and to
Indian philosophy are particularly evident. As I noted in the previous section, the Dalai Lama
XIV’s ethical thought is firmly rooted in Indian Mahāyāna Buddhist thought, particularly that
of Śāntideva. As a consequence, his orientation to ethical thought is structured by several
themes that run through that tradition and that are emphasized by Śāntideva. The first is that
he conceives of ethics as the cultivation of modes of perception and experience and not in the
first instance in terms of conduct or duties. A second is that he thinks about ethics in terms of a
path to cultivation of these modes of perception and so in terms of the techniques one can use
to advance ethically. Third, he conceives of the qualities one cultivates on that path in terms of
the six perfections adumbrated in the context of the bodhisattva stages (generosity, attention,
656
The Dalai Lama XIV
patience, enthusiasm, mediation, and wisdom). Fourth, he conceives the goal of the path as the
attainment of the four divine states (brahmavihāras) that are common to most Buddhist ethical
thought – friendliness or beneficence (maitrī), care (karuṇā), sympathetic joy (muditā), and
impartiality (upekṣā) – with care taking center stage.
On the other hand, there are aspects of the Dalai Lama XIV’s ethical thought that come
from other sources. For instance, his secularization of his arguments for moral cultivation and
moral education lead him to justify his ethical vision not in terms of the goal of awakening but
rather in a reasoned account of human nature and human goals. He argues that taking human
nature and human goals seriously entails adopting this moral perspective. For instance, in Eth-
ics for the New Millennium (1999), he writes:
My call for a spiritual revolution is thus not a call for a religious revolution. Nor is it a
reference to a way of life that is somehow otherworldly, still less to something magical
or mysterious. Rather, it is a call for a radical reorientation away from our habitual preoc-
cupation with self. It is a call to turn toward the wider community of beings with whom
we are connected, and for conduct that recognizes others’ interests alongside our own.
(23–24)
He amplifies this idea both throughout Ethics for a New Millennium and more recently and
with greater emphasis on the secular basis of his ethical thought in Beyond Religion: Eth-
ics for the Whole World (2011), where he cites the European Enlightenment and the French
Revolution explicitly (7). This approach reflects his ecumenism as well as his commitment to
a universal standard of rationality and to argument based on shared premises, ideas that come
both from the Indian philosophical tradition and from European modernity.
There are two other distinctive features of the Dalai Lama XIV’s approach to ethics. First,
there is his commitment to human rights. Second, there is his appropriation of Gandhi’s com-
mitment to non-harm and a representational approach to action which reflects his debt to mod-
ern Indian moral and political theory. The Dalai Lama XIV’s ethical thought is developed in
several books, prominently including The Power of Compassion (1995), Ethics for the New
Millennium (1999), The Compassionate Life (2003), The Art of Happiness at Work (2003), The
Wisdom of Forgiveness (2004), and Kindness, Clarity, and Insight (1984).
Let us begin with the Buddhist roots of the Dalai Lama XIV’s ethics. Śāntideva’s How to
Lead an Awakened Life is a systematic development of a Buddhist moral phenomenology,
characterizing vice and suffering as depending on egocentricity grounded in a primal fear
of death (Cowherds 2016; Garfield 2010/2011, 2019; Heim 2021). The path to fearlessness
and moral maturity involves the accumulation of the six perfections of generosity, attention,
patience, enthusiasm, meditation, and wisdom. Each of these is developed by Śāntideva and
articulated by the Dalai Lama XIV as a way of experiencing oneself, the world, and other
moral agents, and the progressive cultivation of these perfections leads to a more accurate,
salutary, and effective mode of engagement with the world.
One is transformed from an agent who self-reifies, locates oneself at the center of the moral
universe, sees others as mere objects, and as a consequence takes egoism to be a rational mode
of engagement into an agent who experiences herself as an impermanent interdependent par-
ticipant in a vast network of beings with whom one is co-constituted, and so one who takes the
interests of all to be equally motivating. In his own work, the Dalai Lama XIV places special
emphasis on the cultivation of patience as an antidote to anger and aversion and the meditation
on emptiness and interdependence as an antidote to egoism.
657
Jay L. Garfield
In each case, the Dalai Lama XIV argues directly that these are the only rational com-
portments towards the world. Anger and egoism presuppose an incoherent model of our
existence as entirely independent of that of others and of agency as involving autonomy
that ignores our casual dependence on our history and our environment. Egoism involves
the absurd notion that I am far more important morally than anyone else. He argues that
any rational understanding of the world reveals instead that people act due to countless
causes and conditions for which they are not responsible, and so anger is never justified.
And we are no different from others, so our own interests have no special weight in deci-
sion making.
The outcome of this path of cultivation is the achievement of an understanding of ourselves
as embedded, conventionally real, but ultimately empty persons, persons whose goals can only
be achieved through cooperation and who are happier to the degree that they are enmeshed in
bonds of friendship than if they are in competitive strife. This understanding of the world, the
Dalai Lama XIV argues, is not simply that articulated in Buddhist philosophy but is supported
by modern science, including the life sciences and behavioral sciences. As he puts it in Ethics
for a New Millennium:
[w]hen we come to see that everything we perceive and experience arises as a result
of an indefinite series of interrelated causes and conditions, our whole perspective
changes. We begin to see that the universe we inhabit can be understood in terms of
a living organism where each cell works in balanced cooperation with every other
cell to sustain the whole.
(40–41)
In Beyond Religion, he writes that
there is now a reasonably substantial body of evidence in evolutionary biology, neu-

roscience, and other fields suggesting that, even from the most rigorous scientific
perspective, unselfishness and concern for others are not only in our own interests but
also, in a sense, innate to our biological nature.
(5–6)
Understanding oneself and the world in this way, he argues, leads to a spontaneous attitude of
friendship to those with whom we share the world and upon whom our existence is depend-
ent; a care for others, whose suffering we recognize as immediately motivating; happiness at
success, no matter whose it is; and a sense that nobody’s value depends upon their relation to
us in particular. These are the four brahmavihāras, valorized in Buddhist ethics, and argued
by the Dalai Lama XIV to be the only rational attitudes towards ourselves and others. He also
develops this argument in The Compassionate Life (2003, 8–10).
This secular defense of his approach to ethics is one respect in which the Dalai Lama XIV
is developing a modernist version of Buddhist ethics. Another is his commitment to human
rights – an idea with no antecedents in Buddhist or classical Indian thought – and to an inter-
national order that guarantees them as a moral imperative. While this commitment, grounded
in Western political and moral theory in a commitment to individual autonomy and a deonto-
logical ethical framework, might appear to be in tension with a Buddhist ethical framework
committed to interdependence and phenomenological ethics, the Dalai Lama XIV sees the
framework of human rights as having a solid foundation in his Buddhist ethical framework.
658
The Dalai Lama XIV
It is, in his view, introduced as a means to extending care to others, not as an expression of
primary autonomy (see also Garfield 1998). The Dalai Lama XIV puts it this way:
A sense of responsibility toward all others also means that, both as individuals and as
a society of individuals, we have a duty to care for each member of our society. . . .
Just like ourselves, . . . people have a right to happiness and to avoid suffering.
(169)
And in Beyond Religion: Ethics for the Whole World (2011), he writes:
We have a universal declaration of human rights, and awareness of the importance of

such rights has grown tremendously. As a result, the ideals of freedom and democ-
racy have spread around the world, and there is increasing recognition of the oneness
of humanity.
(ix)
Understood this way, we ascribe rights to others as a way of ensuring public policies and inter-
personal behavior that supports the extension of benefits to them, prevents harm, and recognizes
human equality, each manifestations of the brahmavihāras. The Dalia Lama XIV makes much
the same point in Kindness, Clarity, and Insight (70 ff.) and in Perfecting Patience (2018),
where he writes, “one aspect of compassion is to respect others’ rights and others’ views” (6).
Mohandas K. Gandhi argued that the core of moral engagement lies in the conjunction of
the commitment to avoid harm, and the achievement of self-mastery (swaraj). He also insists
that public moral engagement involves the practice of representational action to demonstrate
the moral status of unjust policies (satyagraha, or insistence on the truth). A practitioner of
satyagraha not only behaves in a way that comports with ethical standards but does so in a pub-
lic way to challenge injustice and does so nonviolently. This practice hence not only constitutes
but also represents in a public forum that in which ethical conduct consists and the injustice of
the institution against which satyagraha is launched. This adds an important social dimension
to ethics, one missing in much of traditional Buddhist ethical thought, which focuses more on
the individual perfection of one who has renounced society. The Dalai Lama XIV, as I have
noted, was deeply influenced by Gandhi’s life and writings. This leads him to emphasize the
social face of ethics and the importance to ethical practice of the avoidance of harm.
A recognition of the importance of social engagement has led the Dalai Lama XIV to
become a leader – along with the Vietnamese Zen monk the Ven. Thich Nhat Hanh and the Thai
lay Buddhist activist Sulak Sivaraksa – in the Engaged Buddhism movement, a movement that
has translated Buddhist ethics into a program of social reform (Hanh 1987; Queen 2000; Queen
and King 1996). Buddhist ethics in this tradition positively enjoins social action on behalf of
the poor and the oppressed, on behalf of social justice, and for the benefit of the environment.
Engaged Buddhists reinterpret the basic Buddhist precepts, which are all expressed in negative
terms – enjoining refraining from killing; from harsh, idle, or deceptive speech; refraining from
intoxication; refraining from theft; refraining from sexual misconduct – in positive terms. They
come to be understood as enjoining the protection of life; the use of speech for social good
(including representational action); cultivating awareness of social wrongs; working actively
for the redistribution of wealth and income; and protection of the vulnerable against exploita-
tion. The Dalai Lama XIV has actively promoted this understanding of Buddhist ethics, once
again confirming both his Indian sensibility and his modernist approach to ethical theory.
659
Jay L. Garfield
Engagement with Science

The Dalai Lama XIV’s modernist approach to Buddhist philosophy has also been amply in evi-
dence in his approach to science. He has been committed to taking modern science seriously
for some time. The primary vehicle for his engagement with science has been the long series
of dialogues sponsored by the Mind and Life Institute beginning in 1987 and continuing to
the present day. These dialogues bring together groups of senior monastic scholars, including
the Dalai Lama XIV, with Western natural and social scientists and a few humanities scholars
to provide context. These dialogues have brought the Dalai Lama XIV into close interaction
with physicists, biologists, neuroscientists, and psychologists. These conversations are not
idle exchanges of interest but serious explorations of the relevance of modern science to Bud-
dhist philosophy and doctrine and of Buddhism to problems in modern science. Many have
later been distilled into books summarizing the interactions (Begley 2007; Dalai Lama 1997;
Davidson and Harrington 2002; Goleman 1990, 2003; Harrington and Zajonc 2006; Hayward
and Varela 1992; Houshmand, Livingston, and Wallace 1999; Kabat-Zinn and Davidson 2012;
Louisi and Houshman 2009; Singer and Ricard 2015; Zajonc 2004).
The Dalai Lama XIV argues that science is a partner to Buddhist philosophy, and Mad-
hyamaka philosophy in particular, at both the conventional and the ultimate levels. Follow-
ing the Geluk tradition, he is committed to the view that there is a robust sense in which
conventional truth is truth and that the conventional world is real, even if only convention-
ally. This entails that there is a determinate truth about how things are conventionally, even
if they are ultimately empty of any intrinsic nature. Moreover, in this view, all phenomena
are dependently originated and so can be explained causally and have their identities only
in a network of causal, mereological, and conceptual interdependence. As a consequence,
an analytical investigation into the composition of natural phenomena, and into their causal
relations and theoretical contexts, should lead to a deeper understanding of their conven-
tional nature. The Dalai Lama XIV is convinced that modern science provides the best
possible avenue to such analytical investigation and so into the deepest understanding of
conventional truth.
But this has implications for ultimate truth as well. Since, in this tradition, the two truths are
regarded as extensionally identical but intensionally distinct, with emptiness taken to be iden-
tical with interdependence, any investigation of interdependence is also an investigation into
the nature of emptiness, revealing just how and why things lack intrinsic identity. Moreover,
the demonstration by scientific exploration that everything in the world is in fact interdepend-
ent in these ways, and so lacking in any essence, reinforces the Madhyamaka claim that all
phenomena are empty of intrinsic nature.
In the domain of physics, this dialogue has generally focused on three implications of
quantum mechanics and general relativity. The first is thoroughgoing interdependence. This
is suggested by phenomena such as quantum entanglement. The second is lack of determinate
nature suggested by the probabilistic character of quantum measurement and the uncertainty
principle. The third is the mind-dependence of physical reality, suggested by the dependence
of the collapse of probabilistic wave functions on measurement. All of these ideas are homolo-
gous with central doctrines in Prāsaṅgika Madhyamaka metaphysics.
In the life sciences, the Dalai Lama XIV has emphasized the interdependence of living
things as demonstrated by ecology and the close connection between mental and physical
phenomena. Neuroscience, he argues, provides deep evidence of the efficacy of meditative
practice and of the link between the cultivation of awareness of one’s own cognitive states
and of the nature of mental processes and the development of salutary moral properties such
660
The Dalai Lama XIV
as care, friendliness, and lack of egocentricity. He has argued that cognitive neuroscience and
Buddhist meditative practices are a natural pairing as tools for understanding fundamental
conscious processes and affective response. He has been particularly interested in the neuro-
science of affect and has argued that results in this area demonstrate both the natural arising
of spontaneous affect as a consequence of deep insight into the mind and the benefits of posi-
tive affect for other domains of human functioning and health. This extended philosophical
engagement with science, and the epistemic authority the Dalai Lama XIV grants it, is yet
another mark of his modernism.
Conclusion
The Dalai Lama XIV is one of the most influential contemporary Indian philosophers. He
draws effectively on the classical Indo-Tibetan Buddhist tradition and on modern Indian and
European philosophy. The synthesis he achieves in metaphysics, the philosophy of religion,
ethics, and political philosophy represents a distinctive voice in contemporary philosophy: a
classically grounded modernism that brings Buddhist thought into dialogue with modern sci-
ence and philosophy.
Bibliography
Begley, Sharon. 2007. Train Your Mind, Change Your Brain: How a New Science Reveals Our Extraor-
dinary Potential to Transform Ourselves. New York: Ballantine Books.
Bhushan, N. and J. L. Garfield. 2017. Minds Without Fear: Philosophy in the Indian Renaissance. New
Cowherds. 2011. Moonshadows: Conventional Truth in Buddhist Philosophy. New York: Oxford Uni-
versity Press.
_______. 2016. Moonpaths: Ethics and Emptiness. New York: Oxford University Press.
Dalai Lama XIV. 1984. Kindness, Clarity, and Insight. Boston: Snow Lion Publications.
———. 1988. Transcendent Wisdom. Ithaca: Snow Lion Publications.
———. 1994a. The Path to Enlightenment. Ithaca: Snow Lion Publications.
———. 1994b. A Flash of Lightning in the Dark of Night: A Guide to the Bodhisattva’s Way of Life.
Boston: Shambhala Publications.
———. 1995. The Power of Compassion. New York: HarperCollins.
———. 1996. The Good Heart: A Buddhist Perspective on the Teachings of Jesus. Somerville: Wisdom
Publications.
———. 1997. Sleeping, Dreaming, and Dying: An Exploration of Consciousness. Edited by Francisco
Varela. Somerville: Wisdom Publications.
______. 1999. Ethics for the New Millennium. New York: Riverhead Books.
———. 2000. Dzogchen: Heart Essence of the Great Perfection. Ithaca: Snow Lion Publications.
———. 2003. The Compassionate Life. Somerville: Wisdom Publications.
———. 2005. Essence of the Heart Sūtra: The Dalai Lama’s Heart of Wisdom Teachings. Somerville:
Wisdom Publications.
———. 2009. For the Benefit of All Beings: A Commentary on the Way of the Bodhisattva. Boston:
Shambhala Publications.
———. 2011a. Beyond Religion: Ethics for a Whole World. Boston: Houghton Mifflin Harcourt.
———. 2011b. A Profound Mind: Cultivating Wisdom in Everyday Life. New York: Harmony Books.
————. 2014. Practicing Wisdom: The Perfection of Shantideva’s Bodhisattva Way. Somerville: Wis-
dom Publications.
——. 2018. Perfecting Patience: Buddhist Techniques to Overcome Anger. Boston: Shambhala
Publications.
Dalai Lama XIV, and Victor Chan. 2004. The Wisdom of Forgiveness. New York: Riverhead Books.
Dalai Lama XIV, and Howard C. Cutler. 2003. The Art of Happiness at Work. New York: Riverhead Books.
Dalai Lama XIV, and J. Hopkins. 2007. Tantra in Tibet. Ithaca, NY: Snow Lion.
661
Jay L. Garfield
Davidson, Richard J., and Anne Harrington, eds. 2002. Visions of Compassion: Western Scientists and
Tibetan Buddhists Examine Human Nature. New York: Oxford University Press.
Garfield, Jay L. 1998. “Human Rights and Compassion: Towards a Unified Moral Framework.” In Bud-
dhism and Human Rights, edited by Damien V. Keown, Charles S. Prebish, and Wayne R. Husted,
111–40. Richmond, Surrey: Curzon Press.
———. 2010/2011. “What Is It Like to Be a Bodhisattva?” Journal of the International Association of
Buddhist Studies 33 (1–2): 327–51.
———. 2019. “Seeing Sentient Beings: Moral Phenomenology in Śāntideva’s Bodhicāryāvatāra.” In
Readings of Śāntideva’s Guide to Bodhisattva Practices, edited by Douglas Duckworth and Jonathan
Gold, 192–208. New York: Columbia University Press.
Goleman, Daniel. 1990. “The Psychology of Meditation.” In Proceedings of the First International Con-
ference on Psychotherapy, Meditation and Health, edited by M. G. T. Kwee, 19–35. London: East-
West Publications.
———, ed. 2003. Healing Emotions: Conversations with the Dalai Lama on Mindfulness, Emotions, and
Health. Boston: Shambhala Publications.
Hanh, Thich Nhat. 1987. Interbeing: Fourteen Guidelines for Engaged Buddhism. Edited by F. Epp-
steiner. Berkeley: Parallax Press.
Harrington, Anne, and Arthur Zajonc, eds. 2006. The Dalai Lama at MIT. Cambridge: Harvard Univer-
sity Press.
Hayward, Jeremy W., and Francisco J. Varela, eds. 1992. Gentle Bridges: Conversations with the Dalai
Lama on the Sciences of Mind. Boston: Shambhala Publications.
Heim, Maria. 2021. Buddhist Ethics. New York: Cambridge University Press.
Houshmand, Zara, Robert B. Livingston, and B. Alan Wallace, eds. 1999. Consciousness at the Cross-
roads: Conversations with the Dalai Lama on Brain Science and Buddhism. Boston: Snow Lion
Publications.
Kabat-Zinn, Jon, and Richard J. Davidson, eds. 2012. The Mind’s Own Physician: A Scientific Dialogue
with the Dalai Lama on the Healing Power of Meditation. Oakland: New Harbinger Publications.
Kamalaśīla, and Dalai Lama XIV. 2019. The Stages of Meditation: The Buddhist Classic on Training the
Mind. Boston: Shambhala.
Langritangpa. 2006. Eight Verses for Training the Mind. Translated by R. Sonam and Geshe Sonam
Rinchen. Ithaca: Snow Lion.
Luisi, Pier L. 2019. Mind and Life: Discussions with the Dalai Lama on the Nature of Reality. Edited by
Zara Houshmand. New York: Columbia University Press.
Nāgārjuna. 1995. The Fundamental Wisdom of the Middle Way: Nāgārjuna’s Mūlamadhyamakakārikā.
Translated by Jay L. Garfield. New York: Oxford University Press.
Queen, Christopher S., ed. 2000. Engaged Buddhism in the West. Somerville: Wisdom Publications.
Queen, Christopher S., and Sallie B. King, eds. 1996. Engaged Buddhism: Buddhist Liberation Move-
ments in Asia. Albany: State University of New York Press.
Śāntideva. 1995. Bodhicaryavatara (How to Lead an Awakened Life). Translated by Kate Crosby and
Andrew Skilton. New York: Oxford University Press.
Singer, Tania, and Matthieu Ricard, eds. 2015. Caring Economics: Conversations on Altruism and Com-
passion, Between Scientists, Economists, and the Dalai Lama. New York: Picador.
Thakchöe, Sonam. 2007. The Two Truths Debate: Tsongkhapa and Gorampa on the Middle Way. Boston:
Wisdom Publications.
Tsong-ka-pa. 1977. Tantra in Tibet: The Great Exposition of Secret Mantra. Translated by Jeffrey Hop-
kins. London: George Allen & Unwin.
Yakherds. 2021. Knowing Illusion: Bringing a Tibetan Debate into Contemporary Discourse. New York:
Zajonc, Arthur. 2004. The New Physics and Cosmology: Dialogues with the Dalai Lama. New York:
662
INDEX
Abhayākargupta 615–29; analysis of two truths tradition 211; as phenomenological

620–2; on buddha-nature 625–7; critique of investigation into entities (dharmas) 409;
Yogācāra as realist 622; on inference 623–4; ontological teachings of the Sarvāstivāda
influences on 618; on mind-only 622–3; 180; Sarvāstivāda tradition compared to Pali
Munimatālaṃkāra (Adornment of the Sage’s Abhidhamma 149; Sthiramati on 379–80; two
Thought) 482, 615, 618–27; nature of the models for meta-cognition (svasaṃvedana)
mind 623–5; non-tantric works 615–16; on in 368; see also Mahāvibhāṣā (Large
perception 623; preserver of rare passages Commentary); Vātsīputrīya/Sāṃmitīya
617–18; reliance on Kamalaśīla 482, 523, adarśana (non-seeing): Bhāviveka on 354;
615, 618, 622, 623, 625, 627; reliance on Jñānagarbha on 453; Madhyamaka thinkers
Ratnākaraśānti 617, 618, 626, 627; on on 415; Śāntarakṣita on 453, 456; Yogācāra
scriptural authority 619–20; tantric works 56, versus Madhyamaka interpretation of 453
616–17 adarśanamātra (mere non-observation) 291,
Abhidhamma 143–59, 160–71; 296, 300n24, 300n25, 319
Abhidhammatthasaṅgaha (Compendium on adhyavasāya (determination, perceptual
the Meaning of the Abhidhamma) 149–52; judgment): Dharmakīrti on 311, 315;
Buddhaghosa on 329–30, 331; compared Dharmottara on 441–3; Jñānaśrīmitra on
to Sarvāstivāda Abhidharma 149, 166–8, 558–61, 562, 563, 568, 570; Ratnākaraśānti
172–84; Dhammasaṅgaṇī (Enumeration of on 594; Ratnakīrti on 578, 580, 582, 583–4
Phenomena) 143, 146–8, 189; discovered and agent (kārakā, kartṛ): grammatical category
taught by the Buddha 144; on emptiness 153; 407–8; merely conventional or figurative
as endless phenomenological analysis 152–6, 189, 273, 289; and mind-only 405, 458; and
338; in the Milindapañha (Questions of morality 91, 102, 657; non-Buddhist views
Milinda) 107; as ontology of ultimate reality of 414; reduced to causality 20, 107, 215,
149–52, 154, 329, 342; Paṭisambhidāmagga 273–4, 275, 276; referencing an individual
(The Path of Discrimination) 153, 157n37, cognizer 448; rejected in cognition 289, 399,
160, 330; Vibhaṅga (Analysis) 148–9, 409, 435, 532; in the tantras 49, 50, 57; see
151, 152; see also Kathāvatthu (Points of also Self
Discussion) ākāra (aspect, image in cognition, mental
Abhidharma 172–87, 188–203; image): both conceptual and nonconceptual
Abhidharmasamuccayakārikā (Stanzas of 558; Dharmapāla on 589; four aspect theory
the Compendium of Abhidharma) 191, 194, 368–9; grasper (grāhaka) and grasped
196–7, 198–9; Candrakīrti’s use of 408–12; (grāhya) 289–90, 400; Haribhadra on
criticized by Bhāviveka 349–50, 351; 505–6; Jitāri on 548–9; Jñānaśrīmitra on
criticized by Nāgārjuna 212, 213; as literature 566–70; Ratnakīrti on 579; in Sautrāntika
of sectarian Buddhism 173; Mahāsāṃghika representationalism 367; Śubhagupta on
663
Index
434–5; in Yogācāra ontology 588; Yogācāra mark of particulars 532; as only conventional
theory of 365, 367, 368, 588; see also 311; as qualifying mere entities (vastumātra)
nirākāravāda; sākāravāda 456; as requiring veridical awareness 442
ālambana (mental object, objective support) 36, Āryadeva 236–51; biography of 236–7;
435, 588 Candrakīrti on 240–1, 243–5, 406;
ālayavijñāna (container consciousness, Catuḥśataka (Four Hundred Verses) 236–7;
storehouse consciousness, store mind): as co-founder of Madhyamaka 213, 236;
Abhayākaragupta on 621–2; Asaṅga as “Consequentialist Mādhyamika” 240;
on 74, 208, 255, 256, 257, 259, 260, Dharmapāla on 240–2, 349, 362, 364, 365,
261; Dharmapāla on 367, 370, 389n17; 367; epistemic humility in 246–7, 249n24;
Laṅkāvatārasūtra (The Descent into Laṅka ethics of 242–7; misogyny of 243; “no thesis”
Sūtra) on 621; Ratnākaraśānti on 593, 594; methodology of 238–40; quietism of 238–42;
Sthiramati on 378, 379, 380; Vasubandhu on on scripture and what can be known 244–6;
279–80 tantric author 248n6
alīkākāravāda (false form, theory that images Asaṅga 252–65; Abhidharmasamuccaya
in cognition are false): in embryonic form in (Compendium of the Abhidharma)
Śākyabuddhi 400–1; targeted by Śāntarakṣita 208, 253, 255, 256–8, 261, 378, 379;
468–70; as a Tibetan doxographical Abhisamayālaṃkāra (Ornament of
classification 477n19 Realization) 65, 196, 253, 497, 616;
amanasikāra (cognitive disengagement, Bodhisattvabhūmi (Stage of the Bodhisattva)
non-mentation) 132, 489 29, 39, 256, 384; on the eight consciousnesses
Ambedkar, B. R. 635–49; as architect of Indian 255; as a founder of Yogācāra 459, 587;
constitution 635, 643; biography of 635–7; Mahāyānasaṃgraha (Summary of
on caste 639–40, 643; conception of political Mahāyāna) 208, 253, 256, 258–61; on
faith 636, 639; conversion to Buddhism 646, mental verbalization (manojalpa) 261; on
648n11; critique of Brahmanism 636, 639–40, mind only 255, 458; on non-seeing 453; on
643; Dr. Babasaheb Ambedkar Writings “nothing but presentation” (vijñaptimātratā)
and Speeches 635, 638; emphasis on moral 255, 256, 257, 259, 260, 261; on the potential
cultivation 642; interpretation of emptiness (gotra) for Buddhahood 70–1; questions
643; philosophy of religion 636, 637–9; of authorship 65, 74, 254–7; as receiving
political philosophy of 643–6; The Buddha inspiration from Maitreya in Tuṣita 64,
and His Dhamma 640–3 65, 66, 252, 253; relation to Vasubandhu
anti-foundationalism 219, 310, 465 65, 252, 253, 267–8; on store mind
antirealism 208, 236, 313, 316–18, 469, 470 (ālayavijñāna) 255, 256, 257, 258, 259; as
apoha see exclusion systematizer of Yogācāra 256; on the three
appearances (avabhāsana, nirbhāsa, natures (trisvabhāva) 255, 256, 257, 259;
pratibhāsa): as acceptable as long as not works attributed to 254–5; Yogācārabhūmi
analyzed 456; as constructed 279–81; (Foundation of the Yoga Practitioners) 85,
conventional truth as 460, 471, 484, 492, 253, 261, 355, 367
544; Dharmapāla on 363–4; as dreamlike Aśvaghoṣa 83–96; association with early
54, 505–6, 582, 588; as false 268, 272–3, Sarvāstivāda school 83–4; critique of
396–7, 506, 589, 595, 609; Jñānaśrīmitra on metaphysical entities 92–3; critique of
558, 559–61, 562–3, 568–9, 570; as mere rituals and asceticism 86–7; critique of the
appearances 601–13; as nondual 507, 544, philosopher-teacher Arāḍa 87–90; critique
573; overturned through antidotes 606, 607; of Vaiśeṣika 89–90, 93; philosophy and
Ratnakīrti on 573–7, 581; and reflexive apologetics 84–6; presenting a preclassical
awareness (svasaṃvedana) 401–2; similar Sāṃkhya path to liberation 88; refutation of
to illusions 504–5; as single but variegated the self 90–2; on the role of rational analysis
582–4; Sthiramati on 382–3, 284–5, 389n17; 87–90; Saundarananda (Handsome Nanda)
see also ākāra 83, 84, 85, 86, 87, 92; treatment of Sāṃkhya
Arcaṭa 429–30, 553–4 83, 87–8, 89, 90, 93; Śāriputraprakaraṇa
Aristotle 218, 243, 331, 523, 643 (Drama of Śāriputra) 83, 87
arthakriyā (causal efficacy, causal function, Atiśa 601–14; advocating faith-based
practical activity): ambiguity of the term Madhyamaka 605; on debate 612; Entry
395–6; appearances as having 456; as central to the Two Realities (Satyadvayāvatāra)
to Dharmakīrti’s definition of pramāṇa 310, 604, 605, 611; four great reasons proving
441; imagined entities as lacking 457; as emptiness 606–7; Great Middle Way 601,
664
Index
605, 611, 612; importance of reasoning Kamalaśīla on 487, 490; as one’s own mind
606; influence of Candrakīrti on 605, 607, 54; Śāntideva on 512–14; ultimate 610; as the
609; influence of Nāgārjuna on 604–5, ultimate nature of phenomena 53
610; Lamp for the Path to Awakening bodhisattva: Abhayākaragupta on 621, 625;
(Bodhipathapradīpa) 603, 610; path of Asaṅga on 256, 258, 260–1; Aśvaghoṣa
bodhisattva training 610–12; on reasoning on 84–90; Atiśa on 603, 605, 610, 611;
606–9; as synthesizing teachings of Candrakīrti on 405; Dalai Lama XIV on 651,
Bhāviveka and Candrakīrti 605, 608 656–67; development of ideal 22; Dignāga on
ātman see Self 305; Haribhadra on 502, 503; Jñānagarbha on
atoms (paramāṇu): in the Abhidhamma 147, 451, 459; Kamalaśīla on 483, 488, 492; path
158n39, 329, 338; in the Abhidharma 257, 184, 237, 242, 278; piṭaka (collection) 29;
409–11; borderline between matter and mind Ratnākaraśānti on 588, 595, 597; Śāntideva
379; Dharmakīrti on 316, 532; refuted by on 512–25; Sthiramati on 387
Dharmapāla 363, 365; refuted by Dignāga Brahmanism: critiqued by Ambedkar 636,
288; refuted by Jñānagarbha 460; refuted 639–40, 642, 644–5, 647n8; critiqued by
by Śāntarakṣita 467–8, 471–2, 473, 477n32; Aśvaghoṣa 84, 86; critiqued by Bhāviveka
refuted by Sthiramati 379–80; refuted by 348, 352–3; critiqued by Dharmakīrti 307–8,
Vasubandhu 270, 467; refuted in the tantras 318; critiqued by Dignāga 289, 313; critiqued
49, 52; Śubhagupta’s proof as substances by Śāntarakṣita and Kamalaśīla 428; critiqued
432–3, 345; and the three natures 299n8; as by Śubhagupta 431–2; critiqued by tantric
ultimately real 507 buddhas 48–9
authors: and Abhidharma texts 174; buddhas Buddha/buddhas: as free from appearances 611;
as 4, 9–10; commentators as philosophical as having a mind or no mind after awakening
499–500; consistency as a modern criterion 367; as omniscient, 46; as philosophers 9–10;
for attribution 27–8; and Foucault’s author tantric manifestations of 10, 46–7, 53–4;
function 3; and innovation 381–2, 500; names temporality of 30; as ultimate truth or reality
as designating not persons but workshops in the tantras 10, 46, 47, 53; see also Gotama
326; as related to authority 4, 174; texts Buddha; Siddhārtha Gautama; Tantric Buddha
as taking on the function of 4; as texts Buddhaghosa 328–45; on Abhidhamma 141,
themselves 10; traditional attributions 27–8 144, 145, 148, 189; on the aggregates 338,
Avalokiteśvara 278, 280, 346 341; on the Buddha’s omniscience 333; as
codifying Pali commentarial literature 152;
Bergson, H. 636, 638, 639 on the conventional and ultimate meaning
Berkeley, G. 317, 477n20, 477n30, 507, 531 of the Buddha’s teachings 153; hermeneutic
Bhartṛhāri 285, 287, 298, 389n20 of contextualization 329–30; on joy and
Bhāviveka 346–60; classification as a delight of reading 331, 332; on nidāna
Svātantrika 347–8; criticized by Candrakīrti (context) 332; rhizomatic presentation of the
405, 414; critique of Buddhapālita 213; path 341; on types of understanding 340–1;
critique of pudgala 197; on debate as on understanding 340–2; understanding of
an essential element of spiritual practice emptiness 153; use of the term sabhāva 154,
347; Dharmapāla’s response to 240, 189; Visuddhimagga (Path of Purification)
365–7; doxographical method 349–53; 117, 328
Madhyamakahṛdayakārikā (Verses on the Buddha Nature: Abhayākaragupta on 625–7;
Essence of the Middle Way) 346, 349, 350–6, Maitreya on 65, 66, 69–72; nine similies
384, 475; medical metaphors in 350–2; for 71; not restricted to buddhas 40;
Prajñāpradīpa (Lamp of Wisdom) 227n25, Tathāgatagarbha 74, 212, 625–7; Yogācāra on
229n42, 229n50, 346, 348, 454, 458; response 130–1
to Yogācāra critique by Dharmapāla 365–7; Buddhapālita 213, 348, 373n24, 476n4
Tarkajvālā (Blaze of Reasoning) 197, 346, buddhasaṃjñā (idea of the buddha) 34, 36,
354–5, 384, 605, 627n7; three connotations 38, 40
of ultimate reality 354, 453–4; twofold nature buddhavacana (speech/word of the/a Buddha):
of ultimate reality 355; use of autonomous as Buddhist philosophy 4, 10; as containing
(svatantra) inference 347–9; works attributed multiple voices 33; distinct from question
to 346, 458, 475 of provenance 30, 39–40; as inspired speech
bodhicitta (mind of awakening) 605; (pratibhāna) 34–5, 38, 40; and Mahāyāna
Abhayākaragupta on 618, 619, 621, 626; 212; questioned by Buddhist traditions 28–9;
as epithet for the buddha in the tantras 47; as what is well spoken (subhāṣita) 34–5, 38
665
Index
Candrakīrti 404–20; as aligned with mark of existence 270–1; Nāgārjuna’s

Buddhapālita 348, 406; on buddhas as free critique of 217; in Sarvāstivāda 177–8; see
from appearances 611; Catuḥśatakaṭīkā also dependent arising
(Commentary on the Four Hundred citta see mind
Verses) 240–1, 243–5, 406; classified as cittamātra see mind only
Prāsaṅgika-Madhyamaka 347, 476n4; on compassion (karuṇā): on the bodhisattva
conceptual dependence 416; on conventional path 513, 514, 521, 595, 610; Buddha’s
truth 474, 457; criticized by Ratnākaraśānti authority based on 305, 308, 320n5; in the
587; critique of Bhāviveka 405–6, 454; early discourses 12, 21–2; evoked through
critique of Dignāga 417–18; critique of literature in the Maṇimēkalai Tuṟavu
Yogācāra 405, 406, 458; Madhyamakāvatāra 121; indivisible from nondual gnosis
(Introduction to Madhyamaka) 405, 412–13, 57; limitlessness of 625; motivation for
455; use of Nyāya epistemology 417–18; philosophy 405, 492; nonduality of emptiness
on the path as gradual 458; Prasannapadā and 47, 51, 135; objectless 51, 515; perfected
(In Clear Words) 405–6, 407, 455; on the over lifetimes 305–8, 534; provoked by
principle of common establishment in suffering 520; required for a spiritual teacher
debate 607–8; on the qualifier “ultimately” 134; root of buddhahood 483, 619; stimulated
454; reference to non-Mahāyāna texts 190, through the cultivation of wisdom 523–5
412, 418; and Sanskrit grammar 407–8, concepts (kalpanā, vikalpa; conceptual
418; on spatio-temporally extended entities construction) 56; Dharmapāla on 367;
409–10, 416–17; on the two truths 412–17; Dignāga and Dharmakīrti on 314–15;
on two types of negation 130, 408; use of Haribhadra on 502, 505; Kālacakratantra
Abhidharma, 406, 409–11; use of Sautrāntika on 50, 59; Ratnākaraśānti on 597; ten kinds
theory of cognition 609; on what is according to Asaṅga 260; Vasubandhu on 280
acknowledged by the world 245, 458; works consciousness (vijñāna, viññāṇa): analyzed
405–7; Yuktiṣaṣṭikāvṛtti (Commentary on the in the Vibhaṅga (Analysis) 148; Asaṅga
Sixty Verses on Reasoning) 406, 411 on the Buddha’s transformation of 261;
Carruthers, P. 369–70, 373n31 as dependently arisen 389n17, 409; eight
Cārvāka (Lokāyata): critiqued by Bhāviveka forms or modes of 255, 385; as one of five
349; critiqued by Candrakīrti 414–15; aggregates 180, 268; higher-order theories
critiqued by Dharmakīrti 305–6; critiqued by of 369–70; inanimate objects and trees
Dharmapāla 365; critiqued by Śāntarakṣita lacking 51; as insubstantial 507; as merely
465; critiqued in the Vimalaprabhaṭīkā a convention 567; non-referential 90; in
(Stainless Light Commentary) 51 Sarvāstivāda Abhidharma 180; as one of
caste: Ambedkar on 639–40, 643; Aśvaghoṣa six elements 182; six sense consciousnesses
on 85; Buddhist critiques of 647n8; in Jitāri’s 180, 269–70, 273, 275–7, 279–80, 409, 416;
biography 547–8; Jitāri’s critique of brahmins transformation of 276–7, 367, 382–4, 389n17,
552; Saraha’s rejection of 126 390n22, 390n24, 390n28; as one of twelve
Cāttaṉār 113–23; on Buddhist logic 115–16; links of dependent arising 176, 214–15; in
Cilappatikāram (The Story of the Anklet) Yogācāra ontology 275, 279, 566–7, 588; see
113–14; exemplifying philosophy as narrative also ālayavijñāna; five aggregates; mind-only
117–20; exemplifying philosophy as poetry conventional truth (saṃvṛtisatya; conventional
in 120–1; on interdependent origination reality, customary truth): agreeable as long
in 116–17; Maṇimēkalai Tuṟavu (The as not analyzed 456, 470, 474; Āryadeva’s
Renunciation of Maṇimēkalai) 113–22; as revisionism concerning 242–4, 245;
philosopher 115–22 Bhāviveka on 352, 474, 475; Candrakīrti
catuṣkoṭi (tetralemma): analyzed as on 245, 409–12, 413–16, 455, 474, 609;
non-classical logic 231n66; deployed in Pali Dharmakīrti on 313; Dharmapāla as
discourses 214–15, 228n29, 228n31; one of restricting to the imagined nature 241;
four great reasons proving emptiness 606; and Jñānagarbha on 455–8, 460, 474; Nāgārjuna
the Madhyamaka “no thesis” view 239–40; on 212, 220, 224, 225; relationship to
as key to Nāgārjuna’s project 219–22; ethics 242; Sthiramati’s critique of 386–7;
Nāgārjuna’s use of 214–16 Śāntarakṣita on 45–56, 470–2, 474–5;
causal efficacy see arthakriyā threefold criterion for 470; true versus false
causality: backward causation 529–31, 552–3; conventional 455, 456, 457, 461n4; as “valid
basis for inference 270; Candrakīrti’s analysis fictions” 416; as what is acknowledged in the
of 408–9, 412; and dependent arising 215; world 245, 457, 460
666
Index
critical examination (parīkṣā): as analysis links of 20, 23, 116, 117, 176–7, 185n7, 223,
of reality (tattva) 86; decisive to the 224; see also causality; dependent nature
development of Buddhist philosophy 85, dependent nature (paratantrasvabhāva)
347; as a sine qua non of liberation 93; in 240; Asaṅga on 256, 257; as ceasing at
Śāntarakṣita’s Tattvasaṃgraha 485 Buddhahood 366–7; Dharmapāla on 241–2,
365–6; and Dignāga’s typology of cognition
Dalai Lama XIV 650–62; engagement with 285–6, 299n8; Kamalaśīla on 484; Sthiramati
Christianity 652, 654; engagement with on 387; see also three natures
modern science 652, 660–1; engagement with Devendrabuddhi 393–403; on arthakriyā (causal
Prāsaṅgika-Madhyamaka tradition 651, 660; functionality) 395–8; on causal order 530;
engagement with thought of Mohandas K criticized by Jñānagarbha 452; criticized and
Gandhi 650, 652, 659; as leader in engaged relied on by Kamalaśīla 484, 485; mediated
Buddhism movement 659; interpretation of and unmediated effect 395–8; as part of
Śāntideva 653, 656–9; modernist version the “philological school” of commentators
of Buddhist ethics 656–9; as philosopher 528; relationship to Śākyabuddhi 393, 400;
5, 650–1; religious pluralism of 653–6; Śubhagupta influenced by 433
works 651 Dewey, J. 635–6, 638
Daśabalaśrīmitra 194, 195, 197, 198–200, Dhammapada 110n8, 197, 643
618, 627 dhamma/dharma (element, phenomenon) 145,
deathless state (amata) 12, 13–14, 17, 21, 23, 24 189; analyzed in terms of human experience
debate (vāda): Bhāvaviveka on 365; 148; as conditioned 176–8; defined in
Dharamakīrti on 295; Dharmapāla on 365; terms of particularities 153; definition of
Dignāga on 285, 295, 365; pan-Indian rules 146; as dependently arisen 274; identified
of 348; post-Vasubandhu developments in with sabhāva 151; list of fifty-six 146, 147;
364–5; principle of common establishment organized according to diverse modalities
456, 607–8; represented in the Maṇimēkalai 147; as primary existent object 149;
Tuṟavu (The Renunciation of Maṇimēkalai) psychological 146; as topic of Abhidharma
121; sociohistorical context of 347 172, 189, 266, 379; understood as ultimate
definitive meaning (nītārtha) 46, 70, 343n5, elements of existence 152
412, 451, 458 dharmadhātu (realm of phenomena) 54, 58,
dependent arising (pratītyasamutpāda, 260, 597, 603, 608, 625–7
paṭiccasamuppāda): according to Aśvaghoṣa Dharmakīrti 303–21; antirealism of 312,
93; according to Nāgārjuna 213–15, 217–19, 313, 316–18; apologetic purpose 85, 312;
225, 406, 470, 604; as an antidote to anger argument against the Self 94n31, 307,
522; in Buddhist epistemology 290, 433, 308; on causal efficacy (arthakriyā) 311,
435, 530, 570; as the cause of the world 314; on conceptual construction 315, 502;
in the tantras 49–50; cessation of as the critique of empiricism 291–2, 295, 298;
eradication of ignorance 23; and compassion critique of physicalism 305–7; definition of
653, 658–69; and conventional truth means of knowledge (pramāṇa) 309–12;
(saṃvṛtisatya) 456, 460, 470, 605, 609; as a on exclusion (apoha) 312–14; on inference
correlate to no-self 24, 190; and dependent 291–2, 295, 313, 315, 318–19, 529;
designation 601; and the dependent nature Nyāyabindu (The Drop of Logic) 304, 549;
69–70, 365; as discerned by Gotama Buddha on particulars 314, 441; on perception 287,
20; and emptiness 53, 216, 218–22, 350, 314–16; Pramāṇavārttika (Commentary
416–17, 473, 523, 606–7, 623, 660; as on the Means of Knowledge) 317–18,
foundational principle for the Buddha’s 394–5, 398–402; Pramāṇavārttikasvavṛtti
teachings 20, 274; and Indian philosophical (Auto-Commentary) 304, 309, 312–14, 394–5,
argumentation 530; and no-self 268; as taught 531; Pramāṇaviniścaya (Ascertainment of
in Maṇimēkalai Tuṟavu (The Renunciation the Means of Knowledge) 304, 314, 315, 317,
of Maṇimēkalai) 114–15, 116–17, 119, 120; 433, 493n13, 549, 567; on reflexive awareness
as taught in Abhidhamma/Abhidharma 145, (svasaṃvedana) 317–18, 398–402; relation
148, 155, 176–7, 200, 214–15, 416; as taught to Dignāga 291, 295, 304–5; on religious
in the Śālistambasūtra (Rice Seedling Sūtra) authority 245–6, 305, 308–9; responding to
412, 488; as taught in the tantras 49–50; Kumārila 285, 290–1; theory of causation 529;
as taught in Yogācāra 69–70, 275, 278, on universals as conventional 312–14, 441
281, 289, 384–6; and the transformation of dharmanairātymya (selflessness of dharmas)
consciousness 383, 384–5, 389n17; twelve 409, 621, 625
667
Index
Dharmapāla 361–75; in the Cheng weishi 284, 288, 299n5; on perception 285, 286;
lun (A Treatise for the Establishment of Pramāṇasamuccaya (Compendium of the
Consciousness-Only) 361, 363; commentary Means of Valid Cognition) 295; on the
on Āryadeva’s Catuḥśataka (Four-Hundred principle of invariable concomitance 291–2;
Verses) 240–1, 244, 245, 349, 362, 364, reliance on and critique of Vasubandhu
365, 367; commentary on Dignāga’s 285; on self-awareness (svasaṃvedana/
Ālamabanaparīkṣā (Investigation of svasaṃvitti) 269, 288–91, 368, 369, 370,
Cognitive Objects) 362, 364; commentary on 373n25; Traikālyaparīkṣā (Investigation of
Vasubandhu’s Viṃsikā (Twenty Verses) 361, the Three Times) 285, 287; on two types of
362, 363–4, 371; criticism of non-Buddhist cognition 285–7; unifying epistemology,
schools 365; criticism of Bhāviveka dialectics, and semantics 285
240–1, 361, 365–7; as holding opposite dilemmas: in Aśvaghoṣa 90; in Milindapañha
views to Paramārtha 370; influence on (Questions of Milinda) 101–2; in Śāntarakṣita
East Asian Buddhism 361, 362, 370–1; 466, 467, 469, 470; see also catuṣkoṭi
on meta-cognition (svasaṃvedana) 362, doubt (vicikitsā) 168, 178, 182, 245, 296, 311,
367–70, 373n25; as original and creative or 440, 446–8, 472, 574
conservative thinker 362–3, 370; and theories doxography 2–3, 5, 326, 356n3; Āryadeva’s
of higher-order consciousness 369–70; Jñānasārasamuccaya (Compendium on
theory of the cognition of self-cognition the Essence of Knowledge) 247n4, 548;
(*svasaṃvitti-saṃvitti) 368–70; on the three Bhāviveka’s Madhyamakahṛdayakārikā
natures (trisvabhāva) 240, 365–7, 240–1; use (Verses on the Essence of the Middle
of Dignāga’s logic for the interpretation of Way) 346–56; Cāttaṉār’s list of ten
Yogācāra 361, 363–4 schools in the Maṇimēkalai Tuṟavu (The
*Dharmaśreṣṭhin 180–3, 184 Renunciation of Maṇimēkalai) 113,
Dharmottara 439–49; association with Arcaṭa 115, 119–20; Dharmapāla’s criticism
and Śubhagupta 429–30; cognizer-relative of non-Buddhist schools 365; Jitāri’s
shift in epistemology 443–6; commentary Sugatamatavibhaṅgakārikā (Verses on the
on Dharmakīrti’s Nyāyabindu (The Drop of Classification of Buddhist Systems) 548, 551,
Logic), 439, 441, 444–6; critique by Jitāri 554–5; Kamalaśīla’s Tattvasaṃgrahapañjikā
548–9; critique by Ratnākaraśānti 587; (Commentary on the Compendium of
determination of a continuum (santāna) 442; Realities) 428, 434, 484–6; pudgalavāda
doubt and certainty in inference-for-oneself as presented in 191, 197–8; Śāntarakṣita’s
446–7; divergence from Dharmakīrti 440–3; Tattvasaṃgraha (Compendium of Realities)
inference as involving determination of a 428, 464–5, 484–6; tantric Buddha as
particular 443; influence on Jñānaśrīmitra and doxographer in Kālacakratantra (Wheel
Ratnakīrti 440, 443, 559–61; life and works of Time Tantra) 47–53; Tibetan 211, 459,
439; non-apprehension of the perceptible 476n4, 490–1, 549, 615
(dṛśyānupalabdhi) 443–6; perception and Durkheim, É. 637
inference as having two objects 441–3;
perception as involving conceptual ekānekaviyogahetu see neither-one-nor-many
construction 441–3 argument
dhātu (sense realm, entrance) 269, 411, 458 ekayāna see single vehicle
Dignāga 284–302; as accepting only two emptiness (śūnyatā, suññata): Abhayākaragupta
pramāṇa 284–5, 310; Ālambanaparīkṣā on 625; Ambedkar on 643; Āryadeva on
(Investigation of the Cognitive Object) 288, 245–6, 248n9; Asaṅga on 257; Atiśa on
362, 364–5; on antinomy 295; apologetic 605, 606, 607, 610, 611, 612; Bhāviveka
purpose 85, 285; Bhāviveka’s reliance on 349–50, 353; Buddhaghosa on 153,
on 348, 349, 354, 356; on conceptual 157n37; Candrakīrti on 406, 409–10, 412,
construction 315; critiqued by Candrakīrti 414, 416–17; conscious (ajaḍaśūnyatā)
408, 417–18; definition of pramāṇa 314; 48; Dalai Lama XIV on 651, 657, 660; as
Dharmapāla’s reliance on 361, 362, 363–4, dependent arising 229n51, 416; Dharmakīrti
371; on exclusion of what is other 295–8, on 316; Dharmapāla on 244, 361, 366–7;
313; Hetucakraḍamaru (Two-Headed emptiness of 230n54, 230n57; and epistemic
Drum of the Wheel of Reasons) 288; on foundationalism 221; Haribhadra on 499,
inference 291–5, 318, 319, 363; influence of 504–6; Jñānaŕīmitra on 567; Kamalaśīla
Bhartṛhari on 285; influence on Dharmakīrti on 483–4, 487, 489; Nāgārjuna on 214,
304; Nyāyamukha (Gateway to Logic) 218–22, 223, 229n50; Ratnākaraśānti on 402,
668
Index
594, 595, 597; Śāntarakṣita on 465–6, 470; transformation of 261; Buddhaghosa on 338;
Śāntideva on 515, 523; Saraha on 129–31, Dharmapāla on the lack of externality of
134, 135; in the tantras 46, 53–5; Vasubandhu 363–4; Jñānagarbha on the appearance of as
on 277–9 not negated 455, 460; occurring always with
error (bhrānti): Buddha’s retaining a small other dhammas 147–9; as one of four main
amount of 595, 596–7; distinction between types of phenomena 151; Sthiramati on 379;
error and aspect 505–6; and the identification see also five aggregates
of a self 271–2; inherent in dualistic cognition formations (saṃskāra, saṅkhāra; volitional
400; as lacking causal efficacy 457; as mark constructions, conceptual patterns): absent
of the conventional 456–7; non-conceptual in discriminating gnosis 58; analyzed in the
397; Ratnākaraśānti on 593 Vibhaṅga (Analysis) 148; Asaṅga on the
ethics: and Ambedkar’s understanding of Buddha’s transformation of 261; disassociated
Buddhism 638–9; Āryadeva on 242–7; from thoughts (cittaviprayukta) 180; as
brahmavihāra (divine abiding, immeasurable) impermanent 199; see also five aggregates
148, 637, 657, 658; Dalai Lama XIV on Foucault, M. 3, 27, 490
654, 656–9; Dharmakīrti on the possibility four noble truths: in the Abhidhamma 144–5,
of moral transformation 308; ethical 148; Bhāvviveka on 352; critiqued by
implications in Milindapañha (Questions of Ambedkar 641–2; Dharmakīrti on 305, 309,
King Milinda) 102; moral cultivation 642, 315, 317, 533; and eightfold path 23–4;
643, 656–7; Nāgārjuna on 208, 212, 213, explained by Cāttaṉār 116, 117; as means to
232n79; relationship to conventional truth abandon heterodox views 175; as a means
242; role of ethics in early Buddhist thought to eliminate defilements 176, 178, 180; as
23; Śāntideva on relationship to metaphysics medical diagnostic scheme 22, 24; Nāgārjuna
511, 523–5, 526n2; Saraha on 128 on 217–18; as a practice of diagnosing
exclusion (apoha, anyāpoha): Dharmakīrti on experience 155; as taught by Gotama Buddha
313–14; Dignāga on 284, 295–8, 313; Jitāri 12, 22–4; understood on the path of vision
on 552; Jñānaśrīmitra on 448n7, 560, 561–4, 182–3; see also suffering
602; Ratnakīrti on 448n7, 573, 578–82, 584;
Śubhagupta on 430, 431 Gadamer, H.-G. 335–8
externalism (bāhyārthavāda) 432–3 Gandhi, M. K. 639, 650, 652, 657, 659
Gautama Buddha see Siddhārtha Gautama
feeling (vedanā; feeling tone): according to the God see Īśvara
“Many Types of Feeling Sutta” 154; analyzed Gorampa 248n15, 460, 481
in the Vibhaṅga (Analysis) 148; Asaṅga on Gotama Buddha 11–26; construction of personal
the Buddha’s transformation of 261; and identity 19–20, 21; decision to teach 21–2;
craving (taṇhā/tṛṣṇā) 23; as an “ultimate” figure as close as possible to the historical
(paramattha) dhamma 151; see also five Buddha 11; immaterial attainments 16;
aggregates night of awakening 16–17; pursuit of ascetic
five aggregates (skandha, khanda): analyzed practices 17–19; recollection of past lives
as empty 153, 218; analyzed by Sthiramati 19–20; as represented in early discourses
378–9; Asaṅga on the Buddha’s five kinds 11–12, 24n1
of mastery (vibhutva) relating to 261; gradualism 415, 451, 458–9, 483, 492
associated with the truth of suffering 144; grahaka (apprehender, subject) 53–4, 56, 289,
as the basis for the false construction of a 400, 589
self 268–73, 278, 338; as an ever-changing grāhya (apprehended, grasped object) 53, 56,
continuum 268; examined in terms of 59, 289, 400
modalities in the Analysis (Vibhaṅga) 148–9; Guhyasamājatantra (Secret Assembly Tantra)
taught as part of a gradual path 458; in the 47, 53–4, 56, 450, 487, 547–8, 626
Pañcaskandhakavibhāṣā (Commentary guṇa (quality) 59, 89, 90
on the Treatise on the Five Constituents)
378; Nāgārjuna on 218; as part of the Habermas, J. 339
skandha-dhātu-āyatana framework 189; Hadot, P. 136, 509n9
in Sarvāstivāda Abhidharma 180; in the Haribhadra 497–510; commentaries
tantras 53, 59; as teaching appearance and on the Ornament of Realizations
emptiness 278 (Abhisamayālaṃkāra) 499–504; as exegete
form (rūpa; matter): analyzed in the Vibhaṅga 497, 508; as follower of Śāntarakṣita 464; and
(Analysis) 148; Asaṅga on the Buddha’s Kamalaśīla 434; as metaphysician 506–8; as
669
Index
philosopher of the path 497, 501–4; questions interpretable meaning (neyārtha; provisional
of authorship 499–500; as specialist in meaning) 46, 70, 268, 274–6, 343n5, 412,
prajñāpāramitā literature 497, 498; theory of 451, 458
illusion 504–6 intersubjectivity 275, 387–8, 596
Hegel, G. W. 500, 509n4 Īśvara (God): Aśvaghoṣa’s rejection of 92–3;
Heidegger, M. 136 Bhāviveka’s refutation of 455; Dharmakīrti’s
hetu see logical reason rejection of 305; Dharmottara on 444, 445;
Hevajratantra 47, 55, 56, 57, 58–9, 60, 61, 603 Jitāri’s refutation of 551; Jñānaśrīmitra’s
Hume, D. 477n20 refutation of 558; in Nyāya-Vaiśeṣika 533;
Ratnakīrti’s arguments again 602; refuted in
imagined nature (parikalpitasvabhāva): as the tantras 49–50; Śubhagupta’s refutation of
defilement 72; Dharmapāla on 366; as 429, 430–1
existing only nominally 69; including three Īśvarasena 291, 295, 298, 319
kinds of conceptual cognition according to
Dignaga 286–7; Jñānagarbha’s critique of Jaimini 119, 285
453, 456; Ratnākaraśānti on dissolution of Jains: Bhāviveka critique of 349; in contrast to
589; see also three natures Buddhism 110n1; critiqued by Āryadeva 349;
imagination of the unreal (abhūtaparikalpa) 69, critiqued by Dharmapāla 365; critiqued by
241, 459 Jitāri 550, 554; critiqued by Kamalaśīla 485;
impermanence (anityatā, aniccata): Ambedkar critiqued by Śāntarakṣita 465; mainstream
on the ethical significance of 643; Cāttaṉār meditation practices of 87; as pluralistic 653;
on the liberating insight into 116; Śāntideva represented as opponents of Buddhism 110n7,
on tension between bodhisattva vow and 515; 120, 127; represented as using dilemmas in
Vasubandhu on impernance as conditions debate 101; theory of the soul (jīva) critiqued
280–1 in the tantras 52; Vidyānandin’s critique of
ineffability/inexpressibility: of the awakened Prajñākaragupta 528; view of the pudgala as
state 135; of causal facts 390n24; of the material and non-sentient 191
dependent and perfect natures 240–2; of the Jayanta 320n3, 528, 537n1
“indeterminate self” (pudgala) 190, 274; of Jayanta Bhaṭṭa 439, 448n1
properties 563; of reality 57, 268, 275, 276, Jinendrabuddhi 284
277; of ultimate truth 452–5; of underlying Jitāri 547–57; affiliation with Madhyamaka
ontology in Yogācāra 386 548; affiliation with Yogācāra 554; as author
inference (anumāna): adopted by Bhāviveka of short digests 427, 550; on backward
347, 348, 349, 355; autonomous causation 531, 552–3; on Dharmakīrti as a
(svatantrānumāna) 347, 349, 450, 476n4, Mādhyamika 555n5; life and works of 547–9;
491; and the Buddha’s omniscience 534; tantric practice 547–8; Vādasthānāni (Topics
Dharmakīrti on 446, 529–30; Dharmakirti’s of Debate) 549–5
skepticism of scripturally-based 246; Jñānagarbha 450–62; on reflexive awareness
Dharmottara’s theory of two objects of 441, (svasaṃvedana) 460; on conventional truth
443; Dignāga on 291–5, 363; examples (saṃvṛtisatya) 455–8, 460; as developing
(dṛṣṭānta) in 245, 284, 363; inference-for- Bhāviveka’s ideas 450, 453–4; as embracing
oneself (svārthānumāna) 209, 284, a gradual path 458–9; on imagined versus
304–5, 431, 439, 446–7, 528, 536, 549; mere entities 456–7; on mind-only as
inference-for-others (parārthānumāna) 209, conventionally valid 458–9, 460; on the
284, 299n23, 304–5, 423, 439, 446, 447, principle of common establishment in debate
528, 549; inferential reasoning as used by 456; on the qualifier “ultimately” 454–5;
Cāttaṉār 115–16, 119; governed by a natural Satyadvavibhaṅgavṛtti (Commentary on the
connection (svabhāvapratibandha) 295, Distinction Between the Two Truths) 459–60;
319; Jñānaśrīmitra on 558, 559–61; in the as synthesizing Yogācāra and Madhyamaka
Maṇimēkalai Tuṟavu (The Renunciation 450; tantric works attributed to 451; Tibetan
of Maṇimēkalai) 121; scripturally-based doxographical classifications of 459–60; on
(āgamāśritānumāna) 245, 310, 488; similar the two truths 452–4, 464
and dissimilar examples (dṛṣṭānta) in 291–2, Jñānaśrīmitra 558–72; on appearance and
293–5; threefold nature 363; Vasubandhu on determination 559–61, 570; on conventional
the inference of sense organs 270; see also reality 564–6; on exclusion (apoha)
logical reason 561–3; influenced by Dharmottara 440,
interdependent origination see dependent arising 561; life and works 558–9; similarity to
670
Index
Prajñākaragupta 528, 533; struggle with lists (mātṛkā, mātikā; matrix): as closed
Ratnākaraśānti 567–70; synthesis of Buddhist 150; in the Enumeration of Phenomena
epistemological thought 570 (Dhammasaṅganī) 146–8; as generative
152; as inherently incomplete 147, 155; as a
Kālacakratantra (Wheel of Time Tantra): modal method of analysis 147, 152, 155; as
Abhayākaragupta’s commentaries on 616–17; a view-from-nowhere 151; as reflecting oral/
Buddha as philosopher and doxographer aural context of teachings 172
in 47–55; on the mind as inseparable from logical reason (hetu, liṅga; evidence,
enlightened awareness 58; on nirvāṇa as inferential reason): Dharmakīrti on 529–30;
signless 57; on the world as constructed by Dharmottara on 443, 446–7; Dignāga on
karma 59 446; effect-evidence (kāryahetu) 529, 530;
Kamalaśīla 480–96; acceptance of Yogācāra identity-evidence (svabhāvahetu) 529, 559;
theories 484, 486; doxographical nonperception (anupalabdhi) 443–6, 531–3;
classification of 481, 485–6, 490–2; embodied pseudo-reason (hetvābhāsa) 294, 295, 447,
dimension of authorship 481; Extensive 574, 575; triple conditions for 446–7
Commentary (Pañjikā) 484–6; gradualism logico-epistemological (pramāṇa) tradition 303,
of 480–1, 483, 492; historical accounts of 361, 427, 428–9, 433, 573
480, 481; influence on Abhayākaragupta Lokāyata see Cārvāka
615; influenced by Śubhagupta 428–9; luminosity (prabhāsvara, prakāśa;
Madhyamakāloka (Light of the Middle) illumination): of awareness 486; Dharmakīrti
482–4, 615; neither-one-nor-many argument on the mind as naturally luminous 308;
in 466; provisional acceptance of mind-only in Dharmakīrti’s definition of reflexive
475; questions of authorship 486–7; on awareness 390–1, 402; Jñānaśrīmitra
reflexive awareness (svasaṃvedana) 486; on 567–8; as the nature of the mind in
Stages of Cultivation (Bhāvanākrama) 481, Abhayākaragupta 625, 626; Śākyabuddhi’s
482–3, 651; on term false conventional account of contentless luminosity 402;
(mithyāsaṃvṛti) 461n4 as synonym for reflexive awareness in
Kant, I. 81, 226n12, 249n24 Ratnākaraśānti 402, 581, 590–1, 602, 626;
karma/karman (action): according to Tājiks in the tantras 54, 55, 56; of the ultimate
(Tāyins) 51; Ambedkar on 641, 642; as awakening mind 610
bondage 86, 90; denied by Cārvākas 51; as
incompatible with a self 90, 91–2; as mental Madhyamaka: Āryadeva as co-founder of
588; as not due to God 49; in preclassical 213, 236; Bhāviveka as systematizer of
Sāṃkhya 88–9; as taught through narrative 348; criticized by Yogācāra philosophers
118–20, 188; theorized in Abhidharma 177, 384–7; critique of realist semantics 239;
180, 188; without an agent 189 Dharmapāla’s three-nature reading 240–2;
Karṇakagomin 246, 303 as the elimination of views 218, 221, 350,
Kathāvatthu (Points of Discussion) 160–71; 352, 354; and ethics 212–13, 242–7, 523–5;
citing the nun Śaila on “being” (satva) 190; dependent origination in 213–15, 217–19,
distinction between conventional and ultimate 225, 406, 470, 604; Nāgārjuna as founder
teachings 152; method of argumentation of 213, 405; as rejecting any thesis 221–2,
161–8; scripture and reason as implicitly 239, 364, 607; synthesized Yogācāra 450,
accepted as pramāṇas in 170; see also 472–5, 570; as teaching notational as well as
Abhidhamma; Moggaliputta Tissa existential dependence 416; as therapy 350;
Kātyāyanīputra 172–87; as addressing a see also emptiness
monastic audience 178, 179; as compiler mahābhūta (four great elements) 148, 179, 410
of the *Abhidharmāṣṭagranthaśāstra Mahāsāṃghika 368
(Abhidharma Treatise in Eight Sections) 175; Mahāvibhāṣa (Large Commentary) 172–87;
as founder of the Sarvāstivāda 180; see also Aśvaghoṣa as possible scribe for 83; on
Mahāvibhāṣā conventional cognition as ignorance 457;
(Large Commentary) and Kātyānanīputra 183–4; and Sarvāstivāda
Kuiji (Master Ji) 368, 370 174, 177–8; as a source for Mahāyāna
Kumārajīva 190 developments 184
Kumārila: critique of the Buddha’s omniscience Mahāyāna sūtras 27–40; Abhayākaragupta
533; critique of Dignāga 285, 290; on the authority of 619–20;
definition of pramāṇa 310; responded to by Acintyabuddhaviṣayanirdeśasūtra (Sūtra of
Śāntarakṣita 465 Instruction on the Inconceivable Sphere of
671
Index
Buddhas) 38–9; Adhyāśayasañcodanasūtra memory, 65, 269, 271, 287, 311, 315, 368, 369
(Sūtra which Incites Resolve) 32, 34, mental perception (manovijñāna): Asaṅga on
35, 39; Akṣayamatinirdeśasūtra 453; 255, 261; Dharmakīrti on 315; Vasubandhu
Bhadramāyākāravyākaraṇasūtra on 269–70, 279–80; see also perception
(Prophecy for the Magician Bhadra) 39; mereological (part-whole) analysis 316, 409–10,
as buddhavacana (word/speech of the/a 467, 471, 551, 660; see also ekānekahetu
Buddha) 29, 31–40; Dharmasaṅgītisūtra (neither-one-nor-many argument)
453; Dāśabhūmikasūtra (Ten Stages Sūtra) metaphor (upacāra) 274–81, 289–91, 297, 381–8
393, 610, 619; echoed in the tantras 53; as Milindapañha (Questions of Milinda) 97–112;
featuring multiple voices 33; Gayāśīrṣa and the art of questioning 108–9; appeal
Sūtra (Summit of Mount Gayā Sūtra) 393; to narrative 103–5; dramatic movement
Kāśyapaparivarta (Chapter on Kāśyapa) of 105–7; example of lion in a gilded
377, 381, 489, 608; Laṅkāvatārasūtra cage 107–8; heterogeneity of 98; and
389n19, 457, 621; Mahāvairocanasūtra impersonalism 105–7; invitational quality
(Discourse on the Enlightenment of of 109; Menander and Nāgasena as dramatic
Mahāvairocana) 451; Maitreya Buddha as characters 98; Menander as a historical
teaching 252, 253; Śālistambasūtra (Rice figure of 97; and question of relevance
Seedling Sūtra) 412, 488; Śrīmālādevīsūtra 108–9; as representative of a genre 97; on
(Discourse for Queen Śrīmālā) 71, 625; understanding 340; use of dilemmas 101–2;
Tathāgatagarbhasūtra (Discourse on Buddha use of illustrations in 99–101
Nature) 71; Ugraparipṛcchāsūtra (Questions Mīmāṃsā 119, 314, 351, 353, 464
of Ugra) 517; see also Prajñāpāramitāsūtra; mind (manas, citta; awareness, thought):
Saṃdhinirmocanasūtra (Scripture Clarifying distinguishing good and bad 146; manas
the Intent) (mental organ) 269–70, 279, 285, 379,
Mahīśāsaka school 253 534; defiled (kliṣṭaṃ manas) 256, 259, 260,
Maitreya Buddha 64–77; as author 65; cult of 380, 593, 622; defined as nirvāṇa in the
64; as founder of Yogācāra 64; relation to Kālacakratantra (Wheel of Time Tantra) 57;
Maitreyanātha 74; synthesis of Yogācāra and eliminated in the transcendent (lokuttara)
Tathāgatagarbha thought 72–4; as teacher of mind 279; natural (nijacitta) mind 125; in
Asaṅga 64, 252, 253, 268 the Pali Abhidhamma 146–7, 150, 151, 189;
manas see mind as primordially pure according to Saraha
Mañjuśrī 38, 548 129–31; in Sarvāstivāda Abhidharma 179,
Manorathananandin 303 180; twofold appearance of in tantras 56; see
manovijñāna see mental perception also mental perception
Marx, K. 636, 637, 643, 644, 647 mindfulness (sati; smṛti): on the bodhisattva
materialism see Cārvāka; physicalism path 610; Buddhagosa on 146–7; four
mātṛkā/mātikā see lists applications or foundations of 21, 144, 148,
matter see rūpa 181, 182; as an intervention against craving
means of valid cognition see pramāṇa (taṇhā/tṛṣṇā) 23; in the Pali texts 12, 18, 19,
meditation (bhāvanā; cultivation, habituation): 21, 24, 150, 155; in tantric texts 56, 132; as a
for abandoning mental afflictions 516; on virtue 512
awakening mind (bodhicitta) 618, 619; calm mind-only (cittamātra; consciousness-only):
(śamatha, samatha) 16, 150; on compassion Asaṅga on 255, 256, 257, 259, 458;
651; on the disgusting or repulsive Abhayākaragupta on 622–3; Āryadeva
(aśubhabhāvanā) 153, 182, 243, 248–9n18; on 240; Bhāviveka on 458; Candrakīrti
on the divine abodes (brahmavihāra) 637; on 405–6, 458; Dharmakīrti on 291, 306,
insight (vipaśyanā, vipassanā) 16, 150, 316–17; Dharmapāla on 363–4; Dignāga
409; as giving rise to a vivid appearance on 285–6, 288; Jñānagarbha on 459, 460;
623; as habituation 503; jhāna/dhyāna Jñānaśrīmitra on 566; Kamalaśīla on
14–16, 148, 169; nondual 601, 603, 606, 429, 475, 486; in Mahāyāna Sūtras 256,
610; path of (bhāvanāmarga) 183, 287, 595, 263n38; Maitreya on 72; Nāgārjuna on
606; philosophy and ritual as types of 515; 406; ontological versus epistemological
recollection of the Dhamma 336; salutary interpretation 276; Śākyabuddhi on 402;
practices of 653; stages of described by Śāntarakṣita on 429, 472–5; Sthiramati on
Aśvaghoṣa 85, 87, 89; tantric 59–60, 616; 458, 459; Śubhagupta as refuting 432–6;
wisdom born of (bhāvanāmayīprajñā) 356, Vasubandhu on 268, 274–7, 289; see also
508; see also mindfulness vijñaptimātra
672
Index
Moggaliputta Tissa 160–1, 166, 170, 184; see 387; of the Buddha 352; as characteristic of
also Kathāvatthu (Points of Discussion) perception 442; see also perception
Mokṣākaragupta 447, 448 nonduality: of emptiness and compassion in
momentariness (kṣaṇikatva): arguments for the tantras 51; knowledge as nondual 500–1,
in relation to a self 91–2; as characteristic 506; in nondual meditation 601, 603, 606,
of particulars 441; Dharmakīrti’s proof 610; of saṃsāra and nirvāṇa in the tantras
of 314, 535; Dharmottara on 442, 443; 55–7; variegated (citra-advaita) 469, 545; in
Jitāri’s proof of 552, 553–4; Jñānaśrīmitra’s Yogācāra 280–1
proof of 565–6; and perception 273, 442; nonexistence (abhāva) 271, 314, 623–4
Ratnakīrti’s reconciliation of with universals nonperception (anupalabdhi): Dharmakīrti on
578–80; Ratnakīrti’s proofs of 573, 574–8, 300n25; Dharmottara on 443–6; Dignāga on
583, 584n7; and Sarvāstivāda 177; as tied 291, 298; as a logical reason 531–2, 537n3; of
to appearances 281; Vasubandhu’s proof of the pervader 552; of the perceptible 443–6
270–1 no-self (anattā, anātman; non-self, selflessness):
Mūlasarvāstivāda Vinaya 84, 490; see also according to Aśvaghoṣa 84, 85, 90; as
Sarvāstivāda realized by Gotama Buddha 19–20; taught
in the Abhidharmakośabhāṣya (Treasury of
Nāgārjuna 211–35; authorship defined by Abhidharma Commentary) 268–70; taught in
analytic works 211–12; Bhāviveka’s the tantras 50; unique teaching of the Buddha
interpretation of 346; Candrakīrti’s 20, 188
interpretation of 405–6, 407; as critical of Nussbaum, M. 523
Abhidharma 212, 213, 215–17; compared Nyāya school 464, 465; pramāṇa theory 310,
to Skeptics 221; critique of the epistemic 417; theory of the whole (avayavin) 316;
project 221–2; on dependent arising view of self 90
(pratītyasamutpāda) 213–19; on emptiness
218; and ethics 212–13; as a fundamentalist object (ālambana, viṣaya) 36; as absent in
reformer 213; verses of praise 213–14; and nirvāṇa 57; of the ālayavijñāna 380; as
Mahāyāna 212; Mūlamadhyamakakārikā already perceived 532; external 290; final
(Root Verses of the Middle Way) 213–22; cognitive state as lacking (anālambana,
Ratnāvalī (Precious Garland) 213, 466, nirālambana) 505, 594; mental 36; as
223–4; tantric author 248n6; on the two truths momentary perceptual content 588, 589; as
223, 230n57; Vigrahavyāvartanī (Dispeller of pure 55, 59; six sensory 268; Śubhagupta on
Disputes) 211; works 211, 226n9; Yuktiṣāṣṭika cognition’s need for 433–5; two kinds as basis
(Sixty Verses on Reasoning) 211, 238, 406, for two kinds of pramāṇa 441; as ultimate
411, 477n26, 605 454; Vasubandhu on 36, 268–9, 275–6, 279
narrative 4, 9, 81–2, 103–5, 107, 117–21 omniscience (sarvajña, sarvavid): Bhāviveka on
neither-one-nor-many argument 353; Jitāri on 547, 552; Jñānagarbha on 453;
(ekānekaviyogahetu): in Abhayākaragupta mantras as evidence of the Buddha’s 432;
623; against material simples 467–8; against Prajñākaragupta on 531, 533–5; Śubhagupta
mental simples 468–9; in Kamalaśīla 466, on 428, 430, 431–2; of tantric buddha 46, 47
483; in Nāgārjuna 466; in Śāntarakṣita other-emptiness (gzhan stong) 131, 240
465–70; in Śrīgupta 465, 466
Neoplatonism 506, 507 Pali discourses (sutta): Aṅguttara Nikāya 13,
Nietzsche, F. 4, 81, 124, 129, 135 42n25, 145, 243, 335; Dīgha Nikāya 16;
Nikāya Buddhism 172 Khuddaka Nikāya 109n1, 160, 199; Majjhima
nirākāravāda (theory that cognition is not Nikāya 14–20, 101, 110n13, 333; manuscript
endowed with an image, form free): Dignāga’s tradition 12–13; Saṃyutta Nikāya 22, 214,
critique of 289; Jitāri as adherent of 549; 216, 217, 220; Suttanipāta 108, 214, 215,
Śākyabuddhi’s account of 402; Śubhagupta’s 228, 350; as sources for Abhidhamma
defense of 430, 432; see also sākāravāda 143–56, 162
nirvāṇa: and attainments 16; defined as the mind paradox: Nāgārjuna on 230n57; Saraha on
of awakening (bodhicitta) 55; in Kamalaśīla’s 129–30, 135
Stages of Cultivation 483; non-abiding 625; Paramārtha 64, 173, 184, 194, 196, 252, 366,
ultimate truth as condition for 452 370–1
nonconceptual cognition (nirvikalpajñāna): paramārthasatya see ultimate truth
awareness obtained subsequent to it particular (svalakṣaṇa, bheda; particular
(tatpṛṣṭalabdhajñāna) 387; of bodhisattvas characteristic, defining trait): Candrakīrti’s
673
Index
critique of Dignāga on 408; as dependent undermining 220–1; nonexistence (abhāva)

concepts 410; as existing ultimately 441; as as 300n24; as non non-deceptive 457, 533; in
having causal efficacy (arthakriyā) 532; as Nyāya 310; Prajñākaragupta on omniscience
the object of perception 311, 441, 532; realist as the ultimate pramāṇa 534; in Sāṃkhya
semantic theory of 296; see also universal 310; scripture and reason as implicitly
paryudāsa (affirming negation) 130, 408 accepted in the Katthāvatthu 170; see also
path of vision (darśanamārga) 176, 178, 181, inference; perception
183, 594, 606, 609–10 pramāṇaphala (result of a means of
perception (pratyakṣa, upalabdhi): valid cognition, instrumental effect):
Dharmakīrti’s definition of 623; Dharmotta Devendrabuddhi on 395–8; Dharmakīrti’s
on the two objects of 441–3; Dignāga’s revision of Dignāga’s analysis 290–1;
definition 285, 315; four types according to Dignāga on 289–90, 373n27
Dharmakīrti 315–16; Jñānaśrīmitra on 559; prapañca (conceptual proliferation) 213–14,
mental perception (mānasapratyakṣa) 286; 219, 410, 452, 588–90, 593, 596–8
non-conceptual versus conceptual 310, 442; prasajya-pratiṣedha (non-affirming negation)
Prajñākaragupta’s theory of existence and 130, 241, 408
531–3, 534; types according to Dignāga Prāsaṅgika-Madhyamaka: Āryadeva
285–6; Yogācāra understanding of sense characterized as 240; Candrakīrti characterized
perception 275–6, 279–80 as 230n57, 232n77, 347–8, 476n4, 481; Dalai
perfected nature (pariniṣpannasvabhāva) 69, Lama as a follower of 651, 660; as a Tibetan
240, 365, 484; see also three natures doxographical distinction 491–2
Perfection of Wisdom Sūtras see pratijñā see thesis
Prajñāpāramitāsūtra pratītyasamutpāda see dependent arising
person see pudgala pratyakṣa see perception
Phule, J. 637, 647n2 pudgala (person): associated with term
physicalism 305–7; see also Cārvāka “Pudgalavādin” by opponents 142; canonical
Plato 4, 9, 81, 211, 231n61, 231n62, 231n64 texts on 189–90; compared to Brahminical
prajñā see wisdom understandings of ātman (self) 191; criticized
Prajñākaragupta 528–39; compared to Berkeley by various Buddhist authors 196; Jain
531; critique of Dharmottara 529; on the understanding as material and non-sentient
Buddha’s omniscience 533–5; on Dharmakīrti 191; as neither the same nor different from the
320n3, 395, 528–39, 558, 559, 567, 569;theory elements of being 189; tantric critique of 50–1;
of backward causation 529–31; theory of Vasubandhu on 274; Vātsīputrīya/Sāṃmitīya
existence as defined by perception 531–3 views of 191–2; see also pudgalavāda
Prajñākaramati 34, 199, 405, 475, 620 pudgalanairātmya (personal identitylessness,
Prajñāpāramitāsūtras (Perfection of Wisdom selflessness of persons) 50, 409, 410, 621,
Sūtras): Aṣṭasahāsrikaprajñāpārmitā 623, 625
(Perfection of Wisdom in Eight-Thousand pudgalavāda (doctrine that asserts a person):
Verses) 54, 227n16, 499, 501, 504, 616; critiqued by diverse Buddhist philosophers
Candrakīrti on 405; and development of 197; as a pejorative label 191; refuted
Mahāyāna 212, 348–9; Haribhadra on 497, by Vasubandhu 274; refuted in the
500–1; Kamalaśīla’s commentaries on 489; Vimalaprabhāṭīkā (Commentary on Stainless
Prajñāpāramitāhṛdaya (Heart Sūtra) 278, Light) 51; resources for study of 192, 193; see
489, 653, 654; Vajracchedikā Prajñāpāramitā also Vātsīputrīya/Sāṃmitīya
(Diamond Cutter Perfection of Wisdom)
43n35, 66, 254, 489 quietism: in Āryadeva 238–42, 247; in
prajñaptisat (nominally existent) 69, 286 Nāgārjuna 223, 225, 227n15, 232n77
prakāśa see luminosity
prakṛti (pradhāna, Nature, primordial matter) Ratnākaraśānti 587–600; doxographical
88, 90, 92–3, 455, 456 classification of 590; on the four reliances
pramāṇa (means of knowledge, means of valid 41n13; Instruction on the Perfection of
cognition, source of knowledge): Bhāviveka Wisdom (Prajñāpāramitopadeśa) 286;
on 348–9, 364; Brahmanical theory of 289; ontological account of consciousness 590–5;
Dharmakīrti on 309–12, 441; Dharmapāla’s on the purification of error 55; theory of error
use of 364; Dignāga on 284–5, 289–90; in 587–9, 595, 596–7
Indian philosophy 311, 347; Jñānaśrīmitra Ratnakīrti 573–86; A Proof of Exclusion
on 559; Kamalaśīla on 485; Nāgārjuna as (Apohasiddhi) 578–82; A Theory of
674
Index
Multifaceted yet Non-Dual Appearance sākāravāda (theory that cognition is endowed

(Citrādvaitaprakāśavāda) 578, 582–4; with an image): and the sopalambhaniyama
idealism of 582–4; influenced by Dharmottara argument 433; Dignāga’s adoption of 289;
440; Proofs of Momentariness 574–8; on the Jitāri on 554; Ratnakīrti on 582; terminology
three natures 286 of 477n19, 598n3
Ravigupta 528, 529, 537n1 Śākyabuddhi 393–403; on the causal account
realism: antimetaphysical 226n12; critical 152; of perception 532; criticized by Jñānagarbha
direct 465, 471, 477n19; metaphysical 236, 452; criticized and relied on by Kamalaśīla
238, 246; naïve 506, 507; non-Buddhist 311, 484, 485; on faith-based reasoning 246;
364, 559; ontological 27, 243; representational relationship to Devendrabuddhi 393–5, 400;
468–9, 470, 471, 477n19; robust 651; on mediated and unmediated instrumental
Sarvāstivāda 364; Sautrāntika 460; semantic effect 395–8; as part of the “philological
231, 238, 239, 296–8, 564, 571; scientific school” of commentators 528; on reflexive
652; theory of atoms 363; Vaibhāṣika 465; awareness 398–402; Śubhagupta influenced
Yogācāra 622; see also antirealism by 433; see also Devendrabuddhi
recognition (saṃjñā, saññā; perception, sāmānya/sāmānyalakṣaṇa see universal
classification): analyzed in the Vibhaṅga Saṃdhinirmocanasūtra (Scripture Clarifying the
(Analysis) 148; Asaṅga on the Buddha’s Intent): commentary attributed to Asaṅga 253,
transformation of 261; based on inherited 254, 255, 256; Commentary on the Maitreya
patterns 272; of the Buddha (buddhasaṃjñā) Chapter of 451; denigration of Madhyamaka
34–8, 40, 43n30; Sthiramati on 36; wrong implied in 367; and the Mahāyānasaṃgraha
(mithyāsaṃjñā) 182; see also five aggregates (Summary of the Mahāyāna) 259; presaging
reductio ad absurdum arguments 142, 240, 349, the disagreement between Dharmapāla and
357n10, 406, 476n4, 544 Bhāviveka 367; on the three turnings of the
reflexive awareness (svasaṃvedanā, wheel of Dharma 277
svasaṃvitti; self-awareness, meta-cognition): saṃjñā see recognition
Dharmakīrti’s revision of Dignāga’s theory Sāṃkhya 347, 464; Aśvaghoṣa’s critique of
290–1, 373n28; Dharmapāla on 367–70; 83, 87–8, 89, 90, 93; Bhāviveka’s critique
Dignāga on 285, 288–91, 369; Jñānagarbha as of 349, 352; Dharmapāla’s critique of 365;
accepting conventionally 459; Kamalaśīla’s Vasubandhu’s critique of 267
defense of 486; Mahāsāṃghika view on 369; saṃsāra: actions binding to 91, 176; awakening
Prajñākaragupta as reducing all perception to as transcendence of 21; the bodhisattva
532–3; Ratnākaraśānti on 592; Śākyabuddhi remaining in 212, 513, 515, 522; as fueled by
on 398–402; Sarvāstivāda Abhidharma view grasping 280; Kamalaśīla on the sufferings of
on 368; Vasubandhu as not embracing 269; 490; nonduality of nirvāṇa and 55–7, 128–9,
see also luminosity 218, 219, 230n61
religious authority: Ambedkar’s critique of saṃskāra see formations
divine authority 640; the Buddha as a means saṃvṛtisatya see conventional truth
of knowledge (pramāṇabhūta) 227n22, 305, saṃyojana (fetter) 148, 176, 178
533; Buddhaghosa on 331, 334–5; Buddhist Śaṅkarasvāmin 115, 299n5, 549
skepticism about 246; Gotama Buddha as Śāntarakṣita 463–79; Compendium of True
a 174; Maitreya as a 66; Prajñākaragupta Principles (Tattvasaṃgraha) 92, 463–5,
on 533–5; Śāntarakṣita and Kamalaśīla on 472; incorporation of Yogācāra ontology
487; Siddhārtha Gautama as a 308–9; tantric 464; influenced by Śūbhagupta 428–9;
buddhas as 46, 52; see also buddhavacana; on Jñānagarbha 451, 453, 454, 456, 459;
scriptural authority on the neither-one-nor-many argument
Ricoeur, P. 334 465–70; on non-seeing as the seeing of
Rousseau, J.-J. 636, 638 reality 456; Ornament of the Middle Way
rūpa see form (Madhyamakālaṃkāra) 463, 465–7; synthesis
of Madhyamaka and Yogācāra 472–5; tantric
śabda see verbal testimony works attributed to 476n7; threefold criterion
sahopalambhaniyama (invariable for conventional truth 470–2; use of scales of
co-apprehension): Dharmakīrti on 317, analysis 465
400, 433; Jitāri’s critique of Dharmotta’s Śāntideva 511–27; Bodhicaryāvatāra
interpretation of 548–9; Jñānaśrīmitra on 567; (Introduction to the Practice of Awakening)
Kamalaśīla’s defense of 429, 430, 433, 435; 511, 512–16, 518–25, 653, 656–9; on
Śubhagupta on 433 bodhicitta and the bodhisattva’s vow 512–16;
675
Index
contributions to moral philosophy 525; on semantics: Madhyamaka critique of realist

generosity 516–18, 525; influence on Dalai 239; negative 298; realist theories of 296–7;
Lama 653, 656–9; on patience 518–23, 525; unifying with epistemology and dialectics 291
on relationship between emptiness and the Siddhārtha Gautama 27–45; as distinct from
pāramitās 523; on relationship between Gotama Buddha 10; life according to
metaphysics and ethics 523–5; on relationship Aśvaghoṣa 85–87
between no-self and virtue 524, 525 single vehicle (ekayāna) 70, 484, 619, 625, 627
Saraha 124–37; as anti-philosopher 124, 126–9; skandha/khanda see five aggregates
on four buddha bodies 134–5; influence of Socrates 9, 107, 110n15
buddha nature discourse on 135; as mahāsiddha Spinoza, B. 506
124–6; as mystic philosopher 129–35; and Śrīgupta 470, 463, 464, 470, 474
Nietzsche 124, 129; on non-recollection Sthiramati 376–92; commentary on
(asmṛti) 132; rejection of caste 126; on Vasubandhu’s Triṃśikāvijñaptikārikā (Thirty
subitism and gradualism 131–4; as teacher of Verses on Appearance) 381–8, 459; critique
mahāmudrā 125; on two truths 130–1 of Madhyamaka 384, 385–6; figurative theory
Sāriputta/Śāriputra 110n12, 144, 160, 200, of language 382–8, 389n20; on the grasping
464, 488 of signs 36; as innovative commentator
Sarvāstivāda 142; *Abhidharmāṣṭagranthaśāstra 377–8, 381–2; interpretation of imagination
(Abhidharma Treatise in Eight Sections) (rtog pa) 459; reading of Nāgārjuna 240
173–5, 180–3, 276; Aśvaghoṣa’s association storehouse consciousness see ālayavijñāna
with 83–4; eight Abhidharma topics Strawson, P. F. 242
176–9; emergence as a distinct group 173; Śubhagupta 427–38; as author of short digests
Jñānaprasthāna (Source of Knowledge) as debate manuals 427, 429; demonstration of
83, 173–5, 184; and the path of vision objects external to the mind 432–6; embrace
(darśanamārga) 176; position on past, of nirākāravāda 427, 432, 433; externalism
present, and future entities 166–8, 173–4; rebutted by Śāntarakṣita and Kamalaśīla
rejection of the pudgala (“person”) 190; 429, 433; on omniscience 429; ontological
and the second council 183–4; Vijñānakāya externalism (bāhyārthavāda) of 427; quoted
(Consciousness Group) 166, 167, 170 by Kamalaśīla 433, 434; refutation of
satyadvaya see two truths Vijñānavāda 432–6; reliance on Dharmakīrti
satyākāravāda (theory that cognition is 431, 436; Śāntarakṣita and Kamalaśīla’s
endowed with a true image): Jitāri as adherent debt to 428–9; and Sautrāntika views 434–5;
of 549; targeted by Śāntarakṣita 468–9; as a understanding of the image (ākāra) in
Tibetan doxographical classification 477n19, awareness 435
609; see also alīkākāravāda substance (dravya) 90, 218, 286, 407, 433
Sautrāntika: Aśvaghoṣa as anticipating 84, 93; suffering (duḥkha, dukkha): Ambedkar on
basic doctrines of 286; as emerging from the social suffering 636, 640, 641–3; Aśvaghoṣa
Sarvāstivādins of Bactria and Gandhāra 175; on relationship between suffering and the
on perception 289; provisional adoption by self 87, 91, 92, 94n31; Dalai Lama XIV
Śāntarakṣita 465 on relationship to compassion 657, 658–9;
Sautrāntika-Madhyamaka 459–60, 491 Dharmakīrti on relationship to desire 305,
Sautrāntika-Svātantrika-Madhyamaka 347, 307–8, 319; in early Buddhist teachings 23;
476n4 Kamalaśīla on relationship to compassion
scriptural authority: Abhayākaragupta on 483, 490; Kātyāyanīputra on 178, 180–3;
Mahāyāna Sūtras and 619–20; Buddhist Mahāyāna understanding of 588; in the
philosophers skeptical of 246; Mahāyāna Milindapañha (Questions of Milinda) 103,
philosophers on 627n7; Tibetan “triple 104; Nāgārjuna on 214–15, 225, 228n40;
analysis” (dpyad pa gsum) method for 245; Ratnākaraśānti on ignorance and 588, 596,
see also inference, scripturally-based; 597; Śāntideva on causes and elimination of
religious authority 513–15, 518–25, 526n2; Saraha on 130; the
Self (ātman; soul): arguments against in the tantric buddha on 50–2, 59
tantras 50; Aśvaghoṣa’s arguments against śūnyatā see emptiness
90–2; fundamental question of philosophy svabhāva/sabhāva (self-nature, own nature,
188; Nyāya view of 90; Saṃkhya view of 88, intrinsic nature, particular nature): connection
90; Vaiśeṣika view of 89–90; Vasubandhu to emptiness 67, 219, 241, 351; critiqued
on 271–3; view that there is a self in Sarvāstivāda Abhidharma 177–8, 180;
(satkāyadṛṣṭi) 178 lacking in what is dependently arisen
676
Index
218; in Mahāyāna philosophy 236; in Pali two truths (satyadvaya): Abhayākaragupta on

Abhidhamma 149, 151, 153, 189 620–3; Bhāviveka on 354–5; Candrakīrti on
svasaṃvedanā/svasaṃvitti see reflexive 405, 409–11, 412–17; Jñānagarbha on 450–8;
awareness Kamalaśīla on 484; Nāgārjuna on 230n57;
Svātantrika-Madhyamaka: Maitreya’s Śāntarakṣita on 465–7, 472–5; in Saraha 130;
Abhisamayālaṃkāra (Ornament of Clear see also conventional truth; ultimate truth
Realization) classified as 67; Bhāviveka
classified as 347–8, 476n4; Jñānagarbha Uddyotakara 90, 285, 318, 465
classified as 450; Kamalaśīla classified ultimate meaning (paramattha) 151–2
as 481, 491–2; Śāntarakṣita classified ultimate truth (paramārthasatya; ultimate
as 463, 476n4, 491–2, 615; as a Tibetan reality): Bhāviveka on 352, 354–5, 366, 450,
doxographical distinction 491–2; 453–4; Candrakīrti on 413, 415, 416–17;
see also Prāsaṅgika-Madhyamaka as ineffable (aparyāya) and inconceivable
(acintya) 275, 453; Jñānagarbha on 450,
tantra 46–63; as contributing factor to the 452–4; particulars as existing ultimately 311,
growth of apologetics 347; defined as the 441; Ratnākaraśānti on 286; see also two
cause, result, and method for attaining truths; ultimate meaning
buddhahood 47; focus on bliss and the body universal (sāmānya, sāmānyalakṣaṇa, jāti;
46, 59–60; self-generation as a deity in 50; universal characteristic): Dharmakīrti on
subtle body (sukṣmaśarīra) 125, 127, 132, 312–14, 441; as conventionally existent
133, 135; Vimalaprabhāṭīkā (Stainless Light (saṃvṛtisat) 311; Jitari’s critique of 550, 551,
Commentary) 46, 48, 50–2, 58, 61 553; as the object of inference 311; realist
Tantric Buddha 46–62; having the world as semantic theory of 296
their body 46; indistinct from one another 39; universal momentariness 270, 440, 573
neither male nor female 46; as philosopher
and doxographer 47–52; remaining active in Vaibhāṣika: on atoms 433; Dharmapāla’s
the world 624 critique of 365; on perception 289;
Tathāgatagarbha see Buddha Nature Śāntarakṣita’s critique of 465; subgroup of
tetralemma see catuṣkoṭi the Sarvāstivāda 184; Vasubandhu’s critique
Theravāda: Abhidhamma 141, 143, 149, 160, of 183, 270, 433;
161–9, 196; account of Vātsīputrīya/Sāṃmitīya see also Mahāvibhāṣa; Sarvāstivāda
195; canon 199; and Cāttaṉār 117–18; Early Vaiśeṣika: Āryadeva’s critique of 236, 238;
European scholars of 192; influence of Aśvaghoṣa’s critique of 81, 89–90, 93;
Buddhaghosa on 325, 328; as rejecting the Bhāviveka’s critique of 346, 349, 353;
doctrine of pudgala 192; see also Buddhaghosa Candrakīrti’s critique of 410, 414; Cāttaṉār’s
thesis (pratijñā): Āryadeva’s rejection of any critique of 119, 120; Dharmakīrti’s critique
239; Atiśa on the Madhyamaka rejection of of 312; Dharmapāla’s critique of 364, 365;
607; Dharmapāla on 364; Dignāga on 364; Dignāga’s critique of 285, 288; Śāntarakṣita’s
Nāgārjuna as eschewing 221–2; Vasubandhu critique of 465; Śubhagupta’s critique of 430
as establishing through refutation of vāsanā (habitual propensity, imprint, latent
opponents 364–5 impression): appearances arising from
three natures (trisvabhāva): Abhayākargupta 389n17, 588, 591, 594; as contained by the
on 621; Āryadeva on 240; Asaṅga on 255, ālayavijñāna 593; as contributing to the
256, 257, 259; Bhāviveka’s critique of 366; construction of a continuum (santāna) 442;
Dharmapāla’s response to Bhāviveka’s of expressions 259; of hearing the Buddhist
critique of 365–7; as framework for scriptures 259; of the intention to refrain from
understanding Dignāga’s epistemology wrong deeds 379–80; in the tantras 52, 55, 62;
285–8; Jñānagarbha on 453; Kamalaśīla on of the view of the self 259
484; Maitreya on 75n25; Ratnākaraśānti Vasubandhu 266–83; Abhidharmakośabhāṣya
on 286, 602; three forms of naturelessness (Treasury of Abhidharma Commentary)
(niḥsvabhāvatā) 277–8, 484; and the three 183, 266–8, 434; arguments against a self
turnings of the wheel of Dharma 277; 268–70; commentary on works of Maitreya
Yogācāra interpretation of non-seeing 68–70; on dependent origination 274; as
(adarśana) in light of 453 engaging multiple inherited traditions 266–8;
trisvabhāva see three natures explanatory simplicity in 270; on language
Tsongkhapa 199, 242, 475, 476n5, 482, as figurative 274–81; as a Mahāyāna
616, 651 advocate 275; Pañcaskandhaprakaraṇa
677
Index
(Treatise on the Five Aggregates) 36, 378; Weber, M. 637

as not providing a positive thesis 364; wisdom (prajñā, paññā; understanding):
on reality as inconceivable 276, 280; on Bhāviveka’s definition of 354, 355; born of
scriptural interpretation 273–4; on store hearing or study (śrutamayīprajñā) 355, 508;
consciousness 279–80; Vādavidhi (Method born of meditation (bhāvanāmayīprajñā)
for Argumentation) 284, 285, 292, 299n2; 356, 508; born of reflection, or thought
Viṃśikā (Twenty Verses) 266, 276, 362–4, (cintamāyā, cintāmayīprajñā) 341, 355, 508;
433, 467, 476n7, 588; Vyākhyāyukti (Proper Buddhaghosa on 337–43; Śāntideva on the
Mode of Exposition) 41n11, 70, 273, 489; on relationship between ethics and 523–5; as
the world as mind only 274–6 virtuous practice for a happy life 223
Vatsīputrīya/Sāṃmitīya 188–203; and Wittgenstein, L. 231n64, 276
dependent arising 200; described by Wŏnch’ŭk 366
Xuanzang and Yijing 193; doctrines
presented by Bhāviveka 194, 197; doctrines Xuanzang: and Dharmapāla’s legacy in East
presented by Daśabalaśrīmitra 197, 198–9; Asia 370; disciples of 364, 365, 367; Records
doctrines presented by Vasubandhu 194; of the Western Regions 252; as source for
methodological issues in the study of 192–4; studying Buddhist Vinaya schools 193, 194,
status as Buddhist or non-Buddhist 199–200; 197; translator 185n12, 240, 252–3, 255,
theory of the pudgala rejected by Theravāda 299n5, 362, 361, 362–3, 388n3
192; as umbrella name for diverse affiliated
schools 191; see also pudgalavāda Yāmari 528, 537n1, 558, 529, 572
Veda: Bhāviveka’s reference to 353; Yaśomitra 434
Dharmakīrti’s critique of the authority of 309, Yijing 193, 194, 197, 361
314; Jitāri’s critique of the authority of 551; Yogācāra: antirealist influence on Dharmakīrti
Jitāri’s proof of human authorship of 553; 304; Asaṅga’s contributions to 255–61;
Śubhagupta’s critique of the authority of 430, Aśvaghoṣa and 84, 93; Atiśa as critic of
431–2; tantric critiques of the authority of 605; Candrakīrti’s critique of 405; criticized
48–9 by Bhāviveka 352–3, 384; Dharmapāla’s
vedanā see feeling contributions to 361–71; early authors
Vedānta 349, 350, 353, 464, 465, 485, 653 on the matter-mental distinction 379; as
verbal testimony (śabda): authority critiqued a framework for understanding Dignāga
by some Buddhist philosophers 246; Jitāri’s 285–9; Haribhadra’s integration of
refutation of 553; reduced to inference in 507; as holding there is nothing outside
some Buddhist epistemology 209, 284–5, consciousness 286; influence on Saraha 135;
291, 464; see also scriptural authority Jitāri’s defense of 551, 554; Jñānagarbha’s
Vibhajjavāda 154 critique of 453; Jñānagarbha’s use of
views (dṛṣṭi): erroneous views (mithyādṛṣṭi) 459–60; Kamalaśīla’s integration of 484,
in Sarvāstivāda 179; as illusory in Yogācāra 486; metaphysical discourse of 386–8;
277; Madhyamaka as the elimination of 218, ontological versus epistemological
221, 350, 352, 354; Nāgārjuna on 218, 220–2; interpretation of 276; in contrast to
non-Buddhist speculative as tainted 350 Madhyamaka 385, 390n22; Ratnākaraśānti’s
vijñāna/viññāṇa see consciousness interpretation of 587–91; Śāntarakṣita’s
vijñapti (information, representation): Asaṅga provisional acceptance of 465, 470–2;
on 255–6, 259–61; Sthiramati’s innovative Śāntarakṣita’s synthesis with Madhyamaka
understanding of 378–9 472–5; Sthiramati’s formation of 376–88;
vijñaptimātra (representation only): Asaṅga on Vasubandhu’s contributions to 274–81;
260–1; Dharmakīrti on 316; Dharmapāla on Vasubandhu’s place in 267–8; see also
363; Jitāri’s proof of 551; Sthiramati on 459; ālayavijñāna; three natures
Vasubandhu on 268, 275–7, 280; Yogācāra Yogācāra-Madhyamaka 208, 240, 460, 464,
doctrine 588–9; see also mind-only 472, 474, 491
Vimuktisena 196, 197, 501, 503, 618 Yogācāra-Svātrantrika-Madhyamaka 476n4,
viṣaya see object 491, 615
vyavahāra (everyday activity) 311, 395, 396, yogic perception (yogipratyakṣa) 286, 299n11,
414, 435, 533 315, 444, 445, 483, 623, 624
678

The Routledge Handbook of Indian Buddhist Philosophy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Routledge Handbook of Indian Buddhist Philosophy

Uploaded by

Copyright:

Available Formats

‘A splendid and comprehensive study of Buddhist philosophy in its country of origin.

‘This carefully edited volume provides an understanding of Indian Buddhist philosophy

William Edelglass is Associate Professor of Philosophy at Emerson College and Director of

Sara McClintock is Associate Professor of Religion at Emory University, USA.

Routledge Handbooks in Philosophy are state-of-the-art surveys of emerging, newly refreshed,

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY, POLITICS, AND

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF SEX AND SEXUALITY

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF FRIENDSHIP

THE ROUTLEDGE HANDBOOK OF INDIAN BUDDHIST PHILOSOPHY

For more information about this series, please visit: www.routledge.com/Routledge-Hand

Edited by William Edelglass, Pierre-Julien Harter

Thought is not calmed without speaking words.

1 Gotama Buddha: His Quest and His Teachings 11

2 Siddhārtha Gautama: Beyond the Historical Figure 27

3 The Tantric Buddha: Primordial Buddhas as Philosophical Authors 46

4 Maitreya: The Future Buddha as an Author 64

5 Aśvaghoṣa: The Dawn of Indian Buddhist Philosophy 83

6 The Milindapañha: How to Use a Philosophical Resource and Find a

7 Cāttaṉār: Poet-Philosopher in Tamiḻ 113

8 Saraha: The Anti-Philosopher as Philosopher 124

9 The Dhammasaṅgaṇī and Vibhaṅga: The Perfectly Awakened

10 Moggaliputta Tissa’s Points of Discussion (Kathāvatthu): Reasoning

11 Kātyāyanīputra and the Large Commentary (Mahāvibhāṣā): The

12 The Vātsīputrīya/Sāṃmitīya: Buddhist Personalism as a Mainstream

13 Nāgārjuna: Dependent Arising Without Any Thing Arising 211

14 Āryadeva: Quietism and Buddhist Ethics 236

15 Asaṅga: Great Systematizer of Yogācāra Thought 252

16 Vasubandhu: Mainstream and Mahāyāna 266

17 Dignāga: Early Innovator in Buddhist Epistemology 284

18 Dharmakīrti: Philosopher and Defender of the Faith 303

Introduction to Part 5 325

19 Buddhaghosa: Phenomenology, Hermeneutics, and Understanding 328

20 Bhāviveka: Madhyamaka Dialectic, Doxography, and Soteriology 346

21 Dharmapāla: A Janus-Faced Interpreter of Yogācāra? 361

22 Sthiramati: A Yogācāra Commentator and Innovator 376

23 Devendrabuddhi and Śākyabuddhi: Dharmakīrti’s First Commentators 393

24 Candrakīrti: Gardener of Sky-Flowers 404

Introduction to Part 6 423

25 Śubhagupta: An Externalist Outsider Within the Dharmakīrtian

26 Dharmottara: Systematic and Innovative Commentator 439

27 Jñānagarbha: Two Truths Theory, Gradualism, and Mādhyamika

28 Śāntarakṣita: Climbing the Ladder to the Ultimate Truth 463

29 Kamalaśīla: Mādhyamika Champion of Magical Reason 480

30 Haribhadra: The Voice of Perfect Wisdom 497

31 Śāntideva: Virtue on the Empty Path of the Bodhisattva 511

32 Prajñākaragupta: Buddhist Epistemology as the Path to the Wisdom

Introduction to Part 7 543

33 Jitāri: A Later Buddhist Master of Debate 547

34 Jñānaśrīmitra: Variegated Non-Duality 558

35 Ratnakīrti: Aligning Everyday Experience with Momentariness

36 Ratnākaraśānti: The Illumination of False Forms 587

37 Atiśa: The Great Middle Way of Mere Appearance 601

38 Abhayākaragupta: A Last Great Paṇḍita 615

Introduction to Part 8 633

39 B. R. Ambedkar: Justice, Religion, and Buddhist Political Philosophy 635

40 The Dalai Lama XIV: A Modern Indian Philosopher 650

Allison Aitken is Assistant Professor of Philosophy at Columbia University, USA. Her

Amber D. Carpenter is Associate Professor at Yale-NUS College, Singapore. She publishes

William Edelglass is Associate Professor of Philosophy at Emerson College, USA, and

Stephen Harris is Assistant Professor (Universitair Docent) at Leiden University’s Institute

Sonam Kachru is Assistant Professor in the Department of Religious Studies, University of

Kazuo Kano is an Associate Professor at Komazawa University, Japan. He is working on

Sara McClintock is Associate Professor of Religion at Emory University, USA. Their

Shinya Moriyama is Professor in the Department of Philosophy at Shinshu University, Japan.

1 Gotama Buddha: His Quest and His Teachings 11

2 Siddhārtha Gautama: Beyond the Historical Figure 27

3 The Tantric Buddha: Primordial Buddhas as Philosophical Authors 46

4 Maitreya: The Future Buddha as an Author 64

5 Aśvaghoṣa: The Dawn of Indian Buddhist Philosophy 83

7 Cāttaṉār: Poet-Philosopher in Tamiḻ 113

8 Saraha: The Anti-Philosopher as Philosopher 124

13 Nāgārjuna: Dependent Arising Without Any Thing Arising 211

14 Āryadeva: Quietism and Buddhist Ethics 236

15 Asaṅga: Great Systematizer of Yogācāra Thought 252

16 Vasubandhu: Mainstream and Mahāyāna 266

17 Dignāga: Early Innovator in Buddhist Epistemology 284

18 Dharmakīrti: Philosopher and Defender of the Faith 303

Introduction to Part 5 325

19 Buddhaghosa: Phenomenology, Hermeneutics, and Understanding 328

20 Bhāviveka: Madhyamaka Dialectic, Doxography, and Soteriology 346

21 Dharmapāla: A Janus-Faced Interpreter of Yogācāra? 361

22 Sthiramati: A Yogācāra Commentator and Innovator 376

23 Devendrabuddhi and Śākyabuddhi: Dharmakīrti’s First Commentators 393

24 Candrakīrti: Gardener of Sky-Flowers 404

Introduction to Part 6 423

26 Dharmottara: Systematic and Innovative Commentator 439

28 Śāntarakṣita: Climbing the Ladder to the Ultimate Truth 463

29 Kamalaśīla: Mādhyamika Champion of Magical Reason 480

30 Haribhadra: The Voice of Perfect Wisdom 497

31 Śāntideva: Virtue on the Empty Path of the Bodhisattva 511

Introduction to Part 7 543

33 Jitāri: A Later Buddhist Master of Debate 547

34 Jñānaśrīmitra: Variegated Non-Duality 558

36 Ratnākaraśānti: The Illumination of False Forms 587

37 Atiśa: The Great Middle Way of Mere Appearance 601

38 Abhayākaragupta: A Last Great Paṇḍita 615

Introduction to Part 8 633

39 B. R. Ambedkar: Justice, Religion, and Buddhist Political Philosophy 635

40 The Dalai Lama XIV: A Modern Indian Philosopher 650