JIABS

Journal of the International

Association of Buddhist Studies

Volume 33 Number 1–2 2010 (2011)

The Journal of the International Association of Buddhist Studies (ISSN 0193-600XX) is the organ of the International Association of Buddhist Studies, Inc. As a peer-reviewed journal, it welcomes scholarly contributions pertaining to all facets of Buddhist Studies. JIABS is published twice yearly. As announced at the XVIth IABS Congress in Taiwan, the JIABS is now available online in open access at http://archiv. ub.uni-heidelberg.de/ojs/index.php/jiabs/ index. Articles become available online for free 60 months after their appearance in print. Current articles are not accessible online. Subscribers can choose between receiving new issues in print or as PDF. We are kindly requesting all authors that could be opposed to this decision to inform the Editors by June 2012. Manuscripts should preferably be submitted as e-mail attachments to: editors@iabsinfo.net as one single file, complete with footnotes and references, in two different formats: in PDF-format, and in Rich-Text-Format (RTF) or OpenDocument-Format (created e.g. by Open Office). Address books for review to: JIABS Editors, Institut für Kultur- und Geistesgeschichte Asiens, Apostelgasse 23, A-1030 Wien, AUSTRIA Address subscription orders and dues, changes of address, and business correspondence (including advertising orders) to: Dr Jérôme Ducor, IABS Treasurer Dept of Oriental Languages and Cultures Anthropole University of Lausanne CH-1015 Lausanne, Switzerland email: iabs.treasurer@unil.ch Web: http://www.iabsinfo.net Fax: +41 21 692 29 35 Subscriptions to JIABS are USD 55 per year for individuals and USD 90 per year for libraries and other institutions. For informations on membership in IABS, see back cover.

EDITORIAL BOARD
KELLNER Birgit KRASSER Helmut Joint Editors BUSWELL Robert CHEN Jinhua COLLINS Steven COX Collet GÓMEZ Luis O. HARRISON Paul VON HINÜBER Oskar JACKSON Roger JAINI Padmanabh S. KATSURA Shōryū KUO Li-ying LOPEZ, Jr. Donald S. MACDONALD Alexander SCHERRER-SCHAUB Cristina SEYFORT RUEGG David SHARF Robert STEINKELLNER Ernst TILLEMANS Tom

Cover: Cristina Scherrer-Schaub Font: “Gandhari Unicode” designed by Andrew Glass (http:// andrewglass.org/fonts.php) © Copyright 2011 by the International Association of Buddhist Studies, Inc. Print: Ferdinand Berger & Söhne GesmbH, A-3580 Horn

JIABS
Journal of the International Association of Buddhist Studies
Volume 33 Number 1–2 2010 (2011)

Articles William Chu The timing of Yogācāra resurgence in the Ming dynasty (1368–1643) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vincent EltsChingEr Ignorance, epistemology and soteriology – Part II

. . .

5 27

. . . . . . . . . .

Richard F. nanCE Tall tales, tathāgatas, and truth – On the “privileged lie” in Indian Buddhist literature . . . . . . . . . . . . . . . . . . . . . . . . . Alexander WynnE The ātman and its negation – A conceptual and chronologi­ cal analysis of early Buddhist thought . . . . . . . . . . . . . . . .

. . .

75

. . .

103

2

Contents

Indian Buddhist metaethics
Contributions to a panel at the XVth Congress of the International Association of Buddhist Studies, Atlanta, 23–28 June 2008

Guest editor: Martin T. Adam

Peter harvEy An analysis of factors related to the kusala/akusala quality of actions in the Pāli tradition . . . . . . . . . . . . . . . . . . . . . . Abraham vélEz dE CEa Value pluralism in early Buddhist ethics .

. . .

175 211

. . . . . . . . . . . . . . . . .

Martin T. adam No self, no free will, no problem – Implications of the Anatta­ lakkhaṇa Sutta for a perennial philosophical issue . . . . . . . . Bronwyn Finnigan Buddhist metaethics .

. .

239 267 299

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Stephen JEnkins On the auspiciousness of compassionate violence

. . . . . . . . . .

Jay L. garFiEld What is it like to be a bodhisattva? Moral phenomenology in Śāntideva’s Bodhicaryāvatāra . . . . . . . . . . . . . . . . . . . . . . Tom J. F. tillEmans Madhyamaka Buddhist ethics

. . .

333 359

. . . . . . . . . . . . . . . . . . . . . . . .

Contents

3

Miracles and superhuman powers in South and Southeast Asian Buddhist traditions
Contributions to a panel at the XVth Congress of the International Association of Buddhist Studies, Atlanta, 23–28 June 2008

Guest editor: David V. Fiordalis

David V. Fiordalis Miracles in Indian Buddhist narratives and doctrine

. . . . . . . . .

381 409

Bradley S. Clough The higher knowledges in the Pāli Nikāyas and Vinaya.

. . . . . .

Kristin sChEiblE Priming the lamp of dhamma – The Buddha’s miracles in the Pāli Mahāvaṃsa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Patrick PrankE On saints and wizards – Ideals of human perfection and power in contemporary Burmese Buddhism . . . . . . . . . . . . Rachelle M. sCott Buddhism, miraculous powers, and gender – Rethinking the stories of Theravāda nuns . . . . . . . . . . . . . . . . . . . . . . . . . Luis O. gómEz On Buddhist wonders and wonder-working . • Notes on the contributors

. . .

435

. . .

453

. . .

489 513

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

555

The timing of Yogācāra resurgence in the Ming dynasty (1368-1643)
William Chu

Before proceeding to the main thesis of my paper, I need to review some well known facts about Buddhist-Confucian dynamics in the late imperial era. The rationalist Song Neo-Confucians known traditionally as the Cheng-Zhu school (程朱; after the Cheng brothers Cheng Hao 程顥 1032–1085 and Cheng Yi 程頤 1033–1107, and Zhu Xi 朱熹 1130–1200) had asserted the exclusive orthodoxy of their tradition. They repudiated all Confucians who professed syncretistic interest or sympathies toward Buddhism and Daoism, and highlighted the doctrinal incompatibilities between Confucianism and other “heretical” religions (yiduan 異端). To Zhu Xi in particular, Buddhism was irredeemably at odds with the Confucian “sagely lineage.” Zhu’s staunchly purist stance was targeted at those Confucians who heartily professed dual allegiances to both religions. Another group of Confucians, in contrast, headed by Lu Jiuyuan 陸九淵 (1139–1193), that eventually rose to become Zhu’s major rival for Confucian orthodoxy, also harbored deep reservations about Buddhism and reiterated the need to vigilantly fend off the latter’s spiritual allure. But despite Lu Jiuyuan’s self-proclaimed loyalty to Confucianism, followers of Zhu’s school often branded Lu an apostate, one who was secretly a sympathizer of Buddhism and, in fact, cherished a much buddhicized interpretation of Confucianism. In the Ming dynasty, Wang Yangming 王陽明 (1472–1529) relented on Confucian exclusivism. Although still upholding the supreme status of Confucianism, Wang saw that the Buddhist training was not diametrical to Confucian enlightenment and could in fact serve as a stepping stone. His unapologetic use of Buddhist
Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 5–25

6

William Chu

jargons earned him infamy among his Confucian colleagues, who jeeringly referred to him as the de facto successor to the “Chan of the Five dynasties,” and as one who reverently upheld the teachings of “Bodhidharma and Huineng.”1 Such accusations were not completely baseless, for Wang’s teachings were often direct paraphrase of the words of legendary Buddhist Chan patriarchs, as the following demonstrates:2
The innate conscience knows right from wrong; the innate conscience is also neither of right or wrong. Knowing right from wrong is what constitutes propriety, while being oblivious to (unbounded by) right and wrong and thereby realizing the marvelous, is [what defines] the so-called enlightenment. 良知知是知非, 良知無是無非. 知是知非, 即所謂規矩. 忘是非而得其 巧, 即所謂悟.3

In this case, Wang Yangming summarized his understanding of Confucian morality in verses that were unmistakably appropriated from the Buddhist Platform sūtra. Compare what Wang said with what Hanshan Deqing 憨山德清 (1546–1623 C.E.) – the most prolific Buddhist writer in the Ming – wrote, their similarity becomes apparent:
[People] do not know that the two opposite polarities of good and evil are in fact dualistic dharmas coming from the outside [of one’s innate nature], and have nothing to do with the original essence of our selfnature. That is why those who do evil in the world could at times mend their ways and become good, and good people can also be converted to evil’s way. This is sufficient to prove that [worldly] virtues could not be stable or reliable. For that if one did not cultivate goodness all the way [to eventually consummate in enlightenment], what is apparently moral is really not ultimately moral. As for the presently expounded ‘highest good,’ it constitutes enlightening and illuminating the veriCited in Araki (Rushi) 1978: 181. Although Robert Sharf has argued that that the shared terminology of Confucianism, Daoism, and Buddhism might not have been due to deliberate borrowing (see Sharf 2001), in the case of Wang the assimilation of Buddhist ideas and terms was quite deliberate and self-consciously done, as would be made clear in this study. 3 Cited in Araki (Rushi) 1978: 384.
2 1

Yogācāra resurgence in the Ming dynasty

7

table essence of the self-nature, which is originally devoid of either good or evil. 不知善惡兩頭, 乃是外來的對待之法, 與我自性本體了不干涉. 所以 世人作惡的可改為善, 則善人可變而為惡, 足見善不足恃也. 以善不 到至處, 雖善不善 … 今言至善, 乃是悟明自性本來無善無惡之真 體.4

We may recall, however, that Zhu Xi found precisely such Chan teaching of “there is neither right nor wrong” (wushi wufei 無是無 非) to be logically and morally reprehensible. Zhu interpreted what might be explained by Buddhists as “transcendence from right and wrong” as reckless antinomianism, a disregard for conventional morality. He said:
The difference between we the Confucians and the Buddhists lies in that, we Confucians have reasonable rules and principled guidelines. For the Buddhists, they have none of these. 吾儒所以與佛氏異者, 吾儒則有條理有準則. 佛氏則無此爾.5

Antinomianism and moral laxity were some of the most denigrating and inveterate polemics the Confucians had historically maintained about Buddhism, Wang was conscious of the danger of becoming labeled as an antinomian or moral nihilist. The strategy he employed to differentiate his own position from that of the allegedly amoral Buddhism was, ironically, also the same one many Buddhists resorted to in vindicating Buddhism of this same charge. Wang Yangming posited a bipartite approach where he would put forth both the need to maintain moral rigorousness and the transcendence from inflexible dogmatism. Wang Yangming described his understanding of the “innate knowledge of the good” (liangzhi 良知) very much in terms of the language of “Buddha-nature.” He posited the Mind or the “innate knowledge” as an all-encompassing, ontological basis for both good and evil. Yet the Mind is only functionally “actualized” when good is cultivated. In the same way, Chan followers had always described the Buddha-nature as having both an ontological aspect and a functional aspect. Example: For
4 5

Hanshan dashi mengyou ji, vol. iv, p. 2380. Cited in Araki (Rushi) 1978: 379.

8

William Chu

Zongmi 宗密 (780–841), the Chan patriarch in the Tang dynasty, the “essence” aspect of the mind has to be emphasized along with the “responsive functioning”6 aspect with the same vigor, if both the transcendent and the responsive nature of the Mind are to be maintained.7 And ever since Neo-Confucian’s attack on the alleged Buddhist amoralism, Buddhists had become especially cautious and vocal when articulating their moral stance, arguing that morality without transcendence is ritualism or mundane Dharma, and that transcendence without morality is false enlightenment. Yunqi Zhuhong 雲棲株宏 (1535–1615) was reiterating this dominant Buddhist doctrinal ethos in the Ming dynasty when he made the same point:
Although the mind is originally luminous, yet as one does good or evil deeds, their traces will make the mind soar high or sink to the ground…How can one say that evil deeds do not matter simply because the mind in its essence cannot be designated as good or evil? If one is addicted to the biased view of emptiness, he will deviate from perfect understanding. Once you realize that both good and evil are nonexistent, [it is all the more compelling that] you should stop evil and do good.8

Using unmistakably Chan language, Wang Yangming also explained that at the ontological aspect, the “innate knowledge” transcends the absolutism and dualism of mundane values; but at the functional level, it is ever discerning about good and evil, and actively pursues good and avoids evil. This two-tiered scheme that validates transmundane wisdom without sacrificing the practical need to defer to conventional virtues9 was a common Buddhist motif that could be traced back not only to the Platform sūtra, but

Peter Gregory’s analysis of the historical conditions and mentality that gave rise to this comparable bipartite moral scheme (Gregory 2002: 237– 244) contains many interesting parallels to the Ming scenario. 7 Araki Kengō (1975: 45) suggested that the “solid wall of principle” erected by Zhu runs the danger of “curbing the vitality” of the “Mind.” 8 Cited in Greenblatt 1975: 109. 9 David Kalupahana (1992: 60–67) had explained the sustained Buddhist attempts at avoiding “absolutism.”

6

Yogācāra resurgence in the Ming dynasty

9

also to such early texts like the Dhammapada.10 Although Wang Yangming might be subverting Zhu’s moral absolutism in substantive ways, by bringing such a blatantly Buddhist interpretation into Confucianism, Wang still maintained Zhu’s polemic that [Chan] Buddhism is morally degenerate. In other words, even as Wang simultaneously relegated Buddhism and Zhu Xi’s intellectual legacy – by charging that Buddhism was hopelessly oblivious to and delinquent of secular moral duties, and that Zhu was trapped in inflexible dualistic thinking – Wang’s philosophical views overlapped with important Buddhist ideas popular in the Ming dynasty. His position was much closer to Chan Buddhist ideology than he allowed himself to admit.11 Wang Yangming was responsible for still another Confucian transformation that was to prove conducive to the upsurge of Mingdynasty syncretism. One of the most powerful appeals Buddhism had in distracting, if not converting, some of the best minds from the Confucian establishment was its systematic outline of a spiritual mārga – the cultivational and soteric technology and its promise of ultimate transformations through the application of its prescribed techniques. In the face of Buddhism’s systematic, graduated program of moral and meditative training, many Confucians could not but concede to Buddhist superiority in soteric sophistication. The intricate and elaborate ways in which Buddhists conceived of their spiritual path was the result of the traditional emphasis Buddhists placed on its careful formulation. Robert Buswell and Robert Gimello had explained this emphasis in the following way:
[T]he concept of ‘the path’ has been given in Buddhism an explication more sustained, comprehensive, critical, and sophisticated than that provided by any other single religious tradition…Throughout the
10 The Platform sūtra contained the following passage that illustrates such a two-tiered scheme for morality: 無是無非, 無善無惡…用即了了分明…邪來 煩惱至, 正來煩惱除. 邪正俱不用…悟則剎那間. Similarly, the Dhammapada urged the practice of good yet posited that only by being unattached to even what is the good can one truly be called a “spiritual aspirant (Brahmin).” Compare verse 183 and 412 in Dhammapada, Kaviratna trans. pp. 73 & 159. 11 See some examples of Wang’s critical remarks in Chen (Rongjie) 1984: 77–80.

10

William Chu

two-and-a-half millennia of its pan-Asian career, Buddhism has been consistently explicit in declaring itself to be, above all else, a soteriology, a method of salvation, rather than, say, a creed. Its unflagging concentration on ‘the path,’ whether for the purpose of advocating and charting that path or for the purpose of qualifying and criticizing it, has not only led to the careful and detailed delineation of numerous curricula of religious practice and to the privileging of such delineation over other modes of Buddhist discourse.12

In order to compete, the re-systematization endeavor of the Song Neo-Confucians was to delineate a comparably enticing system of personal cultivation (gongfu lun 功夫論) that could outstrip the near dominance Buddhists had always held in this domain. One of the key leitmotifs Confucians had toiled for centuries to configure in its soteriological system concerned a practice known as the “extension of knowledge” (zhizhi 致知).13 It was one phase of training in a series of graduated steps leading to ever higher spiritual and ethical goals as prescribed in the Confucian classic the Great Learning (Daxue 大學). Though vague and abstruse in its language, the text provided a rare indigenous outline of soteric path for the Neo-Confucians, who were in a desperate need to “discover” something from their own textual tradition rather than appropriating from Buddhism what they supposedly ostensibly lacked. Zhu Xi’s canonization and propagation of the so-called Four Classics (which included the Great Learning) very likely have been propelled by such a mentality. Through this act of canonization, he had undoubtedly changed the character of Confucianism in posterity. The fact that the Four Classics were readily and heartily received by fellow Confucians as the sine qua non primers for the tradition very likely had something to do with their usefulness to buttress Confucianism in those most glaringly deficient areas. With the exception of the Analects, all the Four Classics (the Analects,
Buswell and Gimello 1992: 2–3. It should be noted that Lu Jiuyuan was one of the first to bring about this transformation in the Song. In many ways Wang Yangming simply rediscovered and fine-tuned Lu’s system rather than having single-handedly invented the many implications of an idealist philosophy. See Chan 1963: 572–573; Liu 1964: 165–166; and Qian 1962: 137.
13 12

Yogācāra resurgence in the Ming dynasty

11

the Mengzi, the Great Learning, and the Doctrine of the Mean) in fact were only elevated to their revered prominence from relative obscurity after Zhu Xi’s unreserved commendation of them.14 Each of the three newly promoted Confucian texts served an important soteric purpose. The Mencius (Mengzi 孟子) discusses the Confucian ideal of “sagehood” and the actualization of the “heavenly virtues.” The idea of attaining sagehood as outlined in the Mencius served as a worthy competing ideal to the Buddhist notions of Buddhahood and enlightenment. The Great Learning spelled out the specific praxis involved for that actualization: one sets out to realize the Confucian goal of bringing harmony to the world through a stepby-step self-transformation. The Doctrine of the Mean (Zhongyong 中庸) served to explicate the mechanism and principles underlying those Confucian systems of spiritual cultivation: the principle of emotional and practical moderation was advised in this text as the connecting theme for Confucian practice. It also provided much of the philosophical expressions with which Zhu built his metaphysical theories such as that of the “heavenly nature” and the “universal principle”15 – metaphysical theories that were in one capacity intended to compete with the elaborately constructed Buddhist cosmology. In the Great Learning, there is a passage on the sequence of stages to be accomplished by those cultivators intent on “illuminating the luminescent virtue” and “bringing harmony to the world.” The first two of the stages are the “investigation of things” (gewu 格 物) and, as we have previously discussed, “the extension of knowledge.” As many scholars have noted already, Zhu’s understanding of this process was that it is an inductive one. He believed that only through wide learning and unceasing exploration of the naMencius was beatified by Zhu himself to become the greatest Confucian sage second only to Confucius (yasheng 亞聖). Both the Great Learning and the Doctrine of Mean were excerpts from the Book of Rites by the Younger Zai (Xiaozai liji 小戴禮記) singled out by Zhu for their pertinence to his vision of a new Confucianism. 15 See Zhu Xi’s exegesis on this part of the Book in Chen (Rong) 1984: 22–23.
14

12

William Chu

ture of external things would the seeker arrive at a universal principle which pervades them all.16 He considered Buddhism’s introspective and subitist quest undertaken in the “mind ground” to be too introverted and perfunctory to generate any concrete, reliable, generalizable knowledge. By collapsing what should otherwise be a thorough investigation of the myriad phenomena into a single, idealist principle, Buddhism appeared to Zhu to be oversimplifying a serious learning process and too complacently engrossed in the self. One of Zhu’s supporters in the Ming expressed this distaste toward Buddhism’s allegedly reductionist approach to investigating things:
Some practitioners had only heeded to the aspect of the ‘oneness of principle’ while overlooking the aspect of its divergent ramifications. In the case of Buddhism, even its professed ‘oneness of principle’ [that it so favored] differed from the ‘one principle’ that was advocated by the [Confucian] sages. 學者或只理會這 ‘理一’ 處, 便遺了分殊. 如釋氏之學, 則連所謂 ‘理 一’ 者, 非聖人之 ‘理一’ 17 矣.

This was where Wang Yangming introduced yet another important revolutionary modification of Zhu’s epistemic approach. While Zhu Xi claimed to have “applied his mind in search of the principle amidst the various phenomena and things” 以吾心而求理於 事事物物之中,18 Wang braved the antithetical position. Wang sug-

16 Zhu Xi’s logic was that, “The so-called nature is the myriad principles scattered in various places. This is what makes it the nature [of all things]” 性是許多理散在處為性. Cited in Qian 1962: 117. Therefore only by a tireless investigation into various things would the principle be revealed. Zhu’s conclusion was that, “As for our attempt to extend our knowledge, [the key] lies in searching to the limit of things and thereby fully reveal their principle… This is the reason that the beginning-leveled teaching of the Great Learning always required the apprentice to pursue the investigation of all things without exceptions under Heaven, basing on principles that one had already comprehended, in order to extend it to its utmost limit.” 言欲致吾之知, 在極物而 窮其理也 … 是以大學始教, 必使學者即凡天下之物, 莫不因其已知之理而益 窮之, 以求致乎其極. Ibid., p. 120. 17 Cited in Araki (Rushi) 1978: 389. 18 Cited in Sun, Liu and Hu 1995: 280.

Yogācāra resurgence in the Ming dynasty

13

gested that “the Way of the sages is already endowed in my nature. Those who directed their search for the principle in objects and phenomena are misled” 聖人之道, 吾性自足, 向之求理於事物者誤 也.19 Wang thought that a deductive and introspective “extension of knowledge” could avoid the pitfalls of becoming inevitably “fragmentary and aimless” as was the case in Zhu’s externally-directed system. Instead of looking outward, Wang felt that nothing other than the mind provides a more direct access to the “universal principle,” and that understanding the mind itself would suffice to meet the qualification of “investigating things” and “extending knowledge” as stipulated in the Great Learning.20 Very few Confucians dared to suggest a rearrangement or bypassing of steps on this prescribed sequence, yet Wang Yangming was ready not only to place the process of “rectifying the mind” before “the investigation of things,” thereby reversing its traditional order, he altogether relegated the whole idea of graduated, step-by-step practices as an inferior approach to the recovery of the “original mind.” In such a manner, Wang also seemed to evince the same loathing to externally-directed investigation and entertained the possibility of a direct, sudden realization of this process in very much the same way the mature Chan school would. Wang’s antithetical position to Zhu in this regard was a paradigmatic case for those who are studying the “sudden versus gradual” polarity in religious studies.21 It was fashionable for Buddhist exegetes to brandish their knowledge of literati culture by writing commentaries for the Confucian Classics. The interesting point was that, Wang’s interpretation of the cultivational outline in the Great Learning as a subitist and introspective path was strikingly similar to the way Buddhists had interpreted the Confucian text. This is not to downplay the important, though at times subtle, differences between Wang’s understanding of the Mind and that of the mainstream, Ming-dynasty Buddhist; but for our purpose, their similarities are the focus.
Cited in Wu 1994: 171. For example, Wang emphasized that “the entirety of the investigation should be performed on just this body and mind” 格物之身, 只在身心上做. Cited in Sun, Liu and Hu 1995: 280. 21 Gregory 1997: 8.
20 19

14

William Chu

When the Buddhist monk Deqing commented on another two of the cultivational stages in the Great Learning – the “rectifying of the mind” (zhengxin 正心) and the “making sincere of the intention” (chengyi 誠意) – Deqing, too, saw those stages as perfectible in a single enlightened instant:
The mind is originally luminous. Only because it was clouded by desires that it becomes dull…As soon as the ‘Single Reality’ had been recovered, all delusions would no longer arise…Once enlightened, there would be no further delusions; [In fact,] the extinction of delusions is itself the equivalent [of the realization] of the Single Reality. 心本光明, 欲蔽故暗 … 一真既復, 諸妄不生 … 覺則不妄, 妄息即 真.22

Most of the commentaries written by the Ming Buddhists on the Confucian Classics had in similar fashion attempted to reinterpret the Confucian cultivational stages within the rubric of Buddhist subitist soteriology. Although within the Buddhist tradition itself, there were proponents for both the gradual and the sudden paradigms, the subitist school had decidedly won the ideological battle since the eighth or ninth century. The influential Buddhist systematizer Zongmi categorized the Chan tradition into three major strands, and the one deemed the most advanced on his list of hermeneutical taxonomy was the “School of Directly Revealing the Mind and the Nature” (zhixian xinxing zong 直顯心性宗) – not surprisingly, that was the most “subitist” strand of the three.23 Deqing also echoed this unanimous Buddhist predilection for the subitist and idealist doctrine throughout the Ming dynasty,
The marvel of the great Dao lies in that it could be better accessed through intuitive realization. Even when it comes to mundane knowledge such as the art of elocution, governance of the world, linguistic convention, and various means of livelihood, one could not fail to come to grasp that which is marvelous by simply penetrating into any of these secular enterprises.

Hanshan dashi mengyou ji, vol. iii, p. 1959. Zongmi’s panjiao schemes were recorded in the Chanyuan zhuquan jiduxu. T. 48.2015.
23

22

Yogācāra resurgence in the Ming dynasty

15

大道之妙 … 貴在妙悟. 即世智辯聰、 治世語言、 資生之業, 無有 一法不悟而得奇妙者.24

Another Buddhist monk in roughly the same period, Ouyi Zhixu 蕅益智旭 (1591–1655), in this respect even praised Wang for having reversed Zhu Xi’s claim to Confucian orthodoxy. The Buddhist monk esteemed Wang for having superseded the spiritual accomplishments of all previous Confucians and for that Wang’s enlightenment might not be qualitatively different from Buddhist enlightenment:
Wang Yangming had gone beyond what the Confucians in the Han and the Song had accomplished, and had directly succeeded to the [true legacy] of the ‘Mind School’ as taught by Confucius and Yan [Hui]. The teaching [Wang] had instructed people with throughout his life [could be summed up] in the words of ‘attaining to one’s innate knowledge of the good.’ It corresponds to and perfectly illumines the Self-Nature…It does not direct people to search outside, because the entirety of its practice is rooted in the Self Nature. 王陽明超漢宋諸儒, 直接孔顏心學. 一生示人, 唯有致良知三字 … 稱 性圓照 … 不向外求, 全修在性.25

Most of the points outlined so far are not new in the Buddhist scholarship. But I will now direct the discussion to how these points shed lights on Ming Yogācāra. In all the aforementioned ways Wang brought about the key catalysts for a syncretistic culture – some highly “buddhicized” recast of Confucian ideas. The Buddhists capitalized on these “buddhicized” elements to advance their own polemical agenda, by demonstrating how these elements could best be qualified, harmonized, and criticized in Buddhist hermeneutics. It was against this background of a brooding Confucian intellectual revolution and the resultant change in the inter-religious dynamic, that the return of the Buddhist Yogācāra thought to scholastic spotlight in the high culture of the Ming dynasty becomes more understandable. To the Buddhist syncretists, it must have been apparent that Wang Yangming’s brand of idealism could best be received and countered by Buddhism’s own counterpart of idealism in the
24 25

Hanshan dashi mengyou ji, vol. iv, p. 2405. Ouyi dashi wenxuan, p. 178.

16

William Chu

Yogācāra tradition. One of the most visible common denominators between the two traditions had been their evolved convergence in the discourse on the mind. The newly resurrected Yogācāra tradition assumed a peculiar form in which the traditional category “consciousness” was treated interchangeably with the notion of the “Mind” in the Tathāgatagarbha thought. The so-called Eminent Monks of the Wanli Era (Wanli si ga­ oseng 萬曆四高僧), Hanshan Deqing 憨山德清 (1546–1623),26 Daguan Zhenke 達觀真可 (1543–1603), Yunqi Zhuhong 雲棲株宏 (1535–1615),27 and Ouyi Zhixu 蕅益智旭 (1599–1655) were interesting subjects of study not simply because of their reputable literary and spiritual accomplishments, but also in the unanimity of their fascination with syncretism, Yogācāra, and Tathāgata-garbha. This striking unanimity probably reflected powerful intellectual and cultural trends of their times. The convergence of these seemingly discrete and unconnected subjects of their fascination (again, syncretism, Yogācāra, and Tathāgatagarbha) was in fact the logical conclusion of development within Buddhism and Confucianism as well as their re-engagement in the Ming. I do not wish to unduly downplay the Eminent Monks’ undeniable individual differences, but it is in their striking resemblance – not fully attributable to their mutual influence – that reflected those compelling cultural forces that are in turn the focus of this paper. Though Wang Yangming and his followers had always perceived themselves to be continuing the idealist legacy initiated by none other than Mencius himself, Wang’s originality stood out in his spilling forth the creative, ontological implications in explicit terms:
The ‘innate knowledge of good’ is the numinous Spirit of creation… our numinous Spirit has created the heaven and the earth, as well as the myriad things therein. All things in the universe, however, ultimately return to nothingness. [The Spirit] carries out the creating

For more on Hanshan’s life and syncretistic efforts, see Hsu 1979. Zhuhong’s life and contribution to Buddhist revival movement in Ming is covered in detail in Yü 1981.
27

26

Yogācāra resurgence in the Ming dynasty

17

process in all moments, and constantly transforms [all things], never pausing for even a breath-long instant. 良知是造化之精靈 … 吾之精靈, 生天生地, 生萬物, 而天地萬物, 復歸 於無. 無時不造, 無時不化, 未嘗有一息之停.28

Compare this notion to Buddhism’s own Laṅkāvatārasūtra narrative on the creative Embryo of the Tathāgata, and we can easily discern the likeness between Wang’s idealism and the Buddhist one that was encapsulated in the Yogācāra-Tathāgatagarbha hybrid thought:
The Tathāgatagarbha is the causative force behind both the wholesome and unwholesome [deeds]. It can function everywhere in creating the different realms of rebirth. Just like a conjuring illusionist, it could magically manifest [the appearances of] the various realms of rebirth. 如來藏是善不善因, 能遍興造一切趣生, 譬如伎兒, 變現諸趣.29

The synthesis of Tathāgatagarbha and ālayavijñāna (the “StoreHouse Consciousness” described by the Yogācāra school) was well underway when the first Chinese encountered these disparate Buddhist teachings on the nature of consciousness and mind. The ubiquitous influence of Tathāgatagarbha thought in Chinese Buddhism since the eighth century, as Chinese Buddhist scholars have frequently noted, swept through practically all Chinese Buddhist schools. The result was a reinterpretation of all Buddhist ideas within Tathāgatagarbha light, including the notion of ālayavijñāna. Although Tathāgatagarbha and ālayavijñāna are definitely distinct concepts with very different doctrinal implications, by the Ming they were by and large harmonized and largely read as interchangeable. The Four Eminent Monks’ close alignment with the Tathāgatagarbha thought was not confined to only their understanding of Yogācāra, but was evident in their Huayan, Tiantai, Chan, and Pure Land scholarship as well, where traditionally disparate teachings were collapsed into variants of an all-unifying Tathāgatagarbha
28 29

Cited in Araki (Yang) 1978: 372–373. Dasheng ru lengqie jing, T. 16.619–620.

18

William Chu

idealism. The Eminent Monks’ sweeping reduction of all Buddhist teachings into a uniform doctrine was not without precedent in Chinese Buddhism. In fact the sectarian traditions of almost all surviving Buddhist schools each had its histories of eventually coming to terms with and assimilating the doctrines of an ontological “Buddha-hood” characteristic of the Tathāgatagarbha thought. For the Yogācāra system to be amenably conforming to the then Tathāgatagarbha-permeated Buddhist culture, all the Four Eminent Monks saw the synthesis of Tathāgatagarbha and ālayavijñāna under the rubric of the “One Mind” to be the most convenient and least problematic formulation to that end. Zhenke wrote,
The ‘four divisions’ of the Eighth Consciousness were initially not distinct entities [from the Eighth Consciousness itself]. It was because that the True Suchness [has the tendency to] responsively adapt to conditions, that it as a result permutated into various [manifestations like the ‘four divisions’]. The meaning of ‘True Suchness responsively adapt to conditions’ is a most difficult doctrine to be clearly elucidated. For if the True Suchness was at the beginning unadulterated and free from ‘being perfumed’ [by contaminating influences], what determined its initial activation to responsively adapt to conditions? If one unceasingly investigate this [problem of theodicy], one might suddenly become enlightened to it…At that point one could clearly understand [all] the books written on the subject of the Mind-Only teaching…Therefore, for those who aspire to transcending the world and shouldering the responsibility of propagating the Path, could they afford not to wholeheartedly direct their attention to the study of the “school concerning the Dharma-nature,”30 the “school of Dharmacharacterization,”31 and the school of Chan?
30 Traditionally both the Madhyamaka and the Tathāgatagarbha teachings had been understood by many Chinese to be pointing to a substrative reality, therefore the Chinese often indiscriminately lodged the former with the latter under the category of the “Xing School, or Dharma-nature School.” After the Tathāgatagarbha tradition became the undisputed dominant doctrinal tradition in China, however, the “Dharma-nature School” was increasingly being used to refer specifically to only Tathāgatagarbha-oriented teachings. Many scholars are wrong in assuming that, throughout Chinese history, the designation ‘Xingzong’ was reserved only for the latter group. One example where such an error was made was in Ran 1995: 11. 31 “Xiangzong 相宗” was a rather pejorative descriptive term for the

Yogācāra resurgence in the Ming dynasty

19

八識四分, 初無別體, 特以真如隨緣, 乃成種種耳. 夫真如隨緣之旨, 最 難明了, 良以真如清淨, 初無薰染, 如何瞥起隨緣耶? 於此參之不已, 忽 然悟入…唯識之書 ,便能了了 …是以有志於出世而荷擔法道, 若性、 若 相、 若禪宗, 敢不竭誠而留神哉?32

Zhenke’s Yogācāra disputations were clearly tailored to the doctrinal assumptions of the Awakening of the Faith, an indigenous Tathāgatagarbha text traditionally received by Chinese Buddhists with paramount esteem. The text, though rather succinct, devised deft solutions to reconcile many apparently contradictory doctrines between the Yogācāra and Tathāgatagarbha schools by positing the so-called “nature-origination” (xingqi 性起) formal causal theory. The Awakening of the Faith proved most useful a scriptural authority to call upon whenever someone attempted to equate Yogācāra conception of the consciousness with the Tathāgatagarbha Mind. The text provided (what appeared to be) canonically sanctioned precedent of submitting major Buddhist tenets to the all-subsuming Tathāgatagarbha teaching and was especially useful for the Ming Yogācāras’ purpose. Most of the Eminent Monks’ Yogācāra interpretative frame rested so much on the cardinal themes of the Awakening of the Faith (with some of them expressly acknowledging to have done so), that one could not help but to come to the impression that the Yogācāra tradition in the late Ming was only studied to be rendered compliant to the “school of Dharma-nature” (another common name for the Chinese Tathāgatagarbha tradition). Zhixu was one such person who self-professedly took the apocryphal text to legitimize Yogācāra-Tathāgatagarbha synthesis:
What the Consciousness-Only School described as ‘[From the perspective of] the Real, [dharma] characteristics are undifferentiated,’ it is exactly what the Awakening of Faith explained under the ‘Gate of the True-Suchness aspect of the One Mind.’ As for what the
Yogācāra system coined by its rival traditions. Since the Yogācāra school was perceived by people like Fazang of the Huayan tradition as merely delving into the feature/phenomenal aspect of reality rather than penetrating into the deeper substrative level, they labeled it a “Faxiang zong, or a Dharmafeature/Dharma-phenomena School” in contrast with the “Faxing zong, or a Dharma-nature School.” 32 Man[ji] Zokuzōkyō 卍續藏經 98, pp. 580–581.

20

William Chu

Consciousness-Only School described as ‘[From the perspective of] the mundane, [dharma] characteristics are differentiated,’ it is just pertaining to the ‘Gate of the arising-and-perishing aspect of the One Mind’ as explained in the Awakening of Faith. 唯識所謂‘真, 故相無別,’ 即起信一心真如門也; 唯識所謂‘俗, 故相有 別,’ 即起信一心生滅門也.33

Attempting at harmonizing the “Dharma characteristic” tradition (a common name for the Chinese Yogācāra tradition) with the “Dharma-nature” tradition, Deqing cited the same text and conflated the “defiled” Store-House Consciousness and the “unadulterated Tathāgatagarbha” as different aspects of the same One Mind. As a result he amalgamated what was traditionally perceived as diametrical strands of Buddhist theories of the mind:
[Aśvaghoṣa] authored the Awakening of Faith in order to extirpate [people’s] unwholesome attachments…and [he] synthesized the doctrinal strands of ‘Dharma-nature’ and ‘Dharma-characteristics,’ so that they could be consolidated in [their common] source [in the One Mind]. [馬鳴]著起信論以破邪執…攝性相而會一源.34

These developments within Buddhism, and the rise of Confucian idealism through the work of Wang Yangming, gave Ming Buddhists and Confucians compelling reasons to argue for their syncretistic cause – now that they had found a highly homogenous, mutually inspired discourse on the Mind that served as the perfect medium for their syncretistic dialogue. The catalyzing effect of Wang Yangming’s system not only could explain the philosophical outlook of the Ming Yogācāra scholarship (one that was conflated and reconfigured in such a way it was uncannily consonant with Confucian idealism), it could also account for Yogācāra scholarship’s precise timing of reappearance. In other words, the timing of Yogācāra’s revitalization in the Ming coincided perfectly with the rise of Wang Yangming’s thought. Zhenke’s disciple Wang

33 Ibid., p. 232. I had to redo the punctuation of the Chinese passage in order to make it intelligible. 34 Hanshan dashi mengyou ji, vol. ii, p. 1024.

Yogācāra resurgence in the Ming dynasty

21

Kentang 王肯堂 (?–1613–?) had recounted a semi-mythical story about the origin of the Yogācāra learning in the Ming. This was a little-known story in and since Wang’s time as far as I could tell, but which nonetheless spoke volumes on the timing of Yogācāra resurgence. Upon closer scrutiny, the story could provide us some inkling regarding how the Ming Buddhists themselves explained the sudden resuscitation of the school after such long dormancy:
I have heard the Great Master Zibo (Zhenke) said that, “the transmission of the [Fa]xiang school (the Chinese Yogācāra) had discontinued for long. The Dharma Master Luan [Pu]tai,35 while on one of his ‘learning journeys,’ had stopped underneath the eaves of a household to shelter himself from the falling rain. He heard the sound of someone giving a Dharma talk inside [the house], and upon listening more closely, it turned out to be about the Faxiang (again, the Chinese Yogācāra) teaching. He immediately entered the house to greet the people inside, and saw that it was an old man explaining [the teaching] to an old woman. Master Tai then bowed and asked to be instructed [in the teaching]. He consequently stayed at the place for more than a month, until [the old man] had completed instructing [Tai] of his learning. [Tai] suspected that the old man and the old woman were no ordinary mortals, they should in fact be the magical incarnation of [enlightened] sages.” 余聞紫柏大師言, 相宗絕傳久矣. 魯庵36泰法師, 行腳避雨止一人家簷下, 聞其內說法聲, 聽之則相宗也. 亟入見, 乃一翁為一嫗說. 師遂拜請教, 因留月餘, 盡傳其學而去. 疑翁嫗非凡人, 蓋聖人應化而現者.37

The semi-mythical nature of the story is reminiscent of other Buddhist tales that attempt to relate the emergence of a new teaching. One such example involves Nāgārjuna’s journey into the Dragon King’s Palace where he received Mahāyānasūtras and brought the new teaching to the world. Another story of similar nature: Asaṅga’s ascension into the Tuṣita heaven where he was said to have received instructions from the Bodhisattva Maitreya
Luan Putai 魯庵普泰 (?–1511–?) was an otherwise unknown figure save through his two surviving works on Yogācāra currently included in Man[ji] Zokuzōkyō 98, p. 513. 36 This character is sometimes written with the “艸” radical. 37 Cited in Shi 1987: 201.
35

22

William Chu

on the intricacies of Yogācārabhūmi.38 Such Buddhist tales were likely an admixture of partial historical truths, dramatized fiction, and other projections of culturally specific symbols used to account for the obscure and sometimes questionable background of new teachings.39 It would be hard to separate facts from fictions in stories like these. But the semiotics and circumstantial information contained therein were often predicated on more concrete historical backgrounds. As the provenance of Nāgārjuna’s Mahāyānasūtras and the authorship of the Yogācārabhūmi were concealed by their great antiquity and the hazy memories that reported them, these stories provided assurances of legitimate origins of otherwise questionable teachings. By the same token, as the beginning (if there was such a thing in the singular form and implying a complete prior discontinuance) of the Yogācāra revival in the Ming was shrouded in mystery even to its contemporaries, there must had existed an almost subliminal compulsion to attribute its origin to legitimate and comprehensible sources, even if that attribution could only find expression in symbolic dramatization. Just like Asaṅga purportedly started his Yogācāra career after having received Bodhisattva Maitreya’s personal tutelage on the matter, thereby locating his source of inspiration in an acknowledged authority, the protagonist of our story at hand – Luan Putai – also was said to have rekindled the moribund Chinese Yogācāra school after receiving the divine revelation of the teaching through otherworldly intermediaries. Moreover, Luan Putai published the [only known] Yogācāra works in 1511 (“In the Xinwei year of the
38 Interestingly enough, Deqing also had the experience of dreaming about ascending into the presence of the same Bodhisattva. According to his own relating of the story, this dream was the occasion after which he came to grasp the true meanings of Yogācāra teachings. See the section on ‘age 33,’ Hanshan dashi mengyou ji, vol. iv, pp. 2902–2905. It is puzzling how Western studies of Deqing’s life consistently leave out this richly symbolic event in his life. 39 See, for example, Robert Buswell’s discussion on the symbolism and social-religious factors that were embodied in the dragon king motif in Buddhist mythological lore (Buswell 1989: 51–60).

Yogācāra resurgence in the Ming dynasty

23

Zhengde reign era of Emperor Wu of the Ming dynasty” 明武宗正 德辛未), that is, almost immediately after Wang Yangming had put forth and promulgated his idealism. Since it would only be natural that we place Luan Putai’s reported divine encounter before the publication date of his Yogācāra book – that is, before Luan Putai became so proficient in Yogācāra as to have been able to write about it – the divine source of inspiration that was said to have sparked off a renaissance uncannily coincided with the most opportune time period when important conditions conducive to that renaissance were just becoming ripe. Whether the story originator(s) recounted the story conscious or unconscious of this coincidence, the conclusions we can draw from the story nonetheless corroborate our theory concerning the significant role played by Wang Yangming’s thought in the Ming-dynasty Yogācāra tradition.

Abbreviations and bibliography Abbreviations
T. Takakusu, J. and K. Watanabe, eds. 1924–1932. Taishō shinshū daizōkyō. Tokyo: Taishō issaikyō kankōkai.

Primary sources
Chanyuan zhuquan jiduxu 禪源諸詮集都序. T. 48.2015. Dasheng ru lengqie jing 大乘入楞伽經. T. 16.619–620. Hanshan dashi mengyou ji, 4 volumes. 憨山大師夢遊集(全四冊). 1989. Reprinted in Hong Kong: Puhui Lianshe (reprint). Liuzu dashi fabao tanjing 六祖大師法寶壇經. T 48.2008 Ouyi dashi wenxuan 蕅益大師文選. Compiled by Zhongjing 鐘鏡. 1976. Reprinted in Taipei: Fojiao chuban she. Is the 1976 edition used here the reprint? In this case, change “Reprinted in Taipei: Fojiao chuban she” to “Taipei: Fojiao chuban she (reprint)”

Secondary sources
Araki, Kengō. 1975. “Confucianism and Buddhism.” In The Unfolding of Neo­Confucianism, 39–66, eds. William de Bary and the Conference on Seventeenth-Century Chinese Thought. New York and London: Columbia University Press.

24

William Chu

Araki Kengō 荒木見悟. Trans. with Rushi 如實. 1978. “Yangming xue yu mingdai foxue” 陽明學與明代佛學. In Zhongguo jinshi fojiaoshi yanjiu 中 國近世佛教史研究, ed. Lan Jifu 藍吉富. Taipei: Dasheng Wenhua. Araki Kengō T 荒木見悟. Trans. with Yang Baiyi 楊白衣. 1978. “Yijing yu lengyan jing” 易經與楞嚴經. In Zhongguo jinshi fojiaoshi yanjiu 中國近 世佛教史研究, ed. Lan Jifu 藍吉富. Taipei: Dasheng Wenhua. Buswell, Robert. 1989. The Formation of Ch’an Ideology in China and Korea. Princeton: Princeton University Press. Buswell, Robert, and Robert Gimello. 1992. “Introduction.” In Paths to Liberation: The Mārga and its Transformations in Buddhist Thought, 1–36, eds. Robert E. Buswell, Jr. and Robert M. Gimello. Honolulu: University of Hawai’i Press. Chan, Wing-tsit. 1963. A Source Book in Chinese Philosophy. Princeton: Princeton University Press. Chen Rong 陳戎. 1984. Sishu Jijie 四書集解. Tainan: Zhengyan. Chen Rongjie (Wing-tsit) 陳榮捷. 1984. Wang Yangming yu chan 王陽明與 禪. Taipei: Taiwan Xuesheng. Greenblatt, Kristin. 1975. “Chu-hung and Lay Buddhism in the Late Ming.” In The Unfolding of Neo­Confucianism, 93–140, eds. William de Bary and the Conference on Seventeenth-Century Chinese Thought. New York and London: Columbia University Press. Gregory, Peter. 1987. “Introduction.” In Sudden and Gradual: Approaches to Enlightenment in Chinese Thought, 1–9, ed. Peter Gregory. Honolulu: University of Hawai’i Press. _____ 2002. Tsung-mi and the Sinification of Buddhism. Honolulu: University of Hawai’i Press. Hsu, Sung-pen. 1979. A Buddhist Leader in Ming China: The Life and Thought of Han-Shan Te-Ch’ing. University Park: Pennsylvania State University Press. Kalupahana, David. 1992. A History of Buddhist Philosophy. Honolulu: University of Hawai’i Press. Kaviratna, Harischandra, trans. 1980. Dhammapada: Wisdom of the Buddha. Pasadena, CA.: Theosophical University Press. Liu, Wu-chi. 1964. A Short History of Confucian Philosophy. New York: Dell Publishing. Qian Mu 錢穆. 1962. Songming lixue gailun, vol. i 宋明理學概述(一). Taipei: Zhonghua Wenhua. Ran Yunhua 冉雲華. 1995. Cong Yindu fojiao dao Zhongguo fojiao 從印度佛 教到中國佛教. Taipei: Dongda. Sharf, Robert H. 2001. Coming to Terms with Chinese Buddhism: A Reading of the Treasure Store Treatise. Honolulu: University of Hawai’i Press.

Yogācāra resurgence in the Ming dynasty

25

Shi Shengyan 釋聖嚴. 1987. Mingmo fojiao yanjiu 明末佛教研究. Taipei: Dongchu. Sun Kaitai 孫開泰, Liu Wenyu 劉文雨 and Hu Weixi 胡偉希. 1995. Zhongguo zhexueshi 中國哲學史. Taipei: Wenjun. Wu Guang 吳光. 1994. Rudao lunshu 儒道論述. Taipei: Dongda. Yü, Chün-fang. 1981. The Renewal of Buddhism in China: Chu-hung and the Late Ming Synthesis. New York: Columbia University Press.

Ignorance, epistemology and soteriology1 Part II
Vincent Eltschinger

For Jacques May

The first part of this essay (Eltschinger 2009) concentrated on the basic features and likely sources of Dharmakīrti’s understanding of ignorance (avidyā). Against the Vaibhāṣikas, but with Vasubandhu the Kośakāra, Dharmakīrti defines ignorance as a “counter-” or “anti-knowledge,” i.e., as a cognition that counteracts true (perceptual) knowledge (vidyā) by displaying contrary/erroneous objectsupports and aspects (viparītālambanākāra). According to him, ignorance amounts to pseudo-perception (pratyakṣābhāsa), hence conceptual construction (vikalpa), superimposition (samāropa) and concealment (saṃvṛti). The core of Dharmakīrti’s philosophy, the so-called apoha theory, provides an exhaustive picture of both ignorance as conceptuality and inference as a corrective (though conceptual) principle. This conception of ignorance, however, fails to account for the most dramatic form of the Buddhist ignorance, viz. its being responsible for defilements, rebirth and suffering. In
This study has been made possible by the generous financial support of the Austrian Science Fund (FWF-Projekt P19862 “Philosophische und religiöse Literatur des Buddhismus”). Most sincere thanks are due to Isabelle Ratié, Birgit Kellner, Helmut Krasser and Ernst Steinkellner. Lambert Schmithausen also deserves my wholehearted gratitude for having gone through this essay with incomparably great care and erudition. My most sincere thanks are due to Cynthia Peck, who kindly corrected my English. Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 27–74
1

28

Vincent Eltschinger

order to account for this eschatologically valued form of ignorance, Dharmakīrti equates avidyā with the personalistic false view (satkāyadṛṣṭi). Consistently enough, ignorance as satkāyadṛṣṭi is but a specialization or instantiation of ignorance as conceptuality insofar as the satkāyadṛṣṭi exhausts itself in one’s superimposing such conceptual constructs as “self/I” (ātman, aham) and “one’s own/mine” (ātmīya, mama) on reality. Both Dharmakīrti and his commentators evolved exegetical strategies in order to argue for the orthodoxy of this equation of ignorance with a false view (dṛṣṭi), which Vasubandhu clearly refuses in the Abhidharmakośa (but not in his commentary on the Pratītyasamutpādasūtra). As for the sources of Dharmakīrti’s conception, they are very likely to consist of the Pratītyasamutpādasūtra and its numerous “idealistic” interpretations (Yogācārabhūmi, Vasubandhu’s Vyākhyā). In the second part of this essay, I shall first inquire into Dharmakīrti’s account of dependent origination (pratītyasamutpāda), viz. his interpretation of ignorance as the origin of defilements (craving, etc.), clinging and rebirth. I shall then turn to the philosophical core of this study by attempting to show how Dharmakīrti’s views on ignorance and the two truths/realities provide the basic framework of his epistemological theory. This is tantamount to claiming that Dharmakīrti’s epistemology, in locating ignorance and defining the cognitive means of opposing it and entering the path toward salvation, is Buddhistic in both its inspiration and its finality. As a consequence, his philosophy should cease to be regarded as a dry academic endeavour deviating from the spirit of Buddhism as a salvation system.

2.1. Dependent origination
2.1.1. In his account of the future Buddhaʼs philosophical reflections on the eve of his career, Dharmakīrti presents the cause of suffering (duḥkhahetu) in the following way: “The cause [of suffering, i.e., of rebirth,] is attachment bearing upon the conditioning factors, [an attachment that is] due to the belief in self and  oneʼs  own.”2
PV 2.135ac1: ātmātmīyagrahakṛtaḥ snehaḥ saṃskāragocaraḥ / hetuḥ … sneha = tṛṣṇā according to PVP D56a7/P64a4 and PVṬ D117b3–4/P143b7;
2

Ignorance, epistemology and soteriology (2)

29

According to Devendrabuddhi, craving proceeds from one’s adhering to the painful conditioned factors that are intrinsically free from self and one’s own, under the aspects of self and oneʼs own.3 This is tantamount to saying that defilements such as craving only occur once unreal aspects have been superimposed on dharmas, specifically on the five constituents one clings to, which lack these aspects entirely. While commenting on another passage, Devendrabuddhi claims that defilements such as desire (another equivalent for attachment and craving) proceed from oneʼs superimposing aspects such as permanent, pleasurable, self and oneʼs own on the impermanent, painful, selfless and empty constituents.4 One may adduce here a huge number of passages presenting one and the same idea: The personalistic belief is responsible for oneʼs superimposing contrary aspects such as self and oneʼs own on the selfless and empty constituents.5 As Dharmakīrti himself has it, “desire [arises] from the superimposition of another [i.e., unreal] nature on something (dharma) that does not have this nature.”6 PV 2.270 provides us with Dharmakīrtiʼs most significant statement as to how craving takes place once unreal aspects have been as-

Śākyabuddhi (PVṬ D117b4/P143b7–8) unambiguously explains gocara as viṣaya. 3 PVP D56b1/P64a5–6: sdug bsṅal du gyur pa’i ’dus byas bdag daṅ bdag gi daṅ bral ba la bdag daṅ bdag gi’i rnam par mṅon par źen pas ’jug pa źes bya ba’i don to //. 4 PVP D60b2–3/P69a4–5: mi rtag pa daṅ sdug bsṅal ba daṅ stoṅ pa daṅ bdag med pa’i phuṅ po rnams la rtag pa daṅ bde ba daṅ bdag daṅ bdag gir sgro btags nas ’jug pa ’dod chags la sogs pa de dag … 5 E.g., PVP D88a4–5/P101b4: ñe bar len pa’i phuṅ po lṅa la gaṅ rtag pa daṅ bde ba daṅ bdag daṅ bdag gi rnam pa yod pa ma yin no //. PVP D88a6/ P101b5–6: ñe bar len pa’i phuṅ po lṅa la rtag pa la sogs pa’i rnam par ’dzin pa’i śes pa yaṅ rnam pa med pa ’dzin pa can yin no //. 6 PV 2.196ab: ātmāntarasamāropād rāgo dharme ’tadātmake /. Devendrabuddhi explains (PVP D84a7–b1/P97a1–2): ’dod chags la sogs pa’i raṅ bźin du yaṅ ’gyur ba ma yin te / ’di ltar de bdag med can te / rtag pa daṅ bde ba daṅ bdag daṅ bdag gi daṅ bral ba’i yul du gyur pa’o // chos la ste phuṅ po la sogs pa’i raṅ gi ṅo bo la’o // bdag gźan sgro btags phyir te rtag pa daṅ bde ba daṅ bdag daṅ bdag gi’i raṅ bźin gźan du sgro btags pa’i rgyu’i phyir mṅon par źen pa’i mtshan ñid kyi chags pa skye bar ’gyur ro //.

30

Vincent Eltschinger

cribed to reality: “Having[, due to ignorance,]7 superimposed sixteen unreal aspects, viz. ‘lasting,’ ‘pleasant,’ ‘mine,’ ‘I,’ etc., on the four [Noblesʼ] Truths,8 one experiences craving [for superimposed objects such as delight, etc.].”9 According to Devendrabuddhi and
PVP D116a1/P134b2: sgro btags nas ni mi śes pa’i phyir … At least according to the Vaibhāṣikas, each of the four Nobles’ Truths is to be successively contemplated under four different aspects: the Truth of suffering under the aspects “impermanent,” “painful,” “empty” and “selfless;” the Truth of origin under the aspects of “(distant/material) cause” (as a seed), “arising,” “(serial) causation” and “(joint) condition;” the Truth of extinction, under the aspects of “extinction,” “calm,” “excellent” and “salvation;” the Truth of the path under the aspects of “path,” “fitness,” “access” and “conducive to release” (AKBh 343,16–19 on AK 6.17c1: duḥkhaṃ caturbhir ākāraiḥ paśyaty anityato duḥkhataḥ śūnyato ’nātmataś ca / samudayaṃ caturbhir hetutaḥ samudayataḥ prabhavataḥ pratyayataś ca / nirodhaṃ caturbhir nirodhataḥ śāntataḥ praṇītato niḥsaraṇataś ca / mārgaṃ caturbhir mārgato nyāyataḥ pratipattito nairyāṇikataś ca /. The sixteen aspects are listed at PVP D62a3–7/P71a1–6). The AKBh records a lengthy discussion pertaining to four different ways of interpreting these sixteen aspects (see AKBh 400,1–401,17 on AK 7.13a, Kośa 7.30–39, Pruden 1988–1990: IV.1110–1116). According to the fourth exegetical pattern, each of these aspects aims at counteracting (pratipakṣa) a particular false view (dṛṣṭi): The aspects anitya, duḥkha, śūnya and anātman counteract the false views of permanence, pleasurableness, one’s own, and self; the aspects of hetu, samu­ daya, prabhava and pratyaya contradict the false views of the absence of a cause, of a unique cause such as God or primordial matter (according to AKVy 628,30–31), of an evolution of being, and of an intelligent creation; the aspects nirodha, śānta, praṇīta and niḥsaraṇa oppose the false views that release does not exist, that release is painful, that the bliss of dhyānas is the most excellent, and that liberation, because it is subject to falling again and again, is not definitive; as for the aspects mārga, nyāya, pratipad and nair­ yāṇika, they respectively counteract the false views that there is no path, that this is a wrong path, that there is another path, and that the path is subject to retrogression; see AKBh 401,11–17, Kośa 7.38–39, Pruden 1988–1990: IV.1115–1116. The explanations provided by Dharmakīrti’s commentators are too few to allow us to determine which interpretation, if any, they favoured. Devendrabuddhi and Śākyabuddhi content themselves with listing the four aspects superimposed on each of the last three Truths (see PVP D115b6–7/P134a8–b2 and PVṬ D147b3–5/P182a8–b2). On the sixteen aspects, see Wayman 1980. 9 PV 2.270: sthiraṃ sukhaṃ mamāhaṃ cetyādi satyacatuṣṭaye / abhūtān ṣoḍaśākārān āropya paritṛṣyati //. Note PVṬ D147b5–7/P182b2–4: sgro
8 7

Ignorance, epistemology and soteriology (2)

31

Śākyabuddhi, ignorance,10 i.e., the false view of self, has one grasp aspects that are contrary to the real ones, i.e., superimpose an “I” on what is selfless and a “mine” on what is empty. But ignorance is also responsible for deluded persons taking momentary things to be lasting (sthira) or even unchangeably permanent (kūṭasthanitya),11 or holding intrinsically painful things to be pleasurable, i.e., not to be under the sway of cankers (sāsrava) or dependent on causes (hetuparatantra) in each of their successive phases (pratikṣaṇam).12 2.1.2. According to Dharmakīrti and his commentators, the personalistic false view is the (principal) cause (nidāna), the origin (yoni,

btags nas ni yoṅs su sred ces bya ba’i tshig gis log par sgro ’dogs pa sṅon du soṅ ba can gyi sred pa ñid gsal bar bstan pa yin no // sgro ’dogs pa’i yul la ’jug pa’i sred pa de yaṅ sgro ’dogs pa’i rnam pa ñid yin la / sgro ’dogs pa’i rnam pa can gyi yul can gyi ñon moṅs pa daṅ ñe ba’i ñon moṅs pa thams cad ñid ma rig pa ñid yin pa … “And with the pāda (= PV 2.270d) āropya paritṛṣyati, [Dharmakīrti] clearly indicates craving, which presupposes erroneous superimposition. As for this craving, directed [as it is] to an object of superimposition, it also has the aspect of superimposition, and all the kleśas and upakleśas, which bear on an aspect of superimposition, are [nothing] but ignorance …” 10 PVP D115b3–4/P134a4: ma rig pa des kyaṅ sdug bsṅal la rtag pa źes bya ba’i rnam par ’dzin par byed do //. PVP D115b6/P134a7–8: re źig de ltar sdug bsṅal gyi bden pa la mi śes pa mi rtag pa la sogs pa’i rnam pa las phyin ci log tu sgro ’dogs pa yin no //. See also PVṬ D147a1–2/P181b3–5. 11 According to Devendrabuddhi, all that is produced and lasts more than one moment is permanent (PVP D115b4/P134a5–6: skad cig ma las dus phyis gnas pa’i ṅaṅ tshul can du skyes pa thams cad rtag pa ñid do //. To be compared with Vibh. 102 n. 1: nityam iti vācye kṣaṇāt paraṃ sthāyī sarvo nitya ity arthaḥ /). According to Śākyabuddhi, all that is either unchangeably permanent or lasts for at least a second moment is permanent (PVṬ D147a6–7/P182a2–3: ther zug tu gnas pa’i rtag pa gaṅ yin pa daṅ skad cig ma gñis pa la sogs par gnas pa’i ṅaṅ tshul can dus gźan du gnas pa can gaṅ yin pa de thams cad ni ’dir rtag par ’dod pa yin gyi ther zug tu gnas pa ñid ni ma yin no źes de bstan par ’gyur ro //). 12 According to PVP D115b5/P134a6: bde ba źes bya ba’i zag pa daṅ bcas pa ma yin pa’am skad cig ma re re la rgyu’i gźan gyi dbaṅ la[s] phyin ci log tu btags pa’o //. duḥkha(bhūta) is regularly explained as sāsrava in PVP; see, e.g., PVP D57b7/P66a1 and PVP D58a3/P66a5.

32

Vincent Eltschinger

prabhava), or the root (mūla)13 of all (kinds of) moral faults (doṣa), defilements (kleśa, upakleśa) or moral impurities (mala).14 Among the expressions denoting the fact that defilements such as desire originate from the false view of self, one also meets with “cause” (kāraṇa, alone or with preceding utpatti°, pradhāna°; hetu),15 “arising” (jāti, utpatti)16 and suffixal elements such as °pūr vaka, °maya,17 °hetuka, °ja, °mūla, or °kṛta. Defilements originate from the personalistic false view (satkāyadarśanaja, ’jig tshogs su lta ba’i raṅ bźin), are (causally) preceded/accompanied by the false view of self or by the adherence to self and oneʼs own (bdag tu lta ba sṅon du soṅ ba can, ātmātmīyābhiniveśapūrvaka), arise from the false view of self (bdag tu lta ba las byuṅ ba), or have ignorance for their cause (avidyāhetuka).18 They are all based on the beliefs in “I” and “mine” (ṅar ’dzin pa daṅ ṅa yir ’dzin pa dag la gnas pa) and arise in dependence on a mind that complies with the false view of self and oneʼs own (bdag daṅ bdag gir lta ba’i rjes su ’brel ba’i sems la ltos nas … ’gyur ba).19 2.1.3. As we have seen, the belief in self and one’s own is the cause of suffering, i.e., attachment bearing on the conditioning factors. In other words, ignorance is the cause of craving (tṛṣṇā), which
13 Respectively PV 1.223ab (nidāna gl. pradhānakāraṇa PVSVṬ 402,23– 24), PV 2.211a, PVSV 111,11, PV 2.197ab1 (mūla gl. daṅ po’i rten PVP D84b2/ P97a4), PV 2.212c. 14 E.g., PV 2.197a (doṣa), PV 1.222a (sarvāsāṃ doṣajātīnām), PV 2.214d1 (sarvadoṣa), PVSVṬ 401,24–25 and PVP D91a2/P105a5 ([sarva]kleśa), PVP D60a2–3/P68b4 (ñon moṅs pa daṅ ñe ba’i ñon moṅs), PV 2.212c (malāḥ sarve). On upakleśa, see also PVṬ D133a4–5/P164a4. 15 E.g., PVSVṬ 50,28 (kāraṇa), PVSVṬ 401,29 and PVP D91a2/P105a5 (utpattikāraṇa), PVSVṬ 402,23–24 (pradhānakāraṇa), PVSVṬ 401,21 (hetu). 16 E.g., PV 1.222b (jātiḥ), PVSVṬ 401,22 and 26 (utpatti). 17 Rendered in Tib. as raṅ bźin (can). But note PVṬ D137b3/P916b6: raṅ bźin ni ṅo bo ñid dam rgyu yin no //. 18 Respectively PVSV 111,19, PVP D93b1/P108a1 (on raṅ bźin, see above, n. 17), PVP D60a2–3/P68b2–3, PVSV 8,20, PVP D93a5/P107b5, PVSVṬ 401,24 and 25. 19 Respectively PVP D93b1–2/P108a1–2 and PVP D67b4/P77a6–7.

Ignorance, epistemology and soteriology (2)

33

is nothing but the traditional sequence of dependent origination, where both function as the cause of suffering: As defilements, they give rise both to other defilements (e.g., tṛṣṇā → upādāna) and to act(ion)s (kriyā, e.g., avidyā → saṃskāra, or upādāna → bhava), the latter being in turn responsible for new foundations (vastu) of existence (e.g., saṃskāra → vijñāna, or bhava → jāti).20 Insofar as they give rise to actions leading to new existential foundations, ignorance and craving21 are the two causes of (re)birth ([punar] janman) and transmigration (saṃsāra),22 which are the hallmarks of suffering.23 Whereas Devendrabuddhi simply defines suffering
See AK 3.27 and AKBh 134,26–135,3, Kośa 3.69, Pruden 1988–1990: II.407. 21 PVP D56a6/P64a3: skye ba’i mtshan ñid can gyi sdug bsṅal gyi rgyu; PVP D57b3/P65b4: bdag daṅ bdag gi la chags pa’i mtshan ñid can gyi sdug bsṅal gyi rgyu; PVP D115b6/P134a8: sred pa’i mtshan ñid can sdug bsṅal gyi rgyu; PVP D116a1/P134b3: sred pa sdug bsṅal gyi rgyur gyur pa; PVP D115b2/P134a2–3: sdug bsṅal gyi rgyu ni sred pa yin no źes bstan zin to // de yaṅ ma rig pa las byuṅ ba … According to Śākyabuddhi, craving is kun nas ’chiṅ ba’i rgyu, “the cause of bondage,” and according to PVP D58b1/P66b4, attachment leads to kleśas, punarbhava and janmaparigraha. 22 Dharmakīrti’s commentators provide us with various definitions of saṃsāra. (1) PVP D62b3–4/P71b2–3: ’khor bar ’khor bas na ’khor ba ste / skye ba daṅ ’chi ba’i rgyun no //, to be compared with PVV 62,11–12: janmamaraṇaprabandhaḥ saṃsāraḥ /. (2) PVP D95b6/P110b3: (bdag gir yoṅs su ’dzin pa) rtsom pa la sogs pa’i mtshan ñid can gyi ’khor ba …, which Śākyabuddhi (PVṬ D138b6–7/P171a7–8) comments as follows: bdag gir yoṅs su ’dzin pa la sogs pa rtsom pa la sogs pa’i mtshan ñid can gyi ’khor ba źes bya ba la bdag gi ñid du gzuṅ ba’i srid pa’i loṅs spyod kyi mtshan ñid can gyi dṅos po la mṅon par chags pa sṅon du soṅ ba can gyi ’dzin pa ni yoṅs su (P om. su) ’dzin pa’o // rtsom pa ni mṅon par bsgrub pa’o //. Tib. mṅon par bsgrub pa may translate either abhinirhāra (BHSD s.v., 52b–53a) or (more surely) abhisaṃskāra (BHSD s.v., 57b): Defining “[re]existence” (bhava) in the context of dependent origination, Vasubandhu (Vaibhāṣika definition, AKBh 132,20–21) says: sa … paunarbhavikaṃ karmopacinoti …, “he accumulates action(s) that is/are conducive to rebirth.” Note also TSP Ś230,8–9/K184,21– 22 (unidentified quotation): cittam eva hi saṃsāro rāgādikleśavāsitam /. 23 PVṬ D148a1/P182b6: ma rig pa daṅ sred pa ni sdug bsṅal gyi rgyu ñid yin te / phyin ci log pa’i raṅ bźin can źes bya ba’i don to //. Suffering is also defined in terms of duḥkhatātraya. PVP D62b4/P71b3–4: sdug bsṅal rnam pa gsum gyis dṅos sam brgyud pas sdug bsṅal ba yin no //, which Śākyabuddhi, having named the three “painfulnesses” (PVṬ D120b5/
20

34

Vincent Eltschinger

as (re)birth (skye ba’i mtshan ñid can gyi sdug bsṅal), Dharmakīrti characterizes it as the constituents undergoing transmigration (duḥkhaṃ saṃsāriṇaḥ skandhāḥ).24 It comes as no surprise, then, that Dharmakīrti declares that “as long as (s)he adheres to a self, the [person who experiences craving remains] in saṃsāra.”25 According to Devendrabuddhi, for whom “the personalistic false view is the cause of the connection (pratisandhi) to a new existence (punarbhava),”26 “the [person] who is under the sway of the false view of self has the notion of pleasure (sukhasaṃjñā) with regard to suffering [and] will be connected to a new existence.”27 The link between the false view of self, attachment and rebirth can be summarized as follows: “Thus when there is adherence to a self, a multitude of [moral] faults such as attachment to oneʼs own arise, and the attachment to a self causes [one] to take a [new existential] place (sthāna).”28 2.1.4. Let us consider now the genealogy29 of defilements from the personalistic false view. As we shall see, Dharmakīrti provides a
P147b5), comments as follows (PVṬ D120b6–7/P147b5–7): (1) duḥkhā vedanā is suffering in a direct way as duḥkhaduḥkhatā (its causes and conditions being suffering in an indirect way); (2) sukhā vedanā is suffering in a direct way as pariṇāmaduḥkhatā (its causes and conditions being suffering in an indirect way); (3) asukhāduḥkhā vedanā is suffering in a direct way as saṃskāraduḥkhatā (its causes and conditions being suffering in an indirect way). On duḥkhatātraya, see Schmithausen 1977. 24 Respectively PVP D56a6/P64a3 and PV 2.146c. 25 PV 2.218cd (leaving tena untranslated): tenātmābhiniveśo yāvat tāvat sa saṃsāre //. 26 PVP D85a6–7/P98a3–4: ’jig tshogs lta ba yaṅ srid par ñiṅ mtshams sbyor ba’i rgyur gyur pa … Note also, referring to the sahajaṃ satkāyadarśanam (PV 2.200d), PVP D85b5/P98b2–3: de yaṅ srid pa’i rgyu yin no //. 27 PVP D85a6/P98a3: gaṅ la bdag tu lta ba yod pa de ni sdug bsṅal la bde ba’i ’du śes can yin te / yaṅ srid par mtshams sbyor bar ’gyur ro //. 28 PVP D58a7–b1/P66b3–4: de ltar na bdag tu mṅon par źen pa yod na bdag gir chags pa la sogs pa’i skyon gyi tshogs ’jug par ’gyur źiṅ / bdag tu chags pas kyaṅ gnas yoṅs su len par byed do //. 29 “Genealogy” as a free rendering of Karṇakagomin’s krama (lit. “sequence,” “succession;” PVSVṬ 401,25–26: kena punaḥ krameṇa doṣāṇāṃ satkāyadarśanād utpattiḥ /).

Ignorance, epistemology and soteriology (2)

35

coherent picture of the sequence avidyā–(ṣaḍāyatana–sparśa– vedanā–)tṛṣṇā–upādāna–bhava–jāti, although some items in his account have no explicit equivalent in the traditional twelvemembered chain of dependent origination. In Dharmakīrtiʼs opinion, the false view of self may be held directly responsible for the rise of at least three factors: the notion of otherness, the belief in oneʼs own, and attachment/craving. In an interesting statement, Dharmakīrti points out that “once [the notion of] a self exists, the notion of the other (parasaṃjñā) [arises, and] from this distinction between self and other [is born] grasping and aversion; bound to these two, all the moral faults arise.”30 For reasons that I shall explain below, I am inclined not to follow the traditional explanation that links grasping/attachment to (the notion of) the self and aversion to the notion of the other.31 For the time being, let us leave this problem out of consideration and focus on the genealogy of otherness: “As long as the mind adheres to a self (ātmeti), [it has] the notion of a self (ātmasaṃjñā), and once this [notion] exists, all that [the mind] does not grasp in this way is [held to be] other.”32 In
PV 2.219 (āryā metre): ātmani sati parasaṃjñā svaparavibhāgāt parigrahadveṣau / anayoḥ sampratibaddhāḥ sarve doṣāḥ prajāyante //. Delusion (moha), covetousness (lobha) and hatred/aversion (dveṣa) are traditionally held to be the three root-defilements (mūlakleśa) or roots of evil (akuśalamūla); see AK 5.20c and AKBh 291,8. Note, e.g., AK 5.48a2b: rāgotthā āhrīkyauddhatyamatsarāḥ. “From out of lust there proceeds disrespect, dissipation, and avarice” (Pruden 1988–1990: III.843, Kośa 5.91). For definitions of āhrīkya, auddhatya and matsara, see AKBh 59,19–20 (Pruden 1988–1990: I.200, Kośa 2.170), AKBh 312,17 (Pruden 1988–1990: I.194, Kośa 2.161) and AKBh 312,16–17 (Pruden 1988–1990: III.842, Kośa 5.90). AK 5.48a2b: krodherṣye pratighānvaye. “From out of hatred there proceeds envy and anger” (Pruden 1988–1990: III.843, Kośa 5.91). For definitions of krodha and īrṣyā, see AKBh 312,16 (Pruden 1988–1990: III.842, Kośa 5.90) and AKBh 312,19 (Pruden 1988–1990: III.842, Kośa V.90). 31 PVP D95b1/P110b5–6: bdag ñid du bzuṅ ba la yoṅs su ’dzin pa ni mṅon par chags pa’o // gźan ñid du rnam par phye ba la sdaṅ ba yin te / yoṅs su dor ba’o //. PVV 87,15–16: svaparavibhāgāc ca kāraṇāt svaparayor yathākramaṃ parigraho ’bhiṣvaṅgo dveṣaḥ parityāgas tau bhavataḥ /. 32 PVP D95a7/P110b4–5: ji srid du blo bdag ces mṅon par źen pa de srid du bdag tu ’du śes pa daṅ de yod na de ltar mi ’dzin pa gaṅ yin pa de thams cad gźan yin no //.
30

36

Vincent Eltschinger

another statement, Dharmakīrti declares that “the false view of self generates the belief in oneʼs own (ātmīyagraha).”33 Persons deluded by the false view of self regard the constituents of being both as a self and as belonging to the self, but this feeling of property may well be extended beyond the constituents and range over parts of the world that have been posited as other than the self. The personalistic belief is responsible for yet another factor, which is variously termed “desire” (rāga), “craving” (tṛṣṇā), “grasping” (pari­ graha) or “attachment”/“love” (sneha), and clearly corresponds to the eighth link of dependent origination, i.e., craving. In spite of this functional equivalence, I am inclined not to consider these terms as (always) synonymous, and to believe that Dharmakīrti introduced a causal sequence between them, thus splitting the traditional eighth link into two. If I am correct, from the false view of self arises first attachment or love for the self and oneʼs own, and then craving for the things that are regarded as beneficial or pleasurable to the self. This can be seen in the following stanza: “The one who sees a self has a constant love for this [self, thinking of it as] ‘I.’ Because of [this] love [for the self] he craves for the delights [of this self, and his] thirst conceals [from him] the drawbacks [of the things he deems conducive to these delights].”34 Here, both Devendrabuddhi and Manorathanandin interpret “love” as “love for the self.”35 Whereas attachment is directed to the self (but bears upon the conditioned factors), craving is directed to the delights (sukha) of the self,36 i.e., to the things that are deemed conducive to these delights,37 or to impure (sāsrava) things that are (deemed) favourable (anugrāhaka) in that they are conducive to the delights (of the self).38 Besides the frequent occurrence of expressions such
PVSV 111,18: ātmadarśanam ātmīyagrahaṃ prasūte /. PV 2.217: yaḥ paśyaty ātmānaṃ tatrāsyāham iti śāśvataḥ snehaḥ / sne­ hāt sukheṣu tṛṣyati tṛṣṇā doṣāṃs tiraskurute //. Note that Śākyabuddhi interprets doṣa as jātijarāmaraṇa (PVṬ D138b1/P170b8). 35 PVP D95a6/P111a2, PVV 87,3. 36 PVP D95a6/P111a2: bdag gi bde la sred ’gyur … 37 PVV 87,3–4: sukhasādhanatvenādhyavasitānāṃ vastūnām … 38 PVP D95b1/P111a4–5: bde ba sgrub par byed pa ñid du ñe bar ’gro ba zag pa daṅ bcas pa’i dṅos po … On anugrāhaka, see also PVSVṬ 402,8:
34 33

Ignorance, epistemology and soteriology (2)

37

as ātmasneha,39 ātmātmīyasneha40 or even satkāyasneha,41 we also find Devendrabuddhiʼs definition of sneha: “[We call] ʽloveʼ an inclination for self and oneʼs own which presupposes the [aforementioned delusion].” 42 According to Dharmakīrti, self-love and attachment for what belongs (or ought to belong) to the self is in turn the cause of aversion (pratigha) and hatred (dveṣa): “Indeed, the one who, without grasping (parigraha), sees that there is neither I nor mine, does not love anything and, [being so] unattached, does not hate anything [either], for there is no [aversion] for that which does not hinder the self or one’s own, nor for that which opposes the [said] hindrance.” 43 One can show aversion or hatred only for that which hinders (< uparodha) or harms (< pīḍā) what has been taken as self and oneʼs own:44 “Hatred [arises] with regard to
ātmātmīyatvena tadanugrāhakatvena parikalpya … 39 E.g., PVP D58a1–2/P66a3. 40 PVP D57b3/P65b4. Love for self and one’s own is said to be directed to the object that is clung to as self and one’s own (ātmātmīyatvābhiniviṣṭe viṣaye ātmātmīyasnehaḥ, PVSVṬ 401,26–27). 41 E.g., PVP D90b5/P104b7: ’jig tshogs la chags pa. 42 PVP D60a2/P68b2 –3: de sṅon du soṅ ba can gyi bdag daṅ bdag gir źen pa ni chags pa’o //. Note also PVP D94b7/P109b4–5: chags pa ni bdag tu mṅon par chags pa’o // (maybe: sneha ātmany abhiṣvaṅgaḥ). 43 PVSV 111,15–17: na hi nāhaṃ na mameti paśyataḥ parigraham antareṇa kvacit snehaḥ / na cānanurāgiṇaḥ kvacid dveṣaḥ / ātmātmīyānuparodhiny uparodhapratighātini ca tadabhāvāt /. 44 According to PVSVṬ 402,12: ātmātmīyatvena gṛhītasya ya uparodhaḥ pīḍā /. Note also Devendrabuddhi’s definition of dveṣa at PVP D60a2/P68b3: de (= chags pa) sṅon du soṅ ba can rjes su chags pa’i yul la gnod par byed pa la mnar sems pa ni źe sdaṅ ṅo //. “Hatred is maliciousness with regard to that which injures the object of attachment[, a maliciousness] that presupposes the [afore-mentioned love].” The Sanskrit original for Tib. mnar sems pa is unclear. I would conjecture vyāpannacitta, although, to the best of my knowledge, mnar (ba) is not attested as a translation of vyāpanna(/vyāpāda): vyāpannacitta = gnod sems at AKBh 251,10 and 12 on AK 4.81ac1 (“de pensée méchante” in Kośa 4.178) as well as in the Saṃcetanī yasūtra quoted in AKVy 400,9–15 on AKBh 237,18. Jaini 2001:221: “The kleśas are like roots which produce as well as sustain an evil volition. Abhidhyā, vyāpāda, and mithyādṛṣṭi are not called roots, but are recognized as intensive states of the three roots of evil (akuśalamūla), viz. lobha, dveṣa, and moha respectively.

38

Vincent Eltschinger

that alone which offers opposition (pratikūlavartin) by its hostility to that which love for the self and oneʼs own bears upon (viṣayabhūta). Therefore, there is no hatred without love for the self and oneʼs own.” 45 Dharmakīrtiʼs unambiguous derivation of aversion from love is the reason why I cannot agree with Devendrabuddhiʼs and Manorathanandinʼs interpretation of PV 2.219b (svapara­ vibhāgāt parigrahadveṣau), which presupposes that what is other than the self can only arouse hatred. In Dharmakīrtiʼs eyes, that which is other than the self gives rise to aversion only insofar as it opposes love, but arouses craving as soon as it is regarded as pleasurable to the self. Craving for the delights of the self and that which is conducive to them generally implies oneʼs running around in search of pleasure. This is indeed the Vaibhāṣika definition of the ninth link of dependent origination, appropriation or clinging (upādāna),46 and what Dharmakīrti obviously has in mind in PV 2.218ab: “Seeing [but] qualities [to the things that he deems pleasurable to the self], he craves [for them, thinking of them as having to become] ‘mine,’ and appropriates (upā√dā) the means [that are conducive] to them.” 47 But Dharmakīrti also holds love for the self to be the cause of the three different kinds of craving that the oldest layers of Buddhist canonical literature have made responsible for rebirth (paunarbhavika): craving for (future) existence (bhavatṛṣṇā), craving for sensual pleasures (kāmatṛṣṇā), and craving for non-existence/annihilation (vibhavatṛṣṇā).48 According
All evil volitions are essentially rooted in and spring from one or another of these three basic passions (mūlakleśa).” 45 PVSVṬ 402,13–15: ātmātmīyasnehaviṣayabhūtavirodhena yaḥ sthitaḥ pratikūlavartī tatraiva dveṣaḥ / tasmān nātmātmīyasneham antareṇa dveṣa iti /. 46 AK 3.23cd: upādānaṃ tu bhogānāṃ prāptaye paridhāvataḥ /. 47 PV 2.218ab (āryā metre): guṇadarśī paritṛṣyan mameti tatsādhanāny upādatte /. 48 PVP D79b3–4/P91a7–8: de la sdug bsṅal kun ’byuṅ ’phags pa’i bden pa gaṅ źe na / gaṅ sred pa ’di ni yaṅ srid par ’byuṅ ba can dga’ ba’i ’dod chags daṅ bcas pa de daṅ de la mṅon par dga’ ba’i ṅaṅ tshul can / ’di lta ste ’dod pa’i sred pa daṅ srid pa’i sred pa daṅ ’jig pa’i sred pa yin no źes gsuṅs so //. PVA 134,33–135,2: uktaṃ hi bhagavatā tatra katamat samudaya āryasatyam / yeyaṃ tṛṣṇā paunarbhavikī nandīrāgasahagatā tatratatrābhinandinī / yad

Ignorance, epistemology and soteriology (2)

39

to him, craving for sensual pleasures is to be interpreted as the actions (pravṛtti) of living beings to secure what they hold to be pleasurable (sukhāpti), whereas craving for annihilation refers to those of their actions that aim at avoiding suffering (duḥkhānāpti). This matches again perfectly with the Vaibhāṣika account of the tenth link of dependent origination, viz. bhava (literally “existence”), which is to be understood as the “act(ion) that results in future existence” (bhaviṣyadbhavaphalaṃ karma): bhava refers to the act(ion)s resulting in rebirth (paunarbhavika) that are accumulated by those who run around (under the sway of craving) in order to quench their thirst.49 In these stanzas, Dharmakīrti brings together both meanings of bhava, i.e., action to secure the pleasures of the self, and the (future) existence that they inevitably lead to: “The cause [of suffering] is the longing for [re]existence, because human beings reach a specific [existential] place [and condition] due to [their] hope of obtaining it. The [afore-mentioned longing for existence] is [called] the desire for [re]existence. And since a living being [only] acts with the desire of obtaining pleasure and avoiding suffering, these two [i.e., craving for pleasure and craving for the avoidance of suffering,] are regarded as the desire for sensual pleasures and the desire for annihilation. And since love for the self is the cause [of it, this dual action] pertains to everything for [the living being] who has the notion of [something] pleasurable with regard to [something] unpleasurable. Therefore, craving is the basis of existence [i.e., the cause of bondage].”50

uta kā[m]atṛṣṇā bhavatṛṣṇā vibhavatṛṣṇā ceti … PVV 74,10–11: nanūktaṃ bhagavatā tatra katamaḥ samudaya āryasatyaṃ paunarbhavikī nandīrāgasahagatā tatratatrābhinandinī yad uta kāmatṛṣṇā bhavatṛṣṇā vibhavatṛṣṇā ceti … For the Pāli text, see Vetter 1990: 87, n. 1. 49 AKBh 132,19–21 (together with AK 3.24ab): sa bhaviṣyadbhavaphalaṃ kurute karma tad bhavaḥ / sa viṣayāṇāṃ prāptihetoḥ paridhāvan paunarbhavikaṃ karmopacinoti so ’sya bhavaḥ /. 50 PV 2.183a2–185: hetur bhavavāñchā parigrahaḥ / yasmād deśaviśeṣasya tatprāptyāśākṛto nṛṇām // sā bhavecchā ’’ptyanāptīcchoḥ pravṛttiḥ sukhaduḥkhayoḥ / yato ’pi prāṇinaḥ kāmavibhavecche ca te mate // sar vatra cātmasnehasya hetutvāt sampravartate / asukhe sukhasaṃjñasya tasmāt tṛṣṇā bhavāśrayaḥ //.

40

Vincent Eltschinger

2.1.5. Although the standard formulation of dependent origination is traditionally held to range over three (Vaibhāṣika) or two (Yogācāra, Sautrāntika) lifetimes,51 at least some of its members can also be seen at work on the much shorter sequence of a few interdependent psychological events. According to Vasubandhu, desire follows (anuśete, or: is connected to, samprayukta) a pleasant sensation (sukhā vedanā), whereas aversion follows (or: is connected to) an unpleasant sensation (duḥkhā vedanā).52 Dharmakīrti agrees with this commonsense statement.53 Depending on whether a given tangible object (spraṣṭavya) is considered favourable (anugrāhaka) or unfavourable to the self, the pleasant or unpleasant sensations born from the contact between this object and the sense faculties are conducive to the rise of defilements such as desire or hatred.54 This obviously conforms to the pratītyasamutpāda sequence linking a sensory basis (āyatana), contact (sparśa) between the former and an object, sensation, and craving. But as we have seen, to deem a given object favourable or unfavourable to the self belongs to the personalistic false view. Note should be made here that the erroneous aspects which the personalistic false view consists of overlap in part with those traditionally called “wrong notions” or “misconceptions” (viparyāsa), which amount to four55 and

For a useful overview, see Kritzer 1999: 67–72. AKBh 312,1–2: trivedanāvaśāt trīṇi bandhanāni / sukhāyāṃ hi vedanāyāṃ rāgo ’nuśete ālambanasamprayogābhyām / duḥkhāyāṃ dveṣaḥ /. AK 5.55ab + AKBh 316,6 and 8: sukhābhyāṃ samprayukto hi rāgaḥ / sukhasaumanasyābhyāṃ rāgaḥ samprayuktaḥ / dveṣo viparyayāt / duḥkhābhyām ity arthaḥ / duḥkhena daurmanasyena ca /. 53 See PV 2.151c2d: rāgāder vikāro ’pi sukhādijaḥ /, and the discussion below. 54 According to PVP D66a5–6/P75b5–6: reg bya’i khyad par gyi don phan ’dogs par byed pa daṅ de las gźan pa’i rjes su byed pas bde ba’am sdug bsṅal lam (sic) ’dod chags la sogs pa skye ba daṅ rjes su mthun pa yin pa … 55 To take the impermanent as permanent, the painful as pleasant, the impure as pure, and the selfless as a self (AKBh 283,5–7: catvāro viparyāsāḥ / anitye nityam iti / duḥkhe sukham iti / aśucau śucīti / anātmany ātmeti /). With the exception of the (im)pure, they correspond to the erroneous aspects one superimposes on the Truth of suffering (see above, n. 8).
52

51

Ignorance, epistemology and soteriology (2)

41

are regularly held to be caused by imagination (saṅkalpa).56 Śāntarakṣita and Kamalaśīla provide interesting materials regarding the rise of defilements from wrong notions. According to Śāntarakṣita, “defilements such as desire arise once [erroneous aspects] such as beautiful, oneʼs own, lasting [or pleasant] have been superimposed on a woman, etc.”57 A little later, he says: “[A sensation] such as a pleasant or unpleasant [one] arises in the presence of a [sensory] object[, say a woman]. For those who despise [suspending] wisdom (pratisaṅkhyāna) [and] are subject to improper reflection, this [sensation] gives rise to defilements such as desire or hatred, which are [themselves] born from the ripening of a homologous latent tendency.”58 What does this amount to? The contact between an obOn saṅkalpa, see May 1959: 181n. 586, PrP 451,9 ff., and the following excerpts: PVP D68a4–5/P77b8–78a1: ci ste ’di la yaṅ kun tu rtog pa yaṅ yan lag ñid du rtog par ’gyur ba de’i tshe kun tu rtog pa yaṅ bdag daṅ bdag gi daṅ gtsaṅ ba daṅ bde ba la sogs pa’i miṅ can gyi mtshan ñid kyi sa bon yin no //. PVP D67a3–4/P76b5–6: gaṅ gis bud med ’ga’ źig gi gzugs la sogs pa la kun tu rtog par byed ciṅ ’dod chags kyis gduṅs pa de ni … TSP Ś666,25–667,9/ K547,8–9: atītānāgate ’pi viṣaye saṅkalpavaśād abhivṛddhasukhādiviparyāsasya puṃsaḥ pratisaṅkhyānanivṛttau teṣāṃ rāgādīnāṃ prabalatvaṃ dṛśyate /. MMK 23.1: saṅkalpaprabhavo rāgo dveṣo mohaś ca kathyate / śubhāśubhaviparyāsān sambhavanti pratītya hi //. PrP 452,4–5: tatra hi śubham ākāraṃ pratītya rāga utpadyate / aśubhaṃ pratītya dveṣaḥ / viparyāsān pratītya moha utpadyate / saṅkalpas tv eṣāṃ trayāṇām api sādhāraṇakāraṇam utpattau /. PVSVṬ 166,29–167,2 gl. saṅkalpita (PV 1.70d) as āropita. To sum up, saṅkalpa is the bīja of the wrong notions or, equivalently, of the erroneous aspects, which in turn form the bases (āśraya < āśritya) or conditions (pratyaya < pratītya) of the defilements; to put it as shortly as Candrakīrti, saṅkalpa is the common cause (sādhāraṇakāraṇa) for the rise of the defilements. On saṅkalpa, see also below, nn. 68 and 69. 57 TS Ś1951ac/K1952ac: śubhātmīyasthirādīṃś ca samāropyāṅganādiṣu / rāgādayaḥ pravartante … “Pleasant” according to TSP Ś667,13–14/ K547,12–14 thereon: ātmā*tmīyanityasukhādyākārān abhūtān evāropayanto ’ṅganādiṣu pravartante, na ca śubhādirūpā viṣayāḥ /. *TSPK with no equivalent of ātmā°. 58 TS Ś1953–1954d1/K1954–1955d1 (leaving tu untranslated): viṣayopanipāte tu sukhaduḥkhādisambhavāḥ / tasmāt samānajātīyavāsanāparipākajāḥ  // rāgadveṣādayaḥ kleśāḥ pratisaṅkhyānavidviṣām  / ayoniśomanaskāravidheyānām … Note also PV 2.157ac: sajātivāsanābhedapratibaddhapravṛttayaḥ / … rāgādayaḥ … PVV 66,8–10: sajātivāsanā
56

42

Vincent Eltschinger

ject and a sense faculty generates an affective sensation (pleasant, unpleasant, or neutral). People who do not devote themselves to meditative practices such as the contemplation of the loathsome (aśubhabhāvanā),59 and are therefore under the sway of improper reflection, superimpose erroneous aspects on the object: that it is a women, of course, but also that she is attractive, desirable, (at least virtually) oneʼs own, etc. Affective sensation as well as the superimposed aspects is in turn responsible for the actualization of the latent tendency of desire.60 Commenting on his masterʼs two stanzas, Kamalaśīla provides us with a more systematic account of the sequence at stake: “For such is the sequence [of events]: When an object is present, a pleasure born of the sense faculty arises. And for those who, in the absence of any [suspending] wisdom, abide in the improper reflection consisting of wrong notions such as self, this pleasure brings to maturity (vipāka) the latent tendency imprinted by previous desire, etc. From this [coming to] maturity, defilements such as desire arise. Therefore, the objects [themselves] are not directly the cause [of defilements].”61 How should we un’’tmātmī yagrahamūlasya sajāteḥ (Vibh. 66 n. 1: satkāyadarśanasya) pūrvapūr vābhyastasya rāgāder vāsanā ’parāpararāgādijanikāḥ śaktayas tāsāṃ bhedaḥ parasparatas tatra pratibaddhā pravṛttir janma yeṣāṃ te tathā … Here, sajātivāsanā is analysed as a genitive tatpuruṣa: “latent tendencies of the homologous [defilements which are rooted in the belief in self and one’s own].” But according to Devendrabuddhi and Śākyabuddhi, the compound is to be analysed as a dvandva (PVṬ D123a2–3/P150b7): sajāti refers to the satkāyadṛṣṭi (ātmātmī yadṛṣṭi in PVP D68a6–8/P78a3– 5) whereas the vāsanā(bheda) consists in the pūrvarāgādyāhitabīja. 59 TSP Ś666,22–23/K547,6: aśubhādipratisaṅkhyāna. According to PVP D67a6–7/P77a1–2, rāgādi do not occur in those who have the aśubhādisaṃjñā. Note also Kamalaśīla’s definition at TSP Ś666,23/K547,6–7: aśubhādyālambanā rāgādipratipakṣabhūtā prajñā pratisaṅkhyānam /, which may be compared with AKVy 389,13 on AKBh 226,13–14: pratisaṅkhyānasya tatpratipakṣabhāvanālakṣaṇasya, where tat = kleśa (context: nirvāṇa). Note also AKBh 4,1 on AK 1.6ab1: duḥkhādīnām āryasatyānāṃ pratisaṅkhyānaṃ pratisaṅkhyā prajñāviśeṣaḥ … (see also Kośa 1.8, and AKVy 16,4–7). 60 On latent tendencies and their actualization, see Eltschinger 2009: 57– 58, nn. 53–55. 61 TSP Ś667,19–22/K547,26–548,2: eṣa hi kramaḥ – viṣayopanipāte satīndriyajaṃ sukham utpadyate, tasmāc ca sukhāt pratisaṅkhyānavaikalye

Ignorance, epistemology and soteriology (2)

43

derstand this strong insistence on the responsibility of improper reflection in the rise of defilements? 2.1.6. That improper reflection62 is closely connected with ignorance/personalistic belief and is part of the process leading to the rise of defilements can be easily substantiated.63 The problem raised by the source materials is rather that they testify to contradictory views regarding the relationship between improper reflection and ignorance/personalistic belief. Some sources (mainly Yogācāra) introduce improper reflection in the definition of the personalistic belief, which is held to be the manner deluded people improperly consider the five constituents of being as self and
saty ātmādiviparyāsalakṣaṇāyoniśomanaskāre sthitānāṃ pūrvarāgādyāhitavāsanāparipāko bhavati, tato rāgādayaḥ kleśāḥ pravartanta iti na sākṣād viṣayāḥ kāraṇam /. Note also Prajñākaragupta’s remarks while commenting on Dharmakīrti’s polemics against a Materialist upholding medical ideas (PVA 122,22–23): sukhādijo hi rāgādir na kaph[ā]dibhāvī / sukhaṃ ca kasyacit kathaṃcid upalabdham āntaravāsanāprabodhāt / tato na rāgādayo doṣebhya iti yuktam /. Though Śāntarakṣita and Kamalaśīla cannot be suspected of allegiance toward Vaibhāṣika thought, their views are reminiscent of an interesting passage in the AK(Bh), according to which a defilement arises out of three factors: first, its propensity (anuśaya) has not been eliminated; second, an object (viṣaya, dharma) that is conducive to the actualization of desire for sensual pleasures (kāmarāgaparyavasthānīya) is present and perceived (ābhāsagata); thirdly, an improper reflection occurs with regard to the said object. AK 5.34, together with AKBh 305,19–20: aprahīṇād anuśayād viṣayāt pratyupasthitāt / ayoniśomanaskārāt kleśaḥ – tad yathā rāgānuśayo ’prahīṇo bhavaty aparijñātaḥ kāmarāgaparyavasthānīyāś ca dharmā ābhāsagatā bhavanti tatra cāyoniśomanaskāra evaṃ kāmarāga utpadyate /. 62 AKBh 54,23: manaskāraś cetasa ābhogaḥ /. AKVy 127,33–128,2 thereon: manaskāraś cetasa ābhoga iti / ālambane cetasa āvarjanam / avadhāraṇam ity arthaḥ / manasaḥ kāro manaskāraḥ / mano vā karoty āvarjayatīti manaskāraḥ /. PVSVṬ 50,29–51,12: ayoniśa ityādy asyaiva samarthanam / yoniḥ padārthānām anityaduḥkhānātmādi / samyagdarśanapras[ū]tihetu­ tvāt / taṃ śaṃsaty ālambata iti yoniśaḥ / yoniṃ yoniṃ manaskarotīti saṃkhyaikavacanād vīpsāyām (Pā 5.4.43) iti śaspratyayo vā / tathābhūtaś cāsau manaskāraś ceti yoniśomanaskāro nairātmyajñānam /. 63 On ayoniśomanaskāra and avidyā, see La Vallée Poussin 1913: 8–9, and especially Mejor 2001. On the improper reflection’s conditioning and reinforcing dṛṣṭis, see the passage of AN I.31 alluded to by Mejor (2001: 50 + n. 5); see also AKBh 5.32–33 in Mejor 2001: 51.

44

Vincent Eltschinger

one’s own.64 Some materials regard improper reflection as caused by ignorance: this is the case of the Sūtra quoted in the AKBh, according to which, “depending on the eye and visible [objects,] an incorrect (āvila) reflection born of delusion (mohaja) arises.”65 Much more common seem to be sources viewing improper reflection as the cause of ignorance/personalistic belief: this is the case in a Sutta of the MN and two Suttas from the AN,66 in the MS,67 in the Sat yadvayanirdeśa(sūtra) as it is quoted by Kamalaśīla in BhK 1,68 and in the Sahetusapratyayanidānasūtra as it is quoted in AKVy 288,26–29 and used by Bhadanta Śrīlāta to demonstrate that ignorance (as the first link of dependent origination) has indeed a cause.69 Having quoted and commented on various excerpts
See Eltschinger 2009: 68–69, nn. 92 and 110. AKBh 135,13–14 and AKVy 288,30–31: cakṣuḥ pratītya rūpāṇi cot­ padyate āvilo manaskāro mohaja iti /. Note also AKBh 135,7 (in a quotation): avidyāhetukaś cāyoniśomanaskāraḥ /. 66 MN I.6 ff. (no. 2, Sabbāsavasutta). Here, the ayoniso manasikāra is held to be responsible for the rise (uppajjhanti) and the increase (pavaḍḍhanti) of the three cankers (kāmāsava, bhavāsava and avijjāsava), which are in turn responsible for the rise of false views (diṭṭhi) concerning personal identity in the past (atītaṃ addhānam), in the future (anāgataṃ addhānam) and in the present (paccuppannaṃ addhānam), such as atthi me attā ti and na­tthi me attā ti. On this passage, see Collins 1982: 118–119; for similar expressions of the satkāyadṛṣṭi/sakkāyadiṭṭhi, see Eltschinger 2009: 73–75. AN V.113 ff. (no. 61, Avijjāsutta) and V.116 ff. (no. 62 Taṇhāsutta). According to the Avijjāsutta, ayonisomanasikāra belongs to the eight aliments (āhāra) of avijjā; see Mejor 2001: 52–55. 67 MS 2.20.9 (Lamotte 1973: I.34): mṅon par źen pa’i rnam par rtog pa ni ’di lta ste / tshul bźin ma yin pa’i yid la byed pa las byuṅ ba’i ’jig tshogs la lta ba’i rtsa ba las byuṅ ba lta bar soṅ ba drug cu rtsa gñis daṅ mtshuṅs par ldan pa’i rnam par rtog pa gaṅ yin pa’o //. See also Lamotte 1973: II.115. 68 BhK 1.215[/525],7–14: kathaṃ mañjuśrīḥ kleśā vinayaṃ gacchanti / kathaṃ kleśāḥ parijñātā bhavanti / mañjuśrīr āha / paramārthato ’tyantājātānutpannābhāveṣu (sic, <Tib, but °nabhā° ms) sarvadharmeṣu saṃvṛtyāsadviparyāsaḥ / tasmād asadviparyāsāt saṃkalpavikalpaḥ / tasmāt saṃkalpavikalpād ayoniśomanasikāraḥ / tasmād ayoniśomanasikārād ātmasamāropaḥ / tasmād ātmasamāropād dṛṣṭiparyutthānam / tasmād dṛṣṭiparyutthānāt kleśāḥ pravartante /. 69 anyaḥ in AKBh 135,12, Bhadanta Śrīlāta according to AKVy 289,23; AKBh 135,12–17: anyaḥ punar āha / ayoniśo manaskāro hetur avidyāyā uktaḥ
65 64

Ignorance, epistemology and soteriology (2)

45

of the Sūtra, Yaśomitra comes to the conclusion of a circularity (cakraka), i.e., that improper reflection and ignorance condition each other.70 This is indeed the position most clearly exhibited by the Paramārthagāthās.71 To the best of my knowledge, Dharmakīrti alludes only twice to improper reflection in the context of the rise of defilements. Unfortunately, both statements are far from unambiguous. In PVSV 8,20–21, Dharmakīrti says that “[moral faults] such as desire presuppose [one’s] adherence to self and one’s own, for the rise of

sūtrāntare / sa cāpi sparśakāle nirdiṣṭaḥ / cakṣuḥ pratītya rūpāṇi cotpadyate āvilo manaskāro mohaja iti / vedanākāle cāvaśyam avidyayā bhavitavayam / avidyāsaṃsparśajaṃ veditaṃ pratītyotpannā tṛṣṇeti sūtrāntarāt / ataḥ sparśakāle bhavann ayoniśomanaskāro vedanāsahavartinyā avidyāyāḥ pratyayabhāvena siddha iti nāsty ahetukatvam avidyāyāḥ … See Kośa 3.71n. 4. The whole discussion starts with the Sthavira Vasubandhu’s (AKVy 289,6: sthaviro vasubandhur ācāryamanorathopādhyāya evam āha…) claim that ignorance is not causeless on the basis of a Sūtra (the Sahetu­ sapratyayasanidānasūtra according to AKVy 288,25–26; AKBh 135,7: ayoniśomanaskārahetukā ’vidyoktā sūtrāntare /). As quoted by Yaśomitra (AKVy 288,26–29), this Sūtra runs as follows: avidyā bhikṣavaḥ sahetukā sapratyayā sanidānā / kaś ca bhikṣavo ’vidyāyā hetuḥ kaḥ pratyayaḥ kiṃ nidānam / avidyāyā bhikṣavo ’yoniśomanaskāro hetur ayoniśomanaskāraḥ pratyayo ’yoniśomanaskāro nidānam iti sūtre vacanāt /. This passage is also quoted in PrP 452,7–9 (avidyāpi bhikṣavaḥ sahetukā sapratyayā sanidānā / kaś ca bhikṣavo ’vidyāyā hetuḥ / ayoniśo bhikṣavo manaskāro ’vidyāyā hetuḥ / āvilo mohajo manaskāro bhikṣavo ’vidyāyā hetur iti), but as coming from the Pratītyasamutpādasūtra (PrP 452,6 [but see n. 3 thereon], Kośa 3.70n. 3). Immediately after the quotation, Candrakīrti remarks (PrP 452,9): ato ’vidyā saṅkalpaprabhavā bhavati /. Note also Yaśomitra’s (AKVy 289,1) reference to the Pratītyasamutpādasūtra. Mejor (2001: 61–65) has translated Vasubandhu’s polemics against Śrīlāta (AKBh 134,20–25 and 135,7–27). 70 AKVy 290,5–7: tad etac cakrakam uktaṃ bhavati / ayoniśomanaskārād avidyā / avidyāyāś cāyoniśomanaskāra iti /. This is, indeed, the position of the Sahetusapratyayasanidānasūtra (1. [moha] → āvilo manasikāra → ayoniśomanaskāra → avidyā → tṛṣṇā → karman → cakṣus [but also ear, nose, tongue, body and mind]; 2. cakṣus → karman → tṛṣṇā → avidyā → ayoniśomanaskāra); see above, n. 69, and Mejor 2001: 58 and 65–69 (Mejor’s translation of the Sūtra from Tibetan and Chinese sources). 71 Paramārthagāthā 20 (Wayman 1961: 170): ayoniśomanaskārāt saṃmoho jāyate sa ca / ayoniśomanaskāro nāsaṃmūḍhasya jāyate //.

46

Vincent Eltschinger

all moral faults presupposes improper reflection.”72 A little later, he refers to a “specific condition for the rise of desire, viz. improper reflection that consists in the false view of self/viz. the false view of self and improper reflection.”73 Commenting on the first passage, Śākyabuddhi and Karṇakagomin clearly equate the personalistic belief with improper reflection.74 But commenting on the second passage, they allow both a dvandva and a karmadhāraya analysis of the compound ātmadarśanāyoniśomanaskāra.75 Though I am inclined to interpret these two passages as involving an equivalence between the false view of self and improper reflection, I would like to refrain from any conjecture regarding Dharmakīrti’s position on this issue.76 In the same way, I would like to postpone any attempt at organizing the above-mentioned (§2.1.5–6) psychological events into a sequence of phases exhibiting their mutual relationships. At

PVSV 8,20–21: ātmātmīyābhiniveśapūrvakā hi rāgādayo ’yoniśomanaskārapūrvakatvāt sarvadoṣotpatteḥ /. 73 PVSV 10,11: rāgotpattipratyayaviśeṣeṇātmadarśanāyoniśomanaskāreṇa yogāt /. 74 PVṬ Je D23b1–2/P28a1–2 = PVSVṬ 51,12–13: ātmādijñānam ayoniśomanaskāras tatpūrvakatvāt sar varāgādidoṣotpatteḥ /. 75 PVṬ Je D27a2–3/P32a5–7 = PVSVṬ 55,29–56,12: ātmadarśanaṃ satkāyadṛṣṭiḥ / nityasukhādiviparyāso ’yoniśomanaskāraḥ / dvandvasamāsaś cāyam / ātmadarśanam evāyoniśomanaskāra iti viśeṣaṇasamāso vā /. Interestingly enough, Śākyabuddhi and Karṇakagomin explain “improper reflection” as a “wrong notion such as permanent or pleasant,” which matches perfectly Kamalaśīla’s definition of “improper reflection” as “wrong notion such as self.” According to these authors, then, improper reflection and wrong notions are conceptually equivalent. See above, n. 61. 76 Lambert Schmithausen (personal communication) has drawn my attention to the possibility that in the first passage (PVSV 8,20–21), Dharmakīrti may not be providing a logical justification, but rather a legitimation of his position by resorting to a more traditional phraseology involving a co-extensivity of the two concepts: “d.h. weil sie [bekanntermaßen] ayoniśomanaskāra voraussetzen(, und dieser in nichts anderem besteht als eben dem ātmātmīyābhiniveśa).” By interpreting the compound in the second passage (PVSV 10,11) as a karmadhāraya, one may, then, read the two passages as exhibiting a homogeneous perspective.

72

Ignorance, epistemology and soteriology (2)

47

the present state of research, such an attempt would only be idle speculation.77 2.1.7. Both wrong notions and the personalistic false view consist in the superimposition of erroneous aspects. Both are born of the actualization of a homogeneous latent tendency, which is the hallmark of conceptual construction. In other words, they are but conceptual constructs distorting both internal (the upādānaskandhas) and external reality. Dharmakīrti’s understanding of the personalistic belief harmonizes perfectly well with his overall conception of ignorance as the concealing conceptuality. As for his commentators, they seem to be justified in holding the satkāyadṛṣṭi to be a part, a branch or a specific case of ignorance as a whole. That all conceptual constructs misrepresent reality, and sometimes are even deceiving from a practical point of view, does in no way mean that they are morally and (hence) eschatologically harmful. The su-

77 To the best of my knowledge, no study has ever been dedicated to the issue of the Buddhist epistemologists’ way(s) of dealing with the Abhidharmic cittasamprayuktasaṃskāras. Their assent to Vasubandhu’s treatment of them cannot be taken for granted. To adduce but one example: saṃjñā is classified as a mahābhūmika, and as such, should occur together with vijñāna/citta/ manas; but niścaya(jñāna), the Buddhist epistemologists’ equivalent of saṃjñā, takes place after the sensory awareness (the latter giving rise to the vāsanāprabodha of the conceptual construct). In the present context, I think we should refrain from modelling the epistemologists’ conception of ignorance and improper reflection on Bhadanta Śrīlāta’s above-mentioned (see n. 69) elaborations on this topic. According to him, the improper reflection that is present at the moment of contact (sparśakāle) is the condition (pratyaya) for the ignorance that coexists with sensation (vedanāsahavartiny avidyā) and in turn gives rise to craving. On the contrary, an Arhat’s unbiased (aviparīta) contact does not give rise to a defiled sensation (kliṣṭā vedanā), which in turn does not provide a condition for craving. As both Śrīlāta (at least Vasubandhu’s Śrīlāta) and Yaśomitra describe it, Arhats do have sensations, but these do not generate craving, for only sensations that are accompanied by ignorance (sāvidya) give rise to craving (AKVy 290,13–15: arhatām asti vedanā / na ca sā tṛṣṇāyāḥ pratyayībhavatīti / sāvidyaiva vedanā tṛṣṇāpratyaya iti gamyate /.) Śrīlāta adduces a reasoning (yukti) in order to make his point (AKBh 135,20–22): kayā yuktyā / na hi niravadyā vedanā tṛṣṇāyāḥ pratyayībhavaty arhatāṃ na cāviparītaḥ sparśaḥ kliṣṭāyā vedanāyāḥ / na ca punar niravadyasyārhataḥ sparśo viparīta ity anayā yuktyā /).

48

Vincent Eltschinger

perimposition of ego-related aspects alone results in the rise of defilements and reinforces one’s entanglement in saṃsāra. Dharmakīrti singles out this kind of harmful conceptual distortion as the personalistic belief.78

2.2. Ignorance, inference, and the path toward salvation
2.2.1. Like most Indian systems of salvation, Buddhism traces human beingsʼ unsatisfactory condition back to ignorance, and presents itself as a cleansing and illuminative therapy aimed at uprooting ignorance and the evils it is responsible for. Though the Buddhist epistemologists do not (even pretend to) bring any doctrinal or practical innovation into traditional Buddhist soteriologies, they lay strong emphasis on the means of valid cognition (pramāṇa) as being instrumental in salvation. As is well known, Dignāga reduced the number of genuine pramāṇas from three (perception, inference, and scriptures [āgama]) to two (perception and inference). At the present state of our knowledge about Dignāga, however, it is difficult to estimate the extent to which non-epistemological, i.e., religious (lato sensu) considerations played a role in this epistemological reduction. If one cannot question Dharmakīrtiʼs endorsement and consolidation of Dignāgaʼs two-headed system as far as the epistemology is concerned, one might still argue that Dharmakīrtiʼs religious ideas, as they are known to us, provided, if not the basic framework, at least a strong additional motivation for sticking to this epistemology. This two-headed system could, after all, lay no claim to traditionally sanctioned authority before Dignāga.79 In my opinion, Dharmakīrti was deeply convinced that
PVSV 110,20–21: te [= doṣāḥ] vikalpaprabhavāḥ /. PVSVṬ 398,23–25 thereon: vikalpād ayoniśomanasikāravikalpāt prabhava utpāda eṣām iti vigrahaḥ / tathā hy ayoniśomanaskāram antareṇa saty api bāhye ’rthe notpadyante rāgādayaḥ … 79 According to Frauwallner (1959), Vasubandhu had already restricted the number of pramāṇas from three to two in his Vādavidhi. But this might well be another case of Frauwallner’s use of the argumentum ex/a silentio: the fact that no fragment dealing with (āpt)āgama is available to us does not mean that the original Vādavidhi did not address scripture as a third genuine means of valid cognition. At any rate, Vasubandhu seems to acknowledge
78

Ignorance, epistemology and soteriology (2)

49

perception and inference are enough both to shape and bring about the path to salvation and to provide the basic gnoseological features of the liberated yogin. To put it in a nutshell: Although it is conceptual in nature and thus belongs to ignorance, inference is the means through which perception, which is nothing but “knowledge,” can be brought to function in its most genuine manner. Dharmakīrtiʼs system is in a way analogous to Tathāgatagarbha patterns of thought: though polluted by (ultimately adventitious) false views and defilements, the condition of the liberated mind is already here at hand. To be more precise, perception is basically the same with regard to its operation and objects before and after the revolution of the basis (āśrayaparivṛtti). The only (but admittedly crucial) difference is that, at the completion of the path, it is no longer adulterated and contradicted by the counteracting cognitive factor called “ignorance.” Correcting erroneous superimpositions of all kinds and substituting them with true/validated intellectual contents is the basic task of inference. Far from being a means of investigating the world and improving knowledge, inference aims first and fore-

three means of valid cognition in his AKBh (76,24–25: pramā ṇābhāvāt / na hi … pramāṇam asti pratyakṣam anumānam āptāgamo vā …) as well as in VY 173,16–17: mdor na rigs pa ni ’dir tshad ma rnam pa gsum po mṅon sum daṅ rjes su dpag pa daṅ yid ches pa’i gsuṅ ṅo //). Buddhist eristic-dialectical treatises are at great variance concerning the number (and definitions) of the pramāṇas: four (or five) in the Hetuvidyā Section of the YBh (see, e.g., HV [§3.2] 4*,15–16, where the last five items of the list defining sādhana must be considered as pramāṇas because of their functional similarity (providing evidence [yuktivāda] for the hetu, see HV [§3.22] 5*,3–5): sārūpyaṃ vairūpyaṃ pratyakṣam anumānam āptāgamaś ca; to the best of my knowledge, the HV only uses the term pramāṇa with regard to pratyakṣa; therefore, the number of the pramāṇas here is either five [or four if we consider that sārūpya and vairūpya occur once in a singular dvandva compound] or only one), four in the *Upāyahṛdaya/*Prayogasāra (*pratyakṣam anumānam upamānam āgamaś ca; see *UH 6,10–11 and 13,5 ff.), three in Asaṅga’s Abhidharmasamuccaya (which, maybe on the basis of the BoBh and the Madhyāntavibhāga, sets the standard number for all subsequent Yogācāra treatises), i.e., pratyakṣa, anumāna and āptāgama (see ASBh 152,27, 153,1 and 153,5).

50

Vincent Eltschinger

most at discarding the erroneous superimpositions that ignorance is ultimately responsible for.80 2.2.2. As we have seen, ignorance basically amounts to superimposition, concealment/covering, conceptual construct and pseudoperception. As such, ignorance is of a cognitional character and consists in an “anti-knowledge,” in a mental event counteracting, contradicting or conflicting with “knowledge.” What does, then, “knowledge” consist in? As we have seen, Dharmakīrtiʼs commentators define it as the “vision/perception of a real object” (bhūtārtha°/sadarthadarśana), or the “grasping of a real object” (bhūtārthagrahaṇa).81 In these expressions, darśana and grahaṇa hint at perception and direct cognition (vijñāna), two terms denoting immediate sensory awareness of an object.82 According to Dharmakīrti, the nature of an object is undivided and amenable to sense perception.83 This is tantamount to claiming that a single act of perception is enough to grasp this nature, and that it grasps it in its entirety (sarvātmanā), in all its aspects (sarvākāreṇa), so that no other means of valid cognition is needed for cognizing this nature in a positive way (vidhinā): Perception leaves no part of this undivided nature unknown, so that, say, inference or verbal knowledge might be needed in order to gain access to it.84 In other words, a single perception grasps an object as selfless and momentary, or, to be more precise, grasps a selfless and momentary thing.85 This can,
On the corrective function of inference, see Kellner 2004: 4–9. See Eltschinger 2009: 41–42, n. 6. 82 AK 1.16a: vijñānaṃ prativijñaptiḥ; AKBh 11,7: viṣayaṃ viṣayaṃ prati vijñaptir upalabdhir vijñānaskandha ity ucyate /. AKVy 38,24: upalabdhir vastumātragrahaṇam /. 83 PV 1.43: ekasyārthasvabhāvasya pratyakṣasya sataḥ svayam /; PVSV 26,4: eko hy arthātmā / sa pratyakṣaḥ … 84 PV 1.45: dṛṣṭasya bhāvasya dṛṣṭa evākhilo guṇaḥ /; PVSV 26,5–6: tasya pratyakṣeṇaiva siddheḥ sarvākārasiddheḥ / tadanyasyāsiddhasyābhāvāt /; PVSV 26,9–11: tasmāt pratyakṣe dharmiṇi tatsvabhāvasākalyaparicchedāt tatrānavakāśā pramāṇāntaravṛttiḥ syāt /; PVSVṬ 121,17–18: pratyakṣadṛṣṭāt svabhāvāt ko ’nyaḥ /. Through perception, bare particulars are grasped in their entire true nature (dṛṣṭasarvatattva PVSV 26,14). 85 PVSV 43,8–11: nāpi svalakṣaṇasyānityatvādyabhāvaḥ / yasmān nā81 80

Ignorance, epistemology and soteriology (2)

51

of course, be traced back to Dharmakīrtiʼs “Sautrāntika” assumption that a perceptual awareness results directly from a real thing’s causal efficiency. According to a well-known statement, “experts in reason(ing) hold that [for a given thing] to be a graspable [object] consists in being a cause capable of casting (arpaṇa) [its own] aspect into cognition.”86 That real things cast their own aspect into the consciousness, thus giving rise to perceptual awareness, is the basic meaning of the description of this awareness as arising by the force of something real (vastubalapravṛtta). Dharmakīrti makes it especially clear in the following statement: “The property of a [perceptual] cognition is to grasp an object; [as for] this [object, it] is grasped as it is, and it generates this [cognition of itself] through [its truly] existing nature. Such is the nature [of the cognition and of the object].”87 Devendrabuddhi (as well as Śākyabuddhi and Kamalaśīla) exhibits the rationale behind Dharmakīrtiʼs (provisionityatvaṃ nāma kiṃcid anyac calād vastunaḥ / kṣaṇapratyupasthānadharmatayā tasya tathābhūtasya grahaṇād etad evaṃ bhavaty anityo ’yam anityatvam asyeti vā /. “Neither does the bare particular lack impermanence, etc., for what we call ‘impermanence’ is nothing other than the transient entity [itself. But] this is so because [those who see the last phase* of a continuum] grasp such an [entity] as having the property of being present [during only one] phase, [and thus say, ascribing properties]: ‘This is impermanent,’ or: ‘This has impermanence’.” *The last phase (antyakṣaṇa) is defined in the following way by Śākyabuddhi and Karṇakagomin (PVṬ Je D48a7/P56b8 = PVSVṬ 95,30): sadṛśakṣaṇāntarāpratisandhāyī kṣaṇo ’ntyakṣaṇaḥ … “The last phase [of an entity] is the phase which is not connected with a new (antara) similar phase.” According to PVSVṬ 184,5–6, PVSV 43,8–11 answers the objection formulated in PVSV 42,11–12: svalakṣaṇe cānityatvādyapratīter atādrūpyam / teṣāṃ cāvastudharmatā /. “And since one does not cognize impermanence, etc., in the bare particular, [the bare particular] does not have this nature[, viz. impermanence, etc.], and [hence imper manence, etc.] are not properties of [real] entities.” Note also PVSV 21,4–6: sa eva hi bhāvaḥ kṣaṇasthitidharmā ’nityatā vacanabhede ’pi dharmidharmatayā nimittaṃ vakṣyāmaḥ /. 86 PV 3.247b2d: grāhyatāṃ viduḥ / hetutvam eva yuktijñā jñānākārārpaṇakṣamam //. See Hattori 1968: 53. 87 PV 2.206–207a1: viṣayagrahaṇaṃ dharmo vijñānasya yathāsti saḥ / gṛhyate so ’sya janako vidyamānātmaneti ca // eṣā prakṛtiḥ … See also below, §2.2.6 and n. 136.

52

Vincent Eltschinger

nal) position as follows:88 “When he is asked about the property of a cognition, the one who accepts that a cognition really grasps an object must answer that the property of a [perceptual] cognition is to grasp an object (= PV 2.206ab1). [And] if the property of all the cognitions possessing an object is to grasp an object, then they grasp [their] objects as they [really] are, (…) under an aspect such as impermanence, not under an unreal aspect. For in this way, if it is rationally established that a cognition cognizes (viṣayīkaroti) an object as it [really] is, that which is not cognized in this way is due to an external89 or internal90 adventitious cause of error, just as the [erroneous] cognition of a snake in the case of a rope in a dark place abundant in/suitable for snakes. Therefore, to grasp the real aspect of an object is the nature of a cognition. If on the contrary (atha ca) [its] nature were to grasp [an object] erroneously, then it would not have the property of grasping any object [at all]. Because in this way the object would not be as the cognition cognizes [it], and because [the cognition] would not cognize the object as it [really] is, cognitions would be devoid of object, (…) [and] hence all entities would be unestablished (…) Therefore, the one who accepts a relationship between object and object-possessor has to hold that the property of a cognition is to grasp an object, [and] thus the nature of this [cognition] is to grasp the real aspect of an object. That which is other than this [i.e., unreal,] is produced by a [purely] adventitious condition.”91 This argument draws a sharp delineation
In an introductory statement, Śākyabuddhi reminds his audience that the following argument does not match Dharmakīrti’s final, Yogācāra position in epistemological matters. PVṬ D133b2–3/P164b3–5: don dam par rnam par śes pa ni don ’dzin par ’dod pas źes bya ba la / don dam par rnam par śes pa don ’dzin pa ñid ni ma yin te / gzuṅ ba ma grub pa’i phyir ro // ’on kyaṅ re źig phyi rol gyi don yod par ’dod pa gaṅ yin pa des ’di ltar ’dod par bya’o źes bstan pa’i phyir de skad du brjod pa yin no //. 89 PVṬ D133b3–4/P164b5: phyi rol lam źes bya ba ni ’dra ba gźan daṅ gźan ’byuṅ ba la sogs pa’i ’khrul par byed pa’i rnam pa’o //. See below, nn. 116 and 139. 90 Tib. cig śos = Skt. itara, lit. “other [than external].” 91 PVP D87b5–88a4/P101a2–b3: rnam par śes pa’i chos kyaṅ gaṅ źe na / źes dris pa na don dam par rnam par śes pa ni don ’dzin par ’dod pas rnam śes yul ’dzin pa’i chos śes brjod par bya’o // gaṅ gi tshe rnam par śes pa yul
88

Ignorance, epistemology and soteriology (2)

53

between non-erroneous cognitions, which result directly from their objects’ causal efficiency, and erroneous cognitions, which result from a cause of error (bhrāntinimitta, pratyaya). Whereas the former are termed vastubalapravṛtta, true (bhūtārtha), and (being the mind’s) nature, the latter, which arise, among other factors, from the latent tendencies of erroneous conceptual constructs,92 are decan du gyur pa thams cad kyi chos yul ’dzin pa yin pa de’i tshe / mi rtag (D rtag: P rtag rtag) pa la sogs pa’i rnam pa gaṅ gis … yul yod pa de bźin du ’dzin ’gyur gyi med pa’i rnam pas ni ma yin no // de de ltar na śes pa don ji lta ba bźin du yul du byed par rigs pas thob pa na / de ltar na (D na: P om. na) rtogs pa ma yin pa gaṅ yin pa de ni phyi rol lam cig śos glo bur ba’i ’khrul pa’i rgyu mtshan gyis yin te / dper na sbrul du ’dris pa’i phyogs mi gsal bar thag pa la sbrul gyi (D gyi: P mi) śes pa lta bu’o* // de bas na yul gyi rnam pa yod pa ’dzin pa gaṅ yin pa de ni sems kyi raṅ bźin no // ci ste yaṅ log par ’dzin pa ñid raṅ bźin yin pa de’i tshe yul ’dzin pa’i chos ma yin no // de ltar na ji ltar śes pas yul du byed pa de ltar don de ma yin źiṅ ji ltar don de yin pa de ltar yul du byed pa ma yin pa’i phyir / śes pa dag yul med pa can du ’gyur bas … de ltar na dṅos po thams cad ma grub pa yin te … de bas na yul daṅ yul can gyi dṅos po ’dod pa ñid kyis rnam par śes pa’i chos yul ’dzin pa yin par brjod par bya’o // de ltar na ’di’i raṅ bźin ni yaṅ dag pa’i yul gyi rnam pa ’dzin pa yin no // de rnam pa gźan du ’gyur ba gaṅ yin pa de ni glo bur gyi rkyen gyis byas pa ñid yin no //. * Vibh. 82 n. 4: mandamandaprakāśe sarpopacite pradeśe /. Note also TSP Ś1056,21–1057,5/K872,27–873,7: tathā hi – viṣayaviṣayibhāvam icchatā cittaṃ viṣayagrahaṇasvabhāvam abhyupeyam, anyathā viṣayajñānayor na viṣayaviṣayibhāvaḥ / arthagrahaṇasvabhāvatvenāṅgīkriyamāṇe yas tasya svabhāvas tenaivātmano ’ṃśo ’rthas tena gṛhyata iti vaktavyam / anyathā katham asau gṛhītaḥ syāt / yady asatākāreṇa gṛhyeta tataś ca viṣayaviṣayibhāvo na syāt / tathā hi – yathā jñānaṃ viṣayīkaroty arthaṃ na tathā so ’rthaḥ, yathā so ’rtho na tathā taṃ viṣayīkarotīti nirviṣayāṇy eva jñānāni syuḥ / tataś ca sarvapadārthāsiddhiprasaṅgaḥ / tas­ mād bhūtaviṣayākāragrāhitā ’sya svabhāvo nija iti sthitam / bhūtaś ca svabhāvo viṣayasya kṣaṇikānātmādirūpa iti pratipāditam etat / tena nair­ ātmyagrahaṇasvabhāvam eva cittaṃ* nātmagrahaṇasvabhāvam /. * TSPK reads eveti tan against TSPŚ and TSPTib eva cittaṃ; both the Jaisalmer ms and the Pāṭan ms read eve ! cittaṃ. On this passage of the TSP, see McClintock 2010: 213–214. 92 PVṬ D133b4/P164b5–6: cig śos źes bya ba ni naṅ gi bdag ñid can gyi phyin ci log gi rnam par rtog pa’i bag chags źes bya bas bslad pa’o //. In an etymologizing vein, Devendrabuddhi explains āgantuka as follows (PVP D89a5/P103a2): rkyen gźan gyi rgyu mtshan las ’oṅs pa ñid yin pa’i phyir

54

Vincent Eltschinger

scribed as avastubalapravṛtta, as not agreeing with (means of) valid cognition (pramāṇāsaṃvādin) and as adventitious (āgantu[ka]). According to Dharmakīrti’s followers, this delineation only holds good provided perceptual cognitions cognize their objects in their real aspects. Claiming that a perceptual cognition grasps the real aspect of an object93 is tantamount to saying that it grasps aspects such as impermanence or selflessness.94 As Kamalaśīla nicely puts it, “it is firmly established that the intrinsic nature of the [mind] is to grasp the real aspect of an object; but it has been explained [earlier] that the real nature of an object consists of [its being] momentary, selfless, etc.; therefore, the mind has the grasping of selflessness for its nature.”95 In other words, the nature of the mind is to perceive reality/the true nature (tattvadarśana) of things.96 And granted that selflessness is the true nature of things, the mind turns out to be nothing other than discernment (vipaśyanā) itself,97 which Śākyabuddhi defines as wisdom (prajñā) bearing upon selfless(*pratyayāntaranimittād āgatatvāt). Erroneous cognitions and defilements are due to raṅ daṅ rigs mthun pa’i ñe bar len pa’i rgyu (PVP D89a5– 6/P103a2–3; *svasamānajātīyopādānakāraṇa; note PVṬ Je D251b6/P299a4–5 = PVSVṬ 400,30–431,9: upādānabalabhāvīti vitathavikalpavāsanābalabhāvi). 93 PVP D87b7/P101a6: yul gyi rnam pa yod pa …; PVP D88a3/P101b2 = PVP 89a1/P102b3: yaṅ dag pa’i yul gyi rnam pa … 94 PVP D88b3–4/P102a3–4: mi rtag pa la sogs pa’i rnam pa yod pa’i yul …; PVP D87b6/P101a4 = PVP D90a4/P104a4: mi rtag pa la sogs pa’i rnam pa …; PVP D89a6/P103a3: bdag med pa …; PVP D89b3/P103a8: bdag med pa ñid … 95 TSP Ś1057,2–5/K873,5–7: bhūtaviṣayākāragrāhitā ’sya svabhāvo nija iti sthitam / bhūtaś ca svabhāvo viṣayasya kṣaṇikānātmādirūpa iti pratipāditam etat / tena nairātmyagrahaṇasvabhāvam eva cittam … For the context of this statement, see above, n. 91. 96 PVP D87a7/P100b3: sems kyi raṅ bźin ni de kho na ñid mthoṅ ba’i bdag ñid can yin … (PVṬ D133a3–4/P164a2–3: de kho na ñid mthoṅ ba’i bdag ñid can yin gyi źes bya ba ni dṅos po ji lta ba bźin du gnas pa’i ’dzin pa’i* bdag ñid can źes bya ba’i don to) *Cf. PVV 82,14: yathāvasthitavastugrahaṇam; PVP D89b1/P103a6: sems ni ṅo bo ñid kyis de kho na ñid mthoṅ ba’i bdag ñid can yin … 97 PVP D90a1/P103b8: raṅ bźin yaṅ lhag mthoṅ yin …

Ignorance, epistemology and soteriology (2)

55

ness.98 This “Sautrāntika” epistemology forms the background of Dharmakīrti’s well-known allusion to the canonical topos of the mind’s being radiant (prabhāsvara) by its very nature (prakṛtyā). “Radiant” is to be understood as “having the nature of grasping [entities] as they really are” (yathābhūtagrahaṇasvabhāva), or “consisting in the perception of reality/the true nature [of things]” (tattvadarśanasātmaka).99 “Knowledge” is nothing but direct perceptual awareness, i.e., the mirror-like mind grasping the true nature of real entities.100 What can be regained from Dharmakīrtiʼs understanding of “knowledge” seems to mirror a significant shift from the ideas held by his Yogācāra predecessors. Defining a threefold ignorance, the YBh declares its antidotes (vipakṣa) to be the insights born of audition, reflection and (mental) cultivation.101 In his PrSVy, Vasubandhu defines “knowledge” as the insight born of reflection and (mental) cultivation.102 Dharmakīrti assents, of course, to the fact that ignorance can only be eliminated by the practice of the path and its three (or at least two) successive types of insight. But according to him, soteric practice does not aim at developing entirely new cognitive modalities, but rather, at freeing from all counteracting factors a type of cognition that has already been here at hand.
98 PVṬ D134b3/P166a1: lhag mthoṅ ba yin la źes bya ba bdag med pa la (D la: P la bya ba) dmigs pa’i śes rab bo. Discernment is described in BhK 1.219,23–220,4 as sarvadharmaniḥsvabhāvatālambana, and defined in BhK 3.5,17–20 as follows: bhūtapratyavekṣaṇā ca vipaśyanocyate / bhūtaṃ punaḥ pudgaladharmanairātmyam / tatra pudgalanairātmyaṃ yā skandhānām ātmātmīyarahitatā / dharmanairātmyaṃ yā teṣām eva māyopamatā /. For a French translation, see Lamotte 1987: 340. On vipaśyanā/prajñā, see Eltschinger 2009: 57–58 (§1.2.5) and nn. 26–27. 99 PVP D89a5/P103a1: ’od gsal te / yaṅ dag pa ji lta ba bźin du ’dzin pa’i raṅ bźin yin no //; TS Ś3434ac1/K3435ac1: prabhāsvaram idaṃ cittaṃ tattvadarśanasātmakam / prakṛtyaiva sthitam … 100 On this point, see Eltschinger 2005: 190–192. 101 YBh 206,6–7: śrutamayyāś cintāmayyā bhāvanāmayyāś ca prajñāyā vipakṣeṇa trayaḥ paryāyā yathākramaṃ yojyante /. 102 PrSVy 9a1: bsams pa daṅ bsgoms pa las byuṅ ba’i śes rab ni rig pa źes bya’o //.

56

Vincent Eltschinger

2.2.3. Contrary to “knowledge,” which, qua perception, is a cognition that is free of conceptual construction (kalpanāpoḍha) and non-erroneous (abhrānta),103 the realm of ignorance is coextensive with conceptuality and error. “Error,” however, is not necessarily synonymous with “unreliability” (visaṃvāda, visaṃvāditva): Whereas “erroneous” is to be said of any cognition that does not arise from and hence display a bare particular, “unreliable” denotes those cognitions that are not conducive to a successful practical interaction with the particulars (or, as Dharmottara will say, that do not allow one to reach/obtain [pra√āp] the concrete particular).104 All conceptual constructs are erroneous by their very nature and origin, but some of them are reliable (and hence valid cognitions, pramāṇa),105 whereas others are not. Śākyabuddhi and Karṇakagomin have an opponent ask the following question: “[But] if every conceptual construct is simply erroneous, why [do you hold] conceptual constructs such as [being] impermanent or selfless [to be] valid cognitions, but not conceptual constructs such as [being]

PVin 1.4ab1 ≈ NB 1.4: pratyakṣaṃ kalpanāpoḍham abhrāntam … On kalpanāpoḍha, see Funayama 1992; on abhrānta, see Funayama 1999. 104 See Krasser 1995. Note also TSP Ś479,23–24/K392,7: avisaṃvāditvaṃ cābhimatārthakriyāsamarthārthaprāpaṇaśaktiḥ /. “‘Being non-deceptive’ means the efficacy to realize the attainment of the object which is appropriate for the fulfilment of a desired purpose.” Translation Funayama 1999: 79. On the differences between Dharmottara’s and Kamalaśīla’s interpretations of abhrānta, see Funayama 1999: 80–81. 105 PV 2.5a: vyavahāreṇa prāmāṇyam … “Epistemic validity [is known] through practical activity.” Most important in this connection is the case of inference. PVin 2 46,5–8 (including PVin 2.1cd): tad etad atasmiṃs tadgrahād bhrāntir api sambandhataḥ pramā // svapratibhāse ’narthe ’rthādhyavasāyena pravartanād bhrāntir apy arthasambandhena tadavyabhicārāt pramāṇam /. “Die (Schlußfolgerungserkenntnis) ist wegen der Verbindung [mit dem Gegenstand] eine gültige Erkenntnis (pramā), obgleich sie wegen des Erfassens von etwas als etwas, was es nicht ist, Irrtum ist. (Das heißt:) Obwohl sie Irrtum ist, weil sie in der Weise auftritt, daß sie ihr eigenes Erkenntnisbild, das nicht der (wirkliche) Gegenstand ist, als [diesen Gegenstand] bestimmt, ist sie als mit dem Gegenstand verbundene (dennoch) gültige Erkenntnis.” Translation Steinkellner 1979: 26–27. See also PV 3.55–63.

103

Ignorance, epistemology and soteriology (2)

57

permanent?”106 Dharmakīrtiʼs answer is as follows: “And since all this is an error due to the latent tendencies imprinted by [previous] perceptions of the particulars themselves, [those] conceptual constructs whose arising is [indirectly] bound to these [particulars] are reliable with regard to the thing [itself] although they do not display it, just like the error [consisting] of [cognizing] a gem [is reliable] with regard to the radiance of that gem. [But] others [such as permanence] are not [reliable with regard to the thing itself] because, (…) disregarding (parityajya) the conformity107 with the specific [property] as it has been perceived, they superimpose another[, erroneous] specific [property] by [arbitrarily] grasping any sort of universal (kiṃcitsāmānya). [These conceptual constructs are as unreliable with regard to the thing itself] as the notion of a gem [is unreliable] with regard to the radiance of a lamp.”108 Inasmuch as they do not display bare particulars and owe their existence to latent tendencies, all conceptual constructs are error. Some of them, however, are valid cognitions: Because the aspect they ascribe to the thing exists in it,109 and because they are indirectly related (pratibaddha) to the bare particular, they are reliable with regard to the thing itself, i.e., allow a successful practical interaction with it.110 Other conceptual constructs are not valid cognitions: because
PVṬ Je D95b3–4/P112a8–b1 ≈ PVSVṬ 183,9–10: yadi mithyārtha eva sarvo vikalpaḥ kasmāt … anityānātmādivikalpāḥ pramāṇaṃ nityā[di] vikalpās tu neti … 107 PVṬ Je D96a3/P113a1 = PVSVṬ 183,23–24: anusaraṇaṃ niścayaṃ pa­ rityajya … 108 PVSV 43,2–7: sarvaś cāyaṃ svalakṣaṇānām eva darśanāhitavāsanākṛto viplava iti tatpratibaddhajanmanāṃ vikalpānām atatpratibhāsitve ’pi vastuny avisaṃvādo maṇiprabhāyām iva maṇibhrānteḥ / nānyeṣām / … yathādṛṣṭaviśeṣānusaraṇaṃ parityajya kiṃcitsāmānyagrahaṇena viśeṣāntarasamāropād dī paprabhāyām iva maṇibuddheḥ /. On maṇibhrānti, see Krasser 1991: 65–66n. 121. 109 PVṬ Je D95b6–7/P112b4–5 ≈ PVSVṬ 183,16–17: anityādirūpasya vastuni vidyamānatvāt … 110 PVin 2 48,1–5 (together with PVin 2.7a): ata eva prāmāṇyaṃ vastuviṣayaṃ dvayoḥ pratyakṣānumānayoḥ, arthakriyāyogyaviṣayatvād vicārasya / sukhaduḥkhasādhane jñātvā yathārhaṃ pratipitsavo hi kiṃcit parīkṣante prekṣāpūrvakāriṇaḥ, na vyasanitayā /. “Eben daher bezieht sich
106

58

Vincent Eltschinger

they superimpose an aspect that is not found in the thing itself,111 and because what they ascribe to it is not even indirectly related to it,112 they are unreliable with regard to the thing itself, i.e., are deceiving in practice. 2.2.4. In our philosophersʼ linguistic usage, however, “error” (bhrānti) quite often occurs as a shorter term for “unreliable cognition,” and is equated with “wrong notion” or “misconception” (viparyāsa). A similar semantic shift can be observed in connection with “superimposition” ([sam]āropa), no longer used in the general sense of conceptuality and concealment, but in the sense of a mistaken identification barring determinate cognition (niścaya). In the present context, “error,” “superimposition” (both in this specialized meaning), “wrong notion,” and “lack of determinate cognition” can be considered to be equivalent. Two kinds of situation are responsible for the rise of error: the presence of a cause of error (bhrāntinimitta)113 and the lack of the causal conditions needed for determinate cognition (niścayapratyayavaikalya).114 Together with Śākyabuddhi, we may consider the cause of error as twofold: The internal cause of error consists in the latent tendency of a contrary conceptual construct (viparītavikalpavāsanā);115 as for the external cause of error, it is most often exemplified as the arising of ever new similar phases (sadṛśāparāparotpatti) in a continuum,116
die Gültigkeit der beiden, Wahrnehmung und Schlußfolgerung, auf das Wirkliche, denn eine prüfende Erkenntnis hat ein Objekt, das fähig ist einen Zweck zu erfüllen. Vernünftig handelnde Leute, die (auch nur) ein wenig abwägen, (tun dies), wenn sie die Mittel für Lust und Leid (einmal) erkannt haben, aus der Absicht, [diese] nach Vermögen zu erreichen, aber nicht aus [bloßer] Neigung.” Translation Steinkellner 1979: 29 (slightly modified). 111 PVṬ Je D96a1/P112b7 ≈ PVSVṬ 183,20–21: teṣāṃ [= nityādivikalpānām] … vastuny avidyamānasyaivākārasya samāropāt /. 112 PVṬ Je D96a4/P113a2 ≈ PVSVṬ 183,26–27: pāramparyeṇāpi … apratibaddhatvāt /. 113 PV 1.44a, PVSV 26,15, and passim. 114 PVSV 26,19. 115 See above, n. 92. 116 PVSV 26,20–21; sadṛśāparotpatti at PVṬ Je D61a3/P72a2 = PVSVṬ 122,10–11, PVṬ Je D61a5/P72a5 = PVSVṬ 123,8–9; note also PVṬ Je

Ignorance, epistemology and soteriology (2)

59

D61b4/P72b5 = PVSVṬ 123,27–28: sadṛśasya dvitīyasya kṣaṇasyotpattyā bhrāntinimittena … See also above, n. 89. Locus classicus for sadṛśāparotpatti is PVSV 21,6–9: tāṃ punar asya kṣaṇasthitidharmatāṃ svabhāvaṃ svahetor eva tathotpatteḥ paśyann api mandabuddhiḥ sattopalambhena sarvadā tathābhāvaśaṅkāvipralabdho na vyavasyati sadṛśāparotpattivipralabdho vā /. Translated according to Śākyabuddhi’s explanation (PVṬ Je D46b2– 47a1/P54b6–55a6): “However, although (s)he experiences this property of lasting [only] one phase[, a property which is] the nature of the [entity] since [this entity] is produced such [i.e., momentary,] by its own cause, a [person] of weak intellect fails to determine [it in the same way as (s)he has just experienced it; this failure occurs] either [because this person,] due to having perceived the existence [of this entity at one phase, is] mistaken by the supposition that it permanently (sarvadā) exists in this [very] way, or [because this person is] mistaken by the rise of a new (apara) phase similar [to the former one].” According to Śākyabuddhi’s interpretation (PVṬ Je D47b6–48b1/ P56a6–57a1), the first cause of error (*vipralambhanimitta) is proper to the outsiders (tīrthika) professing the doctrine of non-momentariness (akṣaṇikavāda), and points to their inter nal *kudṛṣṭyabhiniveśavāsanābīja (or else: *anādikudṛṣṭyabhiniveśabīja), which is reinforced by the false views propagated by wrong treatises (*kuśāstradṛṣṭi). As for the second cause of error, it is aimed at explaining why the Buddhists, who follow sound reasoning and scripture (yuktyāgama) professing momentariness, still do not ascertain momentariness upon perceiving the real entity. Karṇakagomin’s explanation (PVSVṬ 91,23) of mandabuddhi is worth noticing: anādisaṃsārābhyastayā nityādirūpāvidyāvāsanayā mandā buddhir yasya … “Whose intellect is [made] weak by the latent tendency, repeated [and reinforced] in the beginningless saṃsāra, of ignorance in the form of [mistaken aspects] such as ‘permanent’.” This ignorance (or rather, its latent tendency) being the internal cause of error, the two causes mentioned by Dharmakīrti point to external causes of error (bāhyam api bhrāntibījam, PVSVṬ 91,27). Note also PVSV 100,4–7 = PVin 2 82,7–9: tam asya mandāḥ svabhāvam ūrdhvaṃ vyavasyanti / na prāk / darśane ’pi pāṭavābhāvād iti tadvaśena paścād vyavasthāpyate / vikāradarśaneneva viṣam ajñaiḥ /. “Weak[-minded people] identify this [transient] nature of the [entity only] later [i.e., at the time of the interruption of the continuum, but] not before [i.e., at the time of the existence of the entity], because even though they [directly] experience [this nature], they lack [intellectual] sharpness. Therefore, [this transient nature] is ascertained [only] later on account of this [determination], just as ignorant [persons identify a poisonous substance that they have seen only] by experiencing a [morbid] affection [such as over-salivation].” See also Steinkellner 1979: 98. Note Karṇakagomin’s explanation of mandāḥ in PVSVṬ 366,27: ā saṃsāram avidyānubandhān mandāḥ … This explanation is borrowed from Dharmottara’s PVinṬ Dze D249b5/P301b3–4: ’khor ba ji srid par ma rig

60

Vincent Eltschinger

which leads to the superimposition of aspects such as permanent (nitya), enduring (sthira), and non-momentary (akṣaṇika).117 Be it internal or external, this cause of error impedes determinate cognition (niścayapratirodhin, °vibandhaka).118 The lack of (conceptual) habitus (abhyāsa) is most often quoted as being among the conditions that, when lacking, prevent determinate cognition from arising.119 Just as determinate cognition bears upon one specificity (bheda) or aspect (ākāra) of a previously cognized particular, wrong notion superimposes one partial erroneous/contrary aspect (aṃśasamāropa)120 and associates (< samyojyeta, PV 1.44b) another, i.e., a false quality (guṇa; glossed as rūpa, dharma),121 to the thing. As Dharmakīrti himself has it, “though it has been perceived as distinct from all [other entities], an entity is not [necessarily] recognized in this way [i.e., in all its aspects], because an obstruction (vyavadhāna) to [the recognition of] a certain specificity [such as momentariness] may occur.”122 Determinate cognition (niścaya, °jñāna, °manas) and superimposition (samāropa, °jñāna; āropamanas) are mutually exclusive and stand in a relationship of mupa daṅ rjes su ’brel pa źan pa … Note also the various inter pretations of the fact that the determinate cognition arises only at the time of pravāhaviccheda: (1) PVṬ Je D227a3–4/P263b7–8: mthoṅ ba’i dus su ṅes pa yod pa ma yin te / ma rig pa’i mun pa ñid kyi phyir daṅ gźan rgyun ’dra ba skye ba’i phyir ro // mthoṅ ba gsal ba med pa’i phyir ro //. (2) PVSVṬ 366,28–29: na darśanakāle ’dhyavasāyo ’sti / avidyā(sāma)rthyāt sadṛśāparotpattyā ca darśanapāṭavasyābhāvāt /. (3) PVinṬ Dze D249b6–7/P301b5: ma rig pa daṅ ldan pa’i źan pa rnams la mthoṅ ba gsal ba med pa’i phyir ro //. Here again, both Śākyabuddhi and Karṇakagomin suggest that the absence of niścaya proceeds from an internal (ignorance) and an external (the rise of a new similar phase) cause. 117 See also above, §2.1.1. and n. 11. 118 PVSV 26,14, PVṬ Je D61a5/P72a5 = PVSVṬ 123,8. 119 PVṬ Je D61b2–3/P72b3 = PVSVṬ 123,21. On niścayapratyayas, see Kellner 2004: 19–32. 120 PV 1.50a; PVSV 27,22–28,1: ākārasamāropa; PVṬ Je D62b3/P73b6 = PVSVṬ 125,28–29: tadviparītākārasamāropī viparyāsaḥ. 121 PV 1.44b, PVṬ Je D64b7/P76a8–b1 = PVSVṬ 131,11, PVV 306,6. 122 PVSV 28,13–14 (leaving hi untranslated): na hi sarvato bhinno dṛṣṭo ’pi bhāvas tathaiva pratyabhijñāyate / kvacid bhede vyavadhānasambhavāt /.

Ignorance, epistemology and soteriology (2)

61

tual annulment (bādhyabādhakabhāva):123 When aspects such as lasting, endowed with a self (sātmaka), or unconditioned (akṛtaka) are superimposed, (real) contrary aspects such as impermanent/ momentary, selfless (nirātmaka), or conditioned (kṛtaka) are not made the objects of determinate cognitions.124 2.2.5. According to Dharmakīrti, the function (vyāpāra) and aim (phala, artha) of inference125 (anumāna, liṅga, sādhana) as a means of valid cognition is not to cognize something in a positive way (vidhinā) or to determine the nature of an entity (vastusvabhāvaniścaya),126 but to rule out, negate, or exclude (vyavaccheda, niṣedha, pratiṣedha, nivṛtti, apoha) unreliable superimpositions and wrong

123 PV 1.49ab: niścayāropamanasor bādhyabādhakabhāvataḥ /; PVSV 28,16–17: samāropaniścayayor bādhyabādhakabhāvāt /. 124 niścitākāras: kṛtakatva (PVSVṬ 124,26 and 125,23–24), anityatva (PVSV 26,5), kṣaṇikatvādi (PVSVṬ 130,28), kṣaṇikatvānātmādi (PVSVṬ 124,12), asthira (PVSVṬ 129,28), nirātmaka (PVSVṬ 129,28); samāropitākāras: sthira (PVSV 28,11, PVSVṬ 122,12), sātmaka (PVSV 28,11), sthiti (PVSV 26,21), akṛtaka (PVSVṬ 125,23–24), nityādi (PVSVṬ 124,13 and 125,24). 125 Note should be made that inference is itself strictly of a conceptual nature, and as such is basically on the side of error and ignorance. An inference indeed mobilizes two properties (dharma, a probans [sādhanadharma, hetu, liṅga] and a probandum [sādhyadharma]) that are thought to belong to a single property possessor (dharmin, or “subject”). Both of these two properties are universals (sāmānya) unduly ascribing a single unitary aspect to the many. At the same time, these two different properties are tied to one and the same subject, thus unduly dividing the indivisible. To unify the many (the seed of the use of universals*) and divide the undivided (the seed of co-reference [sāmānādhikaraṇya]**) are indeed the two main psychological operations giving rise to conceptual constructs. *According to PVṬ Je D101a6/P119a4 and D101a7/P119a5: spyi’i tha sñad kyi sa bon … (sāmānyavyavahārabīja); the psychological genesis of universals is presented in a nutshell in PV 1.82. **According to PVṬ Je D101b4/P119b3: gźi mthun pa ñid … [kyi] sa bon (sāmānādhikaraṇyabīja); the psychological genesis of co-reference is presented in a nutshell in PV 1.83. See also Eltschinger 2009: 59–62 (§1.2.10). 126 Resp. PVSV 27,10 and PVSV 28,20.

62

Vincent Eltschinger

notions:127 “Superimpositions endowed each with its own cause are as many as the alien natures (parabhāva) [wrongly ascribed] to the [entity]. In that they exclude these [superimpositions], the means of valid cognition [named ‘inferences’] can therefore be useful. But these [inferences,] aiming (°phala) [as they do] at the exclusion [of superimpositions,] are not employed in order to cognize a [supposedly still] uncognized part of the entity, because this [part has already been] perceived, and because an indivisible [entity] cannot be perceived in a partial way (ekadeśena).”128 Dharmakīrti spells out the same argument in the following three stanzas: “[If] the undivided (eka) nature of an object is in itself perceptible, which other unperceived part [of it] would there be left for [further positive] investigation by the [other] means of valid cognition [i.e., by inference]? [There would be none,] if another [unreal] quality were not associated [with this nature] due to [some] cause of error, just like the aspect of silver [is associated] with a conch-shell due to oneʼs observing a similarity of colour [between them]. Therefore, all the qualities of the perceived entity are perceived, [but] due to some error, they are not determined. Thus one undertakes an [inferential] proof [in order to determine what the error has left undetermined].”129 To be more precise, inferences, like conceptual constructs and words, perform both a direct, positive (< vidhinā, vidhirūpeṇa)

vyavacchedaphala (PVSV 26,24); samāropavyavaccheda (PVSV 27,13; 27,14); vyavacchedakṛt (PVSV 27,10); anyavyavaccheda (PVSV 27,14); vyavacchedaviṣaya (PVSV 28,9; PV 1.56a); anyavyavacchedaviṣaya (PVSVṬ 127,10); anyasamāropavyavacchedaphala (PVSV 31,12–13); samāropapratiṣedhaphala (PVSVṬ 124,16); bhrāntinivṛttyartham (PVSV 31,12); apohagocara (PV 1.48d; PVSV 28,19); apohaviṣaya (PV 1.47a); an­ yāpohaviṣaya (PVSV 31,13). See Kellner 2004: 4–9. 128 PVSV 26,22–27,2: yāvanto ’sya parabhāvās tāvanta eva yathāsvaṃ nimittabhāvinaḥ samāropā iti tadvyavacchedakāni bhavanti pramāṇāni saphalāni syuḥ / teṣāṃ tu vyavacchedaphalānāṃ nāpratītavastvaṃśapratyāyane pravṛttis tasya dṛṣṭatvāt / anaṃśasya caikadeśena darśanāyogāt /. 129 PV 1.43–45: ekasyārthasvabhāvasya pratyakṣasya sataḥ svayam / ko ’nyo na dṛṣṭo bhāgaḥ syād yaḥ pramāṇaiḥ parīkṣyate // no ced bhrāntinimittena saṃyojyeta guṇāntaram / śuktau vā rajatākāro rūpasādharmyadarśanāt // tasmād dṛṣṭasya bhāvasya dṛṣṭa evākhilo guṇaḥ / bhrānter niścīyate neti sādhanaṃ sampravartate //.

127

Ignorance, epistemology and soteriology (2)

63

and an indirect (< arthāt), negative function.130 In its positive function, inference aims at the conceptual determination of those aspects of the perceived particular that have escaped determination (aniścitaniścaya).131 But inference ipso facto negates the conceptual constructs wrongly ascribed to the perceived entity, and such is its indirect function. In this respect, inference does not differ from words and concepts, which refer simultaneously to positive intellectual constructs and indirectly exclude other, unfitting constructs. It is hardly surprising, then, that Dharmakīrti repeatedly describes inference, too, as the exclusion of another (anyāpoha): Inference aims at determination, but to determine amounts to holding off superimposition (samāropaviveka), i.e., to excluding another, superimposed aspect. That inference always presupposes a wrong notion is the point at stake in the following discussion: “[Objection:] The [inferential] determination of [something previously] uncognized does not necessarily presuppose a wrong notion, as [in the case of one] suddenly (akasmāt) knowing from [the presence of] smoke [that there is] fire [in a certain place], for in this case, the [previous] superimposition of the absence of fire (an­ agni) [in this place] is not possible. Therefore, [inference] does not always (sarvatra) exclude [a previous superimposition]. [Answer:] (…) In this case too, the [person] who sees this [spot] lacks a determinate cognition of its nature [i.e., of this spotʼs indeed possess-

This is made especially clear by Karṇakagomin, who regularly (e.g., PVSVṬ 124,14, 124,22, 124,24, 125,14, 125,15, 125,21, 126,9) adds vidhinā/ vidhirūpeṇa after words denoting niścaya or adhyavasāya, and arthāt after words denoting vyavaccheda, etc. Interestingly enough, close comparison with the PVṬ reveals that this is never done by Śākyabuddhi. Commenting on PV 1.45d (sādhanaṃ sampravartate), Karṇakagomin (PVSVṬ 124,21– 22) says: tanniścayārthaṃ sādhanam anumānaṃ vidhirūpeṇaiva pravartate …, whereas Śākyabuddhi (PVṬ Je D62a2/P73a4) has: sgrub pa źes bya ba ’khrul pa sel bar byed pa’i rjes su dpag pa rab tu ’jug pa yin /. For a similar observation, see Kellner 2004: 5n. 3. 131 Note, e.g., PVSVṬ 184,8–11 (with no equivalent in PVṬ): tena pratyakṣeṇa svalakṣaṇe gṛhyamāṇe ’nityatvaṃ gṛhītam eva kevalaṃ bhrāntinimittasadbhāvād aniścitam / atas tanniścayamātre ’numānavyāpāras / tena tanniścaya eva svalakṣaṇe ’nityatvapratītir iti siddham /.

130

64

Vincent Eltschinger

ing fire. And] why [does he lack it]? Because of a wrong notion!132 And [insofar as] this [person] determines this place as free of [fire] (tadviviktena rūpeṇa) through a cognition that does not presume [by any means] that fire exists [there], how can it be said [that this person is] not mistaken (avipar yasta)? And a [person] who would neither superimpose this aspect nor doubt [the existence of fire] would [certainly] not resort to an inference (liṅga) in order to know that [there is fire in this place].”133 2.2.6. We are now in a position to grasp one of the fundamental trends of Dharmakīrtiʼs philosophy. Perception provides an unmediated and unbiased access to reality, especially to the so-called vastudharmas (impermanence, selflessness, painfulness, emptiness), those ultimately real aspects that entities themselves cast into the consciousness. But ignorance (qua conceptuality and concealment) first has us ascribe erroneous intellectual constructs to reality, both by unifying the many and by dividing the indivisible. Second, ignorance (especially as the personalistic false view) has us fail to identify, recognize, or determine the entities’ real aspects by superimposing contrary qualities. Now, aspects such as self, pleasure, or one’s own are the root causes of craving, appropriating, acting and finally being reborn, i.e., suffering. From this perspective, the value of inference as a correcting, error-eliminating principle cannot be overestimated. In a very interesting passage in PV 3, Dharmakīrti clearly connects error, its elimination by inference, and the (yogic, i.e., Buddhist) strengthening of an (inferentially based) conceptual habitus: “Because of the error that is due to the [immediate] occur132 I.e., because this person grasps this place as identical with a spot without fire. 133 PVSV 27,15–28,1: nanu nāvaśyaṃ viparyāsapūrvaka evāpratītaniścayo bhavati / yathā ’kasmād dhūmād agnipratipattiḥ / na hi tatrānagnisamāropaḥ sambhāvyate / tan na sarvatra vyavacchedaḥ kriyate / … tatrāpi taddarśinas tatsvabhāvāniścayaḥ / kutaḥ / viparyāsāt / sa ca taṃ pradeśaṃ tadviviktena rūpeṇa niścinvann agnisattābhāvanā*vimuktayā buddhyā katham aviparyasto nāma / tadākārasamāropasaṃśayarahitaś ca tatpratipattau na liṅgam anusaret /. *On bhāvanā, see Gnoli 1960 (= PVSV): 27–28n. 22. This passage has also been translated and discussed by Kellner (2004: 10–19).

Ignorance, epistemology and soteriology (2)

65

rence of a new (apara) similar [phase, someone] fails to see [i.e., determine] the difference [between two phases as long as the continuum is not interrupted; this person thus] lacks the [determinate] knowledge of a certain [aspect like impermanence, although (s)he has grasped it perceptually (…) But if the continuum is interrupted by an interval of non-existence,] it is indeed without [resorting to any] inference that down to a child, [any] person determines, upon seeing the rise of a new (uttara) [phase of light] disconnected [from the preceding one], that the light [of a lamp], etc., is perishable. [Or,] failing to see the effect [of an entity] because of the interval [implied by the causal process], an ascertainer [can also], due to dullness (apāṭava), be mistaken with regard to [this entity’s very] capacity [to bring about its effect,] although it is inherent to the entity [itself]. It is in order to remove just this [kind of error] that inference is [so] minutely described. [As for] those of great understanding, they determine all aspects [of an entity] by [just] seeing [it].”134 The intimate connection between inference and the search for the structure of ultimate reality and hence soteriology is emphasized in the following statement by Dharmakīrti: “The differentiation between the probandum and the probans is used by/ allows wise people to penetrate ultimate reality.”135 In determining
PV 3.104ac and 105–107: kvacit tad aparijñānaṃ sadṛśāparasambhavāt / bhrānter apaśyato bhedam … // tathā hy aliṅgam ābālam asaṃśliṣṭottarodayam / paśyan paricchinatty eva dīpādiṃ nāśinaṃ janaḥ // bhāvasvabhāvabhūtāyām api śaktau phale ’dṛśaḥ / anāntaryato moho viniścetur apāṭavāt // tasyaiva vinivṛttyartham anumānopavarṇanam / vyavasyantīkṣaṇād eva sarvākārān mahādhiyaḥ //. See PVP D162b6–163b5/ P189a7–190b1 and PVV 148,19–149,17. Note that both Devendrabuddhi and Manorathanandin analyze the compound asaṃśliṣṭottarodayam as a bahu­ vrīhi. Whereas Devendrabuddhi does not elaborate on mahādhiyaḥ, Śākyabuddhi (PVṬ D178a6/P219b7) explains: blo gros chen pos źes bya ba ni dbaṅ po las ’das pa’i don mthoṅ ba’o (*mahādhiya ity atīndriyārthadarśinaḥ), and Manorathanandin (PVV 149,16), more convincingly: mahādhiyo viparītavyavasāyānākrāntapratyakṣā yoginaḥ. 135 PV 1.86bd: sādhyasādhanasaṃsthitiḥ / paramārthāvatārāya vidvadbhir avakalpyate //. Skt. saṃsthiti is not entirely clear, but must be semantically near vyavasthāna (PV 1.85). Manorathanandin explains saṃsthiti (PV 3.214, 3.315, 3.319, 4.15, 4.64) as vyavasthā (PVV 182,25, 213,14–15, 214,22, 419,11–12, 437,3), “settlement, establishment; statute; fixed rule.” On
134

66

Vincent Eltschinger

what had remained unidentified and hereby excluding wrong notions, inference indeed restores, still on a purely conceptual level, the most fundamental features of reality. The sequence linking the obliteration of perception and an inference’s corrective function is outlined by Dharmakīrti in a highly suggestive statement of PV 2: “The property of [all] cognition is to grasp an object; this [object] is grasped as it [really] is [i.e., as impermanent, etc.], and it generates this [cognition of itself] by [its] real nature. And such is [the object’s and the cognition’s original] nature [i.e., that the object generates a cognition that grasps it as it really is, and that the cognition grasps a real aspect of the object. But] on account of another cause [i.e., on account of a cause of error], the [mind] shifts (skhalat) from this [inherently veracious nature, superimposing such erroneous aspects as permanence on the object,] and becomes uncertain, requiring a [cognitive] condition for the removal [of this state], like the cognition of a piece of rope [as a snake].”136 There is little doubt that the condition alluded to here, explained by Devendrabuddhi as “a means of valid cognition annulling error,”137 is none other than inference. And given the soteriological context (description of the final revolution of the basis, āśrayaparivṛtti) in which this statement occurs, it is no less obvious that Dharmakīrti holds that this condition provides the first impetus toward establishing the mind (vijñāna, i.e., perception), at the completion of the path, in its genuine radiant condition. Taking Dharmakīrti’s epistemological interpretation of the mind’s natural radiance seriously, but also his insistence on perception’s non-erroneousness and its giving access to the ultimate structure of reality, we are left with no other possibility than to hold perception before and after the āśrayaparivṛtti to be one and the same with regard to its content and operation. As we have seen, ignorance as “anti-knowledge”
avatāra, see BHSD s.v., 71a. 136 PV 2.206–207: viṣayagrahaṇaṃ dharmo vijñānasya yathāsti saḥ / gṛhyate so ’sya janako vidyamānātmaneti ca // eṣā prakṛtir asyās tan nimittāntarataḥ skhalat / vyāvṛttau pratyayāpekṣam adṛḍhaṃ sarpabuddhivat //. See above, §2.2.2 and n. 87. 137 PVP D89a2–3/P102b5–6: rkyen la ltos pa yin te / de ltar … ’khrul pa gnod pa can gyi tshad ma la ltos pa daṅ bcas pa yin no //.

Ignorance, epistemology and soteriology (2)

67

neither impedes nor obliterates perception itself, but is responsible for subsequent errors and superimpositions. The main difference between cognition before and after the āśrayaparivṛtti, i.e., between cognition-cum-ignorance and cognition-sine-ignorance, does not pertain to perception itself, or, as Dharmakīrti himself would have it, to the nature of the mind, but to the subsequent treatment of perceptual data. Inference is responsible for bringing out the intellectual contents that correct erroneous superimpositions; it makes determinate cognition possible, and further, endows the yogin with true conceptual counterparts of the entities’ real aspects. In other words, inference sets the path in motion138 that will first enable the yogin to determine the real aspects of entities upon perceiving them,139 and then free his mind from all those adventitious factors that counteracted perception. To the best of my underNote PVṬ Je D252a1–2/P299a8–b1 = PVSVṬ 401,12–13: pramāṇāny anityādibhūtākāragrāhī ṇi pratipakṣamārgam āvahanti /. 139 Note, e.g., PVṬ Je D70b4–5/P83a4–5 = PVSVṬ 142,15: yathā yogināṃ buddhipāṭavād darśanamātreṇa kṣaṇikatvādiniścayaḥ /. That perception as such does not differ between ordinary people (pṛthagjana) and yogins is also Karṇakagomin’s opinion in two interesting statements. (1) PVSVṬ 91,24–25: yogināṃ saty api sadṛśadarśane mandabuddhitvābhāvāt kṣaṇikatvaniścayo bhavati … “The yogins do determine momentariness because, though [their perceptual] experience is the same [as that of ordinary persons], they lack [this] being of weak intellect.” (2) PVSVṬ 92,19–21: mandabuddhir (PVSV 21,7) iti / tena bāhyādhyātmikavipralambhanimittasadbhāvāt pṛthagjanānāṃ [na] niścayaḥ / yogināṃ tu saty api sadṛśadarśane paṭubuddhitvān niścayo bhavaty eva /. “By ‘of weak intellect,’ [Dharmakīrti means the following:] Because of the presence of both external [i.e., the rise of a new similar phase, etc.,] and internal [i.e., ignorance,] causes of error, ordinary persons fail to determine [momentariness in the same way as they have experienced it], but the yogins, though [their perceptual] experience is the same [as that of ordinary persons], do indeed determine [momentariness] because they are of sharp intellect.” According to Karṇakagomin, then, perception itself does not differ between those who have reached the darśanamārga and those who have not; what indeed differs is the degree of their intellectual sharpness, the increase of which can only be due to the habitus (abhyāsa) or cultivation (bhāvanā) that comes along the path. On the context of these statements and the issue of internal as well as external causes of error, see above, n. 116; on abhyāsa as a condition for determinate cognitions to arise, see Kellner 2004: 19–32.
138

68

Vincent Eltschinger

standing, the perception of the liberated saint is to be equated with the paramārthikapramāṇa that Dharmakīrti touches upon at the end of PVin 1.140 I do not intend to claim, in contrast to most scholars and the textual evidence, that Dharmakīrti’s inference has only soteriological meaning and relevance. By pointing out Dharmakīrti’s insistence upon the vastudharmas in his treatment of both perception and inference, and by putting to the fore the corrective function of inference, I would like to emphasize the fact that Dharmakīrti never lost sight of soteriology in his elaborations on epistemology. According to him, there is at least one set of cases (the most important ones indeed) in which the use of inference coincides with, or impinges upon, the precincts of the wisdom born of rational reflection (yukticintāmayī prajñā).141 The wisdom born of rational reflection traditionally consists (at least in connection with the socalled upapattisādhanayukti) in an analysis carried out on the basis of the means of valid cognition. This holds true of the Buddhist epistemologists, according to whom rational reflection basically aims at bringing out intellectual contents that have been thoroughly examined and made immaculate by means of valid cognition (pramāṇaparidṛṣṭārtha, pramāṇapariśuddhārtha), i.e., by inference.142 Though still strictly conceptual in nature, these contents (the vastudharmas again) “co-function” as the antidote (pratipakṣa = nairātmyadarśana, etc.) to the cause of suffering, i.e., ignorance in the form of personalistic belief. Most ordinary people may
PVin 1 44,4–5: cintāmayīm eva tu prajñām anuśīlayanto vibhramavivekanirmalam anapāyi pāramārthikapramāṇam abhimukhīkurvanti /. On this passage, see Krasser 2004: 142–144 and Eltschinger 2005: 155–158. That liberated perception comes about through the yogin’s initially resorting to inferences is clear. How it can be equated with omniscience remains, however, obscure. But does not Dharmakīrti himself term “unfathomable” (acintya) the cognition of (liberated) yogins and the Buddha’s omniscience? PV 3.532d: acintyā yogināṃ gatiḥ //; SAS 94: bcom ldan ’das kyis don thams cad thugs su chud pa ni bsam gyis mi khyab ste / rnam pa thams cad du śes pa daṅ brjod pa’i yul las ’das pa’i phyir ro //. 141 PVin 1 27,9. 142 On the cintāmayī prajñā in the Buddhist epistemologists, see Eltschinger 2010.
140

Ignorance, epistemology and soteriology (2)

69

well show no interest at all for evolving determinate cognitions of momentariness and selflessness. But to the Buddhist yogin still in the stage of being an ordinary person, investigating the most intimate structure of reality by means of inferences is the first significant step towards the path of vision and liberation.

References Abbreviations
BHSD Franklin Edgerton: Buddhist Hybrid Sanskrit Grammar and Dictionary. Volume II: Dictionary. Delhi 1970: Motilal Banarsidass. Jikido Takasaki/Zuiho Yamaguchi/Noriaki Hakamaya: sDe dge Tibetan Tripiṭaka bsTan ’gyur preserved at the Faculty of Letters, University of Tokyo. Tokyo 1977–1981. Manuscript Daisetz T. Suzuki: The Tibetan Tripitaka, Peking Edition, Kept in the Library of the Otani University, Kyoto. Tōkyō/Kyōto 1957: Tibetan Tripitaka Research Institute. sub voce Tibetan

D

ms P

s.v. Tib

Primary sources
AK(Bh) – Prahlad Pradhan: Abhidharmakośabhāṣyam of Vasubandhu. Patna 1975: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series 8). AKVy – Unrai Wogihara: Sphuṭārthā Abhidharmakośavyākhyā, the Work of Yaśomitra. Tokyo 1989: Sankibo Buddhist Book Store (The Publishing Association of Abhidharmakośavyākhyā). AN – R. Morris/E. Hardy/M. Hunt/C.A.F Rhys Davids: Aṅguttara Nikāya. 6 volumes. London 1885–1910: The Pali Text Society. ASBh – Nathmal Tatia: Abhidharmasamuccayabhāṣyam. Patna 1976: K.P. Jayaswal Research Institute (Tibetan Sanskrit Works Series 17). BhK 1 – First Bhāvanākrama (Kamalaśīla). Pp. 497/187–539/229 in Giuseppe Tucci: Minor Buddhist Texts. Delhi 1986: Motilal Banarsidass.

70

Vincent Eltschinger

BhK 3 – Giuseppe Tucci: Minor Buddhist Texts, Part III: Third Bhāvanākrama. Roma 1971: Istituto Italiano per il Medio ed Estremo Oriente (Serie Orientale Roma 43). HV – Hetuvidyā Section of the Yogācārabhūmi. Hideomi Yaita: Three Sanskrit Texts from the Buddhist Pramāṇa-Tradition: The Hetuvidyā Section of the Yogācārabhūmi, the Dharmottaraṭippanaka, and the Tarkarahasya. Narita 2005: Naritsan Shinshoji (Monograph Series of Naritasan Institute for Buddhist Studies 4). The Sanskrit text of the HV can be found on pp. 98/1*–124/27*. MMK – See PrP. MN I – V. Trenckner: The Majjhima-Nikāya. Vol. 1. London 1935: Pali Text Society. MS – See Lamotte 1973: I. NB(Ṭ) – Th. I. Ščerbatskoj: Nyāyabindu. Buddijskij učebnik’’ logiki sočinenie Dharmakirti I tolkovanie na nego Nyāyabinduṭīkā sočinenie Darmottary. Osnabrück 1970: Biblio Verlag. Pā – Pāṇini (Aṣṭādhyāyī). PrP – Louis de La Vallée Poussin: Madhyamakavṛttiḥ: Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna avec la Prasannapadā Commentaire de Candrakīrti. Delhi 1992: Motilal Banarsidass. PrSVy – Pratītyasamutpādavyākhyā (Vasubandhu). D no. 3995, Chi 1b– 61a, P no. 5496, Chi 1–71a. PV 1–4 – Yūsho Miyasaka: Pramāṇavārttika-kārikā (Sanskrit and Tibetan). Acta Indologica 2 (1971–1972), pp. 1–206. See also PVV; for PV 2–3, see also PVA; for PV 1, see also PVSV; for PV 2.131cd–285, see also Vetter 1990. My numbering of the verses in PV 2 follows that of Vetter. PVA – Rāhula Sāṅkṛtyāyana: Pramāṇavārttikabhāṣyam or Vārtikālaṅkāraḥ of Prajñākaragupta (Being a Commentary on Dharmakīrti’s Pramāṇavārtikam). Patna 1953: K.P. Jayaswal Research Institute. PVin 1–2 – Ernst Steinkellner: Dharmakīrti’s Pramāṇaviniścaya, Chapters 1 and 2. Beijing/Vienna 2007: China Tibetology Publishing House/Austrian Academy of Sciences Press (Sanskrit Texts from the Tibetan Autonomous Region 2). PVinṬ – Pramāṇaviniścayaṭīkā (Dharmottara). D no. 4229, Dze 1b1–Tshe 178a3/P no. 5727, Dze 1b1–We 209b8. PVP – Pramāṇavārttikapañjikā (Devendrabuddhi). D no. 4217, Che 1–326b4/P no. 5717, Che 1–390a8. PVSV – Raniero Gnoli: The Pramāṇavārttikam of Dharmakīrti. The First Chapter with the Auto-Commentary. Roma 1960: Istituto Italiano per il Medio ed Estremo Oriente (Serie Orientale Roma 23).

Ignorance, epistemology and soteriology (2)

71

PVSVṬ – Rāhula Sāṅkṛtyāyana: Karṇakagomin’s Commentary on the Pramā ṇavārttikavṛtti of Dharmakīrti. Kyōto 1982: Rinsen Books Co. PVṬ – Pramāṇavārttikaṭīkā (Śākyabuddhi). D no. 4220, Je 1b1–Ñe 282a7/P no. 5718, Je 1b1–Ñe 348a8. Unless otherwise stated, all references to the PVṬ belong to Ñe. PVV – Rāhula Sāṅkṛtyāyana: Dharmakīrti’s Pramāṇavārttika with Commentary by Manorathanandin. Published as an appendix to the Journal of the Bihar and Orissa Research Society 24–26 (1938–1940). SAS – F.I. Ščerbatskoj: Tibetskij perevo’’ sočinenij Saṃtānāntarasiddhi Dharmakīrti i Saṃtānāntarasiddhiṭīkā Vinītadeva. Delhi 1992: Motilal Banarsidass. TS(P) – K = Embar Krishnamacharya: Tattvasaṅgraha of Śāntarakṣita With the Commentary of Kamalaśīla. 2 vols. Baroda 1984: Oriental Institute. *UH – *Upāyahṛdaya/*Prayogasāra. Giuseppe Tucci: Pre-Dignāga Buddhist Texts on Logic from Chinese Sources. Baroda 1929: Oriental Institute (Gaekwad Oriental Series 49). Vibh. – Vibhūticandra’s notes to PVV. See PVV. VY – Jong Cheol Lee: The Tibetan Text of the Vyākhyāyukti of Vasubandhu. Tōkyō 2001: The Sankibo Press (Bibliotheca Indologica et Buddhologica 8). YBh – Yogācārabhūmi, or, followed by page/line numbers: V. Bhattacharya: The Yogācārabhūmi of Ācārya Asaṅga. Calcutta 1957: University of Calcutta.

Secondary sources
Collins 1982 – Steven Collins: Selfless Persons. Imagery and thought in Theravāda Buddhism. Cambridge/New York 1982: Cambridge University Press. Eltschinger 2005 – Vincent Eltschinger: Études sur la philosophie religieuse de Dharmakīrti: 2. L’āśrayaparivṛtti. Journal Asiatique 293/1 (2005), pp. 151–211. Eltschinger 2009 – Vincent Eltschinger: Ignorance, epistemology and soteriology – Part I. Journal of the International Association of Buddhist Studies 32/1–2 (2009 [2010]), pp. 39–83. Eltschinger 2010 – Vincent Eltschinger: Studies in Dharmakīrti’s Religious Philosophy: 4. The Cintāmayī Prajñā. Pp. 553–591 in Piotr Balcerowicz (ed.): Logic and Belief in Indian Philosophy. Delhi 2010: Motilal Banarsidass.

72

Vincent Eltschinger

Frauwallner 1959 – Erich Frauwallner: Dignāga, sein Werk und seine Entwicklung. Wiener Zeitschrift für die Kunde Süd- und Ostasiens 3 (1959), pp. 83–164. Funayama 1992 – Toru Funayama: A Study of kalpanāpoḍha. A Trans­ lation of the Tattvasaṃgraha vv. 1212–1263 by Śāntarakṣita and the Tattvasaṃgrahapañjikā by Kamalaśīla on the Definition of Direct Perception. Kyoto 1992: Zinbun Kagaku Kenkyūsho, Kyōto University. Funayama 1999 – Toru Funayama: Kamalaśīla’s Interpretation of ‘Non-Erroneous’ in the Definition of Direct Perception and Related Problems. Pp. 73–99 in Shoryu Katsura (ed.): Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakīrti Conference (Hiroshima, November 4– 6, 1997). Vienna 1999: Verlag der Österreichischen Akademie der Wissenschaften (Philologisch-historische Klasse, Denkschriften, 281). Jaini 2001 – Padmanabh S. Jaini: On the Ignorance of the Arhat. Pp. 167–179 (= Chapter 9) in Padmanabh S. Jaini: Collected Papers on Buddhist Studies. Delhi 2001: Motilal Banarsidass. Kellner 2004 – Birgit Kellner: Why Infer and not just Look? Dharmakīrti on the Psychology of Inferential Processes. Pp. 1–51 in: Shoryu Katsura/Ernst Steinkellner (eds.): The Role of the Example (dṛṣṭānta) in Classical Indian Logic. Vienna 2004: Arbeitskreis für tibetische und buddhistische Studien Universität Wien (Wiener Studien zur Tibetologie und Buddhismuskunde 58). Kośa – Louis de La Vallée Poussin: L’Abhidharmakośa de Vasubandhu. 6 vols. Bruxelles 1980: Institut Belge des Hautes Études Chinoises (Mélanges Chinois et Bouddhiques 16). Krasser 1991 – Helmut Krasser: Dharmottaras kurze Untersuchung der Gültigkeit einer Erkenntnis – Laghuprāmāṇyaparīkṣā. Teil 2: Übersetzung. Vienna 1991: Verlag der Österreichischen Akademie der Wissenschaften (Beiträge zur Kultur- und Geistesgeschichte Asiens 7). Krasser 1995 – Helmut Krasser: Dharmottara’s Theory of Knowledge in his Laghuprāmāṇyaparīkṣā. Journal of Indian Philosophy 23 (1995), pp. 247–271. Krasser 2004 – Helmut Krasser: Are Buddhist Pramāṇavādins non-Buddhistic? Dignāga and Dharmakīrti on the impact of logic and epistemology on emancipation. Hōrin: Vergleichende Studien zur japani­ schen Kultur 11 (2004), pp. 129–146. Kritzer 1999 – Robert Kritzer: Rebirth and Causation in the Yogācāra Abhidharma. Vienna 1999: Arbeitskreis für tibetische und buddhistische Studien Universität Wien (Wiener Studien zur Tibetologie und Buddhismuskunde 44).

Ignorance, epistemology and soteriology (2)

73

La Vallée Poussin 1913 – Louis de La Vallée Poussin: Bouddhisme, études et matériaux. Théorie des douze causes. Gand 1913: Librairie scientifique E. van Goethem (Université de Gand, Recueil de travaux publiés par la Faculté de philosophie et des lettres 40). Lamotte 1973 – Étienne Lamotte: La Somme du Grand Véhicule d’Asaṅga (Mahāyānasaṅgraha). 2 vols. Louvain-la-Neuve, 1973: Université de Louvain, Institut Orientaliste (Publications de l’IOL 8). Lamotte 1987 – Étienne  Lamotte: Le troisième Bhāvanā-krama de Kamalaśīla. Traduction de la version tibétaine. Pp. 336–353 in Paul Demiéville: 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 1987: Collège de France, Institut des Hautes Études Chinoises (Publication de l’Institut des Hautes Études Chinoises 7). May 1959 – Jacques May: Candrakīrti: Prasannapadā Madhyamakavṛtti. Douze chapitres traduits du sanscrit et du tibétain, accompagnés d’une introduction, de notes et d’une édition critique de la version tibétaine. Paris 1959: Adrien Maisonneuve (Collection Jean Przyluski 2). McClintock 2010 – Sara L. McClintock: Omniscience and the Rhetoric of Reason. Śāntarakṣita and Kamalaśīla on Rationality, Argumentation, and Religious Authority. Boston 2010: Wisdom Publications (Studies in Indian and Tibetan Buddhism). Mejor 2001 – Marek Mejor: Controversy on the mutual conditioning of avidyā and ayoniśomanas(i)kāra in Vasubandhu’s Abhidharmakośa. Journal of the International College for Advanced Buddhist Studies 4 (2001), pp. 49[/292]–78[/263]. Pruden 1988–1990 – Leo M. Pruden: Abhidharmakośabhāṣyam by Louis de La Vallée Poussin. English Translation by Leo M. Pruden. 4 vols. Berkeley 1988–1990: Asian Humanities Press. Ś = Swami Dwarikadas Shastri: Tattvasaṅgraha of Ācārya Shāntarakṣita with the Commentary ‘Pañjikā’ of Shri Kamalshīla. 2 vols. Varanasi 1981: Bauddha Bharati (Bauddha Bharati Series 1). Schmithausen 1977 – Lambert Schmithausen: Zur buddhistischen Lehre von der dreifachen Leidhaftigkeit. Pp. 918–931 in Wolfgang Voigt (ed.): Zeitschrift der Deutschen Morgenländischen Gesellschaft, Supplement III,2 (XIX. Deutscher Orientalistentag). Wiesbaden 1977: Franz Steiner Verlag. Steinkellner 1979 – Ernst Steinkellner: Dharmakīrti’s Pramāṇaviniścayaḥ. Zweites Kapitel: Svārthānumānam. Teil II: Übersetzung und Anmerkungen. Vienna 1979: Verlag der Österreichischen Akademie der Wissenschaften (Veröffentlichungen der Kommission für Sprachen und Kulturen Südasiens 15).

74

Vincent Eltschinger

Vetter 1990 – Tilmann Vetter: Der Buddha und seine Lehre in Dharma­ kīrtis Pramāṇavārttika. Der Abschnitt über den Buddha und die vier edlen Wahrheiten im Pramāṇasiddhi-Kapitel. Vienna 1990: Arbeitskreis für tibetische und buddhistische Studien Universität Wien (Wiener Studien zur Tibetologie und Buddhismuskunde 12). Wayman 1961 – Alex Wayman: Analysis of the Śrāvakabhūmi Manuscript. Berkeley/Los Angeles 1961: University of California Press. Wayman 1980 – Alex Wayman: The Sixteen Aspects of the Four Noble Truths and Their Opposites. Journal of the International Association of Buddhist Studies 3/2 (1980), pp. 67–76.

Tall tales, tathāgatas, and truth On the “privileged lie” in Indian Buddhist literature
Richard F. Nance

1. Introduction
A standard-issue Buddhist view on the practice of lying is that it is harmful. Lies are not told by buddhas and they should not be told by properly observant Buddhists. Those who do tell them are said to suffer various horrid consequences: bad breath, being born without a tongue, and rebirth in hell realms.1 Together with slander,
The view is affirmed even in texts acknowledging that the harmful effects of lying may not be immediately apparent. A passage from the Gāmaṇisaṃyutta (SN iv.47) suggests that while the long-term effects of lying are invariably negative, its short-term consequences may, in fact, appear to favor the liar: “Then, headman, someone here is seen garlanded and adorned, freshly bathed and groomed, with hair and beard trimmed, enjoying sensual pleasures with women as if he were a king. They ask someone about him: ‘Sir, what has this man done?’ … They answer: ‘Sir, this man amused the king with false speech. The king was pleased with him and bestowed a reward upon him. That is why the man is garlanded and adorned … enjoying sensual pleasures with women as if he were a king’.” (trans. Bodhi 2000: 1364) This scenario is immediately followed by one in which a speaker is severely punished for speaking falsely: having brought to ruin a householder or householder’s son by lying, the liar’s arms are bound behind him with strong rope, his head is shaved, and he is led through the streets to the outskirts of the city, where he is beheaded. So, the text concludes, one should not place confidence in “those ascetics and brahmins … who say ‘Anyone at all who speaks falsely experiences pain and grief here and now’.” This is a false view Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 75–101
1

76

Richard F. Nance

harsh speech, and idle chatter, lying is traditionally understood to be a form of unwholesome speech – and to engage in unwholesome speech is to bring pain upon oneself. “A person is born,” we are told, “with an axe in his mouth. One whose speech is unwholesome cuts himself with the axe.”2 Of the various forms of unwholesome speech, lying is sometimes singled out as particularly egregious. One who lies is “the banner of all vices, the producer of all evils: a singular source of darkness.”3 From the Pāli Dhammapada, we learn that “there is no evil that might not be done by a person who tells a lie.”4 The consequences of lying are sufficiently grave that in the Ratnāvalī, Nāgārjuna admonishes his royal addressee to stand fast against duplicity, even if doing so should cost him his life and kingdom: “For your own sake, always (rtag, *nitya) tell the truth – even if it should cause your death or ruin your governance. Do not speak otherwise.”5 In the Abhidharmakośabhāṣya, Vasubandhu counsels his audience that although they may have heard that some forms of lying are harmless – that “a jesting untruth does no harm – nor
– but rejecting such a view does not entail that one thereby rejects the notion that lies are harmful. 2 Sn 657: purisassa hi jātassa kuṭhāri jāyate mukhe, yāya chindati attānaṃ bālo dubbhāsitaṃ bhaṇaṃ. This image recurs elsewhere inside and outside the Pāli canon (cf. SN i.149; AN v.174); Udānavarga viii.2 (Bernhard 1965, Vol. I: 161); Dharmasamuccaya 12.6 (Lin 1969: 372). 3 Dharmasamuccaya 12.7 (Lin 1969: 373): sarvākāryapatākā sā sarvapāpaprasūtikā / tamasāṃ yonir ekā sa yo vācaṃ bhāṣate mṛṣā // 4 Cf. Dhammapada, verse 176 (= Iti 1.3.5): ekaṃ dhammaṃ atītassa musāvādissa jantuno  / vitiṇṇaparalokassa natthi pāpaṃ akāriyaṃ; and 306ab: abhūtavādī nirayaṃ upeti yo cāpi katvā na karomīti cāha. The former passage is partially paralleled by Dharmasamuccaya 12.3 (Lin 1969: 371); the latter, by Udānavarga viii.1 (Bernhard 1965, Vol. I: 161). While the former passage is missing from the Gāndhārī Dharmapada, the latter has been preserved (Brough 1962: 161, verse 269). 5 Translation modified from Hopkins 1998: 129. The Sanskrit for this verse (#274) of the Ratnāvalī is lost. The Tibetan of Hopkins’s edition reads: bden pa gang gis rang don la / ’chi ’am yang na rgyal po’i srid / nyams ’gyur de ni rtag brjod cing / de las gzhan du brjod mi bgyi. Cf. Hahn 1982: 86: bden pa gang gis rang [don la] / rgyal srid nyams par gyur na’ang de / rang gi don la rtag brjod cing / de las gzhan smra mi bgyi’o //

Tall tales, tathāgatas, and truth

77

does one concerning women, one made at the time of marriage, or one made when one’s life or all one’s property is in danger” – such a view is both confused and false.6 The impression that one can get from such passages (and they could be multiplied) is that the Buddhist proscription against lying is absolute: that the tradition does not hold any lie to be “privileged” (i.e., excusable). This view has recently been voiced by the Indologist, comparativist, and legal scholar J. D. M. Derrett, who notes in a 2006 article in the Journal of Buddhist Ethics that for Buddhists, “lies are so injurious that no convenience can excuse lying,” and hence that “privileged lies [are] … totally missing from Buddhism.”7 Yet, pace Derrett, Buddhist views on the subject of lying are far more complex than the sweeping remark above suggests. In what follows, some of these views and their complexities will be surveyed. It may be useful to begin by getting clear on the terms at stake in the relevant Indic languages. The compound musāvāda (Sanskrit mṛṣāvāda) is sometimes translated into English as “false speech,” sometimes as “lie.” At first blush, these notions may appear to be interchangeable, but they are not. To say that someone has spoken falsely need not imply that she has lied. A person might wrongly presume things to be a certain way, and then go on to describe her erroneous impression accurately. In doing so, she will have spoken falsely – but she will not have lied.8 Moreover, to say that someone
6 Abhidharmakośabhāṣya on Abhidharmakośa iv.68 (Vol. 2: 538). The verse quoted by Vasubandhu here is also preserved in the Mahābhārata (1.77.16 of the BORI edition, Sukthankar 1933, Vol. 1, Part 1: 349): na narmayuktam anṛtaṃ (Sukthankar, though acknowledging this variant, here reads vacanaṃ) hinasti na strīṣu rājan na vivāhakāle  / prāṇātyaye sarvadhanāpahāre pañcānṛtāny āhur apātakānīti. Cf. Müller 1883: 273. 7 Derrett 2006: 1. 8 This is not a possibility open to buddhas, since buddhas do not make such mistakes. The distinction between lying and speaking falsely can only be made if a certain kind of mistake is possible: one in which we (unknowingly) fail to grasp how things in fact are, and yet accurately report this mistaken understanding. Given that a buddha unfailingly grasps how things in fact are, the question of whether a buddha is capable of speaking falsely collapses into the question of whether a buddha is capable of lying.

78

Richard F. Nance

has lied need not imply that she has spoken falsely. When a person lies, she misrepresents how she takes things to be – but the falsity of an utterance is typically a matter that has to do not with how a speaker takes things to be, but with how things in fact are. Yet to say that lies involve no more than the misrepresentation of how one takes things to be is still insufficient, since a person can misrepresent how she takes things to be without lying. Such misrepresentations occur regularly for those just beginning the study of an unfamiliar language. Given an exercise in which we are asked to describe our immediate surroundings, we may falter: our descriptions may well misrepresent how we take things to be. These misrepresentations are hardly lies; if anything, they are simple mistakes. For these misrepresentations to become lies, what is required in addition is an element of deliberateness: we lie if, and only if, we deliberately misrepresent how we take things to be. Similar concerns regarding deliberate misrepresentation are broached in Buddhist disciplinary (vinaya) texts, which stipulate a number of conditions that must obtain in order for a particular act of speaking to be judged in violation of proscriptions against mṛṣā-(musā-)vāda. These conditions clarify that what is at issue in such proscriptions is indeed the deliberate misrepresentation of how one takes things to be. At times, this element of deliberateness is made explicit. Far more often, however, such explicit signaling is absent – yet the context makes it clear that an element of deliberateness is being presupposed.9 What is typically at stake in discussion of mṛṣā-(musā-)vāda is, then, not simply false speech, but lying.10
9 One occasionally finds the compound mṛṣā-(musā-)vāda augmented by the term samp(r)ajāna to clarify that what is at issue is the knowing propagation of falsehoods. See, for example, MN 86 (Aṅgulimālasutta), ii.103 (trans. Ñāṇamoli and Bodhi 1995: 714), and the pācittiya proscription against lying (sampajānamusāvāde pācittiya). Yet, as noted below, when the term samp(r)ajāna is absent, an element of deliberateness is sometimes presumed to be signaled by mṛṣā / musā itself: Buddhaghosa, for example, glosses the term musā- as “intentionally misleading” (visaṃvādanādhippāya) (Sumaṅgalavilāsinī i.9). 10 This is not, however, the only way in which the term mṛṣā- (or the Tibetan [b]rdzun pa, which is stipulated as a suitable translation in the Mahāvyutpatti [#7313]) can be used. At times, the terms signal forms of deceptive-

Tall tales, tathāgatas, and truth

79

We can thus rephrase Derrett’s sweeping claim as follows: Buddhist doctrinal texts prohibit the practice of deliberately misrepresenting how one takes things to be, and they do so in absolute terms: there is no case in which such deliberate misrepresentation is held to be permissible. All lies, then, are held to be equally deserving of censure. To say this does not, of course, entail that all lies are held to deserve equal censure – and Buddhist texts do indeed treat lies as falling into various categories, not all of which are judged equally blameworthy. Some lies constitute pārājika offenses, punishable by expulsion from the monastic community. These involve intentionally and falsely representing oneself as one who has seen or known things which are seeable and knowable only by persons of consummate attainment (uttaramanuṣya).11 Most lies are not, however, punished so harshly. Lies that are not counted as pārājika offenses are most often counted as pācittiya (Skt. pāyantika) offenses.12 These are much less onerous; a Buddhist monastic who commits a pācittiya / pāyantika offense may be absolved of fault after the offense is formally confessed. In the Milindapañha, the Buddhist monk Nāgasena is asked about this distinction between forms of lying, and he offers a few clarificatory comments.13 According to Nāgasena, lies can be light (lahuka) or heavy (garuka); the gravity of a particular lie depends on its subject matter (vatthu). In this respect, Nāgasena insists, lyness that are not intentional in the sense taken up here. So, for example, in the Lives of the Eighty-Four Siddhas (Caturaśītisiddhapravṛtti), attributed to Abhayadattaśrī, the siddha Thaganapa, unable to refrain from lying (rdzun smra), is counseled by a monk who tells him “you should contemplate everything as a single deception” (khyod kyis thams cad rdzun gcig por sgoms shig). There is no indication, however, that this practice of contemplation requires Thaganapa to posit a being that deliberately sets out to deceive. Cf. also Ratnagotravibhāga 1.86 (Prasad 1997: 55). 11 Cf. the treatment of pārājika 4 in Pachow 2000. For the relevant Sanskrit, see Bannerjee 1977: 15. 12 False claims made against fellow monastics for the purpose of bringing about their expulsion are, however, categorized as saṅghādisesa (Skt. saṅghāvaśeṣa) offenses. 13 See Milindapañha 192–3 (trans. Horner 1969, Vol. 1: 275–7).

80

Richard F. Nance

ing does not differ from striking another person. Just as the punishment meted out to one who strikes a man on the street differs from the punishment meted out to one who strikes a king, so too are there different kinds of lies, meriting different kinds of punishment.14 Of course, simply knowing that lies are to be distinguished based on their subject matter does not help us to assess the gravity of any particular lie. How does one distinguish between lies that are light and lies that are heavy? The great Buddhist commentator Buddhaghosa offers something like a formal criterion: “when the welfare (attha) that it [i.e., a lie] destroys is slight, it is less blameworthy; when the welfare is great, it is more blameworthy.”15 For Buddhaghosa, assessing the gravity of a lie is not – or at least not simply – a matter of assessing what the lie is about; one must consider the lie’s impact on the welfare of others. And, as Buddhaghosa acknowledges, this impact may depend not only on the subject matter of the lie, but also on the context in which the lie is told. This context encompasses not only the aims or intentions of the liar – what he or she means to accomplish in telling the lie – but also the social situation in which the lie is uttered: Buddhaghosa notes that a lie told during a formal disciplinary proceeding will have a greater impact on the welfare of others than a lie told in jest.16
It is important not to misread the terms of the analogy here: Nāgasena is not implying that lies told to common people are less egregious than lies told to kings. The distinction marked does not derive from the social identity (or political clout) of those to whom a particular lie is told, but from the subject matter of the lie. 15 Trans. Bodhi 1978: 118. Cf. Sumaṅgalavilāsinī 1.9: so yam atthaṃ bha­ ñjati tassa appatāya appasāvajjo mahantatāya mahāsāvajjo. 16 Ibid., “When a householder, reluctant to part with a certain possession, denies that he owns it, it is of little blame; but when he is caused to witness and lies for the sake of destroying another’s welfare, then the blame is heavy. For monks the blame is light when they speak in jestful exaggeration, e.g. if after getting a little oil or ghee they say, ‘Oil flows like a river in the village today.’ But the blame is heavy when they claim to have seen something they did not see.” (gahaṭṭhānaṃ attano santakaṃ adātukāmatāya natthītiādinayappavatto appasāvajjo, sakkhinā hutvā atthabhañjanatthaṃ vutto mahāsāvajjo, pabbajitānaṃ appakampi telaṃ vā sappiṃ vā labhitvā
14

Tall tales, tathāgatas, and truth

81

2. Exceptions to the general rule
The claim that has so far served as a stalking horse – Derrett’s claim that privileged lies are totally missing from Buddhism – does not preclude the notion that Buddhist normative texts present certain lies as somewhat less blameworthy than others. The issue is not whether some lies are presented as less blameworthy, but whether any lie is presented as altogether blameless. The answer to this question would appear to be “yes.” Consider the following, drawn from the Śīlapaṭala section of the Bodhisattvabhūmi:
Although a bodhisattva would not tell a deliberate lie even to save his own life, he speaks that which aims at saving the lives of many sentient beings, at freeing them from bonds, at protecting them from having their hands, feet, noses, or ears cut off, or their eyes gouged out – and [he does so] having individually reckoned the benefit to sentient beings. So, in brief, via this or that [means], a bodhisattva sees precisely [what is of] benefit to sentient beings; he does not see [what is] not [of] benefit to them. His mind is unconcerned with his own gain. And focusing on an idea that is prompted only by the desire to benefit sentient beings, he deliberately speaks in another way. In this way, he does not incur fault, and spreads much merit.17

It is difficult to see this passage as one that does not present certain lies as privileged – viz., lies told by bodhisattvas who, having judiciously appraised the circumstances at hand, are unconcerned with their own benefit, and solely intent on securing the
hasādhippāyena – ajja gāme telaṃ nadī maññe sandatīti pūraṇakathānayena pavatto appasāvajjo adiṭṭhaṃy eva pana diṭṭhantiādinā nayena vadantānaṃ mahāsāvajjo.) 17 Emphasis added. Bodhisattvabhūmi pp. 114–115: yathāpi tad bodhisat­ tvo bahūnāṃ sattvānāṃ jīvitavipramokṣārthaṃ bandhanavipramokṣārthaṃ hastapādanāsākarṇacchedacakṣurvikalībhāvaparitrāṇārthaṃ yāṃ bodhisattvaḥ svajīvitahetor api samprajānan [sic] mṛṣāvācaṃ na bhāṣeta  / tāṃ teṣāṃ sattvānām arthāya pratisaṃkhyāya bhāṣate / iti samāsato yena yena bodhisattvaḥ sattvānām artham eva paśyati / nānarthaṃ paśyati / svayaṃ ca nirāmiṣacitto bhavati  / kevalasattvahitakāmatānidānaṃ ca vinidhāya saṃjñāṃ samprajānan [sic] anyathāvācaṃ bhāṣate / bhāṣamāṇaḥ anāpattiko bhavati / bahu ca puṇyaṃ prasūyate.

82

Richard F. Nance

benefit of others. In telling such a lie, a bodhisattva not only incurs no fault (he is anāpattika); but he also spreads much merit (bahu ca puṇyaṃ prasūyate). Analogous claims are advocated in other Buddhist śāstric texts as well.18 Śāntideva’s Bodhicaryāvatāra famously notes that “even what is proscribed is permitted for a compassionate person who sees it will be of benefit.”19 The point is elaborated and clarified by Prajñākaramati, in his commentary on Śāntideva’s text:
If someone should object, ‘How can he avoid committing an offense (āpatti) while engaged in what is forbidden?’ [The reply is that] the Lord has taught that what is forbidden may be performed by one who perceives with the eye of knowledge a special benefit for beings therein … but the foregoing [exemption] does not apply to everyone: only to [cases of] the exercise of compassion in its highest degree by one who is of a compassionate nature, without a selfish motive, solely concerned with the interests of others and totally dedicated to this [ideal].20

Prajñākaramati goes on to connect the violation of generally applicable ethical principles to the exercise of skilful means, tactical skill, or ingenuity (upāya or upāyakauśalya) – a concept that Damien Keown has associated with later forms of what he terms “Mahāyāna ethics:”
The Mahāyāna allowed monks a limited degree of flexibility … subject to the twofold stipulation that (a) the act should benefit others; and (b) it should be performed from an irreproachable (niravadya) motive. Care is taken specifically to exclude from this provision acts of a grave

18 These claims may have informed the favorable stance taken by Jñānaśrīmitra (10th century) to certain philosophical claims made by Dharmakīrti – claims that Jñānaśrīmitra reads as only partially true, but nevertheless pedagogically useful (Patil 2007). 19 Bodhicaryāvatāra 5.84cd: niṣiddham apy anujñātaṃ kṛpālor arthadarśinaḥ. Translation from Crosby and Skilton 1995: 41. 20 Trans. Keown 1992: 149–50. Bodhicaryāvatārapañjikā p. 84: prati­ ṣiddhārthe pravṛttau kathaṃ na sāpattika iti cet / na / kvacin niṣiddham api sattvārthaviśeṣaṃ prajñācakṣuṣā paśyataḥ karaṇīyatayānujñātaṃ bhaga­ vatā  / …  / tac cāpi na sarvasyāpi tu kṛpāloḥ karuṇāprakarṣapravṛttitayā tatparatantrasya parārthaikarasasya svaprayojanavimukhasya /

Tall tales, tathāgatas, and truth

83

or serious nature, and there is no suggestion that a breach of the fundamental moral precepts would be countenanced. The further development of the principle enunciated here is to be found in the notion of Skilful Means, which can, perhaps, be regarded as the outcome of an attempt to extend the exemption granted in respect of minor offences to serious offences.21

The picture Keown presents is one of change over time: Beginning with an initial reluctance to admit that certain moral precepts – those proscribing great offenses – should ever be suspended or violated, the adherents of Mahāyāna gradually shifted to a view according to which actions previously classed as great offenses may, under certain circumstances, be performed without their performer thereby incurring serious karmic debt. Keown associates the trope of skilful means with this putatively latter stage of Mahāyāna ethical reflection; upāya, used in this specific sense, thus appears on the scene as a consequence of a peculiarly Mahāyāna attempt to reassess the moral dimension of certain actions proscribed in Buddhist Vinaya literature.22 Keown’s description of the way in which the concept of upāya is deployed is surely correct for at least some passages in which the term appears. Upāya does, in certain texts, appear to constitute something like a license to commit actions that would otherwise be impermissible.23 Yet closer examination of the source texts reveals that the notion of skilful means is itself somewhat fluid: Mahāyāna texts differ in their assessment of what practices the notion of skilful means can accommodate. Indeed, when one goes back to look at a story that has become something of a locus classicus for the presupposition that skilful means affords a bodhisattva permission to lie – the Lotus Sūtra’s “Parable of the Burning House” – one finds something like the opposite view expressed. According to the
Keown 1992: 149–50. The term upāya is used in many different ways in Mahāyāna texts, as Nattier (2003: 154–6), among others, has pointed out. Cf. Harvey 2000: 135; Keown 1992: 158–60; Pye 1978: 1–17. 23 On the use of upāya in the sense above, see, for example, certain illustrative stories recounted in the Upāyakauśalyasūtra (trans. Tatz 1994: 34–5; 73–5; cf. Chang 1983: 433–4; 456–7).
22 21

84

Richard F. Nance

analysis presented in the sūtra, one cannot simultaneously lie and engage in skilful means.24

3. Returning to a house on fire
The “Parable of the Burning House” is quite well known, and for this reason I will sketch it only very schematically here.25 It is a parable placed into the mouth of Śākyamuni himself. He tells us of a father who lures his three children out of a burning house to safety by promising to give each of them different gifts once they emerge. All the children, in the end, do receive a gift – but they all receive the same thing: a gift that had previously been promised to one child, but not to the others. Thus, though all the children do receive gifts, two fail to receive what they were told they would receive – and so might well be read to have been lured from the house under false pretenses. Such an interpretation is floated in the sūtra, and quickly dismissed. Having finished his parable, Śākyamuni asks his audience – Śāriputra – whether the father in the parable should not be understood to have told a lie (mā haiva tasya puruṣasya mṛṣāvādaḥ syāt). Śāriputra answers immediately and negatively: such a man would not be a liar (sa puruṣo na mṛṣāvādī bhavet); instead, we should understand the man to have saved his children via skilful means. Note, then, that Śāriputra is not advocating the notion that lying is one form that skilful means can take. On the contrary, he is presenting the two as alternatives: one either lies, or one engages in skilful means.
24 The story is told twice, in prose and in verse, and there are interesting divergences between these two tellings, though these divergences are of little consequence to my concerns in this paper. It is, of course, true that the parable clearly serves more than one function in the sūtra. The sūtra itself encourages us to understand the parable as an allegory for the claim that apparently disparate Buddhist paths are unitary: while the Tathāgata may appear to teach many paths to liberation, he in fact teaches only one. However, this aspect of the story is not directly relevant to the points I am working to make here, and so may safely be left aside. 25 For the Sanskrit, see Vaidya 1960: 51ff. (cf. Kern and Nanjio 1908–12: 72ff.; Wogihara and Tsuchida 1934: 69ff.; Dutt 1953: 54ff.)

Tall tales, tathāgatas, and truth

85

Why, then, should we think of the father’s action toward his children as an instance of skilful means rather than lying? Śāriputra’s account of his reasoning on this issue is sketchy at best, but he works to justify his view by presenting two separable arguments. The first builds on the traditional assumption that lying involves the deliberate misrepresentation of a speaker’s intention. Śāriputra notes that the father’s intention in speaking was precisely to save his children from suffering by the use of some skilful means. The father succeeds in his aim – and thus should not be thought to have told a lie.26 This argument is supplemented by a second: Śāriputra tells us that the father should not be considered a liar, since his actions serve to benefit his children. The reason offered here would seem to be irrelevant to the question of whether the father lies, unless Śāriputra also presumes that lying is incompatible with benefit.27 If he does presume this, then we can fill in the contours of his argument quite easily: given the premises that an utterance cannot be both a lie and a source of benefit, and that a particular utterance is a source of benefit, the conclusion naturally follows that the utterance is not a lie. This argument is valid, but it is likely to strike most of us today as less than sound. It would seem that Śāriputra’s response collapses two separable issues. One has to do with what we might call the moral status of the father’s behavior: whether the father is “doing the right thing” in speaking the way he does to his sons. Śāriputra would, I think, answer this question positively: the father is doing the right

This argument is obviously specious, eliding as it does a distinction between what an utterance is about and the work that it is intended to do – a distinction that informs discussions of abhidheya and prayojana in Buddhist śāstric literature. 27 See below, section 4, and cf. Kambala, Ālokamālā, verse 37: “Even if a statement which leads to injury were to be accurate (bhūta), it would be false (mṛṣā). What is the sense of [categorizing statements] as true or untrue? That [statement] which brings benefit to others is true!” (bhūtam apy upaghātāya yad uktaṃ syān mṛṣaiva tat  / satyāsatyena ko ’rthārthas tat satyaṃ yat parārthakṛt //)

26

86

Richard F. Nance

thing, because his actions benefit his sons.28 So the father’s actions ought to be affirmed. The other issue has to do with descriptive adequacy: how best to capture what sort of “right thing” the father is doing. To describe the father’s speech as an instance of mṛṣāvāda is to commit oneself to the view that certain instances of mṛṣāvāda ought to be affirmed. But to affix the label of mṛṣāvāda onto an utterance is already to impugn it – and thus render such affirmation impossible. It may seem that what we are faced with here is simply an instance of the privileged lie under another name. And perhaps we are – but the attempt at redescription is itself revealing. It suggests that those responsible for the composition of the Lotus Sūtra felt a degree of discomfort with the notion that upāya is finally compatible with mṛṣāvāda: that skilful means can, in fact, be reconciled with the practice of lying. This looks to be a rather different view from the one expressed in the Bodhisattvabhūmi passage cited above – a passage that appears rather more relaxed about claiming that lies can and should, in certain circumstances, be told.29 Both of these views may be mingled in a single text. Consider, for example, the following passage, drawn from the Mahāyāna Mahāparinirvāṇasūtra:
O good man! It is because of all beings that although the Tathāgata knows all things, he says he does not know; although he sees all things, he says he does not see. Why? Because the Tathāgata clearly sees the capacities of individual beings. O good man! Although the Tathāgata speaks in this way, he does not lie. Why not? Because lies involve faults (skyon, *doṣa). How could the Tathāgata lie, being completely free of all blameworthy faults? Although the Tathāgata does not lie, yet in certain cases he may lie for the purpose of benefiting sentient

28 Whether Śāriputra’s view ought to be branded a species of consequentialism is an issue that I will leave to others to debate. 29 Note, however, that the terms used shift over the course of the paragraph, so that explicit reference to the practice of lying drops out. Bodhisattvas do not deliberately tell lies, but a bodhisattva may focus on an idea and deliberately speak in another way (anyathāvācaṃ bhāṣate) in order to bring benefit to others.

Tall tales, tathāgatas, and truth

87

beings by means of the dharma: he teaches and speaks to them with skilful means, as appropriate.30

Two prima facie incompatible assumptions are voiced here. According to the first, lying constitutively involves fault. Hence, the Tathāgata cannot lie, since to admit that he could lie would be to admit that he possesses a trace of fault. According to the second, lying does not constitutively involve fault. Hence, the Tathāgata may lie; in doing so, he practices skilful means. The passage thus immediately juxtaposes both of the standpoints toward lying discussed above; having done so, it appears in the end to accept the notion that the idea of skilful means can accommodate mṛṣāvāda. Against Śāriputra’s assessment in the Lotus Sūtra, this passage strongly suggests that lying is one form that skilful means can take. 31 Mahāyāna sūtra texts would thus appear to differ on the question of whether upāya and mṛṣāvāda can be reconciled. For this reason (among others), we should be cautious about presuming there to be a single “Mahāyāna” attitude to lying – and, by extension, a single
rigs kyi bu de bzhin gshegs pa ni sems can thams cad kyi phyir chos thams cad mkhyen kyang mi mkhyen ces gsung ngo / chos thams cad gzigs kyang gzigs pa med do zhes gsung / … / de ci phyir zhe na / de bzhin gshegs pas ni sems can so so’i dbang po gsal bar gzigs pa’i phyir ro / rigs kyi bu de bzhin gshegs pa ni de skad gsung yang brdzun ma yin no / de ci phyir zhe na / brdzun du smra ba ni skyon dang bcas pa yin te / de bzhin gshegs pa ni nyes pa’i skyon thams cad yongs su bral ba yin na brdzun du smra ba ga la zhig yod / rigs kyi bu de bzhin gshegs pa la brdzun du gsung ba med mod kyi / ji ste sems can dag la brdzun du smra ba’i rkyen gyi / chos kyis phan pa’i don du ’gyur na / ci rigs pa’i thabs kyis de la ston cing gsung ngo // 31 Interestingly, the parable of the burning house is also invoked in the Mahāyāna Mahāparinirvāṇasūtra, though in a highly abbreviated and rather different form – one in which the topic of skilful means is not broached: “For example: a fire rages in the house of a householder, and he emerges from the house. Yet his sons do not escape, remaining inside the house. Then the householder, though unquestionably aware of the conflagration, enters into the house in order to extract his children. It is the same with the BodhisattvaMahāsattva.” (dper na khyim bdag cig khyim du me shor na khyim de nas phyir rol tu ’byung ngo / ji ste khyim bdag de’i bu rnams khyim gyi nang du lus te me las ma thar na / khyim bdag de de’i tshe na gdon mi za bar mes tshig par ’gyur bar shes kyang bu rnams gdon pa’i phyir khyim de’i nang du ’jug go / byang chub sems dpa’ sems dpa’ chen po yang de dang ’dra ste /)
30

88

Richard F. Nance

Mahāyāna ethics (or single “later” Mahāyāna ethics).32 In fact, the Indian Buddhist textual corpus presents a range of views regarding the permissibility of lies. The view expressed in a particular text is not inferable from the knowledge that the text is classed as Mahāyāna. And, importantly, the same point holds true for nonMahāyāna texts as well. In the texts of the Pāli canon, for example, divergent views on the subject of the privileged lie do occasionally surface, though such views are sometimes implied rather than stated outright.

4. Buddhas as perfected speakers: on truth and benefit in Pāli texts
As is well known, texts of the Pāli canon repeatedly affirm the Buddha’s truthfulness. Yet it is also suggested in various places that to speak truthfully does not entail that one has thereby spoken wholesomely. In a passage preserved in the Aṅguttara Nikāya, Gotama advises that a person should not speak of what he or she has seen, heard, sensed or understood, if doing so should cause unwholesome (akusala) states to increase and wholesome (kusala) states to decrease. Apparently, the truth can hurt – and hurt in ways that impede the practice of the path. In such circumstances, one is better off saying nothing.33 On a first reading, this sentiment might appear to be similar to one expressed in Manusmṛti 4.138 – a verse that stipulates what it calls an “eternal dharma” (sanātano dharmaḥ) concerning appropriate brahmanical speech. A brahmin, the text tells us,
Shall say what is true (satya); and he shall say what is agreeable (pri­ ya). He shall not say what is true, but disagreeable; nor shall he say what is agreeable, but wrong. This dharma is eternal.34

Cf. Silk 2002. AN ii.172–3 (Vassakārasutta). 34 Jha 1999, Vol. 1: 378: satyaṃ brūyāt priyaṃ brūyān na brūyāt satyam apriyam / priyaṃ ca nānṛtaṃ brūyād eṣa dharmaḥ sanātanaḥ.
33

32

Tall tales, tathāgatas, and truth

89

However, the prima facie similarity here is, I think, misleading. What is at stake in the passage from the Aṅguttara Nikāya is not the agreeability of an utterance, but the wholesome effects it brings in its wake: the benefit it confers.35 And these two things are not the same, as the tradition acknowledges in the following rather complex passage of the Abhayarājakumārasutta.
Now on that occasion a young tender infant was lying prone on Prince Abhaya’s lap. Then the Blessed One said to Prince Abhaya: ‘What do you think, prince? If, while you or your nurse were not attending to him, this child were to put a stick or a pebble in his mouth, what would you do to him?’ ‘Venerable sir, I would take it out. If I could not take it out at once, I would take his head in my left hand, and crooking a finger of my right hand, I would take it out even if it meant drawing blood. Why is that? Because I have compassion for the child.’ ‘So, too, prince, such speech as the Tathāgata knows to be untrue (abhūta) incorrect (ataccha), and unbeneficial (anatthasaṃhitaṃ), and which is also disagreeable (appiya) and unwelcome (amanāpa) to others: such speech the Tathāgata does not utter. Such speech as the Tathāgata knows to be true and correct but unbeneficial, and which is also disagreeable and unwelcome to others: such speech the Tathāgata does not utter. Such speech as the Tathāgata knows to be true, correct, and beneficial, but which is also disagreeable and unwelcome to others: the Tathāgata knows the time to use such speech. Such speech as the Tathāgata knows to be untrue, incorrect, and unbeneficial, but which is agreeable and welcome to others: such speech the Tathāgata does not utter. Such speech as the Tathāgata knows to be true and correct but unbeneficial, and which is agreeable and welcome to others: such speech the Tathāgata does not utter. Such speech as the Tathāgata knows to be true, correct, and beneficial, and which is agreeable and welcome to others: the Tathāgata knows the time to use such speech. Why is that? Because the Tathāgata has compassion for beings.’36

Cf. the discussion of truth (bhūta / taccha) and benefit (atthasaṃhita) pertaining to covert speech (rahovāda) at MN 139 (Araṇavibhaṅgasutta), iii.234 (Ñāṇamoli and Bodhi 1995: 1083–4). 36 Translation modified from Ñāṇamoli. MN 58, i.394–395: Tena kho pana samayena daharo kumāro mando uttānaseyyako abhayassa rājakumārassa aṅke nisinno hoti. Atha kho bhagavā abhayaṃ rājakumāraṃ etadavoca:

35

90

Richard F. Nance

In this passage, benefit is explicitly distinguished from both agreeability and truth. An utterance’s agreeability (or lack thereof) is presented as irrelevant to the question of whether the utterance is appropriate for a buddha. The salient issue is rather one of benefit: if an utterance is beneficial, a buddha will utter it, whether or not it is also agreeable (and it may not be). Agreeability and benefit are thus presented here as wholly autonomous notions (i.e., the presence of either one implies nothing about the presence of the other). The same autonomy is presented as characterizing agreeability and truth as well. Yet the text’s stance on the conceptual relation between benefit and truth is left somewhat murky. There are four possibilities: An utterance might be untrue and unbeneficial (in which case, a buddha will not utter it); alternatively, an utterance might be true and unbeneficial (in which case, again, a buddha will not utter it); an utterance might be true and beneficial (in which case, a buddha will utter it). Finally, an utterance might be untrue and beneficial. The latter would seem, at least, to be a logical possibility – but it is a possibility that is conspicuously absent from the text; Gotama does not so much as consider it.
taṃ kimmaññasi rājakumāra, sacāyaṃ kumāro tuyhaṃ vā pamādamanvāya dhātiyā vā pamādamanvāya kaṭṭhaṃ vā kaṭṭhalaṃ vā mukhe āhareyya, kinti naṃ kareyyāsīti āhareyy’ assāhaṃ bhante. Sace ahaṃ bhante na sakkuṇeyyaṃ ādiken’ eva āhattuṃ vāmena hatthena sīsaṃ pariggahetvā dakkhiṇena hatthena vaṅkaṅguliṃ karitvā salohitampi āhareyyaṃ. Taṃ kissa hetu: atthi me bhante kumāre anukampā ti. Evameva kho rājakumāra, yaṃ tathāgato vācaṃ jānāti abhūtaṃ atacchaṃ anatthasaṃhitaṃ, sā ca paresaṃ appiyā amanāpā, na taṃ tathāgato vācaṃ bhāsati. Yampi tathāgato vācaṃ jānāti bhūtaṃ tacchaṃ anatthasaṃhitaṃ. Sā ca paresaṃ appiyā amanāpā, tampi tathāgato vācaṃ na bhāsati. Yañca kho tathāgato vācaṃ jānāti bhūtaṃ tacchaṃ atthasaṃhitaṃ sā ca paresaṃ appiyā amanāpā, tatra kālaññū tathāgato hoti tassā vācāya veyyākaraṇāya. Yam tathāgato vācaṃ jānāti abhūtaṃ ataachaṃ anatthasaṃhitaṃ sā ca paresaṃ piyā manāpā, na taṃ tathāgato vācaṃ bhāsati. Yampi tathāgato vācaṃ jānāti bhūtaṃ tacchaṃ anatthasaṃhitaṃ. Sā ca paresaṃ piyā manāpā, tampi tathāgato vācaṃ na bhāsati. Yañca kho tathāgato vācaṃ jānāti bhūtaṃ tacchaṃ atthasaṃhitaṃ. Sā ca paresaṃ piyā manāpā, tatra kālaññū tathāgato hoti tassā vācāya veyyākaraṇāya. Taṃ kissa hetu: atthi rājakumāra tathāgatassa sattesu anukampā ti.

Tall tales, tathāgatas, and truth

91

Why is this possibility ignored? It is hazardous to engage in speculation here, but the passage from the Manusmṛti quoted above suggests one possible rationale. In the Manusmṛti passage, there are likewise four possibilities, only three of which are considered explicitly. (It can, it would seem, go without saying that a brahmin should not utter speech that is both disagreeable and untrue). Is something like the same thing happening in the Abhayarājakumārasutta? Is the notion of a beneficial untruth, or of a buddha’s uttering beneficial untruths, so obviously untenable that it can safely be passed over in silence? If so, then the sutta would appear to present a view of benefit and truth that sees them as distinct but not wholly autonomous notions. In keeping with the view advocated in the Aṅguttara Nikāya passage discussed earlier, truth does not imply benefit – but benefit does imply truth.37 When considering whether an utterance is appropriate for a buddha, all one needs to ask is whether the utterance benefits others; if so, it is ipso facto true.

Cf. the following passage from the Mahāyāna Mahāparinirvāṇasūtra: “All the world loves right speech. I never utter words that are untimely, contrary to the dharma, unbeneficial and non-virtuous. O good man! Certain words are also fundamentally harsh and false, in addition to being untimely, contrary to the dharma, disagreeable to one who hears them, unbeneficial and non-virtuous. I never speak such words either. O good man! Certain words, even though they may be fundamentally harsh, are true and not false; they are timely, concordant with the dharma and bring benefit and virtue to all sentient beings, even though they are disagreeable to one who hears them. I speak such words. And why is that? Because the Buddha, the Blessed One, possesses knowledge: because he knows skilful means.” (yang dag par smra ba la ni ’jig rten pa thams cad ’dod do / dus ma yin pa dang chos ma yin pa dang / phan pa dang bde bar mi ’gyur ba’i tshig ni nam du yang mi gsung ngo / rigs kyi bu tshig kha cig ni shin tu brlang zhing rtsa ba la brdzun pa yang yin te / dus ma yin pa dang / chos ma yin pa dang / gang gis thos kyang mi ’dod pa dang / phan pa dang / bde bar mi ’gyur ba ni ngas nam du yang ma gsungs so / rigs kyi bu tshig tu gsung ba kha cig ni brlang zhing rtsa ba kyang yang dag pa mi brdzun pa ste / dus dang ldan pa / chos dang ldan pa / sems can thams cad la phan pa dang bde ba ste / thos pas mi dga’ bar ’gyur ba yang nges par gsung ngo / de ci phyir zhe na / sangs rgyas bcom ldan ’das ni mkyhen pa dang ldan pa ste / shin tu thabs mkhas pa dang ldan pa’i phyir ro //).

37

92

Richard F. Nance

Perhaps something like this assumption underlies Śāriputra’s peculiar response to Śākyamuni’s question regarding the protagonist of the Parable of the Burning House. Did the father lie? No: given that he benefited his sons, he could not have lied; the very fact that he benefited his sons entails that he told the truth. In distinguishing between benefit and truth and emphasizing the salience of the former in deciding on what is and is not appropriate speech for buddhas, the Abhayarājakumārasutta arguably makes room for a certain ambivalence toward truth per se.38 This ambivalence is also arguably present in the Pāli jātaka literature, to which we now turn.

5. Deception in the Pāli jātaka literature
In Richard Robinson’s well-known and widely used introduction to Buddhism The Buddhist Religion, originally authored in 1972 and now in its fifth edition (having been substantially revised and re-titled as Buddhist Religions), one finds the claim made – and made repeatedly (Robinson et al. 2005: 69, 113) – that although the Pāli jātaka tales present occasions in which the bodhisattva is shown violating various precepts, he is never portrayed as violating the precept against lying.39 If this claim were true, then the jātakas would appear to be the wrong place to look for privileged lies. But the claim is misleading at best, as becomes apparent upon close investigation of the relevant literature – a corpus of texts that comprises not only the skeletal canonical jātaka verses themselves, but the paracanonical Jātakaṭṭhakathā commentary that elaborates the stories traditionally assumed to surround them.

38 Cf. Bhāviveka’s insistence, in the Tarkajvālā, that “[t]he Blessed One seeks the welfare of the world, so he does not always favor reality,” transl. Eckel (2008: 199). The corresponding Tibetan (Eckel 2008: 376) reads: bcom ldan ’das kyi rtsom pa ni ’jig rten la phan pa yin pa’i phyir yin pas de kho na nyid mchog tu ’dzin par mi mdzad. 39 This claim should probably not be attributed to Robinson himself, as it is not present in the first edition of Robinson’s text, but appears to have been introduced in a subsequent edition.

Tall tales, tathāgatas, and truth

93

Here, I want to consider two of these stories. One – the Dumme­ dhajātaka – has been largely ignored by modern scholarship, despite the fact that a story quite similar to it is among those anthologized in the well-known Jātakamālā of Āryaśūra.40 The other – the Vessantarajātaka – needs no introduction to scholars of Buddhism; it is perhaps the most widely known and beloved jātaka tale throughout the Buddhist world.41 In the Dummedhajātaka, the bodhisatta is born as a prince named Brahmadatta. As the story opens, he is 16 years of age. Educated at Takkasilā, he is properly versed in the various branches of brahmanical learning, and is well aware of the protocols for Vedic sacrifice. He is also disturbed by these protocols, insofar as they demand the killing of animals. So he makes a vow: when he comes to power, he will, without harming a single being, cause this killing to stop by means of a stratagem (or a trick: upāyena).42 One day, Brahmadatta is outside the city on his chariot when he notices a crowd of people gathered around a tree, making offerings to the devatā dwelling there. Brahmadatta descends from his chariot and does the same. As time passes, he continues to return to the tree and engages in pūjā “like a worshipper.”43 When he finally attains sovereignty, he calls together the brahmins and householders of the region to remind them of his practice of worshipping the tree-devatā. He then tells them that he has vowed to offer a sacrifice to the tree on the occasion of his becoming king, and that he will need their assistance in preparing the sacrifice. The sacrifice is to consist in the flesh and blood of one thousand people – specifiOn the version of the story preserved in the Jātakamālā, see below, note 44. The title Dummedhajātaka is given to two distinct stories preserved in the Pāli jātaka corpus; the story under consideration here may be found in Jātaka, Vol. 1: 259–6; trans. Cowell et al. 1895–1907, Vol. 1: 128–8. 41 On the popularity of the Vessantarajātaka, see Cone and Gombrich 1977: xv; Collins 1998: 497–8. 42 Jātaka, Vol. 1: 259: ahaṃ pitu accayena rajjaṃ labhitvā ekam pi akilametvā upāyen’ eva pāṇavadhaṃ kātuṃ na dassāmīti. Interestingly, in the version of the tale presented in Āryaśūra’s Jātakamālā, no reference is made to upāya. 43 Ibid.: 259–60: devatāmaṃgaliko viya pūjaṃ karoti.
40

94

Richard F. Nance

cally, those who engage in the five or ten immoral actions. He orders the assembled brahmins and householders to make known that anyone who behaves immorally will be killed and offered to the devatā. They do so, striking fear into everyone. Duly cowed, the community halts the practice of harming living beings and committing any other infraction. Thus, the bodhisatta fulfills the terms of his original vow. This story offers a rather disturbing portrait of righteous kingship. Brahmadatta does succeed in halting the killing of animals, but he does so by enacting a reign of terror. The bodhisatta’s strategem – his upāya – appears to involve protracted play-acting at pūjā, in order to set up the conditions under which his subjects will take seriously the issuance of what amounts to a zero-tolerance policy against ethical infraction, punishable by death. While we, as readers, know that Brahmadatta does not approve of the killing of animals or the harming of human beings, Brahmadatta’s subjects do not. In order for the stratagem to be effective, they need to believe that they will be killed if they behave immorally – even if, given the objections to violence Brahmadatta raises in his vow, no killing will in fact occur. Brahmadatta is thus plausibly read as lying to his subjects. The lie is portrayed as facilitating an end that the tradition affirms – but it is no less a lie for that.44
44 In Āryaśūra’s version of the tale – preserved under the title Yajñajātaka (Kern 1891: 67–73 = Vaidya ed.: 70–6; trans. Speyer 1895: 93–104; Khoroche 1989: 74–80) – the bodhisattva, born as a king, faces the problem of drought, and seeks advice on what to do from senior brahmins. Not surprisingly, they tell him that he needs to prepare a sacrifice such as those described in the Veda – a sacrifice in which many animals will need to be slaughtered. On hearing this answer, the king is appalled; his advisors appear to him “pathetically weak-minded and gullible, unquestioning in their faith and blindly devoted to tradition” (Khoroche 1989: 75). Yet, feigning eagerness to undertake the sacrifice (yajñārambhasamutsuka iva nāma), he agrees to the proposal – with one modification: instead of an animal sacrifice, he will sponsor a human sacrifice (puruṣamedha) in which one thousand victims will be killed. He then calls together the populace of his kingdom and tells them that they should watch what they do, since he will be watching them: he plans to circulate spies around the kingdom who are “sharp-eyed, tireless, and alert.” Persons whom these spies observe behaving unethically will be arrested and added to the pool of sacrificial victims.

Tall tales, tathāgatas, and truth

95

Turning to the Vessantarajātaka, one finds an apparent lie portrayed as facilitating the bodhisatta’s perfected virtue of selfless generosity. The passage in question occurs just after Vessantara’s wife Maddī awakens from a harrowing dream, and approaches Vessantara to relate it and ask him what it means. Cone and Gombrich translate the passage as follows:
She went to the leaf hut of the Great Being and knocked at the door … when she had told him [the dream], just as she had experienced it, the Great Being understood the dream, and knew that he would fulfill the Perfection of giving, and that a suppliant would on the next day come and beg his children from him. He decided to console Maddī and sent her away. ‘Your mind must have been agitated because you were lying uncomfortably, or because of something you had eaten, Maddī. Do not be frightened.’ So he deceivingly consoled her (mohetvā assāsetvā), and sent her away.45

We do not find the compound musāvāda invoked in this passage – and its absence prompts questions. Vessantara has certainly misled Maddī – but has he lied to her? Perhaps not: Maddī’s dream could
Agents are dispatched to monitor the public, and repeated announcements of the impeding sacrifice are made: the population knows that it is being watched very carefully, that a zero-tolerance policy is in effect, and that the consequence of immoral behavior is death. As a result, the text tells us, “all inclination to misbehave left them, and instead they were all eager to vow themselves to a life of virtue. In their readiness to love and respect one another, they turned their backs on petty feuds. Quarrels and disputes ceased, and they abided by their elders’ decisions. Sharing became commonplace, and hospitality too. They took pride in behaving with politeness and modest reserve. It was as though they were living in the Golden Age … no one lapsed at all” (Khoroche 1989: 77–8). The drought is thus brought to an end, the crops are restored, and everyone lives happily ever after. Because he has so successfully modified his subjects’ behavior, the king never has to hold the human sacrifice: he has accomplished a sacrifice without bloodshed – a sacrifice according to the law (dharmayajña). 45 Cone and Gombrich 1977: 54. Cf. Jātaka, Vol. 6, p. 541: paṇṇasālaṃ gantvā Mahāsattassa paṇṇasāladvāraṃ ākoṭesi … sā attanā diṭṭhaniyāmen’ eva kathesi, M. supinam parigaṇhitvā ‘mahyaṃ dānapāramī pūrissati, sve mahyaṃ yācako āgantvā putte yāccissatīti, Maddiṃ assāsetvā uyyojessāmīti’ cintetvā ‘Maddi tava dussayanadubbhojanehi cittaṃ āluḷitaṃ bhavissati, mā bhāyīti’ mohetvā assāsetvā uyyojesi.

96

Richard F. Nance

conceivably have been caused by one or both of by the very conditions that Vessantara identifies as potentially prompting it. Even if the dream turns out to portend future events (as it does), it could indeed have been caused by an awkward physical posture, or by something she ate – and Vessantara might be aware of this. So, in telling Maddī not to worry – that it was just something she ate – Vessantara could conceivably be offering Maddī an accurate account of the antecedent material conditions that prompted the occurrence of her dream. This, however, seems a stretch. Given the scenario as presented here, it seems much more natural to take Vessantara as deliberately misrepresenting how he takes things to be. Vessantara has lied. In doing so, he has not engaged in an action that is ever explicitly identified as blameworthy. The deception of Maddī is not mentioned again – but its effects are important to the narrative: Maddī, having been deceitfully consoled, leaves the children alone with Vessantara the next day, and Vessantara gives them away – thereby fulfilling the perfection of giving.46

6. Concluding reflections
The above remarks barely scratch the surface of some of the issues raised in Buddhist doctrinal texts – and those familiar with such texts will note also certain complexities have been avoided altogether (e.g., the claim, made in the Laṅkāvatārasūtra, the Mūlamadhyamakakārikā and elsewhere, that buddhas do not in fact speak at all). But I hope that the material presented here has been sufficient to show that the sweeping claim that has served as the paper’s stalking horse – the claim that Buddhism (or, more precisely, normative Buddhist doctrine) has no notion of the privileged lie – needs to be seriously reconsidered. So, too, does the claim that the bodhisattva is never portrayed in the jātaka literature as violating the precept against lying. Both of these claims are false. Buddhist doctrine does not, in fact, advocate a uniform stance on
Interestingly, the versions of the story presented in Kṣemendra’s Avadānakalpalatā (Vol. I: 172–5) and Āryaśūra’s Jātakamālā (55–69) omit any reference to Maddī’s dream and her subsequent deception.
46

Tall tales, tathāgatas, and truth

97

the permissibility of lying – nor is a uniform stance on lying implied by the actions of the bodhisatta in the jātakas. One finds, instead, different views stated and implied in the literature – and, as we have seen, such differences emerge even when one confines oneself to the canonical texts of a “single” Buddhist tradition. Lies, though generally discouraged by the tradition, do appear to be presented as blameless in certain circumstances. Despite the prima facie absolutism of certain Vinaya precepts, Buddhist texts suggest that the moral status of lying can be read as varying from case to case. This is a substantive point – but there is also a methodological point at stake here that, although it has been raised before, is worth raising again. Over the past decade, the field of Buddhist studies has seen an explosive growth in searchable e-texts. With the click of a mouse, I can know that the term musā- occurs 653 times in the Pāli canon. This is a salutary development for research in the field – but in addition to the rewards it brings, it also carries with it certain risks. As noted above, certain canonical and paracanonical Pāli texts portray the bodhisatta as engaging in intentional deception (verbal and otherwise). None of these instances is signaled by the term musā-; each would be missed in even the most comprehensive e-text survey of the term and its cognates. For all of the benefits provided by e-texts – and they are considerable – they cannot substitute for the hard work of reading through, and thinking with, these texts. This is especially true if one is interested in detecting conceptual undercurrents: ideas that are not explicitly acknowledged topoi for the tradition. The privileged lie would appear to be one such idea; there is no Sanskrit or Pāli term for it, and this fact has perhaps contributed to the elusiveness of the idea under the philological gaze. Without downplaying the importance of rigorous philological work (we could get nowhere without it), we should, I think, heed the recommendation of the texts themselves, and continually remember to attend not only to their phrasing (vyañjana), but also to their meaning (artha).

98

Richard F. Nance

Abbreviations and bibliography General abbreviations
D – The sde dge mtshal par bka’-’gyur and bstan-’gyur. Delhi: Delhi Karmapae Chodhey Gyalwae Sungrab Partun Khang. Consulted in digital format via the Tibetan Buddhist Resource Center: (http://www.tbrc. org/#library_work_Object-W22084, last visited 07-09-2011) and (http:// www.tbrc.org/#library_work_Object-W23703, last visited 07-09-2011). G – Bstan ’gyur gser gyi lag bris ma. Tianjing. Consulted in digital format via the Tibetan Buddhist Resource Center: (http://www.tbrc.org/#library_ work_Object-W23702, last visited 07-09-2011). Mahāvyutpatti – Sakaki, R. 1916. The Mahāvyutpatti, a Buddhist Dictionary Originally Translated into Sanskrit and Tibetan, Later into Chinese and Japanese. Kyōto: Shingonshū Kyōto Daigaku. Consulted in digital format, in an edition prepared by Mitsuhara, H. et al., eds. 2002. Mahāvyutpatti (http://texa.human.is.tohoku.ac.jp/aiba/archive/mvyut/open/, last visited 07-09-2011).

Primary sources in Pāli
AN – Aṅguttara-Nikāya. Morris, R. (Vols. 1–2) and E. Hardy (Vols. 3–5), eds. 1961–81 [1885–1900]. The Aṅguttara Nikāya. 5 Vols. London: Pali Text Society. Dhammapada – Von Hinüber, O. and K. R. Norman, eds. 1994–5. Dhammapada. London: Pali Text Society. Iti – Windisch, E., ed. 1975 [1889]. Itivuttaka. London: Pali Text Society. Jātaka / Jātakaṭṭhakathā – Fausbøll, V., ed. 1990–1 [1877–96]. The Jātaka Together with its Commentary, Being Tales of the Anterior Births of Gotama Buddha. 6 Vols. London: Pali Text Society. MN – Majjhima-Nikāya. Trenckner, V. et al., eds. 1991–94 [1888–1902]. The Majjhima Nikāya. 4 Vols. London: Pali Text Society. Milindapañha – Trenckner, V., ed. 1986 [1880]. Milindapañha. London: Pali Text Society. SN – Samyutta-Nikāya. Feer, M. L., ed. 1973–80 [1884–1904]. The Samyutta Nikāya of the Sutta-piṭaka. 6 Vols. London: Pali Text Society. Sumaṅgalavilāsinī – Rhys Davids et al., eds. 1968–71 [1929–32]. Dīghanikāya Commentary: Sumaṅgalavilāsinī. 3 Vols. London: Pali Text Society. Sn – Suttanipāta. Andersen, D. and H. Smith, eds. 1990 [1913]. The Suttanipata. London: Pali Text Society.

Tall tales, tathāgatas, and truth

99

Primary sources in Sanskrit
Abhidharmakośa and -bhāṣya (Vasubandhu) – Dwārikādās Śāstrī, S. ed. 1998. Abhidharmakośabhāṣya. 2 Vols. Varanasi: Bauddha Bharati Series 5–8. Avadānakalpalatā (Kṣemendra) – Vaidya, P. L., ed. 1959. AvadānaKalpalatā. 2 Vols. Darbhanga: The Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning. Ālokamālā (Kambala) – Lindtner, C., ed. and trans. 2003. A Garland of Light: Kambala’s Ālokamālā. Fremont: Asian Humanities Press. Bodhicaryāvatāra (Śāntideva) – Dwārikādās Śāstrī, S., ed. 2001. Bodhi­ caryāvatāra. Varanasi: Bauddha Bharati Series 21. Bodhicaryāvatārapañjikā (Prajñākaramati) – Vaidya, P. L., ed. 1960. Bodhi­ caryāvatāra of Śāntideva with the commentary Pañjikā of Prajñākaramati. Darbhanga: The Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning. Bodhisattvabhūmi (attributed to Asaṅga) – Dutt, N., ed. 1966. Bodhisattva­ bhūmi: Being the XVth Section of Asaṅgapada’s Yogacarabhumi [sic]. Patna: Jayaswal Reseach Institute. Caturaśītisiddhapravṛtti (Abhayadattaśrī) – Sanskrit lost. Tibetan: not incorporated in D; G 3090 rgyud ’grel lu 1b1–102a6. Dharmapada – Brough, J., ed. and trans. 1962. The Gāndhārī Dharmapada. London: Oxford University Press. Dharmasamuccaya (Avalokitasiṃha) Ch. 12 – Lin, Li-Kouang et al., ed. and trans. 1969. Dharmasamuccaya: Compendium de la Loi. Recueil de Stances Extraites de Saddharmasmṛtyupasthāna-sūtra par Avalokitasiṃha, Vol. 2 (Annales du Musée Guimet 68). Paris. Jātakamālā (attributed to Āryaśūra) – Vaidya, P. L., ed. 1959. Jātakamālā. Darbhanga: The Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning. Mahābhārata (Ādiparvan) – Sukthankar, V. S. et. al., eds. 1933. The Mahābhārata; for the first time critically edited. Vol. 1: Ādiparvan. Poona: Bhandarkar Oriental Research Institute. Mahāparinirvāṇasūtra – Sanskrit fragments in Yuyama, A., ed. 1981. Sanskrit Fragments of the Mahāyāna Mahāparinirvānasūtra 1: Koyasan Manuscript. Studia Philologica Buddhica, Occasional Paper Series, IV. Tokyo: The Reiyukai Library; Bongard-Levin, G. M., ed. 1986. New Sanskrit Fragments of the Mahāyāna Mahāparinirvāṇasūtra (Central Asian Manuscript Collection at Leningrad). Tokyo: International Institute for Buddhist Studies. Tibetan: D 119 mdo sde nya 1b1–ta 339a7. Prātimokṣasūtra – Bannerjee, A., ed. 1977. Two Buddhist Vinaya Texts in Sanskrit: Prātimokṣa Sūtra and Bhikṣukarmavākya. Calcutta: World Press Private Limited.

100

Richard F. Nance

Ratnagotravibhāga (attributed to Asaṅga/Maitreya) – Prasad, H. S., ed. 1997 [1991]. The Uttaratantra of Maitreya. Delhi: Sri Satguru Publications. Saddharmapuṇḍarīkasūtra – Vaidya, P. L., ed. 1960. Saddharmapuṇḍarīkasūtra. Darbhanga: The Mithila Institute of Post-Graduate Studies and Research in Sanskrit Learning; Kern, H. and B. Nanjio, eds. 1908– 12. Saddharmapuṇdarīka. Bibliotheca Buddhica 10. St. Petersburg; Wogihara, U. and C. Tsuchida, eds. 1934–5. Saddharmapuṇḍarīkasūtram = Kaitei Bonbun Hokekyō (3 Vols). Tokyo: Seigo-Kenkyūkai; Dutt, N., ed. 1953. Saddharmapuṇḍarīkasūtram, with N. D. Mironov’s readings from Central Asian MSS. Calcutta: Asiatic Society. Tarkajvālā (Bhāviveka) – Sanskrit lost. Critical edition of Chapters 4 and 5 in Tibetan translation in Eckel, M. D., ed. 2008. Bhāviveka and His Buddhist Opponents (Harvard Oriental Series, Vol. 70). Cambridge: Harvard University Press. Udānavarga – Bernhard, F., ed. 1965. Udānavarga. 2 vols. Göttingen: Vandenhoeck and Ruprecht. Upāyakauśalyasūtra – Sanskrit lost. Tibetan: D 261 mdo sde za 283b2– 310a7, D 82 dkon brtsegs cha 30a1–70b7.

Secondary sources
Bodhi, B., trans. 1978. The Discourse on the All­Embracing Net of Views: The Brahmajālasutta and its Commentaries. Kandy: Buddhist Publication Society. Bodhi, B., trans. 2000. The Connected Discourses of the Buddha: A Translation of the Samyutta Nikāya. Somerville, MA: Wisdom Publications. Chang, G., et al., trans. 1983. A Treasury of Mahāyāna Sūtras: Selections from the Mahāratnakūtasūtra. University Park: Pennsylvania State University Press. Collins, S. 1998. Nirvāṇa and Other Buddhist Felicities: Utopias of the Pali Imaginaire. Cambridge: Cambridge University Press. Cone, M. and R. Gombrich, trans. 1977. The Perfect Generosity of Prince Vessantara. Oxford: Oxford University Press. Cowell, E. B. et al., eds. and trans. 1895–1907. The Jātaka: or, Stories of the Buddha’s Former Births. 6 Vols. London: Pali Text Society. Crosby, K. and A. Skilton, trans. 1995. The Bodhicaryāvatāra (Oxford World Classics). Oxford: Oxford University Press. Derrett, J. D. M. 2006. “Musāvāda-virati and ‘privileged lies’.” Journal of Buddhist Ethics 13: (http://blogs.dickinson.edu/buddhistethics/ files/ 2010/ 04/derrett-article.pdf, last visited 07-09-2011). Hahn, M., ed. 1982. Nāgārjuna’s Ratnāvalī, Volume 1: The Basic Texts

Tall tales, tathāgatas, and truth

101

(Sanskrit, Tibetan, Chinese). Bonn: India et Tibetica Verlag. Harvey, P. 2000. An Introduction to Buddhist Ethics. Cambridge: Cambridge University Press. Hopkins, J., ed. and trans. 1998. Buddhist Advice for Living and Liberation: Nāgārjuna’s Precious Garland. Ithaca: Snow Lion. Horner, I. B., trans. 1969 [1963–4]. Milinda’s Questions. 2 Vols. London: Pali Text Society. Jha, G., trans. 1999. Manusmṛti with the Mahubhāṣya of Medhātithi. 10 Vols. Delhi: Motilal Banarsidass. Keown, D. 1992. The Nature of Buddhist Ethics. Houndmills: Macmillan. Kern, H. 1891. The Jātaka-mālā: Stories of Buddha’s Former Incarnations, Otherwise Entitled Bodhisattva-avadāna-mālā, by Ārya-çūra (Harvard Oriental Series, Vol. 1). Cambridge: Harvard University Press. Khoroche, P. 1989. Once the Buddha Was a Monkey. Chicago: University of Chicago Press. Müller, F. 1883. India: What Can it Teach Us? London: Longmans, Green, and Company. Ñāṇamoli, B. and B. Bodhi, trans. 1995. The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya. Boston: Wisdom Publications. Nattier, J. 2003. A Few Good Men. Honolulu: University of Hawaii Press. Pachow, W. 2000. A Comparative Study of the Prātimokṣa on the Basis of Its Chinese, Tibetan, Sanskrit, and Pāli Versions. Delhi: Motilal Banarsidass. Patil, P. 2007. “Dharmakīrti’s White Lie – Philosophy, Pedagogy, and Truth in Late Indian Buddhism” In Kellner, B. et al. 2007. Pramāṇakīrtiḥ: Papers Dedicated to Ernst Steinkellner on the Occasion of his 70 th Birthday. 2 Vols. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien. Vol. 2, pp. 597–619. Pye, M. 1978. Skilful Means: A Concept in Mahāyāna Buddhism. London: Duckworth. Robinson, R. 1972. The Buddhist Religion: An Historical Introduction. Belmont, CA: Dickinson Publishing. Robinson, R. et al. 2005. Buddhist Religions: A Historical Introduction (Fifth Edition). Belmont, CA: Wadsworth. Silk, J. 2002. “What, if Anything, is Mahāyāna Buddhism? Problems of Definitions and Classifications” Numen 49, pp. 355–405. Speyer, J. S., trans. 1982 [1885]. The Jātakamālā, or Garland of Birth­Stories of Āryaśūra. Delhi: Motilal Banarsidass. Tatz, M. 1994. The Skill in Means (Upāyakauśalya) Sūtra. Delhi: Motilal Banarsidass.

The ātman and its negation A conceptual and chronological analysis of early Buddhist thought
Alexander Wynne

The denial that a human being possesses a “self” or “soul” is probably the most famous Buddhist teaching. It is certainly its most distinct, as has been pointed out by G. P. Malalasekera: “In its denial of any real permanent Soul or Self, Buddhism stands alone.”1 A similar modern Sinhalese perspective has been expressed by Walpola Rahula: “Buddhism stands unique in the history of human thought in denying the existence of such a Soul, Self or Ātman.”2 The “No Self” or “no soul” doctrine (Sanskrit: anātman; Pāli: anattan) is particularly notable for its widespread acceptance and historical endurance. It was a standard belief of virtually all the ancient schools of Indian Buddhism (the notable exception being the Pudgalavādins),3 and has persisted without change into the modern era. Thus the classical Theravādin view of Buddhaghosa that “there is only suffering, but nobody who suffers”4 is identical to the view of Śāntideva, the famous Indian Mahāyānist, that “the person who experiences suffering does not exist,”5 and both views are mirrored by the modern Theravādin perspective of Mahasi Sayadaw

Malalasekera 1957: 33. Rahula 1959: 51. 3 On the Pudgalavādins see Châu 1999 and Williams and Tribe 2000: 124–28. 4 Vism XVI.90 (Warren and Kosambi 1989: 436): dukkham eva hi, na koci dukkhito. 5 BCA VIII.101 (Tripathi 1988: 164): yasya duḥkhaṃ sa nāsti.
2

1

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 103–171

104

Alexander Wynne

that “there is no person or soul”6 and the modern Mahāyāna view of the fourteenth Dalai Lama that “[t]he Buddha taught that … our belief in an independent self is the root cause of all suffering.”7 This traditional understanding has been accepted by Buddhist scholars past and present. According to La Vallée Poussin, the Buddha did not accept “the existence of a Self (ātman), a permanent individual; he teaches that the so-called Self is a compound of material and spiritual data called skandhas.”8 In a similar vein Norman has stated that “the Buddha denied the existence of the permanent individual self,”9 Collins has spoken of the Buddhist “denial of self,”10 and De Jong has noted that “in early Buddhism impermanence and suffering imply the non-existence of the self as a permanent entity.”11 There is early canonical support for all these statements. In the Vajirā Sutta of the Saṃyutta-Nikāya, where the bhikkhunī Vajirā reports the doctrine to Māra as follows:
Why do you believe in a living being? Is not this your view, Māra? This is nothing but a heap of formations: No being is found here. (553) When there is a collection of parts the word ‘chariot’ is used; In the same way, when the aggregates exist (khandhesu santesu) the conventional term ‘being’ (satto) [is applied to them]. (554) Only suffering (dukkham eva) comes into existence, and only suffering endures. Nothing apart from suffering comes into existence, and nothing apart from suffering ceases to exist. (555)12
Kornfield 1996: 45. Dalai Lama 1994: 111. 8 La Vallée Poussin 1917: 34. 9 Norman 1981: 87ff. 10 Collins 1982: 95. 11 De Jong 2000: 177. 12 SN I.296 (v. 553–55): kin nu satto ti paccesi Māra diṭṭhigatan nu te, suddhasaṅkhārapuñjo ’yaṃ na yidha sattūpalabbhati. (553) yathā hi aṅgasambhārā hoti saddo ratho iti, evaṃ khandhesu santesu hoti satto ti sammuti. (554) dukkham eva hi sambhoti dukkhaṃ tiṭṭhati veti ca, nāññatra
7 6

The ātman and its negation

105

The statement that “only suffering comes into existence, and only suffering endures” is akin to Buddhaghosa’s statement that “there is only suffering, but nobody who suffers,” and Śāntideva’s statement that “the person who experiences suffering does not exist:” all assume that there is no “ghost” in the machine. Such an understanding has been summed up by T. W. Rhys Davids as follows: “Man is never the same for two consecutive moments, and there is within him no abiding principle whatever.”13 According to this definition, and depending on one’s perspective, the doctrine could be taken to mean either that a person has no “soul” (in the sense of a “spiritual part of man in contrast to the purely physical”), or that a person lacks an inherent identity, that which could be termed “self” (in its simple philosophical sense of “that which a person is really and intrisically he (in contradistinction to what is adventitious).” For these definitions see the OED). Given the close correspondence between the Vajirā Sutta, Buddhaghosa, Śāntideva and the more recent Buddhist authorities cited above, this understanding would seem to have been the norm in Buddhist circles for over two thousand years. Indeed the attestation of this idea in a canonical text means that it can most probably be taken back to the pre-Aśokan period, i.e. within roughly 150 years of the Buddha’s death.14

1. The historical problem
Despite its importance and historical endurance, it is odd that the No Self doctrine is hardly attested in the early Buddhist literature, the Vajirā Sutta being perhaps the only Pāli discourse to state the idea explicitly.15 Indeed it is very easy to read a substantial amount
dukkhā sambhoti nāññatrā dukkhā nirujjhati. (555) Buddhaghosa cites some of these verses in his Visuddhimagga (XVIII.25, 27; Warren and Kosambi 1989: 508. 13 Rhys Davids 1877: 94. 14 Assuming that the Buddha’s teaching career began at around 450 BCE, and that the Buddha died in about 404 BCE, i.e. about 136 years before Aśoka’s inauguration (Gombrich 1992: 246). On the pre-Aśokan date of canonical Pāli Suttas, see Wynne 2005. 15 Although see section 7 below on the possibility that the doctrine is as-

106

Alexander Wynne

of the early literature – such as virtually all of the important DīghaNikāya – without encountering anything remotely like it. There are only two plausible explanations for this historical peculiarity. The first is that the doctrine is implicit in the early texts, but for some reason was only explicated in the Vajirā Sutta (and perhaps a few other discourses). The second is that the doctrine was generally unknown to the composers of the canonical texts – either because it emerged at a later date or because it was initially a fringe idea – who therefore failed to record it. Both explanations imply that Buddhist thought changed over time: from implicit to explicit formulations of the doctrine, or so that a later development or minority concern eventually came to dominate the philosophical mainstream of the early saṅgha. In other words, there must have been either a terminological or philosophical change in early Buddhist thinking about the human being. To establish the more likely eventuality, the early Buddhist teachings on personal identity must be reconsidered. This will involve going over much old ground, but since a general consensus has not been reached this is unavoidable.16 Such a study is further necessitated by the fact that a number of important text-critical problems have been ignored: the historical problem noted above has not been properly recognised, little thought has been given to the form of the important Not-Self teaching, the basic vocabulary of the teachings on personal identity has been misunderstood, nonBuddhist parallels to important teachings have been missed, and little attempt has been made to relate the teachings on personal identity to the wider doctrinal concerns of the early texts. In short, there is much scope for a more detailed exploration of this aspect of early Buddhist thought.
sumed by the Mahāhatthipadopama Sutta. 16 A minority but not insignificant view is that of Pérez-Remón (1980), who has argued that since the early texts do not deny the self they must in fact presuppose it; for more scholars who believe that the early Buddhist teachings presuppose a self, see Collins 1982: 3–10. Against this view, Vetter (1988: 41, n.10) has argued that although the early texts do not deny the existence of the self, they do not presuppose it. Oetke’s detailed study (1988) also argues that the early texts neither deny nor affirm the existence of the self.

The ātman and its negation

107

Such a study must begin with a very old and much discussed point: the Upaniṣadic background to early Buddhism. For as we will see, early Upaniṣadic speculation on the ātman was wellknown in early Buddhist circles, and determines the form and content of some important early Buddhist teachings. Of the various senses in which the term ātman is used in the early Upaniṣads, the most important is the “spiritual self or the inmost core of a human being.”17 According to the Yājñavalkyakāṇḍa of the Bṛhadāraṇyaka Upaniṣad, the ātman in this sense is a non-physical substance,18 a principle of life on which a person’s cognitive functions (prāṇa) depend,19 and also the inner subject of perception, i.e. an “unseen seer”20 that consists of nothing but
According to Olivelle (1998: 22) the term is used in two other senses, i.e. as a simple reflexive pronoun and in denoting “a living, breathing body.” The three usages are also well attested throughout Sanskrit literature. According to Monier Williams, in its simplest sense the term ātman refers to the bodily person: “the person or whole body considered as one and opposed to the separate members of the body” (MMW s.v.). In a similar fashion, the term ātman is often used as a reflexive pronoun: “ātman in the sg. is used as reflexive pronoun for all three persons and all three genders, e.g. ātmānaṃ sā hanti ‘she strikes herself’” (MMW s.v.). Monier Williams also cites the expressions ātman (Ved. loc.) dhatte/karoti: “he places in himself, makes his own,” and atmanā akarot: “he did it himself;” he also suggests that something like “essence” is the oldest and most basic definition of the term ātman as “the breath, the soul, principle of life and sensation (RV, AV).” 18 That the ātman is a spiritual principle is made clear in the numerous references to its immortality, e.g. BU IV.3.12, BU IV.4.16, BU IV.4.17, BU IV.4.25. That it is a spiritual principle distinct from the body is made clear in those passages which describe its reincarnation, e.g. BU IV.3.8 and especially the detailed account of BU IV.4.3: … ayam ātmedaṃ śarīraṃ nihatyāvidyāṃ gamayitvānyam ākramam ākramyātmānam upasaṃharati: “Once this ātman has struck the body down and rendered it unconsciousness, it approaches another station and draws itself towards it.” See also BU IV.4.7: tadyathā ahinirvlayanī valmīke mṛtā prayastā śayīta, evam evedaṃ śarīraṃ śete. athāyaṃ aśarīro ’mṛtaḥ prāṇo brahmaiva teja eva. 19 BU IV.4.2: tam utkrāmantaṃ prāṇo ’nūtkrāmati, prāṇam utkrāmantaṃ sarve prāṇā anūtkrāmanti. 20 E.g. BU III.7.23: adṛṣṭo draṣṭrāśrutaḥ śrotāmato mantāvijñāto vijñātā (see also BU III.8.11); see also BU IV.5.15 which makes it clear that the ātman, as the perceiver, cannot be perceived (vijñātāram are kena vijānīyād:
17

108

Alexander Wynne

consciousness.21 The ātman of Yājñavalkya, and henceforth of Vedāntic philosophy in general, is both a spiritual substance and an unchanging inner subject of phenomena. This understanding is not entirely different from the Cartesian “mind-soul,” which is also a spiritual substance as well as the true subject of experience.22� Despite this similarity, however, the ātman of Yājñavalkya is also said to be a nondual consciousness identical to the underlying essence of the cosmos (brahman), the realisation of which is a state of pure bliss.23 This equation of microcosm (ātman) and macrocosm (brahman) is of course philosophically problematic, since it involves the identification of the individual subject of perception with an impersonal essence. In the early Upaniṣads, however, this problem is resolved “mystically” rather than philosophically: as the Bṛhadāraṇyaka Upaniṣad states, by meditating on the ātman (when it is “seen, heard, contemplated and cognised …”), its true nature as a macrocosmic essence will be revealed (“… the whole world is known”).24 Such an understanding belongs more to the realm of religious experience than that of rational enquiry. The peculiar identity of microcosm and macrocosm was not initially problem“By what means might one perceive the perceiver?”). 21 BU IV.3.7: yo ’yaṃ vijñānamayaḥ prāṇeṣu; BU IV.4.22: sa vā eṣa mahān aja ātmā vijñānamayaḥ prāṇeṣu. 22 Although Williams states that “very little” of his discussion about Western concepts of the soul “is relevant to the Buddha,” (Williams and Tribe 2000: 56), his definition of the Cartesian position shows striking similarities with Yājñavalkya’s understanding of the ātman (ibid.): “As is well known, Descartes identified that which gives life to the body, and survives death, with the mind, and he also identified this mind-soul as the true self, of an intrinsically different stuff from the body. The mind-soul is the factor in which lies the identity of the person over time and change.” 23 BU IV.3.32 states that the ātman is the “highest bliss” (parama ānandaḥ), BU IV.5.22 states that the ātman is nondual consciousness (vijñānaghana), BU IV.5.11 states that it is macrocosmic (mahābhūta), and BU IV.5.12 likens the person who unites with the ātman in deep sleep to a “single ocean” (salila ekaḥ), a state equated with the “world of brahma” (brahmaloka). 24 BU IV.5.6: ātmani khalv are dṛṣṭe śrute mate vijñāta idaṃ sarvaṃ viditaṃ: “When the ātman is seen, heard, contemplated and cognised, the whole world is known.” For a parallel teaching see BU II.4.5: ātmano vā are darśanena śravaṇena matyā vijñānenedaṃ sarvaṃ viditam.

The ātman and its negation

109

atic, therefore, but only became so for later generations of Vedāntic thinkers who puzzled it over to varying degrees of success. It is against this conceptual background that the early Buddhist teachings on personal identity must be understood. As will become clear, these teachings refer to the Upaniṣadic ātman in both its microcosmic and macrocosmic apsects (as the inner perceiver and nondual essence respectively). Understanding exactly how this is the case will help resolve the problem of whether the “No Self” teaching of the Vajirā Sutta was implicit in other early teachings, or whether it was a philosophical development from an earlier period dominated by different concerns.

2. The Not-Self teaching
Perhaps the most important source for the early Buddhist critiques of the ātman are the various sectarian accounts of the Buddha’s Second Sermon. The first teachings of this sermon states that the five aggregates (form, feeling, apperception, volitions and consciousness) are “not ātman/attan” (anātman/anattan) since they are beyond a person’s control.25 The precise meaning of this teaching is unclear: quite what the word ātman/attan means, and why the ability to control each of the five aggregates would mean that they constitute an ātman/attan, is difficult to make out.26 The peVin I.13.18: rūpaṃ bhikkhave anattā. rūpañ ca h’ idaṃ bhikkhave attā abhavissa, na yidaṃ rūpaṃ ābādhāya saṃvatteyya, labbhetha ca rūpe: evaṃ me rūpaṃ hotu, evaṃ me rūpaṃ mā ahosī ti. yasmā ca kho bhikkhave rūpaṃ anattā, tasmā rūpaṃ ābādhāya saṃvattati, na ca labbhati rūpe: evaṃ me rūpaṃ hotu, evaṃ me rūpaṃ mā ahosī ti. vedanā anattā, vedanā ca h’ idaṃ bhikkhave attā abhavissa … For this teaching in Buddhist Sanskrit texts, see Mvu III.335.12, SbhV I.138.10 and CPS 15.2 (Waldschmidt 1952: 162). 26 Collins 1982: 97 has suggested that this teaching is directed against the Brahminic notion of the ātman as the “microcosmic reflection of the macrocosmic force of the universe (brahman).” But the teaching does not presuppose that the ātman should be an inner controller, and if so it would not appear to be a “microcosmic reflection” of a world-controlling force. The teaching instead states that if the five aggregates were ātman/attan, a person should be able to change them as he wants. The argument is thus directed against the notion that the five aggregates constitute the ātman/attan, and not
25

110

Alexander Wynne

culiar content of this teaching is matched by the fact that it is mentioned in only one Pāli discourse (the Cūḷasaccaka Sutta) besides the two Pāli texts that record the Second Sermon.27 This peculiarity suggests that the teaching was of little importance in the early Buddhist period. Much more important is the second teaching of the Second Sermon. As Collins has pointed out, “a very high proportion of the discussions of Not-Self in the Suttas consist in various versions of this argument.”28 This teaching is in fact the most important early Buddhist negation of the ātman, and one that has distinct Upaniṣadic undertones:
What do you think, bhikkhus: is form permanent (nicca) or impermanent (anicca)? ‘Impermanent, master.’ Is that which is impermanent unsatisfactory (dukkha) or satisfactory (sukha)? ‘Unsatisfactory, master.’ And is it suitable to regard that which is impermanent, unsatisfactory and subject to change (vipariṇāmadhamma) as “This is mine, I am this, this is my attan”?
against the notion that there is ātman/attan controlling them from within, as is also assumed by Siderits (2007: 46ff.). For a full study of the conceptual and historical implications of this teaching see Wynne 2009a; especially 86–88. 27 Collins 1982: 97 has suggested that this teaching is directed against the Brahminic notion of the ātman as the “microcosmic reflection of the macrocosmic force of the universe (brahman).” But the teaching does not presuppose that the ātman should be an inner controller, and if so it would not appear to be a “microcosmic reflection” of a world-controlling force. The teaching instead states that if the five aggregates were ātman/attan, a person should be able to change them as he wants. The argument is thus directed against the notion that the five aggregates constitute the ātman/attan, and not against the notion that there is ātman/attan controlling them from within, as is also assumed by Siderits (2007: 46ff.). For a full study of the conceptual and historical implications of this teaching see Wynne 2009a; especially 86–88. 28 Collins 1982: 98. Similar teachings begin book IV (Saḷāyatanavagga) of the Saṃyutta-Nikāya: it is stated that all sense faculties and their objects are impermanent (anicca), unsatisfactory (dukkha) and so Not-Self (anattan). Derivatives of this teaching, where the five aggregates are also stated to be anattan, can be found at SN III.20–21, 23–24 and 179.

The ātman and its negation

111

‘No, master.’29

Since the list of five aggregates denotes the different aspects of phenomenal being,30 this teaching therefore states that an ātman/ attan cannot be found in conditioned experience. Although it is not immediately clear how the term ātman/attan is to be taken here, Norman has argued that it is to be understood as the blissful and unchangeable ātman of the early Upaniṣads. This is shown by the fact that the response to the Buddha’s final question can only be given
by those who know, in advance, that the term attā is by definition nicca and sukha, and therefore anything which is anicca and dukkha cannot be attā. This gives us a clear indication of the type of attā which is being discussed. It is the Upaniṣadic idea of an ātman which is nitya and sukha …31

If this teaching negates the Upaniṣadic ātman in the sense of an unchanging, blissful essence, it would seem to be concerned with the ātman in its macrocosmic aspect (as brahman), for this is how
29

Vin I.14: taṃ kiṃ maññatha bhikkhave: rūpaṃ niccaṃ vā aniccaṃ vā ti? aniccaṃ bhante. yaṃ panāniccaṃ dukkhaṃ vā taṃ sukhaṃ vā ti? dukkhaṃ bhante. yaṃ panāniccaṃ dukkhaṃ vipariṇāmadhammaṃ, kallan nu taṃ samanupassituṃ: etaṃ mama, eso ’haṃ asmi, eso me attā ti? no h’ etaṃ bhante. For the various Sanskrit versions of this teaching, see Mvu III.337.11, SbhV I.138.21 and CPS 15.6 (Waldschmidt 1952: 164). 30 Gethin 1986: 49: “the five khandhas, as treated in the Nikāyas and early Abhidhamma, do not exactly take on the character of a formal theory of the nature of man. The concern is not so much the presentation of an analysis of man as object, but rather the understanding of the nature of conditioned existence from the point of view of the experiencing subject. Thus at the most general level rūpa, vedanā, saññā, saṃkhārā and viññāṇa are presented as five aspects of an individual being’s experience of the world …” Hamilton (2000: 27) has similarly written that the five aggregates are “not a comprehensive analysis of what a human being is comprised of … Rather they are factors of human experience.” 31 Norman 1981: 22. To this we might add that by equating “impermanence” (anicca) with being “subject to change” (vipariṇāmadhamma), the Buddha recalls a key feature of the self according to the Yājñavalkyakāṇḍa (e.g. BU IV.5.15), i.e. that it is unchangeable. Bronkhorst (2007: 233) has noted that BU IV.5.15 “introduces the notion of the immutability of the self.”

112

Alexander Wynne

the blissful ātman is considered in the early Upaniṣads.32 It does not make any sense, however, to read this teaching as a negation of a macrocosmic essence. While it might make sense to ask whether consciousness has the characteristics of the macrocosmic ātman (since a number of important early Upaniṣadic passages state that the ātman in its macrocosmic aspect is a nondual consciousness),33 it makes little sense to ask if form, sensation, apperception and volitions have the characteristics of an essence that transcends all phenomena. That the teaching is not a straightforward denial of the macrocosmic ātman is also suggested by the use of the terms sukha and nicca. The primary sense of these terms seems not to be “permanent” and “blissful,” as would be the case in a straightforward negation of Upaniṣadic thought. The term sukha is rather an antonym of the term dukkha, which here has the weak sense of “unsatisfactory,” this being the only way in which the impermanence (anicca) of the five aggregates can be taken: the point is not that the general experiential condition of a person is “suffering,” but rather that it is ultimately unsatisfactory on the basis that something enduring cannot be found. The teaching therefore appears more like an enquiry into phenomenal identity rather than a direct negation of the Upaniṣadic ātman: it seeks to establish that form and so on are affected by causes and conditions (i.e. that they are adventitious: vipariṇāmadhamma) and so cannot constitute what a person “really and intrinsically” is. If so, it would appear that the teaching is a philosophical enquiry into intrinsic identity or “self.”34
See BU II.1.19–20, BU III.9.28, BU IV.3.32–33, TU II.5–9, TU III.6. E.g. BU II.4.12: idaṃ mahad bhūtam anantam apāraṃ vijñānaghana eva; BU IV.5.13: ayam ātmānantaro ’bāhyaḥ kṛtsnaḥ prajñānaghana eva. 34 Opposition to the notion that this teaching denies “soul” has been made by Gombrich (1996: 15): “In Western languages, the Buddha is presented as having taught the doctrine (vāda) of ‘no soul’ (anātman). What is being denied – what is a soul? Western theologians are at home in the Christian cultural tradition. Christian theologians have differed vastly over what the soul is. For Aristotle, and thus for Aquinas, it is the form of the body, what makes a given individual person a whole rather than a mere assemblage of parts. However, most Christians conceive of the soul, however vaguely, in a completely different way, which goes back to Plato: that the soul is other than the
33 32

The ātman and its negation

113

This impression is strengthened by the fact that the term ātman/ attan must here be taken in the philosophical sense of “self.” For it makes no sense to ask whether “form” (i.e. the physical body), sensation, apperception and volitions constitute a “soul,” no matter how the latter is conceived. These are sensible questions in an enquiry into personal identity, however: since a person ordinarily identifies with form and so on as “oneself,” it makes sense to ask if they really can be considered as an intrinsic identity. On this point it is important to note that the Not-Self teaching is often placed directly after the statement that an ordinary person identifies with the five aggregates in the form “This is mine, I am this, this is my self” (ātman/attan).35 Such texts play on the flexibility in the term ātman/ attan, using it first in the sense derived from the reflexive pronoun (where it denotes a person’s phenomenal identity), and second in the more philosophical sense of intrinsic identity. In this way the teaching points out that although a person takes the five aggregates as his individual “self” (ātman/attan), this is unsatisfactory since no intrinsic identity (ātman/attan) can be found therein. It would seem, then, that the teaching addresses the problem of personal identity by questioning the identification with phenomenal being. To this end the Upaniṣadic notion of an ātman that is blissful and permanent/unchanging is certainly invoked, of course, but this would seem to be only for the sake of communicating a new idea in a particular intellectual context. This is far from a statement of the No Self doctrine as described by the Vajirā Sutta, Buddhaghosa and Śāntideva: there is no denial of the self per se, but only a subtle argument that the concept of a “self” does not make sense of conditioned experience. Since the underlying metaphysic of the teaching is not made clear, it is possible that it presupposes a self beyond conditioned experience or the
body, as in the expression ‘body and soul,’ and is some kind of disembodied mental, and above all, moral, agent, which survives the body after death. But none of this has anything to do with the Buddha’s position. He was opposing the Upaniṣadic theory of soul.” A similar point has been made by Williams (Williams and Tribe 2000: 56). 35 This is how the teaching is presented in its most important occurrence in the Pāli discourses, the Alagaddūpama Sutta (MN I.135.27ff.).

114

Alexander Wynne

exact opposite. If so, the difference between this teaching and the “No Self” doctrine of the Vajirā Sutta remains to be determined. A better understanding of the problem requires an investigation of other early Buddhist teachings on personal identity.

3. Self-consciousness in the early Buddhist texts
The Not-Self teaching considers a person’s identification with the five aggregates in terms of the notion “This is mine, I am this, this is my self.” As such, it is closely connected to other early Buddhist teachings concerned with the notion “I” (ahan ti), the notion “I am” (asmī ti), the “conceit I am” (asmimāna), and the “underlying tendency towards conceit with regard to the notions ‘I’ and ‘mine’ (ahaṃkāramamaṃkāra-mānānusaya).” All of the teachings containing these formulations tackle the subject of personal identity by examining what we might call “self-consciousness,” i.e. a person’s awareness of his own “identity,” “acts” and “thoughts.”36 Rather than enquire into whether an intrinsic identity can be found in self-consciousness, as the Not-Self teaching does, such teachings explore the affective and cognitive aspects of reflexive awareness. Its affective nature is most apparent in these texts, this being indicated by the compounds the “underlying tendency towards conceit with regard to the notions ‘I’ and ‘mine’ (ahaṃkāra/mamaṃkāramānānusaya)” and “the conceit ‘I am’ (asmimāna).” Both formulations indicate that “for the unenlightened man, all experience and action must necessarily appear phenomenologically as happening to or originating from an ‘I’.”37 This means, in other words, that self-consciousness is a basic sort of existential conceit, a grasping at individual existence and identity that underpins all conditioned experience. Other texts affirm this affective understanding of self-consciousness, e.g. the Taṇhā Sutta, where eighteen “thoughts caused by thirst” (taṇhāvicarita) are listed with reference to both oneself (ajjhatikassa upādāya) and that which is external to oneself
36 37

I follow the definition of the OED. Collins 1982: 94.

The ātman and its negation

115

(bāhirassa upādāya) respectively.38 Thus the notion “I am” (asmī ti) is the most basic “thought caused by thirst” that paves the way for seventeen further forms of self-consciousness with reference to oneself (e.g. evasmī ti: “I am thus”); and the notion “I am in respect of this” (imināsmī ti) is the most basic “thought caused by thirst” that paves the way for seventeen further forms of self-consciousness with reference to that which is external to oneself (e.g. iminā evasmī ti: “I am thus in respect of this”).39 In contrast to this analysis of the affective aspect of self-consciousness, other texts suggest that self-consciousness is a sort of ignorance. A good example is the Sammādiṭṭhi Sutta, which equates the destruction of the “underlying tendency towards conceit with regard to the view ‘I am’ (asmī ti diṭṭhimānānusayaṃ samūhanitvā)” with the abandonment of ignorance (avijjaṃ pahāya) and the attainment of knowledge (vijjaṃ uppādetvā).40 Other texts combine the affective and cognitive aspects of self-consciousness. Thus the Dutiyanānātitthiya Sutta concludes with the following “inspired utterance” (udāna) of the Buddha:
This generation is obsessed with the notion ‘I’ (ahaṃkāra) and attached to the notion ‘another’ (paraṃkāra). They have not understood this matter, and have not seen that it is a barb. For the person with vision who has removed this barb,
38 It is perhaps possible that the compound taṇhāvicarita is to be read as a dependent determinative (tatpuruṣa) in the dative case, i.e. “movements towards thirst.” But the compound taṇhāvicarita is more likely to be read as a dependent determinative in the instrumental case: a past participle preceded by a substantive suggests some sort of conceptual activity (vicarita) prompted by a cause (taṇhā). If so, self-consciousness would seem to be caused by an underlying affective state termed “thirst.” 39 AN II.212.13: katamāni aṭṭhārasa taṇhāvicaritāni ajjhattikassa upādāya? asmī ti bhikkhave sati … 40 MN I.47.21: yato kho āvuso ariyasāvako evaṃ akusalaṃ pajānāti, evaṃ akusalamūlaṃ pajānāti, evaṃ kusalaṃ pajānāti, evaṃ kusalamūlaṃ pajānāti, so sabbaso rāgānusayaṃ pahāya, paṭighānusayaṃ paṭivinodetvā, asmī ti diṭṭhimānānusayaṃ samūhanitvā avijjaṃ pahāya vijjaṃ uppādetvā, diṭṭheva dhamme dukkhass’ antakaro hoti. ettāvatā pi kho āvuso ariyasāvako sammādiṭṭhi hoti …

116

Alexander Wynne

the notions ‘I am acting’ and ‘Another is acting’ do not occur. This generation is mired, bound and trapped by conceit (māna); it exerts itself over views, and so does not escape transmigration.41

The notions of “I” (ahaṃkāra) and “another” (paraṃkāra) are here related to conceit (māna) and the holding of speculative views (diṭṭhi). Self-consciousness, therefore, is deeply involved in the affective and cognitive causes of a person’s suffering. Given that it is said to be an “underlying tendency” (anusaya), it would seem to be an ever-present factor of conditioned experience, and its transcendence would no doubt effect a substantial personal transformation. Although the early texts generally have little to say about the state achieved through eradicating self-consciousness, there are a couple of illuminating exceptions. A series of Suttas in the SaṃyuttaNikāya describe how Sāriputta attained all nine “gradual abidings” (anupubbavihāra) – from the first jhāna to the “cessation of apperception and feeling” (saññāvedayitanirodha) – despite lacking self-consciousness. These texts begin with a question from Ānanda about the reason for the unusually calm countenance of Sāriputta. The latter explains that this is due to the meditative states he attains without any prior intention, a state of affairs that Ānanda attributes to his lack of self-consciousness:
‘Herein, venerable sir, I pass my time having attained the first jhāna, that state of joy and bliss born of seclusion which is devoid of desire and bad thoughts, but which includes deliberation (vitakka) and reflection (vicāra). It does not occur to me, venerable sir, that I am attaining the first jhāna, or have attained the first jhāna, or have emerged from the first jhāna.’ ‘It is so for the venerable Sāriputta because the underlying tendency towards conceit in the notions ‘I’ and ‘mine’ has for a long time been

Ud VI.6 (v. 70.23): atha kho Bhagavā etam atthaṃ viditvā tāyaṃ velāyaṃ imaṃ udānaṃ udānesi: ahaṅkārapasutā ayaṃ pajā paraṃkārūpasaṃhitā, etad eke nābbhaññaṃsu, na naṃ sallan ti addaṃsu. etaṃ ca sallaṃ paṭigacca passato, ahaṃ karomī ti na tassa hoti, paro karotī ti na tassa hoti. mānupetā ayaṃ pajā mānaganthā mānavinibaddhā, diṭṭhīsu byārambhakatā, saṃsāraṃ nātivattatī ti.

41

The ātman and its negation

117

destroyed. Therefore it does not occur to the venerable Sāriputta that he is attaining the first jhāna, or has attained the first jhāna, or has emerged from the first jhāna.’42

Apart from another text in which Ānanda states that Sāriputta’s “underlying tendency towards conceit in the notions ‘I’ and ‘mine’ has for a long time been destroyed,”43 the only other discourse on the unusual psychology of a person devoid of self-consciousness is the Upasena-āsīvisa Sutta. In this peculiar text, the venerable Upasena is said to have been bitten by a poisonous snake while both he and Sāriputta dwelt in the Sappasoṇḍika mountain cave.44 When he subsequently asks to be taken outside on a couch, before his body “falls apart right here, just like a fistful of chaff,”45 Sāriputta exclaims that he sees no change in his body or decline in his faculties.46 To this Upasena states that his unusual countenance is due to the fact that he lacks self-consciousness with regard to his sense faculties:

SN III.235.22: idhāhaṃ āvuso vivicc’ eva kāmehi vivicca akusalehi dhammehi savitakkaṃ savicāraṃ vivekajaṃ pītisukhaṃ paṭhamajjhānaṃ upasampajja viharāmi. tassa mayhaṃ āvuso na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti vā, ahaṃ paṭhamajjhānaṃ samāpanno ti vā, ahaṃ paṭhamajjhānā vuṭṭhito ti vā ti. tathā hi panāyasmato Sāriputtassa dīgharattaṃ ahaṃkāramamaṃkāramānānusayā susamūhatā, tasmā āyasmato Sāriputtassa na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti vā ahaṃ paṭhamajjhānaṃ samāpanno ti vā ahaṃ paṭhamajjhānā vuṭṭhito ti vā ti. The PTS reading pahatmajjhānā at the end instead of paṭhamajjhānā is clearly an error. 43 SN II.275.1: tathā hi panāyasmato Sāriputtassa dīgharattaṃ ahaṅkāramamaṅkāramānānusayā susamūhatā. 44 SN IV.40.16: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha. purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamuṭṭhī ti. 45 SN IV.40.16: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha. purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamuṭṭhī ti. 46 SN IV.40.20: evaṃ vutte āyasmā Sāriputto āyasmantaṃ Upasenaṃ etad avoca: na kho pana mayaṃ passāma āyasmato Upasenassa kāyassa vā aññathattaṃ indriyānaṃ vā vipariṇāmaṃ. atha ca panāyasmā Upaseno evam āha: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha; purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamuṭṭhī ti.

42

118

Alexander Wynne

Venerable Sāriputta, the person who might think that he is the eye or possesses it … that he is the tongue or possesses it … that he is the mind or possesses it, for him there might be a change in his body or a decline in his faculties. But it is does not occur to me, venerable Sāriputta, that I am the eye or possess it … that I am the tongue or possess it … that I am the mind or possess it. So how could there be a change in my body or decline in my faculties?47

Sāriputta thus concludes that “the venerable Upasena’s underlying tendency to feel conceit in the notions ‘I’ and ‘mine’ has for a long time been destroyed,”48 and the story concludes with the account of how Upasena’s body fell apart “like a fistful of chaff” after he had been taken outside on a couch.49 This text thus claims that Upasena achieved a completely impersonal state, one in which the automatic tendency to identify with conditioned experience had ceased to function. Various texts describe the means of attaining this state, e.g. following the path that leads through the four jhānas and culminates in the three knowledges,50 or concentrating on the thought “this is calm, this is supreme, namely the calming of all mental formations, the relinquishment of all attachment (upadhi), the destruction of thirst, dispassion, cessation, Nirvana,” which is said to lead to the attainment of the “release of mind, a release through understanding.”51 For the purpose of the present enquiry, however,
AN IV.40.29: yassa nūna āvuso Sāriputta evam assa: ahaṃ cakk­ hun ti mama cakkhun ti vā … la … ahaṃ jivhā ti vā, mama jivhā ti vā … ahaṃ mano ti vā mama mano ti vā. tassa āvuso Sāriputta siyā kāyassa vā aññathattaṃ indriyānaṃ vā vipariṇāmo. mayhañ ca kho āvuso Sāriputta na evaṃ hoti: ahaṃ cakkhun ti vā, mama cakkhun ti vā … la … ahaṃ jivhā ti vā mama jivhā ti vā … ahaṃ mano ti vā mama mano ti vā. tassa mayhañ ca kho āvuso Sāriputta kiṃ kāyassa vā aññathattaṃ bhavissati indriyānaṃ vā vipariṇāmo ti? 48 SN IV.41.6: tathā hi panāyasmato Upasenassa dīgharattam ahaṃkāramamaṃkāramānānusayā susamūhatā. 49 SN IV.41.11: atha kho te bhikkhū āyasmato Upasenassa kāyaṃ mañcakaṃ āropetvā bahiddhā nīhariṃsu. atha kho āyasmato Upasenassa kāyo tatth’ eva vikiri, seyyathā pi bhūsamuṭṭhī ti. 50 MN III.32.32ff. 51 AN I.133.1: idh’ Ānanda bhikkhuno evaṃ hoti: etaṃ santaṃ etaṃ
47

The ātman and its negation

119

more important are other texts which relate the transcendence of self-consciousness to the understanding of the five aggregates and the Not-Self teaching. The Dutiyasaññā Sutta of the AṅguttaraNikāya, for example, states that the problem of self-consciousness (ahaṃkāra/mamaṃkāra) is resolved by regarding that which is unsatisfactory (dukkha) as Not-Self (anattan).52 A similar contemplation is outlined in the Mahāpuṇṇa Sutta, which states that seeing the five aggregates as Not-Self (attan) leads to the cessation of selfconsciousness as follows:
O bhikkhus, one should regard whatever form is past, present or future, internal or external, gross or subtle, inferior or superior, near or far – all form – as ‘This is not mine, I am not this, this is not my self (attan)’ … For the person who knows and sees it thus, bhikkhus, the tendency towards conceit in the notion ‘I’ with regards to the body and its consciousness, and towards conceit in the notion ‘mine’ with regards to external objects, does not arise.53

This passage does not state how contemplating the insubstantiality of conditioned experience (the five aggrgegates) aids the transcendence of self-consciousness, but the point is investigated in more

paṇītaṃ yadidaṃ sabbasaṅkhārasamatho sabbūpadhipaṭinissaggo taṇhākkhayo virāgo nirodho nibbānan ti. evaṃ kho Ānanda siyā bhikkhuno tathārūpo samādhipaṭilābho yathā imasmiñ ca saviññāṇake kāye ahaṅkāramamaṅkāramānānusayā nāssu bahiddhā ca sabbanimittesu ahaṅkāramamaṅkāramānānusayā nāssu, yañ ca cetovimuttiṃ paññāvimuttiṃ upasampajja viharato ahaṅkāramamaṅkāramānānusayā na honti, tañ ca cetovimuttiṃ paññāvimuttiṃ upasampajja vihareyyā ti. 52 AN IV.53.7: dukkhe anattasaññāparicitena bhikkhave bhikkhuno cetasā bahulaṃ viharato imasmiñ ca saviññāṇake kāye bahiddhā ca sabbanimitte­ su ahaṃkāramamaṃkāramānāpagataṃ mānasaṃ hoti vidhāsamatikkantaṃ santaṃ suvimuttaṃ. 53 MN III.18.32: yaṃ kiñci bhikkhu rūpaṃ – atītānāgatapaccuppannaṃ, ajjhattaṃ vā bahiddhā vā, oḷārikaṃ vā sukhumaṃ vā, hīnaṃ vā paṇītaṃ vā, yaṃ dūre santike vā – sabbaṃ rūpaṃ: n’ etaṃ mama, n’ eso ham asmi, na m’ eso attā ti, evam etaṃ yathābhūtaṃ sammappaññāya passati … MN III.19.7: evaṃ kho bhikkhu jānato evaṃ passato imasmiñ ca saviññāṇake kāye bahiddhā ca sabbanimittesu ahaṃkāramamaṃkāramānānusayā na hontī ti. For the same teaching see SN II.252.16, 253.11; SN III.80.7, 81.1, 103.12, 136.4, 136.24, 169.12, 170.7.

120

Alexander Wynne

detail in the Khemaka Sutta, where the bhikkhu Khemaka addresses the elders of Kosambī as follows:
Venerable sirs, the Blessed one has spoken of five aggregates of attachment,54 namely: the aggregate of attachment that is form … feeling … apperception … volitions … [and] consciousness. I have no view that any sort of self (attan) or its property (attaniya) is found in these five aggregates of attachment, venerable sirs, and yet I am not an arahant devoid of corruptions. For I still have the notion ‘I am’ (asmī ti) with regard to these five aggregates of attachment, venerable sirs, despite the fact that I do not have the view ‘I am this’ (ayam asmī ti na ca samanupassāmi).55

The logic of this statement is relatively simple. Khemaka knows that he should be detached from the conditioned experience of the five aggregates, this being inherently unsatisfactory since it lacks intrinsic identity (attan). But he is unable to do so because of his automatic tendency to identify with conditioned experience in the form of the notion “I am.” Although Khemaka knows what he should know, according to Buddhist doctrine, and so does not intentionally identify with the the five aggregates, his identification with them runs deeper in the form of a sense of subjectivity (asmī ti) that takes them as its locus. What is required to achieve detachment from the five aggregates, according to Khemaka, is the following contemplation:
Although a noble disciple might have abandoned the five lower fetters, it might occur to him that the conceit (māno), intention (chando) and underlying tendency (anusayo) ‘I am’ (asmī ti) with regard to the five aggregates of attachment has not been destroyed. At another time, he immerses himself in observing the rise and fall of the five aggregates

I give the standard translation of the compound upādānakkhandha, but for a more detailed historical explanation see Gombrich 1996: 67 and Wynne 2007: 84. 55 SN III.128.29: pañc’ ime āvuso upādānakkhandhā vuttā Bhagavatā, seyyathīdaṃ: rūpupādānakkhandho … pe … viññāṇupādānakkhandho. imesu khv āhaṃ āvuso pañcasu upādānakkhandhesu na kiñci attānaṃ vā attaniyaṃ vā samanupassāmi, na c’ amhi arahaṃ khīṇāsavo. api ca me āvuso pañcasu upādānakkhandhesu asmī ti adhigataṃ, ayam aham asmī ti na ca samanupassāmī ti.

54

The ātman and its negation

121

of attachment: ‘form … feeling … apperception … volitions … consciousness is thus, its arising is thus, its fading away is thus.’ In doing this the conceit, intention and underlying tendency ‘I am’ with regard to the five aggregates of attachment that had not been destroyed is destroyed.56

The practice mentioned here – the contemplation of conditioned experience (the five aggregates) as a process – seems to be an attempt to see the truth of the Not-Self teaching at a deeper level: the bhikkhu does not simply think about the insubstantiality of the five aggregates, but attempts to see this truth experientially. According to the Khemaka Sutta, such a contemplation is a more powerful means of overcoming the subtle sense of identification with conditioned experience, but the aim of both contemplations is, however, the same, i.e. detachment from the five aggregates leading to the cessation of identification with them. All of the above passages on self-consciousness elaborate the typically Buddhist understanding that desire and ignorance cause suffering. That self-consciousness is a problem of an affective nature is easy to understand, of course, for self-consciousness implies self-centredness which in turn implies psychological states – selfishness, desire etc. – which are, according to the Buddhist analysis, ethically and spiritually harmful. But if the fundamental problem of desire – the sole cause of suffering according to the Second Noble Truth – can be controlled and suppressed via various religious practices, should it matter that the person who suppresses it is still self-conscious? According to the Khemaka Sutta it does, for the locus of the notion “I am” is the five aggregates. This means that the self-conscious person is inevitably attached to conditioned
SN III.130.28: kiñcāpi āvuso ariyasāvakassa pañcorambhāgiyāni saññojanāni pahīnāni bhavanti atha khv assa hoti: y’ eva pañcasu upādānakkhandhesu anusahagato asmī ti māno asmī ti chando asmī ti anusayo asamūhato. so aparena samayena pañcasu upādānakkhandhesu udayabbayānupassī viharati: iti rūpaṃ, iti rūpassa samudayo, iti rūpassa atthagamo; iti vedanā … iti saññā … iti saṅkhārā … iti viññāṇaṃ, iti viññā ṇassa samu­ dayo, iti viññāṇassa atthagamo ti. tass’ imesu pañcasu upā dānakkhandhesu udayabbayānupassino viharato, yo pi ’ssa hoti pañcasu upādānakkhandhesu anusahagato asmī ti māno asmī ti chando asmī ti anusayo asamūhato, so ’pi samugghātaṃ gacchati.
56

122

Alexander Wynne

experience, and if so it is not enough to conquer one’s desires and abide in a state of altruism, for self-consciousness is by its very nature a subtle form of clinging to conditioned experience: as we have seen, self-consciousness is a subtle, underlying form of grasping after individual experience and identity, a basic existential ‘conceit’ (p. 114). Being self-conscious, in other words, means to be both ignorant and attached, and so subject to suffering. These teachings on personal identity seem to have the same pragmatic point as the Not-Self teaching, i.e. the comprehension and abandonment of the underlying affective and cognitive cause of suffering: self-consciousness. And just like the Not-Self teaching, these teachings seem to have no obvious metaphysical significance: it is not clear if detachment from the five aggregates and transcendence of self-consciousness means that a person transcends intrinsic identity per se or whether he has some other sort of transcendent identity. This is true even of the texts that describe how Sāriputta attained certain meditative states without being aware of it, apparently because he lacked self-consciousness: the focus of these texts is the psychology of Sāriputta rather than more abstract concerns, such as the ontological nature of the state attained by him. Attempts to read a particular metaphysic into such texts are far from persuasive. Pérez-Rémon, for example, has argued that two kinds of personal identity can be indentified in the series of texts on Sāriputta’s meditative attainments. On the one hand there is the socalled “asmimanic self” which is “contained in expressions such as ‘I am attaining the first jhāna, I have attained the first jhāna, I have emerged from the first jhāna’ …”57 On the other hand, PérezRémon believes that forms of the first personal pronoun in statements such as “I dwell having attained the first state of meditation” (idhāhaṃ … pathamaṃ jhānaṃ upasampajja viharāmi) and “it [did not occur] to me thus, venerable sir” (tassa mayhaṃ, āvuso, na evaṃ hoti) is “incompatible with the ‘asmimanic I’ contained in the expressions that follow.” He therefore argues that
Sāriputta is able to say ‘I dwell having attained the cessation of awareness and feeling.’ If the ‘I’ of this sentence cannot stand either for the
57

Pérez-Rémon 1980: 236.

The ātman and its negation

123

the asmimanic self or for the genuine empirical moral agent, what kind of self does it stand for? Nothing is left but to say that it stands for the true self who, in that condition attains to a complete aloofness from the empirical factors in the isolation that is his very being.58

This argument is not very convincing. Pérez-Rémon supposes that there is a metaphysical difference between the subject of the verb in expressions such as paṭhamajjhānaṃ upasampajja viharāmi, which apparently indicates the “true self,” and expression such as ahaṃ paṭhamajjhānaṃ samāpajjāmī ti, which apparently indicates the “asmimanic” or phenomenal self. Such a distinction is entirely arbritary, and does not convince that an intrinsic identity is presupposed by the early Buddhist teachings. Indeed key phrases, such as tassa mayhaṃ āvuso na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti, seem to cancel out the positive language that they follow, such as paṭhamajjhānaṃ upasampajja viharāmi. This would seem to incline towards the position that there is no true subject of experience. A better argument, then, would in fact be that since Sāriputta is unaware of what he experiences, these passages imply the No Self doctrine. Although the Saṃyutta-Nikāya texts on Sāriputta’s meditative attainments perhaps incline towards the No Self doctrine, another argument for intrinsic identity in early Buddhism can be made. This is that since the Not-Self teaching considers only the experiential aspects of the human being (the five aggregates), it might leave room for a non-phenomenal or transcendental self. Indeed the five aggregates are presented as objects of identification or appropriation for the perceiving subject, who is denoted by the term “I” in the expression “I am not this.” It is possible that such an “I” could stand for a transcendent subject of experience. This being the case, the arguments for or against a self would seem to be well-balanced. There are further passages to consider, however, and these are more useful in determining the metaphysical presuppositions of the early Buddhist texts on personal identity.

58

Pérez-Rémon 1980: 237, on which see Oetke 1988: 107–109.

124

Alexander Wynne

4. Dependent origination and self-consciousness
We have seen that the Khemaka Sutta advises a contemplation of the rise and fall of the five aggregates as a means of achieving the cessation of the notion “I am” with regard to them. In this context, the “I” that ceases would seem to be a person’s notion of an individual identity comprised of form, sensation, apperception and so on, for it is this that is undermined by seeing that the five aggregates are impermanent. If so, the text would not seem to deny that subjectivity remains intact in the liberating experience, and the possibility remains that it presupposes a subjectivity abstracted from the five aggregates as a person’s real self, i.e. a sort of transcendent “I” that experiences detachment from the five aggregates. The same could be said for the Not-Self teaching: the statement ‘I am not this’ might presuppose a transcendent “I” beyond the five aggregates. In another respect, however, this point is not so clear. For the Khemaka Sutta goes on to compare the notion “I am” to the subtle persistence of a flower’s scent, a simile in which the “I” seems to indicate simple subjectivity:
‘It is just like the scent of a blue lotus, a red lotus or a white lotus. Would a person be describing it correctly if he were to say that the scent belongs to its leaves, its colour or filaments?’ ‘It is not so, venerable sir.’ ‘How, then, would one describe if correctly? If one were to describe it correctly, one would say that the scent belongs to the flower. In just the same way, venerable sirs, I do not declare ‘I am’ with regard to or apart from form … feeling … apperception and consciousness. And yet I still have the notion of ‘I am’ (asmī ti) with regard to these five aggregates of attachment, venerable sirs, despite the fact that I do not think ‘I am this’.’59
SN III.130.13: seyyathā pi āvuso uppalassa vā padumassa vā puṇḍarīkassa vā gandho. yo nu kho evaṃ vadeyya: pattassa gandho ti vā, vaṇṇassa gandho ti vā, kiñjakkhassa gandho ti vā, sammā nu kho so vadamāno vadeyyā ti? no h’ etaṃ āvuso. yathākathaṃ panāvuso sammāvyākaramāno vyākareyyā ti? pupphassa gandho ti kho āvuso sammāvyākaramāno vyākareyyā ti. evam eva khv āhaṃ āvuso na rūpaṃ asmī ti vadāmi, na pi añ­ ñatra rūpā asmī ti vadāmi, na vedanam … na saññam … na saṅkhāre … na viññāṇam asmī ti vadāmi, na pi aññatra viññāṇā asmī ti vadāmi. api ca me
59

The ātman and its negation

125

Khemaka’s primary point is that the locus of the notion “I am” is the five aggregates as a whole, in a manner comparable to how a scent lingers around the whole flower. But the simile also points out that just as scent emerges from the whole flower, so too does the notion “I am” arise from conditioned experience as a whole. This might indicate the understanding that the notion “I am” is an emergent state of consciousness, one in which the “I” is felt to stand apart from its objective locus as a quasi-independent subject of experience. The text is therefore ambiguous: it is not entirely clear if the term “I” refers to a person’s sense of being a composite entity made up of different phenomenal aspects, or whether it refers to a person’s sense of being a quasi-independent subject of consciousness that observes the different aspects of conditioned experience. Forming a correct understanding of this ambiguity is vitally important. For if the early Buddhist texts understand the “I” in expressions such as asmimāna, ahaṃkāra, asmīti and so on in the sense of a quasi-independent subject of consciousness, they would imply that liberation involves the cessation of subjectivity per se, in which case there would be little possibility that a person has a self. Just as ambiguous as the Khemaka Sutta is the Ānanda Sutta, in which the bhikkhu Puṇṇa teaches that the notion “I am” occurs only in relation to (upādāya) the five aggregates, rather than independently (anupādāya).60 This could indicate the understanding that the “I” is a quasi-independent observer of the five aggregates, rather than a person’s sense of being an “I” made up of the five aggregates. On the other hand, however, this text also includes the Not-Self teaching.61 Since this teaching deconstructs a person’s
āvuso pañcasu upādānakkhandhesu asmī ti adhigataṃ, ayam aham asmī ti na ca samanupassāmi. In the second sentence reading pattassa gandho ti vā, vaṇṇassa gandho ti vā, kiñjakkhassa gandho ti vā with CSCD instead of the PTS pattassa gandho ti, vaṇṇassa gandho pi, kiñjakkhassa gandho ti vā. 60 SN III.105.10: Puṇṇo nāma āvuso āyasmā Mantāniputto amhākaṃ navakānaṃ sataṃ bahūpakāro hoti. so amhe iminā ovādena ovadati: upādāya āvuso Ānanda asmī ti hoti, no anupādāya. kiñ ca upādāya asmī ti hoti, no anupādāya? rūpaṃ upādāya asmī ti hoti, no anupādāya. vedanaṃ … saññaṃ … saṅkhāre … viññāṇaṃ upādāya asmī ti hoti, no anupādāya. 61 SN III.105.25.

126

Alexander Wynne

identification with the five aggregates, and since it is here presented as a means of overcoming the notion “I am,” the term “I” would here seem to denote the phenomenal person as a whole, i.e. the individual “I” understood as an aggregate of five experiential parts. On balance, the presence of the Not-Self teaching in the Ānanda Sutta, probably indicates the latter possibility, i.e. that the term “I” refers to a person’s sense of individuality comprised of different phenomenal aspects. Furthermore, since the Khemaka Sutta’s contemplation of the five aggregates is simply an expansion of the NotSelf teaching, the same conclusion probably applies to it. Other texts definitely do not share this understanding, however, but seem to veer more towards taking the term “I” as a person’s sense of being a quasi-independent subject of experience. They therefore suggest a relationship of dependence or emergence rather than identification between the “I” and the five aggregates. One such text is the Vīṇopama Sutta, which uses the simile of the lute (vīṇā) and its sound to describe the relationship between the subjective aspect of self-consciousness and the five aggregates:
This thing called a lute, venerable sir, consists of many different components so that when played it makes a sound by means of them: dependent on the parchment sounding board, the belly, the arm, the head, the strings, the plectrum and the appropriate effort of the musician, this lute, venerable sir, which consists of many various components, is played and makes a sound by means of them.62

In this simile the different parts of the lute denote the different aspects of a person’s phenomenal being, whereas the sound that emerges from them denotes the notion “I am,” i.e. a person’s sense of subjectivity. The simile of the lute thus suggests that the sense of being an inner perceiver emerges from the different aspects of conditioned experience functioning as a whole, so that an apparSN IV.197.11: ayaṃ kho bhante vīṇā nāma anekasambhārā mahāsambhārā anekehi sambhārehi samāraddhā vadati, seyyathidaṃ: doṇiñ ca paṭicca, cammañ ca paṭicca, daṇḍañ ca paṭicca, upaveṇañ ca paṭicca, tantiyo ca paṭicca, koṇañ ca paṭicca, purisassa ca tajjaṃ vāyāmam paṭicca evāyaṃ bhante vīṇā nāma anekasambhārā mahāsambhārā anekehi sambhārehi samāraddhā vadatī ti. Following the translation of Bodhi (2000: 1254). On this simile see Collins 1982: 101.
62

The ātman and its negation

127

ently independent subject of experience is constructed out of an insubstantial process. This implies that the subjective aspect of selfconsciousness does not exist apart from the five aggregates, despite the cognitive separation of functions (subject vs. object) that might give this impression. In the language of Buddhist philosophy, this means that the “I” is dependently originated and so not ultimately real. If so, a contemplation of the five aggregates could be used to emphasise the fact that the inner subject is dependent on an insubstantial process, and in this way lead to the cessation of a person’s sense of being a quasi-independent subject of experience. Such a contemplation is suggested in the Vīṇopama Sutta, which states that the notions “I” (ahan ti), “mine” (maman ti) and “I am” (asmī ti) can be transcended by investigating the limits of conditioned experience as follows:
In just this way, bhikkhus, the bhikkhu investigates form to its full extent, he investigates feeling … apperception … volitions ... [and] consciousness to its full extent. When he does this, the notions ‘I’, ‘mine’ and ‘I am’ are found in him no longer.63

This passage suggests that by contemplating the limitations of conditioned experience, the subjective aspect of self-consciousness – the “I” that perceives the five aggregates – ceases to function. This contemplation is similar to the contemplation of the rise and fall of the five aggregates outlined in the Khemaka Sutta. But the purpose here seems to be that of emphasising the limitations of that on which the notion “I” is founded. This seems to show that the “I” is limited to the impermanent processes of conditioned experience, and so cannot be separated from them. Dependence on the five aggregates, then, would here seem to indicate that the subjective aspect of self-consciousness is not independent or ultimately real.

SN IV.197.25: evam eva kho bhikkhave bhikkhu rūpaṃ samanvesati yāvatā rūpassa gati, vedanaṃ samanvesati … pe … saññaṃ … saṅkhāre … viññā ṇaṃ samanvesati yāvatā viññāṇassa gati. tassa rūpaṃ samanvesato … pe … saññaṃ … saṅkhāre … viññāṇaṃ samanvesato yāvatā viññāṇassa gati, yam pi ’ssa taṃ hoti ahan ti vā maman ti vā asmī ti vā, tam pi tassa na hotī ti. Reading samanvesati and samanvesato with CSCD for PTS samane­ sati and samanesato respectively.

63

128

Alexander Wynne

Suggestive as they are, the similes of the lute and flower do not explicitly state that the term “I” is to be understood as a quasi-independent subject of experience. There is still some room to doubt, then, that these texts deny the notion of an independent subject of experience per se. Other texts certainly do focus on this understanding, however. We have seen, for example, that the Sāriputta Sutta describes the destruction of the underlying tendency towards conceit in the notions “I” and “mine” (asmimāna/mamaṃkāraanusaya), so that Sāriputta attains each of the nine states of meditation (anupubbavihāra) without an awareness of the fact that he attains, abides in or emerges from them.64 Such a description seems to imply the complete cessation of the subjective aspect of selfconsciousness, as if Sāriputta is in a totally impersonal and “selfless” state in which there is no sense of being an inner perceiver or “I” that observes and comprehends what is happening. In such texts there seems to be little room for a subjective aspect of consciousness that could be taken as the true self in opposition to the phenomenal self (consisting of the five aggregates). A similar understanding is suggested in the Aggivacchagotta Sutta. Like the Khemaka Sutta it advocates the practice of contemplating the rise and fall of the five aggregates, although the end result is expressed slightly differently. The key passage occurs when the Buddha responds to Vacchagotta’s question as to whether he has any views (diṭṭhigata):
‘Does the venerable Gotama have any view (diṭṭhigata)?’ The very notion of ‘view’ has been dispelled by the Tathāgata, O Vaccha, for the Tathāgata has seen this: ‘Form is thus, its arising is thus, its fading away is thus; sensation is thus, its arising is thus, its fading away is thus; apperception is thus, its arising is thus, its fading away is thus; volitions are thus, their arising is thus, their fading away is thus; consciousness is thus, its arising is thus, its fading away is thus.’ Therefore I say that the Tathāgata is released through the destruction, fading away, cessation, abandonment and relinquishment of all thoughts (maññita), agitations (mathita), and every under-

64

See n. 42 above.

The ātman and its negation

129

lying tendency towards conceit (māna) with regard to the notions ‘I’ (ahiṃkāra) and ‘mine’ (mamiṃkāra).65

Understanding the five aggregates as a process is here not said to eradicate a person’s identification with conditioned experience (the notion “I am” with regard to the five aggregates), but rather to eradicate the notions “I” and “mine.” While the understanding of the term “I” is not made clear, this emphatic description of the Buddha’s liberated state seems to indicate a complete transcendence of phenomena: the cessation of all conceptuality (maññita) implies the cessation of the entire contents of consciousness, including the sense of being a quasi-independent subject of experience. If so, the text indicates that in the Buddha’s awakened state the sense of subjectivity per se has been transcended. In support of this interpretation are a number of texts that focus on the dependent origination of the notion “I am.” One such text is the Vepacitti Sutta, which relates the myth of the defeat of the demons (asura) by the gods (deva), and the shackling of their leader Vepacitti. The doctrinal point of this myth is that the bonds of Vepacitti operate as a function of his thoughts:
O bhikkhus, when Vepacitti the leader of the demons thought ‘The gods are righteous, but the demons are not, and so I will go, right here and now, to the citadel of the gods,’ he saw that he was released from the five bonds wrapped round his neck, and being presented and endowed with the five sorts of heavenly sensual pleasure he enjoyed himself. But when, O bhikkhus, Vepacitti the leader of the demons thought ‘The demons are righteous, but the gods are not, and so I will go there right now, to the citadel of the demons,’ he saw that he was shackled by the five bonds wrapped round his neck, and so was deprived of the five sorts of heavenly sensual pleasure. That is how
65 MN I.486.10: atthi pana bhoto Gotamassa kiñci diṭṭhigatan ti? diṭṭhigatan ti kho Vaccha apanītam etaṃ tathāgatassa, diṭṭhañ h’ etaṃ Vaccha tathāgatena: iti rūpaṃ, iti rūpassa samudayo, iti rūpassa atthagamo; iti vedanā, iti vedanāya samudayo, iti vedanāya atthagamo; iti saññā, iti saññāya samudayo, iti saññāya atthagamo; iti saṅkhārā, iti saṅkhārānaṃ samudayo, iti saṅkhārānaṃ atthagamo; iti viññāṇaṃ, iti viññāṇassa samu­ dayo, iti viññāṇassa atthagamo ti. tasmā tathāgato sabbamaññitānaṃ sabbamathitānaṃ sabba-ahiṃkāramamiṃkāramānānusayānaṃ khayā virāgā nirodhā cāgā paṭinissaggā anupādā vimutto ti vadāmī ti.

130

Alexander Wynne

subtle the bonds of Vepacitti are, O bhikkhus, but even subtler is the bond of Māra. The person who thinks, O bhikkhus, is shackled by Māra – but by not thinking he is released from the Evil One.66

There is perhaps no clearer Buddhist text on the notion that thinking or conceptualisation (maññita) causes bondage. The text goes on to state that the foundation of this conceptual bondage is the subjective aspect of self-consciousness:
‘I am,’ O bhikkhus, is a thought (maññita); ‘I am this,’ O bhikkhus, is a thought; ‘I will be,’ O bhikkhus, is a thought; ‘I will not be,’ O bhik­ khus, is a thought; ‘I will possess form,’ O bhikkhus, is a thought; ‘I will not possess form,’ O bhikkhus, is a thought; ‘I will be conscious,’ O bhikkhus, is a thought; ‘I will be unconscious,’ O bhikkhus, is a thought; ‘I will be neither conscious nor unconscious,’ O bhikkhus, is a thought. Thought, O bhikkhus, is an illness, a boil and a barb. Therefore, O bhikkhus, you should train yourselves with the thought ‘I will pass my time with a mind free from thinking.’67

A similar analysis is found in the Dhātuvibhaṅga Sutta, which adds to this the point that thinking itself (maññita), especially in terms of the notion “I am” (asmī ti), is fundamentally problem66 SN IV.202.6: yadā ca kho bhikkhave Vepacittissa asurindassa evaṃ hoti: dhammikā kho devā, adhammikā asurā, idh’ eva dānāhaṃ devapuraṃ gacchāmī ti, atha kaṇṭhe pañcamehi bandhanehi muttam attānaṃ samanupassati, dibbehi ca pañcahi kāmaguṇehi samappito samaṅgībhūto paricāreti. yadā ca kho bhikkhave Vepacittissa asurindassa evaṃ hoti: dhammikā kho asurā, adhammikā devā, tatth’ eva dānāhaṃ asurapuraṃ gamissāmī ti, atha kaṇṭhe pañcamehi bandhanehi baddham attānaṃ samanupassati, dibbehi ca pañcahi kāmaguṇehi parihāyati. evaṃ sukhumaṃ kho bhikkhave Vepacittibandhanaṃ, tato sukhumataraṃ Mārabandhanaṃ. maññamāno kho bhikkhave baddho Mārassa, amaññamāno mutto pāpimato. 67 SN IV.202.20: asmī ti bhikkhave maññitam etaṃ, ayam aham asmī ti maññitam etaṃ, bhavissan ti maññitam etaṃ, na bhavissan ti maññitam etaṃ, rūpī bhavissan ti maññitam etaṃ, arūpī bhavissan ti maññitam etaṃ, saññī bhavissan ti maññitam etaṃ, asaññī bhavissan ti maññitam etaṃ, nevasaññī nāsaññī bhavissan ti maññitam etaṃ. ṃaññitaṃ bhikkhave rogo, maññitaṃ gaṇḍo, maññitaṃ sallaṃ. tasmāt iha bhikkhave amaññamānena cetasā viharissāmī ti, evañ hi vo bhikkhave sikkhitabbaṃ. Reading saññī bhavissan ti maññitam etaṃ with CSCD instead of PTS saññī bhavissan ti; rogo with CSCD instead PTS rāgo; and amaññamānena with CSCD instead of PTS amaññitmānena.

The ātman and its negation

131

atic – “an illness, a boil and a barb” that ought to be transcended (samatikkama).68 The reason for this critique is made clear in the Vepacitti Sutta, which explains that each of the notions beginning with “I am” is “an impulse” (iñjita), “a palpitation” (phan­ dita), a “conceptual proliferation” (papañcita) and “a conceit” (mānagata).69 All these terms indicate that the various manifestations of the subjective aspects of self-consciousness – the “I” as a quasi-independent observer of phenomena – arise in dependence on the conceptual activity of the mind. This is especially true of the term papañcitaṃ. In contrast to the Brahminic notion of prapañca as the manifoldness or diversity of the external world,70 the term in early Buddhist texts refers to the tendency of the mind towards conceptual diffuseness or proliferation.71 If so, it would seem that the subjective aspect of self-consciousness is conceptually constructed in the processes of the dependent origination of consciousness, and thus has no independent reality. The clearest explanation of the dependent origination of conceptual proliferation (papañca) is found in the Madhupiṇḍaka Sutta:
Visual consciousness arises dependent on the eye and forms, the coming together of the three is contact, sensation arises from contact, one apperceives (sañjānāti) what one senses, thinks over (vitakketi) what one apperceives, and conceptually proliferates (papañceti) what
MN III.246.11: asmī ti bhikkhu maññitam etaṃ, ayam aham asmī ti maññitam etaṃ, bhavissan ti maññitam etaṃ, na bhavissan ti maññitam etaṃ, rūpī bhavissan ti maññitam etaṃ, arūpī bhavissan ti maññitam etaṃ, saññī bhavissan ti maññitam etaṃ, asaññī bhavissan ti maññitam etaṃ, nevasaññīnāsaññī bhavissan ti maññitam etaṃ. maññitaṃ bhikkhu rogo, maññitaṃ gaṇḍo, maññitaṃ sallaṃ. sabbamaññitānaṃ tv eva bhikkhu samatikkamā muni santo ti vuccati. 69 SN IV.202.28: asmī ti bhikkhave iñjitam etaṃ, ayam aham asmī ti iñji­ tam etaṃ, bhavissan ti iñjitam etaṃ, na bhavissan ti iñjitam etaṃ, rūpī bha­ vissan ti iñjitam etaṃ, arūpī bhavissan ti iñjitam etaṃ, saññī bhavissan ti iñjitam etaṃ, asaññī bhavissan ti iñjitam etaṃ, nevasaññīnāsaññī bhavissan ti iñjitam etaṃ. The text then repeats this passage but replaces iñjitam etaṃ with phanditam etaṃ, papañcitam etaṃ and finally mānagatam etaṃ. 70 MMW s.v.: ‘expansion, development, manifestation … manifoldness, diversity.’ See also Gombrich 2009: 205–206. 71 The standard study is that of Ñānananda 1971.
68

132

Alexander Wynne

is thought over. From conceptual proliferation comes reckoning, of one’s conceptual proliferations and apperceptions, and this afflicts a man with regard to the past, present and future forms cognised by the eye.72

This passage states that the conceptual forms constructed in the cognitive process ultimately cause a person’s suffering. Selfconsciousness arises in this way: according to the Vepacitti Sutta its subjective aspect – the various forms in which the notion “I” is expressed – is a form of conceptual “proliferation” or “manifoldness” (papañcita). This implies, then, that the subject of self-consciousness does not exist beyond particular cognitive events. Such an analysis leaves little room for an inherently real self denoted by the term “I.” Indeed the Vepacitti Sutta’s comprehensive account of the forms in which this notion occurs seems to indicate that there is no true “I” behind its appearances in thought. At the least, there is very little ground on which this case could be made. That the notion “I am” indicates subjectivity per se, and that this is dependently originated, is made explicitly clear in an important section of the Mahānidāna Sutta which analyses three notions of intrinsic identity.73 The first is the simplest: a self (attan) identical to sensations (vedanā) is dismissed since this would mean that intrinsic identity is changeable, i.e. a contradiction is terms.74 The second and third understandings of the self are more subtle, however. The latter seems to reject the notion that the subjective aspect of self-consciousness is independently real:

MN I.111.35: cakkhuñ c’ āvuso paṭicca rūpe ca uppajjati cakkhuviññā ṇaṃ, tiṇṇaṃ saṅgati phasso, phassapaccayā vedanā, yaṃ vedeti taṃ sañjānāti, yaṃ sañjānāti taṃ vitakketi, yaṃ vitakketi taṃ papañceti, yaṃ papañceti tatonidānaṃ purisaṃ papañcasaññāsaṅkhā samudācarati atītānāgatapaccupannesu cakkhuviññeyyesu rūpesu. It is not clear exactly how the compound papañcasaññāsaṅkhā is to be taken. Ñāṇamoli and Bodhi (1995: 203) translate it as “perceptions and notions [born of] mental proliferation,” although this misses the fact that the compound is declined in the singular number. 73 For an analysis of these teachings see Oetke 1988: 130ff. 74 DN II.67.12: aniccaṃ sukhadukkhavokiṇṇaṃ uppādavyayadhammaṃ attānaṃ …

72

The ātman and its negation

133

‘Therein, Ānanda, to the person who claims that his self (attan) is different from sensation (vedanā) but not without experience (no pi appaṭisaṃvedano), it being able to sense (attā me vediyati) and having sensations as a property (vedanādhammo hi me attā), one should say: “When sensation has completely and utterly ceased without remainder (vedanā va hi āvuso sabbena sabbaṃ sabbathā sabbaṃ aparisesā nirujjheyuṃ), when there is no sensation whatsoever since it has ceased (sabbaso vedanāya asati vedanānirodhā), is it possible in that state to have the notion ‘I am this’ (ayam aham asmī ti)’?’ ‘It is not so, master.’ ‘Therefore, Ānanda, it is because of this reason that it is not suitable to think that one has a self different from sensation but not without experience, it being able to sense and having sensations as a property.’75

This passage rejects the notion of an independently real subject of perception. The problem with such a notion is that although it is possible to conceive this understanding of individual identity when conditioned experience (sensation: vedanā) functions normally, this is not the case in the absence of these conditions. This critique therefore makes the point that the sense of being an inner perceiver only arises under certain conditions, those that pertain in conditioned experience, and thus rejects the notion that a person’s sense of being an independent subject of perception is ultimately real. Such an analysis leaves little room for any sort of self, for what identity could a person have apart from the subjectivity denoted by the term “I?” The only other possibility is perhaps that a person has a transcendent identity beyond conditioned experience. The problem with such a notion is made clear in the second of the Mahānidāna Sutta’s critiques of intrinsic identity. This section of the text points out the impossibility of integrating the concept of individuality (a necessary aspect of any notion of personal identity)
DN II.67.25: tatr’ Ānanda yo so evam āha: na h’ eva kho me vedanā attā, no pi appaṭisaṃvedano me attā, attā me vediyati, vedanādhammo hi me attā ti, so evam assa vacanīyo: vedanā va hi āvuso sabbena sabbaṃ sabbathā sabbaṃ aparisesā nirujjheyuṃ, sabbaso vedanāya asati vedanānirodhā, api nu kho tattha ayam aham asmī ti siyā ti? no h’ etaṃ bhante. tasmāt ih’ Ānanda etena p’ etaṃ na h’ eva kho na kkhamati ‘me vedanā attā, no pi appaṭisaṃvedano me attā, attā me vediyati, vedanādhammo hi me attā’ ti samanupassituṃ.
75

134

Alexander Wynne

with that of transcendence (which implies a completely impersonal state):
‘Therein, Ānanda, to the person who claims “my self (me attā) is beyond sensation (na ... vedanā) and experience (appaṭisaṃvedano),” one should say: “Is it possible to have the notion ‘I am’ (asmī ti) when there is no sensation whatsoever (sabbaso vedayitaṃ n’ atthi)?”’ ‘It is not so, master.’ ‘Therefore, Ānanda, it is because of this reason that it is not suitable to think that one has a self beyond feeling and experience.’76

The problem with the notion of a transcendent self, according to this analysis, is that personal identity is conceived in terms of something completely impersonal. But how could it be claimed that a transcendent self is “one’s own” (me attā) when the individualising factor of self-consciousness (asmī ti) is absent? The notion that something is “one’s own” depends on a person being self-conscious and so able to conceptually appropriate it. The absence of the conditions necessary for self-consciousness, then, would seem to render identification with a truly transcendent state impossible. A state beyond conditioned experience cannot be conceived in terms of personal identity, therefore, the latter pertaining only within certain, limited, cognitive states. These two critiques consider the same problem from different angles: the third critique points out the problem of hypostasising the inner perceiver into a transcendent entity, whereas the second critique points out the problem of individualising the transcendence of conditioned experience. Since notions of a self within conditioned experience or beyond it are both negated, this dual analysis would seem to leave no room for any sort of intrinsic identity. This suspicion is confirmed by the fact that these critiques respond to similar conceptualisations of intrinsic identity stated in
76

DN II.67.17: tatr’ Ānanda yo so evam āha: na h’ eva kho me vedanā attā, appaṭisaṃvedano me attā’ ti, so evam assa vacanīyo: yattha pan’ āvuso sab­ baso vedayitaṃ n’ atthi, api nu kho tattha asmī ti siyā’ ti? no h’ etaṃ bhante. tasmāt ih’ Ānanda etena p’ etaṃ na kkhamati ‘na h’ eva kho me vedanā attā, appaṭisaṃvedano me attā’ ti samanupassituṃ. On this teaching see Collins 1982: 99.

The ātman and its negation

135

the eighth chapter of the Chāndogya Upaniṣad. In the dialogue between Prajāpati and Indra (CU VIII.7–12), Prajāpati presents various ways of understanding the ātman to his pupil Indra, each of which is in turn rejected and replaced by a higher and more sophisticated understanding. The final two conceptualisations of the ātman match the final two notions of the attan criticised in the Mahānidāna Sutta studied above. Prajāpati first presents the ātman in its transcendent or macrocosmic aspect, i.e. as an intrinsic identity beyond conditioned experience, which Indra rejects as follows (CU VIII.11.1):
‘In the state in which a person falls into deep sleep, so that he becomes whole, completely tranquil, and does not perceive even a dream, this is the self,’ said [Prajāpati] – ‘it is the immortal free from fear, it is brahman.’ Indra then left, his heart fully satisfied. Before reaching the gods, however, he saw this problem: ‘This person certainly does not know himself – nor even these beings here – directly in the form ‘I am this’ (ayam aham asmīti); he has become completely annihilated. I see nothing beneficial in this.’77

This conceptualisation of the ātman, and Indra’s reasons for rejecting it, correspond closely to the Mahānidāna Sutta’s second critique of intrinsic identity (attan): both address the notion that a person’s true identity is to be found beyond self-consciousness, the problem being that there is no means of identification with such a state. A similar correspondence can be seen between the Chāndogya Upaniṣad’s final description of the ātman (CU VIII.12.2–5) and the Mahānidāna Sutta’s third and final critique of the attan:
(2). The wind has no body; the clouds, lightning and thunder are also bodiless. Just as these, rising up from space and reaching the highest light, emerge into their true form (3) so too does this tranquil one, rising up from this body and reaching the highest light, emerge into his true form. He is the supreme person and wanders about there laugh77 CU VIII.11.1: tad yatraitat suptaḥ samastaḥ saṃprasannaḥ svapnaṃ na vijānāty eṣa ātmeti hovāca, etad amṛtam abhayam etad brahmeti. sa ha śāntahṛdayaḥ pravavrāja. sa hāprāpyaiva devān etad bhayaṃ dadarśa. nāha khalv ayam evaṃ saṃpratyātmānaṃ jānāty ayam aham asmīti. no evemāni bhūtāni. vināśam evāpīto bhavati. nāham atra bhogyaṃ paśyāmi.

136

Alexander Wynne

ing, playing and enjoying himself with women, carriages and relatives, without being aware of this appendage of the body. Just as a draught animal is harnessed to a cart, so is this lifebreath harnessed to this body. (4). Thus the person whose faculty of vision is fixed on space, he is the subject who sees (cākṣuṣaḥ puruṣaḥ): the faculty of vision merely enables him to see. Thus the one who thinks ‘let me smell this,’ he is the self; the olfactory faculty merely enables him to smell. Thus the one who thinks ‘let me utter this,’ he is the self; the faculty of speech merely enables him to speak. Thus the one who thinks ‘let me hear this,’ he is the self; the faculty of hearing merely enables him to hear. (5). Thus the one who thinks ‘let me think this,’ he is the self, and mind is his divine faculty of vision. It is only this one who – through his mind, the divine faculty of vision – sees the pleasures to be found in the world of Brahma, and enjoys them.78

This passage places the inner perceiver of the Yājñavalkyakāṇḍa within a more developed Brahminic cosmology, so that after its separation from the body such a self is said to exist in the world of Brahma and enjoy its pleasures. Although this cosmology is entirely absent in the Mahānidāna Sutta, this Buddhist text reports the same idea, i.e. the notion that the inner perceiver constitutes a person’s intrinsic identity. There can be no doubt, then, that ideas have been shared between the two texts. Indeed, the Mahānidāna Sutta’s first critique of the attan also corresponds to the formulation of self that precedes the final two teachings of Chāndogya

78 CU VIII.12.3–5: aśarīro vāyuḥ, abhraṃ vidyut stanayitnur aśarīrāṇy etāni. tadyathaitāny amuṣṃād ākāśāt samutthāya paraṃ jyotir upasaṃpadya svena rūpeṇābhiniṣpadyante (2). evam evaiṣa saṃprasādo ’smāc charīrāt samutthāya paraṃ jyotir upasaṃpadya svena rūpeṇābhiniṣpadyate. sa utta­ mapuruṣaḥ. sa tatra paryeti jakṣat krīḍan ramamāṇaḥ strībhir vā yānair vā jñātibhir vā nopajanāṃ smarann idaṃ śarīram. sa yathā prayogya ācaraṇe yukta evam evāyam asmiñ charīre prāṇo yuktaḥ (3). atha yatrai­ tad ākāśam anuviṣaṇṇaṃ cakṣuḥ, sa cākṣuṣaḥ puruṣo darśanāya cakṣuḥ. atha yo vededaṃ jighrāṇīti sa ātmā gandhāya ghrāṇaṃ. atha yo vededam abhivyāharāṇīti sa ātmābhivyāhārāya vāk. atha yo vededaṃ śṛṇavānīti sa ātmā śravaṇāya śrotram (4). atha yo vededaṃ manvānīti sa ātmā, mano ’sya daivaṃ cakṣuḥ. sa vā eṣa etena daivena cakṣuṣā manasaitān kāmān paśyan ramate ya ete brahmaloke (5).

The ātman and its negation

137

Upaniṣad VIII. Prajāpati here teaches that the ātman is the bodily self, a notion that Indra rejects as follows:
Just as this self becomes well-adorned when the body is well-adorned, becomes well-dressed when the body is well-dressed, and decorated when the body is decorated, so too does it become blind when the body is blind, become weary when the body is wearied, and crippled when the body is crippled. It is annihilated in consequence of the body’s annihilation. I see nothing beneficial in this.79

This rejection of a bodily self is akin to the Mahānidāna Sutta’s rejection of a self consisting of sensation, a notion that is rejected because such a self would be changeable and subject to suffering.80 Both critiques thus point out that the notion of identity with a bodily self is unsatisfactory because of its transience. This correspondence confirms that the Buddhist and Brahminic texts are parallel. If so, it would seem that the Buddhist text has drawn from a Brahminic source, for it goes one step further than the Brahminic parallel by criticising the final formulation of personal identity, and as such seems to extend and supplement an already existent teaching. Furthermore, the order of the Buddhist text is peculiar. It would make better sense if the third critique (the notion that the inner perceiver constitutes a person’s intrinsic identity) preceded the second critique (the notion of a transcendent self beyond conditioned experience), for this order – the denial of a self within conditioned experience followed by the denial of a self beyond it – makes better sense of an analysis that begins with the bodily human being. That this order is not followed is odd, but can be easily explained as a response to an Upaniṣadic teaching in which the order had been determined by the need to claim superiority for the notion of the ātman as an inner perceiver. It would seem, then, that an important early Upaniṣad has been used to communicate the early Buddhist critique of the ātman/at­
CU VIII. 9.1: yathaiva khalv ayam asmiñ śarīre sādhvalaṅkṛte sādhvalaṅkṛto bhavati suvasane suvasanaḥ pariṣkṛte pariṣkṛta, evam evāyam asminn andhe andho bhavati srāme srāmo parivṛkṇe parivṛkṇo. asyaiva śarīrasya nāśam anv eṣa naśyati. nāham atra bhogyaṃ paśyāmi. 80 DN II.67.12: aniccaṃ sukhadukkhavokiṇṇaṃ uppādavyayadhammaṃ attānaṃ …
79

138

Alexander Wynne

tan. This helps clarify the meaning of the Buddhist teachings on personal identity. The final teaching of Chāndogya Upaniṣad VIII understands the self as an inner perceiver, i.e. the “unseen seer” of the Yājñavalkyakāṇḍa, and so hypostasises a person’s sense of subjectivity into an intrinsic identity. Since this idea is firmly rejected, based on the argument that it only pertains in a conditioned state of consciousness, there would seem to be no sense in which early Buddhist thought allows a self in the form of the subjective aspect of self-consciousness: a person’s sense of being an independent subject denoted by the term “I” cannot be hypostasised into a self, according to the Mahānidāna Sutta. This indicates that the early Buddhist critiques of the notions “I” and “I am” ultimately address the problem of subjectivity per se. In the end, the Buddhist teachings therefore point out that the separation of subjective and objective aspects of self-consciousness has no ultimate ontological basis. If so, there would seem to be no room in the early Buddhist analysis for any sort of self. Indeed the second critique of the Mahānidāna Sutta even criticises the notion of a transcendent self, and so suggests the impossibility of there being any such thing as intrinsic identity: such an analysis leaves no room for a self abstracted from causes and conditions. This suggests that the point of the early Buddhist critiques of the ātman, as seen in the Not-Self teaching as well as the various other analyses of self-consciousness, is finally that a person has no intrinsic identity and thus is “selfless.” If so, the difference between the Not-Self teaching and the No-Self teaching of the Vajirā Sutta would seem to be only terminological, the implicit point of the former being explicitly articulated in the latter. But before accepting this interpretation, we must first explain why the “No Self” doctrine was not expressed explicitly in the first place.

5. Cognitive conditioning and personal identity
In the above analysis of early Buddhist texts on self-consciousness, the most crucial evidence is provided by the Mahānidāna Sutta. It is this text more than any other that points towards the No Self doctrine, for if intrinsic identity cannot be found in the inner perceiver or beyond it, what sort of “self” could a person possibly have? The

The ātman and its negation

139

apparently obvious implication of this text is that a person has no intrinsic identity. But if so, why does the text not state this explicitly? Why does it instead point out that it is “not suitable” (na kkhamati) to think that there is neither a self beyond conditioned experience, nor that the inner perceiver is a self? The same feature can be seen in the Not-Self teaching, which similarly asks if it is suitable (kallaṃ nu) to regard the five aggregates as one’s self. This is peculiar: if these early critiques really are based on the No Self doctrine, it is odd that they dodge the issue with such evasive formulations. The form of these teachings must be explained before concluding that the No Self doctrine is implicit in them. The only possible reason for the failure to articulate an explicit No Self doctrine is that this abstract philosophical issue is avoided for practical purposes. The pragmatic bent of early Buddhism is well attested in the canonical texts, of course: the simile of the raft and the simile of the arrow make it clear that all unnecessary speculation is better off avoided by the person who wishes to attain Nirvana.81 This pragmatic minimalism is even the reason given for the failure to affirm or deny the self’s existence in one canonical text, where the Buddha explains that he did not answer Vacchagotta’s direct questions because he did not wish to confuse him.82 This approach, in which the psychological well-being of a spiritual seeker is deemed more important than an abstract point, aptly summarises the early Buddhist approach to philosophical discourse: all that is not directly connected to achieving the cessation of suffering is not a proper subject of early Buddhist thought.83 Such a didactic approach would seem to explain the lack of a direct ontological assertion in both the Not-Self teaching and the Mahānidāna Sutta, for the purpose of both is to effect an existential detachment
The simile of the raft is found at MN I.134.30, on which see Gombrich 1996: 23–25. The simile of the arrow is found at MN I.429.2ff., on which see Gethin 1998: 66–67. 82 SN IV.400–401, on which see Gethin 1998: 161. 83 This approach is aptly summarised when the Buddha compares what he has taught to the few leaves in his hand, whereas what he knows is compared to all the leaves in a forest grove (SN V.437–438). For a different interpretation of this simile, see p. 167 below.
81

140

Alexander Wynne

that paves the way for liberation.84 If so, there are good reasons for thinking that the general failure of the early texts to assert the non-existence of the self was due to the pragmatic purpose of early Buddhist discourse, despite the fact that the most relevant teachings point towards this conclusion. This explanation is not entirely convincing, however. The problem is that there is no obvious reason why denying the existence of the self should be regarded as an unnecessary ontological speculation. For if there is no self, and if the belief in and attachment to the self is the ultimate cause of suffering, it would seem wise to indicate its non-existence as a means of helping others attain liberation. After all, subsequent generations of Buddhist thinkers in India did not have any problem in admitting this fact, and it is hard to see why this should not also have been the case in earlier times. While it might not have been suitable for the Buddha to state the non-existence of the self when asked outright by Vacchagotta, it is easy to imagine that this truth could have been revealed on other occasions when the Buddha’s interlocutors were not so likely to have been confused by the answer. Indeed the Mahānidāna Sutta seems eminently suitable as a context in which to reveal this truth. For after the Buddha has presented an opinion about the self and stated an argument against it, it seems natural to follow up this refutation with the conclusion that a person does not possess such a self: if the teaching does indeed presuppose this, there is no reason why the conclusion should not be “it is because of this reason that a person cannot have a self beyond feeling and experience” rather than “it is because of this reason that it is not suitable to think that one has a self beyond feeling and experience.” In short, if the early teachings claim that belief in the self is the principle cause of suffering, the question of its existence or non-existence is not a pointless ontological question of the kind met with elsewhere in the early Buddhist literature, e.g. whether the world is finite/eternal or not, or whether the Tathāgata exists and so on after death.

See the concluding parts of both teachings at MN I.139.11 and DN II.68.4.

84

The ātman and its negation

141

If the argument from pragmatism is not entirely satisfactory, we must try to find another explanation for the refusal to address the issue of the self’s existence. Why is this rather straightforward matter of ontology avoided in both the Not-Self teaching and the Mahānidāna Sutta, two teachings well-suited to make this point? And why is the matter avoided in all the other texts on personal identity studied above? These other texts on personal identity suggest that there is a problem with ontology itself. For we have seen although self-consciousness is sometimes considered in terms of the notion “I” (ahan ti, ahaṃkāra), it is also considered in terms of “the conceit I am” (asmimāna) or more simply “the notion I am” (asmī ti), the latter form being used in the Khemaka Sutta, the Vīṇā Sutta, the Vepacitti Sutta, the Dhātuvibhaṅga Sutta and the Mahānidāna Sutta. Since the formulation “I am” consists of a subject (the personal pronoun “I”) and a verb (-as, “to be”), it surely implies not just a critique of individual identity but also of the notion of individual existence (“I am”). The Vepacitti and Dhātuvibhaṅga Suttas even expand their critique to include the notions “I will be” (bhavissan ti), “I will not be” (na bhavissan ti) and various other ways of conceiving individual existence. The emphasis in these texts is as much on individual existence as it is on individual identity. Furthermore, since the notion of existence is inconceivable apart from the existence of individual entities, it would seem that the Buddhist texts on personal identity have a problem with the notion of existence itself. Why is this? The Vepacitti Sutta states that the various forms in which notions of personal existence are expressed are conceptualisations (maññita). This indicates that such notions – which are forms of conceptual proliferation (papañca) – pertain only under certain cognitive conditions, i.e. that they are dependently originated. This suggests the understanding that “existence” does not pertain independent of human consciousness, but is in fact a reality constructed in the cognitive process, i.e. a conceptual rather than an ultimate truth. While this would seem to provide an extreme solution to the failure to state the non-existence of the self, it at least provides a logical explanation for this peculiarity. For it suggests that the early Buddhist problem with the statement “the self does not exist” is that although it claims to report an ultimate truth (a state of affairs

142

Alexander Wynne

that is true independent of human consciousness), both the notions “self” and “existence” have no reality beyond particular cognitive conditions. And if the terms “existence” and “self” are conceptual constructions rather than ultimate truths, the statement “the self does not exist” cannot describe the way things really are, and must therefore be avoided. Radical as this idea may seem, the notion that “existence” does not pertain beyond human thought fits the Mahānidāna Sutta’s critique of personal identity very well. As we have seen, the second of these critiques denies the notion of intrinsic identity (attan) in a state beyond conditioned experience. Although it would be easy to conclude from this that there is no self beyond conditioned experience, the teaching steers away from drawing an ontological conclusion of this sort. It instead asks if a person can have the thought “I am” (asmī ti) beyond sensation (vedayita), and since this is not so it concludes that a person’s identification with a transcendent identity as “my self” (me attā ti) is impossible. The point of this is that the very notion of individual identity (attan) depends on particular cognitive events, i.e. the sensations that, arising in the process of conditioned experience, lay the cognitive foundations for self-consciousness (“I am”). In other words, in the process of conditioned experience arises the notion of individual existence, and dependent on that arises the notion of intrinsic identity. But both have no reality beyond the conditioning of human consciousness. The ultimate truth, then, is not that the “self” does not “exist,” but rather that the very notions of individual existence and intrinsic identity are dependently originated, and thus that any articulation of such concepts cannot be ultimately true; such concepts do not correspond to the way things really are. Exactly this point about the dependent origination of consciousness is made in the Brahmajāla Sutta, the first Sutta of the Dīgha-Nikāya. This discourse presents and criticises an extensive list of views, attributed to various unnamed ascetics and Brahmins, concerning the ultimate nature of the human being (attan) and the world (loka). The large number of views (sixty-two in total) and the complex manner of their presentation can obscure the ultimate point of the critique. Hayes, for example, has focused on the therapeutic as-

The ātman and its negation

143

pects of the text and so missed the philosophical point entirely.85 The philosophical point has also been missed by Fuller, who had synthesised the Brahmajāla Sutta with the Sammādiṭṭhi Sutta and concluded that the text’s critique of views is concerned only with the “knowledge of the cessation of craving.”86 Even Collins, who has noted that the focus of the Brahmajāla Sutta is the conditioned status of views, believes that this means nothing more than that the views are conditioned by the craving of their proponents. He thus comments that “It is here, par excellence, that the argumentum ad hominem, the denigration of others’ views on the ground of the character of those ‘others,’ and the argumentum ad verecundiam, the appeal to feelings of reverence and respect (for the Buddha), can be seen in Buddhist thinking.”87 Although it is true that the ascetics and Brahmins of the Brahmajāla Sutta are criticised for their craving, the focus is more on the cognitive foundations of their views rather than their affective faults. Moreover, the text does not praise the Buddha simply because of his virtuous character, but also because of his understanding and transcendence of views: faith plays no role in this discourse at all. None of these studies of the Brahmajāla Sutta, nor even that of Rhys Davids (1899: xxv–xxvii), Bhikkhu Bodhi (1978) or the
85

Hayes (1988: 48) seems to view the text as some sort of ancient self-help manual: “In an era in which various teachers are gathering disciples around them and making claims of supernatural powers and access to cosmic information that is beyond the ken of ordinary mortals, the superior knowledge of the Tathāgata consists in no more than a full awareness of his own feelings (vedanā) and the realization that tranquility is possible only by giving up being attached to them.” For a detailed study of the form of the Not-Self teaching in the Alagaddūpama Sutta see Wynne 2010: 210–211. 86 Fuller 2005: 115. Although Fuller does not clarify his understanding of the historicity of the early Buddhist literature, his attempt to synthesise the Brahmajāla Sutta with the Sammādiṭṭhi Sutta is based on the presupposition that the teachings in the Pāli Nikāyas constitute a homogeneous whole; this is also indicated by statements such as “the Pāli canon teaches …” (Fuller 2005: 157). This is seriously misconceived, however, for there is much evidence in the Pāli Nikāyas for divergent views and even debate. For a sample of such views see Wynne 2007: 117ff. and 2009a. 87 Collins 1982: 129.

144

Alexander Wynne

recent analysis of Evans (2009), comment on the fact that the views it criticises are analysed in terms of their connection with the past (pubbanta) or future (aparanta). Although it is easy to overlook this presentation of views according to their temporal significance, it is fundamental to the text’s analysis. Time is an essential aspect of human experience, it being impossible to conceive the fundamental reality of the human being and the world in which he exists apart from the notion of past, present and future. If so, the presentation of views in temporal terms surely indicates that the point is to criticise any attempt to conceptualise the existential reality of the human being and the world. The following statement, occurring immediately after the presentation of views, makes this quite clear:
Whatever ascetics or Brahmins speculate and form views about the past, future or both, declaring various sorts of opinion (adhivuttipadāni) with reference to the past and future, all do so through these sixty-two points or one of them – there is no possibility besides this.88

Whether or not one accepts the comprehensiveness of the views presented, the logic of this statement is that any attempt to understand the reality of the human being and the world in terms of the past and future falls within its critique. This means, then, that the text rejects the notion that the ultimate reality of things can be understood in terms of the concept “time.” For if this were not the case, it would surely be possible to conceive the reality of inner and outer things in terms of some sort of temporal analysis, i.e. another “possibility besides this.” And if the human being and the world cannot be understood in terms of time, their true reality cannot be conceptualised at all. The presentation of views in terms of the past and future is no accidental or convenient way of ordering ideas, then, but is intended to show that any attempt to conceptualise the ultimate existential reality of the human being and the world is impossible, and that the Buddha’s liberated understanding – to which the views are eventually contrasted – is beyond all notions of
88 DN I.39.14: ye keci bhikkhave samaṇā vā brāhmaṇā vā pubbantakappikā ca aparantakappikā ca pubbantāparantakappikā ca pubbantāparantadiṭṭhino pubbantāparanta ārabbha anekavihitāni adhivuttipadāni abhivadanti, sabbe te imeh’ eva dvāsaṭṭhiiyā vatthūhi etesaṃ vā aññatarena, n’ atthi ito bahiddhā.

The ātman and its negation

145

temporal existence. An indication of this is given in the following statement that precedes the presentation of views:
There are, bhikkhus, other matters – profound, hard to see and understand, tranquil, supreme, beyond the scope of logic, subtle, to be known by the wise – that the Tathāgata declares through his own understanding and vision. It is because of these that those who speak correctly would speak the true praise of the Tathāgata.89

The account of views is thus intended to elucidate the nature of the Buddha’s awakened understanding through a simple contrast: whatever the Buddha understands, it is entirely different from any attempt to conceptualise the way things are in terms of the past and future. In fact the sixty-two views are dominated by attempts to understand the human being and the world in spatio-temporal terms. This is most obvious with regard to the views about the world, all of which concern its spatial or temporal limits: the world is imagined to have a beginning or not (and so be eternal), or to be with or without spatial limits, or else to be a mixture of the two.90 The views about the human being are also concerned with the spatiotemporal reality of individual existence: they either comment on the ultimate temporal existence of a person (e.g. that there is an eternal but transmigratory self,91 or that the self has a beginning since it comes into existence spontaneously,92 or that the human being finds ultimate felicity within the bounds of this life),93 or his spatial existence (e.g. that there is a self consisting of consciousness within the body, this being the essential subject that experiences sensations),94 or else his spatio-temporal existence (e.g. that

DN I.12.18: atthi bhikkhave aññ’ eva dhammā gambhīrā duddasā duranubodhā santā paṇītā atakkāvacarā nipuṇā paṇḍitavedanīyā, ye Tathāgato sayaṃ abhiññā sacchikatvā pavedeti, yehi Tathāgatassa yathābhuccaṃ vaṇṇaṃ sammā vadamānā vadeyyuṃ. 90 DN I.22.17ff. 91 DN I.13.11ff. 92 DN I.28.25ff. 93 DN I.36.23ff. 94 DN I.21.16ff.

89

146

Alexander Wynne

the true self – however imagined – is realised after death,95 or that the different constituents of individual existence cease at death).96 The text thus contrasts what it imagines are all conceivable ideas about existence in time and space with the Buddha’s understanding that is “beyond the scope of logic” (atakkāvacara). The text goes on to explain this awakened understanding as follows:
The Tathāgata understands all these [views as follows], bhikkhus: ‘These points of view thus seized and grasped will have such a destiny and such an outcome.’ The Tathāgata understands this, and he understands what is beyond it (uttarītara), but he does not grasp at this understanding, and through not grasping he experiences internal quenching. Having comprehended as it really is the rise and fall of sensations, as well as the pleasure and danger in them and the release from them, the Tathāgata is released without grasping, bhikkhus.97

Rather than present another view about the spatio-temporal reality of the human being or the world, it seems that the Buddha understands “what is beyond” such conceptualisations and so is able to say something objective about them. This objective understanding takes two forms: first, the Buddha understands the effects of the various views in the form of the continued existences to which they lead; and second, the Buddha understands the structure of conditioned experience (the “rise and fall of sensations”), i.e. the cognitive conditions under which views arise. It is the latter aspect of the Buddha’s understanding – the construction and limits of views – that is taken up in the remainder of the text. It is first pointed out that the affective and cognitive state of the various “ascetics and Brahmins” who hold views renders their understanding of primary experience unreliable:

DN I.31.6ff., DN I.32.10ff., DN I.33.1ff. DN I.34.6ff. 97 DN I.39.20: tayidaṃ bhikkhave Tathāgato pajānāti: ime diṭṭhiṭṭhānā evaṃgahitā evaṃparāmaṭṭhā evaṃgatikā bhavissanti evamabhisamparāyā ti. tañ ca Tathāgato pajānāti tato ca uttarītaraṃ pajānāti, ca pajānanaṃ na parāmasati, aparāmasato c’ assa paccattaṃ yeva nibbuti viditā. vedanānaṃ samudayañ ca atthagamañ ca assādañ ca ādīnavañ ca nissaraṇañ ca yathābhūtaṃ viditvā anupādā vimutto, bhikkhave, Tathāgato.
96

95

The ātman and its negation

147

Therein, bhikkhus, whatever ascetics or Brahmins form ideas about the past, future or both, and have views about them, making all sorts of claims by means of these sixty-two statements with reference to the past and future, this is because (tad api) what is sensed (vedayita) by these venerable ascetics and Brahmins, who have no knowledge and vision and are affected by thirst, is subjected to ‘trembling’ and ‘quivering’ (paritasita­vipphandita).98

Although the grammar of this statement is complex, the terms pari­ tasita and vipphandita indicate that views depend on the cognitive processing or elaboration of primary experience (vedayita).99 That
98 DN I.41.29: tatra bhikkhave ye te samaṇabrāhmaṇā pubbantakappikā ca aparantakappikā ca pubbantāparantakappikā ca, pubbantāparantānudiṭṭhino pubbantāparantaṃ ārabbha anekavihitaṃ adhivuttipadāni abhivadanti dvāsaṭṭhiyā vaṭṭhūhi, tad api tesaṃ bhavataṃ samaṇabrāhmaṇānaṃ ajā nataṃ apassataṃ vedayitaṃ taṇhāgatānaṃ paritasitavipphanditam eva. 99 Fuller (2005: 115) translates ajanatam apassatam vedayitaṃ taṇhāgatānaṃ paritasitavipphanditam eva as “only the feeling of those who do not know and do not see […]; only the agitation and vacillation of those immersed in craving.” More recently Evans (2009: 71) writes that “each of the views is merely the feeling [vedayitaṃ] of those who do not know and see, the worry and vacillation of those immersed in craving.” It is unlikely, however, that the past participles in this construction can be taken in a nominal sense: it is more likely that the placing of the “ascetics and Brahmins” in the genitive case (samaṇabrāhmaṇānaṃ) indicates that they are agents of the verb expressed by the past participles, so that they are what Warder (1963: 57) has called “agent-genitives.” Collins’ translation (1982: 128) is closer to this, for he takes the term vedayitaṃ as a past participle and understands the correlative clause beginning tad api as follows: “… something experienced by these ascetics and Brahmins, who neither know nor see, and are subject to craving.” This problem with this translation is that it conventiently avoids the final words paritasita­vipphanditam eva. Moreover, to say that views are “experienced” is odd, and does not add anything to the analysis; it is also implausible to translate tad api as “something,” i.e. as a correlative pronoun with an indefinite sense. The main grammatical problem posed by this sentence is the correlative phrase tad api. The word tad cannot be a correlative pronoun, since such a pronoun would have to be in the plural rather than the singular number, there being a large number of diverse views held by various ascetics and Brahmins. Indeed, the fact that the phrase tad api appears in the subsequent expression tad api phassapaccayā (e.g. DN I.42.2 and following) indicates that tad is not to be taken as a neuter pronoun in agreement with vedayitaṃ. If, then, the

148

Alexander Wynne

this cognitive elaboration is spoken of in terms of “trembling” and “quivering” (paritasita­vipphandita) suggests, in fact, that a person’s conceptual grasp of things is in fact a sort of “distortion” of primary experience. Indeed we have seen that the term phandita is synonymous with the term papañcita in the Vepacitti Sutta, where both define the various forms of the notion “I am.” The Brahmajāla Sutta’s critique would thus seem to be concerned with the cognitive differentiation or conceptual proliferation of a person’s basic experience. Just as the Vepacitti Sutta describes the notion “I am” as a “palpitation” or “distortion,” so the Brahmajāla Sutta states that the same is true of all possible views about individual existence in space-time. The text goes on to point out that these statements of view ultimately depend on sense contact (DN I.43.8: tad api phassapaccayā), and that experience only comes about under these particular cognitive conditions (DN I.44.30: te vata aññatra phassā paṭisaṃvedissantī ti n’ etaṃ ṭhānaṃ vijjati). All ideas about the spatio-temporal reality of the human being and the external world are therefore contingent, the implication being that should the cognitive conditions change so too would a person’s grasp of reality. This is exactly what the text implies has happened to the Buddha: his cognitive state and subsequent grasp of reality are so utterly different that it is impossible to capture in terms of ideas about existence in time and space. Views about the human being and the world are therefore not absolute: they are ideas that pertain only under particular cognitive conditions, those that come about in the process of conditioned experience. It follows from this that the entire content of human
word tad is not a correlative pronoun that picks up a noun in the preceding clause and agrees with the term vedayitaṃ that follows, it is more plausible to take the phrase tad api as an adverbial correlative construction. Such a construction could be used to explain the reason for a preceding state of affairs, i.e. the fact that various ascetics and Brahmins state different views. In other words, it seems to have a meaning close to the adverbial sense noted by noted by Rhys Davids and Stede (ta, s.v. 4c: “therefore … that is why, now, then”); a close translation could be something like “in this case too …” The point is thus that the ascetics and Brahmins are able to state various views because their primary experience (that which they sense: vedayita) is subjected to “quivering” and “trembling.”

The ātman and its negation

149

consciousness – notions of personal identity, existence, non-existence and even space-time itself – are real only in so far as the cognitive conditions for them pertain, and have no essential reality beyond this dependently originated state of consciousness.100 In other words, what human beings assume to be objectively true facts about reality are nothing of the sort.101 The Brahmajāla Sutta thus articulates a philosophy of epistemological conditioning and its transcendence through the doctrine of the dependent origination of consciousness. This constitutes, in other words, a rejection of philosophical realism, as has been pointed out by Ronkin:
What the Buddha rejects is realism, conceptual and ontological alike: the notion that the encountered world is made up of distinguishable substances, and the linguistic theory that words refer to these substances which they represent; the conviction that our language corresponds to or mirrors a mind-independent reality. He points towards conventionalism in language and undermines the misleading character of nouns as substance-words. Whatever we know is part of the

Hamilton (2000: 169–170) has suggested something similar to this: “What is more difficult to grasp, or what is even less obvious, is that if the structure of the world of experience is correlated with the cognitive process, then it is not just that we name objects, concrete and abstract, and superimpose secondary characteristics according to the senses as described. It is also that all the structural features of the world of experience are cognitively correlated. In particular, space and time are not external to the structure but are part of it.” 101 As Ronkin has pointed out (2005: 244), according to this understanding “the boundaries of one’s cognitive process are the boundaries of one’s world: the latter is the world of one’s own experience, dependent on the workings of one’s cognitive apparatus.” Evans (2009: 80) has made the similar point that “by insisting that the 62 positions are vedayita, conditioned by phassa, leading to vedanā and so on, the Sutta implies that they are causally conditioned hence lack fully definite truth-value.” See also Hamilton 2000: 107–108: “the Buddha metaphorically relates the different aspects of what we think as the world around us to one’s subjective experience. In explaining how the khand­ has work, he focuses in particular on the fact that we cannot have access to anything else: all of our experience is mediated to us by means of them. And our ‘world’ is simply that. We cannot have access to an ‘external’ world because we cannot get outside of out experience. Our experience, then, is our world.”

100

150

Alexander Wynne

activity of language, but language, by its very nature, undermines certified knowledge.102

The Brahmajāla Sutta’s philosophy of epistemological conditioning explains the early Buddhist teachings on personal identity very well: it explains the dependent origination of notions of individual existence (as found, for example, in the Vepacitti Sutta), and indicates why the Mahānidāna Sutta and the Not-Self teaching do not deny the self’s existence outright. Furthermore, the position of the Brahmajāla Sutta at the beginning of the Dīgha-Nikāya surely indicates that it is foundational for the early texts. This suggests that the philosophy of epistemological conditioning ought to be generalised to the entire edifice of early Buddhist thought, and since this perspective explains the early teachings on personal identity, there can be little doubt that virtually all these teachings fit into a homogeneous understanding, one that can be ascribed to the same thinker(s) or period of thought. This is also indicated by the fact that the Not-Self teaching and the Mahānidāna Sutta both use early Upaniṣadic thought to elucidate new ideas: the Not-Self teaching alludes to the Upaniṣadic ātman in its transcendent or macrocosmic aspect, whereas the Mahānidāna Sutta uses the dialogue between Prajāpati and Indra (Chāndogya Upaniṣad VIII) in order to criticise the Upaniṣadic ātman in both its microcosmic and macrocosmic aspects (as an inner perceiver and transcendent essence). Furthermore, the philosophy of epistemological conditioning is complemented by a consistent understanding of religious means and ends. This understanding can be discerned in most of the texts studied above, especially those passages of Brahmajāla Sutta that deal with the Buddha’s transcendent state and the insight thought to effect it. This doctrine of transcendence is not confined to the texts on personal identity, however, but is assumed by some of the most important early Buddhist teachings.

102

Ronkin 2005: 245.

The ātman and its negation

151

6. The transcendence of cognitive conditioning
The means of attaining liberation suggested by the Brahmajāla Sutta consists of a discipline in which a person directs attention towards the process of conditioned experience (“the rise and fall of sensations”). A similar means is envisaged by both the Khemaka and Aggivacchagotta Suttas (“form is thus, its arising is thus, its fading away is thus” etc.), this being a more experiential elaboration of the Not-Self teaching: the goal of both is the correct comprehension of conditioned experience. Such a discipline is, of course, consistent with the Brahmajāla Sutta’s philosophy of epistemological conditioning. For if a person’s experiential condition and the suffering encountered therein is ultimately due to the dependent origination of consciousness, a person must understand how this state of affairs comes about in order to be released from it. The end result of such practices is the attainment of a transcendent state, that which the Aggivacchagotta Sutta states is beyond concepts (maññi­ ta) and self-consciousness (the notions “I” and “mine:” ahiṃkāra, mamiṃkāra). The Brahmajāla Sutta similarly understands that this state is beyond views (tato utarītaraṃ) and the scope of logic (atakkāvacarā). Although such statements on religious means and ends indicate that the liberated state is beyond conceptualisation, they also imply that it is beyond conditioned experience per se. Indeed the Brahmajāla Sutta states that the Tathāgata is released not only because he “understands as it really is” (yathābhūtaṃ viditvā) the rise, fall, pleasure and danger of “sensations,” but also because he undestands the release from them (nissaraṇa).103 This point is made explicit towards the end of the text when a bhikkhu is said to be liberated by understanding the “rise, fall, pleasure, danger and release from the six spheres of contact.”104 Perhaps the meaning of this is that in such a mindfully aware person, there is no distortion of primary experience. The implication of this is that transitive consciousness (viññāṇa) must be transcended too, since
See n. 97 above. DN I.45.22: yato kho bhikkhave bhikkhu channaṃ phassāyatanānaṃ samudayañ ca atthagamañ ca assādañ ca ādīnavañ ca nissaraṇañ ca yathābhūtaṃ pajānāti, ayaṃ imehi sabbeh’ eva uttarītaraṃ pajānāti.
104 103

152

Alexander Wynne

this is an essential aspect of sense-contact (phassa).105 This idea is expressed in the Alagaddūpama Sutta – certainly the most important canonical discourse on the Not-Self teaching106 – when the Buddha describes the transcendent state of the liberated bhikkhu who has understood the Not-Self teaching as follows:
Therefore, bhikkhus, I say that when Indra, Brahma, Prajāpati and the gods search for the bhikkhu thus released in mind (evaṃ vimuttacit­ tam), they cannot establish that ‘the consciousness of the Tathāgata is located here.’ What is the reason for this? As soon as the teaching is realised, bhikkhus, I say that a Tathāgata is untraceable.107

This statement would seem to indicate the understanding that the liberated bhikkhu is devoid of transitive consciousness, this being the reason for the failure of the gods to locate him. Such a statement on the inability to find the liberated bhikkhu is akin to the notion that the Tathāgata is indefinable, as implied, for example, by the refusal to answer certain questions about his existential state. These questions form the final four of the well-known set of ten unanswered (avyākata) questions: the first four concern the eternality (or not) and finitude (or not) of the world,108 the next two ask whether the soul or life principle (jīva) is the same as the body (or not), and the final four are concerned with the Tathāgata’s existential status after death (whether he exists, does not exist, both exists and does not exist, or neither exists nor does not exist).109 According to

According to the standard explanation of the Madhupiṇḍaka Sutta (MN I.111.35ff.), conditioned experience begins as follows: cakkhuṇ c’ āvuso paṭicca rūpe ca uppajjati cakkhuviññāṇaṃ, tiṇṇaṃ saṅgati phasso, phassa­ paccayā vedanā. On this passage see n. 72 above. 106 For the argument that this text is probably the source of the Not-Self teaching, see Wynne 2010. 107 MN I.140.3: evaṃvimuttacittaṃ kho bhikkhave bhikkhuṃ sa-Indā devā sa-Brahmakā sa-Pajāpatikā anvesaṃ nādhigacchanti: idaṃ nissitaṃ Tathāgatassa viññāṇaṇ ti. taṃ kissa hetu? diṭṭhe vāhaṃ bhikkhave dhamme Tathāgataṃ ananuvejjo ti vadāmi. 108 On the first four questions, and the textual tradition regarding the unanfirst swered questions, see Collins 1982: 131, n. 1. 109 Collins 1982: 131–133.

105

The ātman and its negation

153

Collins these questions are not answered since they are “linguistically ill-formed,” the problem being that they use
personal referring terms, which according to Buddhist thinking have no real referrent; hence, any answer given directly to them would confirm the misleading presupposition that such terms do refer to some real individual.110

In explaining that the problem posed by these questions is that there is “no real referrent” to terms such as “soul” and “Tathāgata,” Collins reads the classical “No Self” doctrine into the Buddha’s failure to answer them. On the other hand, however, he indicates that the failure to answer questions about the ontology of the world is based on a different reason, i.e. that they are pragmatically pointless.111 This explanation thus assumes the classical ontology in which the term “world” refers to a reality independent of consciousness, whereas personal terms have no ultimate referent in this world. This is problematic not only because such an ontology is not made clear in the early texts, but also because a common set of questions is explained differently. A single explanation should be found for all the questions: if there is a “linguistic problem” with the questions, as maintained by Collins, this should similarly apply to all the points they cover. A single, coherent explanation for the linguistic problem is provided, however, by the Brahmajāla Sutta’s philosophy of epistemological conditioning. According to this philosophy, all linguistic formulation – even basic concepts such as “space-time” and “existence” – have no reality beyond a person’s dependently originated state of consciousness. According to this perspective, the problem with questions about the ontology of the world and the human being is they that assume the mind-independent reality of “spacetime” and “existence,” and to answer them would be to subscribe to such notions. But as Hamilton has pointed out, the unanswered questions are based on false premises:

Collins 1982: 133. Collins 1982: 132 describes these questions as “a standard type of ‘pointless speculation’.”
111

110

154

Alexander Wynne

If, however, … space and time are part of the structural characteristics of the experiential world, and that this is cognitively dependent, then one can see that the presupposition of the transcendental reality of time and space is false, and that the fundamental premises on which the questions rest are therefore also false. What this means is that though the questions are meaningful within a conceptual framework which assumes that space and time are transcendentally real, if space and time are not transcendentally real the questions are in effect unanswerable if one wishes to be truthful. Any formulation of a response within the same conceptual framework as the questions would not truthfully reflect a reality which does not conform to that conceptual framework.112

The unanswering of these questions thus indicates the Buddha’s transcendence of dependently originated states of consciousness, such that notions of “space-time” and “existence” have no ultimate reality to him. It would seem, then, that the philosophy of the Brahmajāla Sutta thus provides a coherent explanation for all of the unanswered questions. Indeed this understanding of the unanswered questions is articulated in at least two canonical texts. In the Avyākata Vagga of the Aṅguttara-Nikāya (Sattaka Nipāta 51), the Buddha explains that each of the final four of these questions is a view endowed with thirst (taṇhāgata), apperception (or ideation: saññāgata), conceptualisation (maññita), conceptual diffuseness (papañcita), attachment (upādānagata) and regret (vippaṭisāra).113 In other words the unanswered questions are conceptualisations that have no correspondence with reality. A similar explanation is found in the Mahānidāna Sutta. After describing the liberation of the bhikkhu who understands its teachings on personal identity (those studied in section four above),114 it states that it is unsuitable (akalla) to consider this bhikkhu “thus liberated in mind” (evaṃ vimuttacittam) in terms of questions about his existence, non-existence, existence and non-existence, and neither existence nor non-

112 113 114

Hamilton 2000: 174. AN IV.68.33ff. DN II.68.4ff.

The ātman and its negation

155

existence, i.e. the final four of the ten unanswered questions.115 The reason for this is given as follows:
The extent of articulation and its range, of utterance and its range, of designation and its range, of understanding and its scope, of existence and its ‘movement’ – the bhikkhu is released from all this through higher understanding.116

This account of transcendence is in agreement with the Brahmajāla Sutta’s philosophy of epistemological conditioning outlined above: questions about the ontological state of the liberated bhikkhu cannot be answered because he has transcended the cognitive conditions on which they are founded. In other words, the problem here is with conceptuality per se: the liberated bhikkhu transcends the “path of articulation” (adhivacanapatha), “the path of designation” (paññattipatha), “the scope of understanding” (paññāvacaram) and even the notion of “existence” (vaṭṭa) and its “movement” (vaṭṭaṃ vattati), i.e. time. Indeed, the equating of the two latter concepts (“existence” and “time”) with those of “articulation,” “designation” and “understanding” indicates the understanding that space-time is only conceptually real. That the problem with the unanswered questions is as much with the notion of “existence” as it is with the notion of “personal referring terms” is similarly suggested in the Aggivacchagotta Sutta. For this text likens the inexplicability of a Tathāgata’s liberated condition to that of an extinguished flame, the point being that the fuel through which both can be designated – the five aggregates for the Tathāgata, grass and firewood for the flame – has ceased. The Tathāgata cannot be conceptualised, therefore, because he has gone beyond the experiential conditions through which an individual is normally understood (the five aggregates). And in this transcendent condition, the notions of “arising” (or being reborn: upapajjati), non arising, both arising and non-arising, and neither arising nor
DN II.68.11ff. DN II.68.18: yāvat’ Ānanda adhivacanaṃ yāvatā adhivacanapatho yāvatā nirutti yāvatā niruttipatho yāvatā paññatti yāvatā paññattipatho yāvatā paññā yāvatā paññāvacaraṃ yāvatā vaṭṭaṃ yāvatā vaṭṭaṃ vattati, tad abhiññā vimutto bhikkhu.
116 115

156

Alexander Wynne

non-arising have no relevance. This account indicates that the conceptual problem is not simply one of defining the Tathāgata, but in conceiving his individual existence. A similar problem with applying the concept of “existence” to the liberated person can be seen in the Buddha’s dialogue with Upasīva in the Pārāyanavagga (Sn 1069–1076): in response to Upasīva’s question about whether the liberated sage (muni) exists eternally or does not after death (Sn 1076), the Buddha states that he cannot be measured (na pamāṇam) because the means of speaking about him have ceased.117 In other words the liberated sage has gone beyond the concepts by which the existence of anything can be known (yena naṃ vajju tassa natthi). What all this means is that the apophatic strand strongly evident in early Buddhist thought can be explained by the notion that space-time is a relative truth transcended by a Tathāgata, the one whose condition is therefore incomprehensible, i.e. “like that.”118 The philosophy of epistemological conditioning and its transcendence thus explains some of the more puzzling aspects of early Buddhist thought: critiques of personal identity that do not commit to any ontology, the refusal to comment on the ultimate reality of the world, the avoidance of questions about the existential status of the liberated sage, the simile of the extinguished flame and so on. The Brahmajāla Sutta therefore provides a coherent explanation for the early Buddhist teachings on personal identity as well as the apophatic strand in early Buddhist thought, and so underpins a religio-philosophical understanding found consistently and extensively throughout the early Buddhist texts. Although this is not a philosophy of realism since it assumes that the reality of space-time does not extend beyond a person’s cognitive conditioning, this does not imply an idealist understanding. For idealism is still an ontology of sorts, and indeed one that can only be imagined under particular cognitive conditions; the authors of the Mahānidāna Sutta would no doubt object to this by pointing out that there would be no means of conceiving such an understanding beyond “sensation”
117 118

On this see Wynne 2007: 90ff. Gombrich 2009: 151.

The ātman and its negation

157

and self-consciousness. Indeed idealism is, basically, an hypostasisation of a person’s subjective awareness into a mind-independent, ultimate reality, an idea which the Mahānidāna Sutta rejects of this system of thought is rather that the way things really are is unspeakable and unthinkable. In other words, the Brahmajāla Sutta’s philosophy of epistemological conditioning implies that reality is ultimately ineffable, as is the state of the person who realises it by escaping his cognitive conditioning.

7. The development of reductionistic realism: Abhidharma origins
This conceptual clarification allows us to see that the metaphysical assumptions of the Brahmajāla Sutta differ considerably from those of the first attempt to systematise the Buddha’s teachings, i.e. the reductionistic realism articulated in the various Abhidharmas. This reductionism is based on the “No Self” doctrine: it assumes that although a person “exists” in the mind-independent reality of the world, he is made up of impermanent “existents” or “events” (dharmas) which lack self. The human being and the world are reduced to their constitutent parts, therefore, these being thought to lack essence but exist transiently in the objectively real domain of space-time. Traces of the change towards a proto-Abhidharma reductionism can be found in the early texts. It is clear, for example, that the Vajirā Sutta is both reductionistic as well as realistic, for it speaks of the aggregates “existing” (khandhesu santesu) and of the failure to “find” an essential being in them (na yidha sattūpalabbhati). This goes beyond the Not-Self teaching and the philosophy of the Brahmajāla Sutta by assuming that ultimate truth can be spoken of in terms of the concepts “existence” and “non-existence.” If so, it follows that the Not-Self teaching and the Vajirā Sutta are separated by an important philosophical change: whereas the former is based on the doctrine of epistemological conditioning and the relative truth of space-time, whereas the latter is based on the realistic assumption that space-time exists independent of human consciousness. This change in thought was probably complex and

158

Alexander Wynne

multi-faceted, but in principle can be simply explained. For it requires only that certain Buddhist thinkers focused on the Not-Self teaching (and related ideas) at the expense of the philosophical framework provided by the Brahmajāla Sutta. In such a scenario the Not-Self teaching could easily have been taken in a realistic sense, and this would have set the foundations for the emergence of the Abhidharmic reductionism. The beginnings of such a development can perhaps be seen in the Khemaka Sutta. As we have seen, the bhikkhu Khemaka was of the opinion that “I see no sort of self (attan) or its property (attani­ ya) in these five aggregates of attachment, venerable sirs” (imesu khv āhaṃ āvuso pañcasu upādānakkhandhesu na kiñci attānaṃ vā attaniyaṃ vā samanupassāmi). Although this does not indicate any departure from the philosophy of the Brahmajāla Sutta, it certainly paves the way for a new sort of enquiry, one in which the contemplative bhikkhu analyses the different aspects of his being in the attempt to find a self. Such an approach leaves Buddhist thought on the verge of an important conceptual change: through this enquiry, factors of experience (the five aggregates) which were thought unsuitable to be regarded as “self” begin to look like impermanent factors of being which lack self. The change from the teaching that “form is not self” (this being an unsuitable conceptualisation) to the similar but subtly different idea that “no self can be found in form” (the latter being something that exists whereas the former does not) is easy to imagine, therefore, on the basis of the Khemaka Sutta. Textual evidence for this change is in fact contained in the Mahāhatthipadopama Sutta. This discourse begins with the simile of an elephants’ footprint but soon turns into an analysis of the first aggregate – form – in terms of the four material elements (earth, water, fire and wind). The Not-Self teaching is applied to these four material elements in the following manner:
What, venerable sirs, is the earth element? It might be internal or external. And what is the internal earth element? That which is internal and personal, i.e. that which is solid, hard and materially derivative, namely: head-hair, bodily hair, nails, teeth, skin, flesh, sinew, bones, bone-marrow, kidney, heart, liver, membrane, spleen, lungs, bowels, intestinal tract, stomach, faeces, and whatever else is internal and personal, i.e. that which is solid, hard and materially derivative, this,

The ātman and its negation

159

venerable sirs, is said to be the internal earth element. This very internal earth element and the external earth element are simply the earth element, which should be seen with correct understanding as it really is: “This is not mine, I am not this, this is not my self.” Once this has been seen with correct understanding as it really is, one becomes disillusioned with the earth element, one cleanses one’s mind of passion for the earth element.119

This teaching does not focus on the process of conditioned experience and its conceptual appropriation, but is rather concerned with a virtually exhaustive analysis of the physical constituents of a human being.120 This is a considerable departure from the Not-Self teaching. The point is no longer that the concept “self” arises in connection with a dynamic process of experience and is ill-suited to it, but rather that when a human being is broken down into his constituent parts all is found to be lacking in self. The Mahāhatthipadopama Sutta thus assumes a realist ontology in which the four elements really “exist” but the self does not. Furthermore, the ultimate truth of things is here captured in words, and is not something beyond logic and the conceptual construction
MN I.185.14: katamā c’ āvuso paṭhavīdhātu? paṭhavīdhātu siyā ajjhattikā siyā bahirā. katamā c’ āvuso ajjhattikā paṭhavīdhātu? yaṃ ajjhattaṃ paccattaṃ kakkhaḷaṃ kharigataṃ upādiṇṇaṃ, seyyathīdaṃ: kesā lomā nakhā dantā taco maṃsaṃ nahāru aṭṭhī aṭṭhimiñjā vakkaṃ hadayaṃ yakanaṃ kilomakaṃ pihākaṃ papphāsaṃ antaṃ antaguṇaṃ udariyaṃ karīsaṃ, yaṃ vā pan’ aññam pi kiñci ajjhattaṃ paccattaṃ kakkhaḷaṃ kharigataṃ upādiṇṇaṃ, ayaṃ vuccat’ āvuso ajjhattikā paṭhavīdhātu. yā c’ eva kho pana ajjhattikā paṭhavīdhātu, yā ca bahirā paṭhavīdhātu, paṭhavīdhātur ev’ esa. taṃ: n’ etaṃ mama, n’ eso ’ham asmi, na m’ eso attā ti evam etaṃ yathābhūtaṃ sammapaññāya daṭṭhabbaṃ. evam ataṃ yathābhūtaṃ sammapaññāya disvā paṭhavīdhātuyā nibbindati, paṭhavīdhātuyā cittaṃ virājeti. 120 Hamilton (1996: 10) is correct to point out of that this list of bodily items in the Mahāhatthipadopama Sutta is “manifestly not comprehensive,” but this does not mean that the point is to “indicate examples of the characteristics being described” in order to emphasise “the characteristics and processes which enable the living body of the human being to function.” (Hamilton 1996: 12–13. The passage is entirely lacking in any words indicating that the contemplative bhikkhu’s attention should be focused on how the human body functions. The point is rather to go into enough detail so that the bhikkhu gets the reductionistic point that a self is not found in the body, despite the fact that every known part of it is not listed.
119

160

Alexander Wynne

of consciousness, as stated in the Brahmajāla Sutta. Concepts thus capture the fact that the human being really exists in the mindindependent reality of space-time, albeit as an aggregate of various elements in which a self cannot be found. The simile of the house, found towards the end of the text, makes this reductionistic realism quite clear:
Venerable sirs, just as an enclosed space is designated ‘house’ dependent on logs, creepers, grass and clay, so too is an enclosed space designated ‘form’ dependent on bones, sinew, flesh and skin.121

This is a statement of the ultimate truth that a person’s physical being is nothing more than an accumulation of different parts in which, as the text goes on to state, are to be found sensation, apperception, volitions and consciousness.122 The reference to an enclosed space (ākāso parivārito) shows that the authors of this text regarded the human being as a construction in space-time, albeit one that lacks intrinsic identity, an idea that comes very close to the chariot simile of the Vajirā Sutta. The truth to be known is here that the human being exists but is an aggregate lacking essence, a fact that the bhikkhu should come to understand by the following means:
He understands thus: ‘Thus indeed is the coming togther, collection and accumulation of the five aggregates of attachment.’ But the Blessed One has said this: ‘The one who sees Dependent Origination sees the Dhamma, and the one who sees the Dhamma sees Dependent Origination.’ These very things are dependently originated, that is to say the five aggregates.’123

121 MN I.190.15: seyyathā pi āvuso kaṭṭhañ ca paṭicca valliñ ca paṭicca tiṇañ ca paṭicca mattikañ ca paṭicca, ākāso parivārito agāran t’ eva saṅkhaṃ gacchati, evam eva kho āvuso aṭṭhiṃ ca paṭicca nahāruñ ca paṭicca maṃsaṃ ca paṭicca cammañ ca paṭicca, ākāso parivārito rūpan t’ eva saṅkhaṃ gac­ chati. 122 MN I.190.28ff. 123 MN I.190.35: so evaṃ pajānāti: evaṃ kira ’mesaṃ pañcannaṃ upādānakkhandhānaṃ saṅgaho sannipāto samavāyo hotī ti. vuttaṃ kho pan’ etaṃ Bhagavatā: yo paṭiccasamuppādaṃ passati so dhammaṃ passati, yo dhammaṃ passati so paṭiccasamuppādaṃ passatī ti. paṭiccasamuppannaā kho pan’ ime, yadidaṃ pañca’ upādānakkhandhā.

The ātman and its negation

161

In this passage the five aggregates (upādāna) really are “aggregates,” i.e. parts out of which a person is made, these being impermanent (dependently originated) and so not intrinsically real. This shows that a realistic reading of the Not-Self teaching did indeed lead to reductionism, i.e. the notion that a person is an accumulation (samavāya) of impermanent, causally-connected elements. This understanding is entirely different from that articulated in the Brahmajāla Sutta and most of the other texts on personal identity: there is no trace in the Mahāhatthipadopama Sutta of the notion that the dynamic processes of conditioned experience is not to be understood in terms of intrinsic identity. The point, rather, is that intrinsic identity cannot be found in the different existential factors of the human being. How is this conceptual divergence to be explained? Whereas the philosophy of epistemological conditioning can be generalised to a great number of early texts, the same cannot be said of the reductionistic realism espoused in the Vajirā and Mahāhatthipadopama Suttas. This indicates that the ideas of the latter texts were marginal in the early period, which can only be explained in two possible ways: either these ideas circulated among a small sub-section of the early saṅgha, and so were not recorded in most of the early texts, or they belong to a later stage of speculation than that recorded in most of the early texts. In support of the latter hypothesis, we can note that the Aṭṭhakavagga and Pārāyanavagga, which are certainly among the earliest Buddhist texts, are generally in agreement with the philosophy of epistemological conditioning.124 In the Kalahavivāda Sutta, for example, the Buddha states (Sn 870) that existence and non-existence depend upon sense contact:
The pleasant and unpleasant originate in sense-contact, but do not arise when there is no sense-contact. I say to you that the fact of nonexistence and existence also originates in this.125

On the antiquity of these texts see Wynne 2007: 73. Sn 870: phassanidānaṃ sātaṃ asātaṃ, phasse asante na bhavanti h’ ete. vibhavaṃ bhavañ cāpi yaṃ etaṃ atthaṃ, etaṃ te pabrūmi itonidānaṃ.
125

124

162

Alexander Wynne

Later on in the same dialogue (Sn 874) the Buddha explains that “form” (rūpa) is a conceptual proliferation that depends on conceptualisation (or apperception, saññā) and can thus disappear for the religious adept:
Not cognisant of conceptualisation, not cognisant of misconceptualisation, not uncognisant but not cognisant of what is untrue: form disappears for the one who has reached this state, for the discernment of manifoldness (papañcasaṅkhā) originates in conceptualisation (saññānidānā).126

According to this enigmatic statement of the Buddha, a person’s physical being is not ultimately real, but depends on the tendency to conceptualise reality in terms of a manifold world of diversity. Elsewhere in the Aṭṭhakavagga, the seeds of the Buddha’s teachings on personal identity can be seen at the beginning of the Tuvaṭaka Sutta (Sn 915–916), when the Buddha responds to a question about the attainment of Nirvana:
‘I ask you, kinsman of the sun, great sage, about detachment and the state of peace: with what sort of vision is a bhikkhu quenched, so that he grasps at nothing in the world?’ (915) The Blessed One said: ‘The contemplative (mantā) should put a complete stop to the notion ‘I am’ (asmī ti), which is the root cause of discerning manifoldness (mūlaṃ papañcasaṅkhāyā). He should ward off whatever inner thirst he has, training himself to be ever mindful.’ (916)127

The notion of individual existence is here said to be the root cause of a person’s diverse perceptions (papañcasaṅkhā), which is simply a more emphatic way of stating the teaching of the Vepacitti Sutta, i.e. that the notion “I am” is a conceptual proliferation (papañca).
Sn 874: na saññasaññī na visaññasaññī, no pi asaññī na vibhūtasaññī: evaṃsametassa vibhoti rūpaṃ, saññānidānā hi papañcasaṅkhā. This statement reminds one of the similar statement in the Aggivacchagotta Sutta that the Buddha has “annihilated” form (rūpaṃ … anabhāvakataṃ, MN I.487.33). On this passage see n. 134 below and Wynne 2007: 95–96. 127 Sn 915–916: pucchāmi taṃ Ādiccabandhuṃ, vivekaṃ santipadañ ca Mahesiṃ: kathaṃ disvā nibbāti bhikkhu, anupādiyāno lokasmiṃ kiñci? (915). mūlaṃ papañcasaṅkhāyā ti Bhagavā, mantā asmī ti sabbam uparund­ he. yā kā ci taṇhā ajjhattaṃ, tāsaṃ vinayā sadā sato sikkhe. (916).
126

The ātman and its negation

163

Besides these Aṭṭhakavagga teachings on existential matters, in the Purābhada Sutta the Buddha explains (Sn 849) that the liberated sage is released from the very notion of time:
‘Devoid of thirst even before death,’ said the Blessed One, ‘not dependent upon the past, immeasurable in the middle, for him nothing is fashioned with regard to the future.’128

The statement that the liberated sage is “immeasurable” in the present is but a poetic way of describing the complete transcendence of time. Indeed in the Aṭṭhakavagga the notion of being “independent” or “unnattached” (anissita) can refer not just to having no inclination or fondness for something, but also to being completely devoid of the notion of something. When the Paramaṭṭhaka Sutta states that the bhikkhu should have no dependency on knowledge (Sn 800: ñāṇe pi so nissayaṃ no karoti), for example, the point is that he should transcend it. Indeed the Kalahavivāda Sutta claims that those who have different opinions about liberation are “dependent” (Sn 877: upanissitā), whereas the Buddha is released, the implication being that he is completely beyond such views. It seems, then, that when the Purābheda Sutta states that the sage is not attached (anissito) to the past, it means that he is free from the very notion of it. Indeed this text goes on to state (Sn 851) that the sage’s freedom from the past and future is connected to his transformed cognitive state and lack of views:
He is without attachment for the future and does not grieve over the past. Perceiving detachment, he is not led into sense-contacts and views.129

All of this evidence is in agreement with the Brahmajāla Sutta’s philosophy of epistemological conditioning: notions of existence, non-existence and time are said to be dependent on a person’s cognitive functioning, the release from which implies the cessation of a person’s awareness of individual existence in space-time. There are reasons for believing that such teachings go back to
128 Sn 849: vītataṇho purā bhedā ti Bhagavā, pubbam antam anissito, vemajjhe nūpasaṃkheyyo tassa n’ atthi purekkhatam. 129 Sn 851: nirāsatti anāgate atītaṃ nānusocati, vivekadassī phassesu diṭṭhīsu ca na niyyati.

164

Alexander Wynne

the Buddha himself, not only because of the Aṭṭhakavagga’s antiquity and originality,130 but also because related teachings in the Pārāyanavagga correspond very closely to the few historical facts about the Buddha that can be deduced from the early literature.131 Similar observations suggest that the Not-Self teaching is also to be ascribed to the Buddha, for it is hard to explain this highly original teaching – especially its occurrence in Alagaddūpama Sutta – as an abstract formulation of later Buddhist teachers.132 A similar antiquity cannot be assumed of the Vajirā and Mahāhatthipadopama Suttas, however. It is surely important that the Buddha does not feature in either text, and that the orator in the Mahāhatthipadopama Sutta is Sāriputta, the patron saint of the Abhidharma. Furthermore, in its statement that “… the Blessed One has said this: The one who sees Dependent Origination sees the Dhamma …,” the Mahāhatthipadopama Sutta even seems to speculate on the meaning of the Buddha’s teaching of Dependent Origination. This text’s highly complex and artificial style of analysis also suggests that it is a proto-Abhidharmic work, a text composed as Buddhist thought progressed towards a fully developed philosophy of reductionistic realism. Further evidence in support of this chronological stratification is found in the various Vinayas accounts of the beginning of the Buddha’s ministry. These texts state that the Buddha’s teaching to the five disciples concluded with the Not-Self teaching, the understanding of which triggered their instantaneous liberation. But it can be shown that the account as a whole draws upon and adapts the earlier account of the Ariyapariyesana Sutta.133 The doctrinal understanding that underpins this adaptation can be seen at the conclusion of the Second Sermon, when it is stated that the minds of the five disciples of the Buddha were “released from the corruptions without grasping” (Vin I.14.34: pañcavaggiyānaṃ
Gómez 1976: 139: “When I first read the Mahāviyūha Sutta of the Suttanipāta I was impressed not only by its freshness and directness, but also by its originality.” 131 See Wynne 2007: 127 for a summary of the evidence. 132 See Wynne 2010. 133 For this analysis see Wynne 2009a.
130

The ātman and its negation

165

bhikkhūnaṃ anupādāya āsavehi cittāni vimucciṃsu). Although this formulation of liberating insight occurs throughout the early Buddhist literature, it deviates from the Ariyapariyesana Sutta’s account of how the five bhikkhus “realised Nirvana” (MN I.173.7ff.: pañcavaggiyā bhikkhū … asaṅkiliṭṭhaṃ anuttaraṃ yogakkhemaṃ nibbānaṃ ajjhagamaṃsu). The authors of the Second Sermon thus drew upon the Ariyapariyesana Sutta but replaced its apophatic account of the five bhikkhus’ liberation with an apparently reductionistic one, the subject of liberation being the minds of the bhikkhus rather than the bhikkhus themselves. This deviation from an older, apophatic description, one that sees no problem in speaking of the human being as a whole realising Nirvana, and its replacement with a reductionistic formulation – which in Buddhist literature past and present complements the “No Self” doctrine – implies that the authors of the Second Sermon believed in the non-existence of the self. It would even seem that in this case the authors of this Vinaya account read the No Self doctrine into the Not-Self teaching. This is strong evidence for a doctrinal change from a philosophy of ineffability to that of reductionistic realism. Regardless of the strength or weakness of this theory of doctrinal change, a couple of facts can hardly be denied. First, a doctrine of ineffability is certainly contained in the early Buddhist texts, and attempts to deny this are implausible.134 And second, this phiSiderits (2007: 70–73) has argued that the Aggivacchagotta Sutta assumes a doctrine of realistic reductionism rather than ineffability. But this understanding cannot be derived from the Pāli text. Most importantly, when the Buddha claims that he, the Tathāgata, has “annihilated” the five aggregates (MN I.487.31), and so cannot be defined, instead of reading tathāgata Siderits reads the term arhat (2007: 71), and understands that it refers to a dead Buddhist saint. Thus he translates the Pāli yena rūpena Tathāgataṃ paññāpayamāno paññāpeyya taṃ rūpaṃ Tathāgatassa pahīnaṃ as follows: “all rūpa by which one could predicate the existence of the arhat, all that rūpa has been abandoned.” By shifting the focus from the living Tathāgata – i.e. the Buddha himself – to the dead arhat, the meaning of the passage is substantially changed. For it would now seem to be saying that nothing can truly be said about the dead saint because the five aggregates of which he was formerly constituted no longer exist. Thus Siderits notes that “[t]he word ‘arhat is a convenient designator,’ just like ‘fire.’ So nothing we say about the arhat can ultimately be true. The only ultimately true statement about the
134

166

Alexander Wynne

losophy is incompatible with the philosophy of reductionistic realism later outlined in the various Abhidharmas, and anticipated in the Vajirā and Mahāhatthipadopama Suttas. It can hardly be the case that both philosophies were devised by a single thinker or the same group of thinkers. One must have developed from the other, and if the philosophy of epistemological conditioning and its transcendence can be ascribed to the very beginnings of Buddhism and perhaps to the Buddha himself reductionistic realism must belong to a later period. This is in fact the most logical explanation of the evidence. The philosophy of epistemological conditioning radically subverts all our most basic presuppositions about life: it is difficult to believe that space-time is merely conceptual rather than an objective, mind-independent reality. It is not hard to imagine that this challenging philosophy was misunderstood and replaced by a sophisticated but simpler realistic philosophy. Indeed teachings based on the philosophy of epistemological conditioning, such as the Mahānidāna Sutta’s three critiques of personal identity, could easily be misunderstood in a realistic manner unless the underlying philosophy is made clear. But it is hard to see how the philosophy of epistemological conditioning could have emerged from that of reductionistic realism. In conclusion, the evidence studied here suggests that the difference between the Not-Self teaching and the Vajirā Sutta is philosophical rather than terminological. It follows from this that although the early Buddhist teachings were not presented in the form of a philosophical system, they are at least philosophically grounded. This is not to say that the Buddha should be regarded

situation will be one that describes the skandhas in a causal series … Does this mean that the arhat is annihilated – that nirvāṇa means the utter extinction of the enlightened person? No. There is no such thing as the arhat, so it lacks meaning to say that the arhat is annihilated. And for exactly the same reason, it lacks meaning to say that the arhat attains an ineffable state after death” (2007: 73). But since the passage is concerned with the living Buddha (Tathāgata) rather than the dead saint (arhat), it therefore is attempting to describe the state of being alive and yet liberated. Its negations, and the simile of the extinguished flame, indicate a completely apophatic understanding of religious experience that points towards a doctrine of ineffability.

The ātman and its negation

167

as a philosopher, however.135 For although the evidence suggests he had worked out a coherent world-view, it also indicates that he applied it as and when he saw fit, i.e. according to the pragmatic demands of the situation.136 Such an interpretation seems to fit with one of the Buddha’s most famous statements on his approach to teaching: in the fifth book of the Saṃyutta-Nikāya (Saccasaṃyutta IV: Siṃsapāvanavagga, no. 1), the Buddha states that although his knowledge is as vast as the numbers of leaves in a forest grove, the teachings he had revealed were comparable to just a few leaves.137 According to the texts studied above, this statement would seem to mean that although the Buddha had worked out a coherent philosophy, he did not teach it directly because of his pragmatic interest in helping others attain the cessation of suffering. This explains why the most important teachings studied above, such as the “Not-Self” teaching and the Mahānidāna Sutta’s critiques of personal identity, are philosophical in their method and argumentation but ultimately avoid a direct statement of philosophical truth. The same is true of the Brahmajāla Sutta: it is philosophical without stating a philosophy directly. All this seems to be the work not of a philosopher interested in abstract ideas for their own sake, but rather of a religious teacher keen to apply a philosophical understanding in order to help his followers achieve the best possible spiritual result. Early Buddhist thinkers were less philosophically parsimonious, however. In contemplating the Not-Self teaching, they came to believe in the non-existence of the self – against the explicit warnings of the Buddha. For in the Alagaddūpama Sutta, perhaps the single most important canonical exploration of the Not-Self teaching, the
For a recent discussion of the philosophical value of early Buddhist thought, see Bronkhorst 2009: 1–7. 136 Richard Gombrich has recently argued something along these lines (2009: 164): “I do not feel that Buddha was interested in presenting a philosophically coherent doctrine: the evidence that his concern was pragmatic, to guide his audience’s actions, is overwhelming. On the other hand, I have also concluded that the evidence that he had evolved such a structure of thought and that it underpinned his pragmatic advice is no less compelling.” 137 SN V.437ff.
135

168

Alexander Wynne

Buddha describes how others responded to his teachings by weeping, beating their breasts and thinking ‘I will be annihilated!’.138 Such people concluded that the Buddha had taught the non-existence of the self, although the Buddha rejected this charge.139 It is ironic that within probably a few generations of his death, the Buddha’s followers had drawn exactly the same conclusion, even if they did so with a little more composure and meditative calm.

Abbreviations
All Pāli citations are from Pali Text Society editions.
AN AV BCA BU CSCD Aṅguttara-Nikāya Atharva Veda Bodhicaryāvatāra (see Tripathi) Bṛhadāraṇyaka Upaniṣad (see Olivelle) Chaṭṭha Saṅgāyana: CD-ROM version of the Burmese Tipiṭika, Rangoon 1954. Dhammagiri: Vipassana Research Institute, version 3. Catuṣpariṣatsūtra (see Waldschmidt) Dīgha-Nikāya Majjhima-Nikāya A Sanskrit-English Dictionary, Monier Monier-Williams. Oxford: Clarendon Press (1899). Mahāvastu (see Senart) Oxford English Dictionary, Second Edition, prepared by J. A. Simpson and E. S. C Weiner. Oxford: Clarendon Press (1989). Pali Text Society Ṛg Veda Saṃyutta-Nikāya Saṅghabhedavastu (see Gnoli)

CPS DN MN MMW Mvu OED PTS RV SN SbhV

MN I.137.1ff: tassa evaṃ hoti: ucchijjissāmi nāmassu, vinassissāmi nāma ’ssu, n’ assu nāma bhavissāmī ti. so socati kilamati paridevati urattāḷiṃ kandati sammohaṃ āpajjati. 139 See the section at the beginning of MN I.140, where the Buddha rejects that he is a nihilist (venayika).

138

The ātman and its negation Ud Vin Vism Udāna Vinaya Visuddhimagga (see Warren and Kosambi)

169

References
Bodhi, Bhikkhu. 1978. The All Embracing Net of Views. The Brahmajāla Sutta and its Commentaries. Kandy: Buddhist Publication Society. Bodhi, Bhikkhu. 2000. The Connected Discourses of the Buddha. A New Translation of the Saṃyutta Nikāya (2 Volumes). Oxford: Pali Text Society. Bronkhorst, Johannes. 2007. Greater Magadha. Studies in the Culture of Early India. Leiden, Boston: Brill. Bronkhorst, Johannes. 2009. Buddhist Teaching in India. Boston: Wisdom Publications. Châu, Bhikshu Thích Thiên. 1999. The Literature of the Personalists of Early Buddhism. English translation by Sara Boin-Webb. New Delhi: Motilal Banarsidass. Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism. Cambridge: Cambridge University Press. Dalai Lama. 1994. The Way To Freedom: Core Teachings of Tibetan Buddhism. San Fransisco: Harper Collins. De Jong, J. W. 2000. “The Buddha and His Teachings.” In Wisdom, Compassion, and the Search for Understanding. The Buddhist Studies Legacy of Gadjin M. Nagao, pp. 171–180. Ed. Jonathan A. Silk. Honolulu: University of Hawai’i Press. Evans, Stephen A. 2009. “Epistemology of the Brahmajāla Sutta.” Buddhist Studies Review, Vol. 26(1), pp. 67–84. Fuller, Paul. 2005. The Notion of diṭṭhi in Theravāda Buddhism. The Point of View. London, New York: RoutledgeCurzon. Gethin, Rupert. 1986. “The Five khandhas: Their Treatment in the Nikāyas and Early Abhidhamma.” Journal of Indian Philosophy, Vol. 14(1), pp. 35–53. Gnoli, Raniero. 1978. The Gilgit Manuscript of the Saṅghabhedavastu, Being the 17th and Last Section of the Vinaya of the Mūlasarvāstivādin, Part I. Roma: Istituto Italiano per il Medio ed Estremo Oriente. Gombrich, Richard F. 1992. “Dating the Historical Buddha: A Red Herring Revealed.” In The Dating of the Historical Buddha Part 2, pp. 237–259. Ed. Heinz Bechert. Göttingen: Vandenhoeck & Ruprecht.

170

Alexander Wynne

Gombrich, Richard F. 1996. How Buddhism Began: The Conditioned Genesis of the Early Teachings. London: Athlone Press. Gombrich, Richard. 2009. What the Buddha Thought. London: Equinox. Gómez, Luis O. 1976. “Proto-Mādhyamika in the Pali Canon.” Philosophy East and West, Vol. 26, pp. 137–165. Hamilton, Sue. 1996. Identity and Experience: The Constitution of the Human Being According to Early Buddhism. London: Luzac Oriental. Hamilton, Sue. 2000. Early Buddhism: A New Approach. Richmond, Surrey: Curzon Press. Hayes, Richard. 1988. Dignāga on the Interpretation of Signs. Dordrecht: Publisher. Kornfield, Jack. 1996. Living Dharma. Teachings of Twelve Buddhist Masters. Boston and London: Shambhala. La Vallée Poussin, Louis de. 1917. The Way to Nirvāṇa: Six Lectures on Ancient Buddhism as a Discipline of Salvation. (The Hibbert Lectures, 1916) Cambridge: The University Press. Malalasekera, G. P. 1957. The Buddha and his Teachings. Colombo: Buddhist Council of Ceylon. Ñāṇamoli, Bhikkhu and Bodhi, Bhikkhu. 1995. The Middle Length Discourses of the Buddha. A Translation of the Majjhima Nikāya. Boston: Wisdom Publications. Ñānananda, Bhikkhu. 1971. Concept and Reality in Early Buddhist Thought: An Essay on Papañca and Papañca-Sañña-Sankhā. Kandy: Buddhist Publication Society. Norman, K. R. 1981. “A Note on attā in the Alagaddūpama Sutta.” In Studies in Indian Philosophy (Memorial Volume for Pandit Sukhlalji Sanghvi), pp. 19–29. Eds. Dalsukh Malvania and Nagin J. Shah. Ahmedabad: L. D. Institute of Indology. Oetke, Claus. 1988. Ich und das Ich. Analytische Untersuchungen zur bud­ dhistisch-brahmanischen Ātmankontroverse. Stuttgart: Franz Steiner. Olivelle, Patrick. 1998. The Early Upaniṣads. Annotated Text and Translation. Oxford: Oxford University Press. Pérez-Remón, Joaquin. 1980. Self and Non-Self in Early Buddhism. The Hague: Mouton Publishers. Rahula, Walpola. 1959. What the Buddha Taught. Bedford: Gordon Fraser. Ronkin, Noa. 2005. Early Buddhist Metaphysics. The Making of a Philosophical Tradition. London and New York: RoutledgeCurzon. Rhys Davids, T. W. 1877. Buddhism: Being a Sketch of the Life and Teachings of Gautama, the Buddha. London: Society for Promoting Christian Knowledge. Republished, New Delhi (2000): Asian Educational Services.

The ātman and its negation

171

Rhys Davids, T. W. 1899. Dialogues of the Buddha Part 1. London: Luzac. Rhys Davids, T. W. and Stede, William. 1921–1925. Pali-English Dictionary. London: Pali Text Society. Senart, Émile. 1897. Le Mahāvastu. Texte Sanscrit, Volume III. Paris: Imprimerie Nationale. Siderits, Mark. 2007. Buddhism as Philosophy: An Introduction. UK: Ashgate Publishing Limited. Tripathi, Sridhar. 1988. Bodhicaryāvatāra of Śāntideva with the Commentary Pañjikā of Prajñākaramati. (Buddhist Sanskrit Texts No. 12). Darbhanga: The Mithila Institute. Vetter, Tilmann. 1988. The Ideas and Meditative Practices of Early Buddhism. Leiden: E. J. Brill. Waldschmidt, Ernst. 1952. Das Catuṣpariṣatsūtra. Teil I: Der Sanskrit Text im Handschriftlichen Befund. Berlin: Akademie-Verlag. Warder, A. K. 1991. Introduction to Pali (Third Edition). Oxford: Pali Text Society. Warren, Henry Clarke and Kosambi, Dharmananda. 1989. Visuddhimagga of Buddhaghosācariya. (Harvard Oriental Series Vol. 41). New Delhi: Motilal Banarsidass. Williams, Paul and Tribe, Anthony. 2000. Buddhist Thought: A Complete Introduction to the Indian Tradition. London and New York: Routledge. Wynne, Alexander. 2005. “The Historical Authenticity of Early Buddhist Literature: A Critical Evaluation.” Wiener Zeitschrift für die Kunde Südasiens, Band XLIX, pp. 35–70. Wynne, Alexander. 2007. The Origin of Buddhist Meditation. London, New York: Routledge. Wynne, Alexander. 2009a. “Early Evidence for the ‘no self’ Doctrine? A Note on the Second anātman Teaching of the Second Sermon.” Thai International Journal for Buddhist Studies, Vol. 1, pp. 64–84. Wynne, Alexander. 2009b. “Miraculous Transformation and Personal Identity: A Note on the First anātman Teaching of the Second Sermon.” Thai International Journal of Buddhist Studies, Vol. 1, pp. 85–113. Wynne, Alexander. 2010. “The Buddha’s Skill in Means and the Genesis of the Five Aggregate Teaching.” Journal of the Royal Asiatic Soicety, Series 3, Vol. 20(2), pp. 191–216

Indian Buddhist metaethics
Contributions to a panel at the XVth Congress of the International Association of Buddhist Studies Atlanta, 23–28 June 2008

Guest editor

Martin T. Adam

Dedicated, with deepest gratitude, to the late Dr. Leslie Kawamura, University of Calgary: Profound teacher, trusted guide, friend. Maṅgalam!

An analysis of factors related to the kusala/ akusala quality of actions in the Pāli tradition
Peter Harvey

Scattered through the texts of the Pāli Canon, and supplemented by some commentarial passages, are a series of discussions of such matters as the nature of action (Pāli kamma, Skt. karman), praiseworthy and reprehensible actions, criteria for determining an action as praiseworthy or reprehensible, and reflections on how the intention and knowledge of an agent should affect the moral assessment of an action. In these one can see a set of complementary principles of ethics. These passages are helpful in ascertaining the closest Western analogue to the ethics of early and Theravāda Buddhism. Is it Utilitarianism, where actions are judged only by the amount of pain or happiness that they cause people, or sometimes also animals? Is it Kantian ethics, where the good will is central? Is it virtue ethics, as argued by Keown (1992), where actions are judged by the virtues or vices that they express, and their contribution to character traits that produce a certain kind of person? Is it simply a form of moral realism, in which ethical statements are seen to report certain moral facts about the world, particularly that certain specified acts are wrong? Or is it that these theories each correspond to only certain aspects of Buddhist ethics? As we will see, the last of these turns out to be the case.

Kusala and akusala actions and their roots
Kusala and akusala are the terms perhaps most commonly used for praiseworthy and reprehensible actions or states of mind in early

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 175–209

176

Peter Harvey

Buddhist texts. Kusala can be translated as “wholesome”1 or “skilful,” and akusala means the opposite of this. For the time being, the translation “wholesome” will be used. In the Sammādiṭṭhi Sutta is a passage (MN I.46–47), which identifies typical akusala and kusala actions and makes clear what lies at the root of such actions:
When, friends a noble-disciple2 understands the unwholesome (akusalañ) and the root of the unwholesome (akusalamūla-), the wholesome (kusalañ) and the root of the wholesome (kusalamūla-), in that way he is one of right view (sammādiṭṭhi), whose view is straight, who has perfect confidence in the Dhamma and has arrived at this true Dhamma. And what is the unwholesome (akusalaṃ)? … Killing living beings is the unwholesome; taking what is not given is the unwholesome; misconduct in sensual pleasures is the unwholesome; false speech is the unwholesome; divisive speech is the unwholesome; harsh speech is the unwholesome; frivolous chatter is the unwholesome; covetousness is the unwholesome; ill-will is the unwholesome; wrong view is the unwholesome. This is called the unwholesome. And what is the root of the unwholesome? Greed (lobho) is a root of the unwholesome; hatred (doso) is a root of the unwholesome; delusion (moho) is a root of the unwholesome. This is called the root of the unwholesome. And what is the wholesome? Abstention from killing living beings is the wholesome; … abstention from frivolous chatter is the wholesome; uncovetousness is the wholesome; non-ill-will is the wholesome; right view is the wholesome. This is called the wholesome. And what is the root of the wholesome? Non-greed (alobho) is a root of the wholesome; non-hate (adoso) is a root of the wholesome; nondelusion (amoho) is a root of the wholesome. This is called the root of the wholesome.

Here, the roots of wholesome action, literally non-greed, non-hatred and non-delusion, are not just the absence of greed, hatred and
1 The sixth edition of the Concise Oxford Dictionary (1976) defines “wholesome” as “promoting physical or moral health.” The idea of kusala actions as nourishing “moral health” seems appropriate. 2 Ariya-sāvako: noble disciple or disciple of the noble ones.

Factors related to the kusala/akusala quality of actions

177

delusion, but states which oppose them: anti-greed (generosity and renunciation), anti-hatred (loving-kindness and compassion) and anti-delusion (wisdom).3 In this passage, “the unwholesome” is explained as consisting of actions of body and speech that are widely recognized as harmful to others, plus mental states that are likely to later lead to such harmful actions. The neuter term akusalaṃ is used with all these actions, showing it is a noun, not an adjective agreeing with the gender of the respective actions. This seems to indicate that these actions are key ingredients of “the unwholesome,” rather than things which happen to have the quality of being unwholesome. In this passage, greed, hatred and delusion, as the “roots” of harmful actions, are not presented as the criteria for labeling something as “the unwholesome,” but as the inner causes of “the unwholesome.” However, at AN I.197, greed, hatred and delusion are not just causes of unwholesome action, but their very existence is an unwholesome situation: an arahant understands: “Formerly there was greed; that was an unwholesome thing (akusalaṃ). Now it does not exist; this is a wholesome thing (kusalaṃ)” (and the same for hatred and delusion). This is also seen at AN I.201–205 which, having identified greed hatred and delusion as each a “root of the unwholesome (akusalamūlaṃ),” says:
Whatever, monks, is greed, that is also is an unwholesome thing (Yad api bhikkhave lobho tad api akusalaṃ). Whatever the greedy one performs by body, speech or mind, overpowered by greed, his thoughts controlled by it (pariyādinnacitto), doing to another by falsely4 causing him pain (asatā dukkhaṃ upadahati) through punishment, imprisonment, loss of wealth, abuse, banishment, (thinking) ‘I am strong, (I have the) advantage of strength,’ that is also is an unwholesome thing. Thus these various evil (pāpakā), unwholesome states born of greed, with greed as cause, greed as origin, greed as condition are assembled together in him.

3 As indicated by the fact that Dhs 32–34 treats them as positive presences, with 34 explaining amoha as paññā, wisdom. 4 The context indicates that the situation concerns lying about someone, so that they are then subject to punishment.

178

Peter Harvey

The same is then said of behaviour based upon hatred and delusion. The passage continues:
Such a person, overpowered by evil, unwholesome states born of greed, being of uncontrolled mind, in this very life lives in pain, with distress, unrest and fever (dukkhaṃ savighātaṃ saupāyāsaṃ sapariḷāhaṃ), and when the body breaks up after death, a bad destiny is to be expected.

The same is then said of someone overwhelmed by states born of hatred or delusion, all such cases being like the case of a tree destroyed by parasitic creepers. On the other hand, non-greed, nonhatred and non-delusion are each a root of the wholesome, and one who abandons states born of greed, hatred and delusion is released in this very life, like a tree that a person completely frees of parasitic creepers. This holds not only that greed etc. are causes for unwholesome action, but are themselves included in what is unwholesome. They can take over the mind so as to lead a person to act in a way that brings suffering on others. This, though, leads to pain and distress to the agent in this life, and a bad rebirth. Showing this as a causal sequence:
unwholesome roots → unwholesome actions that bring pain to others → harm and pain in this and later lives to the agent.

AN I.263 sees both greed, hatred and delusion and non-greed, non-hatred and non-delusion as “causes for the origin of actions (nidānāni kammānaṃ samudayāya).” Furthermore:
… whatever action is of the nature (pakataṃ) of greed (or hatred, or delusion), born of greed, caused by it, that action is unwholesome, it is with fault/blameable (sāvajjaṃ), it ripens in pain (dukkhavipākaṃ).

On the other hand, an action of the nature of non-greed etc., “is wholesome, it is faultless/blameless, it ripens in happiness.” However, while the first kind of action, that born of greed etc., “conduces to the origin of actions (kammasamudayāya), not to the cessation of actions,” the second, born of non-greed etc., does the opposite. Wholesome actions, then, are actions which bring hap-

Factors related to the kusala/akusala quality of actions

179

piness in saṃsāra but also go beyond entrapment in action-fuelled saṃsāra.

The nature and relative effect of attachment, hatred and delusion, and the role of attention
What is the nature of the three “roots” of the unwholesome, and what feeds them? In some passages, lobha, greed, is replaced by rāga, attachment or lust. This covers sexual lust, but also lusting after anything, including subtle meditative states. It encompasses the same ground as “greed” (lobha). At AN I.194, the Buddha calls “greed” “covetousness” (abhijjhā), calls “hatred” “ill-will” (vyāpādo), and calls “delusion” “ignorance” (avijjā). This sees the first two of these as the same as the first two of the three unwholesome mental actions – though a consciously developed line of thought is more of an action than the initial root that it expands from –, but the third as ignorance rather than the mental action of wrong view. Greed and hatred are affective, non-cognitive states and, as directed to certain goals, can be termed “motives.” While the confusion aspect of delusion can have an affective component, delusion is mainly a cognitive state, not itself a motive, in the sense of an impulse directed at a certain kind of goal, now or in the future. In the Abhidhamma, it is equated with spiritual ignorance (avijjā), and both are explained as:
… unknowing (aññāṇaṃ) as regards dukkha (the painfulness of life), … the origin of dukkha, … the cessation of dukkha, … the way leading to the cessation of dukkha;5 unknowing as regards the past … the future … the past and future;6 unknowing as regards dependently arisen states from specific conditionality (idapaccayatā paṭiccasamupannesu dhammesu);7 even all that kind of unknowing which is unseeing, non-achievement, non-wakening, non-awakening, non-penetration, non-apprehension, non-comprehension, non-conThe “unknowing” is not lack of information about these, but lack of direct insight into them. 6 That is, relating to past lives, and to future lives yet to be produced by present and past karma. 7 That is, dependent origination.
5

180

Peter Harvey

sideration, non-reflection, non-clarification – stupidity, folly, lack of clear comprehension,8 delusion, bewilderment, confusion, ignorance, the flood of ignorance, the yoke of ignorance, the latent tendency of ignorance, the besetting of ignorance, the barrier of ignorance, delusion, the root of the unwholesome (Dhs 1061, 1162; cf. Vibh 362).

Delusion, then, is a kind of ingrained misperception, a distorted perspective, even stupidity, a kind of general mental misorientation that gives a view of things – blind to dukkha etc., and conditionality – in which certain motives, such as greed directed at gaining or holding on to something, or hatred, directed at bringing harm to someone, are seen to make sense. Non-delusion is the clarity and wisdom that sustains wholesome motives. AN I.199–201 is another passage that probes the nature of attachment, hatred and delusion, and discusses what causes them and their opposites.
If wanderers of other sects should ask you about the distinction, disparity and difference among these three qualities (dhammā) – attachment (rāgo), hatred (doso) and delusion (moho) – you should answer them thus: ‘Attachment is a small fault (appasāvajjo) but its removal is slow (dandhavirāgī); hatred is a great fault (mahāsāvajjo) but its removal is quicker (khippavirāgī); delusion is a great fault and its removal is slow.’9 If they ask, ‘Now friends, what is the cause and reason for the arising of unarisen attachment (or hatred or delusion), and for the increase and strengthening of arisen attachment (or hatred or delusion)?,’ you should reply: ‘… For one attending unwisely (ayoniso manasi­ karoto) to (an object’s) attractive aspect (subhanimittaṃ), unarisen attachment will arise and arisen attachment will increase and become strong. … For one attending unwisely to (an object’s) irritating aspect (paṭighanimittaṃ) unarisen hatred will arise and arisen hatred will

Clear comprehension is a close complement to mindfulness. While this might refer to the short-term subsidence of these three states, it more likely applies to their long-term eradication. In terms of the ten fetters destroyed by noble persons, while a non-returner destroys ill-will (hatred) and desire for sense-pleasures, other forms of rāga still remain for him: for the subtle elemental form and formless levels. These, along with ignorance (delusion) are only destroyed at arahantship.
9

8

Factors related to the kusala/akusala quality of actions

181

increase and become strong. … For one attending unwisely,10 unarisen delusion will arise and arisen delusion will increase and become strong.’ If they ask, ‘Now friends, what is the cause and reason for the nonarising of unarisen attachment (or hatred or delusion), and for abandoning of arisen attachment (or hatred or delusion)?,’ you should reply: ‘… For one attending wisely (yoniso manasikaroto) to (an object’s) unattractive aspect (asubhanimittaṃ), unarisen attachment will not arise and arisen attachment will be abandoned. … For one attending wisely to the liberation of mind by loving-kindness, unarisen hatred will not arise and arisen hatred will be abandoned. … For one attending wisely, unarisen delusion will not arise and arisen delusion will be abandoned.’

Here we see that the crucial sustainer of attachment, hatred and delusion is attention that is ayoniso: unwise, unsystematic, inappropriate, not focussing on the fundamental nature of its object. Elsewhere, it is said that ayoniso attention supports lack of both mindfulness (sati) and clear comprehension (sampajañña), which then supports non-guarding of the sense-faculties, and this then supports misconduct of body, speech and mind. Wise attention has the opposite effect (AN V.113–116). Wise attention is something that is used in and strengthened by the various forms of Buddhist mind-training or “meditation.” That wise attention feeds the roots of kusala action is reflected in the fact that “kusala,” when applied to a person, can mean “skilled” or “expert.” This is seen at AN I.112, where the Buddha says that in a past life he had been a wheelwright. A wheel he took nearly six months to make was without faults, and when set rolling on its own, did not fall over when the impetus ran out. One he had only six days to make was crooked and with faults and flaws (savaṅkā sadosā sakasāvā), so it did fall over after the impetus ran out. Thus he says,
Then, monks, I was one skilled (kusalo) in [understanding] wood that was crooked, the faults and flaws of wood. Now, monks, I … am one

Note that no specific object of unwise attention is specified here as respecific specified gards delusion (or, below, wise attention as regards non-delusion).

10

182

Peter Harvey

skilled in the crooked ways, the faults and flaws of body… of speech and mind.11

We can thus extend the causal sequence given above, to:
unwise attention → unwholesome roots → unwholesome actions that bring pain to others → harm and pain in this and later lives to the agent.

Why is the moral tone of an action seen to cause certain karmic results? It is said that wrong view leads on to wrong resolve (saṅkappa), and this to wrong speech and thus wrong action, while right view has the opposite effect (AN V.211–212). As wrong actions thus come from the misperception of reality, they can be seen to be “out of tune” with the real nature of things. As they thus “go against the grain” of reality – though they may be the sort of thing that many people do –, they naturally lead to unpleasant results. Thus it is said to be impossible that wrong conduct of body, speech or mind could produce a “ripening that was agreeable, pleasant, liked,” or for right conduct to produce a “ripening that was disagreeable, unpleasant, not liked” (MN III.66).

Corrupt actions of unwholesome volition with painful ripening
As is well known, at AN III.41512 it is said, “It is volition (cetanā), monks, that I call karma. Having willed (cetayitvā), one performs an action (kammaṃ karoti) by body, by speech, by mind.” Moreover: “the ripening (vipāka) of action is in this life, in the next life, or subsequently.” Here, the identifying feature of human “action” is specified as the will or volition (cetanā) that is expressed in a performed action of body, speech or mind (an active line of thought). This seems to be a recognition of the fact that an unwilled movement of limbs is
11 Cousins (1996: 143–144) discusses this and similar passages, linking such passages to the skill required in meditation. 12 Cf. Asl 88.

Factors related to the kusala/akusala quality of actions

183

not, as such, a human “action,” and how an action is morally characterised typically depends on the volition that it expresses. Cetanā encompasses the general motive for which an action is done, its immediate intention (directed at a specific objective, as part of fulfilling a motive), and the immediate mental impulse which sets it going and sustains it (Keown 1992: 213–218). Buddhaghosa explains it thus (Vism 463):
‘It wills (cetayati),’ thus it is volition (cetanā); ‘it collects (abhisanda­ hati)’ is the meaning. Its characteristic is the state of volition (cetanā). Its function is to strive (āyūhana-). It is manifested as co-ordination. It accomplishes its own and others’ functions, as a senior pupil, a head carpenter etc. do. But it is evident when it occurs in marshalling of associated states in connection with urgent work, remembering and so on.

At AN V.292–297, what are elsewhere described as unwholesome actions are seen to express an unwholesome volition:
There are, monks, corrupt and harmful actions (kammantasandosavyāpatti) of unwholesome volition (akusalasañcetanikā) with painful consequences (dukkhudrayā), ripening in pain (dukkhavipākā): three of body, four of speech, and three of mind. …

These actions are the above three forms of wrong action, four forms of wrong speech, and three forms of wrong thought. In the following delineation of these, both the overt action and inner mind state are described:
There is a person who kills living beings; he is cruel and his hands are blood-stained; he is bent on slaying and murdering, having no compassion for any living being. He takes what is not given to him, appropriates with thievish intent … There is one who is a liar … he utters deliberate lies … He utters divisive speech … He is fond of dissension … He is given to frivolous chatter … There is a person who is covetous; he covets the wealth and property of others, thinking: ‘Oh, that what he owns might belong to me!’ There is also one who has ill will in his heart. He has corrupt mental resolve (manasaṅkappo): ‘Let these beings be slain! Let them be killed and destroyed! May they perish and cease to exist!’

184

Peter Harvey

He has wrong views (micchādiṭṭhiko) and a perverted way of seeing (viparītadassano): ‘There is no gift, there is no offering, there is no (self-)sacrifice (that is, these have no worth); there is no fruit (phalaṃ) or ripening (vipāko) of actions well done or ill done (that is, how one behaves does not matter, it has no effect one one’s future); there is no this world, no other world (that is, this world is unreal, and one does not go to another world after death); there is no mother or father (that is, there is no point in respecting the people who established one in this world); there are no spontaneously arising beings (that is, beings without parents in some rebirth worlds); in this world, there are no renunciants and brahmins who are faring rightly, practising rightly, who proclaim this world and the world beyond, having realized them by their own super-knowledge (that is, spiritual development is not possible; people cannot come to have direct meditative knowledge of rebirth into a variety of kinds of world).’

Such actions lead to rebirth in hell. On the other hand there are “beneficial (sampatti) actions of wholesome volition, with happy consequences, ripening in happiness: three of body, four of speech, and three of mind.” These are to abstain from the above ten kinds of action and where appropriate do the opposite kind. For example, “there is a person who abstains from the destruction of life; with rod and weapon laid aside, he is conscientious and kindly and dwells compassionate towards all living beings.” This then leads to a heavenly rebirth. Here the volitions involved in unwholesome actions are themselves unwholesome, and the actions “corrupt and harmful,” They have “painful consequences (dukkhudrayā), ripening in pain (dukkhavipākā).” The term for “ripening,” vipāka, is usually used for future karmic results for the agent of an action,13 but the term for “consequence,” udraya, may refer to the more immediate effects of an action on anyone. The causal sequence now becomes:

An action is often likened to a “seed” and its karmic result to the “fruit” (phala) or “ripening” (vipāka) of this.

13

Factors related to the kusala/akusala quality of actions

185

unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are cor­ rupt and bring pain to others → harm and pain in this and later lives to the agent.

How intention and perception of relevant facts affects the moral assessment of an action
In certain Vinaya passages, it becomes clear that how a monk perceives the facts of a situation affects whether or not he breaks a relevant monastic rule fully, partially, or not at all and, by implication, whether and by how much the act is unwholesome. Most monastic rules can also only be broken when the act is intentionally done (Harvey 1999: 273–280). Vin IV.124–125 concerns the rule against intentionally killing a non-human living being, and comes after a story of a monk shooting crows with arrows:
Whatever monk should intentionally (sañcicca) deprive a living being of life, there is an offence of expiation. … ‘Intentionally’ means a transgression committed knowingly (jānanato), consciously (sañjānanto), deliberately (cecca). …

It is then said that there is an offence of expiation when:
a) he thinks that it is a living being when it is a living thing, and deprives it of life.

There is a lesser offence of wrongdoing when:
b) he is in doubt as to whether it is a living being, and deprives it of life. c) he thinks that it is a living being when it is not a living being, and acts to ‘kill’ it. d) he is in doubt as to whether it is not a living being, and deprives it of life.

There is no offence when:
e) he thinks that it is not a living being when it is a living being,14
14

As when one misperceives a crow as a patch of shade; not as when one,

186

Peter Harvey

f) he thinks that it is not a living being when it is not a living being. g) ‘if it is unintentional; if there is no mindfulness (asatiyā) (of what is being done); if he does not know;15 if he is not meaning death; if he is mad, if he is the first wrong-doer (i.e. a monastic rule cannot be broken before there is such a rule).’

The causal sequence now becomes:
unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are intentional, corrupt and bring pain to others, in a way that is antici­ pated by correct perception of the immediate facts of the situation → harm and pain in this and later lives to the agent.

Unwholesome acts as with fault/blameable, criticised by the wise, not to be done
In the Lakkhaṇa Sutta, it is said that the past karmic cause of the Bodhisatta having smooth skin, to which dust does not stick – this being an indication of his great wisdom –, was that he used to enquire of renunciants and brahmins:
What is it that is wholesome (kusalaṃ), what is it that is unwholesome? What is with fault (sāvajjaṃ), what is not? What course is to be followed (sevitabbaṃ), what is not? What, if I do it, will be to my lasting harm (ahitāya) and pain (dukkhāya), … what to my lasting benefit and happiness? (DN III.157).

This sees understanding the distinction between what is wholesome and unwholesome as linked to wisdom, just as the above Sammādiṭṭhi Sutta associates it with right view. It also sees being “with fault” as closely linked to unwholesomeness, and clearly sees both as not to be pursued. Cousins (1996: 148) comments that:

for example, recognises a worm as a worm, but does not count it as a “living being.” 15 That is, does not know that his action might kill a living being.

Factors related to the kusala/akusala quality of actions

187

In the great majority of cases … whatever other terms are associated with kusala, the term which is always present, usually immediately next to kusala, is blameless (anavajja).

There is a small uncertainty in how best to translate sāvajja and anavajja. Cousins (1996: 139) sees anavajja, as “(originally) not reprehensible, blameless; (later) faultless.” Margaret Cone (2001: 245) says that avajja means, as an adjective, “blameable; low, inferior,” as a noun, “what is blameable; imperfection, fault,” and sāvajja as an adjective meaning “blameable; faulty.” Perhaps a word which captures both aspects of the adjectival meaning is “wrong,” with “right” for anavajja. While micchā and sammā are also translated this way, at AN V.242, the micchā and sammā path factors are respectively called, among other things, sāvajja and anavajja. It is not surprising, then, that an unwholesome action is also criticised by the discerning or wise (viññugarahitā).16 Of course blame from those who are not wise may not be appropriate: the eight worldly phenomena that one needs equanimity in the face of are gain and loss, fame and shame, blame and praise (nidā, pasaṃsā), and pleasure and pain (e.g. DN III.260). As king Pasenadi says in the Bāhitika Sutta (MN II.114):
… we do not recognise anything of value in the praise and blame (vaṇṇaṃ vā avaṇṇaṃ) of others spoken by foolish and ignorant people who speak without having investigated and evaluated; but we recognise as valuable the praise and blame of others spoken by wise, intelligent and sagacious (paṇḍitā vyattā medhgāvino) persons who speak after having investigated and evaluated.

The relevance of criticism from the wise is seen in the Kālāma Sutta, the popular name for the Kesaputta Sutta (AN I.188–193), which sees greed, hatred and delusion as the states identified:
… when you know for yourselves, ‘these states (dhammā) are unwholesome (akusalā) and with fault (sāvajjā), they are criticised by the wise (viññugarahitā); these states, when undertaken and practised, conduce to harm (ahitāya) and pain (dukkhāya)’ …

There is also reference to “virtues dear to the noble ones” (ariyakantehi sīlehi) (SN V.343).

16

188

Peter Harvey

They should, therefore, be abandoned (AN I.189). This is because a person:
… overpowered by greed (hatred or delusion), his thoughts controlled by it, will destroy life, take what is not given, engage in sexual misconduct, and tell lies; he will also prompt others to do likewise.

Hence they “conduce to his harm and pain for a long time.” States of non-greed, non-hatred and non-delusion have the opposite qualities and effects. Indeed, one “devoid of covetousness, devoid of illwill, unconfused,17 clearly comprehending, ever mindful, dwells pervading … the entire world with a mind imbued with lovingkindness,” i.e. he can develop in meditation and be calm and confident. The causal sequence thus becomes:
unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are intentional, corrupt, with fault, and bring pain to others, in a way that is anticipated by correct perception of the immediate facts of the situation, and which are criticized by the wise, not to be done → harm and pain in this and later lives to the agent.

Unwholesome actions as also bringing injury to others, future karmic pain, affliction to self and others, and a bad effect on one’s character
A sutta which gives a series of near equivalents of the term akusala, and then for kusala is the Bāhitika Sutta (MN II.114–115).18 Here Ānanda explains, in response to questioning by king Pasenadi that:
1. An action of body (or speech or mind) which is ‘censured (opārambho) by wise (viññūhi) renunciants and brahmins’ is:

These suggest the overcoming, or at least suspension, of greed, hatred and delusion. Note, though, that such a disciple is only said to be free of enmity and ill will (perhaps temporarily), not delusion. 18 Anālayo (2007) compares this to a Madhyamāgama parallel sutta which he also translates from the Chinese.

17

Factors related to the kusala/akusala quality of actions

189

2. ‘Any bodily behaviour (kāyasamācāro) that is unwholesome (akusalo),’ which in turn is: 3. ‘Any bodily behaviour that is with fault (sāvajjo),’ which in turn is: 4. ‘Any bodily behaviour that is injurious (savyāpajjho),’ which in turn is: 5. ‘Any bodily behaviour ripening in pain (dukkhavipāko),’ which in turn is: 6. ‘(a) Any bodily behaviour, great king, that leads to one’s own affliction (attabyābādhāya), or to the affliction of others, or to the affliction of both, and (b) on account of which unwholesome states increase and wholesome states diminish. – Such bodily behaviour is censured by wise renunciants and brahmins.’19

The passage goes on to say that behaviour uncensured by wise renunciants and brahmins is behaviour that is wholesome. And then, in response to the king’s questions, he explains that wholesome actions are faultless ones; these are ones that do not bring harm; these are ones that ripen in pleasure. Furthermore, these are any that:
… does not lead to one’s own affliction, or to the affliction of others, or to the affliction of both, and on account of which unwholesome states diminish and wholesome states increase. – Such bodily behaviour is uncensured by wise renunciants and brahmins, great king.

Here the faultiness (3) of an action is explained in terms of item 4: it being associated with vyāpajjha: injury, harm, trouble. Presumably this refers to the injury caused to others by killing, stealing, etc.
19 The Chinese parallel refers in turn to conduct that is: i. (= 1 above) “detested by recluses and brahmins,” ii. (= 2) “unwholesome,” iii. (= 3) “constitutes an offence,” iv. (cf. 1) “is detested by the wise,” v. (~= 6) “harms oneself, harms others, harms both; that destroys wisdom and fosters evil; that does not [lead to] attaining Nibbāna, does not lead to knowledge, does not lead to awakening, and does not lead to Nibbāna” (Analāyo 2007: 158–159). Here, item i and iv are close in meaning and v specifically refers to hindering the attainment of nibbāna, while 4 and 5 of the above are not listed. The passage adds that those who carry out such conduct do not know, according to reality, what should or should not be undertaken, accepted, eliminated or accomplished (Analāyo 2007: 159–160). Analāyo sees v/6. as the “basic definition” (2007: 168).

190

Peter Harvey

This is implied by the Kālama Sutta and by the fact that avyāpajjha means to be peaceful and friendly. An injurious action is then explained as item 5: an action “ripening in pain (dukkhavipāko).” The term vipāka usually refers to future karmic results, so here 5 gives not so much the meaning of 4 as refers to its inevitable consequences for the agent. Item 6a then links 5 back to affliction (byābādha): to self, others, or both. “Affliction” to others cannot be referring to future karmic results, as the affliction that one person’s actions may bring to other people cannot mean karmic results that come to them: such results come to the agent of actions, not those they act on. This also implies that the “affliction” to self refers to immediate harm to oneself. Indeed at MN I.342–349, a person who is a “self-tormentor (attantapo),” is one who inflicts harsh asceticism on himself,20 an “other-tormentor” is a butcher, hunter or executioner, while one who torments both is an ascetic king who has large animal sacrifices done by enforced and suffering workers. Nevertheless, 6b brings another reference to the future, not in terms of karmically caused pain for the agent of present actions, but in terms of the affect on their character or virtue: an increased disposition to further unwholesome actions and decrease in wholesome traits. This can also be seen at DN II.279–280, which says that any bodily or verbal behaviour, or the pursuit of goals (pariyesana) should be avoided if it leads to the increase of unwholesome dhammas and decrease of wholesome ones, but followed (sevitabba) if it leads to the decrease of unwholesome dhammas and the increase of wholesome ones.21 Here, dhammas most likely refer to mental states. The causal sequence thus becomes:
unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are intentional, corrupt, with fault, bring pain and injury to oneself or

Nevertheless, SN IV.338–339 says that a person who (generally) “afflicts and torments (ātāpeti paritāpeti) himself” may still develop a wholesome state. 21 My thanks to the reviewer of this paper for drawing my attention to this passage.

20

Factors related to the kusala/akusala quality of actions

191

to others in a way that is anticipated by correct perception of the immediate facts of the situation, criticized by the wise, not to be done → harm and pain in this and later lives to the agent, and un­ wholesome character tendencies.

Afflictive, dark action with afflictive, dark ripening
Item 6a above is also a feature of other passages. In the Amba­ laṭṭhikārāhulovāda Sutta (MN I.415–419), the Buddha advises his son Rāhula, now a novice, to reflect before an action of body, speech or mind and ask himself:
Would this action that I wish to do with the body lead to my own affliction (-attabyābādhāya), or to the affliction of others, or to the affliction of both? Is it an unwholesome bodily action with painful consequences (dukkhudraya-), with painful ripening (dukkhavipāka-)? (MN I.415).

If the answer is “yes,” he should not do the action, if “no,” he can do it. He should likewise break off an action if he realises during its performance that it is leading to these effects and, if after an action he reflects that it has led to such affliction and has painful consequences and ripening, he should confess it and undertake not to repeat it. Along with 6a, factor 5 is mentioned, and “painful consequences” is probably equivalent to 4. The concern not to bring affliction to self or other is also seen in the Veḷudvāreyyā Sutta (SN V.353–356), which urges a nobledisciple to reflect using a negative version of the golden rule:
I am one who wishes to live, who does not wish to die; I desire happiness and am averse to suffering. Since this is so, if someone were to take my life, that would not be pleasing and agreeable to me. Now if I were to take the life of another – of one who wishes to live, who does not wish to die, who desires happiness and is averse to suffering – that would not be pleasing and agreeable to the other either. What is displeasing and disagreeable to me is displeasing and disagreeable to the other too. How can I inflict upon another what is displeasing and disagreeable to me?

192

Peter Harvey

On the basis of this, it is said that a person abstains from the three forms of wrong action and four forms of wrong speech, as well as exhorting others to do so, and praising such abstinence. The context of the passage is that the Buddha is asked by some brahmins for a teaching that will help them live happy lives of many children, beautiful possessions and good rebirths. That is, just as one does not like to be afflicted oneself, do not afflict others – and this abstinence will bring one future good fortune. Indeed, whether he is presently in pain or happy, the wise person does not do any of the three forms of wrong action, four forms of wrong speech or three forms of wrong mental action, as these will ripen in pain (dukkhavipāka) in the future; rather, he or she does actions that ripen in happiness (MN I.313–315). In the Kukkuravatika Sutta, the focus is on not afflicting oneself, as here the Buddha addresses an ascetic who models his behaviour on that of a dog (i.e. he afflicts himself), and explains that this conduct will lead to rebirth as an animal or in hell. He then explains that there are four kinds of action (MN I.389–391), the first three of which are:
• ‘action that is dark22 and with a dark ripening’ (kammaṃ kaṇhaṃ kaṇhavipākaṃ): one ‘generates an activity’ (saṅkhāraṃ abhisaṅkharoti) of body speech or mind that is ‘afflictive’ (sabyābajjahaṃ) and hence is reborn in an ‘afflictive’ world such as hell, with afflictive contacts and afflictive, painful feelings. AN II.234 explains dark actions as breaking the five precepts. • ‘action that is bright (sukkaṃ) with a bright ripening:’ generating unafflictive activities, so as to be reborn in an unafflictive world, such as that of the Subhakiṇṇā heaven (of the elemental form level). AN

22 Cf. DN I.163 refers to “those things that are unwholesome, and reckoned as such, with fault (sāvajjā), to be refrained from (asevitabbā), unbefitting noble ones (nālamariyā), dark (kiṇhā, vl. kaṇhā) …,” with the opposite kinds of things as e.g. “bright” (sukkā). At SN V.104, the nutriment for the arising and development of the discrimination of dhammas factor of awakening is giving wise attention to: “dhammas which are wholesome and unwholesome, with fault and faultless, inferior and superior (hīnapaiṇītā), dark and bright with their counterparts (kaṇhasukkasappaṭibhāgā).” These four pairs also occur at SN I.129 and AN III.165; cf. AN IV.363.

Factors related to the kusala/akusala quality of actions

193

II.234 explains bright actions as keeping the five precepts, and AN II.235 explains them as the ten right actions ending in right view. • ‘action that is dark-and-bright with dark-and-bright ripening:’ generating a mixture of afflictive and non-afflictive activity, so as to be reborn in a world that is both afflictive and non-afflictive, such as the human world.

This passage makes clear that while the nature of an action’s karmic ripening corresponds to the nature of the action, it does not determine its nature. For example a “dark” action is not to be counted as dark because it has a dark karmic result; rather, it has a dark result because it is itself dark: causing affliction here and now.23 Its having dark karmic results is a sign of its dark, afflictive, unwholesome nature, but not the criterion for its being unwholesome in the first place. Nevertheless, the passages we are examining in this article hardly neglect to mention karmic results, in the form of pleasant or unpleasant experiences, – perhaps as additional motivating or demotivating factors for the actions –, and of course do concern themselves with the immediate (non-karmic) effects of actions on both others and oneself. The causal sequence thus becomes:
unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are intentional, dark, corrupt, with fault, bring pain and injury to oneself or to others in a way that is anticipated by correct perception of the immediate facts of the situation, criticized by the wise, not to be done → dark harm and pain in this and later lives to the agent, and unwholesome character tendencies.

The fourth kind of action at MN I.389–391 is:

23 Adam (2005: 68) sees the “affliction” aspect which makes an action “dark” as its “obscuring the mind” and the “dark” (or “bright”) aspect as concerning both “the moral quality and the epistemic character of the action itself” (2005: 69). An epistemic aspect is possible here, but I do not see any direct evidence for it.

194

Peter Harvey

• ‘action that is neither dark nor bright with neither-dark-nor-bright ripening, action that leads to the destruction of action:’ the ‘volition for abandoning (pahānāya yā cetanā)’ the above three kinds of action.

AN II.236 explains this kind of action as right view through to right concentration, and Asl 89 explains it as both the seven factors of awakening (bojjhaṅgas) and the Noble Eightfold Path. AN III.384–385 sees nibbāna as neither dark nor bright, in the sense of neither a bad nor good rebirth. As we have seen above, AN I.263 sees actions born of non-greed, non-hatred and nondelusion – i.e. wholesome actions, as leading to the cessation of (saṃsāra-fuelling) actions. Adam (2005: 72) tentatively identifies the person who does the fourth kind of action as the noble persons who have not yet attained arahantship.24 While the first of these he mentions is the stream-enterer, we should also probably include the “person practising for the fruit that is stream-entry” (sotāpattiphalasacchikiriyāya paṭipanno), who is close to becoming a stream-enterer.25 Wholesome action is in one aspect “bright” action that also leads to “bright” future results in saṃsāra, but at the level of “path” actions it enables a person to go beyond even the “bright” aspects of saṃsāra.

However, at AN V.244–245: right view through to right liberation is the sukkamagga in contrast to wrong view through to wrong liberation as the kaṇha-magga. This implies either that “bright” may sometimes exclude and sometimes include the path, or that only the path’s lokuttara form is beyond both “dark” and “bright.” On this fourth kind of action, see Harvey (2000: 43–46). 25 AN IV.372 sees them as no longer an ordinary person (puthujjana), SN V.23 sees one who is “rightly practising” (sammāpaṭipanno) as having the eight factors of the magga, and SN V.25 sees the eight factors of the magga as what being a renunciant (sāmañña), a samaṇa, entails, whose fruits are the stream-entry-fruit up to the arahantship-fruit. As the magga leads to not only the three higher fruits, but the first one, stream-entry, then those engaged in the magga include the first kind of paṭipanna person. Such a person is endowed with its eight factors to some degree (Harvey 2011).

24

Factors related to the kusala/akusala quality of actions

195

Afflictive action that obstructs wisdom and leads away from nibbāna
The Dvedhāvitakka Sutta (MN I.115–116) echoes the Bāhitika Sutta’s item 6a, but also expands on 6b. Here the Buddha explains that before his awakening, while still a Bodhisatta, he had divided his thoughts (vitakkas) into two groups: those of sensual desire, illwill and cruelty, and those of renunciation, of non-ill-will, and of non-cruelty.26 Whenever one of the first group of thoughts arose, he understood that it:
… leads to (saṃvattati) my own affliction (attabyābādhāya), to others’ affliction, and to the affliction of both;27 it obstructs wisdom (paññānirodhiko), causes difficulties (vighātapakkhito), and leads away from nibbāna (anibbānasaṃvattaniko).

With each of these reflections, such a thought subsided and was then abandoned. The character-forming effect of the nature of one’s thoughts is then emphasised: “Monks, whatever a monk frequently thinks and ponders on (anuvitakketi anuvicāreti), that will become the inclination (nati) of his mind ….” It is then said that when one of the second group of thoughts arose, he understood:
‘This does not lead to my own affliction, or to others’ affliction, or to the affliction of both; it aids wisdom, does not cause difficulties, and leads to nibbāna. If I think and ponder upon this thought even for a night, even for a day, even for a night and a day, I see nothing to fear from it. But with excessive thinking and pondering, I might tire my body; and when the body is tired, the mind becomes disturbed, and when the mind is disturbed, it is far from concentration’. So I steadied my mind internally, quieted it, brought it to singleness (ekodikaromi), and concentrated it.28 Why is that? So that my mind should not be disturbed.
26 The objects of these last three “thoughts” are the same as of the three forms of “right resolve” (sammāsaṅkappa), the second factor of the noble eightfold path. 27 At SN IV.339, it is said that attachment, hatred or delusion lead to willing (ceteti) for the affliction of oneself, others or both. 28 That is, he attained a jhāna, especially the second of these, in which

196

Peter Harvey

Here the emphasis is on mental action that does not lead to afflicting oneself or others (presumably through harmful verbal or physical actions) and which does not make the mind incline away from wisdom or nibbāna. That is, the concern with future results is not in regard to karmically induced pleasures or pains, but on the character trait of wisdom that enables one to go beyond any karmically shaped state. This links with action that is “neither dark nor bright,” and clearly relates to the right resolve factor of the Eightfold Path, itself counted as part of wisdom (MN I.301). The suspension of thought by meditative concentration, even wholesome thought, indicates the more subtle aspects of wholesome/skilful states. The causal sequence thus becomes:
unwise attention → unwholesome roots → unwholesome actions, of unwholesome volition, that are intentional, dark, corrupt, with fault, bring pain and injury to oneself or to others in a way that is anticipated by correct perception of the immediate facts of the situation, criticized by the wise, not to be done → dark harm and pain in this and later lives to the agent, and unwholesome character tendencies, obscuring wisdom and moving one away from nibbāna.

Further reflections on the terms kusala and akusala
Having surveyed what the suttas have to say on these terms and related ones, what do the commentaries say? The Atthasālinī (Asl 38–39; cf. Cousins 1996: 141–142) gives this explanation of the phrase “kusala states” (kusalā dhammā) at the start of the Dhammasaṅgaṇī:
The word ‘kusala’ means healthy (ārogya), faultless (anavajja), skilled (-cheka), ripening in happiness (sukhavipākesu). In such passages as, ‘Is your reverence well (kusalaṃ)? Is your reverence free from ailment?’ etc., kusala means healthy. In such passages as, ‘Venerable sir, what bodily behaviour is kusala? Great king, it is bodily behavvitakka and vicāra, related to “thinking and pondering,” are absent.

Factors related to the kusala/akusala quality of actions

197

iour that is faultless (anavajjo)’ (Bāhitika Sutta, above); and again in, ‘Venerable sir, the Blessed One’s way of teaching Dhamma in regard to kusala states is unsurpassed’ (DN III.102), kusala means faultless. In such passages as, ‘You are kusala at the different parts of a chariot’ (MN II.94); ‘Graceful women who have been trained are kusala in singing and dancing’ (Jat VI.25) etc., kusala means skilled. In such passages as, ‘Monks, it is by the building up of kusala states (that this puñña increases)’ (DN III.58), and ‘from being accumulated (upacitattā) from the doing of kusala actions,’ kusala means ripening in happiness. Now here, in the phrase ‘kusala states,’ either ‘healthy’ or ‘faultless’ or ‘ripening in happiness’ is applicable. … But in regard to word-definitions: kusalas are so called as they cause contemptible evil things (kucchite pāpadhamme) to tremble (salayanti), to shake, to be disturbed, to be destroyed …

Asl 62–63, on a ‘kusala state of mind (citta)’ that is connected with knowledge, says:
‘Kusala:’ kusala in the sense of destroying contemptible states (kucchitānaṃ salanādīhi); or in the sense of healthy (ārogyaṭṭhena), or in the sense of being faultless (anavajjaṭṭhena), or in the sense of produced by skill (kosallasambhūtaṭṭhena). To illustrate: in ‘How are you? Are you kusala sir?,’ kusala is used to mean healthy, i.e. not being ill or sick or unwell in body. So in mental states it should be understood in the sense of ‘healthy,’ i.e. absence of sickness, illness or disease in the form of the defilements (kilesa). Moreover, from the absence of the fault of the defilements (kilesavaj­ ja-), blemish (dosassa) of the defilements, torment (darathassa) of the defilements, kusala has the sense of faultless. Wisdom (paññā) is called skill (kosallaṃ). Kusala has the sense of produced by skill from being produced by skill.

At DN-a 883 (on DN III.102), Buddhaghosa explains the meaning “healthy” as that used in the Jātaka method of exposition, “faultless” as that used in the Suttanta method of exposition, and meanings used in the Abhidhamma method as being: “produced by skill,” “freedom from distress (niddharatha)” and “ripening in happiness.” Cousins (1996: 139–140) comments that “freedom from distress” is otherwise neglected, and says on “ripening in happiness,”

198

Peter Harvey

“the commentators do not in fact often explain the word kusala as having this sense.” He also points out that the ṭīkā on this passage explains “produced by skill” as “caused by appropriate bringing to mind” (yonisomanasikārahetuka), i.e. by what is referred to above as “wise attention.” Cousins (1996: 149) comments that this last meaning is not found only in the Abhidhamma, but “it is intended to suggest that this is in some way a higher or more profound explanation of kusala, or at least one which is more strictly correct.” Here, then, something that is kusala:
a) is morally/spiritually healthy in being unaffected by defilements, i.e. by greed, hatred or delusion. b) is faultless in being free of defilements. c) destroys contemptible states. d) in the case of kusala states connected to knowledge, and perhaps less directly in other cases, is produced by skill, by wisdom. e) more generally, it ripens in happiness.

a)–c) support the translation “wholesome,” d) and the non-delusion-affected aspect of a) support “skilful,” and e) supports something like “beneficial.” Cousins (1996) traces the meaning of kusala/kuśala in pre-Buddhist and Buddhist sources and summarises thus:
1. An original meaning [applied to a person] of ‘intelligent’ ‘wise;’ 2. Expert in magical and sacrificial ritual (in the [pre-Buddhist] Brāhmaṇas); for brahmins, of course, this would precisely constitute wisdom. 3. a) Skilled in meditational/mystic (/ascetic?) practices (in the early Pāli sources and, no doubt, in other contemporary traditions), including skilled in the kind of behaviour which supported meditation, etc., i.e. śīla [keeping moral precepts], etc. b) Skilled in performing dāna [giving] and yañña [sacrifice], now interpreted in terms of central Buddhist ethical concerns; and associated with keeping the precepts and so on. 4. Kusala in later Buddhist and Jain sources becomes generalized to refer to something like wholesome or good states. So there is no reason to doubt that by a later period (i.e. in the commentaries and perhaps later canonical sources) kusala in non-techni-

Factors related to the kusala/akusala quality of actions

199

cal contexts means something which could be translated as “good” (Cousins 1996: 156).

This emphasises “skilful” as the earlier meaning, but it is by no means clear from the above survey that, in the suttas, there is stronger evidence for this than “wholesome” – moreover, this latter meaning cannot simply be equated with “good,” as Cousins implies, as this English term has a rather wide meaning, as Keown argues, though in support of this as the translation for kusala (1992: 119–120). Nevertheless, “skilful” is certainly part of the meaning of the term in the Sutta passages examined here.

Kusala and sīla, samādhi, paññā
It is notable that both the English term “morality” and the Pāli word sīla concern only physical and verbal action: sīla is explained as the right speech, right action and right livelihood aspects of the Noble Eightfold Path (MN I.301). All aspects of the path are kusa­ la, though, so this term covers more than sīla or moral conduct: it also covers aspects of the path pertaining to samādhi and wisdom (Cousins 1996: 144–146). Accordingly: the first jhāna is typically said to be “detached from akusala states” (DN I.73); the four jhānas and other meditative states are referred to as kusala dhammas (Vin I.104 and Vin III.91); the kusala dhammas that the Buddha teaches about are the seven sets later known as the 37 bodhipakkhiya dhammas (DN III.102); and the four right efforts, which pertain to the samādhi aspect of the path concern the ending of akusala dhammas and development of kusala ones (DN II.312). Moreover, the first two of the three forms of right mental action, non-covetousness and non-ill-will, also pertain to mind training, and the third, right view – as either right belief on karma and rebirth etc or noble wisdom (MN III.71–72, below) – pertains to wisdom. Sīla is only part of the path, but the foundation of it all:
So you see, Ānanda, wholesome virtues (kusalāni sīlāni) have freedom from remorse as object and profit; freedom from remorse has gladness; gladness has joy; joy has tranquillity; tranquillity has happiness; happiness has concentration; concentration has seeing things as they really are; seeing things as they really are has turning away

200

Peter Harvey

and non-attachment; turning away and non-attachment have release by knowing and seeing as their object and profit. So you see, Ānanda, wholesome virtues lead gradually up to the summit (AN V.2).

Sīla restrains overt unwholesome actions: those of body and speech (though the mindfulness needed for the more detailed monastic sīla also involves mind-training too). Samādhi counteracts mental unwholesome actions that are expressions of greed/attachment and hatred/ill-will, as well as cultivating their opposites, and paññā digs out the deep roots of greed and hatred and of delusion/ignorance and related wrong view.

Puñña, apuñña and pāpa
Actions of dāna, sīla and bhāvanā (meditative cultivation) are also bases for effecting “puñña” (puññakiriyavatthu; DN III.218; AN IV.214; It 51) and unwholesome actions are often described as “apuñña.” As an adjective, Cousins sees puñña as the “fortunebringing or auspicious quality of an action” (1996: 153), while as a noun “it is applied either to an act which brings good fortune or to the happy result in the future of such an act (1996: 155).” Thus we see:
Monks, do not be afraid of puññas; this, monks, is a designation for happiness, for what is pleasant, charming, dear and delightful, that is to say, puññas. I myself know that the ripening of puññas done for a long time are experienced for a long time as pleasant, charming, dear and delightful. After developing a heart of loving kindness for seven years, for seven eons of evolution and devolution, I did not come back to this world ... (being reborn in a delightful heaven for that time; It 14–15, cf. AN IV.88–89).

As an adjective, puñña can be seen as “auspicious,” “bringing good fortune,” hence “karmically fruitful.” As a noun it refers to the auspicious, uplifting, purifying power of good actions to produce future happy results, or sometimes to such results themselves. The opposite of puñña is apuñña, which one can see as meaning “(an act of) karmic unfruitfulness” or “karmically unfruitful,” i.e. producing no pleasant fruits, but only bitter ones.

Factors related to the kusala/akusala quality of actions

201

A synonym for apuñña is pāpa, which is often given with pu­ ñña as its opposite (e.g. Sn 636). Pāpa is often translated as “evil,” but really means that which is “infertile” (SN IV.315), “barren,” “harmful” (Cousins 1996: 156) or “ill-fortuned” (Cousins 1996: 148). A good way of rendering these meanings would be to see pāpa as an adjective as meaning “(karmically) deadening,” and as a noun as “(karmic) deadness,” meaning that what is so described has a deadening effect on the psyche, making it more constricted and lifeless, rather than having an uplifting, fruitful effect. The term pāpa is fairly frequently used with akusala. AN I.201–205, above, talks of being “overpowered by pāpa, unwholesome states born of greed” and SN V.417 says “do not think pāpaka, unwholesome thoughts (vitakke); that is sensual thought, thought of ill-will, thought of harming.” To have done the kusala is to have “not done what is pāpa” (Vin III.72). The foolish person (bāla) who has done misdeeds, whose past pāpa deeds “envelop” him, thinks:
‘I have not done what is good (kalyāṇaṃ),29 I have not done what is wholesome, I have not made myself a shelter from anguish (bhīruttaṇaṃ). I have done what is pāpa, I have done what is cruel (luddaṃ), I have done what is wicked (kibbisaṃ) …’

Thus he contemplates the bad rebirth he is heading for (MN III.165; cf. AN II.174). Dhp 116 says:
Make haste in doing good (kalyāne); check your mind from pāpa; For the mind of one who is slow in doing the karmically fruitful (puññaṃ) delights in pāpa.30

Cone sees kalyāṇaṃ as meaning “what is good, excellent; virtuous action …” (Cone 2001: 665). At AN II.222–223, pāpa is to do the ten bad deeds starting with killing. Kalyāṇa is to refrain from these. Or these are respectively wrong view to wrong liberation, and right view to right liberation. 30 Note that, just at the English terms “evil” or “bad” can refer both to immoral actions and bad things that happen to one, and puñña can refer to both bad actions and their fruits, pāpa can refer to both actions and things that happen to one: “Even a pāpa person sees good fortune (bhadraṃ) as long as the pāpa does not ripen (paccati); but when pāpa ripens, then the pāpa person sees pāpas” (Dhp 119).

29

202

Peter Harvey

Clearly, the referent of the term puñña overlaps to a considerable extent with that of kusala, but it emphasises the aspect of a good action that is its power to bring future happy karmic results. Cousins comments that in Buddhist canonical texts, puñña “occurs much less frequently than kusala” (1996: 154) and suggests that it “was almost certainly not a technical term in the thought of the Buddha and his early disciples. It was no doubt a part of the background of beliefs current at the time” (p. 155). One can say that in terms of the causal sequence described in this paper, doing an action with an orientation towards its being puñña – a puñña-centric action – emphasises its future karmic results: the happiness that the action will in time engender in one. Typically, in the suttas, such action is associated with giving and other aspects of lay religiosity (Cousins 1996: 154). Doing an action with an orientation to its being kusala – a kusala-centric action – focuses more on the nature of the action itself, and its roots, and its ongoing effect on character, perhaps with wisdom and nibbāna in mind. Ironically, though, the greater the concern with puñña, the lesser the degree of the good result. It is said that, “the mental aspiration of a moral person (sīlavato) is effective through its purity (suddhattā)” (DN III.259), in terms of the rebirth aspired to in giving a gift, but for such an aspiration to truly work, it should not be itself the sole motive of the giving. If a person gives something to a monk “with longing (sāpekkho), with the heart bound (to the gift; paṭibaddha-citto), intent on a store (of karmic fruitfulness; sannidhipekho), thinking ‘I’ll enjoy this after death’ (imaṃ pecca paribuñjissāmi ti),” it is said that he will be reborn for a while in the lowest of all the heavens. A series of what seem to be meant as progressively higher motives is then outlined: giving because one feels “it is auspicious (sāhu) to give!;” wishing to continue a family tradition of giving; wishing to support those who do not cook for themselves; because great sages of the past were supported by alms; because giving leads to mental calm, joy and gladness; or because giving enriches the heart and equips it for meditation (AN IV.60–63). Giving from the last of these motives is then said to lead to rebirth in the first heaven of the realm of (elemental) form, where the brahmās dwell. Thus doing a good action simply because it is seen to have pleasant results is not the highest of motives – it is

Factors related to the kusala/akusala quality of actions

203

better to value goodness in itself, and the peace and wisdom that it facilitates. Accordingly, Buddhaghosa says on sīla:
That undertaken just out of desire (-kāmatāya) for fame is inferior; that undertaken just out of desire for the fruits of karmically fruitful actions (puññaphala-) is medium; that undertaken for the sake of the noble state thus, ‘This is to be done’ is superior (Vism 13).

– with all three involving forms of chanda, or desire-to-do. Any akusala action will also be apuñña, and vice versa. Any puñña action will also be a kusala one, i.e. be harmless, and contribute to skilful, wholesome, faultless states. Even to give from desire for the future karmic benefit of this is kusala to a degree – desire is not necessarily akusala, as is shown by the fact that chanda can be a kusala state, it being the basis of one of the four bases of spiritual success, the iddhi­padas (e.g. DN II.213–214). A person may initially focus on generating puñña, perhaps by acts of dāna, only to contribute to his or her future worldly happiness, but then also to help give a supporting basis for the path to nibbāna. That said, there remains a question of whether all kusala actions are also puñña ones. The three aspects of the nidāna of saṅkhāras are puñña’bhisaṅkhāras, apuñña’bhisaṅkhāras and āneñja’bhisaṅkhāras (DN III.217). The last of these are actions leading to the formless realms (in meditation or rebirth) (Vibh 135), and while such actions will involve kusala states of mind, they lead to a state in which there is only neutral feeling, not happiness: hence they are differentiated from the first, puñña, ones. In relation to the noble path, moreover, the actions which are “neither dark nor bright” will still be kusala, but not puñña, as they pertain to going beyond rebirths, and any happiness they may contain. Indeed we see that the Mahācattārīsaka Sutta (MN III.71–72) differentiates between two kinds of right view:
[T]here is right view that is affected by taints (sāsavā), partaking of puñña (puññābhāgiyā) ripening on the side of attachment (upadhivepakkā), and there is right view that is noble, taintless, transcendent, a factor of the path (ariyā anāsavā lokuttara maggaṅgā).

204

Peter Harvey

The first kind of right view is to believe in karma and rebirth, and that some renunciants and brahmins know of these directly, whereas the second is:
… the wisdom (paññā), the faculty of wisdom, the power of wisdom, the investigation-of-states awakening factor, the path factor of right view in one whose mind is noble, whose mind is taintless, who is possessed of the noble path and is developing the noble path …

By the time a person has attained stream-entry, and so is destined for arahantship within seven lives at most (AN V.120) their sīla is perfected (AN I.231–232), though they do not cling to their sīla (aparāmaṭṭhehi, SN V.343–344) as the whole of the path. Moreover, while their emphasis is no longer on the puñña aspect of good actions, their actions still generate much puñña as a side-effect of their quest for nibbāna: the four factors of streamentry (SN V.343) are also called “streams of puñña (puññābhisandā), streams of the wholesome (kusalābhisanā), nutriments of happiness” (SN V.391), they cannot have any sub-human rebirths, and their future lives may include millions of years in a heavenly realm. The arahant goes wholly beyond puñña (Sn 636; Harvey 2000: 43–46), as he or she will have no future rebirths. Those who do kusala/puñña actions out of concern for the future karmic fruits (though still with regard to not afflicting others) may not be directly aiming for nibbāna, but their actions still help to prepare the ground for later efforts to attain this goal. It is in this sense that Velez de Cea (2004: 129) calls them “instrumental” actions in that they are “actions leading to favourable conditions for cultivating nirvanic virtues,” in contrast to “teleological” actions which are “actually displaying nirvanic virtues or virtues characteristic of the Buddhist ideal of sainthood,” whether or not a person doing such action is consciously aiming for nibbāna. Adam (2005: 74) uses similar language but in a different way, saying that actions aimed at nibbāna, as they still generate puñña as a side-effect, while “teleologically nirvāṇic,” are also “instrumentally karmatic.” On the other hand, actions that are focussed on generating good karmic results are “teleologically karmic” and, as a side-effect, “instrumentally nirvāṇic.” That said, it is perhaps clearer to here

Factors related to the kusala/akusala quality of actions

205

replace “instrumentally” by “indirectly” and “teleologically” by “directly:”
• ‘directly nirvāṇic’ actions directly display virtues that are most fully embodied in the kind of person who has fully experienced nibbāna, the arahant, and directly contribute to attaining arahantship – whether or not this is the conscious aim of such actions. These are also ‘indirectly karmatic’ as they have the side-effect of producing good karmic fruits – except in the case of the arahant, that is, as he or she generates no more karmic fruits, though inspires others to act in ways that generate good karmic fruits. • ‘directly karmatic’ actions consciously aim at good karmic results, and are ‘indirectly nirvāṇic’ as they have the side-effect of helping to prepare the ground for ‘directly nirvāṇic’ actions.

Comparison to Western ethical theories
If we survey the above factors related to an unwholesome/unskilful (akusala) actions, which actions are also misfortune-bringing (apuñña, pāpa) ones, we have:
1. unwise attention, feeding 2. attachment/greed/covetousness, hatred/ill-will, and delusion/ignorance, which are both ‘the unwholesome’ and are roots that that sustain 3. ‘the unwholesome’: specified unwholesome actions of body, speech or mind 4. that are of unwholesome volition, and intentional, 5. that are dark, corrupt, with fault/blameable (by the wise), as they 6. bring pain and injury to oneself or to others 7. in a way that is anticipated by correct perception of the immediate facts of the situation, 8. such that one should not inflict on another what one would not like inflicted on oneself 9. and are criticized by the wise, 10. not to be done 11. and that, as a karmic result, bring dark harm and pain in this and later lives to the agent, as well as 12. unwholesome character tendencies,

206

Peter Harvey

13. obscuring wisdom and moving one away from nibbāna.

Wholesome actions have the opposite qualities. Of these factors:
• realist ethics focuses simply on the body and speech aspects of 3. • utilitarianism focuses on 6, perhaps 7, and the body and speech aspects of 3. • Kantian ethics, with its emphasis on the good will, universalisability, and duty, focuses on 4, 8, 10, 3, and perhaps 5. • Aristotelian virtue ethics focuses on 12, an analogy to 13, 2 and perhaps 3. A further virtue aspect of Buddhist ethics is that in seeking to do acts rooted in non-greed, non-hatred and non-delusion, one is cultivating virtues that have a strong affinity with the destruction of greed/attachment, hatred and delusion that is the goal of the Buddhist path.

None of these particularly focuses on 1, 9, 11, or the nibbāna aspect of 13, all of which except 11 connect to wisdom and skilfulness. Buddhist ethics encompasses all these aspects. A beginner may focus on 11, the particularly puñña aspects, but more serious practitioners focus on the rest, especially 1 and 13. Of course with a number of factors being involved, one may ask whether they can conflict, necessitating a choice of one or more factors over others. For example, can an injurious action (6) originate in a well intentioned action of wholesome volition (the opposite of 4)? I would say that, taking into account aspect 7, a genuinely wholesome volition may result in harm to others or oneself if it is based on incorrect factual information or incorrect perception of the situation. Nevertheless, when insufficient care has been taken to ascertain the facts of the situation, this may bring in culpable carelessness, so as to make the action unwholesome to some extent. There may, though, be cases where the agent of a harmful action thinks of the action as wholesome and beneficial, but where delusion, also leading to wrong view, is one of the action’s roots, rendering it unwholesome. For example, Hitler may have genuinely thought his actions were right. One might also think that a beneficial result can come from an action of unwholesome volition, for example when a starving person is fed by killing and cooking an animal for him. But here, both benefit and harm are actually the result, and the full

Factors related to the kusala/akusala quality of actions

207

action involves both wholesome (to feed a hungry person) and unwholesome (to kill) volitions. How a person deals with potential conflicts of moral criteria may depend on where they are on the path, for example a lie (cf. 3) may help prevent harm to another (6), but strengthen an unwholesome tendency (12) and even have some minor bad karmic results (11); yet with wise attention (1), one may see a way to both avoid lying, and preventing harm to the person. Of course the Mahāyāna is more open to “skilful means” precept breaking, out of compassion. Moreover, potential conflict between certain factors may be skilfully dealt with by bringing in another factor, for example someone’s wish to ordain, or an act of great generosity, may bring suffering to some relatives (cf. 6), but help generate good character traits (12) and facilitate a movement towards nibbāna (13); moreover the harm may be counteracted by sharing puñña with the relative (cf. 11), and the strengthened good character traits, perhaps including wisdom, will also bring benefit to others in the future. While Keown (1992) has argued that a virtue ethic model is the best fit for early Buddhist ethic, this study shows that while there is a virtue-ethics dimension to such ethics, a utilitarian concern for directly caused suffering and happiness, and a Kantian concern for a good will, are also present. There are also plenty of specific ethical rules, which accords with a kind of realist ethics. These aspects are integrated into one system, and as Velez de Cea says, “Early Buddhist ethics is sui generis, that is, one of a kind” (2004: 138). The key content of akusala actions are specified actions that directly (actions of body and speech) or indirectly (mental actions) bring pain and affliction to others, oneself, or both, and the greed/attachment, hatred and delusion that are the motivating or orientating roots of these. Kusala states are actions which avoid these actions and roots, and are the opposite of them; moreover, they include the meditative states involved in the development of the Eightfold Path, which go further than, but still depend on, ethics per se. Kusala states come from wise skill and contribute to wise skill, and are both morally and spiritually wholesome: morally faultless, nourishing further wholesome states, healthily without greed, hatred or

208

Peter Harvey

delusion, and contributing to the end of these. They bring no harm to anyone, and lead to happiness for the agent of them.

Literature Primary sources
The above translations are the author’s own, though they are generally close to those listed below.
AN Asl Dhs Dhp DN DN-a It Jat MN SN Sn Vin Vibh Vism Aṅguttara-Nikāya: 1961, eds. R. Morris and A.K. Warder; 1888, ed. R. Morris; 1897, 1899, 1900, ed. E. Hardy. Atthasālinī: 1897, ed. E. Müller. Dhammasaṅgaṇī: 1885, ed. E. Müller. Dhammapada: 1994, eds O. von Hinüber and K.R.Norman. Dīgha-Nikāya: 1890, 1903, ed. T.W. Rhys Davids and J.E. Carpenter; 1911, ed. J.E. Carpenter. Dīgha-Nikāya-Aṭṭhakathā = Sumaṅgalavilāsinī: 1929, eds T.W. Rhys Davids and J.E. Carpenter; 1931, 1932, ed. W. Stede. Itivuttaka: 1889, ed. E. Windisch. Jātaka with Commentary: 1877–96, ed. E. Fausboll. Majjhima-Nikāya: 1888, ed. V. Treckner; 1898, 1899, ed. R. Chalmers. Saṃyutta-Nikāya: 1884–1898, ed. M.L. Freer. Suttanipāta: 1913, eds D. Andersen and H. Smith. Vinaya Piṭaka: 1879–1883, ed. H. Oldenberg. Vibhaṅga: 1904, ed. C.A.F. Rhys Davids. Visuddhimagga: 1920, 1921, ed. C.A.F. Rhys Davids.

Secondary sources
Adam, M. T., 2005, “Groundwork for a Metaphysic of Buddhist Morals: a New Analysis of puñña and kusala, in the Light of sukka,” Journal of Buddhist Ethics, Vol. 12: 62–85; http://www.buddhistethics.org/12/adamarticle.pdf. Last visited 06-07-2011. Anālayo, 2007, “What the Buddha Would Not Do, According to the Bāhitikasutta and its Madhyama-āgama Parallel,” Journal of Buddhist Ethics, Vol. 14: 153–179; http://www.buddhistethics.org/14/anaalayo-article.pdf. Last visited 06-07-2011.

Factors related to the kusala/akusala quality of actions

209

Cone, M., 2001, A Dictionary of Pāli, Part I, a–kh, Oxford, The Pali Text Society. Cousins, L. S., 1996, “Good or Skilful? Kusala in Canon and Commentary,” Journal of Buddhist Ethics, Vol. 3: 136–164; http://www.buddhistethics. org/3/cousins1.pdf. Harvey, P., 1999, “Vinaya Principles for Assigning Degrees of Culpability,” Journal of Buddhist Ethics, Vol. 6: 271–291; http://jbe.gold.ac.uk/6/harvey991.pdf. Last visited 06-07-2011. Harvey, P., 2000, An Introduction to Buddhist Ethics: Foundations, Values and Issues, Cambridge, Cambridge University Press. Harvey, P., 2011, “The Nature of the Eight-factored Ariya, Lokuttara Magga in the Suttas Compared to the Pali Commentarial Idea of it as Momentary,” Religions of South Asia, Vol. 5 – forthcoming. Keown, D., 1992, The Nature of Buddhist Ethics, London, Macmillan. Velez de Cea, A., 2004, “The Criteria of Goodness in the Pāli Nikāyas and the Nature of Buddhist Ethics,” Journal of Buddhist Ethics, Vol. 11: 123–142; http://www.buddhistethics.org/11/vele0401.pdf. Last visited 06-07-2011.

Value pluralism in early Buddhist ethics
Abraham Vélez de Cea

Introduction
Following Susan Wolf, I understand ethical pluralism as “the view that there is an irreducible plurality of values or principles that are relevant to moral judgment.”1 Wolf distinguishes between two types of ethical pluralism: evaluative and deontic. Evaluative pluralism is concerned with the nature and the lexical order of values or goods. Deontic pluralism on the other hand, discusses principles, not in the sense of specific precepts but rather in the sense of decision procedures and criteria to determine the rightness of actions. I have argued elsewhere that early Buddhist ethics proposes several criteria to determine the rightness of actions.2 In the present essay, I focus on the specific type of value pluralism represented in early Buddhist ethics. The distinction between evaluative and deontic pluralism is important because one can be a pluralist in terms of values and a monist in terms of principles. For instance, contemporary consequentialists tend to be pluralists in terms of values but monists in terms of principles. That is, they accept diverse kinds
Wolf 1992: 785. Vélez de Cea 2004a. I claim that the Pāli Nikāyas contain several criteria to determine the rightness of actions. I use the term goodness instead of rightness to challenge narrow understandings of the realm of the good in Buddhist ethics. Specifically, I question the abhidharmic tendency to reduce the moral fruitfulness of actions to the wholesomeness of the agent’s motivation, that is, the tendency to conflate the rightness of actions with the goodness of the agent’s motivation. In other words, I question interpretations of early Buddhist ethics as agent-based forms of virtue ethics. Instead, I view early Buddhism as presupposing a pluralistic approach to virtue ethics.
2 1

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 211–237

212

Abraham Vélez de Cea

of goods but the rightness of actions depends exclusively on the net amount of value that an action, rule, or virtue promotes. In the first section of this essay I discuss the principal values that can be inferred from moral exhortations and what early Buddhists considered spiritual and non-spiritual benefits. In the second section, I clarify the nature of early Buddhist values. In order to accomplish this goal, I challenge the application to the Pāli Nikāyas of diverse categories common in current western ethical discourse. As an alternative, I distinguish between ultimate and penultimate values, and speak about their enabling, favoring and intensifying functions. In the third and final section, I compare the lexical order of values in early Buddhism to Damien Keown’s (1995) list of three basic Buddhist goods.

1. Deriving values from exhortations and claims regarding spiritual and other benefits
In this article, I do not deduce early Buddhist values from our natural inclinations or from what make humans good and healthy specimens of their kind. Likewise, I do not derive early Buddhist values from a priori accounts of what constitutes human fulfillment or the highest function of humans. Rather, I infer early Buddhist values from the most common moral exhortations found in the Pāli Nikāyas, and from what these texts consider beneficial, whether spiritually or non-spiritually. The non-spiritual benefits to which the Pāli Nikāyas refer reflect the actual values of early Buddhists. These non-spiritual benefits are clearly presented as inferior to spiritual ones. Nonetheless, they are nevertheless presented as valuable and worthy of pursuit. From the early Buddhist perspective, these benefits usually derive from ethical and spiritual practices, but this is not necessarily the case (SN IV.230; AN V.10). Even if it were the case that non-spiritual benefits, such as prosperity, repute, social influence and status, were always the consequence of ethical and spiritual development, this would not render non-spiritual benefits valueless.

Value pluralism in early Buddhist ethics

213

Some scholars of Buddhist ethics tend to reduce the value of non-spiritual benefits to the value of spiritual benefits.3 However, interpreting the value of non-spiritual benefits as mere consequences of spiritual values makes the Pāli Nikāyas “too spiritual.” Besides, this romanticized “too spiritual” interpretation of early Buddhist axiology renders early Buddhism unnecessarily inconsistent with the actual practice of many traditional Buddhists today. Instead of contrasting an ideal abhidharmic Buddhist ethics and a somehow degenerated Buddhist ethics in practice, I interpret the Pāli Nikāyas and early Buddhism as legitimizing from the beginning diverse types of ethical practice, not always motivated by pure and ideal spiritual concerns. Moral exhortations usually appear in the form of precepts, but sometimes they appear as part of simple prudential advice. Here however, I understand the term moral in a broad sense as including not only ethical precepts but also social and spiritual recommendations, which, from a traditional Buddhist perspective, belong to the ethical realm. Precepts point to morally significant items that require from us diverse kinds of response: protection, cultivation, respect, love. Not only positive action, but failure to respond adequately to these items, constitutes “doing” something that is ethically relevant, or using Buddhist terminology, performing an act that is karmically fruitful or unfruitful, wholesome or unwholesome. Thus, precepts indicate the existence of items and actions that are worthy, valuable, good. As Damien Keown states: “precepts gesture beyond themselves in the direction of certain values which it is their function to preserve. Their formulation as negative recommendations flashes an alert that anyone contemplating such actions as killing or stealing is threatening an assault on certain values or ‘goods’.”4 The Pāli Nikāyas contain many sets of moral exhortations. Here, I limit myself to discussing the lists of five and ten precepts, the ten
3 See for instance Damien Keown’s account of early Buddhist ethics, where non-spiritual benefits of actions are reduced to “non-moral secondary consequences entrained by moral acts” (2001: 128). 4 Keown 1995: 41.

214

Abraham Vélez de Cea

wholesome actions, and the three sections on ethical conduct (śīla) of the Brahmajāla Suttanta. The following analysis is not intended to provide a comprehensive list of early Buddhist values, just a sense of their nature. The five precepts are, as Richard Gombrich states, the most obvious principles through which Buddhist values may be examined.5 Even though the Pāli Nikāyas do not correlate the five precepts with values and virtues, it seems possible to correlate the first precept – abstaining from taking life – to the value of life and virtues such as friendliness or loving-kindness and compassion. Similarly, the second precept – abstaining from taking what is not given – can be related to the values of property and social justice. These values can be protected by virtues such as non-greed, generosity, and honesty. The third precept – abstaining from sensual misconduct – can be connected to the values of self-control, moderation, and fulfilling role-dependent duties: celibacy in the case of monks and nuns, and faithfulness in the case of lay people. These values can be guarded by several virtues including temperance, contentment, and non-greed. The fourth precept – abstaining from telling lies – expresses what Peter Harvey calls the value of “seeking truth and seeing things as they are,”6 which can be cultivated through virtues such as wisdom, mindfulness, and investigation of things (dham­ mavicaya). The fifth precept – abstaining from taking intoxicants – reflects the Buddhist concern for mental health, as well as for the value of seeking truth and seeing things as they are. Mindfulness, which in this context includes sobriety, is a fundamental Buddhist virtue because it protects not only the values of mental health and seeing things as they are, but also the values associated with the other precepts. That is, mindfulness is a key Buddhist virtue because it prevents reckless moral conduct and the breaking of precepts in general. The list of ten precepts includes the first five precepts plus five extra precepts. Today, these precepts are usually observed by novices, and nuns who are denied full ordination. On special occasions
5 6

Gombrich 1992: 98. Harvey 2000: 75.

Value pluralism in early Buddhist ethics

215

such as full moon days, lay people may decide to observe some or all of these extra precepts. The sixth precept – abstaining from eating at the wrong time – like the third precept, seems to be intended to preserve the values of self-control and moderation, which relate to the virtues of temperance and contentment. The seventh precept – abstaining from dancing, singing, music, and watching shows – is difficult to understand outside its original cultural context. The seventh precept is not so much about dancing, singing, music, and watching shows per se, but rather about the ambiance that may have surrounded such activities in ancient India, an ambience that can still be experienced in rural South Asia. Thus, I prefer to interpret the seventh precept as expressing the value of circumspection, which relate to the virtues of mindfulness and prudence in the sense of ability to avoid situations that may endanger moral conduct. Similarly, the eighth precept – abstaining from garlands, perfumes, cosmetics, and adornments – does not target concern for beauty as such, or concern for looking good, smelling good, and being clean or elegant. Rather, the precept seems to be intended to undermine attitudes that go against the values of simplicity and humility, for instance, lack of modesty and vanity. Likewise, the ninth precept – abstaining from using high seats or beds – appears to be intended to preserve the values of simplicity and humility, thus, opposing extravagance and vanity. Finally, the tenth precept – abstaining from accepting gold or silver – applies to monks and nuns, and protects the value of right livelihood, renunciation and simplicity through virtues such as contentment and detachment from material possessions. The first four of the ten wholesome actions overlap with the first four precepts. The fifth wholesome action – abstaining from divisive speech – and the sixth – abstaining from harsh speech – seem to relate to social values such as friendship, peace and harmony. These social values can be protected by virtues such as lovingkindness, truthfulness, compassion, non-violence, and even calmness and patience. The seventh wholesome action – abstaining from frivolous speech or idle chatter – demonstrates the Buddhist concern for what is practical in the sense of being directly related to ethical and spiritual development. This focus on what is practical can be related to many virtues, primarily wisdom and diligence.

216

Abraham Vélez de Cea

The eighth wholesome action – abstaining from covetous thoughts – expresses the values of mental health, simplicity, selfcontrol and moderation, undermining negative mental states such as envy, avarice, greed. The ninth wholesome action – abstaining from ill-will – connects to the value of mental health, peace, and harmony, protected through virtues such as loving-kindness and compassion. Finally, the tenth wholesome action – abstaining from wrong views – concerns the values of mental health, seeking truth and seeing things as they are, which can be preserved by the virtues of mindfulness, investigation of things, and wisdom. The moral exhortations that appear in the three sections on ethical conduct of the Brahmajāla Suttanta overlap to a great extent with the aforementioned sets of precepts. Here I will discuss briefly the exhortations that differ from those already discussed. For instance, respect for seeds and plants shows concern for the values of life and the environment. Several virtues including love, compassion, and wisdom may help to guard these values. Abstaining from accepting uncooked grain, raw meat, slaves, animals, and fields, like the former precept, applies to monks and nuns, and relate to the values of right livelihood and fulfilling role-dependent duties, both preserved by the virtue of renunciation. Similarly, abstaining from certain activities such as running messages, buying and selling, dealing with false weights, bribery, cheating, predicting the future by different methods, and reciting charms and incantations to benefit or harm others, embodies the value of right livelihood, though in this case it is less clear that the precept is intended exclusively for monks and nuns. Trading with false weights, bribery, and cheating are also prohibited for lay people. Although many early Buddhist values are linked to moral exhortations and precepts, not all of them are. The second strategy to infer early Buddhist values is to analyze what the Pāli Nikāyas considered benefits, that is something worthy, valuable or good. The number of texts that could be used to infer values from spiritual and non-spiritual benefits is endless. Here, I provide just a few representative examples. Among the texts that connect non-spiritual benefits and virtuous conduct, I will here focus on two: AN IV.197 and DN III.180–193.

Value pluralism in early Buddhist ethics

217

In AN IV.197, the Buddha tells queen Mallikā that material benefits such as wealth, influence, and beauty result from not displaying anger, hatred, resentment, generosity to holy people, and from not being envious of others’ achievements, honor, and respect. In the Sigālaka Suttanta (DN III.180–193), the Buddha explains the dangers of taking intoxicants, haunting the streets at night, attending fairs, gambling, keeping bad company, and idleness. By avoiding this type of conduct, one avoids many negative consequences and achieves many positive ones. Specifically, by not taking intoxicant one avoids “present waste of money, increased quarreling, liability to sickness, loss of good name, indecent exposure of one’s person, and weakening of the intellect” (DN III.182–183). From these diverse benefits, we may infer values such as prosperity, peace, health, beauty, repute, and mental health. The Sigālaka Suttanta is also useful in that it allows us to identify social values such as friendship, respect for status, and fulfillment of role-dependent duties, i.e. fulfillment of the duties of parents toward children, teachers toward students, husbands toward wives, friends toward friends, employers toward employees, holy people toward followers, and vice versa. Among the texts that connect virtuous conduct to both spiritual and non-spiritual benefits, I discuss two: MN III.202–206 and MN I.32ff.. In MN III.202–206, the Buddha explains that killing leads to a bad rebirth, or in case a killer is born as a human being, to a short-life. Conversely, abstaining from killing leads to a happy rebirth, or if born as a human being, to a long life. Injuring beings leads to an unhappy rebirth or if one is reborn as human, to being sick frequently. Conversely, respecting life leads to a good rebirth, or to a healthy life if one is born as human. Being angry and irritable, displaying hate and bitterness makes one ugly; doing the opposite makes one beautiful. Envy of others’ achievements and honors leads to being powerless, doing the opposite makes one influential. Being stingy with holy people leads to poverty and giving to wealth. Obstinance and arrogance and failure to respect those worthy of respect leads to a low-birth; the opposite leads to being high-born. Finally, failure to ask wise and holy people about what is wholesome and unwhole-

218

Abraham Vélez de Cea

some, about what someone should and should not cultivate, and about the actions conducive to harm and suffering, versus welfare and happiness, leads to stupidity in a next life; the opposite leads to wisdom. The foundation of the lists of benefits found in MN III.202–206, appears to be a hierarchy of values: first, biological values such as life and health; second, worldly values such as beauty, influence, prosperity, and social status; third, spiritual values such as wisdom and discernment. This hierarchy of values, demonstrates that spiritual values such as wisdom and discernment are superior to non-spiritual values. However, the fact that biological values appear before worldly values does not seem to suggest that life and health are less important than worldly values. Instead, they seem to come first because they are the preconditions of achieving the other values. That is, without life and health, worldly values cannot be enjoyed. The conclusion of MN III.202–206 expresses two central values of early Buddhism: self-reliance and personal responsibility: “Beings are owners of their actions, student, heirs of their actions; they originate from their actions, are bound to their actions, have their actions as their refuge. It is action that distinguishes beings as inferior and superior.”7 A slightly different hierarchy of Buddhist values can be inferred from MN I.32ff. In this text, the Buddha exhorts his monks to practice the precepts, serenity of mind, meditation, and insight in order to achieve diverse benefits. The first set of benefits is being dear to fellow monks and respected by them, which I interpret as expressing the social values of friendship and respect for spiritual status. Although the Buddha is addressing monks in this text, it can be inferred that the values of friendship and respect for spiritual status are relevant not only for monks and nuns, but also for lay practitioners. The second set of benefits is obtaining the four material requirements of monks and nuns: robes, alms, resting place, and medicine.
kammasakkā māṇava, sattā kammadāyādā kammayoni kammabandhu kammapaṭisaraṇā. kammaṃ satte vibhajati yadidaṃ hīnappaṇītatāyāti.
7

Value pluralism in early Buddhist ethics

219

These benefits presuppose concern for basic material values such as clothing, food, shelter, and health care. Although the text refers to the four requirements of monks and nuns, the values underlying the text might be extrapolated to lay practitioners as well. The third set of benefits is spiritual in nature: achieving great consequences and benefit for those who provide the four material requirements, which reflect to the value of giving and the virtue of generosity. Generosity is generally a virtue to be cultivated by lay people. However, as Gregory Schopen has shown, historically, it has been cultivated by many monks as well. Achieving great consequences and benefits for dead kinsmen and relatives who remember the qualities of holy people, reflect the value of faith, both in monastics and lay people. Conquering the emotional ups and downs, fear and dread, demonstrate the value of self-control and a variety of virtues, primarily temperance. Obtaining the pleasant states of the four jhānas and the four immaterial jhānas, here described as “the peaceful liberations that transcend form and are immaterial,”8 can be related to the value of calm meditation and virtuous mental factors conducive to concentration (samādhi), including bliss and equanimity. The fourth set of benefits, also spiritual in nature, consists in becoming one of the four particular types of persons: a streamenterer (sotāpatti), a once-returner (sakadāgāmī), a non-returner (anāgāmī), or an enlightened being or arahant. The enlightened being is described as someone who acquires nirvanic knowledge and powers: the divine ear, the ability to know the minds of others, recollect former lives, perceive with the divine eye the passing away and reappearance of beings, and liberation of the mind through wisdom and the destruction of taints. One becomes a stream-enterer by destroying the three lower fetters (identity views, doubt, and attachment to rules and observances); a once-returner by attenuating lust (rāga), hate, and delusion; a non-returner by destroying the five lower fetters (the former three plus sensual desire and ill will); and an enlightened being by eradicating the remaining fetters (desire for fine-material existence, desire for inmaterial existence, conceit,
8

santā vimokkhā atikkamma rūpe āruppā.

220

Abraham Vélez de Cea

restlessness, and ignorance). All these spiritual benefits point to the values of insight meditation and spiritual powers, yet overall to the value of character cultivation, which is protected and developed by many virtuous mental factors. The fact that the development of values culminates in becoming a certain type of person suggests that the development of character traits characteristic of enlightened beings is the supreme early Buddhist value. The fact that social values appear before material values does not seem to show a particular ranking of non-spiritual values. Both social and material values can be interpreted as favorable conditions for the cultivation of spiritual values.

2. The complex nature of early Buddhist values
In order to better understand the nature of early Buddhist values, we need to apply to the Pāli Nikāyas the proper set of hermeneutical categories. The objective is not to interpret Buddhist values in terms of non-Buddhist concepts but rather to determine what concepts are most helpful to interpret Buddhist values on their own terms. The most common axiological distinction in western philosophical ethics is that between intrinsic and instrumental values. Intrinsic values are those things valued for their own sake, and instrumental values are those valued for the sake of something else. This distinction – although not in cognate terminology – appears explicitly in the Pāli Nikāyas. For instance, in SN III.189, the monk Rādha asks for the purpose or goal (attha) of seeing correctly (sammādassanaṃ), and the Buddha replies that it is disenchantment (nibbidā). Next, Rādha asks for the purpose of disenchantment, and the Buddha responds that it is dispassion (virāgo). Rādha asks for the purpose of dispassion, and the Buddha replies that it is liberation (vimutti), which in this context refers to the meditative absorptions called immaterial jhānas. Once again, Rādha asks for the purpose of liberation, and the Buddha replies that it is nibbāna (English nirvana). Finally, when Rādha asks for the purpose of nirvana the Buddha replies “You have gone beyond the range of questioning, Rādha. You weren’t able to grasp the limit of your

Value pluralism in early Buddhist ethics

221

questioning. For, Rādha, the holy life is lived with nirvana as its ground, nirvana as its destination, nirvana as its final goal.” Thus, the value of seeing things as they are, the value of spiritual renunciation, and the value of calm meditation would be instrumental, whereas nirvana would be the only intrinsic value. Similarly, in MN I.149–150 the simile of the seven relay chariots seems to suggest that different types of purification are instrumental in leading gradually toward nirvana. The seven are the purifications of 1) virtue, 2) mind, 3) view, 4) overcoming doubt, 5) knowledge and vision of what is the path and what is not the path, 6) knowledge and vision of the way, and 7) knowledge and vision, which is said to be for the sake of reaching final nirvana (parinibbāna) without clinging. However, while these two texts indicate that nirvana is the only thing valued for its own sake, it would be inaccurate to conclude that what is valuable for the sake of nirvana is simply valued as an instrument or means to attain nirvana. That would be simplistic and inconsistent with the Pāli Nikāyas. Damien Keown has conclusively demonstrated that ethical practice is not only instrumental but also constitutive of nirvana.9 Ethical practice is valuable for the sake of attaining nirvana as well as for its own sake. Furthermore, mental qualities constitutive of nirvana can be both instrumentally and intrinsically valuable. For instance, wisdom and compassion are intrinsically valuable and at the same time instrumentally valuable in the sense that they contribute to the achievement, spread, or implementation of other values. To claim that something can be at the same time both instrumentally and intrinsically valuable is neither problematic nor at odds with practice in contemporary philosophical ethics. For instance, Thomas Hurka acknowledges that something intrinsically good or evil can also have instrumental qualities.10 Similarly, Christine Korsgaard speaks about things “that human beings might choose partly for their own sake under the condition of their

9 10

Keown 2001. Hurka 2001: 21.

222

Abraham Vélez de Cea

instrumentality.”11 Nor is this a new development in the West. Plato speaks of the value of justice as desired for its own sake,12 as well as for the sake of something else, namely, for its consequences in this life and the next.13 Similarly, Aristotle speaks of friendship as instrumentally good,14 as well as a good for its own sake whatever other benefits it may yield.15 The Pāli Nikāyas refer to goods that are both intrinsically and instrumentally valuable. But in order to understand the full diversity of early Buddhist values, we need something more than the dichotomy between intrinsic and instrumental values. Christine Korsgaard has expanded the classical distinction between intrinsic and instrumental values, introducing two new distinctions: intrinsic-extrinsic and instrumental-final. The intrinsic-extrinsic distinction concerns the source of a value: intrinsic goods are those that have value in themselves, that is, in virtue of its intrinsic, nonrelational properties; extrinsic goods derive their value from something else, that is, they have value in virtue of its extrinsic, relational properties. The final-instrumental distinction, on the other hand, concerns not to the source of value, but to the reasons for valuing something: final goods are valued for their own sake as ends, whereas instrumental goods are valued for the sake of something else as means.16 (Other philosophers have further challenged the distinction between final and instrumental values, arguing that collapses, and proposing instead a distinction between two types of final values, final intrinsic and final extrinsic values.)17 These conceptual elaborations are philosophically interesting; however, in my view, they are often misleading when applied to early Buddhist values. The final-instrumental distinction faces the same difficulty as the intrinsic-instrumental distinction. From the
11 12 13 14 15 16 17

Korsgaard 1996: 264. Plato. Republic, Books II–IV. Ibid., Book X. Aristotle. Nicomachean Ethics, 1099a31–b6. Ibid., 1155a29–32, 1159a27. Korsgaard 1996: 111. See also Korsgaard 1982. Rønnow-Rasmussen 2002.

Value pluralism in early Buddhist ethics

223

fact that nirvana is valued as the final good, it does not follow that all other goods must be valued as mere instruments or means to attain the values constitutive of nirvana. Moreover, from the fact that something is instrumentally valued by someone, it does not follow that it cannot be at the same time valued for its own sake by someone else. Likewise, something can be final with respect to X but instrumental with respect to Y. For instance, seeing correctly is instrumental with respect to disenchantment, which is a final good with respect to seeing correctly, yet disenchantment is also instrumentally valuable with respect to dispassion. Furthermore, the application of the distinction between intrinsic and extrinsic values to the Pāli Nikāyas is not helpful in understanding early Buddhist values in their own terms. Given the non-substantial, relational ontology of early Buddhism, it would be awkward to say that some things are valuable by virtue of their intrinsic, non-relational properties, as if they were inherently valuable independently of everything else. Even the ethical and spiritual dimension of nirvana, that is, the mental qualities that constitute the state of nirvana, are conditioned in the sense of being dependently originated. Therefore, if all values including the ethical-spiritual qualities of nirvana are dependently originated, does not it follow that, at least to some extent, they all get their value from the multiplicity of factors that condition and contribute to their existence? In other words, strictly speaking, the distinction between what possesses value in virtue of intrinsic, non-relational properties versus what possesses value by virtue of extrinsic, relational properties does not apply to early Buddhist values. It might seem that from a less ontological level of discourse, one may apply the distinction intrinsic-extrinsic to the Pāli Nikāyas. Thus, at a conventional, common sense level of discourse, one may say that if the value of something does not derive from something else, then it has intrinsic value. On the other hand, if the value of something does derive from the value of something else, then it has extrinsic value. Accordingly, all Buddhist values except nirvanic values would be instances of extrinsic values: their value derives from their contribution to the achievement of nirvana. For instance, friendship would not be intrinsically valuable but rather extrinsi-

224

Abraham Vélez de Cea

cally valuable, because of its relationship to nirvana, that is, because it contributes to the attainment of nirvana. The problem with this application of the intrinsic-extrinsic distinction is that the Pāli Nikāyas do not seem to derive the value of non-nirvanic goods exclusively from their relationship or contribution to nirvana. Some early Buddhist goods appear to be valuable independently of nirvana. For instance, life is valuable regardless of its relationship or contribution to nirvana. The precept abstaining from taking life protects the value of life even in the case of non-human animals, which in early Buddhism are understood as incapable of volitional action (kamma), and therefore, as unable to progress toward nirvana. Similarly, the values of respect for spiritual status and fulfilling role-dependent duties do not appear to be dependent on their relationship or contribution to nirvana. I am not saying that these values cannot contribute to the attainment of nirvana. What I am suggesting is that in early Buddhism the values of life, respect for spiritual status, and fulfilling role-dependent duties do not derive only from their conduciveness to or even from other relationships to nirvana. If it is plausible to claim that at least some non-nirvanic values do not derive all their value from their contribution or relationship to nirvana, then they cannot be considered extrinsically valuable alone. In other words, some non-nirvanic values are intrinsically and extrinsically valuable simultaneously. Since some early Buddhist values can be both intrinsically and extrinsically valuable, we cannot apply to early Buddhism the dichotomy intrinsicextrinsic values. Despite their failure to illuminate the domain of Buddhist ethics fully, these distinctions are useful, in that they help us realize that early Buddhist values are complex, at least complex enough to be irreducible to the dichotomies intrinsic-instrumental, final-instrumental, intrinsic-extrinsic. Even though this might be the wrong set of distinctions to apply to the Pāli Nikāyas, they do suggest a model for approaching this rich axiological terrain. Following their lead, I propose an alternative distinction, that between ultimate and penultimate values.

Value pluralism in early Buddhist ethics

225

Ultimate values are the highest. Strictly speaking, for the Pāli Nikāyas only nirvana is ultimate, the end or final destination for the sake of which the entire spiritual path is cultivated (SN III.189; MN I.149–150). However, here by ultimate value I do not mean nirvana in its broadest sense but specifically the mental qualities and character traits constitutive of nirvana. For the sake of simplicity, I call these qualities and traits “nirvanic values.” I break the concept of nirvanic values into two umbrella terms, “nirvanic knowledge” and “nirvanic virtues.” Another possibility would be to explain nirvanic values in terms of more traditional Buddhist concepts such as mental factors (ceta­ sika), perfections (pāramitā), or requisites for enlightenment (bo­ dhipakkhiya-dhamma). However, I prefer the terms “knowledge” and “virtue” to facilitate the understanding of early Buddhist values by non-Buddhists, as well as comparisons of Buddhists and non-Buddhist values. I qualify the terms “knowledge” and “virtue” with the adjective “nirvanic” because not all types of knowledge and not all possible virtues are necessarily constitutive of or conducive to nirvana.18 All values except nirvanic values are penultimate. By penultimate values, I simply mean values not constitutive of nirvana: nonnirvanic pleasures, non-nirvanic knowledge, non-nirvanic virtues, friendship, fulfillment of role-dependent duties, life, health, repute, prosperity, and so on. Given that not all penultimate values are necessarily means toward nirvanic values, and since many intrinsically valuable virtues have also instrumental value, my distinction between ultimate and penultimate values should not be confused with the classical distinction between intrinsic and instrumental values. For instance, health is a penultimate value, and yet it cannot be reduced to a mere means for the attainment of nirvana. Likewise, the virtue of
The distinction between nirvana and nirvanic values might be seen by some as unnecessary. However, I would like to leave open the question of whether nirvana is something more than mental qualities and character traits. In fact, some texts of the Pāli Nikāyas seem to justify an interpretation of nirvana as more than just ethical and spiritual flourishing, that is, as the unconditioned dhamma, a transcendent base or sphere of reality (āyatana).
18

226

Abraham Vélez de Cea

love can both contribute and constitute many other ends other than Buddhist flourishing. For instance, love contributes and constitutes a good relationship; however, that does not make love a mere instrument for human relationships either. Whereas final values in Korsgaard’s sense are valued for their own sake, ultimate values can also be valued for the sake of something else. For instance, nirvanic virtues can be valuable also because they contribute to the well-being of other living beings. Similarly, while Korsgaard’s instrumental values are valuable for the sake of something else, penultimate values do not need to be always valued for the sake of something else. In fact, many penultimate values such as non-nirvanic knowledge and friendship can also be valued for their own sake. Besides understanding early Buddhist values through the concepts of ultimate and penultimate, I propose a critical appropriation of Jonathan Dancy’s distinction between favoring, enabling, and intensifying conditions.19 Unlike Dancy, I use the distinction to explain the functions performed by values. Values function as enablers if they make other values possible; as favorers if they increase the chances of achieving or implementing other values; and as intensifiers if they improve or supplement the value of other goods, for instance, by making them more attractive or enticing. I interpret the functions of enabling, favoring, and intensifying as cutting across ultimate and penultimate values. That is, the functions of favoring, enabling and intensifying can be performed by both achievements valued ultimately or penultimately. The three functions cannot be confused with a mere instrumental function, that is, the function of being a means to an end. As I said, the dichotomies intrinsic-instrumental and final-instrumental might be
19 Dancy 2004. Despite the fact that Dancy’s distinction belongs to his theory of reasons, I think it possible to apply them to the realm of values. In fact, Dancy himself admits the possibility of applying his concept of enabling to his theory of value, though not his concept of favoring. By using the terms enabling, favoring, and intensifying to clarify the lexical order of early Buddhist values, one does not have to share the meaning that Dancy gives to such terms, nor endorse his holistic theory of reasons, nor agree with his moral particularism.

Value pluralism in early Buddhist ethics

227

misleading and unhelpful in appreciating the complexity of some Buddhist values. Accordingly, instead of speaking of instrumental function20 or “being a means to an end,” I prefer to speak more specifically of enabling, favoring, and intensifying functions, which are not the monopoly of either ultimate or penultimate values. When I assign one of these three functions to something we value – a trait or an achievement – I do not want to insinuate that it is the only function that such value can perform. For instance, mental health primarily performs an enabling function but it can also perform a favoring function. Similarly, spiritual pleasures primarily performs an intensifying function, but it can also perform a favoring function in that it may encourage further spiritual practice. Moreover, such valuable traits as repute and influence primarily perform an intensifying function, but they can also perform a favoring function, facilitating the practice of certain virtues. For instance, the fame and the spiritual influence of a holy person can facilitate the practice of generosity and compassion among her/his disciples.

20 Although the application of the concept of instrumental value to the Pāli Nikāyas is problematic, I do advocate a distinction between “teleological” and “instrumental” actions, which correlates to some extent to the early Buddhist distinction between kusala and puñña, and, in a different way, to W.D. Ross’ distinction between the goodness and the rightness of actions. Whereas teleological actions are necessarily performed with a wholesome motivation, instrumental actions do not necessarily presuppose a wholesome motivation, though they are nevertheless morally acceptable and karmically fruitful. From a Buddhist perspective, instrumental actions are right actions, stepping stones toward the eventual performance of teleological actions. I do not deny that actions may be simultaneously good and right, teleological and instrumental, at least in ideal types of ethical practice. However, the goodness and rightness of actions do not always overlap, and therefore, we cannot always reduce teleological and instrumental actions to two different aspects of a single action; at least this is not the case in actions where goodness and rightness do not coincide. For instance, in less ideal types of ethical practice it is possible to do the right thing without a wholesome motivation, or to do what is wrong with a wholesome motivation. Thus, I neither conflate nor fracture the rightness and goodness of actions, and accordingly, I neither identify nor totally separate teleological and instrumental actions.

228

Abraham Vélez de Cea

Life and health primarily perform an enabling function. Values that primarily favor the achievement or implementation of other values can be divided into at least three groups: material, social, and ethical-spiritual. The four requirements of clothing, shelter, food, and medicine or health care, are material values that facilitate other values, for instance, the cultivation of spiritual values. The social values that primarily facilitate other values are right livelihood, friendship, respect for spiritual and social status, and fulfillment of role-dependent duties. Social values may also perform an intensifying function in that they make life in society more enjoyable. Ethical-spiritual values favor the development of material values and enhance the value of social goods. Pleasure, supernatural powers, and worldly benefits such as welth, corporeal beauty, influence, honors, and repute primarily perform an intensifying function. In early Buddhism, pleasure, supernatural powers, and worldly benefits are valued because they enhance the value of other values. For instance, by saying that generosity leads to future prosperity, one intensifies the value of giving. Similarly, the pleasure of nirvana – the highest kind of pleasure (Dhp 203; MN I.508) – intensifies the value of mental qualities constitutive of nirvana. Furthermore, the values of wisdom and certain meditative attainments are intensified by the supernatural abilities they may generate. Intensifying values can also perform enabling or favoring functions. For instance, prosperity enables one to create Dharma centers and thus, favors the cultivation of spiritual practice in oneself and others. Similarly, the pleasure of the first jhāna is praised because it is the path leading to enlightenment, that is, because it facilitates the attainment of nirvana (MN I.247). Someone might object that pleasure, supernatural powers, and worldly benefits are not genuine values but rather side-effects of ultimate values. For instance, health, wealth, beauty and influence are often described as karmic consequences of previous generosity and ethical conduct. Similarly, supernatural powers and spiritual pleasures usually derive from the development of wholesome mental states through meditation practice. Consequently, the objection goes, only ultimate values and not their side-effects should be seen

Value pluralism in early Buddhist ethics

229

as true values. However, the fact that pleasure, supernatural powers, and worldly benefits are consequences of ultimate values does not imply that they possess no value at all, even less that their value is reducible to the value of nirvanic virtues and nirvanic knowledge. The fact that the Pāli Nikāyas use pleasure, supernatural powers, and worldly benefits such as health, wealth, beauty, and influence to motivate ethical and spiritual practice indicates that they provide an incentive that ultimate values cannot provide by themselves, at least for certain kind of practitioners. Such an incentive might not be the best possible motivation, but it is nevertheless valuable for attracting some people to the spiritual path, people who otherwise might not be interested in ultimate values by themselves. Another possible objection is that pleasure, supernatural powers, and worldly benefits are valuable only when they are handled well, that is, if they are enjoyed virtuously. For instance, the objection goes, many texts seem to demonstrate that pleasures associated with unwholesome mental states such as craving and grasping, are not valuable at all. Therefore, their value would be reducible to that of spiritual values. This, however, is true of all Buddhist values except nirvana, not simply those that are primarily intensifying; they all lose worth when associated with unwholesome mental states. All values are more valuable when accompanied by wholesome mental states, but that does not mean that they have no value whatsoever without these states. Even the pleasure of sublime spiritual attainments can become counterproductive for spiritual progress if one develops clinging (upādāna) or ego-conceit (asmimāna) towards them (MN II.264–265; MN II.237). This does not imply that the pleasure of spiritual attainments lacks any value whatsoever when clinging and ego-conceit are present. In fact, MN II.265 speaks about someone who clings to the best object of clinging, namely the meditative base of neither-perception-nor-non-perception. If the value of this pleasant meditative base were dependent exclusively on wholesome mental states, it would make no sense to consider it the best object of clinging (upādānaseṭṭhaṃ). If the objection were plausible, the text should say that the base has no value at all due to the presence of clinging. However, this is not what the text states.

230

Abraham Vélez de Cea

From the early Buddhist perspective, any pleasures, especially sensual pleasures, can become counterproductive for spiritual progress. This misgiving, however, has nothing to do with their value as such, but rather with what early Buddhists call their danger (ādīnava). The Pāli Nikāyas are ambivalent about pleasures since they entail both value and danger. Spiritual pleasures are ranked higher than sensual pleasures, and, overall, they are considered extremely valuable: “the pleasure of renunciation, the pleasure of seclusion, the pleasure of peace, the pleasure of enlightenment. I say of this kind of pleasure that it should be pursued, that it should be developed, that it should be cultivated, that it should not be feared” (MN I.454).21 Strictly speaking, any value, not just pleasures, may hold danger for the unenlightened mind, but this does not imply that their value is dependent on the presence of wholesome mental states and, therefore, that their value is reducible to nirvanic values. Since the Pāli Nikāyas do not reduce all these different kinds of values to an overarching super-value or single good, early Buddhist ethics is pluralistic in terms of values. Moreover, since the Pāli Nikāyas presuppose a hierarchy of values (ultimate and penultimate, nirvanic and non-nirvanic), the evaluative pluralism of early Buddhist ethics is lexically ordered. It is to that lexical ordering that we now turn.

3. The lexical order of early Buddhist values
In order to clarify the lexical order of early Buddhist values, I would like to compare my account to Damien Keown’s list of three basic goods. So far, I have contended that in the Pāli Nikāyas spiritual values surpass non-spiritual values, that non-spiritual values are genuine values irreducible to spiritual ones, and that given the complexity of some early Buddhist values, they cannot be adequately categorized in terms of the dichotomies intrinsic-instrumental, final-instrumental, intrinsic-extrinsic. Instead, I have proposed the distinction between ultimate and penultimate values, and claimed
nekkhammasukhaṃ pavivekasukhaṃ upasamasukhaṃ sambodhasu­ khaṃ āsevitabbaṃ bhāvetabbaṃ bahulīkātabbaṃ. Na bhāyitabbaṃ etassa sukhassāti vadāmi.
21

Value pluralism in early Buddhist ethics

231

that values may perform enabling, favoring and intensifying functions, which are not necessarily reducible to just being a means to an end. In Buddhism and Bioethics, Damien Keown speaks about three fundamental values or basic goods. These goods are life, knowledge, and friendship. By knowledge, Keown means wisdom (pañña, Skt. prajña), “which has as its object the truths of Buddhist doctrine.”22 By friendship, he does not refer to spiritual friendship (kalyāṇamitta) but rather to “a wider complex of ideas…the proper mode of relationship with others…the complex which in early Buddhism is labeled as Morality (sīla), and in the Mahāyāna as Compassion (karuṇā) or alternatively as Means (upāya).”23 By life, he does not mean all kinds of life but specifically what he calls “karmic life.” That is, life that “possesses the capacity to attain nirvana.”24 Keown’s two basic goods of friendship and knowledge overlap to a great extent with what I call ultimate values: nirvanic virtues and nirvanic knowledge. Unlike Keown, I prefer to differentiate clearly between friendship, which in my account is a penultimate value, and friendliness or loving-kindness (mettā), which is an ultimate value or nirvanic virtue. Likewise, I would like to keep separated the value non-nirvanic knowledge, which is penultimate, and nirvanic knowledge, which is ultimate. This distinction is grounded in the Pāli Nikāyas, which distinguish between diverse kinds of knowledge, not all of them equally valuable. Similarly, the Pāli Nikāyas value friendship but never identify it with a particular nirvanic virtue, let alone with morality in the sense of relationships with others. I do not deny that morality regulates our relationships with others, yet morality does much more than that. At least in the Pāli Nikāyas, morality also regulates internal behavior (mental states) and actions that do not necessarily relate to others. For instance, taking intoxicants and engaging in

22 23 24

Keown 1995: 50. Ibid., 43–44. Ibid., 46.

232

Abraham Vélez de Cea

sexual misconduct may take place without necessarily involving a relationship with others. According to Keown, the three goods of life, knowledge, and friendship are basic in the sense of being irreducible to each other; they are also incommensurable: “it is impossible to quantify these things and trade them off against one another as if they could be related on a common scale.” Consequently, for Keown, these goods must be equally valuable: “None of these things can stands as ‘greater’ in relation to another which is ‘lesser’.”25 I agree with Keown’s claim about the irreducible nature of early Buddhist goods and values. I also concur with him when he says that Buddhist values are incommensurable. Nevertheless, from the fact that goods are incommensurable, it does not follow that they must be equally valuable. We must distinguish between commensurability and comparability. Early Buddhist goods are incommensurable but not necessarily incomparable. By saying that values are comparable, I do not mean that they are quantifiable or subject to utilitarian calculations. Rather, the point is that values can be lexically ordered to some extent, and, therefore, they must be somehow comparable. The lexical order of values in the Pāli Nikāyas is undeniable: ultimate values are superior to penultimate values. This hierarchy of values would not exist if values were not comparable. Yet early Buddhist values are incommensurable. For instance, no amount of prosperity equals any specific number of lives. Similarly, it does not make much sense to state that fostering non-nirvanic knowledge in 300 people equals the development of nirvanic wisdom in one person. This way of thinking is foreign to the Pāli Nikāyas. Nevertheless, the value of life seems to surpass the value of material benefits, and the value of nirvanic wisdom is higher than the value of non-nirvanic types of knowledge. The lexical order of values found in the Pāli Nikāyas presupposes not only the comparability of values from different kinds but also of values that belong to the same kind. For instance, the value of spiritual pleasures is superior to the value of sensual pleasures, which are often negatively portrayed (MN I.132ff., MN I.173ff.,
25

Ibid., 55–56.

Value pluralism in early Buddhist ethics

233

etc.). Similarly, human life is more valuable than other types of lives, and the life of holy beings is more worthy than the life of ordinary beings; that is why the karmic consequences of killing a buddha are far more negative than killing any other ordinary being. Even among wholesome mental states, it is possible to perceive some lexical order. For instance, wisdom seems to be the most valuable mental state, even more than loving-kindness and compassion, though according to Gombrich the Pāli Nikāyas are not always consistent in this regard.26 Nonetheless, despite some tensions and minor variations, the Pāli Nikāyas presuppose a lexical order of values. Even if the ranking among values of the same kind is not beyond dispute, it is clear that ultimate values are more important than penultimate ones, and that spiritual benefits surpass non-spiritual benefits. I do not claim that my discussion of early Buddhist values is comprehensive or that my list of values exhausts all things that early Buddhists considered valuable. Like Keown, I admit that there may be other ways of mapping the most important Buddhist values. It would be unfair to question Keown’s list of three basic goods for not being long enough. However, I find some of Keown’s ideas problematic from the early Buddhist perspective, specifically his understanding of friendship and life. Keown’s view of life as belonging to the same category of values as knowledge and friendship cannot be justified on the basis of any early Buddhist texts. Nowhere in the Pāli Nikāyas is the value of life considered similar to the spiritual values of nirvanic wisdom and nirvanic virtue. Similarly, the Pāli Nikāyas do not suggest anywhere that life is constitutive of nirvana beyond death. Keown’s claim that life in nirvana “will take some form, since Buddhist doctrine condemns as heresy the view that nirvana is annihilation,”27 is textually unjustified. I have argued somewhere else that the Pāli Nikāyas are silent about the tathāgata and nirvana beyond death.28

26 27 28

Gombrich 1996: 60ff. Ibid., 49. Vélez de Cea 2004b.

234

Abraham Vélez de Cea

Keown’s suggestion about some form of life in nirvana beyond death is also questionable on philosophical grounds. The Pāli Nikāyas consider eternalism an extreme doctrinal position; it is as extreme as annihilationism. Consequently, Keown’s reasoning to infer the existence of some form of eternal life in nirvana from the rejection of annihilationism is unjustified. One could use exactly the same reasoning to argue that Buddhism is nihilistic because it rejects eternalism. This conclusion, however, would be equally inconsistent with the philosophy of the Pāli Nikāyas.29 Keown seems to be influenced in his interpretation of life and friendship by contemporary neo-Thomist thinkers such as John Finnis and Germain Grisez. For instance, John Finnis lists life and friendship as basic goods and consider them constitutive of the human good. Like Keown, Finnis considers all basic goods incommensurable and equally valuable. Finnis’ concept of friendship is strikingly similar to that of Keown in that both convey the idea of relationships with others and concern for their well-being. In fact, Finnis uses the term sociability as a term equivalent to his understanding of friendship.30 Another possible reason for Keown’s inclusion of life in his list of basic goods is his application to Buddhism of the classical dichotomy between intrinsic and instrumental values, that is, the distinction between what is valued for its own sake as an end, and what is valued for the sake of something else as a means. For instance, Keown distinguishes between the intrinsically valuable life of humans and animals, and the instrumentally valuable life of tiny organisms, vegetables and plant life.31 Since the classical dichotomy intrinsic-instrumental only allows for two options, and since human life is not a mere means or instrument for something else, it is only natural for Keown to place life at the same level as other intrinsic values. However, I have demonSpecifically, inconsistent with the limits that the Buddha of the Pāli Nikāyas puts to language and his teachings (Vélez de Cea 2004b). See also Vélez de Cea 2008. 30 Finnis 1980: 86–88, 141–144. 31 Keown 1995: 46–49.
29

Value pluralism in early Buddhist ethics

235

strated the inappropriateness of applying the dichotomy intrinsicinstrumental to the Pāli Nikāyas. This distinction fails to capture the diversity of early Buddhist values. Once we shift to the alternative hermeneutical strategy of applying the distinction between ultimate and penultimate values, we gain far greater clarity, and this approache undermines the motivation for Keown’s inclusion of life as a basic Buddhist good. In my account, life in the Pāli Nikāyas is a penultimate value that primarily enables or makes possible other values. In the case of humans, life also favors the cultivation of spiritual values, thus intensifying the value of human rebirth. Keown would probably agree with my account because he says explicitly at one point that life is “both a good in itself and a precondition for the fulfillment of other goods.”32 This is precisely what I claim only with a different terminology: that the value of life is not reducible to the value of anything else, and that life performs primarily an enabling function. I prefer to avoid the expression “in itself” because it is usually associated to the concept of intrinsic value, which is ontologically misleading when applied to Buddhism. Keown and I agree that life is not a mere instrument in the restricted sense of the term, that is, as being a mere means without value “in itself.” Attributing irreducible and enabling value to life, does not render it a simple means whose value depends entirely on the performance of a particular function or on being a condition of possibility of other values. And denying that life is a mere means to an end or a mere instrument for the performance of a function, does not entail that the value of life is supreme, at the same level as nirvanic knowledge and nirvanic virtue. Either claim would be foreign to the Pāli Nikāyas. Like the value of life, the value of nirvanic knowledge and virtue is irreducible to other values; they are also similar in that they all can perform enabling, favoring, and intensifying functions. However, unlike life, nirvanic values belong to a higher order of values. The profound respect for life found in the Pāli Nikāyas can

32

Ibid., 44.

236

Abraham Vélez de Cea

be explained without suggesting, as Keown does, that life is an ultimate value constitutive of nirvana. Perhaps the greatest difference between Keown’s account and mine is that unlike his, my account does not view nirvanic values as the only kind of value constitutive of Buddhist flourishing, which I do not identify with enlightenment or human fulfillment from a Buddhist perspective. On the contrary, for Keown: “To say that life, knowledge and friendship are good is to say that these are the things which make for a fulfilled life as a human being. They are fundamental aspects of human fulfillment or flourishing in that each makes a unique contribution to the nature of the being one wishes to become (a Buddha).”33 Thus, Keown seems to identify human fulfillment with flourishing, and these two concepts with the values of life, knowledge and friendship. My account, however, presupposes a difference between Buddhist flourishing and fulfillment: Buddhist flourishing can take place without being enlightened, and it can be spiritual and/or nonspiritual in nature; on the contrary, Buddhist fulfillment is the state of enlightenment, which is constituted only by nirvanic knowledge and virtues. Nevertheless, like Keown, I view spiritual flourishing and nirvanic qualities as the most important sources of value in Buddhism.

Primary sources
All references to the Pāli texts are to the roman-script editions of the Pali Text Society, England. References to the Aṅguttara-, Dīgha-, Majjhima- and Saṃyutta-Nikāya are to the volume and page number.
AN DN Dhp Aṅguttara-Nikāya, ed. R. Morris, E. Hardy, 5 vols. London 1885–1900. Dīgha-Nikāya, ed. T. W. Rhys Davids, J. E. Carpenter, 3 vols. 1890–1911. Dhammapada, ed. O. von Hinüber, K. R. Norman, 1994.

33

Ibid., 43.

Value pluralism in early Buddhist ethics MN SN

237

Majjhima-Nikāya, ed. V. Trenckner, R. Chalmers, 3 vols. London 1888–1899. Saṃyutta-Nikāya, ed. L. Feer, 5 vols. London 1884-1898.

Secondary sources
Dancy, Jonathan. 2004. Ethics Without Principles. Oxford: Oxford University Press. Finnis, John. 1980. Natural Law and Natural Rights. Oxford: Oxford University Press. Gombrich, Richard. 1992. “The Ethic of Intention.” In Charles Prebish, Buddhist Ethics: A Cross-Cultural Approach. Dubuque, Iowa: Kendall/ Hunt Publishing Company: 92–111. _____. 1996. How Buddhism Began: The Conditioned Genesis of the Early Teachings. London: Athlone. Harvey, Peter. 2000. An Introduction to Buddhist Ethics. Cambridge: Cambridge University Press. Hurka, Thomas. 2001. Virtue, Vice, and Value. Oxford: Oxford University Press. Keown, Damien. 1995. Buddhism and Bioethics. London: Macmillan. _____. 2001. The Nature of Buddhist Ethics. New York: Palgrave. Korsgaard, Christinine. 1996. The Sources of Normativity. Cambridge: Cambridge University Press. _____. 1982. “Two Distinctions in Goodness.” The Philosophical Review 2: 169–195. Rønnow-Rasmussen, Toni. 2002. “Instrumental values – strong and weak.” Ethical Theory and Moral Practice 5: 23–43. Vélez de Cea, Abraham. 2004a. “The Early Buddhist Criteria of Goodness and the Nature of Buddhist Ethics.” Journal of Buddhist Ethics 11: 123– 142. _____. 2004b. “The Silence of the Buddha and the Questions about the Tathāgata after Death.” The Indian International Journal of Buddhist Studies 5: 119–141. _____. 2008. “Buddha,” Internet Encyclopedia of Philosophy. http://www. iep.utm.edu/buddha/, last visited 26-09-2011 [last update 2009]. Wolf, Susan. 1992. “Two Levels of Pluralism.” Ethics 102 (July): 785–798.

No self, no free will, no problem Implications of the Anattalakkhaṇa Sutta for a perennial philosophical issue
Martin T. Adam

The free will problem
Are we free agents? The free will problem remains one of the great ongoing debates of western philosophy. This paper investigates the Buddha’s views on human freedom. It suggests that the Buddha’s position is a unique one, implying a negative response to the question of a metaphysically free will but a positive response to the question of moral responsibility and the possibility of human freedom in a spiritual sense. The problem of free will in its most general terms can be formulated as follows. All events are caused. A full understanding of the causes of any particular event and of the laws of nature would allow for the accurate prediction of that event. The actions we perform, including the choices we make, are events. Therefore they are all predictable in principle, if not in fact. Therefore the idea that one can do other than one actually does is false. If one cannot do other than one actually does, one cannot be morally responsible for one’s actions. Therefore human beings cannot be justifiably held morally responsible for their actions.1
This is a modified version of the argument presented by Van Inwagen (1982), who frames the issue in terms of knowledge of the state of the physical world and the laws of physics. Here I have generalized the formulation to exclude such ontological considerations. Strictly speaking they are irrelevant to the deterministic thesis, which depends only on the notion that all events have causes, past and present, sufficient to bring them about (irrespective of Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 239–265
1

240

Martin T. Adam

A common counterargument to this line of thinking is as follows. While it is true that all events are caused, human actions are a special case in that among their causes is the individual’s own will or choice. Because individuals are the causes of their own actions they can be justifiably held morally responsible for them. So free will and determinism are compatible. This is the ‘compatibilist’ position: within a deterministic universe we are nevertheless free in a sense sufficient to justify holding one another morally to account. We do, in fact, hold one another morally responsible when certain specified conditions of freedom are met; a failure to acknowledge this would be to empty the notion of freedom of all meaning. There must be some instances in which we are free. And these instances, by and large, are precisely those in which we recognize each other as being morally responsible.2 Thus the compatibilist argues from the fact of moral responsibility to the assertion of freedom and the irrelevance of the deterministic thesis. The incompatibilist determinist on the other hand, argues from determinism to a lack of freedom and the irrelevance of conventions of moral accountability. Both positions are persuasive; our intuitions seem torn. Before examining what Buddhist teachings might have to offer in connection with this dilemma, it will be helpful to explicitly set out, in brief, its key terms of reference. In discussing the concept of freedom, one of the most basic distinctions philosophers
whether they are physical, mental, spiritual or otherwise). In recent years the argument has been framed in scientific terms: the notion of an action being ‘predictable in principle’ is understood in terms of an imagined third-person observer possessing a complete knowledge of the state of the world at a given point in time and a complete knowledge of the laws of nature. The fact that no one possesses such complete knowledge does not affect the argument, for what we are concerned with is predictability in principle. The classical problem of free will was, of course, framed in terms of God’s foreknowledge: if God knows everything that will happen ahead of time, then everything must happen exactly as God knows it will happen. In spite of the presentday replacement of God by an imaginary super-scientific observer, the basic argument is the same. 2 Clear cases would be those in which one is rational, knows what one is doing, has reflected on the alternatives, is not being coerced, and approves of the action without doubt or hesitation.

No self, no free will, no problem

241

have drawn is that between empirical and metaphysical freedom. Empirical freedom can be understood in a number of ways, but in its most general positive conception it refers to the ability of an individual to act as she wants, or to do as she wills. Philosophers have, of course, offered varying accounts of what such a positively conceived freedom actually means once it is spelled out in detail − for example politically, in term of specific rights and freedoms (‘freedoms to’ do one thing or another). It is, however, beyond the scope of the present study to explore such differences; for our purposes it is enough to note the general concept. We must also observe that the concept of empirical freedom can be formulated negatively, in terms of an absence of constraints or impediments that might potentially obstruct an individual’s ability to act as she wants (‘freedoms from’ one condition or another). In this case clarity would demand that a particular set or sets of constraints be specified. Here we should note that the notion of a ‘constraint’ may be conceived of as either external or internal to a person. Philosophers have considered various sets of external and internal constraints in spelling out their own particular understandings of freedom. Political philosophers, for example, have tended to focus on restrictions placed upon the individual by external forces such as other persons, governments, political classes, or even material conditions. Psychologically minded thinkers, on the other hand, have focused on internal constraints such as compulsions, obsessive thoughts, depression, confusion and so forth. But for philosophers working in the area of metaethics, concerned as they are with establishing foundations for our judgments of moral responsibility, it has usually been the idea of metaphysical freedom that has been called upon to do the work. Metaphysical freedom, like empirical freedom, can be conceived negatively as an absence of constraints. But in this case, the constraints are regarded as those of causality itself. Moral responsibility is thought to require some kind of freedom from, or exception to the necessity that characterizes the normal cause and effect operations of nature. Attaching a clear meaning to this idea of freedom is problematic, but in general either one of two basic approaches is taken. On the one hand, to say that a person is metaphysically free is interpreted

242

Martin T. Adam

as asserting that at least some of her actions or decisions are un­ caused.3 On the other hand, to make this statement is taken as asserting that at least some of her actions or decisions are self­caused. One final distinction needs to be recognized, even though it is not one that is often explicitly drawn. A broad distinction can be drawn among the kinds of object to which the concept of freedom is applied. For our purposes we can enumerate three: actions, wills, and persons (or individuals).4 Freedom of the will is sometimes equated with freedom of action and sometimes with freedom of the person. In fact, authors often slide between these three ways of speaking about freedom, assuming that to talk of one is to talk of the others, and that the predication of freedom in one these categories ipso facto implies a statement of the same truth value in the others. But as we shall see, this assumption is worth questioning. At a minimum it must be ensured that the same set of constraints in terms of which freedom is understood is being applied across categories. Further, each of these categories themselves admit of various conceptions and therefore require careful analyses. In this paper I will argue that the Buddha’s teachings do not allow for the possibility of a metaphysical freedom of the will. Nevertheless persons are regarded as morally responsible for their actions. In order to explain this view I will examine the implications of the Anattalakkhaṇa Sutta, comparing these with an influential account of free will provided by the western philosopher, Harry Frankfurt. The overall discussion is framed with reference to Harvey (2007), which is probably the most comprehensive individual survey of Theravāda Buddhist teachings on this topic to

We shall not explore this possibility in great detail in this paper. It has commonly been observed that such indeterminism does not appear to be of much use when it comes to the grounding of moral responsibility. If a decision or action were uncaused, would it not then be random? How can a person be justifiably held responsible for a random decision or act? Such a conception of freedom would actually seem to undermine the foundations of moral responsibility. 4 A fourth possibility can here only be mentioned in passing, namely, freedom of a group or society.

3

No self, no free will, no problem

243

date. This is where we begin our own exploration of Buddhism and free will.

Buddhism and free will
Harvey’s review of the primary and secondary sources on this topic is extensive and it cannot be my aim here to provide a detailed critique of the account he provides. The conclusions he derives are complex, but by and large fall into two parts. Initially he concludes that Buddhism accepts a form of compatibilism.5
“On the whole, it can be said that the implied position of Theravāda Buddhism on the issue of ‘freedom of the will’ is a middle way between seeing a person’s actions as completely rigidly determined, and seeing them as totally and unconditionally free […] It accepts a variable degree of freedom within a complex of interacting mental and physical conditions. This freedom of action is such that present awareness always offers the possibility of not being wholly determined by past patterns of internal or external conditioning […]” (Harvey 2007: 86)

Nevertheless he also maintains that there is a second sense in which the Buddha’s teachings imply that neither free will nor determinism can be true:
“In a different way […] if a person is wrongly seen as an essential, permanent self, it is an ‘undetermined question’ as to whether ‘a person’s acts of will are determined’ or ‘a person’s acts of will are free.’ If there is no essential person-entity ‘it’ can not be said to be either determined or free.” (Harvey 2007: 86)

Harvey’s answer is difficult conceptually. In this paper I will mainly take issue with its second part: I will argue that if there is no essential person-entity, the implication is not that the will is neither free

This would also appear to be the view of Thanissaro Bhikkhu who speaks of “some room for free will” in Buddhism (1996: 13). This author also provides a similar rationale for the attribution of freedom, in terms of the reflective capacity of present awareness − or what he calls “feedback loops” (1996: 40−42). This is indeed an important factor in understanding the human capacity for freedom. We will return to it below.

5

244

Martin T. Adam

nor determined, but rather that there is no metaphysical freedom of the will, which is to say no will that is free of ordinary causality.6 At least two general points arise from Harvey’s conclusions. First, with regard to its first part, it is important to note the qualification that Theravāda Buddhism has an ‘implied’ position on the free will debate. This needs to be emphasized. While I differ with Harvey as to what that implied position is, he is undoubtedly correct in this. The problem of moral responsibility in relation to the deterministic thesis has never been a burning issue in traditional Buddhism.7 This is not to suggest that the issue is not important to Buddhists today, nor is it to say that the question of human freedom was not important in early Buddhism. It is simply to suggest that the problem presented to the foundations of morality by determinism is not a ‘live’ problem in the Buddha’s teachings.8 There are reasons for this, which we shall examine below.9 The second point relates to the first part of Harvey’s conclusion as well. In what sense are we to understand the notion of a
The impulse to read Buddhism as a form of compatibilism possibly stems from the assumption that moral responsibility must be regarded as contingent upon some kind of metaphysical freedom of the will. (Note, however, that Harvey does not himself make this assumption). In the present paper I argue that this is not the case for the Buddhist tradition at least. 7 We shall see however that there may have been a peripheral awareness of this tension on the part of the compilers of the Pali Canon. 8 While the Buddha was concerned, on occasion, with the refutation of fatalism, this cannot be equated with a rejection of determinism. Oddly Harvey seems to make this equation at one point, in his discussion of the Buddha’s response to the Ājīvikas (2007: 40). Fatalism is the view that individuals do not have control over what happens to them, in the sense that certain events will happen to them irrespective of what they decide to do. In some versions of fatalism this is attributed to the design of an outside power or Fate. Determinism is simply the view that every event has causes sufficient to bring it about; our choices are both causes of the events that occur to us as well as caused events themselves. Thus one can consistently deny that there is any external force of destiny, such as the Ājīvikas’ niyati, and still be a determinist. 9 The explanation, we shall see, lies in the fact that freedom was principally understood as a quality of persons, dependent on their knowledge and mental purity, rather than as a quality of volitions or actions.
6

No self, no free will, no problem

245

Buddhist “middle way” in this context? Harvey seems to suggest that Buddhism adheres to a compromise position, i.e. one that lies “between” strict determinism and complete freedom from causality (or as he puts it “between seeing a person’s actions as completely rigidly determined, and seeing them as totally and unconditionally free”). But how can this be so? Surely, the two positions are mutually exclusive. And what might it mean to say that persons possess a “variable degree” of freedom? Is freedom the kind of quality that admits of degrees? This needs explanation. It would appear that this middle way solution runs the risk of incoherence (a charge often leveled at compatibilist accounts). It may be that each of Harvey’s statements is correct, and that in point of fact the Buddha’s various teachings do imply these rather different positions at different places in the scriptures. If so, one might be inclined to conclude that the Buddha of the Pali scriptures is simply inconsistent and that his teachings on freedom constitute a colourful, but philosophically unprofitable area of investigation. More charitably, one might wish to suggest that the Buddhist concept of freedom is one of religious or mystical paradox, beyond rational comprehension. Here, however, I will argue that there is a systematic and consistent rationale underlying the Buddha’s teachings concerning freedom. The concept of freedom plays a critical role within what is a highly sophisticated soteriological system. The Buddha’s implied position on freedom of the will cannot be properly understood outside the confines of this framework.

Free will and the Anattalakkhaṇa Sutta
To investigate these matters, I will now turn to an examination of one of the most famous of the Buddha’s discourses, the Anatta­ lakkhaṇa Sutta.10 This is the Buddha’s second sermon, delivered to his first five disciples at the Deer Park in Sārnāth. In this sutta the Buddha systematically argues for the impossibility of identifying a Self with any of the five aggregates that together constitute a person. While the Anattalakkhaṇa Sutta is not normally considered
For another interpretation of this sutta in relation to the issue of free will, see Federman 2010.
10

246

Martin T. Adam

as addressing the matter of free will, the teachings it contains do have implications that bear on this topic. The Buddha offers two distinct arguments to support the thesis that no aspect of a person is a wholly autonomous, essential, permanent Self; the first of these is directly relevant to our present concerns. In this argument, the Buddha suggests that none among the five aggregates that together constitute a person can be identified with such a Self because none among them is subject to control.11 Beginning with the body or form (rūpa) the Buddha makes his case:
“Bhikkhus, form is nonself. For if, bhikkhus, form were self, this form would not lead to affliction, and it would be possible to have it of form: ‘Let my form be thus; let my form not be thus.’ But because form is nonself, form leads to affliction, and it is not possible to have it of form: ‘Let my form be thus; let my form not be thus’.”12 (SN III 66)

A similar line of reasoning is offered for each of the five aggregates. To appreciate the implications of this argument for the question of free will, we need to see that the Buddha is relying on a conceptual connection between the notion of Self and the notion of control. If there were a Self, he asserts, it would be that aspect of the person over which one has control.13 We do not have control over any of
11 The second argument reasons from the impermanence of each of the aggregates to the fact that they are each dukkha. From these two considerations the conclusion is drawn that none among the aggregates are fit to be regarded as ‘Self.’ 12 Rūpam bhikkhave anattā // rūpañ ca bhikkhave attā abhavissa nayidaṃ rūpaṃ ābādhāya saṃvatteyya // labbhetha ca rūpe Evaṃ me rūpaṃ hotu evaṃ me rūpaṃ mā ahosīti // Yasmā ca kho bhikkhave rūpaṃ anattā tasmā rūpam ābādhāya saṃvattati // na ca labbhati rūpe Evam me rūpaṃ hotu evaṃ me rūpaṃ mā ahosīti // // Translations are those of Bhikkhu Bodhi (2000). 13 In a paper delivered to the annual meeting of the UKABS (2009) Tse-fu Kuan has argued that this way of conceiving the self as “what comes under control” constitutes a deliberate “twisting” of the classical Brahmanic concept of the self as “inner controller” found in in the Upaniṣāds. His argument is principally focused on a comparison of Majjhima-Nikāya Sutta 35 and two Chinese versions of this text, where the same argument is found. According to the author, an earlier version of the argument, employing the concept of Self as inner controller, can be found in Chinese Ekottarika-āgama version.

No self, no free will, no problem

247

the five aggregates. The five aggregates are all that a person is. The implication is clear: there is no self.14 In his notes on this sutta, Bhikkhu Bodhi makes some insightful observations about the basis of this argument. The five aggregates’ lack of selfhood is demonstrated, he says:
“on the ground that they are insusceptible to the exercise of mastery (avassavattitā). If anything is to count as our ‘self’ it must be subject to our volitional control; since, however, we cannot bend the five aggregates to our will, they are all subject to affliction and therefore cannot be our self.” (Bodhi 2000: 1066-1067)15

From these comments we can see how the Buddha’s argument can be connected to the issue of free will − there is a conceptual link

While I cannot evaluate this argument here, the need to establish an explanation for the apparent oddity of conceiving the Self as “what comes under control” would appear to be eliminated if one recalls that the fourth aggregate, the saṅkhāras, encompasses the function of directing or controlling one’s actions (cetanā). Thus the phrase “what comes under control” includes the “inner controller.” This very function, the only candidate for inner controller that there is, is itself not subject to control. See below. Kuan’s paper has recently been published (2009). 14 The initial argument structure is Modus Tollens and is formally valid. If there were Self, it would be controllable. None among the five aggregates are controllable. Therefore none among the five aggregates are Self. There is an additional, unstated assumption required to reach the conclusion that there is no Self, viz., that the five aggregates are all that there is. This is a safe enough assumption from the Buddhist perspective. The complete argument also involves one other assumption, viz., if we could control our states we would choose those that are not afflicted by suffering and its causes. This seems a safe enough assumption, for all but the masochist. I bring it up because it reveals an important teleological aspect to the Buddha’s thinking, which is not argued for here: we are naturally oriented away from suffering and towards happiness. In point of fact it is this aspect of our constitutions that makes it possible to attain spiritual freedom, the realization of nibbāna. The general point that we are naturally predisposed towards wanting to be free from suffering is explicitly taken up in the following sutta, Mahāli (SN III 68−71). We will return to it below. 15 Elsewhere he writes that “[…] the aggregates are suffering because they tend to affliction and cannot be made to conform to our desires.” (Bodhi 2000: 842)

248

Martin T. Adam

between the idea of ‘Self’ and the idea of ‘volitional control.’ If there were a Self, whatever else it might be, we would be able to control its states. Thus in the above passage, concerning rūpa, the idea is that we would all choose not to suffer and to be well in our bodies if we could; indeed this is our natural wish and predisposition. In spite of this, we remain afflicted and disposed to affliction. Suffering is inherent to rūpa. It is not possible to simply wish it away. If rūpa were Self we would be able to do this. It is important to notice that the sense in which it is said that we do not have control over rūpa seems to be one of direct control over its states, in particular its state of being subject to affliction. In the passage above, there is no denial of the idea that we can do as we wish with respect to the actions we perform with and through our bodies; the denial is of the notion that we can be as we wish with respect to the presence or absence of affliction. The wish that the Buddha describes as impossible to fulfill is “Let my form be thus, let my form not be thus,” not “Let my form do thus, let my form not do thus.” If free will is simply understood as the empirical ability of persons to act voluntarily or to do as they want within a specified set of constraints, then the Buddha’s position does not here imply any denial of this. All it suggests is that we cannot directly wish away the suffering associated with the first aggregate. In point of fact, the Buddha’s teachings are premised on the idea that it is possible to do something about suffering, and indeed to eliminate it. But we cannot simply do away with it directly. Are we then to conclude that the Buddha’s doctrine implies a limited or qualified free will, one in which we can do as we will if not actually be as we will with respect to suffering? Is this the end of the story? Actually, the Buddha’s position turns out to be considerably more complex than this. This can be seen if we pause to analyze the notion of ‘will’ a bit more carefully.

Buddhism and the will
What is the will? What do we mean to refer to when we employ this word? From a Theravāda Buddhist perspective, if the five aggre-

No self, no free will, no problem

249

gates are all that a person is, we must locate the idea of ‘will’ within them. The problem of matching concepts is, of course, always a difficult one for those engaged in the enterprise of cross-cultural philosophy. That being said, whatever aspect of the human person we might consider the will to be, in the Buddhist context it will have to correspond to some aspect or aspects of the five aggregates − for this is all that a person is considered to be. Among Pali terms, the most obvious candidate for ‘will’ is cetanā. This word is usually translated as ‘intention’ or ‘volition.’ Keown (1992: 212−221) has also rendered it as ‘choice.’ These English words each carry very different implications, a few of which can be brought out here. The term ‘volition’ implies an actual effort, an impulse towards action, perhaps even a ‘trying’ or exertion; the English word ‘intention’ does not. It simply indicates a plan or desire to act, but not necessarily one that has been initiated or set in motion. The notion of ‘choice’ implies the conscious entertaining of alternatives and an actual mental event in which one alternative is favoured over the others. Of these terms, ‘will’ seems to be the most general, encompassing all of the others in its potential meaning. We will have more to say concerning some of the western variations of the notion of the will in the next section of this paper; but for the moment let us accept the tentative identification of the English language concept ‘will’ with the Pali concept of cetanā.16 However inexact the match may be, the concept of the will must correspond to some aspect or aspects of the five aggregates − and this is actually all we need for our argument to proceed. Here it is necessary to observe that cetanā is considered part of the fourth aggregate, the saṅkhāras. The latter term has commonly been translated as ‘volitional formations,’ a heading meant to capture those mental events that direct one’s actions − physical, mental and

16 From a more Mahāyāna perspective, the will might be viewed as an abstraction, a reification of diverse events, possessing no inherent nature of its own. There are, of course, many ways to pick out and group these events into a concept of ‘will’ (e.g. desires we are moved by, desires we identify with, our intentions, etc).

250

Martin T. Adam

vocal.17 It would appear, then, that volitional formations, qua voli­ tional formations, constitute the very aggregate in virtue of which action can be said to be voluntary. Keeping this understanding in mind allows us to raise a deeper question regarding the freedom of the will. For, as mentioned, in the Anattalakkhaṇa Sutta an analysis identical to that carried out on rūpa is carried out on each of the aggregates in turn, including that of the saṅkhāras.
“Volitional formations are nonself. For if, bhikkhus, volitional formations were self, they would not lead to affliction, and it would be possible to have it of volitional formations: ‘Let my volitional formations be thus; let my volitional formations not be thus.’ But because volitional formations are nonself, volitional formations lead to affliction, and it is not possible to have it of volitional formations: ‘Let my volitional formations be thus; let my volitional formations not be thus’.”18 (SN III 67)

In effect, then, this analysis suggests that the very aggregate that includes the will is itself unfree; it is not subject to control. What could this mean? If we follow our earlier analysis with respect to rūpa, the lack of freedom here would simply amount to our inabili17 Mahasi Sayadaw provides the following explanation. “Saṅkhāras are the mental states headed by cetanā, volition. There are fifty two kinds of mental states. With the exception of feeling and perception, the remaining fifty constitute the aggregate of volitional formations, saṅkhārakkhandha. In Sutta discourses, only cetanā, volition, is specified as representing the saṅkhāra activities, but according to the Abhidhamma, we have other volitional formations that can produce kamma, such as attention (manasikāra), initial application of thought (vitakka), sustained application (vicāra), zest (pīti), greed (lobha), hatred (dosa), delusion (moha), nongreed, nonhatred, and nondelusion. These fifty kinds of volitional formations are responsible for all kinds of activities, such as going, standing, sitting, sleeping, bending, stretching, smiling, and speaking. These actions, as well as mental activities such as thinking, visual consciousness and auditory consciousness, are carried out and directed by saṅkhārā.” (1996: 49−50) 18 Saṅkhārā anattā // saṅkhārā ca hidaṁ bhikkhave attā abhavissaṃsu // na yidaṃ saṅkhārā ābādhāya saṃvatteyyuṃ // labbhetha ca saṅkhāresu Evaṃ me saṅkhārā hontu evaṃ me saṅkhārā mā ahesunti // yasmā ca kho bhikkhave saṅkhārā anattā tasmā saṅkhārā ābādhāya saṃvattanti // na ca labbhati saṅkhāresu Evaṃ me saṅkhārā hontu evaṃ me saṅkhārā mā ahe­ sunti // //

No self, no free will, no problem

251

ty to make saṅkhāras unafflicted directly by wishing them to be so. In glossing this passage, Mahasi Sayadaw indicates the manner in which we would change our volitional formations if we only could: we would make them all wholesome (kusala) and not unwholesome (akusala) respectively.19 Unfortunately this is impossible. This is a critical consideration, for it suggests that the very mental factors determining the morality of action are not subject to control. If the saṅkhāras are not subject to control, this means that we are unable to directly determine their composition. The mental states that direct our actions − the very desires, attitudes, and values we identify with and which determine the morality of our actions − are themselves not under control. In this case, it might be said that we are unfree with respect to the volitional aspect of ‘who or what we are,’ rather than with regard to the aspect of what we do. If this is indeed the implication, then it would appear that the Buddha probably would not have disagreed with the following assertion, famously attributed to Schopenhauer: “A man can do what he wants, but not want what he wants.”20 The Buddhist analysis suggests that the problem of free will is not simply firstorder issue as to whether we can do what we want. There is a much deeper problem − one that turns on second-order considerations as to whether we can be what we want to be, or, put another way, whether we can have the wills we want to have. The issue of the freedom of the will is a question regarding whether we have freedom with respect to our own constitutions. The Buddha’s answer appears to be negative. While it may be the case that we can be judged empirically free to the extent that we can do as we want, we are not metaphysically free in the sense of being able to directly
“Monks, were volitional factors self, they would not inflict suffering and it should be possible to say of them, ‘Let volitional formations be thus (all wholesome), let volitional formations be not thus (unwholesome),’ and manage them accordingly.” (Sayadaw 1996: 49) 20 Quoted in Einstein 1982: 8. Although I cannot locate an exact, original source for this quotation, Schopenhauer’s position is expressed throughout his essay On The Freedom of The Will (1985). See note 25 below. Along the same lines, Bertrand Russell is also reputed to have quipped that while we can do as we please, we can not please as we please.
19

252

Martin T. Adam

determine the constellation of factors we identify with, and out of which our actions proceed. In the context of this sutta, the reasons for this assertion are clear: the will is not subject to control in this way, because, quite simply, there is no independent entity over and above the shifting configuration of mental factors to do the controlling. There is no self-controlling controller. There is no one (i.e. no single unified being) holding the reins. There is no Self.

Harry Frankfurt and the Buddha
The Buddha’s implied position on the freedom of the will can be fruitfully analyzed by comparing it with a recent and influential account of the will’s freedom provided by the philosopher Harry Frankfurt. As is the case for Buddhism, in Frankfurt’s analysis second-order considerations are the critical factor in assessing the will’s freedom. Frankfurt notes what he takes to be a unique feature about human beings:
“Besides wanting and choosing and being moved to do this or that, men may also want to have (or not to have) certain desires and motives. They are capable of wanting to be different, in their preferences and purposes, from what they are. Many animals appear to have the capacity for […] ‘desires of the first order,’ which are simply desires to do or not to do one thing or another. No animal other than man, however, appears to have the capacity for reflective self-evaluation that is manifested in the formation of second-order desires.” (Frankfurt 1982: 82−83)

These observations concerning the self-reflective capacities of human beings are directly pertinent to the Buddhist analysis, where they find an obvious resonance in the capacity of human beings to reflect on the nature and composition of the aggregates themselves. Interestingly, however, the conclusions Frankfurt arrives at are rather different from those we have just reached. In brief, Frankfurt argues that a coherent account of free will can be given in terms of the capacity of human beings to form second-order desires and volitions about their first-order desires. To understand Frankfurt’s account we must note that he identifies the

No self, no free will, no problem

253

will with the first-order desire that actually moves, or would move, an individual to act.21 This he terms the agent’s effective desire.
(The notion of the will) is the notion of an effective desire − one that moves (or will or would move) a person all the way to action. Thus the notion of the will is not coextensive with what an agent intends to do. For even though an agent may have a settled intention to do X, he may none the less do something else instead of doing X because, despite his intention, his desire to do X proves to be weaker or less effective than some conflicting desire.22 (Frankfurt 1982: 84)

Frankfurt’s account of free will turns on the notion that one is free only if one wants to be moved by the desire that actually does move

This conception of the will is not entirely dissimilar to the general Buddhist understanding of cetanā as the mental factor lying behind voluntary behaviour, in virtue of which such behaviour is considered action (kam­ ma). It should be noted, however, that the Buddhist has a broader conception of action, one that encompasses acts of mind as well as those of body and speech. Cetanā is itself a mental action; it might therefore be said that in its case, its being is its doing. Cetanā is ‘intention’ in the sense of being the ‘intention’ that occurs while doing an action (i.e. a volition or mental impulse). However, as Buddhism accepts the idea of mental action, the activity of planning/intending-to-do a future action is itself a current mental action, with its own cetanā. Thanks to Peter Harvey for the latter observation. 22 In identifying the will with effective desire, Frankfurt is adhering to a conception of the will and willing that goes back at least as far as John Locke (1632−1704). Locke defines ‘the Will’ as the power to command and to prefer one option over another. Interestingly, from a Buddhist perspective, Locke warns his readers not to fall prey to the tendency of thinking of this power as an actual faculty that has autonomous and real being in the souls of men (1959: 314−15). This ‘power’ is distinguishable from the instances in which an individual exercises his will. Instances of ‘willing’ or ‘volition’ are those in which a desire actually moves a person to act. “This power which the mind has thus to order the consideration of any idea, or the forbearing to consider it; or to prefer the motion of any part of the body to its rest, and vice versa, in any particular instance, is that which we call the Will. The actual exercise of that power, by directing any particular action, or its forbearance, is that which we call volition or willing.” (Locke 1959: 313−314) Thus Locke draws a distinction between the Will and the activity of willing. To will is to be moved to act in some manner.

21

254

Martin T. Adam

one to act. If one does not want to be moved to act by that desire but is nevertheless moved by it then the will is unfree. The example Frankfurt employs as an illustration is that of an unwilling drug addict. Frankfurt’s analysis of the condition of such a person is that he is the subject of conflicting first-order desires and a second-order volition towards one of these. He both wants and does not want to take the drug. But in taking the drug he is being moved to act in a way that he desires not to. His desire to take the drug on these occasions, because it moves him to act, may be identified with his will. And in this case it is unfree. It is unfree because the agent does not want it.23

In refining his account Frankfurt employs the notion of second-order volitions as a special kind of second-order desire. Second-order desires, in the most general sense, are simply desires for desires. A second-order volition is a second-order desire that has as its object the efficacy of a particular first-order desire. This is an important distinction, insofar as it is possible for someone to want to possess a particular first-order desire without wanting it to be effective. To see this we can imagine the case of another addict, a gambler, who is actually quite happy with his habit, who yet wants to have the desire to give it up, but who does not want this latter desire to be effective. “If I didn’t want to give it up at least a little bit,” he might reason, “then my friends wouldn’t be sympathetic and lend me the money I need.” This person has a second-order desire (a desire for a desire), but not a second-order volition. If, contra Frankfurt (and Locke), we choose to conceive of the will as the desire we identify with, then for this case we could maintain that the unwilling addict’s will is free while his action is not. This is, in fact, another well-attested usage of the term ‘will;’ in saying that one wills something, there is no necessary implication of effort by the agent. Rather, the notion of will is linked with our deepest wishes or values, or even our self-concept. The manner in which Bhikkhu Bodhi speaks of the will (see above) seems to reflect this usage: the will is identified with a very deep desire, in this case the desire to be free from affliction − ineffective though this may be. Augustine could be taken as another example of someone who thinks of the will in this way. In general, most philosophical discussions of free will can be usefully divided along the lines of these two different ways of conceiving the will. It is important to be clear about which concept is being presupposed in any case where free will is being discussed; obviously, these two different conceptions of the will will lead to two very different ways of talking about free will.

23

No self, no free will, no problem

255

But it is clear that most cases in life are not like that of the unwilling addict. Most of us, most of the time, are moved to act by ordinary desires that we want to have move us to act. In such cases we can be said to possess second-order volitions directed towards our wills. We approve of our will. Hence for the most part, on Frankfurt’s analysis, our actions are freely willed. In fact, this way of thinking about free will provides a good explanation for these cases, in which we ‘feel free’ in acting and are therefore willing to take responsibility for what we do. Our actions reflect our choices and the values we identify with. In brief, they reflect ‘who we are’ (or at least who we take ourselves to be). In spite of the refreshing clarity of Frankfurt’s account of free will, it is not without its difficulties. Here I will mention only two that are particularly relevant to our present concerns. The first difficulty is that an individual’s second-order desires and volitions are not consistent through time. A person’s deeper values and wishes are subject to change depending on a great variety of internal and external conditions. In Frankfurt’s terms, we may say that we are inconsistent as to what we want our will to be. Which of one’s various ‘selves’ does one identify as being one’s ‘true self’? On what basis? This issue is clearly relevant in the context of Buddhism. A second and more serious problem stems from a basic ambiguity in Frankfurt’s conception of freedom. While it is clear that Frankfurt regards the will’s freedom as contingent on the presence of a second-order volition it not is not at all clear that this is a sufficient condition. It would appear that while Frankfurt’s account may provide a good analysis of those actions we feel free in undertaking and for which we feel responsible, it may not actually address the deeper metaphysical problem presented by determinism.24 This problem can be brought out through the following considerations. A moment’s reflection reveals an infinite regress that threatens to result when the predication of freedom is made to turn
24 A close reading of Frankfurt suggests that he holds that it must also be the case that the agent ‘could have done otherwise’ than constitute his will as he did (1982: 94). If this is indeed Frankfurt’s view, then his account might best be seen as providing insight into the psychology of freedom, rather than addressing its metaphysics.

256

Martin T. Adam

on the presence of higher-order volitions. If the freedom of the will is dependent on the presence of a second-order volition towards it, are we free with respect to that second-order volition? Do we not then require a third-order volition to ensure the freedom of the second? Once this sequence gets started we are quickly faced with the prospect of requiring an infinite number of higher-order volitions, each needed to guarantee the freedom of the one below it; ultimately an infinite series of volitions would be required to guarantee the freedom of the will. But this is impossible.25 Perhaps this difficulty could be dealt with by arguing that, as a point of empirical fact, all we ever really do have are desires of the first and second-order or, at most, of the third order. If we choose to speak of even further, higher-order desires and volitions, it is not really clear that we would be referring to anything at all. The thirdorder statement, “I want to have the desire to have the effective desire to do X” seems rather dubious in terms of its possible point of reference. And it certainly does not appear that by adding another “I want” to the beginning of the sentence we would be adding any new information about the subject’s actual mental life. At some point there is no further “I want;” the causes for one’s desires are impersonal. One’s desires just are, they arise without any choice, or even reflection, being involved.

25 Cf. Schopenhauer 1985: 6: “The empirical concept of freedom signifies: ‘I am free when I can do what I will.’ Here in the phrase ‘what I will’ the freedom is already affirmed. But when we now inquire about the freedom of the willing itself, the question would now take this form: ‘Can you also will your volitions?,’ as if a volition depended on another volition which lay behind it. Suppose that this question is answered in the affirmative, what then? Another question would arise: ‘Can you also will that which you will to will?’ Thus we would be pushed back indefinitely, since we would think that a volition depended on a previous, deeper lying volition. In vain would we try to arrive in this way finally at a volition which we must think of and accept as dependent on nothing else. But if we were willing to accept such a volition, we could as well accept the first as the one we happened to make the last. Consequently, the question would be reduced to a simple: ‘Can you will?’ But whether a mere affirmation of this question decides the problem of the freedom of the will, is what we wanted to know. So the problem remains unresolved.”

No self, no free will, no problem

257

These considerations serve to underline the limitations of Frankfurt’s account of free will. They return us squarely to our earlier observations, in the Buddhist context, regarding the freedom of the mental states upon which actions are based. As we have seen, it is indeed possible to sensibly ask whether a person has the will they want to have. A determinist will argue that the causes that give rise to the mental states upon which one’s actions are based are not subject to control; they are, when one traces them back, ultimately impersonal in nature (in the sense of being e.g. historical, genetic, cultural, etc). Determinists take the fact that choices are caused events very seriously; even if our present awareness can reflect on and evaluate our choices, the thoughts and values entering into these evaluations are, in the last analysis, themselves beyond control − at some point they just are; we do not choose them or their causes. The Buddhist position accords with such considerations. There is no final, independent Self at which point the chain of causes and conditions magically comes to a halt. In the last analysis it is not possible to have it of the will, ‘Let my will be thus, let my will not be thus.’

The foundations of morality
If this is so, should it then be concluded that the Buddhist position, like that of the incompatiblist determinist, undermines the foundations for moral responsibility? If there is no essential Self to which responsibility may ultimately be attributed, is there then no moral responsibility at all? Interestingly, from the Buddhist perspective the answer would appear to be no. In fact the Buddha appears to have held the unusual view (from the western philosophical perspective) that while the will is not metaphysically free, moral responsibility is just a fact about the way things are. Although ultimately there is no autonomous, permanent self-essence or Self, persons’ actions do have results that accord with the moral character of those actions. Moral causality is simply one kind of causality operational in the universe. So moral responsibility is simply one kind of causal responsibility. Like it or not, results flow from actions; happiness and suffering are the inevitable results of moral (kusala) and immoral (akusala) action. Such action (kamma) is dis-

258

Martin T. Adam

tinguishable as mental, physical, and vocal behaviour that is voluntarily performed or willingly done (i.e. accompanied by cetanā); this is the key factor in determining moral responsibility. Freedom of the will is not. The point is that the action is voluntary, not that the will is free.26 In any case, it can be seen that the problem of the compatibility of universal causality and moral responsibility does not appear to have been a concern to the Buddha. Causality − in terms of such things as motivations, and karmic results − itself is a necessary correlate of morality from the Buddhist perspective.27 What did seem to concern the Buddha, however, was perhaps a not altogether unrelated problem, which may be stated as follows. If a person is ultimately only a series of causally interrelated events, some of which are identified with, how is it that freedom, qua liberation (nibbāna) is possible? Put another way: if the five aggregates are ultimately beyond our ability to control, how is it possible that we would ever begin to strive for, much less reach, the goal which is the end of suffering? In the Mahāli Sutta, the Buddha provides an answer. The context is a question posed by one Mahali, who has been listening to the teachings of the samaṇa Pūraṇa Kassapa. The latter has been espousing the view that beings are defiled and purified without

A case might be made that it is precisely in these instances of willingly performed behaviour that the will may be said to be free. Thus the will is empirically free, rather than metaphysically free, and this is, from the Buddhist perspective, the only kind of freedom necessary for moral responsibility. But this position is problematic. The notion that kamma is free by definition or nature does not correspond to the manner in which it is actually treated in the Buddhist tradition. It does not rest easily with the basic Buddhist orientation that one should aim to become free from kamma, and indeed that one does become progressively more free from it as one advances on the spiritual path. The idea of free action in Buddhism can probably be best understood as action that is performed when there is freedom from delusion on the part of the agent. Such freedom is, however, not usually attributable to the ordinary person (see below). 27 Buddhism differs from western compatibilist theories that seek to ground judgments of moral responsibility in some kind of intersubjective agreement or social consensus. Morality is objectively grounded in the way things are. Dhamma, qua moral order, is not conventional.

26

No self, no free will, no problem

259

cause or condition. Such a view would seem parallel that of the indeterminist described at the outset of this paper. The Buddha’s explanation as to how it is that one is purified with cause and condition runs as follows:
“If Mahali, this form were exclusively pleasurable, immersed in pleasure, steeped in pleasure, and if it were not [also] steeped in suffering, beings would not experience revulsion towards it. But because form is suffering, immersed in suffering, steeped in suffering, and is not steeped [only] in pleasure, beings experience revulsion towards it. Experiencing revulsion, they become dispassionate, and through dispassion they are purified. This, Mahali, is a cause and condition for the purification of beings; it is thus that beings are purified with cause and condition.”28 (SN III 70)

The Buddha goes on to give identical analyses with regard to each of the other four aggregates. It is, perhaps, merely fortuitous that this sutta is placed in the Saṃyutta-Nikāya immediately following the Anattalakkhaṇa Sutta. On the other hand, its placement there could suggest that the early compilers of the canon were aware of the issue raised by the preceding sutta. They may well have sensed the possibility of doubt arising with regard to the compatibility of spiritual freedom and universal causality, given the extraordinary claim that none among the five aggregates are subject to control. This is admittedly speculative, but it is not, perhaps, entirely implausible. In any case, we can see that the Buddha taught that beings are actually constituted in such a way as to allow for the possible at28 Rūpaṃ ca hidam Mahāli ekantasukhaṃ abhavissa sukkānupatitaṃ sukhāvakkantam anavakkantaṃ dukkhena // nayidaṃ sattā rūpasmiṃ nibbindeyyuṃ // // Yasmā ca kho Mahāli rūpaṃ dukkhaṃ dukkhānupattitaṃ dukkhāvakkantam anavakkantam sukhena // tasmā sattā rupāsmiṃ nibbindanti nibbindaṃ virajjanti virāgā visujjhanti // // Ayaṃ kho Mahāli hetu ayam paccayo sattānaṃ visuddhiyā // evam pi sahetu-sapaccayā sattā vis­ sujjhanti // // Essentially the same sequence of causes can be found in the Anattalakkhaṇa Sutta, with the difference that rather than beginning with suffering it begins with the recognition, for each of the aggregates, ‘This is not mine, this I am not, this is not my self.’ A further difference is that the Mahāli Sutta explicitly frames the discussion in terms of causes and conditions.

260

Martin T. Adam

tainment of purification and liberation. Among the possible realms of rebirth, it is the realm of human beings in particular that is considered to have just the right balance of pleasure and suffering as to generate the motivation to aspire for freedom. We are lucky!

Towards an account of freedom in Buddhism.
Freedom in Buddhism is not conceived of as a quality of the will. If there is no independent originary source over and above our mental, physical and vocal actions, then there certainly cannot be any free will. Thus the assertion that from a higher perspective the will cannot be said to be either free or unfree is, it seems to me, off the mark. It is precisely from the higher perspective that the will can be seen to be, in truth, unfree. Our lack of free will in this case exactly parallels the Buddhist understanding of personal identity. While it makes sense to talk of persons, there is no Self. It is not the case that such a Self neither exists nor does not exist. From a Theravāda Buddhist perspective, this formulation is mistaken. The notion of free will is conceptually bound up with the notion of Self. Just as the Self is ultimately seen to be a delusion, so too is the will’s freedom. No Self, no free will. This is a difficult point. It is interesting to note that the Anatta­ lakkhaṇa Sutta is not addressed to ordinary persons. It is a discourse directed to an audience of learners or disciples in higher training (sekhas), individuals who have attained the higher perspective that sees things as they really are. There is an important sense in which these individuals, beginning with the stream-enterer (sotāpanna), are understood to be free already − in a way that is not true of ordinary people (puthujjanas). They are free from the false view of Self.29 The very notion of the sekha as a kind of agent is defined in terms of having undergone a moment of transformative insight into the truth of nonself. There is nothing that it is right
29 They are free of sakkāya-diṭṭhi, ie beliefs or views that see any of the aggregates as Self, owned by a Self, in Self, or containing Self. Sekhas are not, however, free of the more diffuse, non-specific ‘I am’ conceit (asmi­ māna). This only disappears when one reaches arahathood. Thanks to Peter Harvey for this observation. See Harvey 2004: 32.

No self, no free will, no problem

261

to view as in some way related to ‘Self.’ The first five disciples are said to have experienced this insight some days earlier, upon hearing the Buddha’s first sermon, the Dhammacakkappavattana Sutta.30 Upon hearing the second it is said that they became arahats (Bodhi 2000: 1066).31 In closing this essay I would suggest that these considerations concerning the reported audience of the Buddha’s second sermon offer us a clue as to how freedom in Buddhism may be best understood. Freedom is principally a predicate of persons and consists in an absence of suffering and its causes. It is thus dependent on the state of knowledge and purity of the awareness of the agent. The ultimate aim of Buddhism is, of course, freedom from suffering. Suffering is a reality that, first and foremost, is to be under­ stood. Thus freedom means understanding, and then abandoning, the causes and conditions of suffering within oneself. It is this very knowledge of causality (paṭiccasamuppāda) that allows the mind to become liberated. Because different degrees of insight and mental purity may be attributed to the various kinds of spiritual actor described by the Buddha, so too corresponding levels of freedom may be attributed to them. The ordinary person (puthujjana) is not a free person, operating as she does from within the deluded perspective of being an independent actor in control of her life in saṃsāra. Although such an agent may be reflexively aware of her actions, and although such actions may be voluntary, or empirically free in some other sense, they occur in the context of the basic delusion of an underlying ‘I,’ whence they are regarded as originating independently. Being out of touch with reality in this way, such a person’s mind will ineviThis would explain the argument’s apparent presupposition that the five aggregates are all that a person is. 31 In earlier articles I have argued that early Buddhist ethics can best be understood as a kind of agent based moral contextualism, in which the key classes of agent are the puthujjana, the sekha and the arahat. Rather than beginning with an examination of the nature of moral action per se, this approach to Buddhist ethics begins by investigating the phenomenology of moral experience belonging to each kind of person. See Adam 2005 and 2008.
30

262

Martin T. Adam

tably be trapped in confusion, inconsistent and conflicting desires, and suffering. From the Buddhist perspective this kind of person must be regarded as unfree. The sekha, on the other hand, is a free person in a certain way, having rid herself of the basic delusion of self (and with it, we should add, any notion of an independent will). Being irreversibly oriented away from suffering and its causes and towards nibbāna, such a person can be characterized as more consistently having the desires she wants to have and on a firmer basis than is the case for ordinary persons.32 An internal order has been irreversibly established in such a person and she cannot do otherwise than act from within a psychological orientation that is turned towards nibbāna. Although the mind of the sekha remains obscured to some extent, afflicted by residual defilements, the complete freedom of nibbāna is assured. The arahat has realized nibbāna; she has attained spiritual freedom.33 She is a completely free person, being free from all mental defilements including any trace of self-centred desire; indeed because of this she is free from kamma itself.34 In fact, because the arahat is entirely free from desire that she identifies with, she might
32 Different gradations of freedom can be associated with each of the variDifferent ous subdivisions of sekha i.e. the stream-enterer (sotāpanna), once-returner (sakadāgāmin), and non-returner (anāgāmin). In general, although their sīla is perfect, they remain subject to various fetters (samyojana), including such unskillful desires as craving for subtle rūpa and arūpa states. Streamenterers and once-returners are still subject to sense-desires and ill-will. 33 See Harvey 2007: 81−84 for a more detailed discussion of the qualities associated with the arahat’s spiritual freedom. 34 In an earlier article (2005) I have attempted to spell out these distinctions among kinds of persons in relation to some of the key vocabulary of morality employed in the texts, principally the antonymous pairs of kusalaakusala, puñña-apuñña, and sukka-kaṇha. I have argued that the latter dyad serves as a conceptual bridge between the former two pairs, carrying connotations of each. Whereas kusala is a term with strong epistemic implications, puñña principally indicates moral goodness. The two are coextensive. More recently (2008) I have moved towards providing a phenomenological account of the same distinctions. I characterize the action of the sekha as principally (or teleologically) kusala and secondarily (or intrumentally) puñña.

No self, no free will, no problem

263

even be described as being free from the will. To put the matter in this way depends, of course, on a conception of the will as ‘desire one identifies with.’ On the other hand, if we follow Frankfurt and identify the will with the ‘desire that moves one to act’ then the arahat can also be described as having the will she wants to have, and therefore a ‘free’ will.35 While no agent can be said to possess freedom of the will in the metaphysical sense of self-causation sought by some western philosophers, the arahat can be said to have a free will, indeed a perfectly free will, in the empirical sense of this expression proposed by Frankfurt. More generally, freedom in Buddhism can be regarded negatively as a freedom from constraints upon a person − either internal or external depending on one’s focus. That is to say, just as the ordinary person, the learner and the liberated being are free from mental defilements to varying degrees, so too they are free from saṃsāra. The sekha may be contrasted with the ordinary person in that while the latter wanders aimlessly in saṃsāra, the former is consistently oriented towards nibbāna − a goal she is destined to attain. Alternatively, if we wish to characterize their respective states in positive terms, one may say that each type of person possesses a different degree of knowledge and mental purity. Thinking of freedom principally as a quality of these ideal types, dependent on their respective levels of spiritual realization,
35 One may also say that they conduct themselves freely, in a way that isn’t true of the ordinary person − their conduct is informed by a veridical awareness of the way things really are. As a consequence their activities are accompanied by a natural sense of freedom. Similar feelings would be associated with the experience of the sekha. As well, we would expect to find in the sekha a new sense of freedom qua relief at being so recently rid of the burden of a belief in Self. Although we cannot examine the idea in any detail in this essay, it should be noted that Buddhism is not unique in maintaining a teleological conception of freedom, one in which freedom is at least in part conceived in terms of conduct that is in accord with one’s higher nature. In the case of Buddhism, of course, the paradoxical aspect is that there is no one, ultimately, who possesses such a nature. A case can be made that because there is no Self to serve as a repository of hidden desires, the Buddhist position avoids the dangers associated with such an idea − on which see Berlin 1984: 22−25.

264

Martin T. Adam

allows us to gain a clearer understanding of some of Harvey’s observations, mentioned at the outset of this paper. We can now see how it is that freedom may be thought of as possessing varying “degrees.” The suggestion here is that such variation may best be regarded as occurring among kinds of person and not (or at least not principally) within an individual person over the short term (although, of course, an ordinary person may become a sekha and so on). Thus the Buddha’s teachings do in fact suggest that freedom admits of degrees. But they do not imply that human beings are possessed of a will that is metaphysically free, or one that is both metaphysically free and unfree, or even one that is neither. From a Theravāda Buddhist perspective it would be more accurate to say that while a person’s will may be judged empirically free in one sense or another, it definitely is not possible for anyone to possess a metaphysically free will. But for the Buddhist this presents no problem.

Abbreviations
SN Saṃyutta-Nikāya, ed. L. Feer. 5 vols. London 1884–1898 (Pali Text Society). English translation: see Bodhi (2000).

Bibliography
Adam, Martin T. (2005) “Groundwork for a Metaphysic of Buddhist Morals: A New Analysis of puñña and kusala, in light of sukka.” Journal of Buddhist Ethics 12, 62−85. Adam, Martin T. (2008) “Classes of Agent and the Moral Logic of the Pali Canon.” Argumentation 22, 115−124. Berlin, Isiah. (1984) “Two Concepts of Liberty.” In Liberalism and Its Critics. Ed. Michael Sandel. New York: New York University Press, 15−36. Bodhi, Bhikkhu. (2000) The Connected Discourses of the Buddha. 2 vols., Boston: Wisdom Publications. Einstein, Albert. (1982) Ideas and Opinions. New York: Three Rivers Press. Federman, Asaf. (2010) “What Kind of Free Will did the Buddha Teach?” Philosophy East and West 60, 1−19. Frankfurt, Harry. (1982) “Freedom of the Will and The Concept of a Person.” In Free Will. Ed. Gary Watson. New York: Oxford University Press, 81−95.

No self, no free will, no problem

265

Harvey, Peter (2004) The Selfless Mind. Oxon: RoutedgeCurzon. Harvey, Peter (2007) “‘Freedom of the Will’ in Light of Theravāda Buddhist Teachings.” Journal of Buddhist Ethics 14, 35−98. Keown, Damien. (1992) The Nature of Buddhist Ethics. London: Macmillan Press. Kuan, Tse-fu. (2009) “Rethinking Non-Self: A New Perspective from the Ekottarika-āgama.” Buddhist Studies Review 26.2, 155−175. Locke, John. (1959) An Essay Concerning Human Understanding. Ed. A.C. Fraser. New York: Dover Publications, Inc. Sayadaw, Ven. Mahasi. (1996) The Great Discourse on Not Self. Bangkok: Buddhadhamma Foundation. Schopenhauer, Arthur. (1985) On the Freedom of the Will. Oxford: Basil Blackwell. Thanissaro, Bhikkhu. (1996) The Wings to Awakening. Barre: Dhamma Dana Publications. Also available at: http://www.accesstoinsight.org/lib/authors/thanissaro/wings/index.html, last visited 02-06-2011 Van Inwagen, Peter. (1982) “The Incompatibility of Free Will and Determinism.” In Free Will. Ed. Gary Watson. New York: Oxford University Press, 46−58.

Buddhist metaethics
Bronwyn Finnigan

Introduction
The nature and scope of Buddhist ethics is a topic of inquiry that is increasingly garnering the critical attention of contemporary Buddhist thinkers. Significantly, much recent work in this field involves attempts to determine which contemporary Western normative ethical theory a Buddhist ethic most closely resembles. For instance, Keown (2001) and, later, Cooper & James (2005: 68) claim that Buddhist ethics is a type of virtue ethics. Siderits (2003, 2007) argues that it is unmistakably consequentialist, as do Williams (1998) and Goodman (2008). Velez (2004) and Clayton (2006) argue that Buddhist ethics is best understood as a combination of virtue ethics and utilitarianism whilst Harvey (2000: 49), though acknowledging the analogies, nonetheless maintains that a Buddhist ethical theory is significantly distinct.1 While a somewhat analogical exposition of a Buddhist ethic may suggest simplification and conflation (even reduction), it is not difficult to see why one might find such a strategy appealing. As Georges Dreyfus (1995) reminds us, the Tibetan Buddhist tradition (to take just one Buddhist lineage as an example) did not develop systematic theoretical reflections on the nature and scope of ethics; nor, indeed, do Indic Buddhist texts provide systematic analyses of ethics and ethical concepts. This is not to overlook the fact that
1 Note that all of the above positions come with various limitations and qualifications. For instance, Goodman’s position is limited to Indian Mahāyāna Buddhist ethics whereas Clayton’s position is presented in the context of an exposition of Śāntideva’s Śikṣāsamuccaya.

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 267–297

268

Bronwyn Finnigan

Indo-Tibetan traditions had substantive ethical views; moral precepts are advanced and discussed at great length throughout the entire Buddhist tradition. Dreyfus’ point, however, is that such discussion is often conducted in practical terms (i.e. concerning the implementation and application of precepts) rather than with the kind of rich and systematic reflection about their nature and scope that is typical of contemporaneous treatises focusing on epistemology and philosophy of language. As Dreyfus rightly points out, this dearth in Buddhist ethical theorization causes problems for modern scholars who want to explore the nature of Buddhist ethics. To a large extent, Buddhist ethicists find themselves having to construct Buddhist ethical theories. Given that Western philosophy has quite a robust tradition of ethical theorization, it seems reasonable to exploit analogies with these systems in the attempt to weave together the various Buddhist moral threads.2
2 One might wonder whether Buddhism is in need of ethical theorization. The fact that systematized theorizations were not developed within the Indo-Tibetan traditions may be taken to suggest something distinctive about Buddhist approaches to ethics. For instance, Hallisey (1996) takes this to indicate that Buddhist practitioners and intellectuals employed a kind of ethical particularism which recognises a diversity of values. Hallisey contrasts this notion with ‘ethical theories,’ interpreted as aiming to provide singular criteria or methods for the resolution of moral conflicts, moral decision-making and action-guidance. Whether or not we infer from the historical lack of ethical theorization in the Buddhist tradition that Buddhism, therefore, has no need for ethical theory will depend on what we mean by ‘ethical theory.’ Hallisey’s characterisation of this notion seems consistent with many contemporary Western views on the nature and purpose of normative ethical theorizing; the predicate ‘normative’ may be taken to suggest that ethical theories aim to provide criteria and methods for figuring out what we ‘ought’ to do. In response, one might argue that Buddhist canonical texts already provide sufficient resources aimed at advising practitioners on the recognition, instantiation and/or implementation of moral precepts and/or qualities without additionally requiring the introduction of ethical theories, so understood. This is not the only way to think of ethical theory, however. Alternatively, it might be understood as the attempt to coherently explain the nature and role of ethics, in its variety of aspects, within the wider context of the Buddhist

Buddhist metaethics

269

As in the contemporary Western ethical tradition, questions concerning (normative) ethical theorizing are only one species of ethical concern.3 An additional area of interest is metaethics.4 Metaethical questions primarily concern the nature and status of, and assumptions that support, ethical claims and theories. Such assumptions include, inter alia, various epistemological and metaphysical commitments. One particular metaphysical commitment that has been explicitly considered by most aforementioned Buddhist ethicists is the
philosophical tradition. Ethical theory, on this broader interpretation, seeks to answer questions specifically about ethics (such as its function in Buddhist soteriology; the function of moral precepts for Buddhist ethical practice; the role of analogies and stories for the cultivation of ethical practice) rather than provide instructions or methods for how practitioners should engage in ethical practice. On this broader conception of ethical theory, Hallisey’s notion of ethical particularism, itself, counts as an ethical theory. For instance, Hallisey claims that ‘ethical particularism’ is a necessary heuristic for explaining how ethical stories function for Buddhist practitioners (i.e. Hallisey argues that they help sensitise practitioners to morally salient features of situations in order to negotiate conflicts without appeal to criteria). To the extent that ethical particularism seeks to provide an explanation of a certain aspect of Buddhist ethics, ethical particularism, on our broader definition, is an ethical theory. However, Hallisey is right to argue that it is not a normative ethical theory, understood as providing specific methods or criteria to guide moral decision-making as oftentimes conceived by contemporary Western ethicists. This might be taken to indicate that the contemporary Western ethical taxonomies are unduly limited. Indeed, on this broader definition of ethical theory, the division between ethical theory and metaethical concerns becomes much less distinct. However, I shall leave further reflection on this issue for a different paper. 3 See previous note. 4 The contemporary Western ethical tradition also pays considerable attention to practical or ‘applied ethics.’ Applied ethics proceeds, for the most part, via application of ideas developed in ethical theories to practical situations and cases. Although I shall not discuss applied ethical questions in this paper, it is notable that Buddhist scholars have recently ventured into this territory, offering accounts of how a Buddhist ethic may provide solutions to issues concerning the environment (Cooper & James 2005), abortion (Keown ed. 1998), and bioethics (Levin 2004).

270

Bronwyn Finnigan

bearing certain metaphysical views concerning the nature and status of an ontological entity ‘self’ (Pali atta, Skt. ātman) has on ethical theorizing. One of the central tenets of Buddhism is that there is no self (best known in its Pali form as the anatta doctrine). How best to understand this doctrine is subject to much dispute. For many, this doctrine constitutes the ontological thesis that there is no substantial ego underlying our experience – ‘we’ are nothing but a sequence of causally linked psychological and physical events and processes. Whether such events and processes themselves have substantive metaphysical status is a moot issue. For Siderits (2003), the anatta doctrine is not only an ontological thesis (which, he argues, commits us to a reduced level of events and processes with substantive metaphysical status); it also has direct implications for putative Buddhist ethical theorizing. In particular, Siderits argues that the anatta doctrine rules out theorizing Buddhist ethics by analogy with contemporary Aristotelian virtue ethics. His reason for this is the view that virtue ethics presupposes a virtuous agent as bearer of particular virtues. This, he takes it, can be reduced to a metaphysical commitment to a ‘self,’ which Buddhists would deny. Keown (2001), in stark contrast, defends a virtue ethical model of Buddhist ethics. He argues that Buddhism provides sufficient criteria to allow the individuation of subjects (as ‘persons,’ perhaps), which, on his view, is all that is required by his Buddhist virtue ethical theory.5 For Keown, pursuing the issue of the ulti5 Indeed, Siderits himself acknowledges a distinction between the concept of ‘person’ and that of ‘self;’ whilst Buddhists clearly reject the latter, Siderits argues that they need not deny the pragmatic utility of employing the former. Despite these concessions, I believe that Keown’s particular exposition of Buddhist virtue ethics does, indeed, commit him to a substantive account of the nature of self (above and beyond that of merely individuated ‘persons’). This is because Keown conceives of action as essentially involving choice which, though not necessarily problematic in itself, is defined by reference to that “core of the personality” which, Keown maintains, is the final resort for explanation of action (2001: 221). How consistent Keown is in endorsing this view is arguable. However, while Keown may be committed to a substantive view of the self, it does not necessarily follow that all theorizations of Buddhist ethics in virtue ethical terms will necessarily be so

Buddhist metaethics

271

mate ontological constitution of persons as part of a study of ethics confuses ethics and metaphysics and, hence, “does not make for a fruitful line of enquiry” (19). Moreover, he protests that “although facts are not irrelevant to values they have no priority, and ethical issues must be addressed with ethical arguments: they cannot be brushed aside by reference to facts of a scientific, ontological or metaphysical nature” (162). Metaethics certainly should not be pursued or conceived of as merely the joint pursuit of ethics and metaphysics (or science or epistemology). Nor should it be conceived of as a simple-minded attempt to derive evaluative conclusions from purely factual considerations (nor, indeed, to ‘demonstrate’ the impossibility of ethics by citing the failure to provide such a derivation). Keown is right to protest that this would lead to a number of confusions. However, as Siderits’ challenge makes clear, legitimate and important questions arise, and need to be addressed, concerning the relationship between metaphysical and epistemological considerations and ethical theories, questions which, in my view, constitute the proper domain of metaethics. Ethical theories are not simply collections or lists of values, they are attempts to explain the nature and role and relationships between these ethical values and thoughts and practices within the wider context of the aims and projects of the Buddhist philosophical tradition. As such, they take for granted various epistemological and ontological views and commitments. If these commitments conflict with ‘received’ Buddhist theories on the nature of epistemology and ontology, the fact of such conflict seriously challenges the status of the putative Buddhist ethical theory. Having said this, however, Keown is surely right to point out that the mere presence of such conflict does not, in itself, automatically confer priority to the epistemological and ontological theories; we do not simply ‘brush ethics aside’ in the face of opposition. Arguably all Buddhist lineages recognize there to be an important relationship between the practical and theoretical domains; between ethics (Pali sīla, Skt. śīla) and compassion (karuṇā) on the one hand, insight or wisdom (Pali paññā, Skt. prajñā ) on the
committed.

272

Bronwyn Finnigan

other. Given that there is no definitive explanation of the nature of these relationships, why should we assume from the outset that when push comes to shove ethical theories and considerations will, and should, give way? What, then, is the nature of the relationship between Buddhist theories on metaphysics and epistemology (i.e. the insight and wisdom traditions of Buddhist thought) and ethical theorizing? In this paper, I shall investigate certain aspects of this relationship. I shall not seek a definitive and comprehensive answer. Moreover, I shall not proceed by addressing this relationship in terms of the bearing certain metaphysical views about the nature of self may have on certain forms of Buddhist ethical theory. Rather, I shall focus on the bearing certain epistemological considerations concerning the possibility of ‘action’ have on Buddhist ethical theorizing. My point of departure shall be Keown’s highly influential virtue ethical theorization of Buddhist ethics (2001). I shall discuss this ethical theory in relation to a contemporary exposition of two prominent Buddhist epistemological theories, namely, Dunne’s exposition of the views of Dharmakīrti and Candrakīrti (1996). I shall highlight certain points of conflict between these ethical and epistemological theories and will argue that the resolution of this conflict requires revision (either in interpretation of theories or in the theories themselves) by all parties. I shall conclude by arguing for substantive revision to these theories via an engagement with this conflict; in so doing, I hope to exemplify some of the virtues of engaging with a metaethical methodology for the advancement of the respective domains of inquiry.

1. Keown’s (virtue ethical) nature of Buddhist ethics
Fundamental to Keown’s virtue ethical theorization of Buddhist ethics is the idea that Buddhism is centred on a teleological goal or summum bonum of human endeavour, namely, nirvāṇa (or, more specifically, nirvāṇa-in-this-life).6 Crucial to this view is the idea
I follow Adam in using Sanskrit words (i.e. nirvāṇa, śīla, karman) in place of the Pali (i.e. nibbāna, sīla, kamma). One exception, however, is that
6

Buddhist metaethics

273

that the perfection of ethical conduct (śīla), together with the perfection of knowledge or insight (paññā), is jointly constitutive of nirvāṇa. On almost all accounts, a buddha is the paradigm of one who has achieved nirvāṇa. It follows from Keown’s account that buddhas not only achieve intellectual perfection but engage in ethical conduct; indeed, the perfection of ethical conduct. Moreover, for Keown, the moral precepts that are presented in Buddhist treatises “circumscribe the conduct of the Buddha” (54). That is, the moral precepts of ethical action developed in the early treatises take descriptions of the Buddha’s behaviour as their paradigm. These moral precepts, in turn, form the basis for the preceptual codes common to both the Theravāda and Mahāyanā traditions. “To observe the precepts, therefore, is to model one’s behaviour on that of the Buddha” (31). More specifically, to pursue the goal of ethical perfection, on Keown’s account, is to pursue the goal of acting as the Buddha would act. “The Buddha’s śīla, or moral perfection, becomes an essential goal for all who aspire to his status” (55). For Keown, moral precepts serve a dual function; they both encapsulate “condensed descriptions” (54) of how the historical Buddha did act as well as provide a model for how a buddha would act (or, how we would act were we to attain nirvāṇa). Thus, for instance, given that compassion (karuṇā), generosity (dāna) and courage (vīrya) are all considered to fall under the “umbrella term” śīla (19), and “śīla circumscribes the conduct of the Buddha” (54), it follows that the conduct of the historical Buddha not only instantiated these virtues or qualities (i.e. he acted compassionately, generously and courageously) but that buddhas, more generally (i.e. those who attain nirvāṇa) also act in ways that are compassionate, generous and courageous. This introduces an important element of normativity into Keown’s theory. Insofar as we seek to attain nirvāṇa, and a partial constituent of nirvāṇa involves acting

I shall retain the original Pali form of paññā rather than its Sanskrit form prajñā. This is because Keown argues for conceptual differences between these two terms and, thus, his claims with respect to one will not automatically transfer to the other.

274

Bronwyn Finnigan

in ways that instantiate these qualities or virtues, we should (or ‘ought’ to) attempt to act in these ways. Crucial to Keown’s theory is the rejection of transcendental accounts of buddhahood and nirvāṇa. For Keown, nirvāṇa is the highest and best form of human life and, hence, the Buddha achieved the “fulfillment of human potential, not its transcendence” (113); he “lived an exemplary moral life” (75) with merely a difference in degree, and not in kind, of cultivated ethical goodness from one who is still following the path. What distinguishes a buddha from ordinary fallible beings following the path is not the transcendence of human activity, but the fact that his “moral conduct, that is to say his interaction with other beings, is perfect” (114). The idea that a buddha’s actions or behaviour is to be characterised in much the same way as followers of the path (albeit with a profoundly higher degree of perfection) is controversial. Much recent debate has focused on Keown’s claim that every virtuous action (including that of a buddha who has attained nirvāṇa) is both kuśala (i.e. morally good or skillful) and puṇya (i.e. karmically meritorious) (123). One locus of controversy concerns the canonical view that an arhat (i.e. a liberated or perfect being) is free from, or has passed beyond the domain of, karma and rebirth. How can a virtuous action be considered both kuśala and puṇya while the agent of the action has passed beyond the domain of karma, puṇya and apuṇya? Keown (2001) responds to this challenge by arguing that puṇya is a function of progress towards kuśala; once kuśala is achieved, puṇya becomes redundant (124). According to Velez (2004) this response contradicts the claim that virtuous action, by definition, is both kuśala and puṇya (given that the virtuous action of a buddha is not puṇya). Nonetheless, Velez can be seen to endorse Keown’s putative solution as an important correction to his view. For Velez, actions with the quality of puṇya are merely instrumental towards nirvāṇa whereas actions with the quality of kuśala are genuinely constitutive. Significantly, this dispute is conducted on the basis of a common assumption that a buddha (or arhat; i.e. one who has attained nirvāṇa) has the capacity to act. The debate concerns what quali­ ties can be predicated of a buddha’s actions (i.e. whether kuśala

Buddhist metaethics

275

and/or puṇya) not whether or not action is possible. Adam (2005) attempts to defuse this debate by pointing out that the term for ‘action’ in the Buddhist canon is karman (Pali kamma). Given that the arhat is considered to have reached the goal of having “destroyed action (kamma),”7 Adam argues that an arhat cannot be considered to act at all.8 According to Adam, an arhat’s “enlightened conduct or awakened activity” (77) falls entirely outside the scope of kuśala and puṇya. Note that while Adam denies arhats the capacity for ‘action’ (karman), he nonetheless acknowledges that they engage in “good conduct” (76) and “activity” (77). Thus, even with Adam’s qualification, it remains the case that the debates that arise in response to Keown’s account concern the kind or quality of a buddha’s conduct (i.e. whether or not such conduct is karmically efficacious or, in Adam’s terms, whether or not such conduct counts as ‘action’). However, such debates do not question the underlying assumption that it is possible for a buddha or arhat to engage in conduct or activity (i.e. action that is not karmically efficacious). In what follows I shall focus on this underlying assumption and shall set aside the issues of whether or not a buddha’s conduct is karmically efficacious; whether or not it should be termed conduct or activity or ‘action;’ whether or not kuśala or puṇya can be predicated of such conduct or action.

2. Keown continued…
As mentioned, Keown’s theory of Buddhist ethics is committed to the view that a buddha has the capacity to engage in ethical conduct insofar as ethical conduct is constitutive of nirvāṇa. This theory is also committed to the idea that a buddha’s capacity for ethical
This is a quote from the Kukkuravatika Sutta in the Majjhima-Nikāya, i, 390. 8 This argument depends on the close etymological connection between karman/kamma, as the term used for ‘action,’ and karman/kamma as denoting the cycle of saṃsāra. Note also that this is not a complete characterisation of Adam’s argument. At the core of Adam’s argument is the idea that the terms bright (sukka) and dark (kaṇha) can be employed to “bridge the conceptual gap between puñña and kusala.” This issue, like that of puṇya and kuśala proper, is outside the scope of this particular paper.
7

276

Bronwyn Finnigan

conduct is characterised in much the same way as that of a follower of the path, albeit with a higher degree of perfection. How does Keown characterise this assumed capacity for ethical conduct (and action, more generally) which underlies his theory?9 There is no simple answer to this question. The dominant thread of Keown’s argument focuses on an account of ethical conduct as requiring choice. This notion is first introduced in the context of discussing the Buddha’s decision to teach the dharma. According to the Vinaya, after the Buddha attained enlightenment he was asked three times to teach. It is only when he is asked for the third time that he “surveys the world with his Buddha-eye out of compassion (kāruññatā) for beings” and decides to teach rather than remain silent (Vin 1.6). Keown generalizes from this event the idea that the Buddha’s “initial hesitation” (during which time the Buddha “recognized alternative conceptions of human good”) and “subsequent decision” to teach represent a “new scale of values introduced by Buddhism” (42). Ethical conduct, on this ‘new’ view, paradigmatically involves choice and choice requires the recognition of alternatives. The idea that ethical conduct involves recognition of alternatives and choice resurfaces in Keown’s later discussions of the notion of cetanā. Although, again, there are many threads to Keown’s discussion of cetanā, his focal argument centres on the view that “the faculty of moral choice […] for Buddhism is cetanā […] it is the pivot around which virtue and vice revolve” (195). Indeed, Keown considers cetanā, defined as capacity for choice, to be that which determines the moral quality of an action. Keown’s argument for this idea turns on an appeal to Abhidharmic literature. For Keown, virtues and vices are dharmas;
9 Henceforth I shall use the terms ‘conduct’ and ‘action’ interchangeably insofar as conduct does not have a present perfect form which leads to clumsy expression. Notice, also, that I will assume that what holds for ethical conduct must hold for conduct/action insofar as ethical conduct/action is a subset of conduct/action proper (i.e. it is action of a particular kind, or action performed in a particular way, but action for all that). In this paper I shall focus on ethical conduct, but most of the argumentation can be extended to a concern with action, more generally.

Buddhist metaethics

277

they are real and objective elements “‘found’ within the psyche” (64). On his account, however, the virtues and vices do not, of themselves, motivate or cause action. Rather, they “influence specific moral choices and decisions” (214). Cetanā, for Keown, is the “preceding motivation […] which determines the moral quality of the action” (178); cetanā is the “compass-needle of moral choice” (211). A Buddhist virtue ethical theory, on Keown’s view, is thus concerned with “virtuous choices” (222); virtues are dispositions to “choose rightly” (207). Indeed, by the end of Keown’s argument it is no longer the case that the early Buddhist treatises describe the Buddha’s conduct; rather, they contain a record of the Buddha’s “important moral choices” (226). Significantly, the notion of cetanā defined as ‘choice’ is intimately connected with that of representing and considering alternative courses of action; there is no such thing as choice without alternatives between which one can choose. This is both recognized by Keown when he introduces the notion of a Buddha’s decision to teach as occurring subsequent to his recognition of alternatives, and theorized in his claim that cetanā is the outcome of a process that involves “initial attention to the matter in hand (vitakka), reflection upon it (vicāra), and an intellectual decision or resolution (adhimokkha)” (211). Indeed, Keown proceeds to offer cetanā as a suitable candidate for the Western philosophical term ‘practical reason’. According to this thread of Keown’s thought, cetanā comes to denote the entire process that generates action, starting from the initial intuition of a good end, subsequent deliberation (cetayitvā) about practical choices, and conclusion in choice and action (218). Insofar as cetanā is morally determinative, it follows that behaviour that is not produced via this “process of cetanā” is “not ethically charged” (220). It is especially important to notice that the above characterisations of ethical conduct require that a buddha engages in conceptual thought. The characterisation of cetanā in terms of ‘choice’ (what I call Keown’s ‘minimal conception’ of cetanā) requires that a buddha has capacities to represent and choose between alternative courses of action. A ‘course of action’ is a conceptualized mental item that is entertained or represented, rejected or

278

Bronwyn Finnigan

endorsed. The characterisation of cetanā as ‘process’ (what I call Keown’s ‘maximal conception’ of cetanā) requires that a buddha also has capacities to reflect, compare and deliberate about these mental items. Both conceptions of cetanā presuppose conceptuality. Significantly, however, conceptual thought is held suspect by most Buddhist epistemological theories. Hence, there is reason to think that the epistemological commitments that underlie Keown’s theory conflict with the ‘received’ views of Buddhist epistemology.

3. A dilemma: Dunne’s thoughtless Buddha
In order to grasp the precise nature of the conflict that arises between the definitions of ethical conduct that underlie Keown’s ethical theory and Buddhist epistemology, we need to consider and address particular epistemological theories. ‘Buddhist epistemology’ is an umbrella term for a number of sophisticated and highly systematic theorizations. While there is much agreement among these theories (insofar as their proponents each conceive themselves to be consistent with the teachings of the Buddha) there is also considerable disagreement. Hence, in order to show that the presuppositions underlying Keown’s theory do, indeed, conflict with ‘received’ Buddhist epistemological views, we need to locate this disagreement in the context of particular epistemological theories. In what follows I shall focus on Dunne’s exposition of two highly influential epistemological theories (i.e. that of Dharmakīrti and Candrakīrti) and shall demonstrate the extent to which they do, indeed, come into conflict with Keown’s ethical theory.10 Both Dharmakīrti and Candrakīrti are part of an epistemic tradition that considers concepts (vikalpa, kalpanā) to be closely linked to ignorance (avidyā) and ignorance to be at the root of all suffering. Though the respective analyses of the relationship between concepts, ignorance and suffering vary, that there is such a
10 I shall only provide an overview of these theories and, even then, shall focus only on those aspects that are relevant to this discussion. In particular, I shall not address Dharmakīrti’s theory of apoha as putative solution to the possibility of inferential knowledge. For detailed discussion of this theory see Katsura 1984; Dreyfus 1997; Tillemans 1999.

Buddhist metaethics

279

relationship is a shared assumption. It is also important to recognize that both thinkers agree that conceptual thought and inferential knowledge has objects or referents in much the same way as perceptions are considered to have objects or referents. For Dharmakīrti, the objects of perception are considered to be unique and momentary particulars (svalakṣaṇa). These objects count as ‘real’ insofar as they can be individuated in terms of the performance of causal functions and are detectable by sense faculties. Unlike perception, however, the objects or referents of conceptual thought and inferential knowledge are not considered to be ‘real;’ they are words, concepts, universals (sāmānya), i.e. generally characterised phenomena (sāmānyalakṣaṇa), which have no inherent existence. As a result, Dharmakīrti considers such cognitions to be ultimately erroneous or distorted (bhrānta). This is not to undermine the practical benefits of engaging in conceptual thought and inference.11 The salient problem, however, is that, according to Dharmakīrti, conceptual thought essentially involves determining the apprehended intentional object as being a real particular when, in fact, the only real particulars are unique and momentary events in causal relations.12 In conceptual thought we combine things that are, in fact, distinct and we direct our anger (krodha), fear (bhaya), and craving (tṛṣṇā) towards these pseudo-objects and, thereby, perpetuate suffering. It is not merely the case that, in conceptual thought, we make a mistake about what there really is, “conceptual knowledge is not even capable of apprehending (particulars) just as they are because conceptuality is controlled by ignorance” (PVSV, 49; Dunne 1996: 532; emphasis added). Dunne (1996) observes that “for Buddhists of Dharmakīrti’s ilk ignorance is the beginningless source of suffering. And on Dharmakīrti’s view ignorance is also the mental mechanism – the internal cause – that compels people to lump things together conceptually and suffer thereby” (532). Hence, it is not merely conceptuality that is false; it is false because it is (or is rooted in) ig-

11 “Although all (conceptual cognitions) are confused, we still define some as instrumental and some as spurious.” (PVSV, 49) 12 For a discussion of this point see Tillemans 1999.

280

Bronwyn Finnigan

norance. As Dharmakīrti himself writes, “ignorance is defined as conceptuality. That is, ignorance is conceptuality. Ignorance leads one astray by its very nature” (PVSV, 49). In identifying ignorance with conceptuality, argues Dunne, Dharmakīrti implies that what is said of ignorance applies to conceptuality. Hence, just as ignorance is an essential cause of suffering, so too is conceptuality. It follows from this that “just as ignorance must be overcome to become a buddha, so too must conceptuality” (Dunne 1996: 533). Candrakīrti, like Dharmakīrti, also held the view that conceptuality is incompatible with buddhahood. Candrakīrti’s major work, the Madhyamakāvatāra, focuses on investigating the distinction between ‘conventional’ reality (vyavahāra, Tib. tha snyad; saṃvṛti, Tib. kun rdzob) and ultimate reality (paramārtha; Tib. don dam).13 Although Candrakīrti discusses this distinction in a number of ways, that which is relevant to the current discussion is the idea that things which exist in terms of ultimate reality are ‘truly real’ and things that exist in terms of conventional reality are only provisionally real (i.e. they do not have ultimate or true reality). To this extent, Candrakīrti’s view is seemingly compatible with Dharmakīrti’s epistemological distinction between the objects of perceptual knowledge, which are ultimately real, and the objects of inferential knowledge, which do not have ultimate reality but, nonetheless, have a provisional or conventional status given their important pragmatic role in ordinary, everyday discourse. For Dunne, Candrakīrti’s view is much more radical than that of Dharmakīrti insofar as Candrakīrti argues that the objects of perceptual knowledge (i.e. svalakṣaṇa), themselves, cannot be ultimately real. This is because svalakṣaṇas are defined in terms of their causal capacities or functions; the fact of having causal capacities or functions serves to individuate them as particulars. For Candrakīrti, the possibility of individuating particulars in terms of their definitive function presupposes that particulars have an essence. However, as Candrakīrti argues, that which arises from
13 I here follow Dunne in his citations of probable equivalences between Sanskrit terms and original Tibetan translations (only ‘probable’ insofar as it remains the case that the Madhyamakāvatāra and its commentaries are available only in Tibetan).

Buddhist metaethics

281

causes and conditions cannot have an essence. Hence, the objects of perceptual knowledge are only conventionally real and the only possible ultimate reality is essencelessness (niḥsvabhāvatā, Tib. ngo bo nyid med pa) or emptiness (śūnyatā, Tib. stong pa nyid ). According to Dunne, one implication of Candrakīrti’s extension of conventional reality to the objects of perception is that any experience of the world (i.e. whether it be of persons, tables, the color blue, or a sensation of pain) is necessarily an experience of an ultimately unreal, conventionally constructed thing. If we extended Dharmakīrti’s identification of ignorance and conceptuality to ignorance and conventionality, it would follow that conventionality, like ignorance, is a source of suffering. Hence, a buddha must not only abandon ignorance, he/she must also abandon conventionality in order to attain nirvāṇa. I shall not, here, critique Dunne’s exposition of these epistemological theories. Notice, however, that on the assumption that it is correct, these theories have significant implications regarding the cognitive capacities of a buddha. For instance, it follows from this exposition of Dharmakīrti’s view that a buddha does not engage in any form of conceptual thought. In particular, he does not (or cannot) engage in forms of deliberation or reasoning (which would require the possibility of entertaining, and relating conceptualized thoughts). Nonetheless, for Dharmakīrti, a buddha has the capacity for direct, sensory perception (pratyakṣa) insofar as the objects of perception are real. This is not the case for Candrakīrti, however. For Candrakīrti, on Dunne’s analysis, a buddha does not even have the capacity for perception insofar as the objects of perception are also conventional constructs and, hence, rooted in ignorance. As Dunne argues: “not only does such a buddha not see the ordinary things of the world, he does not even know ultimate reality because nothing at all occurs in a buddha’s mind. Indeed, it would seem that Candrakīrti’s buddhas do not know anything at all […] one might even conclude that such a buddha is simply dead” (1996: 548). While we need not go as far as Dunne in concluding that a Candrakīrtian buddha is (cognitively like one) dead, it should be clear that the implications of these epistemological theories for the possibility of a buddha’s capacities for action will be incred-

282

Bronwyn Finnigan

ibly steep. In particular, the theories of both Dharmakīrti and Candrakīrti, on the above exposition, would rule out the possibility that a buddha could act or engage in ethical conduct as characterised by Keown. Both Keown’s minimal and maximal characterisations of cetanā require that a buddha engage in conceptual thought as part of the process that produces action. Indeed, a buddha’s conduct does not count as ethical (it is not “ethically charged”) unless it is produced by just such a process. And, yet, according to the epistemological theories we have just canvassed, a buddha does not have the capacity to engage in such a cognitive process insofar as a buddha does not have capacities for conceptual thought. If we prioritize the conclusions of these epistemological theories, it seems to follow that Keown’s virtue ethical theory is not a plausible theorization of Buddhist ethics insofar as it requires that a buddha can act (or engage in ethical conduct) and is committed to a characterisation of action (or ethical conduct) which presupposes that a buddha has the very capacities that are denied him on epistemological grounds. Hence, we should reject Keown’s theory. However, if we alternatively prioritise Keown’s theory we could argue that the epistemological theories must be wrong in their characterisation of a buddha’s cognitive capacities (or lack thereof). Hence, we should reject the epistemological theories. How do we resolve this dilemma?14

One problematic way to approach the problem might be to appeal to the authoritative status of the disputants (a move not unfamiliar to the Buddhist commentarial tradition). That is, one could argue that we have a choice between accepting the views and theory of Keown, a contemporary Buddhist ethical theorist, or that of Dharmakīrti and Candrakīrti, Buddhist thinkers whose epistemological theories constitute the background of the entire IndoTibetan logico-epistemological tradition. This is not an uncommon move in Buddhist debates both contemporary and ancient. Such an approach would be fallacious, however; rejecting Keown’s theory for reasons concerning relative authority in the tradition would be an ad hominem and has no bearing on the actual substance of his theory, which is what is at issue. Certainly, appeals to authority in support of the truth of particular claims are legitimate argumentative moves (albeit with limitations). However, appealing to the (relative) lack of authority of a disputant to disprove their theory is fallacious.

14

Buddhist metaethics

283

In what follows, I shall argue that we can resolve this dilemma only if both parties of the conflict revise their positions in certain crucial respects. I shall argue that these revisions are substantive advances in both ethical and epistemological theorization. I shall conclude that this result reflects favorably on metaethics as a methodology for inquiry.

4. Resolving the dilemma
My first approach to resolving this dilemma involves a return to Keown’s virtue ethical theory of Buddhist ethics. The idea that puts Keown’s theory in direct conflict with the epistemological theories of Dharmakīrti and Candrakīrti is the idea that ethical conduct (and action, more generally) requires representation of alternatives and choice of a course of action as an essential part of the process that directly productive of action. Need a virtue ethical theory of Buddhist ethics conceive of ethical conduct in these terms? I think not. The first reason is phenomenological and stems from a reflection on typical actions of ordinary fallible beings. In the course of a typical day we do a multitude of things: we breathe almost continually; blink from time to time; walk from place to place; eat and drink; forget various things; talk to our friends. Some of these things do not count as intentional (e.g. blinking and forgetting). For many philosophers, it is only intentional, volitional action that genuinely counts as action.15 Such is true for many Buddhist thinkers. Indeed, cetanā is often translated as volition or intention. “It is intention/will (cetanā), O monks, that I call action (kamma); having intended/willed, one acts through body, speech or mind” (AN III, 415). Now, we might concede that action must be intentional action (or conduct that involves cetanā) for it to genuinely count as action or conduct (and not mere behaviour). However, it does not follow that this necessarily commits us to a phenomenology of reflection and choice as directly productive of action. For instance, I typically do not reflect and choose what I say before every utterance I make to my friends. Nonetheless, I (often) intend what I
15

See, for instance, Davidson 1980.

284

Bronwyn Finnigan

say when I say it. Similarly, I do not typically reflect and choose to brush my teeth before getting into bed, or to walk down the stairs to the kitchen in the morning, or many other instances of habitual behaviour. Of course I could, and sometimes do. But not always, and this is the point. Although it is beyond the scope of this paper to mount a substantive argument in defense of this view, it seems at least plausible to suppose that some, if not much, of our intentional action does not involve a phenomenology of reflection and choice directly prior to action. Now, a defender of Keown’s theory may respond to this by conceding that not all instances of intentional action involve conscious reflection and choice immediately prior to, and necessarily productive of, action. However, she may argue that Keown’s theory is only committed to the view that ethical conduct is to be characterised in this way. Virtues are dispositions to choose; an action only counts as virtuous if chosen in the light of the virtues. If we deny that reflection and choice are constitutive of the process that generates ethical conduct then there is no way for the virtues to get a grip on our behaviour and, hence, for conduct to count as ethical. It seems, however, that if we have already granted the possibility that intentional action can count as intentional without prior reflection and choice, why can’t we also grant the possibility that at least some instances of ethical conduct can be characterised in the same way? Consider, for instance, an ordinary, fallible person writing a monthly cheque to Plan International. Imagine that the person in question has been sponsoring a child for a number of years. Her income is healthy (she is the chair of a wealthy academic department and in no threat of redundancy); she is satisfied that Plan is managing her funds well and that her donation is genuinely contributing to the well being of the child. It is reasonable to suppose that this person does not need to represent alternatives and choose or form the intention to write her monthly cheque every time she writes the cheque; in many cases she simply sees that the payment is due (it might be marked on her calendar, for instance) and responds by writing the cheque. One might say that the intention to send a donation every month was formed way back when she originally chose to sponsor a child and does not have to be remade every time

Buddhist metaethics

285

she writes the cheque. Writing a cheque for Plan has become a habit, one that she fully endorses. Moreover, even though the person does not represent alternatives and form the intention to send the donation every time she writes a cheque for Plan, if queried, she would freely endorse the action as something she intended to do and could provide reasons that explained why she did it. It does not follow, however, that these reasons and intentions need to have been mentally represented and formed as part of the process that directly produced the action (writing a cheque) for the action to count as intentional. Most importantly, if the act of writing a monthly cheque for Plan genuinely counts as virtuous, this is not because the person was motivated to choose to act in this way. It counts as virtuous because the action, itself, instantiates the virtue in question. A virtue is, characteristically, a disposition for a type of action, not a disposition for a certain type of choice. Notably, if we analyse Keown’s own choice-model of ethical conduct, we can see that it already presupposes that types of actions are characterised by virtues. On Keown’s account, virtues motivate or inspire or cause us to choose to enact virtuous courses of action. The paradigm object of choice is a course of action. Thus, action-X counts as virtuous, on Keown’s account, if we choose to do course-of-action-X (rather than course-of-action-Y). However, it would seem that we choose to do course-of-action-X (rather than course-of-action-Y) insofar as X counts as a virtuous course of action, and Y does not (or X counts as more virtuous than Y, or is more virtuous in this situation or is more relevant etc.). Virtues could only motivate choices if courses of action can be recognized to instantiate virtues. Hence, to use Keown’s example, compassion could only motivate the Buddha’s choice to teach the dharma if the Buddha recognized that this course of action, ‘teaching the dhar­ ma,’ counted as a compassionate action whereas keeping silent did not (or keeping silent was less compassionate or not compassionate given the intended audience, the situation, etc.). Of course what counts as an instantiation of a virtue will only be vaguely demarcated insofar as some action types will instantiate some virtues in some circumstances, but not in others, or at certain times, but not

286

Bronwyn Finnigan

others, or in certain ways, but not others.16 Hence, the identification of an action as instantiating a particular virtue will necessarily take a number of factors into consideration. However, if an account of virtuous choice already presupposes the idea of a virtuous type of action (which we can recognize as being virtuous in order to choose to act in that way and, hence for our action to count as virtuous) why can’t we simply dispense with the mediating act of choice and speak simply of virtues as being dispositions to perform certain types of actions or to respond in certain types of way (i.e. virtuously, compassionately, courageously), rather than merely dispositions to make certain types of choices to act in certain types of ways? That Keown seems to acknowledge some version of this response-account of virtuous conduct (in contrast to the choiceaccount of virtuous conduct) is arguably evident in his claim that “moral virtue is manifested by making the appropriate ethical response in different situations” (208) and that “generosity and the other virtues involve not merely the bare realization that a practice is good, but also the instantiation of the practice” (213). Keown also admits that the term cetanā does not merely designate the act of making a choice or decision but may be taken to “describe the general moral stance or posture adopted by the psyche and its orientation with respect to ends” (213). If we adopt a response-account of ethical conduct, we might interpret this claim as follows: the fact that a person consistently and reliably responds compassionately to certain kinds of situations (and, thereby, instantiates a disposition to respond compassionately), expresses their moral stance or the posture of their psyche and its orientation. Moreover, given that this moral stance or posture was cultivated while they were a follower of the path, and perfected once they became a buddha, one could reasonably argue that their compassionate responses are intentional; the fact that they can respond compassionately as they do (i.e. in the right way, at the right time, to the right object etc.) is the
16 For instance, giving a child some bread might be compassionate in one circumstance but not in another (e.g. if they are allergic); at one time but not another (i.e. if given before a child is about to run); in one way but not another (e.g. if thrown into the dirt at their feet).

Buddhist metaethics

287

result of a process of self-cultivation in which they have intentionally engaged. This position is consistent with the claim that actions involve cetanā, but does not thereby commit us to an account of a buddha explicitly representing options and making choices immediately prior to ethical conduct. Cetanā is still required for behaviour to count as action, but the scope of cetanā extends over the process of cultivating dispositions for virtuous response rather than constituting choices (or choosing) that necessarily produces the action. Hence, my suggestion is that we revise Keown’s choice-model of virtues and, instead, focus on a response-model of virtues. Virtues are dispositions for response, first and foremost, and not limited to dispositions for choice.17 If we accept this revision of Keown’s account, we are no longer committed to a view of buddhahood that requires a buddha to represent and choose their actions in order to be considered to engage in ethical conduct. Of course, it may have been necessary for them to represent and choose actions while they were a follower of the path in order to cultivate the dispositions that enable them to respond the way they do. Nevertheless, the perfection of these dispositions is instantiated in their direct and spontaneous ethical responses, their actions are not mere manifestations of represented courses of actions that have been chosen and enacted.

17 Note that the ‘response-model’ alternative to Keown’s choice model uses Keown’s presupposed definition of choice, which involves a phenomenology of representation, consideration and weighing up of alternative courses of action. It is an open question whether this is the best way to conceive choice and, hence, whether the response-model may be compatible with alternative conceptions. Choice is a function. It is unclear, however, whether or not the presentation and consideration of alternative courses of actions need be the relevant input and the performance of the function (i.e. choosing) need be a conscious mental event. It is the definition of choice qua deliberatively choosing, presupposed by Keown’s ethical theory, that is problematic and that I take as the ‘choice-model’ in contrast to the response-model advanced in this paper.

288

Bronwyn Finnigan

5. Resolving the dilemma (cont.)
Is this response-model of a Buddhist virtue ethics compatible with the epistemological conclusions of Dharmakīrti and Candrakīrti as presented by Dunne? If so, have we resolved our dilemma? If we consider the view of Candrakīrti, as interpreted by Dunne, the answer seems to be ‘no.’ While a buddha with capacities for virtuous response, as characterised above, need no longer engage in deliberation or represent courses of action as options for choice prior to response (and, hence, need not engage in conceptual thought, broadly construed), a buddha would still need to be able to perceive situations in order to respond to them. A Candrakīrtian buddha, on Dunne’s analysis, is not capable of perceiving anything and, hence, does not have the required cognitive capacities. A Dharmakīrtian buddha, however, is capable of perception. One might argue that if we opt for a Dharmakīrtian epistemology over that of Candrakīrti, then compatibility between our ethical theory and epistemology is assured and, hence, the dilemma resolved. Our response-model of Buddhist virtue ethics, on this view, is compatible with a Dharmakīrtian epistemology and, hence, retains its status as a plausible theorization of Buddhist ethics. Moreover, given an apparent standoff between the epistemological theories of Dharmakīrti and Candrakīrti, it might seem that the appeal to ethical considerations functions as a means for adjudicating this dispute. As promising as this might sound, however, things are not quite so straightforward. While Dunne’s Candrakīrtian buddha clearly does not have sufficient cognitive capacities to count as a buddha capable of ethical response, it is not entirely clear that a Dharmakīrtian buddha does either. A buddha, according to the response-model, does not merely require perceptual capacities; they need to be able to perceive and respond to situations or composite objects (e.g. disciples, suffering sentient beings etc.) which are, essentially, general. Even if these situations or objects of virtuous actions are not represented and, thereby, recognized as being objects of certain general kinds (i.e. a buddha does not need to think that the perceived object is a sentient being, least of all a suffering

Buddhist metaethics

289

sentient being, which would be required if, in order for his conduct to count as ethical, the buddha had to choose what to do), nevertheless, what is perceived is an object of a certain kind (i.e. the buddha sees ‘a suffering sentient being’) and it is that which stimulates the relevant kind of ethical response. A Dharmakīrtian buddha, however, merely has capacities to perceive particulars. What counts as a particular is a matter of much commentarial dispute (particularly in the Tibetan tradition).18 It seems, however, that if Dharmakīrti is consistent in maintaining that the objects of perception are free from all conceptuality then he must commit to a view of perceptual objects being unique, momentary and non-persisting events. Such objects, however, do not seem to be the proper objects of ethical response. Dharmakīrti might be seen to disagree with this analysis. He writes, “even though buddhas and bodhisattvas may act out of compassion for the sake of others, ‘they are faultless because they do not make false impositions’” (PVSV, 9; Dunne 1996: 539). Dunne interprets this to mean that buddhas do not engage in conceptuality insofar as they do not need to compulsively assume that the object of perception (i.e. the other; a disciple; a suffering sentient being) is anything more than a group of psycho-physical aggregates. However, even on this interpretation, it seems to be only when a group of psychophysical aggregates is seen as a particular object (i.e. a suffering sentient being) that a particular kind of compassionate response is stimulated. If seen as a different object (i.e. an enlightened being), a different kind of response is stimulated. Seeing a ‘group’ of psychophysical particulars as an object of a certain kind, however, essentially involves combining particulars. If the combination of particulars is essential to conceptuality and, thereby, bound to (if not identical with) ignorance, it follows that the relevant kind of perceptual objects for ethical response is bound to ignorance and, hence is beyond the ken of a buddha. Should we, thus, conclude that the dilemma still holds and that our response-model of Buddhist virtue ethics is incompatible with the demands of epistemology? The answer to this, I propose, turns
18

For extended discussions of this point see Dreyfus 1997; Dunne 2004.

290

Bronwyn Finnigan

on how we understand the notion of conceptuality in Dharmakīrti’s thought. Contemporary commentators seem to divide on what I will call a narrow and a broad interpretation of conceptuality in Dharmakīrti’s thought. Aspects of the narrow interpretation can be found in Dunne (2004, 2006). According to this view, conceptuality (vikalpa) merely involves the linguistic interpretation of ‘non-conceptual’ mental images generated by perception.19 That is, perception produces a mental image through contact with sensory objects (2006: 508). This mental image is non-conceptual but has phenomenal content. Conceptuality consists in the subsequent superimposition (samāropa) of an unreal conceptual image (ākāra) – such as sameness – onto the original image and, thereby, “compulsively imput[ing] sameness onto entities that are in fact not the same” (2004: 61). Conceptuality is tied to ignorance, on the narrow view, insofar as it involves the (erroneous) imposition of universals onto the non-conceptual images afforded by perception. Crucial to this view is the idea that the non-conceptual images generated by perception are not flawed. Rather “it is the perceiver’s inability to interpret the image, and not the image itself, that is causing the error” (2004: 87, my italics). If we pursue the narrow interpretation of conceptuality, it follows that a buddha has capacities for entertaining the non-conceptual mental images afforded by perception but does not super-impose conceptual constructs onto these images. If conceptuality simply involves the explicit thought or linguistic designation (prajñapti) of the affordances of perception, then this view of conceptuality may be compatible with the response-model of Buddhist virtue ethics. All that is required for the ethical responsiveness of a buddha is that she is able to perceive (and respond to) objects of certain kinds;
This might be thought to contradict Dunne’s earlier claim that “on Dharmakīrti’s view ignorance is also the mental mechanism – the internal cause – that compels people to lump things together conceptually and suffer thereby” (1996: 532). One might reply, however, that this claim identifies ‘lumping things together conceptually’ as the problem, not the mere act of ‘lumping things together.’ Whether this is a plausible reply, I shall not here seek to answer.
19

Buddhist metaethics

291

she does not also need to subsequently think or reflect or superimpose linguistic designations onto the objects of perception. If the non-conceptual mental images afforded by perception can count as objects of perception and, thereby, be sufficient for the stimulation of behavioural response (an equivalence that would need to be established) and if the formation of such objects does not involve conceptuality, understood as above, then it seems that our ethical theory may be compatible with a Dharmakīrtian epistemology and, thus, the dilemma may be resolved. The narrow interpretation of conceptuality in Dharmakīrti’s thought is not without challenge, however. On the broader interpretation, conceptuality plays a critical function for the very combination of multiple affordances of perception into a single image or object in the mind (which, on this account, occurs prior to superimposition or linguistic designation as conceived on the narrow view). As argued by Katsura (1984), Dharmakīrti’s particulars are momentary (kṣaṇika) and, as such, beyond the scope of perceptual awareness. What we ordinarily experience, however, are undifferentiated continua (santāna) of moments, or persisting perceptual objects, and not moments themselves (216). Conceptuality (vikal­ pa), according to the broader view, is the critical factor involved in unifying the unique, momentary perceptual inputs into persisting, perceptual objects. Given that conceptuality is bound to (or identical with) ignorance, it would seem that perceiving an object (let alone responding to it correctly) is, itself, beyond the ken of a buddha. Hence, on the broader interpretation of conceptuality, our response-version of Keown’s theory is not compatible with a Dharmakīrtian epistemology. The dilemma remains. Dharmakīrti allows for the possibility of ‘yogic perception’ (yogipratyakṣa). Could one not argue that a buddha has yogic perception and, thereby, the relevant cognitive capacities? Given recent exegeses of this notion, however, the answer seems to be ‘no’. If we follow Dunne’s analysis, for instance, the proper objects of yogic perception are truths; in particular, the four noble truths (2006: 497). While the perception of the four noble truths may be advantageous for a buddha, a buddha with capacities for virtuous response, as we have argued, also requires the capacity to perceive

292

Bronwyn Finnigan

objects of certain kinds in order to respond appropriately. Thus, yogic perception, on Dunne’s analysis, will not help. If we follow Katsura’s analysis, by contrast, the proper objects of yogic perception are the innumerable point-instants or moments of material objects and mental phenomena that elude the sense perception of ordinary fallible beings (1984: 216). While this account provides a buddha with (innumerable) objects of perception, it does not present them in a form that is consistent with an account of virtuous responses as generated by dispositions that were acquired through a process of cultivation.20 Again, it seems that an appeal to yogic perception, based on contemporary analyses of this notion, is not going to solve our particular problem.21 Significantly, Katsura (1984) points out an inconsistency in Dharmakīrti’s thought, which is relevant to our discussion. He notes that while Dharmakīrti identifies perception as a form of pramāṇa (i.e. a means/source of true knowledge) he nonetheless defines pramāṇa as instrumental to the ‘fulfillment of human purpose’ (arthakriyāsthiti)22 ‘human activity’ (pravṛtti) and ‘experience’ (222). Practical human activity, Katsura points out, requires perceptual judgment and determination (adhyavasāya). Indeed, without these, Katsura argues, perception would have no practical significance at all (226). Perceptual judgment, as characterised by Katsura, is similar to our notion of the objects of perception being of certain kinds. Katsura exemplifies this notion in terms of the perception of a woman as a skeleton (by a monk), as an object of lust (by a lustful man), as a nice dinner (by a dog). Where, for Katsura, perceptual
I discuss this point in more detail in Finnigan 2011a. The notion of yogic perception has only recently received the critical attention of contemporary Buddhist thinkers and, hence, there is reason to anticipate future interpretative variation. Whether or not these variations will be compatible with our response-version of Buddhist virtue ethics is yet to be seen. 22 “Pramāṇa is non-contradictory knowledge (avisaṃvādijñānam). Noncontradictoriness [here] means the existence of the fulfillment of a human purpose (arthakriyāsthiti).” (Pramāṇavārtika, Chapter 2, v1a‒c, as quoted in Katsura 1984: 219)
21 20

Buddhist metaethics

293

judgment is formed subsequent to the perceived image (i.e. the perceptual judgment ‘a nice dinner’ is formed after the perception of ‘a woman’), according to our response-version of Keown’s theory, the perception of ‘a woman’ is, already, the perception of an object of a certain kind. Despite this difference, it is notable that, on both accounts, it is the perception/perceptual judgment of an object of a certain kind that is the significant factor for the stimulation of conduct or response. For Katsura, perceptual judgment and adhyavasāya are required to prompt human activity (224). Moreover, Katsura argues that both of these notions are based in linguistic conventions (saṅketa, 226). Hence, Katsura may be read as presenting a dilemma for Dharmakīrti based on considerations that are internal to his own theory.23 Either Dharmakīrti needs to allow for certain forms of conceptuality to be involved in perception to satisfy his own pramāṇic criteria (i.e. such that perception can be considered instrumental to the fulfillment of human purposes), or he needs to omit instrumentality for the fulfillment of human purposes from his criteria of pramāṇa such that an entirely non-conceptual perception can be considered a form of pramāṇa. The above dilemma can be seen to mirror the dilemma posed by our ethical theory.24 On both counts, it seems that Dharmakīrti should either allow some rudimentary form of conceptuality into his account of perception (in order to properly capture the notion of a perceptual ‘object’ of response which is required for human activity, in general, and ethical response, in particular) or he should redefine conceptuality much more narrowly, possibly akin to the narrow interpretation of conceptuality, such that the perception (or perceptual judgment) of an object of a certain kind is no lonThis is not to make any claim about Katsura’s actual intentions in raising these concerns. 24 Further pressure is added by considerations concerning the possibility of a buddha teaching the dharma as discussed in Dunne 1996, Finnigan 2011a and 2011b. As Dunne points out, Dharmakīrti allows that Śākyamuni Buddha spoke. How can we make sense of this, however, if the Buddha does not employ concepts? Dunne attempts to resolve this inconsistency by arguing that a buddha perceives concepts (and, thereby, only employs perceptual knowledge). However, this does not seem to solve the problem insofar as a buddha would need to employ concepts in uttering speech acts.
23

294

Bronwyn Finnigan

ger identified with ignorance and, hence, no longer eliminated on the pathway to buddhahood. Either revision would require a much more sophisticated account of the relationship between conceptuality and perceptions. The upshot of this is that Dharmakīrti’s theory would not only be compatible with his own definition of pramāṇa (and thereby resolve the internal dilemma) it would also resolve the external dilemma that arises when Dharmakīrti’s epistemological theory is brought into dialogue with Buddhist ethical theory.

6. Conclusion
In this paper, I focused on the relationship between Buddhist epistemological theories and Buddhist ethical theorizing. In particular, I focused on certain epistemological commitments concerning the possibility of action that underlie one prominent theorization of Buddhist ethics (i.e. Keown’s Buddhist Virtue Ethics) and demonstrated the ways in which these commitments come into direct conflict with the epistemological theories advanced by two prominent Buddhist thinkers (i.e. Dharmakīrti and Candrakīrti). Rather than ethical theory being ‘brushed aside’ or collapsing under the pressure of the conclusions of epistemology (as warned against by Keown), I demonstrated a certain mutuality of revisionary pressure bearing on both parties of the conflict. On the one hand, I demonstrated that insofar as an ethical theory assumes certain underlying epistemological commitments, there is pressure on the theory to be revised in the face of conflict with the established positions of Buddhist epistemological theories. As I have argued, Keown’s virtue ethical account of Buddhist ethics can relevantly meet this pressure for revision if it alters from a choice-model of virtue and action to a response-model of virtue and action. Such a revision is both internally consistent with Keown’s account remaining a virtue ethical theory and also meets certain aspects of the epistemological demands. In order to completely satisfy the epistemological demands, I have argued, would require some revision on the part of the epistemological theories. Nonetheless, our proposed revision of Keown’s theory, from a choice-model to a responsemodel, is a substantial improvement in the theory. Hence, while the pursuit of metaethics may not have resulted in a perfect resolution

Buddhist metaethics

295

of the conflict that arises between ethical theory and epistemology, it has resulted in the advancement of our ethical theory and this is a significant outcome. On the other hand, insofar as there are internal conflicts with­ in each of the epistemological theories addressed in this paper as well as conflicts between these epistemological theories, I demonstrated that ethical considerations may be utilized both in the attempt to adjudicate these disputes as well as to add support to particular strategies for the resolution of internal conflicts within these theories.25 I argued that Dharmakīrti’s epistemological theory could relevantly meet the pressures of our revised responsemodel of Buddhist virtue ethics only if it revises its account of the relationship between perception and conceptuality (or, at least, if scholars can provide a more sophisticated analysis of their relationship). Whether or not this epistemological theory can be revised, or a more sophisticated analysis provided in a way that both meets the demands of our ethical theory and retains consistency with its other theoretical commitments, is beyond the scope of this paper to conclusively answer. One might argue, however, that the arguments raised by Katsura suggest there is already sufficient pressure from considerations internal to Dharmakīrti’s theory that the theory be revised in ways that would be compatible with the requirements of our ethical theory. Hence, while it may be arguable whether the requirements of our ethical theory, by themselves, provide sufficient reason to motivate revision, consideration of these requirements may be mobilized to support extant revisionary claims. The Buddhist tradition is singularly distinctive in its expressed value of an intimate relationship between ethics and insight, com-

25 In this paper we employed an indirect strategy to adjudicate the dispute between a Candrakīrtian and Dharmakīrtian epistemology, in Dharmakīrti’s favor, by appeal to the plausibility of their theories relative to the cognitive requirements of a buddha for ethical response. This strategy would only genuinely adjudicate these epistemological theories, it seems to me, if each theory were established as internally consistent and there being no further point of appeal within the ken of epistemological theorizing to demonstrate one theory as better than the other. That this is the case, for either of these theories, remains to be established.

296

Bronwyn Finnigan

passion and wisdom. This value, I believe, introduces a certain demand for mutual attentiveness to, if not drive for, compatibility between the respective domains of Buddhist thought. An epistemological theory that excludes the possibility of ethical conduct is just as problematic, in the context of Buddhist soteriology, as a theory of ethical conduct that requires that which is necessarily excluded on epistemological grounds. While in this paper I have only been able to highlight and put pressure on certain elements in prominent epistemological theories, I have demonstrated that the challenges raised by pursuing an investigation of the relationship between epistemology and ethics stimulates an important revision to one of the most prominent Buddhist ethical theories. This revision, I contend, constitutes a significant improvement to the theory and, hence, exemplifies one of the virtues in pursuing a metaethical methodology.

References Primary sources
PVSV Pramāṇavārttikam: The First Chapter with the Autocommentary. 1960. Serie Orientale Roma, Rome: Istituto Italiano per il Medio ed Estremo Oriente. Aṅguttara Nikāya. The Book of Gradual Sayings. 1932‒6. F. L. Woodward and E. M. Hare tr. London: Pali Text Society. Vinaya, The Vinaya Pitakam. The Book of Discipline. 1964. H. Oldenberg. London: Luzac.

AN Vin

Secondary sources
Adam, M. T. 2005. “Groundwork for a Metaphysic of Buddhist Morals: A New Analysis of puñña and kusala, in light of sukka.” Journal of Buddhist Ethics Vol. 12, 61‒85. Clayton, B. 2006. Moral Theory in Śāntideva’s Śikṣāsamuccaya: Cultivating the fruits of virtue. London: Routledge. Cooper, D. E. and S. P. James 2005. Buddhism, Virtue and Environment. Hants Burlington: Ashgate. Davidson, D. 1980. Essays on Actions and Events. Oxford: Oxford University Press.

Buddhist metaethics

297

Dreyfus, G. 1995. “Meditation as Ethical Activity.” Journal of Buddhist Ethics Vol. 2, 28‒54. _____ 1997. Recognising Reality: Dharmakirti’s Philosophy and Its Tibetan Interpretations. New York: State University of New York Press. Dunne, J. D. 1996. “Thoughtless Buddha, Passionate Buddha.” Journal of the American Academy of Religion Vol. 64(3), 525‒556. _____ 2004. Foundations of Dharmakīrti’s philosophy. Somerville, MA: Wisdom Publications. _____ 2006. “Realizing the unreal: Dharmakīrti’s theory of yogic perception.” Journal of Indian Philosophy Vol. 34, 497‒519. Finnigan, B. 2011a. “How can a buddha come to act? The Possibility of a Buddhist account of Ethical Agency.” Philosophy East and West Vol 61(1), 134‒159. Finnigan, B. 2011b. “A Buddhist Account of Ethical Agency Revisited: Reply to Garfield and Hansen.” Philosophy East and West Vol. 61(1), 183‒194. Goodman, C. 2008. “Consequentialism, Agent-Neutrality, And Mahāyāna Ethics.” Philosophy East and West Vol. 58(1), 17‒35. Hallisey, C. 1996. “Ethical Particularism in Theravāda Buddhism.” Journal of Buddhist Ethics Vol. 3, 32‒43. Harvey, P. 2000. An Introduction to Buddhist Ethics. Cambridge: Cambridge University Press. Katsura, S. 1984. “Dharmakīrti’s Theory of Truth.” Journal of Indian Philosophy Vol. 12, 215‒235. Keown, D. ed. 1998. Buddhism and abortion. Hampshire: Palgrave Macmillan. _____ 2001. The Nature of Buddhist Ethics. Hampshire: Palgrave Macmillan. Levin, J. 2004. “Functionalism.” Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/functionalism/, last visited 30-07-2011. Siderits, M. 2003. Empty Persons: Personal Identity and Buddhist Philosophy. Aldershot, Burlington: Ashgate. _____ 2007. Buddhism as Philosophy. Aldershot, Burlington: Ashgate. Tillemans, T. J. F. 1999. “On the So-called Difficult Point of the Apoha Theory” in Scripture, Logic, Language: Essays on Dharmakirti and his Tibetan Successors. Boston: Wisdom Publications. Velez, A. 2004. “The Criteria of Goodness in the Pāli Nikāyas and the Nature of Buddhist Ethics.” Journal of Buddhist Ethics Vol. 11, 123‒142. Williams, P. 1998. Altruism and Reality: Studies in the Philosophy of the Bodhicaryavatara. London: Curzon.

On the auspiciousness of compassionate violence
Stephen Jenkins

In light of the overwhelming emphasis on compassion in Buddhist thought, Buddhist sources that allow for compassionate violence have been referred to as “rogue sources” and equivocations. A recent article states that, “Needless to say, this stance [that one may commit grave transgressions with compassion] is particularly favored by the Consciousness-Only school and in esoteric Buddhism.”1 However, the same stance is presented in the Mādhyamika tradition by Bhāviveka, Candrakīrti, and Śāntideva, as well as in a variety of sūtras.2 Allowances for compassionate violence, even killing, are found among major Buddhist thinkers across philosophical traditions and in major scriptures. It is also remarkable how broadly influential a singular source like the Upāyakauśalya-sūtra can be. This paper reflects on the question of whether killing can be auspicious in Mahāyāna Buddhism with secondary reflections on the problems that arise in attempting to apply Western metaethical categories and modes of analysis.3 Studies so far have been reluctant to accept that compassionate killing may even be a source of making merit, choosing instead to argue that even compassionate killing has negative karmic consequences.4 If it is true that the
Kleine 2006: 80. Nāgārjuna and Āryadeva do not explicitly allow for deadly violence, but do allow for inflicting pain or performing normally inauspicious action based on intention. 3 This is a preliminary report on one dimension of a long-term research project, on compassionate violence. Other dimensions of this study of Buddhist ethics of violence will include a reappraisal of Aśoka’s edicts, comparison with the Dharmaśāstras, “mainstream” and abhidharmic traditions, tantric ethics, and the violence of warfare. See also Jenkins 2010. 4 See Harvey 2000: 135–138.
2 1

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 299–331

300

Stephen Jenkins

compassionate bodhisattva killer takes on hellish karmic consequences, then it would seem that this is an ethic of self-abnegating altruism. Buddhist kings would seem to be in an untenable ideological situation in which even the compassionate use of violence and deadly force to maintain order and security will damn them to hell. Buddhist military and punitive violence, which has historically been a consistent feature of its polities, often including monastic communities, appears to be radically and inexplicably inconsistent with the values expressed by its scriptures and inspirational figures. If there are negative karmic consequences to compassionate killing, then these acts must be read at best as necessary or “lesser evils.” However, altruism and negative karmic consequences rarely go together in Buddhist thought. A review of the remarkable spectrum of great Buddhist thinkers who have discussed this issue, many of them with reference to the Upāyakauśalya-sūtra, shows general agreement that compassionate violence can be an auspicious merit-making opportunity without negative karmic consequences. Since I started working on this issue, which was integral to my doctoral dissertation, others have written on compassionate violence basing their thoughts primarily on Asaṅgaʼs Bodhisattvabhūmi and Mahāyānasaṃgraha, and the Śikṣāsamuccaya and Bodhicaryāvatāra attributed to Śāntideva. Building on the pioneering work of Mark Tatz, I am going to add examples from Candrakīrtiʼs commentary on Āryadevaʼs Catuḥśatakam, and examine the views of Bhāviveka brought to light by David Eckel’s recent work.5 I also highlight some overlooked details of the Upāyakauśalya-sūtra, which has been misread on this issue, and take a fresh look at Asaṅga’s foundational work in the Bodhisattvabhūmi.6
5 Special thanks to David Eckel for directing me to Bhāviveka’s treatment in his then unpublished translation. 6 Any merit of this work is largely due to Dr. Sangye Tandar Naga, of the Library of Tibetan Works and Archives in Dharamsala, and the scholars of the Central College of Higher Tibetan Studies, particularly Venerable Lobsang Dorjee Rabling and Professor K. N. Mishra. Geshe Ngawang Samten kindly granted me free housing during an extended research period at what was then called the Central Institute for Higher Tibetan Studies. I am

On the auspiciousness of compassionate violence

301

Auspicious killing
Without question the writing of Asaṅga has been one of the principal sources on the ethics of compassionate killing both within Mahāyāna traditions and the academic study of religion. He describes a hypothetical situation in which a bodhisattva observes a thief about to commit a mass murder of persons of the highest moral status for the sake of a pittance. Although he does not directly cite the sūtra, Asaṅga is almost certainly referring to the Ship Captain Jātaka of the Upāyakauśalya-sūtra, the focus for most of the discussion of compassionate killing.7 Killing people such as arhats, bodhisattvas and pratyekabuddhas is the worst kind of murder, which results after death in immediate rebirth in a hell realm. There are no intermediate rebirths in which to moderate the effects of such terrible crimes. They are known in Buddhist traditions simply as the “immediates.” The bodhisattva, seeing this imminent tragedy, realizes that if he kills the thief then he himself may go to hell. But he decides that it is better that he go to hell than allow this person to suffer such a fate.8
With this attitude, the bodhisattva, having discerned either a neutral or auspicious mind; regretting and employing a mind of empathy alone, then takes that living being’s life. [That bodhisattva] becomes blameless and produces abundant merit.9

also indebted to Professor Premasiri Pahalawattage for guiding my reading of many of the Pāli sources. 7 The commentator, Jinaputra, merely says that the argument is the same as in the sūtra (Tatz 1986: 323). 8 My use of the masculine pronoun is based on the fact that Asaṅga speaks in ways that assume the bodhisattva is male, for instance in discussing sexual transgression. 9 BoBh 113.24–114.2. evam āśayo bodhisattvas taṃ prāṇinaṃ kuśalacitto ’vyākṛtacitto vā viditvā ṛtīyamānaḥ anukampācittam evāyatyām upādāya jīvitād vyaparopayati  / anāpattiko bhavati bahu ca puṇyaṃ prasūyate  /; cf. BoBh–Wogihara 166; Peking Bstan-ḥgyur, Sems-tsam, Shi, 100b3; Tatz 1986: 70–71; note that upādāya has a strong idiomatic relationship with anukampācittam.

302

Stephen Jenkins

In the discussion of a variety of cases of compassionate ethical transgression, Asaṅga drives his point home by repeatedly closing with the final phrase expressing the bodhisattva’s faultlessness and generation of abundant merit, anāpattiko bhavati bahu ca puṇyaṃ prasūyate, a total of nine times. One aspect of his description is not entirely clear. He describes the bodhisattva, before the act of killing, as observing a mind that is either auspicious, kuśala, or neutral, avyākṛta [often translated as “indeterminate”].10 This refers to a common abhidharmic classification that distinguishes between auspicious, neutral and inauspicious states of mind. Only the last are affected by the kleśas, attraction, revulsion, and delusion, and so have negative karmic outcomes.11 There is disagreement in both modern and classical scholarship about whether this represents a concern for the killer’s state of mind or the victim’s.12 Both interpretations have some
In a highly recommended article, Rupert Gethin elaborates the Theravādin view that killing can never be based on auspicious, kuśala, or neutral, avyākṛta, states of mind. Therefore killing can never be based on compassion, nor can it be auspicious. The key example is of a king who seems to take pleasure in ordering the execution of a criminal. On a subtle level, the commentaries say his mind is still qualified by aversion. However, all killing is not equally inauspicious; he also shows the broad range of conditions that qualify an act of killing, including the moral status of the victim (Gethin 2004). 11 See Rahula 2001: 149, n. 169, avyākṛta defined; see 49 on a mind neither bad nor pure; Holt 1981: 80. Referring to this threefold division, he discusses how actions which are not affected by the kleśas do not have karmic outcomes; see Samyuktābhidharmahṛdaya i.153–154, on three types of citta-samprayukta, i.e. avyākṛta, kuśala and akuśala. He says that “neutral awarenesses [citta] are weak and only strong awarenesses can produce bodily and verbal action.” The context is unclear and there is disagreement on this theme, but if this point were generally agreed, then the concern for avyākṛtacittasamprayukta could not be related to the compassionate killer, since killing would not be possible from a neutral perspective. 12 Bhāviveka, discussed below, identifies the concern with the bodhisattva’s state of mind (Eckel 2008: 188); Tsong-kha-pa notes disagreement, but without identifying the sources, and states that it makes no sense to attribute it to the bandit (Tatz 1986: 215); Paul Demiéville reads it as a concern for the bandit’s sake, perhaps based on the Chinese (1973: 379); in an expansive article that should be the starting point for all interested in these issues, Lambert
10

On the auspiciousness of compassionate violence

303

merit. Regarding the thief, the concern would be to assure that the victim die in at least a morally neutral moment. Rebirth has a strong relationship to a person’s dying thought, maraṇacitta. Out of compassionate concern for the victim, one would attempt to take their life in a positive or neutral state, rather than in an inauspicious one. Killing a murderer while they are in a moment of homicidal rage would defeat the purpose, because death in that state of mind would lead to a bad rebirth. On the other hand, if the compassionate killer did not maintain at least a neutral state of mind, they themselves would go to hell. By affirming that they generate abundant merit, Asaṅga makes clear that the bodhisattva acts with an auspicious intention. One would not expect a neutral intention to result in great merit or for Asaṅga to advocate killing with a neutral, rather than auspicious, intention. The Upāsakaśīla-sūtra, though not cited by our sources, puts this in striking terms.
Someone may say that one commits an offense of killing whether one’s mind is good, bad, or neutral, just as anyone who is burnt by fire or takes poisons will die even if the mind is good, bad, or neutral. Such an argument is not true. And why is it not? Just as some people in the world do not die even if they are burnt by fire or drink poison, so one who kills without a vicious mind does not commit the crime.13

Jinaputra’s commentary on the Bodhisattvabhūmi, from several centuries after Asaṅga, indicates that the bodhisattva is concerned with his own state of mind.14 This suggests a concern for the compassionate killer’s own karmic wellbeing. This may seem inconSchmithausen also identified the concern with the bandit’s state of mind. He gives preference to manuscripts that support this, but without explanation (1999: 59 and n. 67); Tatz also notes differences in Sanskrit manuscripts, but follows the commentary of Jinaputra which appears to identify the concern with the bodhisattva’s state (Tatz 1986: 326 and n. 403). 13 Shih Heng-ching 1994: 171. For an extended discussion of killing see Chapter XXIV. The text was translated into Chinese in the early fifth century. Its origins are unclear, but Paul Groner takes it as an “authentic Indian source.” See Groner 1990: 244; for a similar discussion using the same metaphors in an abhidharmic source see Samyuktābhidharmahṛdaya i.188–189. 14 See Tatz 1986: 326.

304

Stephen Jenkins

sistent with the bodhisattva ideal, but all the treatments of compassionate killing show a strong concern for the protection and benefit of the killer. This is a crucial point for understanding what often appear to be acts of self-abnegating altruism. The benefit of both self and other is one of the strongest themes in Buddhist thought. In the “Benefit of Self and Other Chapter,” Svaparārthapaṭala, of the same text, Asaṅga explicitly rejects an intention that is strictly interested in benefiting others as inferior to one that benefits both self and others.15 Using an old model from the Nikāyas, he divides the possibilities into four: interest in benefiting nobody, only oneself, only others, and both oneself and others.16 Being interested only in others is superior to being interested in only oneself or being interested in nobody’s well being. But being interested in the benefit of both self and other is best, since developing oneself is necessary in order to have the ability to benefit others. This is simply expressed in the bodhisattva’s vow to attain the pinnacle of self-empowerment, buddhahood, for the sake of benefiting others. A circularity between the benefit of self and others is evident in the fact that it is only through helping others that a bodhisattva can accumulate the vast merit required to attain buddhahood. This relationship can become highly ironic, as the benefit of self and other are profoundly interrelated. As Śāntideva famously put it:
… upon afflicting oneself for the sake of others, one has success in everything. The desire for self-aggrandizement leads to a miserable state of existence, low status, and stupidity. By transferring that same desire to someone else, one attains a fortunate state of existence, respect, and wisdom. …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.17

As I have discussed at length elsewhere, although compassion should be disinterested, it is also regarded throughout Buddhist
15

BoBh 21–22; Jenkins 2003: 57–62. AN ii.95; tr. Woodward 1933: 104. This is a common theme. For crossreferencing and commentary see Jenkins 2003: 56–64. 17 Bca Chapter VIII, verses 126–9; Wallace and Wallace 1997: 105–106.
16

On the auspiciousness of compassionate violence

305

traditions as highly beneficial to the agent, providing karmic and even physical self-protection.18 The passage above is naturally and correctly read as encouraging suffering for the sake of others. On the other hand, it presents this as a key to happiness, status, respect, good rebirth and wisdom for the agent. Self-benefit is based on benefiting others. This can easily lead to misreading, since the most dramatic examples of self-sacrifice, feeding oneself to a starving tigress or offering one’s head, are also incredibly beneficial to the apparent martyr. So when the bodhisattva-killer takes care to have only empathy, anukampā, as he performs the action, he is concerned to protect both himself and his victim from falling into the hell realms.19 The concern for states of mind also bears on another difficult point that bears on the attitude toward this action. I translate Asaṅga above as saying that the bodhisattva is “regretting” as he kills. The Sanskrit term here is ṛtīyamānaḥ, and it has been previously understood either as “full of horror”20 or “feeling constrained.”21 The object of negative emotion is unclear, and surely the intention is to express that the situation is regrettable. Demiéville loosely, but elegantly, renders this as “full of both horror for sin and mercy for the sinner.22 But, in English at least, “horror” is too strong here
See Jenkins 2003, 2010. The Mahāvaṃsa gives an amusing example from the Sri Lankan myth of origin. Sihabahu drew his bow to murder his father the lion, progenitor of the Sinhalas. But, because the sight of his son aroused affection in the lion, his arrows only bounced off. It is only after he realized what was happening that: “Anger weakened his compassion and made him vulnerable. The third arrow pierced his body and killed him” (tr. Geiger 1986: 53). 19 Anukampā is a common substitute for both maitrī and karuṇā. When specifically defined, it signifies emotional sensitivity to the suffering of others. 20 For extensive notes on this obscure and difficult term, see artiyati in Edgerton 1985. It can be understood as meaning ‘grieved,’ ‘pained,’ ‘perturbed,’ ‘disgusted,’ ‘offended,’ also, when used as a noun, as meaning ‘shame,’ ‘humility,’ ‘distress’ etc., often in regard to morality. 21 Tatz 1986: 70. 22 “plein à la fois d’horreur pour le péché et de pitié pour le pécheur” (Demiéville 1973: 379).
18

306

Stephen Jenkins

and an equivalent for “sin” does not occur in the text. If the bodhisattva were experiencing horrific revulsion, then this could not be a merit-making, and therefore auspicious, action. Tatz apparently takes ṛtīyamānaḥ/ʼdzem bzhin du” as “feeling constrained,” with the sense of moral constraint, or lacking options.23 There is no question that in all accounts the Ship Captain Jātaka is framed in a way that makes killing a last resort. So this does no violence to the meaning.24 However, I use “regret” with the purpose of relating a more literal meaning of the Sanskrit to the English idiomatic sense of regretting what one must do or that something is one’s painful duty. All this emphasis on intention and states of mind sometimes leads to an exaggeration of its importance in Buddhist ethics to the point of claiming that killing is the mental intention to kill. In fact, all sources seem to agree that for there to be killing, there must be an actual living being, and the intention to kill must result in death. Unintentional killing is not murder, but the intention to kill alone does not entail the karma for murder.25 The moral status of the murder victim also has a crucial effect on the karmic repercussions. This is generally true in Indian thought. In the Manusmṛti, for instance, killing an untouchable may have no more karmic cost than killing an animal.26 So Asaṅga presents this as an extremely dangerous situation for the bandit by describing the people he is about to murder as bodhisattvas, pratyekabuddhas, etc., i.e. persons of the highest moral quality.
In this passage, Tatz translates “feeling constrained.” The same Tibetan verb occurs twice more in the root text and in the commentary. There Tatz first translates it as “embarrassed” and the second time as “feeling constrained.” Tatz 1986: 72, cf. BoBh 115.22 and Tatz 1986: 75, 222, cf. BoBh 117.16. In both cases the meaning associated with the Sanskrit is more fitting. The same Tibetan term is, however, used to render lajjā, shame, in Asaṅga’s Mahāyānasaṃgraha. See Nagao 1994. 24 Jinaputra comments that this is because the bodhisattva has no other means (Tatz 1986: 326). 25 For abhidharmic cross-referencing see Samyuktābhidharmahṛdaya ii.171, n. 502; for Theravāda sources see Gethin 2004: n. 19. 26 tr. Olivelle 2004: 199–200.
23

On the auspiciousness of compassionate violence

307

He notes that to kill such persons is one of the grave sins called “immediates.” [The moral calculus becomes even more complex when we consider that the killer’s relationship to the victim matters too. Killing one’s own mother is an “immediate,” but not killing someone else’s.] By the same logic, the situation is entirely reversed for the bodhisattva who is about to kill the depraved bandit. The karmic liability for killing such a person is the lowest possible. This is extremely important for understanding Buddhist penal codes, which have almost always included capital punishment, and Buddhist warfare.27 To give an extreme example cited by Harvey, the Mahāparinirvāṇa-sūtra describes killing the morally hopeless, icchāntika, as less than killing an ant.28 In any case, the more morally depraved or potentially harmful a person is, the less karmic demerit there is in killing them. So military enemies or slanderers of the dharma are in a dangerous moral category. Even those who merely hold wrong views are destined to be reborn as animals or in hell. This should be taken into account in interpreting the famous case of the Sri Laṅkan King Duṭṭhagāmaṇi.29 According to the Mahāvaṃsa, he marched to battle against the Tamils with a relic in his spear and a great company of monks, not for the sake of conquest, but to establish the dharma of the Saṃbuddha. Seeing that he took no joy in the bloody victory, eight arahants flew through the air to comfort him. They reassure him that having killed millions will be no obstruction to his entry into heaven, because his non-Buddhist war victims were accounted as being no more than animals.30 These people were active enemies of the Buddhadharma. The Buddhists he killed count for only one and a half persons. One who had taken refuge counted as half a person, while the other, who had taken precepts, counted as a full person. The story closes with a commitment by Duṭṭhagāmaṇi to never take a meal without offering to the saṅgha. The spear with
Florida 2005: 57. Harvey (2000: 138) cites Taishō 12.562b. He thanks Victor He. 29 Mhv 170–178. 30 The term used for “animals” here, pasu/Sanskrit paśu, is also the technical term for a sacrificial animal. In the Hindu homologization of warfare with sacrifice, this term is often used for victims killed in battle.
28 27

308

Stephen Jenkins

the relic became the axis of a great stūpa. This emphasizes that the king also has an ability to make massive merit through his support of the saṅgha that helps compensate for his negative acts. There is no question that this is an exceptional text that seems shocking, but these are basic Buddhist arguments that particularly support the violence of kings.31 Bhāviveka, the great sixth century Mādhyamika, brought out another aspect of karmic causality, which ameliorates the karmic vulnerability of the compassionate killer and harkens back to the Nikākyas.32 A person of high moral quality may be affected far less by a grave sin than a degenerate may be affected by a minor one. As heavily salted water may be rendered undrinkable by just a little more salt, while pure water may take much more and still be drinkable, so a small crime could lead to bad rebirth for a degenerate, while a large one might not for a saint.33 This would also be important for kings, who are generally considered to have large stores of merit. The points drawn out above only begin to explicate the complexity of the factors that condition the act of killing.34 The vulnerability of the compassionate killer is ameliorated by the fact that he has empathy, auspiciousness, and a reservoir of positive merit. Furthermore, his target is the worst sort of person, and the intention is to benefit both himself and his potential victims.35 The vulnerability of the villain, on the other hand, is enhanced by the fact that he is pitiless, has defiled inauspicious intentions, and has no
31 I do not see this story as being as inconsistent with normative Buddhist values as it is often perceived to be. For a different perspective and a variety of contrasting views, see Gethin 2007. 32 Eckel 2008: 186. 33 This seems to be a generally held idea. Loṇaphala Sutta, AN i.249, offers the same analogy, except that the large body of water is the Gaṅga. Eckel (2008: n. 327) directs us here to a rich note by La Vallée Poussin on the various contingencies on karmic outcomes (La Vallée Poussin 1990: 730, n. 217). 34 For a detailed technical discussion in abhidharmic style see La Vallée Poussin 1990: 642–666. 35 In Asaṅga’s case, we can say the ideal intention would also include benefiting himself.

On the auspiciousness of compassionate violence

309

reservoir of counterbalancing merit. Furthermore his targets are the best sorts of persons, and he seeks petty personal gain for himself alone. A last point we should note from the Bodhisattvabhūmi is that it is also a moral downfall to refrain from engaging in various harsh actions, when they are called for to benefit others.36 Candragomin, a seventh century Yogācāra remembered as a competitor of Candrakīrti, included this point when he famously and influentially summarized Asaṅga’s “Śīlapaṭala” in only twenty verses.37 Actions performed out of compassion or love with an auspicious intention are without fault. Indeed, even to refrain from harsh or threatening action when it is necessary to benefit others is a moral failure.38

Amputation with kindness
Āryadeva, who is considered the next great figure after Nāgārjuna in the Mādhyamika lineage, wrote in the third to fourth century C. E., “Because of their intention both bad, aśubham, and good, śubham, [actions] become auspicious for a bodhisattva.”39 Many
bodhisattvo yena kaṭukaprayogena tīkṣṇaprayogena sattvānām arthaṃ paśyati taṃ prayogaṃ daurmanasyārakṣayā na samudācarati  / sāpattiko bhavati[…] BoBh 116; Tatz 1986: 74, 221; following Tatz’s translation of Tsong-kha-pa’s commentary, Harvey claims that such an assertion does not occur in the Bodhisattvabhūmi (Harvey 2000: 140); however, both Tatz’s translation of the root text and Tsong-kha-pa’s commentary contain this statement. Yet, Tatz also reads Tsong-kha-pa a bit earlier, in commenting on “With mercy there is no [deed] without virtue,” as saying, “[…] the two commentaries on the Chapter on Ethics teach that there are occasions when the seven of body and speech – murder and the rest – are permitted. Aside from this, they do not state that to not engage in them for the sake of others is a fault.” Tatz observes here that three Chinese translations of the Bodhisattvabhūmi omit “this passage,” while that of Hsüan Tsang includes it. The extent of the omission is not clear (Tatz 1986: n. 396). 37 For rich textual cross-references and commentary see Tatz 1985: 36–38; for cross-referencing on compassionate transgression see Tatz 1982: 38, 64. 38 Tatz 1985: 28, v. 12, “not to give treatment even comprising affliction.” See also 29, v. 20. 39 CŚ Chapter V, v. 105, pp. 249–250 (tr. Sonam 1994: 136): bsam pas by­ ang chub sems dpaʼ la // dgeʼam gal te mi dgeʼang rung // thams cad dge legs
36

310

Stephen Jenkins

cases of similar statements can be cited from both the sūtras and śāstras and all the figures considered here, Asaṅga, Vasubandhu, Śāntideva, Āryadeva, Bhāviveka, and Candrakīrti agree on the basic point that a bodhisattva may do what is normally forbidden or inauspicious, akuśala.40 Candrakīrti, in the early seventh century, first comments on Āryadevaʼs verse by defining the inauspicious as that which leads to lower forms of rebirth and the auspicious, kuśala, as that which reverses the process of saṃsāra. Auspicious actions result in good births and happiness, while inauspicious actions result in the suffering of birth, old age, and death etc.41 This is a general principle, but it is one which raises the level of ambiguity. Since, for Candrakīrti, any act that reverses the cycle of rebirth becomes auspicious, the possibility is opened that any action may be auspicious depending on a variety of factors. If an act of killing may make merit, then it is neither a necessary evil, nor merely value free, but is clearly auspicious. In typical Buddhist fashion, he then proceeds to offer a catalogue of narrative case studies, rather than an abstract analysis. The first example is of a physician, certainly one of the most important and pervasive metaphors for a bodhisattva, amputating a finger that has been bitten by a poisonous snake, thus preventing the spread of greater suffering. Jinabhadra, a sixth century Jain, used the same example:

nyid ʼgyur te // gang phyir yid deʼi dbang ʼgyur phyir // Karen Lang has been very generous in sharing her forthcoming translation (see Lang tr. 2011) of this section of Candrakīrti’s commentary and has supported my work on this text for years. 40 To cite some examples that do not otherwise occur in the text: “As long as a Bodhisattva does not give up bodhicitta he has not broken the precepts” (Upāliparipṛcchā-sūtra, tr. Chang 1983: 269); “Even that which is proscribed is permitted for a compassionate person who sees it will be of benefit.” (Bca Chapter V, v. 84; Crosby and Skilton 1996: 41) 41 dge ba yang bde ba dang bde ʼgroʼi rnam par smin paʼi ʼbras bu can yin du zin kyang skye ba dang / rga ba dang ʼchi ba la sogs paʼi sdug bsngal sgrub par byed pa nyid kyi phyir na dge legs ma yin no/ / (CŚ 250).

On the auspiciousness of compassionate violence

311

A doctor has to cause pain, but is still non-injuring and innocent because his intention is pure … There can be nonviolence even when an external act of violence has been committed.42

The awkwardness of this translation, in which there is no violence even when there is violence, is eased somewhat if we substitute “non-harm” for “nonviolence,” which is a misleading translation of ahiṃsā in Buddhist, Hindu, or Jain thought. Henk Bodewitz points out that the term “non-violence,” which was never taken before modern Indian times to forbid war or capital punishment, is absent in older English dictionaries and is strongly associated with Gandhi.43 With the use of this term, the Gandhian conception, inspired by Tolstoy, is projected onto the past. In many examples below, it is important to recognize that being harmless may actually require violent action and that restraint from violent action may be harmful. Dictionary definitions of “violence” often include not only harmful physical force, but also the sense of being morally unwarranted or unjust. We would not normally describe surgery as violence, because it is neither harmful nor unwarranted. For this reason, Tibetan scholars I have worked with have sometimes objected that the compassionate killing of bodhisattvas is not violence. However, the interpretive problem in Indian thought in general is that warfare, torture, animal sacrifice, the horrific punishments of the dharmaśāstras etc. all may fall within the definition of ahiṃsā, since they are both warranted and beneficial.44 The same text from which one may pluck apparently unqualified statements of support for ahiṃsā may also advocate torture. This is not usually a failure of internal inconsistency. Therefore the word “violence” is being used in the context of this paper without any moral connotation, since the question is whether violent action, such as killing, may be moral. This also avoids a use of the term that would require one
Dundas 1992: 140. Bodewitz 1999: 17. He also notes that many scholars have misinterpreted ahiṃsā as the desiderative of the verb root han, to kill. He suggests “noninjury” as a translation, but this would not work with examples like killing. 44 See Jenkins 2010 on compassionate warfare and torture.
43 42

312

Stephen Jenkins

to call morally warranted killing, such as a bodhisattva stabbing to death a thief, nonviolent. A morally justified war would even have to be called nonviolent. My use of the term “violence” indicates injurious physical force, including killing, warfare, punishment and torture regardless of its moral character. An authoritative Buddhist precedent for Candrakīrti is found in Nāgārjuna’s Ratnāvalī, an epistle by the foundational Mahāyāna figure to a Buddhist king. Citing the word of the Buddha from an unknown text,45 Nāgārjuna writes:
It is called beneficial to cut off a finger when it has been bitten by a [poisonous] snake. So the Buddha says to even cause extreme pain, if it will help another.46

Chopping off a finger is a painful and violent act that brings to mind the famous Buddhist criminal Aṅgulimāla, “Finger-garland,” who decorated his neck with fingers cut from his victims. The simple act of cutting off a finger might be very similar, but, because of the differing intentions and outcomes, the moral implications are completely different. The action itself is morally neutral, even though we might assume that a compassionate doctor would perform such an amputation with a sense of regret and as a last resort, as described by Asaṅga. Candrakīrti emphatically states that the physician certainly does not accrue demerit for preventing the spread of even greater harm.

Bodhisattvagocaropāyaviṣayavikurvaṇanirdeśa 111.a.1, cited in the Sūtrasamuccaya attributed to Nāgārjuna, also uses the example of a doctor inflicting pain in advising a king to discipline the unruly, but not in relation to snakebite. 46 Ratnāvalī 181–182. Hopkins translates “mi bde ba yang bya bar” as “One should even bring discomfort.” This accords with the literal Tibetan, but seems mild for the example of cutting off a finger. For mi bde ba, Lokesh Chandra’s Tibetan-Sanskrit Dictionary gives śūla or simply duḥkha, which both seem stronger than discomfort; see Hopkins 1998: v. 264, 128; cf. the Mahāyānasaṃgraha on bodhisattvas assuming the role of a king “even inflicting torment on sentient beings to establish them in the code of discipline” (Keenan 1992: 88). It is also worth noting, in light of the fact that Nāgārjuna is addressing a king, that amputation was a common form of punishment in ancient and more recent Buddhist polities.

45

On the auspiciousness of compassionate violence

313

In the Majjhima-Nikāya, the Abhayarājakumārasutta uses a very similar analogy of saving a choking child to explain that the Buddha may sometimes use harsh speech. It was well known that the Buddha had spoken harshly to Devadatta and angered him by saying that he was incorrigible and destined for hell. Speech is a form of karmic action capable of causing harm. Most discussions of compassionate transgression include harsh speech and often begin with it. Here the Nigaṇṭha Nātaputta prompted Prince Abhaya to ask Gotama whether he would speak unwelcome and offensive words to others, a question which he predicted would be like an iron spike stuck in the Buddha’s throat. The Buddha explains his use of harsh speech as follows:
Now on that occasion, a young tender infant was lying on Prince Abhayaʼs lap. Then the Blessed One said to Prince Abhaya: “What do you think, prince? If, while you or your nurse were not attending to him, this child, were to put a stick or a pebble in his mouth, what would you do to him?” “Venerable sir, I would take it out. If I could not take it out at once, I would take his head in my left hand, and crooking a finger of my right hand, I would take it out, even if it meant drawing blood. Why is that? Because I have compassion for the child.” So too prince…Such speech as the Tathāgata knows to be true, correct, and beneficial, but which is unwelcome and disagreeable to others: The Tathāgata knows the time to use such speech. … Why is that? Because the Tathāgata has compassion for beings.47

This should not be taken as a general endorsement of compassionate transgression, but, for Mahāyānists, for whom harsh speech is a basic example, this would be very recognizable in terms of their own ideas. Compassion leads a person, who skillfully knows when it is appropriate, to cause pain in another when it has practical benefit. The Buddha makes use of the prince’s ordinary common sense ethics, rather than a supererogatory model, to illustrate his point. The sutta makes clear that, even if he is correct, the Buddha does not use harsh speech if it will not benefit others. To the degree that we can regard this as an earlier stratum of Buddhist thought, this appears to be a precedent for the basic type of thinking employed by the Mahāyāna.
47

MN i.392, Abhayarājakumārasutta (tr. Ñāṇamoli and Bodhi 1995: 499).

314

Stephen Jenkins

Candrakīrti offers another example of a hunter who kills one of his sons to prevent both from dying. The two sons are arguing at the edge of a precipice and one of them grabs the other with the intention of hurling them both over. Since he cannot reach them, and so has no other option, the hunter shoots one son with an arrow to prevent them both from dying. This case shows a concern for reducing the proportional extent of harm, as in the example of amputation. The Buddha is also often compared to a caravan leader, and in another example we find one whose fellow travelers are cornered by a lion. The caravan leader shoots the lion in the head to protect his company. Demiéville cites another caravan story, from the Mahā-Upāyakauśalya-sūtra, to be distinguished from the Upāyakauśalya-sūtra, which appeared very early in China and has had enduring influence.48 In this account a Brahmin is traveling with a caravan, which comes into proximity with a horde of five hundred bandits. The Brahmin kills the scout of the bandits, who was apparently his own personal friend, to prevent him from alerting the murderous band of thieves about his caravan’s location. Part of his consideration is that, if he tells his companions about the scout, they will kill the scout and become murderers themselves. In this way he prevents 999 people from becoming murderers, i.e. the 500 bandits and the 500 merchants minus himself, by taking on the karma of murder himself.49 Candrakīrti also relates the story of a bodhisattva born among lions who saves a large group of people caught in the coils of a

48 He cites Taishō 156, vii, 161b–162a. Demiéville 1973: 379. According to Lewis Lancaster, the Mahā-Upāyakauśalya-sūtra was first translated into Chinese in the third century (Lancaster 1979: 140). 49 It is not clear if the killer actually goes to hell. One would expect this tale to be a jātaka or avadāna, but the Brahmin is not identified as a bodhisattva. Demiéville does not give the karmic outcome of the story, except to say that the bandits and travelers are all converted. If he does go to hell, it would be a strong exception to my argument, and the first case I have found of compassionate transgression resulting in karmic penalty. This would also make it an irrational choice for motivating Chinese Buddhists to kill in war, since the assumption would be that they go to hell as a result.

On the auspiciousness of compassionate violence

315

huge snake. The bodhisattva frightens the snake by mounting the head of an elephant and releasing a great roar. In terror, the snake relaxes its grip and its captives are freed. This is an example of harsh speech as a violent act. In another example, a father accidentally kills his own beloved son. His only son had returned from a long period abroad in a very fragile state of health. The father brings about his son’s death by strongly embracing him. This clever example illustrates the fundamental importance of intention by making deep affection result in killing. Most of these stories are told in just a few lines, as if he takes for granted that his readers know the tales.

The ship captain
Candrakīrti also uses one of the most famous and influential, yet often misread, passages in Buddhist thought from the Upāyakauśalya-sūtra of the ship captain who kills a bandit. This jātaka is cited by both Madhyamaka and Yogācāra sources and has continued to be important in modern times.50 It seems to be a combination of two older stories, one in which the Buddha under the same name, “Greatly Compassionate,” saved five hundred passengers at sea and another in which, as a king, he stabbed a man to death with a spear.51 In this example, Captain Compassionate is faced with the knowledge that a thief intends to kill the five hundred merchant bodhisattvas riding in his ship. He gives this long reflection. If he tells the merchants, they will kill the thief and so suffer the bad karmic results.52 So, forming the compassionate intention to take the negative results upon himself, the ship captain
For examples, see Williams 2009: 152 and 340, n. 12; Welch 1972: 272– 288. Thanks to Chris Queen. 51 See n. 56 below. 52 The early sūtras have many examples of stupid and backsliding bodhisattvas, even bodhisattvas who have forgotten they were bodhisattvas. The fact that the text sees bodhisattvas as capable of killing in anger shows that it does not just indicate near deities with this term. Texts on bodhisattva ethics show a general concern for the fact that bodhisattvas make regular mistakes that require confession and contrition.
50

316

Stephen Jenkins

stabs the thief to death with a short spear. In this way, he skillfully benefits the potential mass murderer by saving him from eons in the hell realms. In fact, the thief is reborn in a heaven. [Perhaps this is an early source for the idea seen later in tantric contexts that a compassionate killer can direct the continuum of their victim to a heavenly rebirth.] In the case of someone about to commit a heinous crime, not only is there less sense of negative consequence for the killer, there is even the sense that one is benefiting them by executing them before they can accrue more time in the richly described Buddhist hell realms. This raises the issue of what other crimes also have such bad karma that it would be better to kill the person rather than allow them to be performed. For instance, the “immediates” often include splitting the saṅgha and sometimes slandering the dharma. That would imply that enemies from both within and without Buddhism could merit the same violence as someone about to kill a parent or saint. The Mahāparinirvāṇa-sūtra, for instance, says that if oneʼs motivation is pure, it is possible to kill someone who is persecuting Buddhists or deriding the Mahāyāna without incurring karmic retribution.53 The captain also saves the bodhisattva-merchants either from being murdered or becoming murderers themselves by attacking the thief.54 This is highly double edged; the very motivation for killing is based on the devastating negative consequences of murder. One would be better off to be murdered, than to kill without compassion. All these sources agree that killing may be used to prevent others from taking on the karma of murder. The entire story, like many in the Upāyakauśalya-sūtra, is framed as an explanation of a problematic event in the Buddha’s hagiography. Here the issue is that his foot was punctured by a thorn, which seems to suggest that the Buddha could be affected
Taishō No. 375, 12.676b5–6). Thanks to Jan Nattier. For more examples see Schmithausen 1999: 57–58 and n. 60. 54 This is a very potent example in the age of terrorism. On August 11, 2000, the Associated Press reported that Jonathan Burton, a teenage passenger who became combative on a Southwest Airlines flight, was killed by the passengers.
53

On the auspiciousness of compassionate violence

317

by karma. The thorn is homologous with the spear with which he stabbed the thief in a past life. As part of a general effort to show that compassionate killing remains an evil, albeit a necessary one, Harvey argues that the thorn shows that “the act had various bad karmic consequences, though not as bad as if it had been done without a compassionate motivation.”55 But the final word of the sūtra’s account explicitly rejects this interpretation. The Buddha merely shows himself to be punctured by the thorn as a skillful technique to teach the law of karma.56 In the process, he prevents another murder by demonstrating the law of karma to some potential killers. “For those reasons the Thus-Come-One has a thorn of Acacia stuck in his foot. That also is the skill in means of the ThusCome-One; it is not an obstacle caused by past deeds.”57 In another episode from the Upāyakauśalya-sūtra, not long after the story of Captain Compassionate, the Buddha knowingly allows a non-Buddhist female ascetic to be murdered.58 Part of the explanation for this is based on the common idea that our days are numbered. The Buddha saw that her lifetime was exhausted in any case. But what about her murderers, who will certainly go to one of the fantastically horrific hells so elaborately described in Buddhist texts? Killing just anyone is not an “immediate,” but surely the killers of this ascetic will suffer a horrible fate in the hell realms. Shouldn’t they be protected by the Buddha’s compassion? The murderous death of the ascetic will also have a negative karHarvey 2000: 136. Ap verses 21–22, gives another story in which the Buddha’s foot is hurt as the result of a past life as a king in which he killed a man with a spear. “I became a king and killed a person with a spear. By the ripening of that karma, I was boiled vigorously in hell.” In the present life, the Buddha is shown to still experience pain in his foot for that past killing and the karmic effects are not yet exhausted. This Apadāna is a catalogue of past-life misdeeds of the Buddha, including several murders. There is no sense that these were compassionate or dharmic acts. A central purpose of the Upāyakauśalyasūtra is to reread such tales, which seem to indicate that the Buddha could continue to suffer karmic consequences, in terms of Mahāyāna buddhology (Ap i.300). 57 tr. Tatz 1994: 77. 58 Ibid., 460.
56 55

318

Stephen Jenkins

mic effect, if she dies in terror. The sūtra argues that, in this case, bringing dishonor to the opponents of the dharma is a compensating benefit.59 It turns out that the killers were religious competitors of Buddhism. They are referred to as tīrthika, a name often erroneously translated as “heretic,” which probably refers to the antecedents of traditions we call Hindu today. The sūtra explains that the Buddha allowed the woman to be murdered, so that the discredit would fall on her tīrthika killers. Perhaps this should be read in the light of the fact that, since early times, holding wrong views is in itself sufficient to result in rebirth as an animal or in hell. The opponents of Buddhism, or a misguided Buddhist, would be understood to be leading others to such misfortune. So, both allowing and preventing murder is validated. No specific outcomes or actions are essentially evil. Killing, preventing murder, and allowing a murder are all auspicious within one narrative context.

Making merit with murder and mercy sex
Like Asaṅga, Candrakīrti also says that the Ship Captain benefits himself as well by reversing saṃsāra by myriad ages.60 On this point, I suspect an old mistranslation has been influential. In his Śikṣāsamuccaya, Śāntideva directly cites the Ratnamegha-sūtra on the allowance to kill someone who intends to commit an “immediate” [and also points out that the Śrāvaka Vinaya allows for the euthanasia of animals.] However a large part of his discussion of permitted transgressions is focused on the Upāyakauśalya-sūtra. He cites the jātaka of Jyotis, a Brāhmaṇa youth who broke his vow of abstinence in order to save the life of a woman who threatened to kill herself, if he would not engage in sex with her. In his translation of the Śikṣāsamuccaya, Bendall rendered a key phrase from the Upāyakauśalya-sūtra, cited by Śāntideva as “And so I myself
In the Bodhisattvabhūmi, Asaṅga often makes allowances for more negative behavior in the case of tīrthikas, for instance in terms of harsh speech or returning abuse. See Tatz 1986: 221–223. 60 dge baʼi rtsa ba des kyang bskal pa stong phrag brgyar ʼkhor ba la rgyab kyis phyogs par byas so // (CŚ 253.1).
59

On the auspiciousness of compassionate violence

319

young sir, by an impulse of pity, though vile, and full of desire, was set back for ten thousand ages.”61 He thus reversed the meaning and presented compassionate transgression as an enormous karmic setback. In his translation of the sūtra itself, Tatz rendered this instead, “Because I generated a thought that was endowed with great compassion but conjoined with transitory passion, birth and death was curtailed for ten thousand years.”62 Asaṅga says in the Mahāyānasaṃgraha that: “Even if a bodhisattva in his superior wisdom and skillful means should commit the ten sinful acts of murder etc., he would nevertheless remain unsullied and guiltless, gaining instead immeasurable merits.”63 Śāntideva, again quoting the Upāyakauśalya-sūtra, similarly says in the Śikṣāsamuccaya, “Behold, son of good family, the very action which sends others to hell sends a bodhisattva with skill in means to the Brahmaloka” heaven realms, [a traditional result of generating compassion].64 There is no question of the compassionate bodhisattva killer going to hell in these sources. This is consistent with a general pattern in Mahāyāna thought wherein the more pure a bodhisattvaʼs intention is to go to hell, the less likely she is to do it. The bodhisattva dramatically shortens the path to buddhahood, precisely because of being willing to sacrifice hiser own spiritual progress. The motivational conception and its actual results can be completely different. In fact the motivation can produce the opposite of what is intended; those who intend to endure hell realms do not, precisely because they are willing to do so.

Bendall and Rouse 1971: 163; So ’haṃ kulaputra mahākāruṇyacittotpādena-itvareṇa kāmopasaṃhitena daśakalpasahasrāṇi saṃsāram akārsaṃ; Tatz 1994: 35, n. 49; Śikṣ 167. 62 Tatz 1994: 34; for a similar phrase, see Suvarṇaprabhāsottama-sūtra, tr. Emmerick 1970: 31. Thanks to Mark Tatz for supplying me with the unpublished manuscripts of his Tibetan editions of the Upāyakauśalya-sūtra. Note that this also accords with Chang’s translation from the Chinese (Chang 1983: 456–457). Another translation of this episode from the Chinese can be found in Welch 1972: 284–286. 63 Keenan 1992: 88. 64 paśya kulaputra yad anyeṣāṃ nirayasaṃvartanīyaṃ karma, tad upāyakauśalyasya bodhisattvasya brahmalokopapattisaṃvartanīyam // (Śikṣ 167).

61

320

Stephen Jenkins

I have not yet located an example where a compassionate killer suffers negative karmic consequences. Bhāviveka may offer a highly qualified exception. In arguing that even great evil, pāpa, can be overcome, he points to the famous cases of the mass murderer, Aṅgulimāla, the patricidal King Ajātaśatru, and the wicked King Aśoka who turned their lives around by subsequently forming positive intentions.65 As with the thorn in the Buddha’s foot, Bhāviveka argues that it is only taught that they were reborn in hell to generate confidence in the law of karma, in fact their negative karma had been completely eliminated. They were born there, he then says, “like a silk ball that falls down and rises up. They were not touched by the flames of hell. In this way evil can be uprooted without denying the laws of karma.”66 The objection is then raised that even the Buddha suffered negative karmic consequences, such as his foot being pierced by a thorn. Bhāviveka specifically rejects this, referring directly to the Upāyakauśalya-sūtra. He then follows with a discussion of killing with compassion.
Others see someone on the verge of committing a heinous crime (ānantarya), know that this action will cause suffering for a long time, and kill that person out of compassion. They certainly know that they will be born in hell, but they adopt a wholesome or indeterminate (avyākṛta) motivation (citta) and kill in order to protect [others]. They accept their own rebirth in hell, but their wholesome [motivation] is sustained by wholesome thoughts like: “This is great suffering, but it will not last long.” This [motivation] is wholesome, because it is like a

65 The fact that a thematic study of Aṅgulimāla and Ajātaśatru could easily comprise a book length study shows how important the issue of avoiding the fruition of past inauspicious action was to Indian Buddhists. There are entire sūtras and sections of others focused on them. The Mahāparinirvāṇasūtra has an extensive discussion. The study of the theme of overcoming the karma of murder will have strong implications for the understanding of Buddhist ethics. This is particularly true for tantric texts, which, with their claim to achieve liberation in the present lifetime, can even avoid the fruition of the “immediates” which lead directly to hell on rebirth. 66 Eckel 2008: 185. The reference to either neutral or auspicious states of mind still seems unclear to me, since he immediately follows by indicating that the killer’s intention is auspicious.

On the auspiciousness of compassionate violence

321

thought that is free from desire and so forth.67 … A bodhisattva who commits murder out of compassion, cannot be reproached for this action, because it is not generated by hatred, …68

Bhāviveka’s exposition, including the reference to wholesome [auspicious] or neutral motivation, is very close to that of the Bodhisattvabhūmi, with which he was familiar. He does not say whether the killer actually goes to hell. Nothing would prevent this more than the intention to do so. However, his description of how the compassionate killer considers that the hell experience will not last long may be related to the description of bouncing in and out of hell without being touched by the flames. We can be certain that, if Aṅgulimāla is untouched by the flames of hell, that a bodhisattva would have at least as positive an outcome. It would seem incongruous to even correlate the karmic outcomes for a reformed mass murderer and a compassionate bodhisattva killer. But if we take these two together and assume that Bhāviveka is indicating that even such a bodhisattva bounces in and out of hell, it would explain the broadly held view of contemporary Tibetan scholars that compassionate killers have an extremely brief experience of hell.69 The
According to Eckel, Vasubandhu presents the identical paragraph in his Vyākhyāyukti. This seems remarkable, since this is not a citation of the sūtra (Eckel 2008: 187, n. 333). Demiéville notes that the ship captain story is recited in the commentary to Asaṅga’s Mahāyānasaṃgraha as an example of gāṁbhīriya śīla (Demiéville 1973: 380). He cites Lamotte 1939: 215–216. 68 Eckel 2008: 188. The term “wholesome” is commonly used as a translation alternative for “auspicious.” 69 As Eckel acknowledges, the comparison to the bounce of a silk ball, which suggests a momentary contact, remains a difficult translation problem. In a rich footnote, Eckel observes that Sthiramati uses a similar metaphor in commenting on verse 3.8 of Asaṅga’s Mahāyānasūtrālaṃkāra (Eckel 2008: 185, n. 324); in any case, the passage commented on by Sthiramati asserts that anyone possessing the bodhisattva­gotra, who takes rebirth in the lower realms, has a brief stay and minimal suffering. The intention of the metaphor is clearly to minimize either the extent or duration of suffering. For the purpose here, the fact that they do not experience the flames of hell is sufficient. Xuanzang offers a similar idiom of the time it takes for a ball of thread to fall to the ground after being tossed up (tr. Li Rongxi 1996: 105). The repeated closeness of Bhāviveka’s treatment to that of Asaṅga, Vasubandhu, and Sthiramati, is remarkable. See also Mahāyānasūtrālaṃkāra, tr. Jamspal,
67

322

Stephen Jenkins

great Tsong-kha-pa cites Bhāviveka here with approval.70 However Bhāviveka does not emphasize the production of great merit as do Asaṅga, Śāntideva and the Upāyakauśalya-sūtra itself, instead focusing on overcoming karmic negativity.71

Some qualifications
Candrakīrti clearly realizes the possibilities for exploitation in this idea. Later in the same commentary he launches into a jeremiad against a king who seeks to justify violence for the sake of maintaining moral order.72 For Candrakīrti, the reason bodhisattvas are not destroyed by such violence, while others are, is that they possess a controlled mind with compassionate intent.73 The opposite is also obviously true. Those who do these things without these qualities face fantastically negative consequences. The tension between these two is perfectly expressed in the Upāyakauśalya-sūtra itself, where the express purpose of the story is to actually discourage others from murder, rather than to validate compassionate violence. The Buddha demonstrates the power of karma to a group of potential murderers by showing himself to be pierced by a thorn as an outcome of spearing the thief in his earlier life as Mahākāruṇika the Ship Captain.74 So, even the portrayal of compassionate murder
L. et als. 2004: 27. 70 tr. Lamrim Chenmo Translation Committee 2000: 256. 71 The apparent contrast between Bhāviveka and his Prāsaṅgika opponents raises the interesting question of the relationship between ethics and ontology. Śūnyavāda traditions seem to be more ethically liberal than abhidharmic ones. We would expect Bhāviveka, with his concern for firmly establishing conventional norms, to perhaps be more conservative than Candrakīrti. There is a striking correlation between tathāgatgarbha traditions and vegetarianism that is based perhaps on their strong sense of a base consciousness. Sources that validate killing on the basis of emptiness are another further area of exploration. 72 See Lang 1992: 232–43. 73 gzhan dag la yang de ltar ciʼi phyir mi ʼgyur zhe na / sems la dbang thob pa med paʼi phyir dang / sems kyi rgyud nyon mongs pa mkhrang zhing nye bar sad pas bzung ba nyid kyi phyir ro // (CŚ 251.3). 74 Tatz 1994: 34–9; 73–4. cf. Chang 1983: 431–440, 456–457.

On the auspiciousness of compassionate violence

323

is used to discourage murder by malicious people. However, the sūtra has already shown, consistent with the later interpretations of Asaṅga, Śāntideva and Candrakīrti, that the ship-captain in fact made enough merit through this murder to reverse saṃsāra by one hundred thousand kalpas and the thorn is merely an upāya of the Buddha, not an actual karmic outcome.

Compassionate violence as common sense
In general, compassionate killing is a supererogatory ethic, not one of imitation. It is double edged in opening the possibility for murder precisely to prevent its horrific karmic outcome. Yet Candrakīrtiʼs earthy examples also suggest that there is something commonsensical about compassionate violence. Part of the power of Candrakīrti’s hypothetical cases is that they appeal to natural human responses to protect children and companions. They draw on issues and choices that doctors, leaders, parents or pilots may face in everyday life and derive their force from the fact that they make intuitive sense to people. If bodhisattvas are like ordinary folk, then ordinary folk may be like bodhisattvas. The possibility that this discourse merely elaborates a supererogatory ethic without general significance seems dubious. All the sources view compassionate killing as dangerous. But one would expect Buddhists to attempt, as far as they were able, to behave like bodhisattvas when faced with difficult moral choices. If a bodhisattva is like a physician cutting off a poisoned finger, then a physician is also like a bodhisattva. As in the teachings for bodhisattvas, a good doctor must know what she is doing, have a compassionate intention, and would regret the pain that she causes. Surely, as in the Jain understanding, a doctor performing an amputation need not be a great bodhisattva to avoid terrible karma. When Nāgārjuna uses the finger amputation analogy to advise a king that he may have to inflict great pain, he is not speaking to a bodhisattva, but to a very dangerous person. Kings routinely used amputation as a punishment in ancient India. In the broadly cited Bodhisattvagocara-upāyaviṣaya-vikurvaṇanirdeśa-sūtra, the same thinking is applied to penal codes and war-

324

Stephen Jenkins

fare.75 A king is encouraged to compassionately punish and even torture the unruly in order to discipline them and protect society, but he is not to kill or permanently damage them. He may go to war to protect his family and his people. He should try to avoid war in the first place and carefully consider how his policies are responsible for the creation of enemies. But even if he kills the enemy, as long as he avoids the destruction of life, infrastructure, and nature, he will be blameless and produce great merit. This is stated with almost the same phrasing as Asaṅga’s. There is no sense that the king, his warriors, or law enforcement officials must be bodhisattvas. As we consider these sources, all of them framed within or focused on narrative, we should remember that even the early mainstream narrative traditions of Buddhism are full of stories of Buddhist warfare that feature the Buddha in past lives as a weapons master, king, warhorse, execution elephant, elephant mahout engaged in a siege etc.76 In Āryaśūra’s Jātakamālā, a Mahāyānist collection of birth-stories, the Buddha is described as being born as Śakra, i.e. Indra, in a past life. Indra is king of the devas and a model of the ideal king. The demonic daityas, another class of lesser deities, challenge him to battle. The battle is described in vivid dramatic detail.
Despite his scruples, everything inclined the bodhisattva to engage in the frenzy of battle: the enemies’ presumption; the fear people felt, which put an unpleasant curb on their amusements; his own dignity; and the course of action that prudence dictated. … There then took place a battle that shattered the nerve of the cowardly and in which armor splintered at the clash of weapon on weapon. … “Watch out!” “Now how are you going to escape me?” “Attack!” “That’s the end of you!” – such were the cries as the combatants killed one another.77

See Jenkins 2010. Ibid. 77 trans. Khoroche 1989: 81–82. Pāli texts also refer to this battle. See for instance SN i.221. Here Indra’s conduct toward the defeated and bound enemy king is lauded.
76

75

On the auspiciousness of compassionate violence

325

The devas broke ranks and fled under a shower of arrows, and finally Indra himself turned his chariot in retreat. But as he turned his huge chariot to flee the field, he saw that he was about to overrun some nests full of baby birds. “I would rather that the demon chiefs battered me to death with their terrible clubs than that I live on with my reputation ruined, under the reproach of having slaughtered these creatures who are distraught with terror.”78 In order to avoid crushing them and at the risk of his life, he turned directly back at the pursuing daityas. Shocked by this turn of events, the demonic forces broke rank in turn and were routed by the devas. Victory turned on the compassionate response to baby birds and once again karuṇā proved to be protective. The story is typically ironic in simultaneously validating both deadly warfare and that “all decent men should cultivate sympathy for living things.” The bodhisattva-king of the devas surely provides a model for the good Buddhist king here.

Closing reflections on metaethics
This paper has been an effort to begin to understand what Indian Buddhist texts say about compassionate killing. I think we are going to keep discovering a Buddhism very different than the one we think we know. Important sūtras and large bodies of narrative literature are in many cases untranslated. Even major figures such as Candrakīrti, Asaṅga, and Bhāviveka have only been partially translated, not to speak of the commentaries. I was fortunate to have the very recent work of David Eckel on Bhāviveka. Paul Harrison has recently shown that in many cases we do not even know when Śāntideva, who has been at the absolute center of the study of Buddhist ethics, is composing or quoting.79 It seems critical for the inherently comparative application of metaethical analysis to Buddhist thought to have a clear object of analysis or pole of comparison, but we have not yet clarified what we intend to analyze even in regard to individual thinkers.

78 79

trans. Khoroche 1989: 83. Harrison 2007: 215–248.

326

Stephen Jenkins

The ethics of compassionate violence are a complex matrix of multiple interrelated and competing concerns, including proportionality, intention, virtue, situation, and consequences conceived from a multiple-life perspective. The basic principle that the auspicious is defined by that which leads to positive karmic outcomes only increases the level of ambiguity by removing the possibility that any action is essentially inauspicious. Although there are many warnings of hell and promises of heaven for specific acts in Buddhist ethical rhetoric, there are as many reminders that the workings of karma are ultimately inconceivable. If karma is inconceivably complex, then the auspicious is equally inconceivable, and so follows Buddhist ethics. I do not mean by “ambiguity” to say that Buddhists are befuddled or that they do not have clear moral principles. I mean this in the positive sense that lack of moral certainty, appreciation for narrative complexity, rejection of oversimplification, and a high toleration for the almost unfathomable complexity of moral situations can be positive things. I suspect that Buddhist ethics constantly resort to narrative, because it is capable of maintaining tensions and ambiguities and representing diverse voices and multiple levels of concern.80 In Buddhist thought, narrative is as likely to be the commentary itself as it is to be the object of analysis.81 This makes the application of Western metaethics especially challenging, since it tends to function in the opposite way, that is, by clarifying narrative through systematic analysis. The jātakas, avadānas, hagiographies etc. are at least as important for the understanding of Buddhist ethics as any subtle psychological or philosophical analysis, but these are the most neglected texts in modern studies. Legends of Śāntideva and Asaṅga may tell us more about how Buddhists understood their ethics than the Bodhicaryāvatāra or Bodhisattvabhūmi, at least in regard to the contexts that held those legends dear. Certainly, in my experience, relatively few Buddhists know the commentarial literature to
I am indebted here to conversations with John Strong. I acknowledge the influence of my dissertation advisor, Charles Hallisey, on this point.
81 80

On the auspiciousness of compassionate violence

327

which most of this paper is devoted, while the story of the ship captain is known throughout the Buddhist world. It is remarkable that Asaṅga, Vasubandhu, Bhāviveka, Candrakīrti, and Śāntideva all resorted to the same brilliant little story of Captain Compassionate, which holds the possibility of auspiciously killing with compassion in dynamic tension with the horror of killing without it. One gets the feeling that Buddhist thinkers are deliberately enhancing the ambiguity, as if only an ambiguous ethic could do justice to lived reality.

References Primary sources
AN Aṅguttara-Nikāya. Edited by R. Morris, E. Hardy. 5 vols. London: Pali Text Society, 1885–1900. Ap The Apadāna. Edited by Mary Lilley. Oxford: Pali Text Society, 2000. Bca Bodhicaryāvatāra of Śāntideva with the Commentary Pañjikā of Prajñākaramati. Buddhist Sanskrit Texts no. 12. Edited by Sridhar Tripathi. Darbhanga: Mithila, 1988. BoBh Asaṅga. Bodhisattvabhūmi. Edited by Nalinaksha Dutt. Patna: Jayasawal Research Institute, 1978. BoBh–Wogihara – Asaṅga. Bodhisattvabhūmi. Edited by Unrai Wogihara. Tokyo: Seigo Kenyukai, 1930–1936. Bodhisattvagocaropāyaviṣayavikurvaṇanirdeśa – [Better known and cited as, but not catalogued as, the Āryasatyakaparivarta-sūtra.] Byangs chub sems ba’i spyod yul gyi thabs kyi yul la rnam par ’phrul pa bstan pa. Derge, mDo-sde, Text 146, bKa’ ’gyur, Volume Pa 203. 82a–141b. CŚ Catuḥśatakam: Candrakīrtipraṇītaṭīkayā Sahitam. Sanskrit and Tibetan. Edited with Hindi translation by Gurucharan Singh Negi, PhD Dissertation. Sarnath: Central Institute Higher Tibetan Studies, 2005. MN Majjhima-Nikāya. Edited by V. Trenckner, R. Chalmers. 3 vols. London: Pali Text Society, 1888–1899. Ratnāvalī Nāgārjuna. Ratnāvalī of Ācārya Nāgārjuna with the Commentary of Ajitamitra. Edited by Ācārya Ngawang Samten. Sarnath: Central Institute Higher Tibetan Studies, 1990.

328

Stephen Jenkins

Samyuktābhidharmahṛdaya – see Dessein (tr.) 1999. Śikṣ Śikṣāsamuccaya: A Compendium of Buddhistic Teaching Compiled by Śāntideva Chiefly from Earlier Mahāyāna Sūtras. Edited by Cecil Bendall. The Hague: Moutons, 1957. SN Saṃyutta-Nikāya. Edited by L. Feer. 5 vols. London: Pali Text Society, 1884–1898.

Secondary sources
Bendall, Cecil and Rouse, W.H.D. tr. 1971. Śikṣāsamuccaya: A Compendium of Buddhist Doctrine Compiled by Śāntideva Chiefly from Earlier Mahāyāna Sūtras. Delhi: Motilal Banarsidass, 1971. Bodewitz, Henk. 1999. “Hindu Ahiṁsā and Its Roots,” in Violence Denied: Violence, Non­Violence and the Rationalization of Violence in South Asian Cultural History. Edited by Jan E. M. Houben and Karel R. Van Kooij. Leiden: Brill, 17–42. Chandra, Lokesh. 1959–1961. Tibetan-Sanskrit Dictionary. New Delhi: International Academy of Indian Culture. Reprinted 1990, Kyoto: Rinsen Book Company. Chang, Garma C.C. ed. 1983. A Treasury of Mahāyāna Sūtras. London: Penn State. Crosby, Kate and Skilton, Andrew. tr. 1996. Śāntideva. The Bodhicaryāvatāra. Oxford: Oxford University Press. Demiéville, Paul. 1973. “Le Bouddhisme et la Guerre,” in Choix d’études bouddhiques 1929–1970. Edited by Paul Demiéville. Leiden: Brill, 347– 385. Dessein, Bart. tr. 1999. Samyuktābhidharmahṛdaya. Heart of Scholasticism with Miscellaneous Additions. Delhi: Motilal Banarsidass. Dundas, Paul. 1992. The Jains. London: Routledge. Eckel, Malcom David. 2008. Bhāviveka and His Buddhist Opponents. Cambridge: Harvard University. Edgerton, Franklin. 1985. Buddhist Hybrid Sanskrit Grammar and Dictionary. Vol. 2. Delhi: Motilal Banarsidass. Emmerick, R.E. tr. 1970. The Sūtra of Golden Light: Being a Translation of the Suvarṇaprabhāsottamasūtra. Translated by R. E. Emmerick. London: Luzac and Co. Ltd. Florida, Robert. 2005. Human Rights and the World’s Major Religions: Vol. 5, The Buddhist Tradition. Westport, Connecticut: Praeger Publishers. Geiger, Wilhelm. tr. 1986. Mahāvaṃsa or the Great Chronicle of Ceylon. Translated by Wilhelm Geiger into German. Translated from German into English by Mabel Haynes Bode. New Delhi: Asian Educational

On the auspiciousness of compassionate violence

329

Services, 1986. Gethin, Rupert. 2007. “Buddhist monks, Buddhist kings, Buddhist violence: on the early Buddhist approaches and attitudes to violence,” in Religion and Violence in South Asia: Theory and Practice. Edited by John Hinnells and Richard King. New York: Routledge. _____ 2004. “Can Killing a Living Being Ever Be an Act of Compassion? The analysis of the act of killing in the Abhidhamma and Pali Commentaries,” in Journal of Buddhist Ethics. Volume 11, 166–202. Groner, Paul. 1990. “The Fan­wang ching and Monastic Discipline in Japanese Tendai: A Study of Annen’s Futsū jubosatsukai kōshaku.” in Chinese Buddhist Apocrypha. Edited by Robert Buswell. Honolulu: University Hawaii, 251–290. Harrison, Paul. 2007. “The case of the vanishing poet: new light on Śāntideva and the Śikṣā-samuccaya,” in Indica et Tibetica. Edited by Konrad Klaus and Jens-Uwe Hartmann. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien, 215–248. Harvey, Peter. 2000. An Introduction to Buddhist Ethics. Cambridge: Cambridge University Press. Holt, John. 1981. Discipline: The C anonical Buddhism of the Vinaya. Varanasi: Motilal Banarsidass. Hopkins, Jeffrey. 1998. Buddhist Advice for Living and Liberation: Nāgārjuna’s Precious Garland. Ithaca: Snow Lion. Jamspal, L. et als. tr. 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. Edited by R. Thurman, New York: American Institute of Buddhist Studies, 2004. Jenkins, Stephen. 2010. “Making Merit through Warfare and Torture According to the Ārya-Bodhisattva-gocara-upāyaviṣaya-vikurvaṇanirdeśa Sūtra,” in Buddhist Warfare. Edited by Mark Juergensmeyer and Michael Jerryson. New York: Oxford University Press, 59–75. _____ 2003. The Circle of Compassion: An Interpretive Study of Karuṇā in Indian Buddhist Literature. Cambridge Buddhist Institute Series, series ed. R. C. Jamieson. Ayrshire, Scotland: Hardinge Simpole Publishing. Currently out of print, but available under the same title as Harvard University Doctoral Dissertation, Ann Arbor: University Microfilms International, 1991. Keenan, John. tr. 1992. The Summary of the Great Vehicle. Translated by John Keenan. Berkeley: Numata Center. Khoroche, Peter. tr. 1989. Once The Buddha was a Monkey, Ārya Śūra’s Jātakamālā Chicago: University Chicago Press.

330

Stephen Jenkins

Kleine, Christoph. 2006. “Evil Monks with Good Intentions,” in Buddhism and Violence. Edited by Michael Zimmermann. Kathmandu: Lumbini International Research Institute, 65–98. Lamotte, Étienne. tr. 1939. La somme du grand véhicule d’Asaṅga (Mahāyānasaṃgraha). Vol. 2, 1939; repr. Louvain-la-Neuve: Institut Orientaliste, 1973. Lamrim Chenmo Translation Committee. tr. 2000. Tsong-kha-pa, The Great Treatise on the Stages of the Path to Enlightenment: Lam Rim Chen Mo. Vol. 1. Translated by the Lamrim Chenmo Translation Committee. Edited by Joshua Cutler and Guy Newland. Ithaca: Snow Lion Publications. Lancaster, Lewis. 1979. The Korean Buddhist Canon: A Descriptive Catalogue. Berkeley: University California. Lang, Karen. tr. 2011. Catuḥśatakam. Unpublished manuscript. _____ 1992. “Āryadeva and Candrakīrti on the Dharma of Kings,” in Asiatische Studien: Zeitschrift der Schweizerischen Gesellschaft für Asienkunde Études Asiatiques: Revue de la Société Suisse d’Études Asiatiques. 46.1, 232–243. de La Vallée Poussin, Louis. tr. 1990. Vasubandhu, Abhidharmakośabhāṣyam. Translated by Louis De La Vallée Poussin. Translated into English by Leo Pruden. Berkeley: Asian Humanities Press, 1990. Li Rongxi. tr. 1996. The Great Tang Dynasty Record of the Western Regions: Translated by the Tripiṭaka-Master Xuanzang under Imperial Order, Composed by Śramaṇa Bianji of the Great Zonchi Monastery. Taishō, Volume 51, Number 2087. Translated into English by Li Rongxi. Berkeley: Numata Center, 1996. Nagao, Gadjin. 1994. An Index to the Mahāyānasaṃgraha. Studia Philologica Buddhica IX, Part 1: Tibetan-Sanskrit-Chinese, Tokyo: International Institute for Buddhist Studies. Ñāṇamoli, Bhikkhu and Bodhi, Bhikkhu. tr. 1995. The Middle Length Discourses of the Buddha. Boston: Wisdom Publications. Olivelle, Patrick. tr. 2004. The Law Code of Manu. Oxford: Oxford University Press. Rahula, Walpola. 2001. Le Compendium de la super­doctrine Philosophie d’Asaṅga. Paris: École française d’Extrême-Orient, 1971. Translated from French by Sara Boin-Webb. Fremont, CA: Asian Humanities Press, 2001. Schmithausen, Lambert. 1999. “Aspects of the Buddhist Attitude to War,” in Violence Denied: Violence, Non­Violence and the Rationalization of Violence in South Asian Cultural History. Edited by J. E. M. Houben and K. R. van Kooij Leiden: Brill, 45–67. Shih Heng-ching, Bhikṣuṇī. tr. 1994. Upāsaka Precepts. The Sūtra on

On the auspiciousness of compassionate violence

331

Upāsaka Precepts. Translated by Bhikṣuṇī Shih Heng-ching. Berkeley: Numata Center. Sonam, Ruth. tr. 1994. Catuḥśatakam. Yogic Deeds of Bodhisattvas. Translated by Ruth Sonam. Ithaca: Snow Lion. Tatz, Mark. 1986. Asaṅga’s Chapter on Ethics with the Commentary of Tsong-kha-pa. New York: Edwin Mellen Press. _____ 1985. Difficult Beginnings. Boston: Shambhala. _____ 1982. Candragomin’s Twenty Verses on the Bodhisattva Vow and its Commentary by Sakya Dragpa Gyaltsen. Dharamsala: Library of Tibetan Works and Archives. _____ tr. 1994. The Skill in Means Sūtra. Delhi: Motilal Banarsidass. Wallace, Vesna and Wallace, Alan. tr. 1997. A Guide to the Bodhisattva’s Way of Life. Ithaca: Snow Lion. Welch, Holmes. 1972. Buddhism Under Mao. Cambridge: Harvard University. Williams, Paul. 2009. Mahāyāna Buddhism: the Doctrinal Foundations. Second edition. London: Routledge. Woodward, F. L. tr. 1933. Aṅguttara Nikāya. The Book of Gradual Sayings. New York: Oxford.

What is it like to be a bodhisattva? Moral phenomenology in Śāntideva’s Bodhicaryāvatāra1
Jay L. Garfield

1. Introduction: the text, its title and some readings
Bodhicaryāvatāra was composed by the Buddhist monk scholar Śāntideva at Nālandā University in India sometime during the 8th Century CE. It stands as one the great classics of world philosophy and of Buddhist literature, and is enormously influential in Tibet, where it is regarded as the principal source for the ethical thought of Mahāyāna Buddhism. The title is variously translated, most often as A Guide to the Bodhisattva’s Way of Life or Engaging in the Bodhisattva Deeds, translations that follow the canonical Tibetan translation of the title of the book (Byang chub sems pa’i spyod pa la ’jug pa) and the commentarial tradition of Tibet. But that translation itself represents only one of two possible original Sanskrit titles, and a more natural English rendering of either Sanskrit title is simply How to Lead an Awakened Life, and that indeed describes the content of the text admirably. Taking this as the meaning of the title can issue in a kind of gestalt shift in our view of the text, allowing us to see it not so much as a characterization of the extraorThanks to Nalini Bhushan, Caroline Sluyter, Susanne Mrozik and Jan Westerhoff for useful comments on an earlier draft of this essay, and to Barbra Clayton for insightful critique at the International Association of Buddhist Studies XVth Congress. Thanks especially to Steve Jenkins for an extensive, challenging critique, from which I have learned much, and which has improved this paper. Any errors that remain are my own, and probably reflect my inattention to what my colleagues have pointed out. Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 333–357
1

334

Jay L. Garfield

dinary moral life of a saint, but as a guide to moral development open to any of us. I therefore recommend that as an English translation of the title and as an understanding of the subject of the text. Śāntideva’s understanding of how to lead such a life is distinctive, and is very different from accounts of the moral or the exemplary life familiar in the Western tradition. It is, I will argue, primarily a phenomenological account, and that is why I think it important for this account of Buddhist ethics to be taken as a voice in the contemporary philosophical conversation about the nature of ethics and about the proper form of moral theory. When read in the context of contemporary metaethics, and only as an important text in the history of Buddhism, How to Read an Awakened Life gains new importance. The central moral phenomenon taken up in the text is that of bodhicitta, a term I prefer to leave untranslated. This term is usually translated either as the awakened mind or as the mind of awak­ ening. But that’s not very helpful, in part because of the different connotations of citta/sems and mind in Buddhist and Western philosophy, respectively, and in part because of the unclarity of the bare genitive construction in English. Avoiding the temptation to follow attractive philological and metaphysical byways, let me offer this preliminary reading of the term: Bodhicitta is a complex psychological phenomenon. It is a standing motivational state with conative and affective dimensions. It centrally involves an altruistic aspiration, grounded in compassion, to cultivate oneself as a moral agent for the benefit of all beings.2 That cultivation, as we shall see, demands the development of skills in moral perception, moral responsiveness, traits of character, insight into the nature of reality so deep that it transforms our way of seeing ourselves and others, and what we would call practical wisdom. In short, bodhicitta entails a commitment to attain and to manifest full awakening for the
We must tread with care, here however. As Susanne Mrozik has pointed out to me, Stephen Jenkins argues (1998) that “altruism” may be a bit strong, since, as we shall see below, bodhicitta and the motivations and skills connected to it, are beneficial to the bodhisattva as well as to others. It is, as Śāntideva will emphasize, always in the end in one’s own interest to cultivate bodhicitta.
2

What is it like to be a bodhisattva?

335

benefit of others. A bodhisattva is one who has cultivated bodhi­ citta in at least one of two senses adumbrated by Śāntideva, and distinguished below. How to Lead an Awakened Life addresses the nature of bodhi­ citta and the means of cultivating it. We can read it as a treatise on the distinction between the phenomenologies of benighted and of awakened moral consciousness. Śāntideva’s account of morality has been read in the West as a distinctively Buddhist theory of moral virtue, that is, as structurally Aristotelian, even if very different in content from Aristotle’s account of virtue and the good life (Keown 2005, 2007). It has also been read as consequentialist (Goodman 2008, Siderits 2007). Each of these readings, I fear, is a symptom of a dangerous hermeneutic temptation to force Buddhist ethics into a Western mould, and while each reading reflects something of the content of Śāntideva’s approach, each misses the heart of the matter. It is true, as proponents of the areteic reading note, that Śāntideva focuses in How to Lead an Awakened Life on the cultivation of traits of character, and it is true that he contrasts moral virtues such as patience and compassion with moral vices such as a irascibility and selfishness, and recommends virtue over vice, focusing on states of the agent as opposed to actions or obligations. On the other hand, for Śāntideva the point of all of this is not to lead a happy life, or even to be a good person: bodhicitta does not take the moral agent as its object. The point is to benefit others, as well as oneself.3 Perfection itself, in other words is, for Śāntideva, neither an end in itself, nor final, nor self-sufficient. This is no virtue theory, and awakening, while analytically related to the perfections, is not a kind of eudemonia analytically related to virtues. It is also true that Śāntideva urges us to care for the happiness of others, and to reduce their suffering. And it is even true that we can find verses in the text that enjoin us to compare how little our own happiness or suffering is in comparison to that of all sentient
I thank Steve Jenkins for emphasizing that for Śāntideva, as for most Buddhist ethicists, there is no difference between self and others as objects of moral concern.
3

336

Jay L. Garfield

beings in order to motivate us to sacrifice our own interests for that of others. Here is an example many (e.g. both Goodman 2008 and Siderits 2007) cite in support of a consequentialist reading:
Without any hesitation, I relinquish My body, my pleasures, And all virtues achieved throughout all time In order to benefit all sentient beings. III: 104

But to take such verses to be the expression of a kind of consequentialism would be to take them seriously out of context, and to miss the heart of Śāntideva’s account. This verse in particular, as we will see below, occurs in the context of a resolution to abandon selfishness, and to broaden my moral gaze to universal scope, to cultivate a way of seeing myself in the context of a much broader whole in which my own interests are a rather small affair. Śāntideva does not argue that bodhicitta or the perfections cultivated by the bodhisattva are valuable because of the consequences they entail; and there is never a suggestion that the suffering of one can be balanced against the happiness of another. Whereas for a consequentialist, balances of benefits and harms are the ground of the value of actions or attitudes, but not necessarily their objects, for Śāntideva, the good of others is an object of bodhicitta, but not the ground of its value. Its value is grounded instead in the fact that it is the only rational way of taking up with the world. And comparison, or tradeoff of suffering and benefit is never on the table.5
All translations are my own, from the sDe dge edition of the Tibetan text, as reprinted in rGyal tshab 1999. As Wallace and Wallace note in Śāntideva 1997, the Tibetan version of the text differs in many – usually minor – ways from the available Sanskrit edition of the text. It appears that even early on there were at least two versions of the text, and there is no way of determining whether the Sanskrit edition from which the Tibetan and Indian translators worked was in some respect preferable to that which survives today or not. But it is worth noting that because of the importance this text attained in Tibet, most of the significant commentarial literature refers to the Tibetan version. The notable exception is Prajñākaramati’s Bodhicaryāvatāra-pañjikā, which follows the available Sanskrit. 5 Moreover, as Steve Jenkins points out (1998), in the end there never is a tradeoff – most accounts of virtuous action end up with the claim that every­ one benefits, even those who apparently suffer temporary adversity, although
4

What is it like to be a bodhisattva?

337

Tillemans (2008) sees How to Lead an Awakened Life as centrally concerned with the problem of akrasia, arguing that Śāntideva is concerned with the problem of how to overcome the conflict between his knowledge of what is best – both for himself and for others – and his desire for vice. There are certainly passages, for instance those in which Śāntideva attempts to cultivate revulsion for sexual behaviour that support this interpretation:
If you do not lust for the impure, Why do you repeatedly embrace another Who is only flesh-smeared bones Bound together with sinew? VIII: 52

In passages such as these, Śāntideva clearly aims to counteract vicious desire by reminding himself of the knowledge of virtue. Nonetheless, as I hope will be clear from the remainder of this discussion, just as the conception of virtue, vice and moral perfection at work in this text are non-Aristotelian, the conception of moral conflict at work in How to Lead an Awakened Life is not Aristotelian. Śāntideva is more concerned with the conflict between desire or aversion and impersonal aspiration – or, perhaps more clearly, the conflict between attachment and freedom – than he is in a conflict between knowledge and desire. The akratic desires one thing, but knows that another is better, and the Aristotelian puzzle concerns reconciling rationality, knowledge of the good and desire for the ill. The solution to akrasia is the cultivation neither of more knowledge, nor of other desires, but of moral strength. Not so for Śāntideva: although confusion, on his account, is the root of desire and aversion, and eliminating that confusion is the ground of awakening, moral conflict for him is not so much cognitive as conative. It is to be resolved by firmly establishing metaphysical knowledge in one’s mode of taking up with the world, a knowledge which issues in the relinquishing of desire and the arising of the appropriate aspiration, not through cultivating moral strength. Moreover, for Śāntideva, since vice is always ultimately rooted in confusion, and the elimination of confusion issues in virtue, there can never be a situation in which one really
perhaps one must take the long view to see those benefits.

338

Jay L. Garfield

knows what is right but chooses what is wrong. There is always a failure of knowledge, not just of will, in vicious action. It will hence be better to set aside the doxography that helps us to sort Western ethical theories, and to approach Śāntideva on his own terms in the context of Buddhist ethical thought. The insights we will gain from reading him in this context will repay forbearing to locate him in our landscape. We will focus on the place of the bodhisattva ideal in Mahāyāna ethics as a preliminary to exploring the phenomenology of morals in How to Lead an Awakened Life.

2. The Buddhist moral outlook and the bodhisattva ideal
Buddhist moral theory is not Western moral theory. What is it? I have argued elsewhere (Garfield unpublished) that Buddhist ethics is best thought of as an attempt to solve a deep existential problem – the problem of the ubiquity of suffering – and as an attempt to solve that problem by developing an understanding of our place in the complex web of interdependence (pratītyasamutpāda) that is our world. Buddhist ethics is grounded in the so-called “four noble truths.” The first two are particularly important for present purposes: (1) that the universe is pervaded by suffering and the causes of suffering, a truth obvious to anyone on serious reflection, though one that escapes most of us most of the time precisely because of our evasion of serious reflection, an insight that, as we shall see, Śāntideva takes very seriously; (2) that suffering arises as a consequence of actions conditioned by attachment and aversion, each of which in turn is engendered by confusion regarding the nature of reality, a confusion that is a kind of primal cognitive instinct, which includes a tendency to reify ourselves, and that which pertains to ourselves; to take that which is impermanent to be permanent; that which is insubstantial to be substantial; that which is interdependent to be independent. This triune root of suffering is represented in the familiar Buddhist representation of the Wheel of Life with the pig, snake and rooster at the hub, the six realms of transmigration representing aspects of the phenomenology of suffering – brutality; pain and despair; insatiable need; arrogance and the need for recognition;

What is it like to be a bodhisattva?

339

insensitivity to the pain of others in our own happiness; and the vulnerability and imperfection that comes with being human – revolving around them, structured by the twelve links of dependent origination (a detailed psychology of perception and action), all of which is depicted as resting in the jaws of death, the great fear of which propels so much of our maladaptive psychology and moral failure. In the present reading of How to Lead an Awakened Life, I take that representation, one we an imagine Śāntideva walking past every day on his way to work, as inspiring the text in a very deep way. The most important innovation in Mahāyāna moral theory, of course, is not the well-known framework of the six perfections, but the installation of compassion as the central moral value and the model of the bodhisattva’s compassionate engagement with the world as the moral ideal. The compassion at issue is not a passive emotional response, and not a mere desire. Instead it is a genuine commitment manifested in thought, speech and physical action to act for the welfare of all sentient beings. Compassion in this tradition is founded upon the insight to which Śāntideva gives voice, that suffering is bad, per se, regardless of whose it is. To fail to take another’s suffering seriously as a motivation for action is, he argues, itself a form of suffering and is irrational.
“Self and others are the same,” One should earnestly meditate: “Since they experience the same happiness and suffering, I should protect everyone as I do myself.” VIII: 90 Divided into many parts, such as the hands, The body is nonetheless to be protected as a single whole. Just so, different beings, with all their happiness and suffering, Are like a single person with a desire for happiness. VIII: 91 Even if my own suffering Does no harm to anyone else’s body, It is still my own suffering. Since I am so attached to myself it is unbearable. VIII: 92 Just so, even though I do not experience The sufferings of others, It is still their own suffering.

340

Jay L. Garfield

Since they are so attached to themselves, it is hard for them to bear. VIII: 93 I must eliminate the suffering of others Just because it is suffering, like my own. I should work to benefit others Just because they are sentient beings, as am I. VIII: 94 Since I am just like others In desiring happiness, What is so special about me That I strive for my happiness alone? VIII: 95 Since I am just like others In not desiring suffering, What is so special about me That I protect myself, but not others? VIII: 96 If, because their suffering does not harm me, I do not protect them, When future suffering does not harm me, Why do I protect against it? VIII: 97 The idea that this very self Will experience that suffering is false: Just as when one has died, another Who is then born is really another. VIII: 98 If another should protect himself Against his own suffering, When a pain in the foot is not in the hand, Why should one protect the other? VIII: 99 One might say that even though it makes no sense, One acts this way because of self-grasping. That which makes no sense with regard to self or to others Is precisely the object you should strive to abandon! VIII: 100 The so-called continuum and collection, Just like such things as a forest, or an army, are unreal. Since the sufferer does not exist, By whose power does it come about? VIII: 101 As the suffering self does not exist, There are no distinctions among anyone. Just because there is suffering, it is to be eliminated.

What is it like to be a bodhisattva?

341

What is the point of discriminating here? VIII: 102 “Why should everyone’s suffering be alleviated?” There is no dispute! If it is to be alleviated, all of it is to be alleviated! Otherwise, I also am a sentient being! VIII: 103

This is a deep insight, and one over which we should not pass too quickly: the bodhisattva path is motivated in part by the realization that not to experience the suffering of others as one’s own and not to take the welfare of others as one’s own is to suffer even more deeply from a profound existential alienation born of a failure to appreciate one’s own situation as a member of an interdependent community.6 Interdependence guarantees that our joys are social joys; our sorrows are social sorrows; our identity is a social identity; the bounds of our society are indefinite. We either suffer and rejoice together in the recognition of our bonds to one another, or we languish in self-imposed solitary confinement, afflicted both by the cell we construct, and by the ignorance that motivates its construction. Compassion, grounded in the awareness of our individually ephemeral joint participation in global life, hence is the wellspring of the motivation for the development of all perfections, and the most reliable motivation for morally decent actions. Compassion is also, on the Mahāyāna view, the direct result of a genuine appreI must concede that, as Steve Jenkins has urged (personal communication), although dependent origination plays a central role in Indian and Tibetan Buddhist metaphysics, and although confusion is often glossed as the failure to understand or to apprehend interdependence, and although vice is held to be grounded in confusion, there is no explicit assertion in classical Indian Buddhist literature or in Tibetan commentarial literature that virtue arises from the appreciation of community membership, or of interpersonal, or of inter-sentient being connection. The closest an Indian text comes is Śāntideva’s analogy at VIII: 90–92 of the moral community to the body. More explicit uses of interdependence as a moral idea await the engaged Buddhism of the 20th and 21st centuries. Nonetheless, it should be clear that the material for the rational reconstruction I offer is present in the classical texts. Moreover, canonical meditation practices in the Tibetan tradition intended to cultivate compassion, such as the visualization of all sentient beings as one’s mother, emphasize this point.
6

342

Jay L. Garfield

ciation of the essencelessness and interdependence of all sentient beings. Once one sees oneself as nonsubstantial and existing only in interdependence, and once one sees that the happiness and suffering of all sentient beings is entirely causally conditioned, the only rational attitude one can adopt to others is a compassionate one. This is the Mahāyāna philosophical framework that sets the more specific context for Śāntideva’s project.

3. The structure of the text and the structure of moral experience
As I noted at the outset, the structure of How to Lead an Awakened Life is a clue to Śāntideva’s account of the structure of moral experience. Let us examine that structure in more detail. As a reminder, the chapters run as follows 1. In Praise of Bodhicitta 2. Eplanation of Vice 3. Adopting Bodhicitta 4. Caring for one’s Attitude 5. Maintaining Awareness 6. The Perfection of Patience 7. The Perfection of Enthusiasm 8. The Perfection of Meditation 9. The Perfection of Wisdom 10. The Dedication of Merit. Many commentators note immediately that Śāntideva’s presentation of the Mahāyāna list of the six perfections appears incomplete, leaving out what is generally regarded as the first – the perfection of generosity. (The fourth and fifth chapters are really a presentation of the second – the perfection of mindfulness.) Tibetan commentators, including the fifteenth century exegete rGyal tshab (1999), and the present Dalai Lama in oral teachings, present a twofold solution to this apparent problem. First, they note, the third chapter,

What is it like to be a bodhisattva?

343

on bodhicitta, is replete with references to generous intentions and so can be construed as a presentation of that perfection. Second, they point out, the final chapter, the dedication of merit, addresses generosity at a much deeper level. Here Śāntideva discusses the experience of giving even one’s own moral attainments and aspirations, and the universal scope of moral concern. This return to the theme of generosity is taken by these commentators to present a deliberate contrast between the way generosity is experienced at the beginning of the bodhisattva path in the context of what Śāntideva calls aspirational bodhicitta and the way it is experienced in the context of the engaged bodhicitta achieved in the cultivation of an awakened life. In chapter III, for instance, we encounter resolutions such as these, as Śāntideva cultivates the aspiration to lead the awakened life. The generosity he imagines cultivating involves depersonalizing his own motivations and developing a commitment to benefit all beings in direct material ways:
And so I will perform these deeds. Through the virtue I thereby acquire, May I completely alleviate All of the suffering of all sentient beings. III: 6 May I be the medicine For all who are ill. May I be both medicine and physician Preventing the recurrence of illness. III: 7 Through showers of food and drink May I alleviate the pain of hunger and thirst! At times of famine May I be both food and drink! III: 8 To all destitute and miserable beings May I be an inexhaustible treasure! May I be the one who presents them With all their many necessities. III: 9 Without any hesitation, I relinquish My body, my pleasures, And all virtues achieved throughout all time

344

Jay L. Garfield

In order to benefit all sentient beings. III: 10

In the tenth chapter, on the other hand, the generosity described is that of one who has cultivated engaged bodhicitta. Śāntideva is now depersonalizing not only his ends, but his own state of being. The transference of merit he envisions involves conceiving of his own virtue not as a state pertaining to him, but as a more general feature of the moral universe, and hence his own experience of himself, as of generosity, is transformed through the cultivation of engaged bodhicitta.
Through the merit attained By my composition of How to Lead an Awakened Life May all beings lead awakened lives. X: 1 Through my merit, may all those In all directions, who suffer In mind or in body Attain oceans of bliss and happiness. X: 2 May my life be just like that of Mañjuśrī, who strives to benefit All sentient beings Dwelling in the ten directions of space. X: 54 For as long as space endures, For as long as there are beings in cyclic existence, So long may I endure, In order to dispel the suffering of those beings. X: 55

The generosity embodied in aspirational bodhicitta is personal, taking as its intentional object my own contribution to the welfare of the world; the generosity embodied in engaged bodhicitta is im­ personal, taking as its intentional object only the benefit of others, with my achievements, not myself serving as its condition. This observation provides one interpretative key to a reading of How to Lead an Awakened Life as a treatise on moral phenomenology. But we can go further. Śāntideva opens the text with reflections on the moral experience of one contemplating serious moral development, writing, as he does throughout the text, in an intensely first-person confessional voice enabling deep phenomenologi-

What is it like to be a bodhisattva?

345

cal reflection, and often calling to the Western reader’s mind the Confessions of Augustine:
There is nothing here that has not been said before, Nor do I have any skill in composition. Thus, I have no concern for others and I have composed my text solely to cultivate my own mind. I: 2 Since my virtue is cultivated, My faith thereby increases in power. Nonetheless, if someone else with an outlook like my own Sees this, it would be meaningful. I: 3

Śāntideva emphasizes here not only his own moral deficiency, but more importantly that moral practice is aimed principally at self-cultivation. What is to be cultivated, he tells us, is bodhicit­ ta. Śāntideva emphasizes its importance, and immediately distinguishes two important degrees of bodhicitta. The first, to be cultivated at the outset of moral development is aspirational bodhicitta, the serious intention to lead an awakened life with a commitment to cultivating one’s moral capacities; the second is engaged bodhic­ itta, the set of spontaneous moral perceptual skills and dispositions that lead one to act in beneficial ways.
When bodhicitta has arisen, in an instant, Even a wretch who is bound in the prison of the cycle of existence Comes to be known as a child of the tathāgatas, And becomes an object of reverence in the realms of gods and men. I: 9 In brief, bodhicitta Should be known to be of two kinds: Aspirational bodhicitta And engaged bodhicitta. I: 15 Just as one can tell the difference between One who aspires to travel and a traveler, The learned can tell The analogous difference between these two. I: 16

The second and third chapters explore the nature of vice and the motivation for aspiring to an awakened life. Once again, the exploration of these themes is undertaken through a reflection on

346

Jay L. Garfield

Śāntideva’s own experience, and emphasizes the interior quality of vice and of the desire to transcend it.
Thus, since I have not realized That I am ephemeral, Through confusion, attachment and aversion, I have committed many kinds of vicious deeds. II: 38 O Protectors, through being inattentive And heedless of this danger, For the sake of this impermanent life I have achieved much that is vicious. II: 42 With distressed glances, I seek protection in the four directions. Which good person Will protect me from this great fear? II: 45 Having seen that there are no protectors in the four directions, I fall into total confusion. With no protectors anywhere, What shall I do in such a state? II: 46

Here we see Śāntideva taking a state of moral immaturity to be a state of intense suffering, conditioned by confusion and permeated by fear. We will return to this central role of fear in Śāntideva’s distinctive analysis of moral vice below, but the general point to observe at this stage is simply that Śāntideva’s approach to the question, “why be moral?” at this stage of the text is directly phenomenological: vice feels just terrible, and terrible in characteristic ways.7 And the motivation to moral progress is characterized in similar terms:
Two things need clarification here: first, the sense of the word “vice,” that I use to translate pāpa/sdig pa (often translated as sin – see Wallace and Wallace in Śāntideva 1997: 24, n. 22 for a discussion of this issue; second, the sense in which it feels so terrible. First, “vice” denotes any state of character, motivation or mind that conduces to maintaining confusion, attachment and aversion, and hence constitutes an obstacle to the liberation from suffering. It thus contrasts with virtue, which conduces to liberation through reducing confusion, attachment and aversion. Vices include such mundane states as selfishness, carelessness and sloth as well as more florid states such as murderous rage or boundless avarice. Second, the fact that vice feels terrible does
7

What is it like to be a bodhisattva?

347

Overwhelmed by fear, I offer myself to Samantabhadra, And of my own accord, I offer this, my body to Mañjughoṣa. II: 49 In despair, I cry out for help to The protector, Avalokiteśvara, Who acts compassionately and inerrantly, Begging him to protect my vicious self. II: 50 Now, having experienced the great terror, Heeding what you once told me, I approach you for refuge so that you Might quickly dispel my fear. II: 53 It makes no sense for me to enjoy the present day Saying to myself, “I will not die just now.” The time when I will cease to exist Will inevitably arrive. II: 58

The subsequent chapters take up the path to an awakened moral life. The account is rich – far too rich to take up in full detail. But the sequence of the chapters demonstrates the strikingly phenomenological approach to ethics adopted by this text. Śāntideva begins by considering how one, having developed aspirational bodhicitta, cares for and nurtures the attitude; he then turns to how one develops the concentration required to maintain introspective awareness of one’s own motivational and affective states; to the development of patience and then enthusiasm for ethical practice. The final chapters of the text address the role of meditation in stabilizing the qualities and ways of seeing cultivated earlier, and finally the importance of a particular kind of wisdom as the foundation of the engaged bodhicitta that is the foundation of awakened life – that is, the ability to see all phenomena – including oneself, that to which one is intimately related, and other moral agents – as
not entail that it is always explicitly experienced as terrible. Śāntideva anticipates by a bit over a thousand years Freud’s insight that our deepest pain may be unconscious – that everyday life is often full of pathology precisely because of underlying, unacknowledged, unconscious suffering. Often, as Śāntideva recognizes here, and as Freud was to discover much later, the recognition of one’s own pain is the first step to its resolution.

348

Jay L. Garfield

empty of inherent existence, as interdependent and as impermanent. For Śāntideva, the culmination of ethical practice is a cognitively rich perceptual skill – a new way of experiencing oneself in the world.

4. Fear and refuge: Suffering, aspiration and awakening as moral development
The role that fear plays in Śāntideva’s account of the phenomenology of moral life is striking, and the diagnosis of this fear is subtle. This point connects deeply with the centrality of the practice of taking refuge in Buddhist life, and one can read this text profitably as an extended meditation on refuge. The very need for refuge itself suggests an overarching experience of fear – perhaps a fear whose dimensions and objects are to a large extent opaque to the sufferer. The ultimate source and object of this fear is depicted graphically in the Tibetan representation of the wheel of life, in which all of existence takes place in the jaws of death. The iconography suggests that our cognitive and emotional lives, the constant cycling between states of mind and the experience of being buffeted about by events – whether external or internal – that are beyond our control, gives rise to so much suffering and is driven in large part by the unconscious awareness and fear of the inevitability of death. That fear engages us psychologically at the hub. Although the awareness of our own impermanence and that of all about which we care constitutes the horizon of our experience, we suppress that fear in confusion, living our lives as though these impermanent phenomena are permanent. This is why Tsong kha pa (2006: 34– 35) remarks that confusion, in this sense (avidyā/ma rig pa) is not simply the absence of knowledge, but the direct opposite of knowledge – a psychologically efficacious and destructive denial of the truth. In this case, as Śāntideva is aware, it is a denial of what we at a deeper level know to be true, of a troubling knowledge. This confusion born of fear generates attraction and aversion. Our attachment to ourselves, to our own well-being, to our possessions, to our conventions and practices, and in general to all that in the end is a source of suffering, Śāntideva urges, is at bottom a

What is it like to be a bodhisattva?

349

reflexive defensive reaction to the fear of loss. Our aversion to that which we find distasteful is at bottom a reaction to the fact that it reminds us of our own impermanence and vulnerability. Our conviction that we are independent agents interacting with other independent agents – a feature of our moral experience that runs both so very deep and so contrary to all that we know upon reflection – he urges, is a way of warding off the fear of interdependence, of being out of control, of being subject to the natural laws that issue in our aging, infirmity, reliance on others, and eventual demise. And around the hub cycle our emotions, desires, actions, and experiences, as expressed in II: 38 and II: 42 translated above on p. 346. But fear and awareness of fear are two very different things. Though we all live in fear, we are not, Śāntideva thinks, all aware of that background of fear, or of its impact on our lives. Moral sensibility properly so-called, according to the account of How to Lead an Awakened Life, arises when one becomes truly aware of the terror that frames one’s life and that lies at the root of selfdeception and vice. That awareness generates the impulse to take refuge and to strive for awakening, and as a consequence, the cultivation of aspirational bodhicitta.
From this very moment I go for refuge to The victors, protectors of all beings, Who strive for the purpose of protecting all, And who have great power to completely eliminate all fear. II: 47 Likewise, I honestly go for refuge to The Dharma in which they are completely engaged, Which completely eliminates the fear of cyclic existence, As well as to the assembly of bodhisattvas. II: 488

Moral development is hence a transformation of moral experience; a transition from a life conditioned by terror and unreason – albeit perhaps unconscious terror and unrecognized unreason – to a life conditioned by confidence and clarity; from a life constituted by phenomenological self-deception to a life constituted by introspective awareness; from a life in which vice is inevitable and taken to be unproblematic just because it is not recognized, or recognized
8

Cf. further II: 49, 50 and 53 as translated above on p. 347.

350

Jay L. Garfield

as vice, to a life in which the cultivation of virtue is at the centre of one’s consciousness.

5. Moral phenomenology as moral theory in Buddhism
One might, if one were a Western metaethicist, think that this rich moral phenomenology is an adjunct to explicit ethical theory, and search for the account of the right or the good, or of virtue or vice that underlies this account of the contrast between the experience of the morally immature and the morally mature agent. One would, however, look in vain. How to Lead an Awakened Life is a text on ethics; indeed, it is one of the most important Indian Mahāyāna Buddhist ethical treatises, the central ethical treatise for the Tibetan tradition and indeed is firmly grounded in theory articulated in pre-Mahāyāna śrāvakayāna texts. Nonetheless, How to Lead an Awakened Life is exclusively devoted to an account of the cultivation of moral sensibility and moral experience. Let us explore some examples. Chapter V on maintaining awareness, opens with these verses:
One who wishes to protect his practice Should be careful to protect his mind. If one does not protect one’s mind It is impossible to protect one’s practice. V: 1 The elephant of the mind Causes much harm and degradation. Wild, mad elephants Do not cause so much harm. V: 2 Nonetheless, if the elephant of the mind Is restrained by the rope of mindfulness, Then all fear is banished, And every virtue falls into our hands. V: 3

Maintaining the focus on the relationship between the dissipation of fear and moral development, Śāntideva argues that the cultivation of a moment-to-moment awareness of one’s own cognitive and emotional states is central to leading an awakened life. The morally benighted are characterized by an inattention to their own mental lives (even if they may be obsessed with an idea of morality); awak-

What is it like to be a bodhisattva?

351

ening consists in part in replacing that inattention with mindfulness. Later in the same chapter, Śāntideva emphasizes, using the metaphors of Buddhist hell imagery (including the notorious hideous women who appear in the trees) that our own suffering is entirely psychological, and that moral development is entirely mental cultivation:
Who so purposefully forged The implements of sentient beings’ hell? Who constructed the floor of burning iron? And whence have those women come? V: 7 The Sage has explained that The vicious mind gives rise to all of these. So, there is nothing whatever in the triple world More frightening than the mind. V: 8

Perhaps the most widely studied and most beautiful chapter in How to Lead an Awakened Life is that on patience. In this chapter Śāntideva emphasizes the pervasiveness of anger and aversion in the morally immature state, and the enormous – though often unconscious – suffering they bring in train, feeding on a cycle linking anger to fear and aggression.9 The predominance of these emotions prior to the cultivation of bodhicitta contrasts with the patience that characterizes awakened moral experience.
All of the virtuous actions Amassed over a thousand eons, such as Giving alms and making offerings to the tathāgatas, Are destroyed by a single instance of anger. VI: 1 There is no vice like aversion, and There is no aescetic practice like patience. Hence one should assiduously

9 The unwary Western reader might think that there is a confusion here between emotional immaturity and ethical immaturity. This thought should be resisted, as it rests on a presumption that our emotional life is independent of our moral life. It is central to Śāntideva’s – and indeed, any Buddhist – conception of the domain of the moral that our emotions are morally significant and morally evaluable. Emotional immaturity is one dimension of moral immaturity; emotional maturity one dimension of moral maturity (see Dreyfus 2002).

352

Jay L. Garfield

Cultivate patience in a variety of ways. VI: 2 When the thorn of aversion sticks in the heart, The mind finds no peace, Nor can it achieve happiness or joy. Sleep does not come, and one’s strength ebbs away. VI: 3

The account of the development of moral consciousness continues with an examination of the role of enthusiasm. Once again, the story of ethical growth is told not in terms of obligations or actions, but in terms of the development of character, which in turn is analyzed experientially.
Thus, after patience one should cultivate enthusiasm. For awakening depends on enthusiasm. Just as without wind nothing moves, Virtue cannot arise without enthusiasm. VII: 1 What is enthusiasm? It is determination to attain virtue. What should we call its opposite? It is to cling to the base, To deprecate oneself and one’s tradition. VII: 2 I have not completed this; I have just started this, but it remains half done. When death suddenly arrives, I will think, “Alas, I am defeated!” VII: 8 When vices are abandoned, there is no more suffering. If one is wise, one is without sorrow. For mistaken conception and vice Are what harm mind and body.VII: 27

The final two chapters of How to Lead an Awakened Life address the more directly cognitive aspects of the awakened moral life, those concerned with establishing how to see reality properly. Meditation is central to Śāntideva’s account, for it is through meditation that one embeds discursive knowledge into one’s character. In the following three verses he focuses on various aspects of attachment, its consequences, and how to relinquish it. The first of these considers the impact of the understanding of impermanence on the release from attachment to others, and the development of equanimity; the second addresses attachment to self, and once again, its connection to fear; the third emphasizes more graphically the role of medita-

What is it like to be a bodhisattva?

353

tion in reconstructing not our behavior or sense of duty, but our way of seeing the world.
What impermanent being Attaches himself to what impermanent being? For sadly, he will not see her For thousands of lifetimes. VIII: 5 If one thinks such thoughts as “I am very rich and respected, And many people like me,” When death approaches, fear will arise. VIII: 17 If you do not lust for the impure, Why do you repeatedly embrace another Who is only flesh-smeared bones Bound together with sinew? VIII: 52

Finally, in the chapter on wisdom, Śāntideva emphasizes both the importance of a deep understanding of metaphysics for moral life, and, more specifically, the fact that the relevant metaphysical view is the Madhyamaka view according to which all phenomena are empty of essence, interdependent, and have only conventional identities. It is important to note here not just that Śāntideva recommends this metaphysical position as the foundation of awakened life and morality, but, especially in the context of the preceding chapter, that it is a foundation for such a life precisely because once internalized, this view (and for Śāntideva only this view) transforms one’s very experience both of the external world and of oneself, generating a metaphysical vision that enables a moral engagement, thus enabling engaged bodhicitta.10
The Sage taught all of these matters For the sake of wisdom. Therefore, if one wishes to avoid suffering, And to attain peace, one should cultivate wisdom. IX: 1

10 Of course this means that it is not as easy as it sounds to attain moral perfection. Perfect compassion requires perfect wisdom. Nonetheless, this does not mean that it is impossible to cultivate any virtue: the cultivation even of mundane compassion increases wisdom; the cultivation of even a basic understanding of emptiness increases compassion.

354

Jay L. Garfield

Without an understanding of emptiness A mental state that has ceased will arise once again, Just as when one engages in non-conceptual meditation. Therefore, one should meditate on emptiness. IX: 48 Pride, which is the cause of suffering, Increases due to delusion regarding the self. Since from one, the other necessarily follows, Meditation on selflessness is supreme. IX: 77

Taking the treatise in this way allows us much better to understand the place of the ninth chapter in the text. Its presence and form are puzzling at first. Most of the rest of the text addresses issues immediately recognizable as ethical, such as generosity, patience, etc. And most of the rest of the text is written in a direct, accessible poetic style. And the rest of the text speaks quite directly, without the scholastic device of an interlocutor to whose arguments the text responds. The ninth chapter, on the other hand, is concerned with difficult questions in metaphysics and epistemology, such as the ultimate nature of all phenomena, the relationship between the way things appear and the way they are, the question of the relation between mind and the external world, and about whether consciousness is necessarily reflexive – issues not so obviously ethical. Its verses are difficult, highly abstract, and often written in a formal scholastic debate style. It is very difficult to parse, let alone to understand, without a good commentary. One might even reasonably suspect that it is a late graft. The present reading, however, shows why this excursion into abstract metaphysics and epistemology is so important to Śāntideva, and why it occurs at the conclusion of the treatise, and not, as one might think, given the foundational role of insight in Buddhist theory, at the beginning. It is a central theme of How to Lead an Awakened Life, as it is a central theme of the Buddhist diagnosis of the existential problem of suffering generally, that suffering and the egocentric tendencies it generates and which in turn perpetuate it, are grounded in a fundamental confusion about the nature of reality – taking what is in fact interdependent, impermanent and essenceless, on both the subjective and objective side – to be independent, enduring, and substantial. This attitude, Śāntideva urges,

What is it like to be a bodhisattva?

355

is not the result of careful metaphysical reflection, but an innate cognitive instinct. A truly awakened life requires its extirpation. This extirpation requires philosophical reflection, but such reflection is not sufficient, given the depth of the cognitive set.11 Even receptivity to that argument requires the cultivation of a moral sensibility that loosens the attachment and aversion implicated with the metaphysical error. But meditative practice is also necessary, in order that reflective thought can become a spontaneous cognitive set, a way of being in the world, rather than a way of thinking about the world, in which we experience ourselves and all around us as we are, interdependent, impermanent, insubstantial.12 This transformation of vision, and consequent transformation of mode of being, even though it is cast in How to Lead an Awakened Life as a direct understanding of ultimate reality, and an understanding of the relation between this ultimate reality and conventional reality, amounts not to seeing behind a world of illusion, but rather to coming to see a world about which we are naturally deceived just as it is, not being taken in by the cognitive habits that issue in that deception. For this reason, just as the historical Buddha, in the presentation of the eightfold path at Sarnath, emphasized that one’s view of the nature of reality is a moral matter, Śāntideva, in his analysis of an awakened life, urges that our metaphysics and epistemology is central to our moral lives. It is because it is this vision that finally transforms aspirational to engaged bodhicitta that this chapter comes at the end, not at the beginning of the text.

6. A distinctive view of moral life
This whirlwind tour through How to Lead an Awakened Life of course cannot do justice to the richness of Śāntideva’s insight and moral thought. On the other hand, I hope that it convinces the reader that in Śāntideva’s account of Buddhist ethics we find a disA mere addition of insight to injury, as a psychoanalyst friend of mine used to say about some cognitive therapies 12 Once again, the fact that meditative practice may be necessary to achieve virtuoso ethical status, this does not mean that meditation is the sine qua non of any moral progress.
11

356

Jay L. Garfield

tinctive approach to ethical thought, a perspective not available in mainstream Western ethical thought. When Śāntideva asks about moral life, he asks not what our duties are, nor what actions are recommended, nor what the relation is between the good and our actions, nor even what would make us individually happy. Instead, he starts with a problem that is to be solved – that of the ubiquity of suffering – and the standard Buddhist diagnosis of that problem in terms of attachment and aversion rooted in fundamental ontological confusion. He asks how to solve that problem. His solution is distinctive in that it develops a deeper diagnosis of the problem through an analysis of our own experience of ourselves and of our place in the world. It seeks the solution to this problem not – at least not directly – in a transformation of the world, or even of our conduct, but rather in a transformation of that experience. The task of leading an awakened life – a morally desirable life – is the task of transforming our phenomenology. It may be tempting either to force this account into a familiar Western form, or even just to graft it on to one, as the phenomenological adjunct to a more familiar metaethical account. I suggest that we resist that temptation, and open ourselves to the possibility of a very different way of understanding ethical aspiration and engagement – the path of the bodhisattva.

Bibliography Primary sources
rGyal tshab (1999). Byang chub sems pa’i spyod pa la ’jug pa’i rnam bshad rgyal sras ’jug ngogs. Sarnath: Gelukpa Student Welfare Committee.

Secondary sources
Drefyus, G. (2002). “Is Compassion an Emotion: A Cross-Cultural Exploration,” in R. Davidson and A. Harrington, eds., Visions of Compassion. Oxford: Oxford University Press. Garfield, J. (unpublished). “Buddhist Ethics.” http://www.smith.edu/philosophy/documents/BMT.pdf, last visited 30-07-2011

What is it like to be a bodhisattva?

357

Goodman, C. (2008). “Consequentialism, Agent-Neutrality and Mahāyāna Ethics,” Philosophy East and West 58 (1), pp. 17–35. Jenkins, S. (1998). The Circle of Compassion: A Comparative Study of Karuṇā in Indian Buddhist Literature. PhD Dissertation, Harvard University. Keown, D. (2005). “Origins of Buddhist Ethics,” in W. Schweiker, ed., Blackwell Companion to Religious Ethics. Oxford: Blackwell. _____ (2007). “Buddhist Ecology: A Virtue Ethics Approach,” Contemporary Buddhism 8 (2), pp. 97–112. Śāntideva (1997). A Guide to the Bodhisattva’s Way of Life. V. Wallace and A. Wallace, trans. Ithaca: Snow Lion. Siderits, M. (2007). “Buddhist Reductionism and the Structure of Buddhist Ethics,” in P. Bilimoria, J. Prabhu and R. Sharma, eds., Indian Ethics. London: Ashgate, pp. 283–297. Tillemans, T. (2008). “Reason, Irrationality and Akrasia (Weakness of Will) in Buddhism: Reflections on Śāntideva’s Arguments with Himself,” Argumentation 22, pp. 149–163. Tsong kha pa (2006). Ocean of Reasoning: A Great Commentary on Nāgārjuna’s Mūlamadhamakakārikā. N. Samten and J. Garfield, trans. New York: Oxford University Press.

Madhyamaka Buddhist ethics1
Tom J. F. Tillemans

What is this elusive discipline called “Buddhist ethics?” As is often the case in modern interpretations and analyses, it is not easy to characterize exactly what a western-inspired term corresponds to in traditional Indian culture, and sometimes it’s not even clear how a particular term is being used in a burgeoning modern secondary literature on the subject. If we look at Indian and Tibetan literature, the closest term to the western notion of “ethics” seems to be śīla, “moral discipline,” something that is the subject of monastic Vinaya codes, Abhidharma scholastic, bodhisattva literature, Jātaka tales, narrative Avadāna literature, some Madhyamaka2 treatises, even tantric texts and so on and so forth – in short, a little bit everywhere. Modern scholars have devoted significant efforts to the question as to whether there is a recognizable Western ethical theory – be it utilitarianism or virtue ethics – that is implicit in all, or at least the most, significant works in this literature. This debate will not be my concern here, although like Jay Garfield (2011) (year) I too think it is difficult to meaningfully attribute such an overriding ethical theory to Buddhism.3 That said, almost all Buddhist literature is certainly profoundly ethical in orientation, even if it is
This article grew out of a lecture to the annual conference of the Center of Buddhist Studies of the University of Kathmandu. My thanks to the Center and its students for their continued informed interest in substantive issues. Thanks also go to Mark Siderits for his helpful feedback. 2 In what follows, for convenience rather than conviction, I adopt the modern convention of using “Madhyamaka” for the thought and “Mādhyamika” for the thinkers. 3 For a recent vigorous defense of Buddhist ethics as utilitarianism, see Goodman 2009. Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 359–378
1

360

Tom J. F. Tillemans

not clearly and consistently theoretically oriented. The śīla discussions do attempt to tell us what we ought to do and why – to take a very rough and ready characterization of what ethics is about (Thomson 2001: 6). The codes and advice and obligations involve a sense of “should/ought,” one no doubt weaker than a Kantian duty, but an ethical demand nonetheless. And śīla literature does often tell us why, that is, it gives justificatory reasons as to why one should think that such and such ethical demands are genuine and well-founded and others are not. The Buddhists are not just moralizers; they give rationales for what they say ought to be done.4 Is there in any interesting sense a Madhyamaka Buddhist ethics, i.e., an ethics that would be particular to the Middle Way school and follow from or somehow be linked to its subtle analyses of metaphysics? In other words, does the Madhyamaka anti-realist philosophy that all things are empty (śūnya) of intrinsic nature (svabhāva) make any difference to discussions about what people ought to do and why? In many respects Madhyamaka ethics is just general Mahāyānist Buddhist ethics, no more no less. A radical and purely text-based answer thus might be to say flatly, “No, there’s no evidence in the texts that would suggest it makes any significant difference at all.” One could point out that the position Nāgārjuna and other Indian authors seem to espouse is that canonical ethical distinctions remain thoroughly intact in Madhyamaka, all be they transposed from the level of ultimate (paramārtha) to conventional reality/truth (saṃvṛtisatya). The monastic rules, bodhisattva precepts, love, compassion, and the attention to the law of karma and its often unfathomable consequences remain unchanged. In the auto-commentary to Nāgārjuna’s Vigrahavyāvartanī verses 7-8 the opponent is depicted as arguing that without real moral intrinsic natures, the typical Abhidharma list of virtuous and non-virtuous mental factors (caitta) could not exist, nor could there be any liberating (nairyāṇika) tropes (dharma) or any of the other factors needed for the path. Nāgārjuna, in reply, doesn’t contest anything within
4 “That ‘why’ is important: moralizers are happy to tell you what you ought to do – moral philosophers differ in that they aim to tell you also what makes it the case that you ought to do the things they say you ought to do (Thomson 2001: 6).”

Madhyamaka Buddhist ethics

361

the classifications of what is virtuous, non-virtuous, liberating and binding. Instead he seeks to show that the whole Abhidharma-style list remains possible for a Mādhyamika providing it is suitably transposed to the proper level of truth. While he thus consecrated very significant efforts to showing that ethics would not be simply precluded in toto by a philosophy of emptiness (śūnyavāda), he did not seem to even entertain the idea that some important aspects of ethics or ethical reasoning would have to be affected or that some new approach to ethical questions would be demanded.5 Often it is said that understanding Buddhist ultimate reality leads to one being convinced of the interconnectedness of all life and that it hence reinforces environmental ethics or universal responsibility; these are themes frequent in popular presentations. In fact this is not what I shall focus upon for the simple reason that talk of interconnectedness and the transformative effect of realizations of the ultimate is found in writings of virtually all Mahāyānist schools and is not linked exclusively, or even principally, to the Madhyamaka. The potential changes I wish to take up stem from systemic tensions in Madhyamaka positions on worldly reality. I insist on “systemic” to emphasize that what is at stake is rational reconstruction of the system of Madhyamaka thought and not the discovery of some hitherto unknown textual data. (Not surprisingly, rational reconstructions are predominant nowadays in discussions of Buddhist ethics and especially so when it comes to applied ethical issues, like contraception, responsibilities to future generations, environmental ethics, human rights, etc., that don’t have clear textual discussions in canonical literature.)6 The Mādhyamikas’ śūnyavāda should, if carried through, have significant implications
5

As Jan Westerhoff succinctly put it: “[Analyses] dealing with the specific ethical consequences of Madhyamaka thought are virtually absent” (Westerhoff 2009: 209).” 6 Rational reconstruction seems to be what e.g. Damien Keown (2005) is doing in taking up Buddhist positions on issues such as cloning and others that medieval Indian Buddhists certainly were not aware of. Another striking example: investment guidelines for ethical investing by Buddhists in the stock market. The result of the deliberations of Richard Gombrich et al. (2007) is what you find in the Dow Jones Dharma Index.

362

Tom J. F. Tillemans

for their conceptions of what worldly reality is and on the epistemology that governs knowledge claims and justification on the level of worldly truths. For ethics, this means a break with certain types of justificatory reasoning that Buddhists use on the worldly, or conventional (saṃvṛti), level. In a chapter on ethics in a book on Indo-Tibetan Buddhist notions of conventional truth/reality, Bronwyn Finnigan and Koji Tanaka (2011) also examine justification in a Madhyamaka approach to ethics, arguing that in a thoroughgoing anti-realist śūnyavāda genuine justification becomes impossible; in its stead we supposedly only find a weaker type of reasoning to tell Buddhists why they should act and think in certain ways rather than others as Buddhists. Now, it is true that in the Madhyamaka and other schools there are many discussions that are little more than homiletics and that involve reasoning which would not stand up, or even be intelligible, outside the church. Such is often the case in Candrakīrti’s first five chapters of Madhyamakāvatāra (often cited by Finnigan and Tanaka); it is also what we find in the enormously complicated SvātantrikaMādhyamika scholastic treatments of Prajñāpāramitā ethical schemata discussed in the Indian commentarial literature centered on the Abhisamayālaṃkāra or in the corresponding Tibetan literature. Nonetheless, not all Madhyamaka ethical reasoning is purely or even essentially destined for the already committed Buddhist – far from it. In texts like Bhāviveka’s Madhyamakahṛdaya, we find direct polemical confrontations with non-Buddhists to show that their ethical pronouncements are wrong and unjustified, and that the Buddhists’ views alone are justified. There is also a recurring insistence, by Svātantrika-Mādhyamika authors aligned with Dharmakīrti’s school, such as Śāntarakṣita and Kamalaśīla, that ethical argumentation should be accessible to an open-minded rational being who provisionally suspends religious commitment; this is the so-called “judicious person” (prekṣāvat) who represents an ideal figure in that she adopts positions, ethical and otherwise, purely on the basis of sound justificatory reasoning alone. It is thus incontestable that there were such rational strategies for justification in non-Madhyamaka and Madhyamaka Buddhist ethics alike. Mādhyamikas not only thought they were important, but that they functioned unproblematically on the level of conventional truth/re-

Madhyamaka Buddhist ethics

363

ality.7 Of course, one could still maintain that Indo-Tibetan Mādhyamikas were somehow badly wrong about this and that an appeal to conventional truths/realities in justificatory reasoning simply is not compatible with anti-realism at all, philosophically speaking. But this, if right, would end up being the very strong claim that there is a fatal flaw in the numerous sorts of anti-realism East and West, namely, that such philosophies and ethical justification just simply can’t mix a priori. I don’t see that that general position is established in Finnigan and Tanaka (2011) or elsewhere. I think that Finnigan and Tanaka (2011) were certainly on the right track in focussing on issues of justification. Where I disagree with them, however, is in their position that the Mādhyamikas’ anti-realism constrains these thinkers to a type of ethics without justification. We’ll leave aside the Svātantrika-Mādhyamikas, like Kamalaśīla et al., who clearly rely heavily on the robust justificatory reasoning of the Epistemological School (pramāṇavāda) transposed onto the domain of conventional truth. The more interesting case is that of the Prāsaṅgika-Mādhyamika. Does Finnigan and Tanaka’s position apply better here? Briefly: I don’t think we end up with no justification; on the contrary, justification for a Prāsaṅgika can remain strong and not simply an affair of the church preaching to the faithful. Justification would, however, become significantly different from what it is generally in Buddhist ethics. This is because evaluation of actions largely in terms of their humanly unfathomable karmic consequences – which is a major part of Buddhist ethical reasoning – becomes especially problematic for the Prāsaṅgika. What problems does a Prāsaṅgika have with this that others don’t? The Prāsaṅgika, more than other Buddhists, is caught between accepting that justification proceeds via the usual tallying of karmic consequences, on the one hand, and a fundamental methodological constraint, on the other, stemming from their views on
7 The debate about the possibility or impossibility of justification and its compatibility with śūnyavāda is certainly not a new one, and the Madhyamaka answers follow Nāgārjuna’s well-worn strategy of arguing that ethical reasoning functions on the level of the conventional. See Nāgārjuna’s replies to the objections in Vigrahavyāvartanī 7-8.

364

Tom J. F. Tillemans

metaphysics. This methodological constraint is that conventional truth and the reasoning based upon it must, in some sense, be in conformity with the world’s ideas and intuitions – Prāsaṅgikas must remain in keeping what the world acknowledges (lokaprasid­ dha), and should not, as do the other Buddhist schools, propose radical alternatives.8 Related to this is the idea that argumentation must proceed in keeping with principles recognized (prasiddha/ abhyupagata) by the other party. Indeed Prāsaṅgika-Mādhyamikas notoriously claim that a consequence of their śūnyavāda is that they see no reason to go fundamentally against the world’s views, and thus stress that they accept what the world accepts and reject what it rejects as conventional truth, as for a śūnyavādin there is nothing deeper grounded in real intrinsically existing facts.9 The prescriptive accounts of the other schools, accounts that seek largely to change what the world thinks in order to better conform to entities, end up for the Prāsaṅgika as being a reformulation of realism. Let’s look at the details. (I will rely on some previous publications giving a more detailed treatment of the Indo-Tibetan ideas and the textual data; the point here is to assess the implications for ethics). What an emphasis on conformity with the world and its ways of reasoning means for Buddhist ethics is, in effect, that the world’s fundamental moral intuitions, epistemic practices and norms are

8 I have taken up the idea of lokaprasiddha and its problems of interpretation in Tillemans 2011. There are obviously better and worse exegeses, the worst being that truth simply is equated with what the majority of people in the world believe to be true. 9 The Prasannapadā of Candrakīrti cites a famous passage from the Ratnakuṭasūtra to show that Madhyamaka acknowledges just what the world acknowledges (lokaprasiddha): “The world (loka) argues with me. I don’t argue with the world. What is generally agreed upon (saṃmata) in the world to exist, I too agree that it exists. What is generally agreed upon in the world to be nonexistent, I too agree that it does not exist.” loko mayā sārdhaṃ vivadati nāhaṃ lokena sārdham vivadāmi / yal loke ’sti saṃmataṃ tan mamāpy asti saṃmatam  / yal loke nāsti saṃmataṃ mamāpi tan nāsti saṃmatam. Trisaṃvaranirdeśaparivarta (chapter 1) of the Ratnakūṭa. The source is traceable back to Saṃyutta Nikāya III, p. 138. Sanskrit found in Candrakīrti’s Prasannapadā 370,6-8 (ed. L. de la Vallée Poussin, St.-Pétersbourg 1903/1913).

Madhyamaka Buddhist ethics

365

reinstated and legitimized as the grounds for justification.10 And indeed Mādhyamikas do often justify their ethical positions in this way. A good example of Prāsaṅgika ethical argumentation that proceeds in terms of the world’s moral intuitions and that uses justificatory reasoning destined for the unconvinced is found in the first four chapters of the Catuḥśataka (CŚ).11 In this text of the third century C.E. Mādhyamika Āryadeva and in the Catuḥśatakaṭīkā (CŚṬ) of Candrakīrti (6th century) we find an elaborate discussion of the four “illusions” (viparyāsa) that, according to canonical literature (i.e., Abhidharma), are supposedly present in the minds of worldlings: taking transitory life as permanent, what is painful as pleasurable, what is dirty as clean and what is selfless as having a self. The argumentation follows a pattern. The authors try to show that the world’s superficial attitudes on these matters are in conflict with its deep-seated intuitions – if the world reflected it would recognize the four illusions as indeed illusions. Candrakīrti here, in effect, uses the methodology he uses everywhere else, i.e. a form of prasaṅga-method using reductio ad absurdum and “opponentaccepted reasons” (paraprasiddhahetu) invoking principles that the world implicitly accepts, and will recognize upon reflection.12
10 On the tension between moral theory and intuitions, see McMahan 2000. 11 Translated and studied in Lang 2003, Sanskrit fragments in Suzuki 1994. 12 To résumé the critique in the CŚ and CŚṬ concerning the illusions (viparyāsa): the ordinary person’s confidence about the future is based on self-deception about his mortality (CŚ I.6-7); worldlings’ attitudes to mourning are inconsistent (’gal ba = viruddha) so that they mourn what on reflection does not deserve it (CŚ I.13); pleasure and happiness are rare, contrary to widespread opinion; upon reflection we see it is actually pain that is prevalent (CŚ II.4); people might think that work is a source of happiness, but it is better seen as largely meaningless and slavish exertion to survive (CŚ II.18); attitudes about beauty and cleanliness are confused and would be seen to be wrong if we reflected upon them (CŚ III.3-5); possessiveness makes no sense (CŚ III.11); kings (and other so-called “superior individuals”) are more like social parasites, dependent upon others’ work – they have no reason to feel justified of their status (CŚ IV.2); a king who is violent, corrupt or cruel deserves to be denounced, even though he claims to provide protection or to be the “father of the people” (CŚ IV.11-13), and so on and so on.

366

Tom J. F. Tillemans

Not only did Candrakīrti in CŚṬ appeal to propositions that the world recognized, but he proposed an account of the pramāṇas (reliable sources of knowledge) that stayed as close as possible to worldly epistemic practices; he deliberately rejected the prescriptive epistemology of the Dignāga school. Indeed in chapter XIII of CŚṬ he ridiculed prescriptive epistemologists as being “completely unversed in mundane objects” and “intoxicated through imbibing the brew of dialectics” (Tillemans 1990, vol. 1: 177, 179). The commitment to lokaprasiddha in all things epistemic and ethical is abundantly clear. What difference would such a professed conformity with the world’s moral intuitions and epistemic practices make to Buddhist ethics? It is obvious that Āryadeva and Candrakīrti were seeking to rationalize and conserve intact a canonical ethical schema and that they did not see their prasaṅga-method nor their epistemology as in any way placing it in jeopardy. However, unlike the CŚ’s discussion of the four illusions, which is largely argumentation about intuitions and does not involve controversial facts, much of Buddhist ethical argumentation, does crucially rely on problematic facts, typically when actions are evaluated because of their total set of karmic consequences across several lives. While some karmic consequences will be accessible to ordinary reflecting individuals – e.g., the general rule that he who lives by violence tends to die by it – many will be completely unfathomable by any ordinary human beings, in that they are supposedly unobservable and not inferable from anything observable. These so-called “radically inaccessible facts” (atyantaparokṣa) – like the details of why one comes to have the particular destiny and rebirth one has – are thus supposedly only understandable through scriptures authored by individuals with extraordinary understanding (Tillemans 1999 and 2000). In later Indian and Tibetan Buddhism this extraordinary understanding is increasingly taken to be full omniscience: knowing everything about all. Although Prāsaṅgika-Mādhyamikas professed lokaprasiddha across the board on conventional matters, they certainly did not, for all that, abandon recourse to radically inaccessible facts about karma to justify ethical positions. In a sense, they tried to have the

Madhyamaka Buddhist ethics

367

usual Buddhist cake and eat it, and the result was unsatisfactory. Buddhists, Mādhyamika or not, were certainly aware that persuading people by citing scriptures that weren’t understood or believed in, or by invoking the idea that the Buddhists’ teacher was a superior individual (atiśayapuruṣa) with supra-normal understanding, were indeed highly problematic in the eyes of the world.13 However, instead of simply siding with the world on this, Mādhyamikas sought an ingenious argumentation strategy whereby critical thinkers could supposedly come to accept scriptural propositions and accept that they be used as justificatory reasons in ethical debate. This Buddhist strategy goes back at least to Āryadeva, after which it is taken up by Dignāga (5th century) and Dharmakīrti (7th century); most modern Tibetan Buddhists continue to promote it as being a critical approach leading to rational proof of otherwise inaccessible facts; it is regularly espoused by the Fourteenth Dalai Lama. In brief, the procedure in the “triple analysis” (dpyad pa gsum) and “scripturally based inferences” (āgamāśritānumāna)14 is that one first ascertains the truth of a scripture’s pronouncements on all the observable matters (pratyakṣa) it treats as well as on those matters which are rationally accessible (even though unobservable) – an immaculate record in these two categories allows one to infer the scripture’s accuracy on otherwise completely unknowable matters like karma. This method is indeed present in Āryadeva’s Catuḥśataka XII.5, which tells us that when, in an ethical deliberation, there is doubt about the veracity of the Buddha’s descriptions of completely obscure karmic consequences we should be confident in his teachings about them because he was right in other areas, notably the teaching on emptiness.
13 For example, Dharmakīrti in PV I.218 says that we would accept what such a superior person says “if we could know that he is superior” (śakyeta jñātuṃ so ’tiśaya yadi), the point being that short of us having clairvoyance we simply couldn’t know who had such extraordinary knowledge and who didn’t. 14 On the so-called “triple analysis” (dpyad pa gsum) and “scripturally based inferences,” see chapters 1 and 2 in Tillemans 1999; see also the introduction to Tillemans 1993.

368

Tom J. F. Tillemans

buddhokteṣu parokṣeṣu jāyate yasya saṃśayaḥ / ihaiva pratyayas tena kartavyaḥ śūnyatāṃ prati // “When someone entertains doubt concerning the imperceptible things (parokṣa) taught by the Buddha, he should develop conviction in these very things on account of emptiness (śūnyatā).”

The point is that we supposedly can, with our own critical acumen, determine that the teachings on emptiness are an example (dṛṣṭānta) where the Buddha got the facts perfectly right. Therefore, because of his reliability on something essential like emptiness, it is also rational to believe his statements even when we cannot ourselves determine their truth. As Candrakīrti puts it in his ṭīkā to Catuḥśataka XII.5:
“Now the [adversary] cannot state even the slightest reason for any uncertainty, and thus this example [i.e., emptiness] is indeed proven. Therefore, you should understand by means of your very own principles alone (svanayenaiva) that the other statements of the Illustrious One, which establish unobservable states of affairs, are also true, for they were taught by the Tathāgata, just as were the statements setting forth [that] state of affairs which is the emptiness of intrinsic nature. How then could there be any place for doubt concerning the imperceptible things taught by the Buddha?”15

In short, this (and the slightly more elaborate later idea of āgamāśritānumāna) is an appeal to the Buddha’s track-record in teaching. It’s a justificatory argument that one finds in Prāsaṅgika-Mādhyamikas like Candrakīrti, who commented upon Āryadeva, as well as in Svātantrika-Mādhyamikas like Kamalaśīla and Śāntarakṣita, who relied on Dharmakīrti’s method in Pramāṇavārttika (PV) I to prove unfathomable facts inferentially. What is striking for our purposes is that Candrakīrti says that the appeal to the Buddha’s reliability proceeds “by means of your very own principles alone (svanayenaiva).” It is clear that Candrakīrti wishes to say that this
na ca śakyam anena svalpam apy aniścayakāraṇaṃ kiṃcid abhidhātum iti siddha evāyaṃ dṛṣṭāntaḥ  / tataś cānyad apy asamakṣārthapratipādakavacanaṃ bhagavato yathārtham iti pratīyatāṃ svanayenaiva tathāgatopadiṣṭatvāt svabhāvaśūnyatārthābhidhāyakavacanavad iti kuto buddhokteṣu parokṣeṣu saṃśayāvakāśaḥ / Translation and text in Tillemans 1990 vol. 1: 120 and vol. 2: 17-19..
15

Madhyamaka Buddhist ethics

369

argument strategy is in conformity with the world and its own epis­ temic norms. Candrakīrti, and perhaps Āryadeva, and certainly later Tibetan writers, thus thought that a Madhyamaka philosophy which took conventional truth as “what is acknowledged by the world” (lokaprasiddha), could also engage in a method that stretched the world’s acceptance to include propositions about unfathomable things. And this is where one gets the distinct impression of a stratagem to have the cake and eat it too. Is Candrakīrti’s idea nonetheless somehow defensible? Do we actually conform to the world’s own epistemic standards and practices if we invoke such track-records to persuade the unconvinced and doubt-ridden on matters where there are not, and indeed supposedly cannot, be any other empirical evidence or rational arguments. I don’t think so. First of all, it’s implausible to claim that any major scripture will actually pass the test of getting all empirical and all other humanly verifiable matters right. It might get some right, but not all, given that empirical knowledge changes and grows. True, Buddhist writers, like Dharmakīrti and probably Āryadeva and Candrakīrti too, recognized that what really counted was not a track-record of one-hundred percent accuracy on every conceivable trivial matter treated in a scripture or treatise, but accuracy on important or difficult principal topics, such as the four noble truths, emptiness and the like.16 But what kind of rational “conviction” (pratyaya) could one develop in this way? How would all doubts be eliminated, as Āryadeva and Candrakīrti say, by reminding oneself of the rightness of the Buddha’s teaching on emptiness, for surely the mere fact of getting one very important thing right is not ipso facto a guarantee on anything else. Dharmakīrti was quite skeptical about these “scripturally based inferences,” and said clearly in his Pramāṇavārttikasvavṛtti to PV I.217 that
See PV I.217: “Because he/it is reliable on the principal matters, we can infer [reliability] on the others” (pradhānārthāvisaṃvādād anumānaṃ paratra). For Dharmakīrti and his commentators the principal matter is the four noble truths. It is striking that Āryadeva and Candrakīrti speak of the significance of getting emptiness right: it seems clear that the Mādhyamika in CŚ XII.5 is using a reasoning very similar to PV I. 217, but that the principal matter is indeed emptiness.
16

370

Tom J. F. Tillemans

they weren’t actually bona fide inferences at all – na ... anumānam anapāyam – as they lacked certainty (niścaya). As he recognized, there are just too many counterexamples where a person is correct on one set of things, be they important or not, but falls down hopelessly on other things. (See e.g., his Svavṛtti to I.318: 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.”) Any observer of human foibles can come up with examples where people get significant things right and fall down on less significant and even comparatively easy matters.17 Now, the first impression one gets in reading this type of critique of scripturally based inference is that the problem raised is just the usual one of the fallibility of induction: we cannot arrive at certainty about the truth of generalizations on the basis of a finite number of confirmations. Indeed, the weakness of induction is a constant theme in Dharmakīrti. But if it were only a matter of the well-known failings of induction to generate certainty, one might well reply, “So much the worse for Buddhist demands of certainty,” and thus become a resigned or even cheerful fallibilist. In that case a fallibilist could say that demonstrated accuracy in certain important matters constitutes at least reasonable grounds for supposing accuracy in another duly related type of matter. After all, it is indeed so that we proceed by such transfers of credibility in many cases where someone’s proven expertise in one set of matters serves as grounds (all be they fallible) for trusting him or her in another closely related set. (One need only think of expert witnesses in court cases to see that this is indeed common practice.) A suitably fallibilist Candrakīrti would then at least be right in saying that this is in keeping with the world’s epistemic standards. Arguably, though, the problem with scriptural inference is not simply the ever-present uncertainty one finds in inductive reasoning. And so merely embracing fallibilism is not the remedy either.

To update things a bit: the math department may be brilliant on the significant theoretical aspects of topology but unable to add up their phone bill correctly.

17

Madhyamaka Buddhist ethics

371

The real catch in transferring credibility from one area to another, if we wish to conform to the world’s own norms, is that there must be a type of connection or relevant similarity between the different areas so that the expertise is transferable in a reasonable, albeit fallible, fashion. We are all too familiar with people who are highly qualified in one area – say, economics – and who then think they can expound on virtually everything else, related or not, and usually with disastrous results. We would say that those would-be experts become unreliable, not because they hadn’t at some point understood a lot of significant things rightly or because of some general problem about induction, but because they overstep their qualifications. They take on subjects not clearly related to the area in which they have been recognized to be reliable. Alas, the flaw in a fallible appeal to the Buddha’s track-record is similar: it is not at all clear that reliability concerning important general 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,18 – knowing that much would hardly suggest that one somehow knows the specific details of what causes what. Now, while Āryadeva and Candrakīrti seem to have thought that scriptural inferences were an unproblematic way for Buddhists to argue with any opponents, Buddhist or not, in that such inferences were just another case of the world’s extending credence to people on the basis of their past performances, Dharmakīrti’s position was no doubt much more nuanced. Dharmakīrti, in Pramāṇavārttika I and IV, maintained that this type of faith-based reasoning about the specifics of karma would not be persuasive to non-believers; not only is the reasoning extremely uncertain, but many opponents would simply refuse to recognize its subject matter (dhar­ min), i.e., Buddhist accounts of the details of karmic causality.
See Mūlamadhyamakakārikās XXIV.18ab: yaḥ pratītyasamutpādaḥ śūnyatāṃ tāṃ pracakṣmahe “Dependent arising, that we declare to be emptiness.”
18

372

Tom J. F. Tillemans

Nonetheless, Dharmakīrti’s view was that scriptural inferences, even if they weren’t fully bona fide, were still somehow to be used, at least in the “private” context of Buddhists who wished to set out (pravṛttikāma) on the spiritual path. While scripture may provide direction (albeit very fallible) in a closed context of believers, it is not at all probative (i.e., a sādhana) in a public context when appeals to authority are contested. I think that some such distinction between private and public is indeed important to the Buddhist position. On the one hand, Dharmakīrti would steer clear of outright denial of karma – which no scholastic Buddhist writer can just flatly deny – but on the other hand he knows that he cannot and should not use scriptural descriptions of karmic consequences to clinch debates in the public domain, where rebirth and the causality between lives are either contested issues or too obscure to be admitted in “fact-based” (vastubalapravṛtta) debates. How might such a public-private distinction work? If we take another type of subject matter, viz. the nature of mind, there is a great difference between a discussion in a private domain, amongst convinced Buddhists, that cites sūtra or tantra passages on the luminous nature of mind, the subtle consciousness and so forth, and such a discussion in the public domain where the other party is a well-meaning, but non-Buddhist, cognitive scientist working in a secular university. To mix up the private and public domains and say that Buddhist scriptural quotations about otherwise inaccessible features of mind should also be probative for the cognitive scientist would be seen as a rather comical violation of norms of rationality. And if one then goes from bad to worse and persists in somehow saying that the scripture is in any case right because it is the words of the Buddha, etc., this would be seen as not far from fundamentalism. Now, instead of cognitive science, let’s take an example of an ethical argument on a contested issue in applied ethics: animal welfare and vegetarianism. People on both sides can and will invoke considerations that are publicly debatable: harm to animals, ecological consequences, health benefits of eating meat or not eating meat, suffering, perhaps rights of animals, speciesism, etc. This is recognizably normal argumentation on such an issue. Things are much different when someone seeks to prove their case by a scriptural quote concerning the unfathomable karmic con-

Madhyamaka Buddhist ethics

373

sequences of eating meat. For example, a vegetarian may invoke passages from the Laṅkāvatārasūtra to the effect that a meat-eater will be reborn as a carnivore, or alternatively a meat-eating Tibetan Buddhist might reply to a vegetarian opponent by invoking a (tantric?) teaching that the meat-eater will establish a “karmic connection” with the animal whose flesh he eats and will in a future life lead the being to enlightenment, etc., etc. Perhaps such ways of thinking would still somehow be meaningful in the private context of committed Buddhist exegetes wondering what their scriptures advise. However, they would certainly become deeply suspect if they left the purely private context of intra-Buddhist religious exegesis. Clearly there are huge philosophical problems about holding any dual perspectives, be they in philosophy of mind or on ethical and religious issues. While private perspectives do sometimes co-exist with what is publicly acceptable, it is another matter as to whether and when they should co-exist, if at all.19 The question thus remains whether some dual perspectives are to be preserved, and which ones, or whether norms of rationality dictate that we always seek unity. Dharmakīrti, if I read him rightly, thought that at least in ethics a purely intra-Buddhist perspective on some key matters, while publicly unarguable, could be more than simple irrationality. For our purposes, however, we shall have to leave that larger philosophical problem on hold. In any case, the shift to such a private perspective will not enable scriptural argumentation about karma to be used convincingly in an adversarial debate on ethics, as Āryadeva and Candrakīrti thought it could be used and as Buddhists often do try to use it. Instead, following the world’s norms, this would be an illegitimate shift from one perspective to the other, i.e., an attempt to promote certain Buddhist ethical ideas

It is of course no secret that many intelligent, scientifically minded peoscientifically ple hold weird private ideas – e.g. on medicinal remedies, on ways to ensure good luck, on CIA conspiracies, etc., etc. – ideas that they know would be unacceptable in a public discussion with their peers. Such types of dual perspectives, however, are not of much interest to us here, as they would usually be dismissed as simply irrational.

19

374

Tom J. F. Tillemans

in the public domain all the while insulating them from criticism by stressing people’s epistemic incompetence. Recently the philosopher Owen Flanagan (2006), in a lecture on Buddhism and science, confronted much the same problem where Buddhists priced certain problematic subjects outside critical debate: he coined the term “epistemological protectionism.” (This may capture the problem better than the accusations of dogmatism that are sometimes bandied about). Such a protectionist approach would claim that one could cite reasons – e.g, scriptural passages – to provide proof in a public debate, but at the same time preserve them from criticism by making them unassailable to anyone but the largely convinced alone. I think it is clear that in a critical approach to ethics appeals to humanly unfathomable facts are also protectionist in this way. They too fail to conform to what the world, on reflection, values and demands in debate on contested issues, namely, that a discussion, if not between the already convinced, must be open for both parties to criticize and evaluate on the basis of publicly accessible information. Once one puts into question the Mādhyamika’s use of scriptural inferences to give knowledge of unfathomable karmic consequences, his appeal to the omniscience of the Buddha is also going to be in considerable trouble. In both cases one is pricing much ethical debate outside the realm of criticism, or worse, engaging in a type of protectionism. However, there are also other reasons as to why a Mādhyamika should feel particularly uncomfortable in invoking a literal notion of omniscience to clinch an ethical debate. The problem arises that if the Buddha supposedly has knowledge of all things in all their details, including the unfathomable effects of karma, this looks to be tantamount to him understanding how things are in themselves, by their own intrinsic nature (svabhāvena). And that would be precluded by a Mādhyamika’s thorough-going antirealism – realist Buddhist schools might have some way to accept it, but it seems that the Madhyamaka would be in an especially delicate position. Literal omniscience seems in effect to be a “God’s Eye point of view of the world,” a view sub specie aeternitatis, and

Madhyamaka Buddhist ethics

375

that is one way to formulate what an anti-realist holds to be impossible.20 Of course, the subject of omniscience and the different views that Buddhists have had about it historically are matters upon which I cannot reasonably embark here. In later Indian Buddhism and in its Tibetan successors we find in one way or another several versions as to what full-fledged omniscience is, and some versions would be closer to a crypto-realism than others. (There has also been, over time, an evolution from a limited conception of omniscience where an omniscient being supposedly knows one or a few essential things about everything, to a much stronger type of knowledge.)21 For our purposes we’ll contrast two quite different versions of omniscience. If we take, for example, a conception such as that of the 14th century Tibetan writer Klong chen rab ’byams pa who distinguishes between “dualistic mind” (sems) and “primordial gnosis” (ye shes) or Mind itself (sems nyid), omniscience is taken as primordial gnosis/Mind itself rather than “dualistic mind” – dualistic mind is in fact something to be eliminated.22 Omniscience for Klong chen pa, like primordial gnosis, is a special type of understanding free from all objects (yul). In such a case, there seems to be no question of an omniscient being knowing each and every thing in all its details – that would per impossibile be a sort of dualistic mind (sems). Gnosis/omniscience is a type of transcendent understanding in which the very idea of a thing/object is absent. On the other hand, if we take omniscience as it seems to be understood by later Indian commentators on Dharmakīrti or by Tibetan
The characterization of metaphysical realism as involving the God’s Eye perspective is due to Hilary Putnam (see, e.g., Putnam 1981: chapter 3). 21 See McClintock 2010. 22 Cf. Klong chen rab ’byams pa, Sems dang ye shes kyi ’dri lan, gSung thor bu, p. 384: mdor bsdu na khams gsum pa ’i sems sems byung cha dang bcas pa thog ma med pa nas brgyud pa ’i bag chags can sgrib pa gnyis kyi ngo bo ’dzin cing / bskyed par brten pas spang bya yin zhing dgag dgos par bshad pa yin no / “In short, the three realms’ dualistic minds (sems = citta), mental factors (sems byung = caitta), and their qualities are subject to imprints coming down from beginningless [time], have as nature the two obscurations and rely on production. As such, it is explained that they are to be eliminated and should be stopped.”
20

376

Tom J. F. Tillemans

Mādhyamika writers such as Tsong kha pa, it does seem to involve essentially an amplification ad infinitum of people’s cognition of objects to arrive at all of them in all their details. It is this latter, more literal, version that would seem to run thoroughly afoul of anti-realism. Where does this critique of the Mādhyamika appeals to trackrecords and omniscience leave us? I think we have to admit that Prāsaṅgikas attempted the impossible in professing lokaprasid­ dha all the while trying to recuperate unfathomable karma in ways that the world itself would supposedly accept. My argument so far has been that this is not likely to work and that some hard choices therefore have to be made. Now, one of the more attractive features of Buddhism is, and has been for a long time, its openness to discussion and its emphasis on reason. We see this strongly in the present Dalai Lama’s commitment to dialogue based on empirical methods and rationality – what Flanagan (2006) called a “welcome mat” to open, public, debate.23 What I think needs to be offered in ethics, in largely the same constructive spirit as that of Owen Flanagan’s article on science, is a sober note: this welcome mat will attract few long term visitors if contemporary Buddhists, Mādhyamikas included, continue to justify their ethical views on the basis of facts knowable only via a scripture whose author had omniscient knowledge. What then would śūnyavādin Buddhist ethics look like without such a heavy reliance on radically inaccessible facts and God’s Eye omniscience? I think a partial answer is something like the following: if moral intuitions and personal attitudes become more dominant in ethical debate, rather than scripture and omniscience, ethics becomes humanized and to quite a degree secularized; the extended welcome mat will be genuinely attractive. The opposite tendency however will keep Buddhist ethics oriented towards religious
Here is how Flanagan (2006) reformulated the Dalai Lama’s invitation to dialogue. “The Welcome Mat: ‘Come sit by my side, my Western scientific and philosopher friends. Tell me what you know. I will teach you what I know. We can debate. But in the end it is our duty, on both sides, to change our previous views if we learn from the other that what we believe is unfounded or false’.”
23

Madhyamaka Buddhist ethics

377

fundamentalism and increasingly out of touch with contemporary concerns. Both these competing visions were in fact combined in various degrees in Buddhist schools, just as they are in most major world religions. The important feature of the Madhyamaka for Buddhist ethics is that it, more than any other school, insists on the primacy of what the world acknowledges, rather than upon humanly unfathomable facts that are as they are for all time. This, if carried through, would be a significant move away from fundamentalism and towards the humanization of ethics.

Bibliography Primary sources
CŚ CŚṬ PV Catuḥśataka, see Suzuki 1994, Tillemans 1990. Catuḥśatakaṭīkā, see Suzuki 1994, Tillemans 1990. Pramāṇavārttika, chapter on inference for oneself (svārthānumāna). See PVSV. PVSV Pramāṇavārttikasvavṛtti. Ed. R. Gnoli, The Pramāṇavārttikam of Dharmakīrti. The First Chapter with the Autocommentary. Rome, 1960. Mūlamadhyamakakārikās. Ed. L. de la Vallée Poussin. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna, avec la Prasannapadā commentaire de Candrakīrti. Osnabrück: Biblio Verlag, 1970. Sems dang ye shes kyi ’dri lan of Klong chen rab ’byams pa. In Kun mkhyen klong chen pa dri med ’od zer gyi gsung thor bu. Reproduced from xylographic prints from A ’dzom ’brug pa chos sgar blocks. 2 volumes. Pema Thinley, Gangtok, 199?.

Secondary sources
Finnigan, B.; Tanaka, K. 2011. “Ethics for Mādhyamikas.” In: The Cowherds (eds.). Moonshadows. Conventional Truth in Buddhist Philosophy. Oxford: Oxford University Press, pp. 221-231. Flanagan, O. 2006. Science for Monks: Buddhism and Science. Written version of the keynote lecture to the symposium on “Mind and Reality” at Columbia University. See www.usc.edu/schools/college/crcc/private/flanagan_lectures/Science_for_Monks.pdf, last visited 23-02-2011. Garfield, J. L. (2011). “What is it like to be a Bodhisattva? Moral pheno-

378

Tom J. F. Tillemans

menology in Śāntideva’s Bodhicaryāvatāra.” Journal of the International Association of Buddhist Studies 33 (2010) [2011] 333–357. Gombrich, Richard et al. 2007. Dow Jones Indexes and Dharma Investments to Launch New Faith­Based Indexes. See http://www.asria.org/news/ press/1201231179, and

http://www.djindexes.com/mdsidx/html/pressrelease/press_hist2008. html, last visited 23-02-2011.
Goodman, C. 2009. Consequences of Compassion. An Interpretation and Defense of Buddhist Ethics. Oxford: Oxford University Press. Keown, D. 2005. Buddhist Ethics: A Very Short Introduction. Oxford: Oxford University Press. Lang, K. C. 2003. Four Illusions. Candrakīrti’s Advice to Travelers on the Bodhisattva Path. Oxford: Oxford University Press. McClintock, S. L. 2010. Omniscience and the Rhetoric of Reason:

Śāntarakṣita and Kamalaśīla on Rationality, Argumentation, and Religious Authority. Boston: Wisdom Publications.

McMahan, J. 2000. “Moral Intuition.” In: LaFollette, H. (ed.). The Blackwell Guide to Ethical Theory. Oxford: Blackwell, 92-110. Putnam, H. 1981. Reason, Truth and History. Cambridge: Cambridge University Press. Suzuki, K. 1994. Sanskrit Fragments and Tibetan Translation of Candrakīrti’s Bodhisattvayogācāracatuḥśatakaṭīkā. Tokyo: Sankibo. Thomson, J. J. 2001. Goodness and Advice. Princeton: Princeton University Press. Tillemans, T. J. F. 1990. Materials for the Study of Āryadeva, Dharmapāla and Candrakīrti. 2 volumes. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien. Tillemans, T. J. F. 1993. Persons of Authority. Stuttgart: Franz Steiner Verlag. Tillemans, T. J. F. 1999. Scripture, Logic, Language: Essays on Dharmakīrti and his Tibetan Successors. Boston: Wisdom Publications. Tillemans, T. J. F. 2000. Dharmakīrti’s Pramāṇavārttika. An Annotated Translation of the Fourth Chapter (Parārthānumāna). Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Tillemans, T. J. F. 2011. “How Far Can a Mādhyamika Buddhist Reform Conventional Truth? Dismal Relativism, Fictionalism, Easy-Easy Truth, and the Alternatives.” In: The Cowherds (eds.). Moonshadows. Conventional Truth in Buddhist Philosophy. Oxford: Oxford University Press, pp. 151-165. Westerhoff, Jan. 2009. Nāgārjuna’s Madhyamaka. A Philosopical Intro­ duction. Oxford: Oxford University Press.

Miracles and superhuman powers in South and Southeast Asian Buddhist traditions
Contributions to a panel at the XVth Congress of the International Association of Buddhist Studies Atlanta, 23–28 June 2008 Guest editor

David V. Fiordalis

Miracles in Indian Buddhist narratives and doctrine1
David V. Fiordalis
“False prophets can bring about false miracles through their magic arts, which are the same in appearance as true miracles wrought by true prophets through divine power.” Pierre d’Ailly, De falsis prophetis2

Despite the fact that scholars have recognized for a long time that Buddhist literature contains numerous marvelous and fantastic events, there have been reservations about the use of the word “miracle” in the context of Buddhism. When the word has been used to speak about Buddhism, scholars have tended to note that Buddhist miracles are not “miracles in our Western sense,” speaking as though there were a single Western understanding of the concept that could be used to measure the Buddhist understanding. By contrast, scholars have generally been less reticent to speak of magic and magical powers in Buddhism. What they generally mean by preferring magic to miracle is that, in Buddhism, extraordinary powers are not thought to be violations of natural law. They
1 I would like to thank Luis Gómez and John Strong for reading earlier drafts of this essay and making helpful comments. I also want to thank them, as well as Bradley Clough, Kristin Scheible, Rachelle Scott and Patrick Pranke, for their participation in a panel at the IABS Congress in Atlanta, 2008, dedicated to the exploration of miracles and superhuman powers in South and Southeast Asian Buddhist texts and traditions. That conference panel and the essays currently collected in this issue of the Journal, which are based on the presentations made in Atlanta, testify to what we feel has been an enjoyable, timely and fruitful confluence of scholarly interest in the topic of the wonders and wonderment in Buddhist literature and beyond. 2 Quoted in Caciola 2006: 295.

Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 381–408

382

David V. Fiordalis

can be acquired through the use of mantras or meditation, and these are thought to be quite natural processes. Therefore, goes the argument, it is more appropriate to speak of marvels or wonders or magical powers in Buddhism.3 More work is needed to determine whether this generalization accurately portrays the complex and nuanced Buddhist discourse on the miraculous. One should note initially that these terminological preferences strongly reflect the Western history of the concepts of miracle and magic. These concepts have complicated and intertwining histories, having been conceived in tandem for many centuries. As the 14th century French Catholic theologian Pierre d’Ailly suggests in the quote above, discerning between miracles and magic has not always been easy. Traditionally, “miracle” has had a positive connotation, used to refer to wondrous events that a particular tradition, the Christian tradition, holds to be authentic and authoritative, while magic has been used pejoratively to refer to other traditions (or others within one’s own tradition whose motives and authority one wishes to question). The application of the concept of magic has therefore been criticized on the grounds that it implies skepticism about the claims and motives of “others.”4 In a connected but distinct usage, magic also evokes images of maFor example, chapter twelve of Andy Rotman’s recently published translation of the Divyāvadāna is the Prātihārya-sūtra, which Rotman translates “The Miracle Sūtra.” In a footnote (Rotman 2008: 429, n. 580), he cites T. W. Rhys Davids’ opinion that Buddhist miracles are not “miracles in our Western sense. There was no interference by an outside power with the laws of nature. It was supposed that certain people by reason of special (but quite natural) powers could accomplish certain special acts beyond the power of ordinary men.” See Rhys Davids 1899, Vol. 1: 272. Thus, throughout his translation, Rotman speaks of the Buddha performing “miracles” (prātihārya) by means of his “magical powers” (ṛddhi), but for a Medieval Christian reader like Pierre d’Ailly, this would seemingly imply the falsehood of the displays. 4 For one recent discussion and some useful references, see Burchett 2008. For another critique of the anthropological category of magic, see also de Sardan 1992. For the use of this terminology in modern anthropology, cognitive theory and psychology of religion, see, for instance, Rosengren, Johnson, and Harris, eds., 2000, and the various publications of Pascal Boyer. For references to Boyer’s work and more, see Luis Gómez’s discussion and bibliography in his article below.
3

Miracles in Indian Buddhist narratives and doctrine

383

gicians who entertain and amaze the gullible with illusions and trickery. Yet, in our Post-Harry Potter age, who would argue that “the magical” does not also have positive senses in many people’s minds, and that the term does not overlap considerably with miracles and marvels in describing a broad range of “supernatural” phenomena? These terms exist within a constellation of concepts that includes the natural and supernatural, divine and diabolical, as well as science and religion.5 Their complex, interrelated history ought to make us pause before we sketch our generalizations too broadly across cultures, time periods and religious traditions. As in Western discourse, one finds no single Buddhist understanding of the miraculous, but a plurality of voices and perspectives that likely changed over time. These Buddhist voices share a common, though complex and differentiated vocabulary, which they employ to describe an array of different types of “miracles,” “magic arts” and “superhuman powers.” Rather than eliminating one or another of our terms for translation at the outset, we need to consider all our vocabulary as we seek to understand the Buddhist discourse on the miraculous and to convey our understanding to others. Our words may still prove insufficient, but keeping our options open will at least allow us to show that Buddhist literature does sometimes draw its own distinctions between miracles and magic, even though these distinctions are inconsistently maintained. This may be seen though analysis of the common threefold classification of miracles in non-Mahāyāna Buddhist literature.6 In some Buddhist sources in Pāli and Sanskrit, rival ascetics call the Buddha a magician. In the Upāli-sutta of the Majjhimanikāya, for instance, a Jain ascetic tries to dissuade Mahāvīra from sending his lay disciple, Upāli, to refute the Buddha’s doctrine, saying: “For the ascetic (i.e., the Buddha) is a magician (māyāvin). He knows a concealing magic by which he deceives the disciples of
For further explorations of this history, in addition to Caciola’s work cited above, see also Flint 1994 and Bartlett 2008. 6 Here, the designation “non-Mahāyāna” also excludes Tantra. For recent treatment of the concepts of miracle and magic in the Vimalakīrtinirdeśa and selected Mahāyāna sūtras and śāstras, see Fiordalis 2008 and 2012. See also Luis Gómez’s article in the current volume of JIABS.
5

384

David V. Fiordalis

other ascetics’ teachers.”7 In the Sphuṭārthavyākhyā, Yaśomitra’s sub-commentary on the Abhidharmakośabhāṣya, the question is raised, “How do the other rival ascetics, such as Maskari Gośāla and others, criticize the Buddha?” Yaśomitra then gives two citations:
In a treatise of the Nirgranthas, it is said: “Who displays his superhuman powers (ṛddhi)? The magician Gotama does.” Also, they say, “Every hundred ages a magician of this type appears in the world and cause the people to be consumed by his magic.”8

In these examples, being called a magician carries a pejorative connotation, but it also portrays the grudging respect and resentment that the Buddha’s rivals appear to feel towards him, at least as seen through the eyes of the Buddha’s own followers. The Buddha is called a magician, because he displays his superhuman powers, but the quotes also suggest that the Buddha succeeds at winning a large following by doing so. One may contrast such accusations with the Buddhist monastic rule that prohibits monks and nuns from displaying their superhuman powers in front of laypeople. This well-known rule appears in the Cullavagga of the Pāli Vinaya, where the Buddha criticizes the Arhat, Piṇḍola Bhāradvāja, for using his superhuman powers to fly up and retrieve a sandalwood bowl placed upon a scaffold by a skeptical layperson.9 Alternate versions of the story
MN i.375. AKVy 319. 9 Vin ii.112. In fact, the episode elicits two rules in the Pāli Vinaya. Monks and nuns are prohibited from displaying superhuman powers in front of the laity, and they should not use alms-bowls made of wood. This creates an ambiguity about precisely why the Buddha found Piṇḍola Bhāradvāja’s act objectionable. Another version of this story, found in the Dhammapada commentary, provides more insight into the motivations of the characters, but interestingly remains ambiguous on precisely why the Buddha found the act objectionable. In that version, Piṇḍola Bhāradvāja seems to display his powers in order to prove the validity of the dharma, and not for the sake of the bowl, but when the Buddha asks Piṇḍola why he displayed his powers, no response is given before the Buddha prohibits the display of superhuman powers. See Dhp-a iii.203. In both these versions, but far more strongly in the Vinaya, the Buddha appears to prohibit displays of superhuman powers
8 7

Miracles in Indian Buddhist narratives and doctrine

385

are also found in the Vinaya collections of the Dharmaguptakas, Mahīśāsakas, and Sarvāstivādins, as well as the Pāli commentary on the Dhammapada, where it prefaces the narrative cycle of the Buddha’s miracles at Śrāvastī and Sāṅkāśya.10 Yet, despite the fact that the Buddha claims in the Cullavagga that “miraculous displays of superhuman powers (iddhi-pāṭihāriya) will not generate faith in those without faith, nor increase the faith of the faithful,” and castigates Piṇḍola for displaying his powers for the sake of a mere bowl, many exceptions to the rule appear throughout Buddhist literature. Do these many exceptions justify the claim made by rival ascetics (through Buddhist sources) that the Buddha did, in fact, display his superhuman powers in order to win disciples, or at least that Buddhists told many stories that he did? Certainly, there are many stories of the Buddha’s miracles, and some doctrinal discussions of miracles and superhuman powers put forth the view that such displays can and did bear positive results. However, another scripture that criticizes such displays, and that has garnered a fair amount of attention from scholars in search of Buddhist views on the miraculous, is the Kevaṭṭa-sutta of the Dīgha-Nikāya.11 It has been suggested that this scripture expresses a “rationalistic” perspective,12 which it may, but it also may be said to distinguish between true miracles and false miracles or magic. In it, a lay follower named Kevaṭṭa suggests to the Buddha that he instruct his monks to perform “miraculous displays of superhuman power” (iddhi-pāṭihāriya) or “superhuman feats” (uttarimanussa-dhamma) so that the prosperous people of Nāḷandā will
in front of laypeople, because he feels that Piṇḍola retrieved the bowl out of covetous desire for it. 10 For a French translation of the Vinaya versions, see Chavannes and Lévi 1916: 233–247. For an English translation of the version in the Dhammapada commentary, see Burlingame 1921, Vol. 30: 35–38. For another discussion of Piṇḍola, see Strong 1979. 11 DN i.211ff. The alternate spelling of the title and of the name of the Buddha’s chief interlocutor in the discourse is Kevaḍḍha. 12 Gómez 1977: 221. The passage does indeed seem to rationalize the wondrous in one sense, and that is by classifying it. For other discussions of the Kevaṭṭa-sutta, see Gethin 1987: 187, and Granoff 1996.

386

David V. Fiordalis

develop even more faith in him.13 The Pāli commentary tells us what Kevaṭṭa might have had in mind: flying above the city and performing a pyrotechnic display for the townspeople below. The implication seems to be that the laypeople of Nālandā who bear witness to such a display would give even more support, material and otherwise, to the Buddhist saṅgha. The Buddha’s initial response sounds somewhat reluctant, if not downright perfunctory: “I don’t give such instruction to the monks, saying, ‘Go, monks, and perform miracles or superhuman feats for the white-clothed laypeople.’” Kevaṭṭa is persistent, however, and after he asks a third time, the Buddha gives the following explanation:
Kevaṭṭa, I have declared that there are three types of miraculous display, having directly realized them by my own higher knowing (abhiññā). What are the three? They are the miraculous display of superhuman powers (iddhi-pāṭihāriya), the miraculous display of telepathy (ādesanā-pāṭihāriya), and the miraculous display of instruction [in the dharma] (anusāsanī-pāṭihāriya).

Here the Buddha mentions one of the most common classifications of miracles in Buddhist literature. It is found in a wide range of sources, with an important alternate list also attested in the Mahāvastu.14 This standard threefold typology also appears in the
13 The Pāli term, iddhi, for which the Buddhist Sanskrit is ṛddhi, not sid­ dhi, has no simple equivalent in English. Since it is derived from the root, ardh, to grow, the term literally means success or flourishing. My preference for translating it as “superhuman power,” at least in this context, derives partly from the fact that Buddhist texts like this one sometimes gloss the term with uttarimanussa-dhamma, which can be translated more literally as a superhuman quality or characteristic. Another text that does so is the Prātihārya-sūtra of the Divyāvadāna. For a different, though related, sense of the term uttarimanussa­dhamma, see the fourth pārājika rule in the Pāli Vinaya (Vin iii.91ff.), which dictates that the punishment for those who lie about possessing “superhuman qualities” that they do not really possess is expulsion from the monastic order. 14 See, for instance, Ps ii.227–229. The Mahāvastu contains the same threefold listing as above (Vol. 1: 238), but also attests a variant list (Mv iii.137–138): “miraculous displays of superhuman power” (ṛddhi-prātihārya), “miraculous displays of instruction” (anuśāsanī-prātihārya), and “miracu-

Miracles in Indian Buddhist narratives and doctrine

387

Abhidharmakośabhāṣya and the Catuṣpariṣat-sūtra, as we will see below. As the dialogue with Kevaṭṭa continues, the Buddha criticizes the display of superhuman powers and telepathic ability:
[The Buddha said, “Suppose] someone who has faith and trust sees [a monk] doing these things. He tells this to someone else lacking in faith and trust, saying, ‘Isn’t it marvelous, sir, isn’t it amazing (acchariyaṃ vata bho, abbhutaṃ vata bho), the great power and great might of this ascetic!’ The one who lacks faith and trust would say, ‘It is only by means of a Gandhārī spell (Gandhārī nāma vijjā) [or] a Maṇikā amulet (Maṇikā nāma vijjā) that he can perform [such things].’ What do you think, Kevaṭṭa, wouldn’t someone lacking in faith and trust say that to the man who possesses faith and trust?” “Reverend Sir, he would say that,” [answered Kevaṭṭa.] “That is why,” [the Buddha responded,] “I see danger in [such miraculous displays], and am troubled, ashamed and disgusted [by them].”15

Here, the Buddha raises doubts about the efficacy of displaying superhuman powers to impress skeptical people. He then goes on to laud teaching of the dharma, apparently suggesting that it is the true miracle. Is the Buddha of this discourse speaking metaphorically when he calls teaching the dharma a type of miracle? If not, then how can teaching the dharma be considered a type of miracle? Perhaps the discourse means to align the display of superhuman powers with magic, separating the act of teaching the dharma from “superstition” in the same way that “rational science” is sometimes distinguished from “magic” (and “religion”). Although plausible, one problem with this interpretation is the fact that the semantic range of vijjā (Skt: vidyā), the term translated here as “magic,” includes the variety of mundane “sciences,” as well. Another problous displays of teaching the dharma” (dharma-deśanā-prātihārya). This alternate suggests, perhaps, that the most significant underlying opposition in the tripartite scheme of miracles is that between superhuman powers and teaching the dharma. 15 DN i.212–214. For stylistic reasons, my translation condenses two nearly identical exchanges into one.

388

David V. Fiordalis

lem is that the Kevaṭṭa-sutta does not suggest that the Buddha and other Buddhist monks do not actually possess various types of superhuman powers. Rather, it implies the opposite, while raising the question of how a skeptical person can discern between true miracles and false miracles or magic, an issue that we have seen also occupied Medieval Christian theologians. It would seem that both Buddhists and non-Buddhists accused each other of being magicians. Here, the Kevaṭṭa-sutta suggests that “magical powers” are ubiquitous, and thus their display does not necessarily prove the superiority or uniqueness of the Buddha and his message, as teaching the dharma seems to do. The Abhidharmakośabhāṣya also discusses the three types of miracles, arguing like the Kevaṭṭa-sutta that teaching the dharma is the best kind of miracle.16 While the reasoning of the Kevaṭṭasutta is compressed into a few statements, Vasubandhu elucidates two lines of argument to explain why the miracle of teaching the dharma should be considered the best of the three miracles. Firstly, echoing the logic and terminology of the Kevaṭṭa-sutta, he argues that the first two types of miraculous display are also achievable by means of “magic spells.” For instance, the Gāndhārī vidyā gives one the power of flight, while the Īkṣaṇikā vidyā grants the power to know the thoughts of others. Not so for the miracle of teaching the dharma; one cannot teach the dharma simply by using a spell. Vasubandhu draws a parallel between the three types of miracle and the various capabilities that the Buddha achieves as part of his awakening. It is not possible, Vasubandhu argues, to perform the miracle of teaching the dharma without truly possessing the knowledge of the destruction of the defilements (āśrava-kṣaya-jñāna), which is the third of the so-called three “knowledges” (vijjā, vidyā) and the sixth of the so-called “super-knowledges” or “superpowers” (abhiññā, abhijñā).17

Chapter seven, verse forty-seven and commentary. AKBh 868–869. Note as well that the Bodhisattvabhūmi also articulates this parallel between the three types of “miracle” and the three relevant “super-knowledges.” See BoBhū, p. 54. For a detailed analysis of the abhiññās in Pāli canonical literature, see Bradley Clough’s article in the present volume of JIABS.
17

16

Miracles in Indian Buddhist narratives and doctrine

389

However one classifies it, the knowledge of the destruction of the defilements seems to involve both certainty about one’s own attainments and true knowledge of the nature of reality (yathā-bhūtajñāna). Such knowledge forms the basis for teaching the dharma, and perhaps this is why descriptions of it often make reference to basic Buddhist doctrines like the Four Noble Truths. Teaching the dharma would then provide evidence that a person has truly achieved “sainthood” (arahantā, arhattvā), for it is the unique possession of buddhas and Buddhist saints. Thus, for Vasubandhu, teaching the dharma is the best form of miracle, because it provides evidence of true sainthood, being based on knowledge of the true nature of reality. To this extent, Vasubandhu seems to support the reasoning of the Kevaṭṭa-sutta. However, Vasubandhu gives a second line of reasoning to explain why the miracle of teaching the dharma is the best of the three miracles. He argues that the first two types of miraculous display are useful only for impressing people as to what is preeminent (pradhāna-āvarjana-mātra).18 With the miracle of teaching the dharma, however, it is possible to obtain the preeminent, what one truly desires, the ultimate good (hita), freedom from suffering. “For,” Vasubandhu writes, “it is said that true ‘success’ (ṛddhi – the term I have been translating ‘superhuman power’) is [achievable] only through teaching the means [of achieving freedom from suffering].”19 Vasubandhu’s second line of argument seems to be that teaching the dharma is foremost among the miracles, because it will ultimately lead the faithful person along the path to the ultimate good. While teaching the dharma is still seen by Vasubandhu as the best type of Buddhist miracle, he also seems to accept on this second line of reasoning that miraculous displays of superhuman powers and telepathic ability can play an important role. They are efficacious for the purposes of initial conversion. This more moderate

18 For further discussion of this phrase, see Luis Gómez’s article below (Gómez 2011: 517). 19 anuśāsanaprātihāryeṇa tu hitena iṣṭena phalena yogo bhavaty upāyopadeśād ity evāvaśyaṃ ṛddhir ity ucyate. AKBh 869.

390

David V. Fiordalis

position on the display of superhuman powers is also indicated by Vasubandhu’s discussion of the meaning of the word, prātihārya, the word that I have been translating as “miracle.” Vasubandhu defines prātihārya as “at the outset, carrying away (haraṇa) people who are ready to be disciplined (vineyamanas).” He explains the verbal prefix prāti- as a combination of two prefixes, pra + ati, the former signifying “the beginning” and the latter “extreme intensity.”20 Or, Vasubandhu tells us, miracles are called prātihārya because they “seize” (pratiharanti) people who hate or are indifferent to the dharma.21 One may doubt the philological accuracy of these etymological explanations, but there can be little doubt that they are intended to draw a clear connection between miracles and religious conversion.22 In these Buddhist discussions of miracles and superhuman powers, one can perhaps hear an echo of another way in which miracles have sometimes been distinguished from magic. In the words of
20 vineyamanasāmādito ’tyarthaṃ haraṇāt prātihāryāṇi prātisabdayor ādikarmabhṛśārthatvāt. AKBh 869. 21 pratihatamadhyasthānāṃ manāṃsyebhiḥ pratiharantīti prātihāryāṇi vā. AKBh 869. 22 Here, the term “conversion” is used primarily to cover the sense of the Sanskrit terms given above, namely, prātihārya and āvarjana. Nevertheless, the use of the term also brings up issues of comparative analysis not dealt with here. Though a full exploration is beyond the scope of this article, it bears mentioning that the connection between miracles, displays of superhuman power and religious conversion in the sense used here is emphasized and further developed in Mahāyāna Buddhist narrative and doctrine. According to the Mahāprajñāpāramitā-upadeśaśāstra, the bodhisattva develops the “superpowers” (abhijñā) in the interest of other beings and performs miracles so that the minds of other beings may become pure. If the bodhisattva did not perform miracles, he would not be able to inspire as many beings to strive for the ultimate good. See Lamotte 1944–1980, Vol. 4: 1819–1820. The Bodhisattvabhūmi (BoBhū p. 46) also explains that one of the purposes of “superhuman power” (ṛddhi) is to introduce people into the Buddha’s teaching by “converting” (āvarjayitvā) them with a miraculous display. See Gómez’s discussion and translation of this passage in his article in this volume of JIABS (Gómez 2011: 531). There are nevertheless Mahāyāna critiques of displays of superhuman powers, particularly in the Chan/Zen traditions. See Gómez 1977 and Bielefeldt 2002.

Miracles in Indian Buddhist narratives and doctrine

391

Richard Swinburne, miracles must have “religious significance,” that is, they must “contribute significantly towards a holy divine purpose for the world.” By contrast, “extraordinary events lacking religious significance are more appropriately characterized as magical or psychic phenomena rather than as miracles.”23 While the distinction between religion and magic remains problematic (as does the term “conversion”), along with the meaning and application of these terms in a Buddhist context, some Buddhist narratives and doctrinal discussions seem to suggest that certain events are made miraculous by reason of their connection to the Buddha’s sacred mission to lead people beyond suffering. Acts of teaching the dharma and displays of superhuman powers are clearly among them. A miracle story from the Avadāna-śataka, a 1st century collection of Buddhist narratives, includes a passage on miracle suggestive of this sacred purpose. The tenth story of the eighth varga features a woman named Virūpa so distressed that she tries to commit suicide by hanging herself in a cave. Just as she is about to do so, the Buddha (apparently using his superhuman powers) becomes aware of her and emits a golden light from his body, which cuts the rope. At this point, the story lapses into a semi-scholastic list seemingly meant to explain the miracle. According to the story, six “conditions” (sthāna), when manifested in the world, “produce” (prādurbhāva) a “miracle” (āścāryādbhuta): “1) the Buddha, 2) the dharma and Vinaya taught by the Buddha, 3) a human being, 4) born in the land of the Āryas, 5) having all the sense organs in working order, 6) (and) freely desiring the good dharma.”24 Among other things, this list suggests that Buddhist miracles are not mere wonders, but somehow derive from the Buddha’s mission or purpose
23 Swinburne 1970: 8–9. One should also note that Swinburne, having made this point, proceeds with his analysis based on Hume’s definition of miracle as a violation of natural law. One may usefully compare with the definition and treatment of miracles found in MacCulloch 1908, Vol. 8: 676– 690. 24 Avś ii.55. The passage does not clearly differentiate the six conditions. My division reflects my own interpretation, but there are certainly other ways of dividing this list into six components.

392

David V. Fiordalis

to lead living beings to freedom from suffering. After performing the miracle, the Buddha teaches the dharma to Virūpa, who then becomes an arhantī. Notice how the story seems to integrate the three types of miracle (manifesting superhuman powers, telepathic ability, and teaching the dharma) within a single narrative. In order to get a fuller sense of a Buddhist miracle, the threefold typology of miracles found in the Kevaṭṭa-sutta and elsewhere ought to be compared with another typology present in the Pāli commentaries, but also found in a variety of forms elsewhere. This second typology lists specific events that constitute the skeleton of the Buddha’s final lifetime. For instance, in the commentary on the Mahāpadāna-sutta of the Dīgha-Nikāya, one finds the following statement:
[As Bodhisattvas in our final birth], we will display miracles (pāṭihāriya) that will, among other things, shake the earth, which is bounded by the circle of ten thousand mountains, when (1) the allknowing Bodhisattva enters his mother’s womb, (2) is born, (3) attains awakening, (4) turns the wheel of the dharma, (5) performs the “Twin Miracle” (yamaka-pāṭihāriya), (6) descends from the realm of the gods, (7) releases his life force, [and] (8) attains cessation.25

In his ground-breaking article on the miracle of Śrāvastī, Alfred Foucher highlights another important example of an eightfold series of such events, a 5th century stele found at Sārnāth with eight panels depicting a slightly different set of events.26 These lists of events in the last life of the Buddha can be related to a wider set of lists of varying lengths, including lists of four, five, six, twelve, and even a list of thirty. Classified under different names, from “miracles” (pāṭihāriya) to “wondrous and amazing things” (ac­
DN-a ii.412. It is interesting to compare this series of events with the list of causes of earthquakes found in the Mahāpārinibbāna-sutta (DN ii.108– 109). At some point between sutta and commentary, natural causes and the superhuman powers of śramaṇas, brāhmaṇas and gods have been replaced by the miracle of Śrāvastī and the descent from the Heaven of the Thirty-three. 26 Foucher 1909. English translation in Foucher 1917. For the image, see Foucher 1917, fig. 19.1. In the stele, the Buddha’s conception and the release of his life-force are replaced with episodes in which the Buddha tames the maddened elephant and receives a gift of honey from a monkey.
25

Miracles in Indian Buddhist narratives and doctrine

393

chariya-abbhuta-dhamma, āścārya-adbhūta-dharma) to “events that the  Buddha must perform” (avaśya-karaṇīya), these events collectively form what John Strong has called the “Buddha-life blueprint.”27 As the quote above suggests, the fact that earthquakes and other “wondrous and amazing things” occur during these events is one of the characteristics connecting them as miracles. Such signs and wonders help to indicate their “religious significance.”28 If one compares the threefold typology with this list of eight miracles, certain events come into focus, such as the first sermon and the miracle cycle at Śrāvastī and Sāṅkāśya, examples in which the Buddha teaches the dharma or displays his superhuman powers in order to win converts and further his mission. Another event that is not contained in the above lists, but nonetheless sheds an important light on the tension contained in the threefold typology of miracles, particularly between teaching the dharma and displaying superhuman powers, is the story of the conversion of the three Kāśyapa brothers. Described as a “thaumaturgical impasse,”29 the story of the conversion of the Kāśyapa brothers, three rival Brahmin ascetics with many followers of their own, occurs in the narrative cycle at the beginning of the Mahāvagga of the Pāli Vinaya. The sequence begins just after the Buddha’s awakening and tells the story of how the Buddha establishes his ministry and wins his first disciples. Thus, it includes other well-known events, such as the first and second sermons, as well as the conversions of Yaśas, King Bimbisāra, Śāriputra and Maudgalyāyana. The narrative cycle is also present in the Vinaya collections of other mainstream Buddhist schools and in the Catuṣpariṣat-sūtra, while the Mahāvastu contains an alternate version of the story of the three Kāśyapas.30 Comparing
John Strong 2001: 10–13. One might compare this with the Chinese concept of “resonance,” discussed in Kieschnick 1997: 96ff. 29 Gómez 1977: 222. 30 The passages in the Vinaya collections of the Mahāvihāra Theravāda (i.e. the Pāli version), Mahīśāsaka, and Dharmaguptaka have been ana28 27

394

David V. Fiordalis

several versions of this story not only clarifies the relationship that was thought to exist between the three types of miracles; it illustrates that different Buddhists conceived this relationship differently. The basic story of the impasse, as it appears in the Mahāvagga and in the Catuṣpariṣat-sūtra, is as follows. In his attempt to convert the eldest Kāśyapa brother, the Buddha performs many displays of his superhuman powers. After the performance of each miracle, which the Mahāvagga suggests numbered 3,500 in total,31 Kāśyapa thinks to himself, “Certainly, the great ascetic possesses great superhuman power (mahā-iddhika), great, wondrous presence (mahā-anubhāva)….but he is no saint (arahant) like I am.”32 Thus, the Buddha and Kāśyapa reach an impasse. After many displays of superhuman power, Kāśyapa remains unconverted. The point thus far appears to be that displays of superhuman power are insufficient to establish the superiority of the Buddha and convert the rival ascetic. Interestingly, however, the first of the Buddha’s displays of superhuman power in the Mahāvagga, and by visual accounts the one most closely associated with the occasion, sufficiently impresses
lyzed in detail by André Bareau (1963). In addition to these versions is the Caṭuṣpariṣat-sūtra, which is nearly identical to the version found in the Mūlasarvāstivāda-vinaya. Bareau does not include an extensive analysis of the Mūlasarvāstivāda-vinaya or the Caṭuṣpariṣat-sūtra. For an explanation of his reasons for not doing so, see Bareau 1963: 8–9. For the Sanskrit edition of the Caṭuṣpariṣat-sūtra, as well as the Tibetan, see CPS. For the Mahāvastu version, see Mv iii.425ff. 31 Vin i.34. The early translators, T. W. Rhys Davids and Hermann Oldenberg, suggest that the passage containing this reference to the total number of miracles was interpolated, perhaps due to the fact that the comment seems rather out of place in the flow of the narrative. See Rhys Davids and Oldenberg 1885, Vol. 3: 133. There is no such phrase in other versions of the story. See Bareau 1963: 316. Whether it is a later addition or not, the passage nevertheless emphasizes the fact that the Buddha performed many marvelous displays of superhuman power on the occasion. In the Catuṣpariṣatsūtra, the number of miracles totals eighteen. 32 Vin i.32. In the Catuṣpariṣat-sūtra, sometimes Kāśyapa alternatively says, “….but I am a saint like him” CPS 244. Other versions also alternate between the two phrases.

Miracles in Indian Buddhist narratives and doctrine

395

Kāśyapa that he invites the Buddha to stay with his community and agrees to provide him with material support. For his first miracle, the Buddha asks to spend a night in Kāśyapa’s fire-lodge. Kāśyapa is reluctant to allow it, because a fierce, fire-breathing snake also lives in the fire-lodge, and he is concerned the Buddha might be hurt. The Buddha insists, however, and Kāśyapa finally relents. So, the Buddha spends the night in the fire-lodge and succeeds in taming the fire-breathing serpent through a miraculous display of his power over the element of fire. The snake and the Buddha engage in a duel of fire, with the snake breathing smoke and fire, and the Buddha, by means of the “fire-element concentration” (tejodhātusamādhi), emitting flames from his own body, until finally the snake is subdued. The next morning, the Buddha emerges from the fire-hut with the serpent coiled up in his alms bowl, providing an image that is used to depict the event in Buddhist art.33 The language used in the Mahāvagga to describe Kāśyapa’s reaction to this miraculous display is important to consider. Kāśyapa becomes “serene (abhippasanna) as a result of this miraculous display of superhuman power” (iminā iddhipāṭihāriyena), and invites the Buddha to stay. One might say that he becomes receptive to the faith. The word I have translated as serene is merely an adjectival form derived from the Pāli term pasāda or prasāda in Sanskrit. Prasāda is a complex concept without a simple equivalent in English. It means faith and trust, but also beauty and serenity, as when the mind is unclouded and free of doubt. It also means mental receptivity to faith, as characterized by awe and veneration. The literal meaning of the term, prasanna, “settled,” also evokes a sense of purity and clarity. When dirty water has been allowed to sit and the sediment has sunk to the bottom, the clear water that remains is prasanna. Returning to our story, only the Pāli version contains this description of Kāśyapa’s reaction. Curiously, the entire episode of the taming of the snake is told twice in the Pāli, first in prose and then again in verse. In the prose portion of the Mahāvagga, as in the versions found in the Vinayas of other mainstream Buddhist
33

Rhi 1991: 71. For a few images, see Ingholt and Lyons 1957, figs. 87–89.

396

David V. Fiordalis

schools and in the Catuṣpariṣat-sūtra, Kāśyapa’s reaction remains merely skeptical. The Buddha may have power, Kāśyapa thinks, but he is not my equal, or alternatively, he is merely my equal. Only the sentence that concludes the versified version in the Pāli describes Kāśyapa as serene as a result of the miraculous display of superpower. By using the term, abhippasanna, the text may simply be trying to say that Kāśyapa is pleased with the Buddha’s performance, but it may also reflect the fact that Kāśyapa is becoming receptive to the Buddha’s advances. One could say that this miraculous display of superhuman power does what Vasubandhu suggests that such displays ought to do. It makes an initial impression that leads to Kāśyapa’s conversion. Kāśyapa is not converted by the snake-taming miracle, however. Only in the version found in the Mahāvastu is Kāśyapa converted by the snake-taming miracle, but that version differs from the others in several other important respects as well. For one thing, the snake-taming miracle occurs at the end of the sequence of miracles, not the beginning. Another important difference is that, in the Mahāvastu, the Buddha and Kāśyapa both display their superhuman powers in a wonderworking contest, which the Buddha eventually wins. Most versions suggest, on the contrary, that Kāśyapa does not display his own powers. His conversion occurs only after a series of unsuccessful displays of the Buddha’s superhuman powers. In the Mahāvagga, for example, the Buddha finally uses his telepathic powers to become aware of Kāśyapa’s continued insistence that he is equal to or greater than the Buddha, and determines that “the confused man will continue to think thus for a long time.” Therefore, the Buddha decides to “shock” (saṃvejayan) him.34 He says to Kāśyapa, “You are not a saint, Kāśyapa, nor have you attained the path to sainthood. The path you walk will not lead you to become a saint, nor will it lead you to the path to sainthood.”35 It
For a classic discussion of the concept of saṃvega, see Coomaraswamy 2004. For further discussion of the concepts of pasāda and saṃvega, and their importance in another interesting miracle story of the Buddha, the story of the Buddha’s journey to the island of Sri Lanka, see Kristin Scheible’s article in this volume. 35 Vin i.32.
34

Miracles in Indian Buddhist narratives and doctrine

397

seems that this trenchant demonstration of the Buddha’s ability to read Kāśyapa’s thoughts has the desired effect. In the Catuṣpariṣatsūtra, although no reference is made to the Buddha’s intentions, once the Buddha has contradicted him outright, Kāśyapa thinks to himself, “The great ascetic knows my mind with his mind!”36 He then converts to Buddhism. It may seem that the employment of telepathic ability is somewhat anti-climactic after the wondrous displays of so many superhuman abilities. Yet, the Buddha’s ability to read Kāśyapa’s mind, coupled with Kāśyapa’s inability to read the Buddha’s, provides a clear indication of his superiority.37 In the story of the Buddha’s miracles at Śrāvastī in the Divyāvadāna, it is stated that even lesser beings, such as tiny biting ants, can read the Buddha’s “mundane” (laukika) thoughts, but not even solitary buddhas, the Buddha’s disciples or the gods can perceive his “supermundane” (lokottara) thoughts.38 In any case, if an important theme of the story were the ineffectiveness of superhuman powers to convert Kāśyapa, then one would expect to find a clearer contrast between the displays of superhuman powers, telepathy and teaching the dharma. If one ignores the verse section and focuses only on the prose, then perhaps the Pāli version suggests such a contrast, but in the Mahīśāsakavinaya, for instance, the Buddha contradicts Kāśyapa while levitating in the air.39 If the point were to emphasize the ineffectiveness
CPS 302. The significance of the Buddha’s ability to read the minds of others in the Pāli nikāyas is amply demonstrated in Brad Clough’s article in this volume. 38 See Divy 161. Again, although it lies beyond the scope of the current article, the distinction between laukika and lokottara is highly operative in Buddhist discussions of miracles and superhuman powers, and it may be one of the best places to begin exploring Buddhist notions that roughly approximate Western distinctions between the natural and supernatural. I hope to explore this distinction in further detail in a future article, but for the moment one may refer to discussions of it in the articles by Pat Pranke and Brad Clough in the present volume. 39 Bareau 1963: 304. For discussions of the importance of levitation in South Asian religions, see Hocart 1923, Brown 1928, Mahony 1987 and Strong 1983.
37 36

398

David V. Fiordalis

of superhuman powers, then this would seem decidedly odd. Thus, alternate versions of the story suggest that not all Buddhists drew a sharp contrast between the different types of miraculous displays. In fact, although it is true that the Buddha converts Kāśyapa in the Mahāvagga and Catuṣpariṣat-sūtra by means of the display of his telepathic ability, the Buddha uses all three types of miracle in these stories, one after the other, and more or less equally. The Buddha first employs his superhuman powers, then utilizes his telepathic abilities, and finally teaches the dharma. Specifically, the Buddha preaches the famous “Fire Sermon.” The theme of fire is an important one throughout the story of the Kāśyapa brothers. The Kāśyapa brothers are said to make fire sacrifices before their conversion. The Buddha chooses to impress Kāśyapa by taming his fire-breathing snake, and he does so by relying on his own mastery of fire. Finally, the Buddha returns to the theme of fire in the Fire Sermon, expressing the teaching that “everything is on fire” (sabbaṃ ādittaṃ). Here burning is made into a metaphor for the fact that all conditioned things are conjoined with passion, hatred, and confusion, the three root afflictions. So, the Buddha uses the metaphor of fire to teach the fire-worshipping Kāśyapa brothers about detachment. It seems that the editor of the Catuṣpariṣatsūtra also recognized the importance of fire in the story, because the description of the Buddha’s awakening that prefaces the text specifically mentions the fire-element concentration, the only description of the Buddha’s awakening to do so, so far as I am aware. After hearing this teaching, the three Kāśyapa brothers and their followers all become Arhats. Thus, from the various versions of the story, it seems not only that superhuman powers and telepathic abilities are effective and possibly necessary (though maybe insufficient) means of conversion, but also that teaching the dharma is a type a miracle insofar as it leads people to the “greatest good,” to use Vasubandhu’s phrase. Although the Pāli version does not explicitly use the typology of three types of miracle to refer to the various events that occur in the story, the Catuṣpariṣat-sūtra and alternate versions in other Vinaya collections do structure the narrative in terms of the three types of miracles, with the Fire Sermon explicitly described as a miracle of teaching the dharma.

Miracles in Indian Buddhist narratives and doctrine

399

When the Fire Sermon is described as a miracle of teaching the dharma, something curious also occurs in the description of the miracle of mind-reading in the Catuṣpariṣat-sūtra. If one compares it with the description of the miracles of mind-reading and teaching the dharma in the Kevaṭṭa-sutta, it appears that the two types of miracles have been condensed into one, or perhaps vice versa. Here is the description of the miracle of mind-reading in the Catuṣpariṣat-sūtra, restored with the help of the Tibetan translation of the Mūlasarvāstivāda-vinaya:
Monks, this is your mind (cittam). This is [your] mind (manas). This is [your] consciousness (vijñānam). Consider (vitarkayata) this. Don’t consider that. Think about (manasikuruta) this. Don’t think about that. Abandon (prajahata) this. Don’t abandon that. Having taken up (upa­ saṃpadya) this and realized [it] directly with the body, practice it.40

Now, compare this with the description of the miracles of mindreading and teaching the dharma in the Kevaṭṭa-sutta. First look at the description of mind-reading:
This, Kevaṭṭa, [is the miracle of mind-reading ability.] A monk points out (ādisati) the very mind (cittaṃ pi) of another being, another person. He points out the very mental state (cetasikaṃ pi). He knows the gross thoughts; he knows the subtle thoughts (vitakkitaṃ pi ādisati; vicāritaṃ pi ādisati). [And he says,] “This is in your mind (cittam). This is your mind (mano). Your mind is here.”

And here is the description of the miracle of instruction or teaching the dharma in the Kevaṭṭa-sutta:
Consider (vittakketha) this. Don’t consider that. Think (manasikaro­ tha) this way. Don’t think that way. Renounce (pajahatha) this. Having taken up this (idam upasampajja), practice [it].41

The parallel phraseology of these passages is striking in that the Catuṣpariṣat-sūtra includes, as one single passage, a slightly condensed, but clearly parallel version of the explanations of both the second and the third types of miracle, which are so strongly

40 41

CPS 320. DN i.213–214.

400

David V. Fiordalis

opposed in the Kevaṭṭa-sutta.42 This parallel raises some difficult questions about the intertextual history of the tripartite scheme, as well as the interrelationship between the three types of miracles, particularly the miracles of mind-reading and instruction. In this context, we should recall the alternate classification of the three miracles found in the Mahāvastu, where the second and third types of miracles are the “miracle of instruction” (anuśāsanī-prātihārya) and the “miracle of teaching the dharma” (dharma-deśanāprātihārya), respectively.43 This classification seems to bear a strong resemblance to the descriptions of the three types of miracles that we find in the Catuṣpariṣat-sūtra. What are we to make of these alternate classifications and descriptions of the three types of miracles, and can we assume beyond doubt that one is necessarily earlier or later than the others? The story of the conversion of the Kāśyapa brothers in the Mahāvagga and Catuṣpariṣat-sūtra depicts the three miracles in an ordered sequence. While the Kevaṭṭa-sutta suggests that some Buddhists tried to create a tension between displaying superhuman powers or telepathic abilities and teaching the dharma, the story of the conversion of the Kāśyapa brothers suggests that not all Buddhists perceived this tension in the same way. Some may not have perceived any tension, at all.44 In the case of the so-called
In addition, a sequence of miracles more or less corresponding to those of the Catuṣpariṣat-sūtra occurs on the 11th, 12th and 13th days, respectively, of the miracle at Śrāvastī in the version found in the Dharmaguptaka-vinaya. Indeed, on the thirteenth day, the Buddha is said therein to preach the Fire Sermon. See Takakusu and Watanabe, eds., 1924–1932, Vol. 22, number 1428: 949c–950a. English translation is in Rhi 1991: 234–235. 43 See n. 14 above. 44 Indeed, this point can be strengthened by returning to the Kevaṭṭa-sutta. As we now possess it, the sutta appears to contrast the “miraculous display of instruction” (anusāsanī-pāṭihāriya) with the miraculous displays of superhuman powers and telepathy, as we saw above, but then directly afterwards it equates the miracle of instruction with various other attainments that are commonly found throughout the suttas of the Dīgha-Nikāya. Among them are mastery of the four states of meditation (jhāna), the ability to create mind-made bodies, the various superhuman powers (iddhi), the other “superknowledges” (abhiññā), including telepathy and so forth. All these attain42

Miracles in Indian Buddhist narratives and doctrine

401

“Twin Miracle” (yamaka-pāṭihāriya), performed by the Buddha at Śrāvastī, one finds another instance in which the three types of miracles appear to be employed equally and more or less simultaneously. Here is a pertinent passage from the Dhammapada commentary’s description of the event:
On that day, the Teacher walked up and down performing the [Twin] Miracle, and as he did so, he taught the dharma to the audience, not wearying them with uninterrupted discourse, but giving them sufficient opportunity to refresh themselves from time to time. Thereupon, the audience sent up shouts of applause. Hearing the shouts of applause that arose from the audience, the Teacher straightaway looked into the hearts of the great multitude, and in sixteen ways perceived the disposition of mind of each one. So quick is the movement of the mind of the Buddhas, that in case any person took pleasure (pasanna) in any teaching or in any display of superhuman power, the Buddha taught the dharma and displayed his superhuman power in accordance with the temper and disposition of each person. As he thus taught the dharma and displayed his superhuman powers, a great multitude of living beings obtained clear comprehension of the dharma.45

Here, displaying superhuman powers, telepathic abilities and teaching the dharma are combined into one miraculous display, with the passage concluding by emphasizing its great efficacy. At this point in the story, the Buddha again appears to read the minds of those in the audience, and thus concludes that no one is capable of asking him appropriate questions. So, he magically produces a double of himself, and he and his double engage each other in a dialogue on the dharma, resulting in many more millions
ments, including those the miraculous display of which the Buddha criticizes just beforehand, are equated with the miracle of instruction in the Kevaṭṭasutta. Despite this fact, virtually all modern translations, from Rhys Davids onwards, have been abbreviated in such a way as to make this contradiction less apparent. For the redactors of the Dīgha-Nikāya, however, perhaps there was no clear contradiction between displaying superhuman powers and teaching the dharma. For an alternative discussion of this issue and different conclusions, see Meisig 1993. I owe John Strong for this reference and for impressing upon me the irony of this apparent contradiction in the text. 45 Dhp-a iii.215. My translation modifies slightly that of Burlingame 1921, Vol. 3: 47.

402

David V. Fiordalis

of beings attaining comprehension of the dharma. Thus, on one interpretation, even the Twin Miracle itself combines the display of superhuman power and the act of teaching the dharma. While the miracle performed at Śrāvastī is arguably the paradigmatic example in Buddhist literature of the Buddha displaying his superhuman powers, and despite the fact that the display is greatly embellished in this and other versions of the story, later artistic representations of the event continue to depict the Buddha in the gesture of teaching, perhaps indicating that teaching the dharma remained an important component of the miracle in the minds of some Buddhists.46 This essay has explored doctrinal discussions of the three types of miracles, using this typology to analyze a few Buddhist miracle narratives, which in turn, simultaneously clarify and obscure the tripartite classification. Not all Buddhist miracle tales refer to the three types of miracles, however. Nor do all of them seek to distinguish displays of superhuman power or telepathic ability from acts of teaching the dharma, or to criticize the display of the former. Among other things, miracle tales of the Buddha are concerned with demonstrating in one way or another his superiority vis-à-vis divine beings of various types, Buddhist disciples of various levels of attainment, rival teachers, and others, while highlighting the religious significance of the Buddha’s sacred mission.47 Oftentimes, these goals are achieved precisely through the display of superhu46 Many such images from Gandhāra are listed in Rhi 1991: 197–206. For a few examples, see Ingholt and Lyons 1957, figs. 252ff.; Kurita 1988–1990, Vol. 1, figs. 400–401, Vol. 2, fig. 79; Foucher 1917, figs. 19.1, 20, 21.1, 23.1–2, 25.2–26.2; Huntington 1986, figs. 2–4. Brown 1984, figs. 6, 9–10. 47 John Strong has also studied a range of stories that feature the failure of the power of flight on the part of Buddhist monks and even the Buddha himself in a former birth. One would think that such stories would work against the idea that the Buddha and his monks and nuns are superior to others who possess superhuman powers. Rather than reducing the prestige of the Buddha and his monks, however, Strong argues that such stories also emphasize their superiority, paradoxically, by indicating that their powers are the real thing. See Strong 1983. A revised and unpublished English version, “When Magical Flight Fails: A Study of Some Indian Legends about the Buddha and his Disciples,” was presented by Professor Strong at the University of Michigan on March 28, 2008.

Miracles in Indian Buddhist narratives and doctrine

403

man powers. Rivals may be exposed as magicians or frauds, while the Buddha dominates or outpaces them with his own superhuman powers, which are greater, in terms of raw power, than their own.48 In a similar vein, miracle tales featuring eminent Buddhist monks and nuns who display their superhuman powers may, in addition to demonstrating their own attainments, establish by extension the preeminence of the Buddha, his teachings and institutions.49 In conclusion, work remains to be done to understand the nature and significance of “miracles,” “magic,” and “superhuman powers” throughout a fuller range of Buddhist narrative and doctrine. Yet, scholars have been too quick to conclude from a few scant criticisms of displaying superhuman powers that Buddhism rejects the miraculous wholesale in favor of some sort of rational humanism that reflects modern predilections. For one thing, making this argument requires that one disregard the many Buddhist stories in which the Buddha or his eminent disciples perform acts of conversion by displaying their superhuman powers. Scholars have suggested that such stories are merely “popular” or represent “later” (often a euphemism for degenerate) traditions. Yet, these are problematic conclusions. This investigation has focused primarily upon what are, by most accounts, elite narrative and scholastic traditions. Some of the stories considered here, such as those from the Vinaya collections of the mainstream Buddhist schools, may even date from a relatively early period.50 Just how early is difficult to say and
An example of this is the story of Sirigutta and Garahadinna, found in the Dhammapada commentary (Dhp-a i.435–445). The similarity of this story to the Mahāyāna sūtra known as “The Prophecy of the Magician Bhadra” (Bhadramāyākara-vyākaraṇa) has been recognized by Konstanty Régamey (1990). 49 The story of Piṇḍola Bhāradvāja in the Dhammapada commentary is one such example. See n. 9 above. For another, see Rachelle Scott’s discussion in this volume of the important story of Mahāpajāpatī Gotamī’s nirvāṇa, which was translated for the first time into any Western language by Jonathan Walters (1995). 50 There has been some scholarly debate on the age of the relevant section of the Vinaya relative to other portions of the collection. For one theory, see Frauwallner 1956: 154. For a critical assessment of his theory, see Lamotte
48

404

David V. Fiordalis

I will not proceed to speculate here, but more and more, scholars of Buddhism are coming to see the “popular” reflected in the “elite” and vice versa. When we calibrate our reading of Buddhist stories to the terms that Buddhists use, then even ostensibly familiar events, like the first sermon, the paradigmatic example of teaching the dharma and also an act of conversion, can take on new shades of meaning. The other major argument that has been used to deny the significance of miracles in Buddhism is comparative. If a miracle is defined as a violation of the laws of nature by a supernatural agent, then perhaps displays of superhuman power in Buddhism do not fit this definition. Yet, one may question the adequacy of this definition, even with respect to a Western context. There are other important elements to the concept of the miracle, such as the wondrous signs and portents that accompany such events. Above all, one could argue, miracles provide evidence of the beneficence and holiness of the agent. That is, they have religious significance for the world. When defined in such a way, this concept of miracle then becomes relevant to Buddhism. However, discerning miracles from magic also then becomes important. The ambivalence towards the display of superhuman powers in Buddhist literature derives, at least in part, from questions about its evidentiary value in a world in which it is difficult to discern between “true” miracles and “false” ones. While the distinction is not always clearly maintained, we have seen that Buddhist literature does indeed sometimes distinguish between miracles and magic in order to argue for the supremacy and unique holiness of the Buddha, and by extension, his doctrines, institutions and eminent disciples.

References
Primary sources
AKBh Abhidharmakośabhāṣya s. AKVy.

1988: 177–179.

Miracles in Indian Buddhist narratives and doctrine AKVy

405

Avś

BoBhū CPS

DN DN-a

Dhp-a Divy

MN Mv Ps Vin

Śāstrī, Svāmī Dvārikādāsa, ed. 1998. The Abhidharmakośa & Bhāṣya of Ācārya Vasubhandhu with Sphutārtha commentary of Ācārya Yaśomitra. Varanasi: Bauddha Bharati. Speyer, J. S., ed. 1906–1909. Avadānaçataka: A Century of Edifying Tales Belonging to the Hīnayāna. Bibliotheca Buddhica 3. St. Petersburg: l’Académie Impériale des Sciences. Dutt, Nalinaksha, ed. 1978. Bodhisattvabhūmi. Patna: K. P. Jayaswal Research Institute. Waldschmidt, Ernst, ed. 1962. Das Catuṣpariṣatsūtra: Text in Sanskrit und Tibetisch, Verglichen mit dem Pāli nebst einer Übersetzung der Chinesischen Entsprechung im Vinaya der Mūlasarvāstivādins. Teil III: Textbearbeitung: Vorgang 22–28, Abhandlungen der Deutschen Akademie der Wissenschaften zu Berlin. Klasse für Sprachen, Literatur und Kunst, Jahrgang 1960: Nr. 1. Berlin: Akademie-Verlag. Rhys Davids, T. W. and J. Estlin Carpenter, eds. 1890–1911. Dīgha-nikāya. 3 vols. London: Pali Text Society. Rhys Davids, T. W., J. Estlin Carpenter and William Stede, eds. 1968–1971. The Sumaṅgalavilāsinī, Buddhaghosa’s Commentary on the Dīgha-Nikāya. 3 vols. London: Pali Text Society. Smith, H. and H. C. Norman, eds. 1906–1915. Commentary on the Dhammapada. 5 vols. London: Pali Text Society. Cowell, E. B., and Robert A. Neil, eds. 1970. The Divyāvadāna: A Collection of Early Buddhist Legends. Amsterdam: Philo and Oriental Presses. Trenckner, Vilhelm, ed. 1888–1925. The Majjhima Nikāya. 3 vols. London: Pali Text Society. Senart, Émile, ed. 1977. Le Mahâvastu. Tokyo: Meicho-FukyūKai. Taylor, A. C., ed. 1905–1907. Paṭisambhidāmagga. 2 vols. London: Pali Text Society. Oldenberg, Hermann, ed. 1879–1883. Vinayapiṭaka. 5 vols. London: Pali Text Society.

Secondary sources
d’Ailly, Pierre. 1706. De falsis prophetis. In I. L. Ellies Dupin, ed., Joannis Gersonii Opera Omnia. Antwerp: Sumptibus Societatis. Bareau, André. 1963. Recherches sur la biographie du Buddha dans les Sūtrapiṭaka et les Vinayapiṭaka anciens: De la quête de l’éveil a la conversion de Śāriputra et de Maudgalyāyana. Paris: École Française

406

David V. Fiordalis

d’Extrême-Orient. Bartlett, Robert. 2008. The Natural and the Supernatural in the Middle Ages. Cambridge: Cambridge University Press. Bielefeldt, Carl. 2002. “Disarming the Superpowers: The abhijñā in Eisai and Dōgen.” In Daihonzan Eiheiji Daihonki Kyoku, ed., Dōgen zenji kenkyū ronshū. Fukui-ken: Eiheiji, pp. 1018–1046. Brown, Robert L. 1984. “The Śrāvastī Miracles in the Art of India and Dvaravati,” Archives of Asian Art 37, pp. 79–95. Brown, William Norman. 1928. The Indian and Christian Miracles of Walking on the Water. Chicago: Open Court. Burchett, Patton. 2008. “The ‘Magical’ Language of Mantra,” Journal of the American Academy of Religion 74.4, pp. 807–843. Burlingame, Eugene W., trans. 1921. Buddhist Legends: translated from the original Pali text of the Dhammapada commentary. 3 vols. Cambridge: Harvard University Press. Caciola, Nancy. 2006. Discerning Spirits: Divine and Demonic Possession in the Middle Ages. Ithaca: Cornell University Press. Chavannes, Éduoard and Sylvain Lévi. 1916. “Les seize arhat protecteurs de la loi,” Journal Asiatique 8, pp. 5–50, 189–304. Coomaraswamy, Ananda K. 2004. “Samvega: Aesthetic Shock.” In The Essential Ananda K. Coomaraswamy. Bloomington: World Wisdom, pp. 193–199. Fiordalis, David. 2008. Miracles and Superhuman Powers in South Asian Buddhist Literature. Ph.D. Dissertation: University of Michigan. _____. 2012. “The Wondrous Display of Superhuman Power in the Vimalakīrtinirdeśa: Miracle or Marvel?” In Jacobsen, Knut, ed., Yoga Powers: Extraordinary Capacities Attained Through Meditation and Concentration. Leiden: Brill Academic Publishers, pp. 97–125. Flint, Valerie. 1994. The Rise of Magic in Early Medieval Europe. Princeton: Princeton University Press. Foucher, Alfred. 1909. “Le ‘Grand miracle’ du Buddha à Çrâvastî,” Journal Asiatique 13, pp. c. _____. 1917. The Beginnings of Buddhist Art. Paris: Paul Geuthner. Frauwallner, Erich. 1956. The Earliest Vinaya and the Beginnings of Buddhist Literature. Rome: Istituto Italiano per il Medeo ed Estremo Oriente. Gethin, Rupert. 1987. The Path to Awakening. Ph.D. Dissertation: Manchester University. Gómez, Luis O. 1977. “The Bodhisattva as Wonder-worker.” In Lewis Lancaster, ed., Prajñāpāramitā and Related Systems. Berkeley: University of California, pp. 221–262.

Miracles in Indian Buddhist narratives and doctrine

407

_____. 2011. “On Buddhist wonders and wonder-working.” Journal of the International Association of Buddhist Studies 33 (2010) [published 2011]: pp. 513–554. Granoff, Phyllis. 1996. “The Ambiguity of Miracles: Buddhist Understandings of Supernatural Power,” East and West 46, pp. 79–96. Hocart, A. M. 1923. “Flying Through the Air,” Indian Antiquary 70, pp. 80–82. Huntington, John. 1986. “Sowing the Seeds of the Lotus: A Journey to the Great Pilgrimage Sites of Buddhism, Part III,” Orientations 17.3, pp. 32–46. Ingholt, Harald and Islay Lyons. 1957. Gandhāran Art in Pakistan. New York: Pantheon. Kieschnick, John. 1997. The Eminent Monk. Honolulu: University of Hawaii Press. Kurita, Isao. 1988–1990. Gandāra Bijutsu. Tokyo: Nigensha. Lamotte, Étienne. 1944–1980. Le traité de la grande vertu de sagesse. 5 vols. Louvain: Bureaux du Muséon. _____. 1988. History of Indian Buddhism from the Origins to the Śaka Era. Louvain: Institut Orientaliste. MacCulloch, J. A. 1908. “Miracles.” In James Hastings, ed., Encyclopedia of Religion and Ethics. Vol. 8. New York: Charles Scribner’s Sons, pp. 676–690. Mahony, William K. 1987. “Magical Flight.” In Mircea Eliade, ed., The Encyclopedia of Religion. Vol. 5. New York: MacMillan, pp. 349–53. Meisig, Konrad. 1993. “On the Precanonical Shape of the Kevaddha Sutta as Compared with Kien-ku-king.” In Premier Colloque Etienne Lamotte. Louvain: Institut Orientaliste, pp. 63–70. Régamey, Konstanty. 1990. The Bhadramāyākāravyākaraṇa: Introduction, Tibetan Text, Translation and Notes. New Delhi: Motilal Banarsidass. Rhi, Ju-hyung. 1991. Gandharan Images of the “Sravasti Miracle:” An Iconographic Reassessment. Ph.D. Dissertation: University of California, Berkeley. Rhys Davids, T. W., trans. 1899. Dialogues of the Buddha. 3 vols. London: Pali Text Society. Rhys-Davids, T. W. and Hermann Oldenberg, trans. 1885. Vinaya texts. 3 vols. Oxford: Oxford University Press. Rosengren, Karl, Carl Johnson, and Paul Harris, eds. 2000. Imagining the Impossible: Magical, Scientific, and Religious Thinking in Children. Cambridge: Cambridge University Press. Rotman, Andy, trans. 2008. Divine Stories: Part One. Boston: Wisdom.

408

David V. Fiordalis

de Sardan, Jean-Pierre Olivier. 1992. “Occultism and the Ethnographic ‘I:’ The Exoticizing of Magic from Durkheim to ‘Postmodern’ Anthropology,” Critique of Anthropology 12.1, pp. 5–25. Strong, John. 1979. “The Legend of the Lion-Roarer: A Study of the Buddhist Arhat Piṇḍola Bhāradvāja,” Numen 26, pp. 50–88. _____. 1983. “Wenn der magische Flug misslingt,” In H. P. Duerr, ed., Sehnsucht nach dem Ursprung. Frankfurt: Syndikat, pp. 503–18. _____. 2001. The Buddha: A Short Biography. Oxford: Oneworld. _____. 2008. “When Magical Flight Fails: A Study of Some Indian Legends about the Buddha and his Disciples.” Paper read at the University of Michigan on March 28, 2008. Swinburne, Richard. 1970. The concept of miracle. New York: St. Martins Press. Takakusu, J. and K. Watanabe, eds. 1924–1932. Taishō shinshū daizōkyō. Tokyo: Taishō issaikyō kankōkai. Walters, Jonathan, trans. 1995. “Gotamī’s Story.” In Donald Lopez, ed., Buddhism in Practice. Princeton: Princeton University Press, pp. 113– 138.

The higher knowledges in the Pāli Nikāyas and Vinaya
Bradley S. Clough

He who knows his previous lives, Sees heavens and hells, And has attained the exhaustion of birth; A sage who has mastered the higher knowledges, Who knows his mind is purified, freed from all passions; Who has eliminated birth and death; Wholly given up to the pure life; Who has transcended all things; Such a one is called Buddha.1

This passage from the Pāli Nikāyas, which celebrates one who has achieved Buddhism’s highest goal of awakening, is like many other writings in the Buddhist traditions of South Asia and elsewhere, in its inclusion of the abhiññās (Sanskrit: abhijñās) or “higher knowledges” as one of the integral factors in the process of liberation. The overall acceptance of the idea of attainment of certain extraordinary psychic powers by those adepts who have reached advanced stages of meditation is one of the most ancient and consistent features of South Asian Buddhism, from early Pāli texts to treatises of the latter phases of the Mahāyāna in this region. Despite the large amount of significance given to the achievement of these higher knowledges by Buddhist traditions, and the importance assigned to their use in soteriological processes, scholarship has largely steered clear of the study of the abhiññās, perhaps due to a reluctance to

1

MN ii 144. Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 409–433

410

Bradley S. Clough

acknowledge their presence in systems frequently characterized as rationalist and non-mystical. This article is the first in a series of three that will address the role that the abhiññās played in various phases of South Asian Buddhist history. Later pieces will address the place of the abhiññās, first in the Abhidharma literature and meditation manuals of nonMahāyāna schools, and second in the sūtra and śāstra literature of the Mahāyāna. Our focus here will be on the importance of the abhiññās in the Pāli Nikāyas and Vinaya, with an eye towards reasons for the abundant discussion of them in these texts. As a result it is hoped that some further light will be shed on certain pedagogical, epistemological and soteriological concerns of the early centuries of South Asian Buddhism, as preserved by the Theravāda tradition. Granting that the central emphasis in Buddhism has never been on meditation as a method for attaining supernormal powers, and also that the self-serving manifestation of them, as well as their direct pursuit, has usually been condemned, the basic premise here is that the abhiññās nevertheless have long been an important part of Buddhist systems, and have had a greater religious role than previous scholarship has appreciated.

I. The abhiññās as presented in the Sutta- and Vinaya-Piṭakas of the Pāli canon
In the earliest strata of the Pāli canon, one usually finds five abhiññās listed: 1) knowledge of the varieties of supernormal power (iddhi­ vidhāñāṇa). 2) divine ear element (dibbasotadhātu). 3) knowledge comprehending the mind or cognizing others’ thoughts (cetopariyañāṇa/paracittavijānana). 4) knowledge of recollection of past lives (pubbenivāsānussatiñāṇa). 5) divine eye, or knowledge of passing away and rebirth (dibba­ cakkhu/cutu papātañāṇa). ̄

The higher knowledges in the the Pāli Nikāyas and Vinaya

411

From the earliest tradition on, these five higher knowledges have been regarded as mundane (lokiya) achievements, attainable by Buddhist and non-Buddhist adepts alike. When possessed by someone who did not follow a Buddhist path, they were seen as inferior because such a person remained tainted with the āsavas or “influential defilements,”2 which Buddhist paths aimed at removing in the pursuit of liberation. Thus, in the Pāli suttas one frequently finds a sixth abhiññā added, which is “knowledge of the destruction of the influential defilements” (āsavakkhayañāṇa). This higher knowledge was said to be accessible only to Buddhists and was considered supermundane (lokuttara), since it was essentially the equivalent of the paññā (Sanskrit: prajñā) or the liberating insight that realized nibbāna (Sanskrit: nirvāṇa). However, this sixth abhiññā was dropped from lists of the post-canonical Theravāda tradition, due in most part, I believe, to an increasing distinction made between results of vipassanābhāvanā or “insight meditation” as the only effective means of finally eliminating the āsavas, and the fruits of cultivating the jhāna states of meditative absorption (cultivated through another meditative process, samathabhāvanā or “tranquility meditation”), which are said to produce the abhiññās.3 In general, the five abhiññās besides iddhi imply the internal subjective power of intellectual facilities, which are purified and developed by a method of systematic extension from immediate experiences to more distant ones, or from material and gross sensations to ethereal, refined, and even divine ones. In contrast, iddhi signifies the controlling power of both the subjective and objective, manifesting itself in control of both mind and matter. Before turning to this abhiññā and the perceived spiritual benefits of it and the other higher knowledges, we will first take into consideration, in the interests of presenting an accurate and balanced picture of

The Pāli texts typically list three āsavas, namely: 1) kāmāsava – the āsava of (craving) sensory pleasure; 2) bhāvāsava – the āsava of (craving) continued exisitence; and 3) avijjāsava – the āsava of ignorance. Sometimes a fourth āsava is added, namely diṭṭhāsava – the āsava of (wrong) views. 3 To be specific, the texts uniformly agree that it is in the fourth level of jhāna that one can develop the five abhiññās.

2

412

Bradley S. Clough

these powers as presented in early Pāli works, ways in which they were considered to be potentially problematic.

A. The abhiññās: dangers and limitations
Most of the scholarship that has dealt with the abhiññās has tended to focus on Buddha’s apparent ambivalence towards, and denunciation of, such pursuits. Certainly any complete treatment of our topic must consider this aspect, as well as the potential for misuse of the powers and the limitations of their use, as perceived by tradition. Perhaps the most frequently cited episode involving Buddha and the miraculous is the story of an encounter between Buddha and a fellow samaṇa ascetic. Buddha asks the ascetic what he has gained from 25 years of austerities, and the ascetic’s proud response is that he could cross over water, like the river that both were standing by, by walking on it. The bemused Buddha replies that this seems like little gain for so much effort, since for the price of one coin, he could cross the river by boat. But this story may well betray the anxiety many scholars feel about wonder-working in Buddhism, for this much cited story cannot be found in any primary Buddhist source; rather, it is an uncited story taken from Conze 1959: 104. Still, there are also clear signs in primary sources of the Buddha being presented as being wary of dependence upon of excess attention to supernormal powers. In one telling sutta, the dim-witted disciple Sunakkhata, who is constantly awestruck by the miraculous displays of charlatans whom he considers arahants, threatens to leave the san ̇gha unless Buddha agrees to teach him the iddhis. The Buddha scolds, “So, Sunakkhata, whether miracles are performed or not, the purpose of my teaching the dhamma is to lead whoever practices it to total cessation of suffering. Then what purpose would the performance of miracles serve? Consider, you fool, how great this fault of yours is.”4 There was also danger of mistaking the attainment of iddhi for high states of spiritual accomplishment. To regard the iddhis
4

DN iii 4.

The higher knowledges in the the Pāli Nikāyas and Vinaya

413

as signs of incipient success had other potential dangers, in that they were available to non-Buddhist yogins as well. Even the great villain of the Pāli Nikāyas, Devadatta, possessed these abilities. Worse yet would be to mistake acquisition of the iddhis for the final abode of nibbāna. It was probably with these dangers in mind that tradition has Buddha making the distinction between “noble” (ariya) and “ignoble” (anariya) iddhis. The supernormal powers available to Buddhist adepts and non-Buddhist yogins alike were labled “ignoble,” since they were seen as being concomitant with the āsavas and worldly attachment. The only truly “noble” iddhi is the “accomplishment” (one meaning of iddhi) of abiding mindfully in equanimity (upekkhā), indifferent to both that which is unpleasant and that which is pleasant.5 Instead of furthering one’s involvement in saṃsāra as the ignoble iddhis can do, the noble iddhi of equanimity overcomes the polarities which disturb one’s life in the world. Thus it is said to be without impurities and attachment. Along very similar lines, Buddha is presented as putting the wondrous in proper perspective in his formula of the three “miracles” (pāṭihāriya). The three are the miracles of supernormal powers (iddhis), mind-reading (ādesanā), and instruction (anusāsani). In the Kevaḍḍha Sutta, a householder from Nālandā encourages Buddha to perform miracles in order to increase people’s faith. Buddha replies that he does not teach dhamma by telling monks to perform miracles, and proceeds to outline the three pāṭihāriyas. As for iddhis, he says that skeptics will dismiss them as a magical “Gandharan charm,” and concludes, “That is why, O Kevaḍḍha, seeing the danger of [such supernormal powers], I dislike, reject, and despise [them].”6 As for the power of mind-reading, it is dismissed on similar grounds. Teaching is the best miracle, Buddha says, and he then proceeds to discuss the dhamma in terms of the path of training culminating in arahanthood, a path that still, it is worth noting, includes cultivation of the jhānas, which therefore entails realizing the supernormal powers and mind-reading, and the rest of the abhiññās! Still, the main message here seems to be
5 6

DN iii 112. DN i 213.

414

Bradley S. Clough

that it is only instruction on what to do and what not to do on the path that is truly valuable to the faithful.7 Likewise, Buddha was also portrayed as being wary of performing miracles that might confuse the uninitiated about what was truly important in his teaching. In the Cullavagga section of the Vinaya, Buddha scolds the disciple Piṇḍola Bhāradvāja, who has been encouraged by the irrepressibly wonder-working Moggallāna to engage in a contest of iddhi powers. Clearly opposed to any purposeless exhibition of iddhi for its own sake, Buddha calls Bhāradvāja’s actions unfitting for a samaṇa and likened them to performing a strip-tease for money. It was this very event that led Buddha to declare the precept that forbids monks from displaying supernormal powers.8 The only other Vinaya rule involving the abhiññās concerns falsely claiming possession of supernormal powers. This is a pārājika (one who has committed a grave transgression against the monastic rules) offense, warranting expulsion from the san ̇gha.9 Finally, it must be pointed out that a thorough reading of the Pāli Nikāyas reveals that the abhiññās were not always deemed a necessary part of the path to nibbāna. The sources indicate that actually a relatively small percentage of disciples were advanced enough in cultivation of the jhānas to have attained them. Most arahants seem to have been liberated by insight meditation alone. In one passage, Buddha says that out of 500 monks, 60 have what are called the three vijjās, which are synonymous with the final three abhiññās; 60 have followed the path that has resulted in them becoming what is called “liberated in both ways” (ubhatobhāgavimutta);10 and the
For a full treatment of the pāṭihāriyas (Sanskrit: prātihāryas), see David Fiordalis’ article in this volume. 8 Vin ii 112. 9 Vin iii 91. 10 Both in the Pāli Nikāyas and among modern academics, there is no universal agreement on what it means to be “liberated in both ways.” Scholars have tended to assert that it means a person described in many Nikāyan passages as “liberated by insight” and “liberated in mind,” but the evidence does not bear this out. A definition found only once in the Nikāyas says that when a monk is able to attain the “eight liberations” (meaning the jhānas of the form and formless realms [Paṭis ii 38–40]) in forward and reverse order, he
7

The higher knowledges in the the Pāli Nikāyas and Vinaya

415

rest are said to have attained what is called “liberation by insight” (paññāvimutti) alone, which is the result of a path that does not involve cultivation of the higher jhānas and the abhiññās.11

B. Iddhividhāñāṇa
How light is my body, Touched by abundant joy and bliss! Just as a tuft of cotton floats on the breeze, In the same way my body seems to float.12

The word iddhi is derived from a verb root meaning “to prosper or succeed” (Pāli: iṣ; Sanskrit: ardh/ṛdh). So while the “common” meaning of iddhi is “success”/“achievement”/“prosperity,” in the context of South Asian meditative traditions, it usually denotes supernormal or wonder-working power, implying both psychic and physical powers which lead to accomplishment or success. When discussing these so-called “supernormal powers,” it is important to bear in mind that these abilities should not be considered miraculous in the sense of being contrary to the Buddhist understanding of mental processes. Iddhi powers are not the results of some suspension of the processes of mind or nature, but are rather the by-products of the mastery of these processes. In this sense iddhi could be translated simply as “success,” meaning that attainment of it is indicative of progress along the path, particularly meditative attainment. These supernormal powers are therefore more important for what they signify, which is the attainment of a certain important level of soteriological success in practice (namely realization of the fourth jhāna of meditative absorption, the state which enables one to cultivate these powers), than they are as powers in and of themselves. That the iddhi powers are considered non-miis considered “liberated in both ways” (DN ii 70–71). But the most frequently occurring and only other definition says that such a person is one who has attained all nine meditative absorptions of the form and formless realms, and then goes on to see their impermanence by means of insight (AN iv 453; MN i 477–481; et freq.). 11 SN i 191. 12 Th i 104.

416

Bradley S. Clough

raculous by-products indicating progress in meditation is seen in the gloss of the great Theravāda commentator Buddhaghosa, which renders iddhi as “effectiveness of means” and “success in the sense of attainment.”13 The Pāli term iddhividhā literally means “kinds of success.” To see what kinds of success were held to be available to the meditation adepts of early Buddhism, attention must be given to the locus classicus for the presentation of the abhiññās, the Sāmaññaphala Sutta, where the iddhis are presented as both the immediate results of systematic meditation, and as the hidden powers of the human mind. According to this sutta, after the five hindrances (nīvaraṇas),14 – which are considered conducive to unreflective involvement in the realm of desire – have been eliminated, absorption (appanā) in the four jhāna meditative states is possible. It is upon attaining the fourth jhāna that one is able to cultivate the supernormal powers and the other abhiññās. What follows is the oft-repeated stock description of how the abhiññās are acquired, as it is given in this sutta:
When the mind is thus concentrated, pure, cleansed, free from stains, free from corruptions, supple, pliant, steady, and unperturbed, one directs and thoroughly turns the mind to knowledge and vision.15

It is noteworthy that the abhiññās occur in the fourth jhāna level of the form realm, where not only has the normal perception (saññā) of the desire realm been left behind, but also where the four mental factors of the previous jhānas, which are initial thought (vit­ akka), sustained thought (vicāra), mental joy (pīti) and bodily bliss (sukha) have also been eliminated. Only in the fourth jhāna, characterized by equanimity (upekkhā) and purity of mindfulness (sati­ parisuddhi), is the mind concentrated, purified, and flexible enough to possess a higher and more direct knowledge of the true nature of things. Indeed, what is implied here will be explicitly stated by
Vism 12.21. They are: 1) sensual desire (kāmacchanda); 2) ill-will (byāpāda/ vyāpāda); 3) sloth and torpor (thīnamiddha); 4) mental agitation and worry (uddhaccakukkucca); and 5) doubt (vicikicchā). 15 DN i 76.
14 13

The higher knowledges in the the Pāli Nikāyas and Vinaya

417

later tradition: It is these very eight factors of mind just mentioned in the passage given above – factors such as concentration, purity, pliancy, etc. – that are the proximate cause for realization of the abhiññās.16 Next, after having observed that one’s mind is dependent on one’s impermanent, disintegrating body, one directs one’s mind to the creation of what is called a “mind-made body” (manomayakāya). One creates from this mind-made body “another body” (aññakāya), which “has form and is complete in all its limbs and faculties.”17 This process is likened to a person pulling a reed from a sheath, a sword from its scabbard, or a snake from its skin.18 It is this “other body,” created by the mind in the fourth jhāna, as opposed to the body subject to its usual frailties, that one employs to acquire iddhi and the other abhiññās. A major issue that has been raised here is the scholarly argument that says that since the abhiññās involve this mind-made body and not one’s “normal” body, they can be interpreted as mere ideations or mental projections of sorts. In this attempt to explain away this abhiññā, we see here another example of scholars’ anxiety, in this case about the supernormal powers as truly realizable. While this argument may be valid to a degree, it is important to remember that according to Buddhism, the mind is the sixth sense, and as such it is not different from the other senses. Therefore these “ideations” are no less “real” than other types of sensory experience.19 Having created the body made of mind, one then directs attention to iddhividhā, the various kinds of supernormal power. The version given in the Sāmaññaphala Sutta is stereotyped, found without variation in eighteen other places in the canon, as well as many times in the later commentaries and meditation manuals of the Theravāda tradition, and also in Mahāyāna literature. The passage goes as follows:

16 17 18 19

Vism 12.9. Paṭis ii 210–211. Paṭis ii 211; Vism 12.139. Johansson 1969: 48.

418

Bradley S. Clough

Being one he becomes many; being many he becomes one again. He appears and vanishes. He goes unhindered through walls, enclosures, and mountains, as if through air. He dives in and out of the earth, as though in water. He walks on water as though on broken ground. Seated cross-legged, he flies in the sky like a winged bird. With his hand he touches and strokes the moon and sun, so mighty and powerful. He travels as far as the Brahma world.20

Unfortunately, the Pāli canon does not provide much more information regarding the techniques by which one would obtain each of these eight extraordinary abilities. However, later meditation manuals, particularly the Visuddhimagga, do.

C. Dibbasotadhātu
As is presented in the Sāmaññaphala Sutta, the meditator, in the same concentrated and purely mindful state of the fourth jhāna as was described above, directs his attention to the cultivation of the “divine ear element.” The rather brief stereotyped passage, which is also found frequently in the Pāli canon, provided the model upon which later texts expanded, is as follows:
…he directs and turns the mind to the divine ear element. With the divine ear element, which is purified and surpasses the human, he hears both kinds of sounds, the divine and the human, those that are far as well as near.21

This power is likened to the ability to distinguish between the various kinds of drums heard at the same time. It is a tremendous expansion of auditory perception in both extent and depth, without the medium of the actual sense organ. The formula implies not only the ability to perceive sounds from realms extremely far away, but also the ability to understand non-human beings in other realms. Unfortunately, the suttas and Vinaya provide very little information about the potential uses of this power or the techniques applied for attaining it. Divine ear appears to be the least important of the abhiññās. Unlike the other abhiññās, its role and usage in Buddhist
20 21

DN i 78. DN i 79.

The higher knowledges in the the Pāli Nikāyas and Vinaya

419

epistemological and soteriological schemes is unclear. Except from the point that the ability to hear sounds from other realms might confirm the Buddhist cosmological teaching of the existence of beings in these realms (the Buddhist worldview of the five or six destinies), it does not really serve to verify key Buddhist teachings, the way the other abhiññās do, as we will see. Most importantly, it seems to lack the revelatory nature of the other abhiññās, in that it adds little additional information to one’s understanding of the way things truly are. Furthermore, there is not much pedagogical use for the divine ear. This too separates it from the other remaining higher knowledges, which can be used to help others enter into or progress along the path. However, it is certainly conceivable that the power could be used to impress others, and thus be a possible skillful use of means (upāya) that might convert others. However, here we need to bear in mind the Buddha’s usual discouragement of using supernormal powers to convert, as displays of them could mislead others about what is truly important in the dhamma.

D. Cetopariyañāṇa/paracittavijānana
As presented in the Sāmaññaphala Sutta, the meditator, still equipoised in the fourth jhāna, is next able to read others’ minds in the following manner:
He knows and distinguishes with this mind the minds of other beings or other persons. He knows the mind of passion to be with passion; he knows the mind without passion to be without passion. He knows the mind of hate to be with hate; he knows the mind without hate to be without hate. He knows the deluded mind to be deluded; he knows the undeluded mind to be undeluded. He knows the distracted mind to be distracted; he knows the attentive mind to be attentive. He knows the unexpanded mind to be unexpanded; he knows the expanded mind to be expanded. He knows the surpassed mind to be surpassed; he knows the unsurpassed mind to be unsurpassed. He knows the concentrated mind to be concentrated: he knows the unconcentrated mind to be unconcentrated. He knows the liberated mind to be liberated; he knows the unliberated mind to be unliberated.22
22

DN i 79–80.

420

Bradley S. Clough

The text goes on to say that one who possesses this abhiññā knows the workings of another’s mind with the same familiarity as one who can observe one’s face in a mirror or pan of water and notice whether there is a mote on it or not.23 The Pāli canon speaks of several different degrees of telepathic mind-reading. The early stereotyped formula given above seems to imply that an adept could come to know only the general character of another’s mind. But in the same stratum of thought, a much more detailed knowledge is suggested: “One can know the mind (citta), the states of mind (cetasika), the initial thoughts (vitakka), and the trains of thoughts (vicārita) of other beings and persons.”24 In another passage, the Buddha claims to know by this abhiññā the “specific thoughts” (parivitakka) in the mind of a brāhmaṇa student. In the Aṅguttara-Nikāya, there are said to be four ways of knowing another’s mind. They are: 1) observing external signs; 2) obtaining information from others; 3) listening to the vibrations of thought of another as he thinks; and 4) comprehending the mind of another and observing how the mental dispositions are ordered so as to be able to predict which subsequent thoughts might arise.25 The great attention given to the different types of knowledge of others’ minds in the suttas suggests that this abhiññā was considered extremely important by early tradition. Indeed, one can see its usefulness in a variety of applications. As the references to knowing “trains of thoughts” and the order of thoughts indicates, there seems to be a conviction that through this abhiññā one can gain a more certain knowledge of mental phenomena. The AṅguttaraNikāya passage just quoted even says that Buddha and his disciples were able to discern how certain thoughts followed other thoughts. This clearly suggests that one could gain a vipassanā-like awareness of the mental continuity and the ever-changing nature of thought processes. This also could be seen as a precursor to Abhidhamma and its explantions of how the mind works, of how certain thoughts follow other specific thoughts. This insight into the transient nature
23 24 25

DN i 80. DN i 213. AN v 170–171.

The higher knowledges in the the Pāli Nikāyas and Vinaya

421

of thoughts that make up the mind is a key, liberating realization along the Buddhist path, and it is intriguing that this type of awareness of momentariness or impermance, phenomenologically quite similar to the third foundation of mindfulness, could be gained while meditating in the fourth jhāna. This affirms the description of the fourth jhāna as being a state of mindfulness, and at the same time seems to counter the not infrequent Buddhist claim that such insight only comes from vipassanā practice. Not only did this abhiññā provide valuable insight into the workings of the mind – and thus quite likely contributed to Buddhist theories of mental connections eventually articulated in full in the Abhidhamma, as well as verify the Buddhist understanding of mind as a continuity or succession (santati/santāna) of momentary thoughts – but it also could serve as a verificatory tool for affirming attainment of certain states of spiritual advancement. When a disciple attained arahanthood (arahattā), it was common for him or her to “declare gnosis” (aññām vyākaroti) by announcing, “Destroyed is rebirth, lived is the holy life, done what is to be done. I am assured that there is no more life in these conditions.”26 The Buddha set forth six ways to scrutinize this claim, which involved such methods as simply asking the person what he or she had realized. But because a person could conceivably lie or be mistaken in response to such an inquiry, it was said that the only certain means for determining another’s attainment was for a Buddha or disciple to read the other’s mind with his or her own mind (cetasā ceto paricca parivittakam aññāya).27 Knowledge of others’ minds could also be combined with divine eye to gain insight into the workings of saṃsāra. One first applied divine eye in order to see other beings in other destinies faring according to their kamma, and then would employ knowledge of their thoughts. In this way one could come to know what beings in hell, animal, and hungry ghost realms suffer. This could provide powerful incentive to take up or practice harder the path to liberation, just as could the knowledge of the minds of those abiding in human and
26 27

DN i 84; MN i 184, 255–256, 348, 496, 522, ii 239, iii 36, 287; et freq. AN v 161–164.

422

Bradley S. Clough

divine realms who had attained an almost nirvanic state of calm, equanimity, joy, and insight. Knowledge that those whose kammas have given them less advanced but still very good human or even divine destinies could likewise provide powerful incentive to live a morally good life. A section from the Nikāyas on mind-reading that seems to have been entirely overlooked by scholars discussing the abhiññā of mind-reading is from the Mahāsīhanāda Sutta of the MN. It is remarkable in that like what has been discussed in the last two paragraphs, the Buddha speaks of using this abhiññā both in a way that could confirm that nibbāna is a truly existent reality and provide insight into the workings of kamma and the different kinds of actions that lead to rebirth in either good or bad destinies. But it addition to this, the Buddha is said to have gained the ability to become aware of the minds of even those who had attained nibbāna! In this discourse, the Buddha describes in some length how he, by “understanding” (or, perhaps better, “encompassing”) mind with mind (cetasā ceto paricca), can see beings passing into the five destinies28 and the kinds of action that lead to rebirth in each. Here again a way into seeing not only that the law of kamma is actually in effect, but also a way of coming to know in a very practical way what kinds of behaviors typically lead to either bad or good destinies, is provided by this soteriologically important abhiññā. But of at least equal importance in this section on the results of “encompassing mind with mind” is the Buddha asserting that this supernormal ability can affirm a person’s attainment of nibbāna (to be precise, the Buddha sees that any being who “here and now enters upon and abides in the liberation of mind and liberation by wisdom that are without the influential impurities because of the destruction of influential impurities”).29 So, once again we find this abhiññā leading to key realizations that certain very central Buddhist beliefs – in this case including the conviction that nibbbāna truly exIt seems that in the earlier strata of discourses in the Nikāyas, the destiny of life as a demon (asura) is not yet included as one of the possible destinies of living beings. Later texts add the asura destiny to make up the better known list of six destinies. 29 MN i 76. The whole section just discussed is MN i 74–76.
28

The higher knowledges in the the Pāli Nikāyas and Vinaya

423

ists (making this abhiññā a direct, experiential confrimation of the third noble truth) – are actual, true realities. Receiving affirming knowledge that nibbāna is indeed attainable also could function very well to motivate one to greater effort in practice, knowing that such a state certainly exists. Finally, this abhiññā could also be used as a teaching tool to benefit others. The tradition holds that Buddha and his chief disciples used it to preach the dhamma with the most success, because they knew the mental states of the individuals they were instructing. So this abhiññā could be a powerful form of skillful means (upāya). And, as any student of Buddhism knows, tradition attributes to Buddha the constant concern with students’ mental disposition, so that he could respond in the most soteriologically useful way.30 In this respect, this abhiññā is quite likely to have been a highly valued skill for teachers of Buddhist traditions.

E. The three vijjās
The next three abhiññās – divine eye, recollection of past lives, and knowledge of the exhaustion of the impure influences – are collectively known as the “three knowledges.” Before I treat each individually, I will discuss them as a group. These three abhiññās are most important for the tradition, as they stand at the very core of descriptions of Buddha’s experience of liberating awakening. The three vijjās can be distinguished from the other abhiññās in that they are considered “clear visions” (parisuddhañāṇadassana) experienced by Buddha under the bodhi tree. As represented in the Verañja Sutta of the Aṅguttara-Nikāya and the Bhayabherava and Mahāsaccaka Suttas of the Majjhima-Nikāya, Buddha received these clear visions on his night of awakening by attaining the four jhānas, but not proceeding to the meditations of the “formless realm.” In the first watch of the night, he directed his mind to the knowledge of many thousands of his own previous births, in each case recalling his name, class, food ingested, experiences of pleasure and suffering, and death. By this knowledge it is said that
30

See, for example, Dhp-a iii 215.

424

Bradley S. Clough

“ignorance was overcome, clear vision arose, darkness was overcome, light arose,”31 and that he became “diligent, ardent, and selfcontrolled.”32 In the second watch, by means of the pure divine eye, Buddha saw beings reaching destinies according to their deeds. Again, “ignorance was overcome, clear vision arose,” etc. In the third watch, he directed his mind to the exhaustion of influential defilements (āsavas) and to the realization of the four noble truths. The recognition that these higher knowledges were factors that essentially and directly led to Buddha’s liberating awakening is crucial to an appreciation of the role of the abhiññās in Buddhist tradition. It is quite noteworthy that, as set forth in these particular traditional accounts, it is these three knowledges in particular, not insight into selflessness, dependent origination, the four noble truths33 or any of the other insights considered central to Buddhism, which function to eliminate ignorance and consequently liberate Buddha. That these three knowledges have been seen as absolutely essential to the path to liberation is brought out in several passages found elsewhere in the the Pāli Nikāyas. Perhaps the most telling example of the value attributed to these knowledges is the TevijjaVacchagotta Sutta, in which Buddha is asked if it is correct to say that he is omniscient (sabbaññā). Buddha replies that it is not correct, but that it would be correct to describe him as “one who possesses the three knowledges.”34 He goes on to say that he recollects former lives in all modes and details, comprehends that beings are inferior, superior, beautiful, ugly, well-faring and ill-faring accordMN i 22; 248–249. MN i 22; 248. 33 The versions of the enlightenment found at SN 22:26 and 12:65 focus respectively on insight into selflessness and dependent origination as the central realizations of the Buddha’s experience of awakening. DN ii 30–35 presents the former Buddha Vipassī as attaining enlightenment by virtue of contemplating both dependent origination and the selfless nature of the aggregates.The famous first sermon of the Buddha, on “Setting in motion the wheel of the dhamma” (Dhammacakkappavattanasutta) presents Buddha basing his enlightenment on insight into the four noble truths. 34 MN i 482.
32 31

The higher knowledges in the the Pāli Nikāyas and Vinaya

425

ing to their actions, and abides in the liberation of mind and the liberation by insight realized by the higher knowledge of the destruction of the āsavas.35 Thus, the significance given to the three knowledges (at DN ii 275 they are said to be the “things to be realized,” and at AN ii 65 they are said to be the characteristics of a noble monk [ariyabhikkhu]) expands the role and importance of the abhiññās even further. Not only are they of illuminating, verificatory (confirming as they do the Buddhist understanding of the workings of kamma and saṃsāra), and pedagogical value, but they also describe the types and limits of knowledge which Buddha and the arahants possessed after liberation. At other points in the canon, Buddha’s scope of knowledge appears to be far broader than these abhiññās, but this threefold knowledge formula is clearly represented as being of more soteriological significance to him than omniscience. In the suttas, Buddha often directs his discourse to the three vijjās as a topic of great spiritual value. In fact, many passages equate the attainment of arahanthood itself with the realization of the three vijjās. For example, the Mahāssapuru Sutta says that a bhikkhu who acquires the three vijjās in the fourth jhāna is called “a samaṇa, a brāhmaṇa, one who is cleansed, an attainer of knowledge, wellversed in sacred learning, a noble person, and an arahant.”36

F. Pubbenivāsānussatiñāṇa
The stereotyped formula, found in the Sāmaññaphala Sutta and elsewhere,37 is as follows:
I remembered many previous existences: one, two, three, four, five births, ten, twenty, thirty, forty, fifty births, a hundred, a thousand, a hundred-thousand births, many eons of destruction, many eons of creation, and many eons of world destruction and creation. In such a place such was my name, such my clan, such my appearance (or “class:” vaṇṇa), such my food, such experiences of pleasure and pain,

35 36 37

ibid. MN i 278. Such as MN i 22 and i 248.

426

Bradley S. Clough

and such limits of my life. When I passed away from there, I arose elsewhere, and there such was my name, such was my clan…. and when I passed away from there, I arose elsewhere. Thus I remembered many previous existences, with their conditions and particulars.

There are some curiosities and contradictions that occasionally arise with respect to this abhiññā. Some passages even speak of the ability to “remember” (anussarati) future lives!38 And there was a minor controversy – that would reach near inconceivable proportions in Abhidharma discussions – over the length of different beings’ memories, that had its seeds in the earlier suttas. At one point, Buddha says that he can recall only as far back as 91 kalpas.39 At other points, he claims to remember an “immeasurable” amount of kalpas.40 Besides the tales in the Jātaka, which always end with the Buddha exercising this abhiññā to recall that he was that very living being discussed in the story, there are numerous instances in the suttas where the Buddha recalls his former lives, although unlike the distinctive pattern found in the Jātaka, they do not always carry the emphasis on improved spiritual development over many lifetimes. From a Buddhist perspective, clearly the most useful knowledge to be gained from memory of past lives is confirmation of the teachings of cyclic existence and rebirth. As already stated, such memories could provide powerful incentive to be free from saṃsāra. There is, however, an interesting sutta which shows that a Buddhist interpretation was not the only possible conclusion taken from this abhiññā. In the Brahmajāla Sutta, Buddha admits that certain other samaṇas and brāhmaṇas, by means of their attainment of certain states of samādhi, are able to call to mind several hundred-thousand past lives. This experience, however, only
Demiéville 1927: 291. Bodhi (2000) says that the Sāratthappakāsinī (SN-a), in commenting on SN v 176, speaks of the elder who remembers both a thousand eons of past live and a thousand eons of future lives. 39 MN i 483. 40 Jā i 25, 83; Pv iv 17; Dhp-a i 88.
38

The higher knowledges in the the Pāli Nikāyas and Vinaya

427

serves to confirm their reifying, eternalist views. This memory of theirs confirms their beliefs that the self and the world are eternal, like a mountain peak which gives birth to nothing new, set firmly like a post. These beings rush around, circulate, pass away, and are born, but their self remains eternally.41 Thus it appears that even one of the three vijjās can possibly further misunderstanding from a Buddhist point of view.

G. Dibbacakkhu/cutūpapātañāṇa
The following stereotyped formula as it appears in the Sāmaññaphala Sutta is as follows:
He directs and turns his mind to the knowledge of the passing away and rebirth of beings. With the divine eye, which is purified and surpasses the human, he sees beings passing away and being reborn: inferior and superior, fair and ugly, happy and unhappy in their destinies, as kamma directs them. He understands, ‘these beings, on account of of bad conduct in body, speech, and mind, reviling noble ones, have wrong view, and will suffer the fruits of kamma due to their wrong view. Upon the break-up of their bodies after death, they are reborn in a lower world, in a bad destiny, a state of suffering, hell. But those beings, on account of good conduct of body, speech and mind and not reviling noble ones, have right view, and will receive the fruits of kamma due to their right view. Upon the break-up of their bodies after death, they are reborn in a good destiny, a divine world.’42

This knowledge is likened to a person standing on a terrace looking over a crossroads, viewing others as they pass in and out of doorways.43 Divine eye, as presented in the suttas, seems to have two distinct aspects to it. The first is the ability to see contemporaneous events beyond the range of normal vision. For example, Anuruddha, considered foremost among Buddha’s disciples in mastery of divine eye, is said to have the power “to see a thousand worlds.”44
41 42 43 44

DN i 13. DN i 82. DN i 83. MN i 213 says “he sees a thousand worlds” (sahassaṃ lokānam vo­

428

Bradley S. Clough

Second, divine eye is also directed towards gaining a knowledge of the passing away and rebirth of beings, which would give one not only an understanding of the workings of kamma, but also the liberating insight into the fluctuating nature of things, as one witnesses the ever-changing conditions of the lives of those faring according to their kammas. So, once again, the tremendous soteriological import of an abhiññā can be seen. In the Mahāsīhānāda Sutta, Buddha declares that one with divine eye comprehends each of the five destinies,45 the course of action leading to each destiny, and how that being fares according to kamma, so that upon death the next fitting destiny is reached.46 This obviously also has ramifications for Buddhist ethics, as again such knowledge would provide one with an increased understanding of what kinds of deeds lead to better rebirths and what kinds of deeds lead to worse ones. In a later Pāli text, the chief disciple Moggallāna wants to use this abhiññā to “lift the veil” of an unknown deva-world, so that he might benefit and encourage beings by reporting back that beings are faring very well there, due to their previous good deeds on earth.47 One could well see this abhiññā, with the insights it brings into kamma and that which leads to rebirth in higher and lower destinies, as well as into the ephemeral nature of samsaric conditions, as a central basis from which Buddha designed his course of practice for the spiritual advancement of sentient beings.

H. Āsavakkhayañāṇa
Finally, at the end of the Sāmaññaphala Sutta, the meditator, still concentrated in the fourth jhāna, directs and turns his mind to the destruction of the āsavas or influential impurities. The passage,
loketi), and SN v 176 and v 299 say “I directly know a thousand worlds” (sahassaṁ lokam abhijānāmi). 45 This discourse is interesting, in that it first describes how divine eye leads to insight in the five destinies, but also, as we saw above, describes how mind-reading leads to the same insight. 46 MN i 73. 47 Vimalatthavilāsinī (commentary on the Vimānavatthu of the KhuddakaNikāya), p. 4.

The higher knowledges in the the Pāli Nikāyas and Vinaya

429

which is identical to certain descriptions of Buddha’s awakening in the third watch of the night, is as follows:
He knows as it really is, ‘this is suffering.’ He knows as it really is, ‘this is the cause of suffering.’ He knows as it really is, ‘this is the cessation of suffering.’ He knows as it really is, ‘this is the path leading to the cessation of suffering.’ He knows as it really is, ‘these are the āsavas.’ He knows as it really is, ‘this is the cause of the āsavas.’ He knows as it really is, ‘this is the destruction of the āsavas.’ He knows as it really is, ‘this is the path leading to the destruction of the āsavas.’ And through this knowledge and vision, his mind is liberated from the āsava of sensual desire, his mind is liberated from the āsava of desire for continued existence, his mind is liberated from the āsava of ignorance. The knowledge arises in him, ‘this is liberation,’ and he knows: ‘Birth is ended, the pure life has been led, done is what is to be done, there is no further life in these conditions.’48

This is likened to a person with good eyesight seeing the oysters, gravel, and fish in a clear pool.49 This abhiññā is said to be the final fruit of the samaṇa’s life, more excellent and perfect than the previous fruits, which include the five other abhiññās. The āsavas or influential impurities of sensual desire (kāma), desire for continued existence (bhava), and ignorance (avijjā), which, along with hatred (dosa), are the main causes of attachment to saṃsāra, and with their elimination comes nibbāna. Ancient South Asian Buddhists and modern Western scholars alike distinguish this as the distinctively Buddhist abhiññā, knowledge of which prevents misuse or misinterpretation of the other five higher knowledges. Not surprisingly, it is called the supermundane (lokuttara) abhiññā, since it is equivalent to the attainment of nibbāna. The sixth abhiññā is the criterion of arahanthood; the arahant had to know for himself/ herself that his/her āsavas, his/her obsessions, were destroyed. The Pāli commentaries call it arahattasādhana, the “proving or producing of arahanthood.” So, essentially, liberating awakening is finally gained through the sixth abhiññā, since the final requirement is that one verify it

48 49

DN i 84. ibid.

430

Bradley S. Clough

by knowledge of the destruction of the āsavas. But what is most intriguing about this part of the Sāmaññaphala Sutta and identical sections of some other suttas – something quite unusual which scholars apparently have failed to see as noteworthy – is that these texts describe that by directing one’s mind to the destruction of the āsavas while abiding in the fourth jhāna, liberation can be attained. This would seem to directly contradict the paths of those who are said to be “liberated by mind” (cetovimutta) and “liberated in both ways” (ubhatobhāgavimutta), since these forms of liberation involve attainment of even higher states of meditative absorption – the aru pa samāpattis or “formless states of meditative attainment” ̄ – as a means to liberation, as well as the path of those adepts said to be “liberated by insight” (paññāvimutta), since they do not even attain the lower jhānas as part of the liberating discipline. So, what we seemingly have here is a description of another way to nibbāna. Furthermore, it also should be noted that the descriptions in these texts do mirror certain descriptions of both Buddha’s awakening experience and his parinibbāna, which are achieved via the fourth jhāna.50 It must also be noted that it is only in the context of a few texts that liberation is presented as being attained solely through realization of the abhiññās in the fourth jhāna. Much more frequently found are passages which acknowledge that it is necessary to cultivate insight (vipassanā), in addition to ethics (sīla) and the high states of samādhi or concentration that are the hallmark of samathabhāvanā or “mental cultivation of tranquility.” For example, we have the stock phrase, “he should fulfill the moral precepts, be intent upon internal serenity of mind, be posssesed of insight… ”51

The description of the Buddha’s awakening taking place while absorbed in the fourth jhāna is found at MN i 247–249, and the description of the Buddha entering into parinibbāna via the fourth jhāna is found at DN ii 156 and SN i 157–158. 51 MN i 33–36; et freq.

50

The higher knowledges in the the Pāli Nikāyas and Vinaya

431

II. Conclusion
While we have seen at the outset of this paper that the early Pāli materials display an ambivalent and even critical attitude towards the use of supernormal powers, it is clear that no picture of Buddhism’s ideal figures (not only does the chief disciple Moggallāna come to mind here, but Buddha himself is associated with many miraculous feats which space has not permitted me to look at here), nor any picture of its pedagogical, epistemological, and soteriological concerns, would be complete without consideration of the abhiññās. While we have seen that there were path alternatives that apparently did not involve cultivation of the abhiññās, such is the diversity of the early Pāli discourses that there are very important discussions, such as those found in the significant Sāmaññāphala Sutta (certainly a major statement on early Indian Buddhist and Theravādin soteriology) that place the abhiññās at the very heart of the Buddhist endeavor. The description of no less crucial an event than Buddha’s own awakening experience also makes at least some particular higher knowledges central realizations. Essentially, the abhiññās confirm an underappreciated conviction of much of these texts, which is that certain crucial Buddhist forms of knowledge and insight can be developed not just through vipassanābhāvanā but through samathabhāvanā and its jhāna states of deep concentration and mindfulness as well. As is said on several occasions in the suttas, “It is natural that one in a state of concentration knows and sees things as they really are.”52 The abhiññās were reckoned by compilers of the suttas as both the corollary and sometimes crowning features of arahanthood, and in many stock phrases and formulas, abhiññā is listed among the chief values of Buddhism. For example, they are virtually equated with the religion’s final goal in the stock phrase, “this pure life leads to complete detachment, to freedom from desire, to cessation, to peace, to higher knowledge, to complete awakening, to nibbāna.”53

52 53

AN v 3, v 313; et freq. DN ii 251; et freq.

432

Bradley S. Clough

Although some of the abhiññās, particularly iddhi and divine ear, could not always be seen as highly valuable in and of themselves, they certainly could function at least as reliable signs of progress along the path of meditation practice, as indications that that one is transcending the normal limits of the phenomenal world to which one is bound. As for the power of mind-reading, it could be a beneficial teaching device and was clearly distinguished as the only foolproof way of assessing others’ levels of spiritual attainment. We have also seen that certain abhiññās, especially knowledge of past lives and divine eye, were deemed highly useful by the tradition for the purpose of experiential illumination and confirmation of certain Buddhist truths, such as saṃsāra, kamma, dukkha, anicca, and nibbāna. And, as is often repeated in the suttas, the sixth abhiññā affirms the Four Noble Truths. The Buddha is famous for saying that his teaching is verifiable by experience (ehipassika).54 Through meditation, the adept is deemed able to extend the normal limits of human experience and knowledge and, by attaining the abhiññās, one is enabled to experientially substantiate key Buddhist teachings. Thus on many levels, some significant early Pāli material appears to insist that the abhiññās provide insights into many of the most significant truth claims of Buddhism.

References Primary sources
References to Pāli texts are to the roman-script editions of the Pali Text Society, England, by volume number and page.
AN Dhp-a DN Aṅguttara-Nikāya, ed. R. Morris, E. Hardy. 5 vols. London 1885–1900. Dhammapada-aṭṭhakathā, ed. H. C. Norman. 5 vols. London 1906–1914. Dīgha-Nikāya, ed. T. W. Rhys Davids, J. E. Carpenter. 3 vols. 1890–1911.

54

DN ii 217, iii 5, 227; SN i 9, iv 41, 272; v 343; AN i 158, ii 198.

The higher knowledges in the the Pāli Nikāyas and Vinaya Jā MN Paṭis Pv SN Th Vin Vism

433

Jātaka, together with its commentary, ed. V. Fausbøll. 6 vols., London 1877-1896. Majjhima-Nikāya, ed. V. Trenckner, R. Chalmers. 3 vols. London 1888–1899. Paṭisaṁbhidāmagga, ed. A. C. Taylor. 2 vols. London 1905– 1907 Petavatthu, ed. J. P. Minayeff. London 1888. Saṃyutta-Nikāya, ed. L. Feer. 5 vols. London 1884–1898. Theragāthā, ed. H. Oldenberg and R. Pischel, rev. K . R. Norman, L. Alsdorf. Londen 21966. Vinayapiṭka, ed. H. Oldenberg. 5 vols. London 1879–1883. Visuddhimagga, ed. C. A. F. Rhys Davids. 2 vols. London 1920– 1921.

Secondary sources
Bodhi, Bhikkhu tr. 2000. The Connected Discourses of the Buddha, Volume II. Boston: Wisdom Publications. Conze, Edward 1959. Buddhism: Its Essence and Development. New York: Harper and Row. Demiéville, Paul 1927. “Sur la mémoire des existences antérieures,” in Bulletin de l’École Française D’Extrême-Orient. Tome XXVII: 283-298. Johansson, Rune 1969. The Psychology of Nirvana. London: George Allen and Unwin.

Priming the lamp of dhamma The Buddha’s miracles in the Pāli Mahāvaṃsa
Kristin Scheible

Stories of the Buddha’s miracles certainly function as narrative hooks for an engaged audience, and serve as reminders of his potent presence. But I read the Buddha’s miracle stories, or at least those that occur at the outset of the fifth century Pāli Mahāvaṃsa, as more than signs of his aptitude and authority. I argue that they function on a deeper level than strategic literary exemplification of the persuasiveness of the Buddha to convert. The Buddha’s miracles serve to engender particular emotional responses that in turn incite ethical transformation. There has been a lot of scholarly interest of late in relics and images, and how they function to compel individuals to perceive a productive proximity to the Buddha even in his absence. Just as the efficacy of Buddha images is not ascribed to human agency, so, too, the narratives of the Buddha’s miracles perform work that transcends mere didacticism or representation of the miraculous nature of the Buddha. Robert Brown suggests, “With their wondrous size and priceless material, [Buddha images] appear not so much as artistic or archaeological facts, but rather as part of a miraculous epistemology, a way of knowing through the miraculous.”1 I assert that this miraculous epistemology is fundamentally operative in Pāli narrative literature, whether it is evident in how the environment responds to the Buddha through earthquakes, garlands and perfume, or how animals are moved to serve the Buddha; or how items of use behave bizarrely, such as bowls moving upstream; or at times when the Buddha himself is performing a miracle, straight1

Brown 1998: 26. Journal of the International Association of Buddhist Studies Volume 33 • Number 1–2 • 2010 (2011) pp. 435–451

436

Kristin Scheible

up and unmediated.2 In other words, the miracles employed in Pāli texts aren’t simply another clever narrative device to hook a reader’s attention; they work upon that reader. Miracle stories in fact reflect the power they portray within the text in the work they do outside the text. A reader sufficiently prepared to expect miracles is worked upon, or primed, for the dhamma, much in the way the characters within the stories are terrified, subdued, awed, transformed, calmed, convinced, converted, and compelled to achieve higher spiritual stations by their exposure to the Buddha’s miracles. Much as a soaking in good oil will prime a lamp’s wick for the lighting, miracle stories prepare the audience for the cultivation of potent emotions and resultant ethical transformation. Miracles have often been dismissed by scholars as little more than a method of pandering to the “bhaktic and magical beliefs”3 of the masses, and as mythical and secondary accretions. But miracles are employed throughout Buddhist texts to titillate productively; they prime audiences for the profundity of religious truths and cultivation of a miraculous epistemology. The presence of miracle stories is not a marker for establishing a line between monk and lay interests. We have abundant evidence of monkish interest in miracles; for example, Xuanzang was the epitome of a monk who was nonetheless, and obviously, undeterred by magic and who recounted many miracles without hesitation.4 And paramount monk Buddhaghosa himself becomes the central character in a miraculous meeting of monks recounted in Buddhaghosuppatti.5 In this paper, I look at a literary product of the religious virtuosi of fifth century Sri Lanka written for the consumption by other monks. Sure enough, in its very opening chapter it launches immediately
2 Abundant examples of the forms and functions of such miracles can be found in the articles by Bradley Clough (2011) and David Fiordalis (2011) within this volume. 3 This is the stance taken by Conze in his Buddhist Thought in India (1973: 32), quoted in Brown 1998: 29. 4 Ibid. More work needs to be done with the slippery categories of magic and miracle; see Fiordalis (2011). 5 Buddhaghosuppatti, attributed to Mahāmaṅgala. Edited by James Gray. London: Luzac & Co. 1892.

Priming the lamp of dhamma

437

into a narrative rich in the miraculous, a literary strategy to prime its readers. The Pāli Mahāvaṃsa opens with a narrative account of the Buddha’s three visits to the island of Laṅkā. Certainly these postEnlightenment escapades of the Buddha serve to lend an air of legacy and legitimacy to the Buddhism of Sri Lanka, and the narrative that places the Buddha on the island of Laṅkā proper serves to relocate the light of the dhamma, a narrative argument for the recentering of the Buddhist world. The Buddha visits the island at specified moments on the timeline in his biography.6 The appearance of the Buddha’s physical presence in Laṅkā in the midst of his accepted biographical trajectory may be miraculous in and of itself – he must fly there, flight being a common miracle. Or can we consider the power of flight a miracle at all? Is it simply far along on the spectrum of possible human endeavors, a superpower to be sure, but one that may be humanly developed?7 It could be argued that some miracles are not in fact miraculous, but instead completely reasonable effects of extraordinary human agency. As Robert Brown suggests,
It seems that the power that explains the ability to produce miracles is rather clearly identified in this regard, with a not particularly mysterious nor mystical way of achieving it. In short, when we are talking about human beings (including the Buddha) performing miracles, the superhuman power needed for their performance is more of an understandable human achievement than an incomprehensible characteristic of a divinity. It appears to me that within the context of the Buddha’s life, the possession of iddhi, the ability to perform miracles,

The visits occur in the ninth month, fifth year, and eighth year after the Buddha’s Enlightenment. 7 As noted by Bradley Clough in his article (2011: 418), flight of the Buddha is frequently listed in the Pāli cano