Tsur, Reuven - Poetic Conventions As Cognitive Fossils

i
Poetic Conventions as Cognitive Fossils

ii
COGNITION AND POETICS

Cognition and Poetics (CAP) fosters high-quality interdisciplinary research
at the intersection of cognitive science, literature, the arts, and linguis-
tics. The series seeks to expand the development of theories and method-
ologies that integrate research in the relevant disciplines to further our
understanding of the production and reception of the arts as one of the
most central and complex operations of the human mind. CAP welcomes
submissions of edited volumes and monographs in English that focus on
literatures and cultures from around the world.
Series Editors:
Alexander Bergs, University of Osnabrück
Margaret H. Freeman, Myrifield Institute for Cognition and the Arts
Peter Schneck, University of Osnabrück
Achim Stephan, University of Osnabrück
Advisory Board:
Mark Bruhn, Regis University Denver, CO, USA
Peer Bundgard, Aarhus University, Denmark
Michael Burke, University College Roosevelt Middelburg, the Netherlands
Wallace Chafe, University of California Santa Barbara, USA
Barbara Dancygier, University of British Columbia Vancouver, Canada
Frank Jäkel, Universität Osnabrück, Germany
Winfried Menninghaus, Freie Universität Berlin, Germany
Keith Oatley, University of Toronto, Canada
Jan Slaby, Freie Universität Berlin, Germany
Reuven Tsur, Tel Aviv University, Israel
Mark Turner, Case Western Reserve University Cleveland, OH, USA
Simone Winko, Georg-August-Universität Göttingen, Germany
Dahlia Zaidel, University of California Los Angeles, USA
Cognitive Approaches to Early Modern Spanish Literature
Isabel Jaén and Julien Jacques Simon
Cognitive Literary Science: Dialogues between Literature and Cognition
Edited by Michael Burke and Emily T. Troscianko
A Prehistory of Cognitive Poetics: Neoclassicism and the Novel
Karin Kukkonen
Expressive Minds and Artistic Creations
Szilvia Csábi
Reuven Tsur
iii
Poetic Conventions
as Cognitive Fossils
Reuven Tsur
1
iv
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2017
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

Names: Tsur, Reuven, author.
Title: Poetic conventions as cognitive fossils / Reuven Tsur.
Description: New York : Oxford University Press, [2017] | Series: Cognition and poetics |
Includes bibliographical references and index.
Identifiers: LCCN 2016039315 (print) | LCCN 2017000222 (ebook) |
ISBN 9780190634698 (pbk. : alk. paper) | ISBN 9780190634681 (cloth : alk. paper) |
ISBN 9780190634704 (pdf) | ISBN 9780190634711 (online course)
Subjects: LCSH: Poetics—Psychological aspects. | Cognitive grammar.
Classification: LCC P311 .T66 2017 (print) | LCC P311 (ebook) | DDC 808.1—dc23
LC record available at https://lccn.loc.gov/2016039315
9 8 7 6 5 4 3 2 1
Paperback printed by WebCom, Inc., Canada
Hardback printed by Bridgeport National Bindery, Inc., United States of America
v
CONTENTS
Preface vii
Preface by the Author xi
Acknowledgments xvii
About the Companion Website xix
1. Where Do Conventions Come From? 1

2. Some Implications of D’Andrade’s Assumptions 19
3. Poetic Conventions as Fossilized Cognitive Devices: The Case of
Medieval and Renaissance Poetics 57
4. Frozen Formulae and Expressive Force: The Ballad “Edward” 87
5. Artistic Devices and Mystical Qualities in Hebrew Devotional Poems
with Idit Einat-Nov 99
6. Figurative Language and Sociocultural Background: Hebrew Poetry as
a Test Case 131
7. The Translated Poem as an Aesthetic Object: How Conventions
Constrain One Another in a Poem 157
8. More Is Up—Some of the Time 187
9. Some Remarks on the Nature of Trochees and Iambs and Their
Relationship to Other Meters 209
10. Poetic Language and the Psychopathology of Everyday Life 231
References 261
Index 271
vi
vi
P R E FA C E
In developing his pioneering theory of cognitive poetics over the past half
century, Reuven Tsur has embraced an astounding range of research and
approaches, including Gestalt psychology, Russian formalism, analytic phi-
losophy, stylistics, and theories of prosody. Even more remarkable is the
scope of the poetic forms, genres, and strategies to which he has applied his
theoretical and conceptual methodologies. They include approaches both
to basic themes and to aspects of rhyme, genre, archetypal patterns, period
style, and sound symbolism. In his books and articles Tsur has applied his the-
ories to English, French, German, Hungarian, Italian, Spanish, and Hebrew
poetry, ranging from the Bible through medieval to Renaissance, classic to
Romantic, and modern to contemporary poetry. Chapters in his monumen-
tal work Toward a Theory of Cognitive Poetics, now in its second expanded and
revised edition, cover issues that include the intricacies of poetic rhythm,
the complexities of poetic metaphor, meditative poetry and altered states
of consciousness, translation theory, the implied critic’s decision style, and
the thorny question of critical competence. Poetic Conventions as Cognitive
Fossils draws its illustrations from Tsur’s research into the poetries of Middle
and Modern English; Old, Middle, and Modern French; German; Hungarian;
and Medieval and Modern Hebrew. And it draws its intellectual foundations
from literary criticism, linguistic theory, phonology, cognitive psychology,
and anthropology, to name only the most prominent of its resources.
Tsur defines cognitive poetics as a theoretical methodology that explains
how poetic language and literary form are shaped and constrained by
human cognitive processes. In his contribution to Brône and Vandaele’s
book Cognitive Poetics: Goals, Gains, and Gaps, Tsur begins with an anecdote
about human information processing:
There was an old joke in Soviet Russia about a guard at a factory gate who at the
end of every day saw a worker walking out with a wheelbarrow of straw. Every
vi
day he thoroughly searched the contents of the wheelbarrow, but never found
anything but straw. One day, he asked the worker: “What do you gain by taking
home all that straw?” [The worker responded:] “The wheelbarrows.” (2009, 23)
Tsur’s point in telling the joke was to introduce the basic gestalt rules of fig-
ure and ground. But the joke certainly also resonates with Tsur’s own life
and work in finding unconventional and innovative ways to accomplish one’s
objectives. Tsur has always been smuggling out wheelbarrows, while the rest
of us were examining the straw. Following are just some of the many theo-
retical and methodological wheelbarrows he has brought out over the years.
Noam Chomsky aimed to develop a rule system for language that would
generate all and only grammatical sentences. In an article titled “Chaucer
and the Study of Prosody” published in 1966, Morris Halle and Jay Keyser
set out to do something similar for English poetry, to develop a rule sys-
tem that would generate all and only metrical lines. Having noticed that
the occurrences of many unmetrical lines in English poetry, according to
their theory, shared certain regularities, Tsur instead developed a theory
of rhythmical performance that focuses on explaining poetic effects, cul-
minating in the publication of Poetic Rhythm: Structure and Performance
(1998). Similarly, whereas linguistic research, in analyzing poetic meter,
applied phonological stress rules to distinguish rhythmic grouping in iam-
bic and trochaic feet, Tsur focused instead on their versification patterns.
His findings explain more precisely the tendency in English poetry toward
iambic groupings. His perception-oriented theory of meter is based on
empirical investigation using the latest technological resources.
Tsur’s ability to place linguistic studies within the larger perspective of
research in other disciplines and his unwavering focus on the nature of the
poetic text distinguish his work from other cognitive research. Cognitive
metaphor studies, for example, inspired by the work of George Lakoff,
Mark Johnson, Gilles Fauconnier, and Mark Turner, have tended, with
some notable exceptions, to focus on the metaphorical underpinnings of
conventional cognitive processes. For that very reason, such studies have
not been designed to illuminate the qualities that distinguish poetry from
prose or to explain how poets can use the same metaphor with different
effects. In contrast, Tsur provides convincing explanations for the differ-
ences among poetic metaphors that depend on such aspects as their discor-
dant versus concordant elements, low versus high differentiation, and split
versus integrated focus and the literary effects of synaesthetic transfer.
In Poetic Conventions as Cognitive Fossils, Tsur draws upon much of his
past research to show how conventional poetic styles originate from cog-
nitive rather than cultural principles. Whereas traditional scholarship
[ viii ] Preface
ix
assumes that cultural—and poetic—conventions can be accounted for by

tracing their developments through social transmission, Tsur seeks the
sources of poetic convention in the cognitive constraints of the brain. Over
time, he argues, we turn expressive features into fossils of cognitive orna-
ment, into characteristics of style. In the ballad tradition, for example, con-
ventions such as interjections and vocatives come to be perceived as the
rigid formulae of the genre.
Similarly, the placement of the caesura in English iambic pentameter (usu-
ally after the fourth metrical position out of ten in the iambic pentameter
line) is seen, in the “influence-hunting” model, as a development from French
poetry. But the “cognitive-constraints” model attributes the fact that this
phenomenon occurs in both the French and English traditions to the limited
channel capacity of short-term memory, such that these lines occur in unequal
segments, ordered short to long, a principle that has been shown to operate
in English prose as well and is explained by widely documented research in
perception. Tsur extends the principles of this analysis of cognitive origins
of poetic form not only to the writing systems of the Western world but also
to Egyptian hieroglyphs and Chinese and Japanese writings, to aspects of
figuration in medieval and Renaissance love poetry in English and French, to
the formulaic mystical poems of the Hebrew Middle Ages, to the metaphysi-
cal conceit, to theories of poetic translation, to the contemporary theory of
metaphor, and to slips of the tongue and the tip-of-the-tongue phenomenon,
showing the workings and disruption of psycholinguistic mechanisms.
Poetic Conventions as Cognitive Fossils sets a high standard for the principles
of scholarly learning—formidable yet available, humanistic in the best sense,
and a fitting capstone to the career of an illustrious scholar. We are pleased to
acknowledge the immense contribution Reuven Tsur has made to the field of
cognitive poetics with this latest publication in Cognition and Poetics.
Alexander Bergs
Margaret H. Freeman
Peter Schneck
Achim Stephan
Biographical Note: Reuven Tsur was born in 1932 in Nagyvárad, Transylvania

(now Oradea, Romania). He holds a B.A. in English and Hebrew literature
from the Hebrew University, Jerusalem, and a D.Phil. in English from the
University of Sussex. He is a professor emeritus of Hebrew literature and
literary theory at Tel Aviv University. In 2009, he was awarded the Israel
Prize in “General Literature,” and he received an honorary doctorate from
Osnabrück University in 2013.
P r e fac e [ ix ]
x
xi
P R E FA C E B Y T H E A U T H O R
One of the most exciting minutes in one’s research record is when one
discovers that a wide variety of one’s writings have grown together into
one coherent whole. That is what I experienced when I wrote my Toward a
Theory of Cognitive Poetics (Tsur 2008a): I discovered that all my work over
the decades had grown together into a coherent, comprehensive theory of
cognitive poetics. I had a similar experience when, in writing this book, a
subset of my work coalesced into a coherent subfield of cognitive poetics
that conceives of poetic conventions as cognitive fossils shaped and con-
strained by cognitive constraints, rather than accounting for them through
an infinite regress of ever-earlier influences.
SYNOPSIS OF THIS BOOK
This book explores the questions of where conventions come from, how
they work, how they could be understood by their first audience, and how
they could become conventions. Chapter 1 presents a “constraints-seeking”
or “cognitive-fossils” approach to conventions as against the prevalent
“migration,” “influence-hunting,” or “culture-begets-culture” approach.
Two of its main assumptions are derived from Roy D’Andrade and
Anton Ehrenzweig. According to the former, cultural programs (in this
case poetic inventions) may become conventions when in the course of
repeated social transmission they take on forms that are consistent with
the natural capacities and constraints of the human brain. According to
the latter, in order to avoid the more “dangerous” expressive elements of
an artistic work, we tend to neutralize expressive features by formalizing
them—we turn them into characteristics of style, into harmless ornament.
The culture-begets-culture approach, by contrast, accounts for conventions
by pointing at earlier occurrences in other cultures, assuming that the
xi
earlier occurrence influenced the later one, merely transferring the mys-
tery from one place to another.
I use the term fossils in a weaker and a stronger sense. In the weaker
sense, for example, such verbal devices as interjections, exclamations,
vocatives, and repetitions have emotive force owing to their relation to
processing in the right hemisphere of the brain. In Elizabethan drama,
interjections followed by repeated vocatives became a hackneyed conven-
tion that affects the audience’s beliefs about the character’s emotion rather
than a perception thereof. In the stronger sense, in the ballad “Edward”
the same conventions turn into a rigid formula that is perceived as “ballad
style” rather than expressive devices: they affect the reader’s beliefs about
the genre rather than about the speakers’ emotions.
The opposition between the constraints-seeking and influence-hunting
approaches may be illustrated through their handling of caesurae.
According to the cognitive approach, the rhythmic processing of poetry is
constrained by the limited channel capacity of the cognitive system. This
has shaped a wide range of versification conventions in a number of lan-
guages and versification systems. Thus, in the English iambic pentameter
line, the caesura may occur after the fourth, fifth, or sixth position (out of
ten), but the vast majority occur after the fourth. The influence-hunting
approach attributes this distribution to the influence of French poetry,
where the majority of caesurae occur after the fourth position in deca-
syllabic lines. The cognitive-constraints approach, by contrast, assumes
that owing to the limited channel capacity of immediate memory, when
there are two parallel segments of unequal length, the longer coming last
is felt to be the well-shaped order in both English and French poetry,
and in many other languages, as well as in music. The influence-hunting
approach cannot explain why French poetry places the caesura after the
fourth position or why English poets have adopted this practice in a dif-
ferent versification system—it merely transfers the mystery from one
place to another.
In c hapter 2 I elaborate on the implications of Roy D’Andrade’s concep-
tion that in the course of repeated social transmission cultural programs
take on forms that are consistent with the natural capacities and con-
straints of the human brain. I explore two crucial cognitive constraints and
their literary implications: (1) The cognitive system has a limited channel
capacity and therefore demands great parsimony; and (2) language is typi-
cally logical and conceptual, whereas some poetry at least is supposed to
convey emotional or mystical qualities that are nonconceptual and illogi-
cal. As to repeated social transmission, I quote Bartlett, who staged a con-
trolled experiment with an Eskimo story that in the course of repeated
[ xii ] Preface by the Author

xi
transmission assumed a good fit to schemata entertained by Western

readers. Then I follow Stanislas Dehaene, who discovered, in the brain
area responsible for reading, neurons shared with other primates that
had evolved for other survival purposes (visual constancy) that are sensi-
tive to certain letter-like shapes. These shapes have been shown to be the
graphic primitives that constitute the characters of all the world’s writing
systems, except Egyptian hieroglyphs, which display a good fit to the physi-
cal objects out there in the world. After the alphabetic revolution, it took
about half a millennium for Semitic writing to relinquish, in the course of
repeated social transmission, the mimetic shapes and assume a good fit to
the shapes to which neurons in primate brains are sensitive. Finally, I point
out that going against those cognitive constraints may have its own expres-
sive value as well, even when it is an established convention.
In chapter 3 I argue that in medieval and Renaissance poetry the use of
genres is highly conventional and is determined by vogue or social conven-
tions rather than the poets’ psychological needs. The first-person singu-
lar is frequently regarded as a rhetorical device or a device whose meaning
changes with the requirements of the genre rather than as evidence of
personal experience. The notion of “the spontaneous overflow of powerful
feeling” as an indication of excellence is alien to these poetries. For this rea-
son it is wrong to assume that we have explained a poet’s subjective experi-
ence by the application of some psychological theory to his or her poems.
Rather, it is more productive to apply D’Andrade’s cognitive-constraints
and Ehrenzweig’s style-as-defense conceptions. The ambivalence inher-
ent to love is a potential source of undesirable stress. Splitting the mother
figure into a dead good mother and a straightforward wicked stepmother
in dreams and fairy tales is less threatening than an ambivalent attitude.
The same applies to the convention of catalogs of contradictions, which
was widespread in medieval and Renaissance love poetry: the undesirable
stress of an ambivalent feeling can be made less threatening by “sharpen-
ing” it into a straightforward logical contradiction. Logical contradictions
can be threatening as well, but this disturbing element can be alleviated by
inclusion in a catalog of contradictions. Finally, I explore the varieties of
poetic effects resulting from aesthetic manipulations of these catalogs of
contradictions.
Chapter 4 is a close reading of the popular ballad “Edward.” In this
ballad such emotion-laden verbal devices as vocatives, interjections, and
repetitions are fossilized into “lifeless” formulae and leave little room for
expressing complex meanings. A detailed analysis demonstrates how an
exceptionally complex human situation nevertheless emerges from these
fossilized poetic devices.
P r e fac e b y t h e Au t h or [ xiii ]
xvi
Chapter 5, written in collaboration with Idith Eynat-Nov, explores even

more rigid formulaic poetry in the Hebrew Middle Ages. These poems are
said to be derived from earlier Merkabah mysticism. Some scholars even
claim that the poets actually underwent mystical experiences. We argue
that such frozen formulae cannot be evidence of fervent mystical expe-
riences. In contrast, we quote from the description of an actual mystical
experience, in which the experiencing subject lost control of his physical
and mental self. We argue that these rigid formulae are elements derived
from such fervent mystical experiences after the process of fossilization
suggested by Ehrenzweig expunged their disquieting element.
Chapter 6 investigates the complex relations between the mind and the
world that determines it and is determined by it and the cultural forms
that spring from this interaction and feed back into it. It discusses medi-
eval, Romantic, and modern Hebrew poetry and focuses on such Mannerist
stylistic devices as metaphysical conceit and the arbitrary genitive, both of
which can be described in Dr. Johnson’s phrase as “the most heterogeneous
ideas yoked together by violence” (1951 [1779]). I treat them as adaptation
devices adapted to aesthetic ends: they may shock the reader out of tune
with the world, and the reader must then readjust to it in a more adequate
manner. Between shock and readjustment the reader’s own adaptation
mechanisms become perceptible, indicating that they are functioning prop-
erly. Such “adaptation devices” tend to occur in sociocultural and historical
periods marked by strife and social instability, in which the world picture
has been undermined or is dominated, simultaneously, by more than one
scale of values, with no firm ground to choose between them.
While this book is about the ways in which the natural constraints and
capacities of the human brain shape and constrain poetic conventions,
chapter 7 is about how the various conventions constrain each other in
an aesthetic object. Initially, it was written as a counterproposal to Clive
Scott’s claim that translation, irrespective of what kind of verse is being
translated, should always opt for free verse, but it also has wider implica-
tions for the nature of the aesthetic object in general and poetic rhythm
in particular. This chapter reviews three possible approaches to the issue,
in theory and practice. The first is literal translation plus close reading;
the second opts for free verse; the third seeks an elegant solution to a
problem, one that results in an aesthetic object in its own right. This is
the proposal expounded here. None of these approaches is right or wrong
per se. Each one consists of a specific package of gains and losses that can
provide the reader with different kinds of experience. Thus, each approach
is legitimate in its own way and should be assessed in view of the kind of
experience it offers.
[ xiv ] Preface by the Author

xv
In chapter 8 I explore the cognitive rationale of why we call sounds

with greater frequency “high” and those with smaller frequency “low” and
why we perceive higher sounds as “thinner.” The problem is set by Zohar
Eitan and Renee Timmers’s intercultural studies of pairs of antonymic
adjectives used to describe music. I compare and contrast two cognitive
attempts to account for the terminology used in our culture: Roger Brown’s
and George Lakoff’s “mediated-association” approach and the present
“cognitive-constraints” approach. The latter assumes that the spatial per-
ception of sound is a cognitive fossil, originating in a solution to a problem
posed by cognitive constraints, in this case the inconstancy of sound and
the limitation on the resolving power of the ear.
In chapter 9 I explore some perceived qualities of iambic and trochaic
meters. I argued above that in two or more parallel linguistic units, “lon-
gest comes last” is perceived as the well-ordered structure of the utterance.
In versification I showed this on both the line-segment and foot levels.
This principle, however, cannot account for such subtleties as Seymour
Chatman’s observation “that lighter syllables seem more readily ictic in tro-
chaic than in iambic feet” (1965). Theoreticians from Aristotle and Horace
through Jespersen and Chatman to Halle and Keyser have noted that the
iambic meter has a very different effect than the trochaic. In a wide variety
of languages the iambic is felt to be more natural than the trochaic, even in
Hungarian, where stress falls invariably on the first syllable of a word. The
culture-begets-culture approach accounts for this phenomenon in English
poetry by such extraneous reasons as “Germanic strong-stress verse” and
“the comparatively short history of the trochaic mode” in English poetry.
This chapter offers a cognitive explanation, based on two sets of experi-
ments. One of them is H. Woodrow’s tick-tack experiment in the 1920s,
paraphrased as “durational differences tend to result in ‘end-accented
rhythms,’ and intensity differences tend to result in ‘beginning-accented
rhythms.’ Pitch has neither a group-beginning nor a group-ending effect.”
Curt Rice’s (1992) more recent experiments suggest that variations in pitch
lead to a significant shift toward iambic groupings. The other set involves
D. B. Fry’s (1958) experiments with stress perception: the acoustic cue for
stress is a complex of pitch, duration, and loudness, in this order of decreas-
ing effectiveness. If pitch differences are irrelevant to grouping direction
and duration differences are more effective in stress perception than ampli-
tude differences, end-accented meters should be more natural in poetry. If
variations in pitch also lead to a significant shift toward iambic groupings,
that should reinforce this effect.
Chapter 10 speculates on the nature of poetic language, on the one hand,
and the Freudian psychopathology of everyday life (slips of the tongue, the
P r e fac e b y t h e Au t h or [ xv ]
xvi
tip-of-the-tongue phenomenon, etc.), on the other. Both are organized

violence against language—the former for aesthetic purposes, the latter
for psychopathological purposes. I assume that the same psycholinguistic
mechanisms underlie both and may generate similar experiences. Those
psychopathologies allow us to catch a glimpse of the invisible workings of
psycholinguistic mechanisms and their disruption.
Briefly, this book expresses dissatisfaction with the prevalent “migra-
tion” or “influence-hunting” approach to literary conventions, which
merely transfers problems from one place to another. It proposes, instead,
a “constraints-seeking” conception of “cognitive fossils,” which is pregnant
with implications.
Now, a final personal note. I usually work alone, and hardly ever get
feedback. This time, however, I benefited from a sustained and highly fruit-
ful dialogue with my old friend Don Freeman. The final version of this book
reflects his ability to encourage me to clarify certain obscurities, which in
many cases led to incorporating new exciting insights that illuminate my
argument from unexpected angles.
[ xvi ] Preface by the Author

xvi
ACKNOWLEDGMENTS
Parts of my 2010 article “Poetic Conventions as Cognitive Fossils” appear

in chapters 1 and 2, with the kind permission of Penn State University
Press and the editor of Style.
Chapter 3 has been reproduced from PsyArt, with warm encouragement
of the editor.
Chapter 4 and parts of chapter 6 have been reproduced from the second
edition of my Toward a Theory of Cognitive Poetics (Tsur 2008a), with the
kind permission of Sussex Academic Press.
The discussion of Merkabah hymns and some of the theoretical assump-
tions in chapter 5 as well as of Herbert’s “Anagram” in chapter 10 have been
reproduced here with the kind permission of Imprint Academic.
With the kind permission of Duke University Press I have repurposed
three issues from my What Makes Sound Patterns Expressive? The Poetic
Mode of Speech Perception (Tsur 1992a): my criticism of “mediated associa-
tion” in c hapter 8, my discussion of Brill’s work in chapter 10, and my dis-
cussion of Verlaine’s poem in chapters 7 and 10.
Chapter 7 has been reproduced with minor changes from Thinking Verse,
with warm encouragement of the editor.
The discussion of the TOT phenomenon has been reproduced here
from my Playing by Ear and the Tip of the Tongue by the kind permission of
John Benjamins Publishing Company.
The publisher and I gratefully acknowledge the assistance and agree-
ment of the publishers and editors cited above. The publishers apologize
for any errors or omissions in the above list and would be grateful to be
notified of any corrections that should be incorporated in the next edition
or reprint of this book.
xvi
xi
ABOUT THE COMPANION WEBSITE
www.oup.com/us/tsurpoeticconventions/
Oxford has created a companion website that houses sound files accompa-
nying Chapters 7 and 8. The reader is encouraged to consult this resource.
x
xxi

xxi
1
CHAPTER 1
Where Do Conventions Come From?
Tongs are made with (other) tongs, but who made the original [pair of] tongs? It was,
perforce, a Heavenly creation!
—Babylonian Talmud, Tractate Pesachim, 54a1
I n this book I contrast two approaches to cultural conventions in gen-

eral and poetic conventions in particular: the prevalent “migration” or
“influence-hunting” or “culture-begets-culture” approach and the recently
emerging “constraints-seeking” or “cognitive-fossils” approach. The for-
mer, associated with so-called traditional scholarship, assumes that cul-
tural programs can be accounted for by identifying their roots in earlier
cultural phenomena and providing a map of their migrations. The latter,
associated with cognitive and evolutionary criticism, assumes that cul-
tural programs originate in cognitive solutions to adaptation problems
that have acquired the status of established practice. This criticism can-
not account for adaptation solutions. But once these adaptation solu-
tions are accounted for by work in other disciplines, they can account
for poetic conventions as well. As will be seen, both conceptions assume
“repeated social transmission,” but with very different implications.
According to the “constraints-seeking” or “cognitive-fossils” approach,
repeated social transmission can do everything that the “culture-begets-
culture” approach attributes to it; but, as we shall see, it can account for
much more.
1. I am indebted to Ronit Shoshany for this citation.

2
The “migration” or “influence-hunting” approach sees a convention as

given. It assumes that cultural phenomena can be satisfactorily explained
as reflecting some influence or by being a convention and that no questions
need be asked about the mediating processes in the black box of the head.
In terms of culture begets culture, poetic conventions are seen as linguistic
and conceptual constructs to which earlier generations attributed certain
meanings and effects; later generations are somehow conditioned to attri-
bute to them the same meanings and effects. This approach does not ask
why the first audience should have attributed the poet’s intended mean-
ings or effects to them or how those attributions acquired the status of
widespread conventions.
On the other hand, the “constraints-seeking” or “cognitive-fossils”
approach assumes that cultural and poetic conventions are verbal con-
structs that reflect active cognitive (sometimes depth-psychological) pro-
cesses and constraints that have become fossilized over time. For example,
poetry is frequently expected to convey emotions. But language is con-
strained by its conceptual and logical nature. This creates a problem for
some poets. One solution to this problem is to have recourse to verbal
devices that in one way or the other mobilize cognitive resources associ-
ated with emotional processes. The use of such verbal devices may remain
sporadic or become in due course poetic conventions. This approach must
thus also account for how the sporadic, sometimes ingenious inventions
of creative individuals can become generally accepted conventions, some-
times even in unrelated cultures.
COGNITIVE FOSSILS
Fossil refers to the remnant or trace of something constituted to carry

on the activities of life, but only some of its outward shape has been pre-
served, not the activities of life it was constituted to carry on. I am using
the term cognitive fossils in a weaker and a stronger sense. The weaker sense
suggests artistic devices, structures, motives, and so on that originated in
solutions to problems posed by cognitive constraints and may convey, to a
considerable extent, perceptions and experiences related to the cognitive
processes involved. In time they have acquired persistence and a publicly
accessible status. At this stage, the cognitive problems that prompted them
are forgotten, and the perceptual and experiential effects are considerably
diminished and sanctioned by common practice—the devices become con-
ventions. Eventually, such conventions can also be used without any aware-
ness of these perceptual and experiential effects.
[ 2 ] Poetic Conventions as Cognitive Fossils

3
As an epitome of “cognitive fossils” in the weaker sense, consider some

of the illustrious speculations on the putative origins of language:
I cannot doubt that language owes its origin to the imitation and modification,
aided by signs and gestures, of various natural sounds, the voices of other ani-
mals, and man’s own distinctive cries. . . . [W]‌e may conclude from a widely
spread analogy that this power would have been especially exerted during the
courtship of the sexes, serving to express various emotions, as love, jealousy,
triumph, and serving as a challenge to their rivals. The imitation by articulate
sounds of musical cries might have given rise to words expressive of various
complex emotions. (Darwin 1871: 56)
Let us complement Darwin’s passage with another one by Gombrich, from

his essay “Meditations on a Hobby Horse”:
Even the origin of language, that notorious problem of speculative history,

might be investigated from this angle. For what if the “pow-wow” theory, which
sees the root of language in imitation, and the “pooh-pooh” theory, which sees
it in emotive interjection, were to be joined by yet another? We might term it
the “niam-niam” theory postulating the primitive hunter lying awake through
hungry winter nights and making the sound of eating, not for communication
but as a substitute for eating—being joined, perhaps, by a ritualistic chorus try-
ing to conjure up the phantasm of food. (1985: 5)
If Darwin and Gombrich are right, we have in the origins of language a

perfect model of fossilization. According to this model, words originate in
sounds emitted to serve as imitation, expression of emotions, or substitu-
tions for real things—for emotional or magical purposes. One didn’t need
to explain the meanings of those vocalizations; they were self-explanatory
to everybody, as they are to courting animals—knowledge unlearned and
untaught. Today we have an enormous vocabulary for all possible needs
of communication, and if Darwin and Gombrich are right, all words are
derived from those three types of vocalization: imitation, expression, and
substitution. Today most of the words are completely detached from all
those emotions or magic and accompanying gestures—and are publicly
accessible.
The stronger sense is derived from Anton Ehrenzweig’s (1965) concep-
tion that in order to avoid the more “dangerous” expressive elements of
an artistic work, we tend to suppress expressive features by formalizing
them—we turn them into style, into harmless ornament, as in the exam-
ples in chapters 4–5.
Where Do Conventions Come From? [ 3 ]

4
To give just one continuous example of the weaker and the stronger
senses to be elaborated later, one of the main problems facing poets is
that they have to convey information that is nonlogical and nonconceptual
through language, which is logical and conceptual. One of their preferred
tools for doing this is to transfer some of the linguistic information process-
ing to the right, emotional hemisphere of the brain. One possible device
for accomplishing this is to have recourse to interjections and vocatives,
arousing right-hemisphere information-processing activities. When this
practice became a poetic convention in much Elizabethan drama, it came
to affect the audience’s beliefs concerning the characters’ emotions rather
than stimulating direct perception of the emotional experiences. When this
convention was overused, it was parodied by Shakespeare in the Pyramus
and Thisbe interlude in A Midsummer Night’s Dream (see c hapter 2). Finally,
the vocatives and interjections became a rigid poetic formula in the ballad
“Edward,” in the stronger sense of the term cognitive fossils (see c hapter 4).
It would appear that such fossilizing processes are ubiquitous in human
experience and take place outside art and language too, in the social
domain. I propose to adduce here an illuminating, nonliterary example of
fossilization concerning customs that originated in solutions to very real
problems posed by the constraints of certain historical circumstances.
Those customs became detached in the course of repeated social transmis-
sion through the centuries from the causes that had prompted them, until
their practitioners had no idea why they would observe them at all.
Recently I met a young woman, originally from Spain, whose grand-
mother had observed a custom of lighting candles in the cupboard. It’s in
memory of the dead, she used to tell her. Her grandfather would go to the
butcher, buy pork meat, go home, and throw it out into the garbage. When
asked why he did this, he said that that was what his father had taught
him to do. Her mother was an inquisitive soul who eventually discovered
that they were descendants of the Marranos—Jews living in medieval
Spain who converted to Christianity but continued to practice Judaism
in secret. Those “New Christians” whose chimney was observed not to
smoke on Shabbat were subjected to the Inquisition. Obviously, they also
had to buy pork meat to demonstrate that they didn’t practice Judaism.
They preserved these customs over the centuries, even when there were
no persecutions anymore, even when they no longer remembered their
Jewish origins—when those practices had become mere fossils. This fam-
ily has settled now in Israel and converted to Judaism, and they are happy
orthodox Jews.
I am going to present three main points in the thesis of this book through
three passages quoted from different sources. The first one is taken from

5
the introduction to the Literature and Cognition website, which is no lon-

ger available online:
The past twenty years of cultural studies have focused on the ways in which cul-
ture begets culture, ignoring the cognitive capacities that are the preconditions
for cultural change. We investigate the complex relations between the mind, the
world that both determines it and is determined by it, and the cultural forms
that arise from this interaction and feed back into it.2
“Culture begets culture” suggests infinite regress, just as tongs are made
with tongs. But who made the original cultural program? Or who made the
original (pair of) tongs?
The second passage is by the cognitive anthropologist Roy D’Andrade:
An important assumption of cognitive anthropology is that in the process of

repeated social transmission, cultural programs come to take forms which have
a good fit to the natural capacities and constraints of the human brain. Thus,
when similar cultural forms are found in most societies around the world, there
is reason to search for psychological factors which could account for these simi-
larities. (1981: 182)
I will argue that “cultural programs” have solid cognitive foundations and
are shaped and constrained by the natural capacities and limitations of the
human brain, which results in certain significant regularities.3 I will also
argue that the “culture-begets-culture” paradigm cannot explain some of
the most significant aspects of cultural processes that are easily accounted
for by the “cognitive-constraints” paradigm. I will attempt to explore the
ways in which an infinite variety of cultural forms can arise in cultural pro-
grams constrained and shaped by the same cognitive capacities.
The third passage is derived from Ehrenzweig: “In music as in the visual
arts, the secondary elaboration succeeds in transforming a chaotic primi-
tive form into a rigid and ornamental style” (1953: 118). I will argue, then,
following Ehrenzweig, that there is a cognitive (or depth-psychological?)
mechanism of “secondary elaboration” that transforms chaotic forms
and stylistic devices that have “too much” emotive power into rigid orna-
ments, into distinctive characteristics of some style. The vocatives repeated
2. I assume that by “cultural studies” the paragraph means simply “the study of cul-
ture,” and not “a field of theoretically, politically, and empirically engaged cultural
analysis.”
3. In this sense, cultural programs include poetic conventions.

6
throughout “Edward,” for instance, lose their emotional force and may be
regarded as distinctive characteristics of the ballad style. Nevertheless, I do
not deny the validity of the “culture-begets-culture” paradigm; I simply argue
that it should not usurp the legitimate place of the “cognitive-constraints”
paradigm.
This book is devoted to the question of how cognitive processes shape
and constrain cultural and literary forms. It assumes that the generation
of cultural forms has to do with the deployment of devices that adapt to
the individual’s physical and social environment, whereas the response
to poetry, a literary form, involves adaptation devices turned toward aes-
thetic ends. The ensuing argument contrasts the view that “culture begets
culture” with the claim that the generation of culture is governed by adap-
tation devices exploited for cultural and aesthetic ends.
The theoretical choice between these two approaches is, to a very large
extent, a matter of prevalent intellectual and academic trends; these trends,
in turn, are no mere freaks of intellectual history but are related to more
general attitudes. According to Morris Weitz (1962), the role of theory in
aesthetics is to make a “crucial recommendation” as to what to look for
and how to look at it in works of art. In the present book I give two ways
of looking at cultural history in general and literary history in particular.
I suggest that the acceptance or rejection of one point of view or the other
is determined not only by prevalent intellectual trends but also by the crit-
ic’s cognitive complexity. The theoretical foundations for this conception
in terms of “perception,” “information processing,” and “personality style”
have been described in detail elsewhere (Tsur 1975, 1979, 2006c: 11–77).
The verb and the notion of “influence” are an illuminating case. I will
not review or document the controversies concerning this notion in the
twentieth century or the resulting crisis in comparative literature. Rather,
I will make only two brief comments: this verb entails a semantic-syntactic
anomaly, and the conflicting approaches discussed here are related to wider
issues—namely, to conflicting worldviews. Specifically, the use of the verb
influence is anomalous in this context and suggests that its subject is the
agent and its object is the recipient. When we say “A influenced B,” the sug-
gestion is that A is active, whereas B is passive. The “older” culture forces,
so to speak, the “younger” culture to accept some of its tenets. This is, in
fact, an authoritarian conception of cultural processes: namely, that the ear-
lier the culture, the greater its authority. This view denies B the possibility
of autonomous (let alone creative) action. In fact, the opposite is true: In
intercultural contacts B is active; it is B that selects the cultural elements it
needs from neighboring cultures to achieve its exigencies and yearnings.
Underlying such cultural exigencies and yearnings are tensions, pressures,

7
and wants generated within the system, whether cultural or cognitive. In this
sense, cultural change is motivated by a tendency to resolve such tensions,
pressures, and wants by accommodating them in a wider structure. This is a
self-realization conception of cultural interactions. In this type of view, it is not
enough to realize that in cultural interactions the recipient is active; one must
also realize that such activities arise from certain problem-solving needs. This
approach is shared by cognitive poetics, on the one hand, and cultural studies
as implemented by Itamar Even-Zohar (1978) and his followers, on the other.
The influence-hunting approach assumes that when the same conven-
tions occur in different places at different times, there must have been
an influence; the scholar’s task is to map how this influence proceeded in
space and time. The cognitive-constraints approach allows for such a pos-
sibility but also that the same convention can be reinvented several times,
independently, in several places, for similar reasons.4 As early as 1890, Sir
James Frazer observed in his monumental comparative studies The Golden
Bough (1922) and Folk-Lore in the Old Testament (1919) that “many of these
4. This seems to be true of biological evolution as well, as Deric Bownds claims:
During evolution there have been frequent reinventions of similar adaptations,

and there have been losses, shifts, and even reversals of function. For example,
the wings of birds and the wings of bees are analogous traits that perform the
same function but arose independently on different branches of the evolution-
ary tree. In contrast, the limbs of vertebrates descended from a common ances-
tor and are referred to as homologous traits. The fins of ancestral fish, used for
swimming, evolved into the legs of ancestral reptiles, birds, and mammals, used
for running or hopping on land. The front legs of certain ancestral mammals and
reptile-birds then evolved into the wings of bats and modern birds, respectively,
and came to be used for flying. Bird wings and mammalian legs then evolved
independently into the flippers of penguins and whales, respectively, thereby
reverting to a swimming function and effectively reinventing the fins of fish. At
least two groups of fish descendants independently lost their limbs to become
snakes and legless lizards. (1999: 32–33)
Even humans could have been “invented” more than once over the course of evolution:
A controversial question has been whether H[omo] erectus evolved into H. sapi-
ens, or a separate line that became H. sapiens displaced H. erectus worldwide as
it did the Neanderthals in Europe. Two main hypotheses have been proposed.
One hypothesis argues for a multiregional evolution of humans whereby H. erec-
tus, H. neanderthalensis, and other populations semi-independently became
the modern races of H. sapiens, with some gene flow between them. The other
hypothesis contends that all modern humans derive from a small group of com-
mon ancestors that lived in Africa. These individuals evolved into the modern
races of humans during their emigrations from Africa and displaced all other
hominid lines. (Bownds 1999: 99)
In either case, the various kinds of hominids sprang from different branches of the
evolutionary tree.

8
resemblances are to be explained by simple transmission, with more or less

of modification, from people to people, and many are to be explained as
having originated independently through the similar action of the human
mind in response to similar environment” (1919: 104). For instance, there
are myths all over the world that tell the story of a flood, in European,
Amerindian, Asian, and Australian aboriginal cultures. In some of the Old
World cultures the Sumerian-Babylonian-biblical influences can be traced,
but most of them are independent inventions. How does this happen?
Gradually I realized that the question “How?” although usually defined
as an adverb of mode, is, in fact, a request for more detailed information.
Frazer does not go into the processes of the human mind. He presents
a very general hypothesis that may account for the observable facts. Its
main importance is to point to the need for two alternative approaches.
D’Andrade gives a more detailed account of this process. In c hapter 2 I will
examine in greater detail the implications of D’Andrade’s position.
Similar cultural or literary conventions in different societies can be
accounted for in several ways. The least interesting (and perhaps the least
plausible) possibility is that one society adopts, in a servile manner, the
conventions and institutions of another society with an authoritative
status. The other extreme, as suggested by Frazer (1922), is that in simi-
lar circumstances, the human mind functions in similar ways; hence the
similar cultural or literary conventions can be (re)invented more than
once. An interesting variant on Frazer’s idea is reflected in the assertion by
D’Andrade cited above, “that in the process of repeated social transmission,
cultural programs come to take forms which have a good fit to the natural
capacities and constraints of the human brain” and thus become similar,
even though they may initially have been very different. In between these
two poles is the third possibility that a culture can “borrow” cultural pro-
grams from neighboring societies to ease the cognitive accommodation of
tensions within its social and cultural structures. This seems very plausible
but tells only a small part of the story. These borrowed cultural programs
also have to arise in one way or another. To account for this we must have
recourse to Frazer and D’Andrade or some similar approach.
COGNITIVE CONSTRAINTS VERSUS INFLUENCE HUNTING
There is a scholarly tradition, the so-called Russian school of verse study,

whose members assert that their approach is based not so much on insight
and intuition as on wide quantitative analyses of observable facts. In what
follows, I will seek to show that “quantitative analyses of observable facts”

9
can also provide some indications on the “insights and intuitions” of poets
dead for hundreds of years. As to “how those attributions acquired the
status of widespread conventions,” let us consider the cognitive principles
that govern the placement of caesurae.
Such verse lines as the iambic tetrameter, pentameter, and hexameter
are divided by a caesura into two roughly equal segments. Classical prosody
requires a caesura at midline; in my view, this convention reflects certain per-
ceptual needs (see Tsur 1977a: 66–82, 1992b: 134–139, 1998: 113–139; see
chapter 2). Iambic lines, whose number of positions is divisible by four, can
be divided into segments of equal length and equal structure. The tetrameter
and the hexameter can be divided into 4 + 4 and 6 + 6, respectively, with each
segment beginning with a weak position and ending with a strong position.5
In these meters there is a very active caesura at the exact middle, even if it
is not marked by punctuation and even if it is overridden by the absence of
a word boundary. The iambic pentameter, by contrast, can be divided into
5 + 5 positions, but in this case the first segment begins and ends with a weak
position, whereas the second segment begins and ends with a strong posi-
tion. Alternatively, there can be two segments of unequal length, containing
4 + 6 or 6 + 4 positions, where each segment begins with a weak position and
ends with a strong position. So a caesura is more flexible in the pentameter
than in the tetrameter and hexameter, where it is rigidly fixed. As a result,
in the tetrameter, short rigid units follow fast upon each other, giving rise to
certain perceptual qualities especially when coupled with the trochaic meter.
Aristotle, for instance, perceived a quality in them that is “too much akin to
the comic dance” as opposed to the “characteristic rhythm of people as they
talk” (1932). I also claim that in the syllabotonic system in English, Hebrew,
and Hungarian poetry, 4 + 6 lines are overwhelmingly more frequent than
6 + 4 lines.6 The same asymmetry is found in French syllabic verse. I con-
clude, therefore, that the 4 + 6 line contains the unmarked caesura and the
6 + 4 line contains the marked caesura (see Tsur 1977a, 1992b; Tsur 1998 also
provides the cognitive reasons for this conclusion).
A classic instance of the “culture-begets-culture” approach to cultural
processing is to be found in the oft-heard claim, for example, with respect to
5. I use the terminology of Halle and Keyser (1966, 1971), who define meter as an
abstract pattern of regularly alternating weak and strong positions actualized in vari-
ous systems by strings of alternating stressed and unstressed syllables, or long and
short syllables.
6. In Hungarian iambic pentameter lines, the majority of the caesurae occur after
position V; but regarding the 4 + 6 or 6 + 4 division, the former division is the
dominant one.

01
poetic meter, that “the hemistich form 4 + 6 of the English iambic pentam-
eter is derived from the pattern of the French decasyllable” (Anonymous
reviewer, personal communication). But a far stronger argument can be
found for the “cognitive-constraints” approach in the seemingly unrelated
findings of an overwhelming tendency to place the caesura after the fourth
rather than the sixth position of decasyllabic lines in languages as remote
from each other as English, French, Hungarian, and Hebrew and in such
different versification systems as syllabic and syllabotonic meters. This set
of sufficiently unrelated findings clearly calls for a cognitive rather than an
influence-based explanation. It is much easier to show that this was the
case in French poetry and in sixteenth-century English poetry than that
the latter was derived from the former.
A number of important questions follow from these research findings.
Why does the same pattern recur in Hebrew and Hungarian poetry? Why
did the French poets, in the first place, prefer the 4 + 6 division so emphati-
cally rather than a random distribution? And why did this French practice
become such a widespread convention, in widely different versification sys-
tems? In English, French, Hungarian, and Hebrew I have not found a single
poet who uses a majority of verse lines with a segmentation of 6 + 4. The
culture-begets-culture approach merely transfers these mysteries from one
place to another.
There is also an interesting phenomenon in these languages: namely,
that in all these poets (with the notable exception of Shelley) the propor-
tion of “marked” caesurae increases with the proportion of marked forms
in other prosodic respects. In what follows I will claim that the marked
nature and unmarked nature of caesurae are determined by cognitive fac-
tors and are independent variables. But they may be associated with a vari-
ety of other independent variables, marked or unmarked, on the prosodic,
semantic, syntactic, and thematic levels; interact with them; and contrib-
ute to the generation of a wide variety of “regional qualities”7 that may
have great aesthetic significance. The culture-begets-culture approach, by
contrast, maintains that, as an anonymous reviewer from the Russian tra-
dition of verse studies commented on an earlier paper of mine,
the hemistich form 4 + 6 of the English iambic pentameter is derived from the
pattern of the French decasyllable; it is characteristic of periods where there
7. Regional quality is a Gestaltist term taken from Beardsley’s Aesthetics. It refers
to a perceptual property of a whole that is not a property of its parts (Beardsley
1958: 83–88).

1
was a more rigorous iambic canon in English literature. In the later works of
Shakespeare, in the verse of Romantic and Post-romantic poets the “hemis-
tich boundary” moves to the right, positions 6/7. What makes the 4 + 6 form
unmarked? The fact that it appears in a more rigorous variant of the English
iambic pentameter?
This and my approach represent two opposite lines of thought. In one line
of thought, the caesura derives its cognitive and aesthetic nature from the
other neighboring elements; in the other, certain cognitive factors deter-
mine the structural and aesthetic potential of the caesura, which, in turn,
contributes to the overall effect. Thus, in certain poetic styles the caesura
tends to move to the right (the later Shakespeare, Milton), back to the left
(Pope), and then back to the right (some Romantics). This is not a random
process but, rather, is governed by certain pervasive cognitive and aes-
thetic principles.
The exception in the context of the relative frequency of caesurae is
Milton, where there is no significant difference in the number of caesurae
after positions IV, V, and VI. This forces the conclusion that the unmarked
form is the caesura after position IV. Had a similar distribution occurred
in Pope, we would have to conclude that caesurae after positions IV and
VI are equally unmarked. These different conclusions relate to a distinc-
tion I make elsewhere (e.g., Tsur 2008a: 100–104) between convergent
and divergent poetry. In Milton’s divergent poetry, contrasts are typi-
cally blurred (for instance, by stressed syllables in weak positions and
unstressed ones in strong positions and by divergent patterns of allitera-
tion). Statistically, the difference in deviant stresses between Milton and
Pope is negligible: twenty-four deviations per one hundred lines, that is,
per one thousand positions, which hardly suffices to account for the intu-
ition that Milton is the prototypical deviant poet and Pope is the prototypi-
cal regular poet. However, the divergent effect of Milton’s poetry cannot
be accounted for merely by the number of deviating stresses but, rather, by
his recourse to the marked options in a wide range of versification devices,
of which stressed syllables in weak positions is only one and a marked
caesura appears to be another. He has recourse not infrequently to such
marked forms as strings of stresses that end in weak positions, disyllab-
ics with their stressed syllables in weak positions, and last but not least,
stress maxima in weak positions.8 In Paradise Lost, book I, fifteen out of
twenty compounds have their first (strongest) stress in a weak position.
8. A stress maximum is a syllable that bears lexical stress between two syllables that
do not. “A gárden” contains a stress maximum; “a stóne gárden” does not.

21
All these marked forms are virtually nonexistent in Pope. So we should not
be surprised that with respect to the caesura as well, Milton resorts to the
marked form more frequently than most other poets.
As to the relative frequency of unmarked and marked caesurae, I have
concluded that this is an instance of a wider linguistic principle, which
Jakobson subsumes under the poetic function:
“Why do you always say Joan and Marjory, yet never Marjory and Joan? Do
you prefer Joan to her twin sister?” “Not at all, it just sounds smoother.” In a
sequence of two coordinate names, as far as no rank problems interfere, the
precedence of the shorter name suits the speaker, unaccountably for him, as a
well-ordered shape of the message. (1960: 356–357)
This principle seems to hold for the “precedence of the shorter hemistich”
as well. This, in turn, intimates an even wider linguistic principle, which
has to do with the limitations of short-term memory that constrain many
speech-processing and rhythm-processing activities. Paraphrasing Cooper
and Ross (1975: 92), members that are easier to process (in this case, which
are shorter) tend to occupy the first place(s), enabling the listener to han-
dle the preliminary processing of this member while new information is
still being presented by the speaker. Geoffrey Leech formulated this as fol-
lows: “There is a general tendency for the weight of syntactic structure to
occur later rather than earlier in the sentence, so as to avoid strain on a per-
son’s short-term memory in the course of constructing and interpreting
sentences” (1974: 197). This principle can also be formulated as “the item
hardest to process comes last” and can account for a wide range of phonetic
preferences, including for such expressions as sing-song and ding-dong to
song-sing and dong-ding, of trick or treat to treat or trick, of walkie-talkie to
talkie-walkie, and so on (Cooper and Ross 1975; Tsur 1992a: 84–86).
In the traditional view, we are used to these expressions, and this is why
their reversal sounds odd. Here, I claim that there are certain general prin-
ciples (all derived from the limitations of short-term memory) that govern
these preferences in a variety of languages, in unforeseen combinations
as well. Consider, for instance, the title of Joseph Haim Brenner’s Hebrew
novel Shkhol vǝkhishalon (Bereavement and Failure). Shkhol is shorter
than kishalon, and, accordingly, it precedes it. Hillel Halkin, in his English
translation, reversed the nouns: Failure and Bereavement, establishing the
well-ordered shape. The following anecdote illustrates how intuition works
in such cases. A few years ago I met an Israeli poet in the corridor of a pub-
lishing house. “What are you doing here?” he asked. “Publishing a vol-
ume of poetry translations.” “What’s the title?” “The Cuckoo and the Flute.”

3 1
“It doesn’t sound good. I don’t know why, but The Flute and the Cuckoo would
sound better. It’s your job to find the explanation.” The director general of
the publishing house, himself a professor of literature, called out from his
office: “Joan and Marjory!”
Far from being used to this order of the two nouns, “the precedence
of the shorter name suits the speaker, unaccountably for him, as a
well-ordered shape of the message.” These two principles are certainly not
mutually exclusive. Rather, the existence of “conventional” pairs of nouns
is an additional illustration of the central claim of this book that rigid for-
mulae reflect the freezing of cognitive processes into “style.”
Cooper and Meyer (1960: 61) found that this principle prevails out-
side verbal communication, in music. Likewise, according to Woodrow’s
experiments, which date to the 1920s (Chatman 1965: 26–27; Meyer
1956: 106–107; Tsur 1977b: 88–89; and chapter 9 below), nonlinguistic
sound stimuli are perceptually grouped into end-accented groups when
there are differences of duration and into beginning-accented groups when
there are differences of amplitude. Differences of pitch do not affect the
direction of grouping.9 Among a variety of explanations, Cooper and Meyer
(1960: 61) referred to this finding to account for their own discovery that
musical sequences tend to divide into a shorter followed by a longer musi-
cal phrase.
Prosodists have pointed out for years that the longest verse line that has
no obligatory caesura is ten syllables long. The received view until recently
was that it is difficult to pronounce a longer verse line in one breath. The
truth is that even twenty syllables can be pronounced in one breath. In
fact, there are two different perceptual reasons that account for a caesura at
midline. One is the limitation on short-term memory, which, according to
George Miller (1956), is the “magical number” of 7 ± 2 monosyllabic words.
The other one is the Gestalt rule that the simpler the parts, the more they
tend to stand out as independent entities, such as, for instance, in sym-
metric wholes that have two equal parts. Iambic lines of eight and twelve
syllables can be divided into two segments of equal length and equal struc-
ture. Therefore the tetrameter, though well within the 7 ± 2 limit, forcefully
9. Woodrow’s experimental findings seem to be consistent with some of my other

cognitive assumptions. That durational differences tend to give rise to end-accented
rhythms can be interpreted as “whenever durational differences arise, the longest
segment comes last.” His finding that differences in amplitude tend to give rise to
beginning-accented rhythms can be explained by assuming, first, that, other things
being equal, greater amplitude at the beginning of a sequence more efficiently elicits
alertness on the part of the perceiving system than at the end and, second, that greater
amplitude does not constitute a greater load on short-term memory.

41
demands a break after the fourth syllable, whereas the hexameter both
exceeds the limit and divides into two segments of equal length and equal
structure.
To sum up this part of the controversy, the “culture-begets-culture” view
cannot explain how cultural models arise or why these models, once cre-
ated, become models to be imitated. It can merely refer us to ever-earlier
occurrences of the model. The cognitive approach, by contrast, can account
for the need for the caesura in the first place and the need to prefer the 4 +
6 segmentation to the 6 + 4 segmentation in decasyllabic lines, irrespective
of language and versification system. Furthermore, it can be shown that
this cognitive dynamic applies to additional linguistic phenomena, as well
as to nonverbal communication in music and mechanical beats. Even if it
is assumed that “the hemistich form 4 + 6 of the English iambic pentam-
eter is derived from the pattern of the French decasyllable,” we can now,
at least, give a reasonable answer to the questions of why a French poet
would want to segment a verse line into 4 + 6 and why an English poet in
the syllabotonic versification system would ever want to adopt “the hemis-
tich form 4 + 6” from the French syllabic system, rather than be satisfied
with a random distribution of caesurae. The compelling reason is that the
English poet intuitively feels that this is the well-ordered shape of the line,
irrespective of what French poets do.
TURNING EXPRESSIVE FEATURES INTO HARMLESS ORNAMENTS
Anton Ehrenzweig (1965), in his sometimes speculative but always illumi-

nating book on the psychology of music and the visual arts, put forward
the view that to avoid the more “dangerous,” “overly” expressive elements
of an artistic work, we tend to suppress expressive features by formalizing
them: namely, we turn them into the distinctive characteristics of a style.
He elaborates on the defense mechanisms human society musters to pro-
tect itself from the excessive expressive force of artistic devices by turning
them into harmless ornaments. Ehrenzweig illustrates his contention with
examples of erotic symbolism in primitive art and various stages in the
evolution of Western music. He calls this “the secondary elaboration into
style and ornament.” Here I take up Ehrenzweig’s notion, but with a shift
in emphasis, and apply it in directions that were not necessarily originally
implied. This view is coupled with a form of “unpenetrating reading”: When
we recognize the style of an artistic work, that is, the features it has in com-
mon with other comparable works, we tend to be less discriminating, less
keenly aware of differences, the infinitesimal traits that, indeed, are the

5 1
deviations from this style; to use George Klein’s term (see 1970: 134–136),
we tend to “level out” the differences between unique works. In such cases,
an explication of a poem may be of decisive importance.
The characteristics of ballad style, for instance, are typically defined as
repetition (often merely decorative) and ellipsis (often quite easily com-
pleted). When we ask, “What does this repetition or that ellipsis contribute
to the ballad?” we are often handed the diagnosis that “this is a character-
istic of ballad style” or that “it contributes to the balladic atmosphere.”
Whenever there is a threefold repetition, one can expect to be told that “it
contributes to the popular-ballad style.” I by no means deny that repeti-
tion does contribute to the balladic character of the ballad. I would even
admit that these characteristics add a certain charm, a naïveté, to the bal-
lad. Nevertheless, this naïveté is sometimes the consequence of repressing
some utterly nonnaïve and expressive feature of the poem.
This principle can be applied to poetry and, among other poems, to the
popular ballad “Edward.” Let us take a fleeting look at its first stanza:
Why dois your brand sae drap wi bluid,

Edward, Edward,
And why so sad gang yee O?”
“O I hae killed my hauke sae guid,
Mither, mither,
O I hae killed my hauke sae guid,
And I had nae mair bot hee O.10
As Jakobson (1960, 1980) and others have noted, interjections and voca-

tives, which are heavily laden with emotions, tend to be processed by the
right hemisphere, and this may endow their repetition with even greater
expressive force. The simple repetition of a line (as in lines 1 and 3 or 5
and 7) enhances the emphasis and may be considered a redoubled voca-
tive, as in “Edward, Edward” or “Mither, mither,” as well as the addition
of the interjection “O.” All these features tend to lose their emotive force
when they occur regularly, stanza after stanza, in their preassigned places
throughout the poem. When this occurs the reader tends to stop assigning
10. “ ‘Why does your sword so drip with blood, / Edward, Edward? / Why does your
sword so drip with blood? / And why so sad are ye, O?’ / ‘O, I have killed my hawk so
good, / Mother, mother: / O I have killed my hawk so good: / And I had no more but he,
O’ ” (http://www.poemhunter.com/poem/edward-edward-a-scottish-ballad/).

61
emotive emphasis to them and instead sees them as ballad-style features.

This is what I mean by “freezing into style.” As Ehrenzweig put it, we turn
the expressive features into style. Outside the world of this poem, the use
of vocatives and interjections may elicit perceptions and beliefs concerning
the speaker’s emotions. Here, by contrast, these devices elicit perceptions
and beliefs concerning the genre of the poem rather than the charac-
ters’ emotions. In c hapter 4, this poem is discussed at length, and I show
how this rigid formula does in fact convey very sophisticated meanings.
Chapter 5 takes an extended look at even more rigid formulae in some
genres of medieval Hebrew liturgical poetry.
Ehrenzweig defined three stages in the development of artistic devices.
In Stage 1, artistic devices are perceived subliminally and affect what he
termed the “depth mind.” A good example is music where certain devices
can be subliminally inserted between notes. As these devices become more
emphatic their emotional appeal increases, so long as they do not become
consciously perceptible. If they do become somewhat perceptible, they are
considered bad taste or cheaply emotional. Ehrenzweig called this the sec-
ond stage. In the third stage, these devices turn into ornaments, with dras-
tically reduced emotional appeal. If the “inarticulate” glissandi and vibratos
of singers and the great masters of the violin are not consciously audible,
they have a strong and valued emotional appeal. When second-rate sing-
ers and violinists exaggerate them so that they become semiconsciously
audible, these devices are considered offensive or in bad taste. But in the
third stage, this offensive emotional force vanishes when the devices are
“sharpened” into fully conscious but rigid and lifeless ornaments. A possi-
ble fourth stage is when such “dead” ornaments are revived through poetic
manipulation. In chapter 4, I show how the restricted formulae of these
fossilized devices are revived again in the ballad “Edward” and exploited
as a major expressive resource. Likewise, chapter 5 discusses, among other
things, how two poems that have recourse to the same rigid poetic formula
(anadiplosis, where the last word of each line is repeated as the first word
of the next line, or concatenation) either reinforce the rigid formula or sug-
gest some mystical experience.
The transition from Ehrenzweig’s second to his third stage draws on lev-
eling and sharpening, a conspicuous cognitive device that serves a wide vari-
ety of literary purposes. When a stimulus pattern contains certain kinds of
ambiguity (such as a slight deviation from some symmetric pattern), sim-
plification and strong shape may be achieved, according to Arnheim, “by
changing a figure in which two structural patterns compete for dominance
into another that shows clear dominance of one of them” (1957: 57). This
tendency is called “sharpening.” The opposite tendency is called “leveling.”

7 1
“Leveling” attempts to minimize or even eliminate the anomalous detail.

In Figure 1.1, both a and d deviate slightly from symmetric patterns. When
such figures are presented under conditions that keep the stimulus control
weak enough to leave the observer with a margin of maneuver, two types
of reaction ensue. Some people perfect the symmetry of the model (b, e),
whereas others exaggerate the asymmetry (c, f) (Arnheim 1957: 57). These
ambiguous percepts in the form of semiconsciously audible expressive
devices “sandwiched” between the tones are sharpened into consciously
perceived, formal ornaments.
This explanation is strengthened by Gombrich’s (1954) analysis of a
related diachronic process. In some medieval religious art, gold repre-
sented the kingdom of light, suggesting (sometimes even evoking) power-
ful mystical experiences. Renaissance artists considered this to be in bad
taste: gold gives direct gratification to the senses, instead of arousing these
experiences using complex artistic structures. Later generations felt suffi-
ciently remote and secure in these artistic trends and regarded them more
favorably: they considered gold as a naïve trait of these paintings—they
turned it into a characteristic of a style. Gombrich noted a similar change in
taste with reference to our attitude toward Victorian furniture. It is reveal-
ing to compare his comments with those of Ehrenzweig on the same topic:
The main objection against the Victorian interior is that it is “vulgar,” by which is
meant that it allows a too easy, childish gratification in the gaudy and “cheap.” . . .
In design Victorian “atrocities” have acquired the distinction of being “amusing”—
which must be taken to mean that they have receded far enough into the back-
ground no longer to be dangerous. (Gombrich 1954: 20)
(a) (b) (c)
(d) (e) (f)
Figure 1.1 “Leveling” and “sharpening”: when a stimulus pattern shows a slight devia-
tion from some symmetrical pattern (as in a and d), subjects tend to restore symmetry (as
in b and e), or exaggerate the asymmetry (as in c and f), so as to avoid ambiguity. (After
Arnheim 1957.)

81
The very disgust we feel at the “cheap,” the “gaudy,” the “sloppy,” proves our
strong emotional involvement. Nor is the nature of this involvement hard to
guess. We react as if we resisted seduction. (Gombrich 1954: 20)
Ehrenzweig would certainly agree with the second paragraph; but he puts
forward an argument that is rather different from what is said in the first
paragraph:
Until recently, Victorian fashions were held almost in disgust. To us, the
Victorian drawing-room appeared to be crowded with useless furniture and
bizarre ornament. But in the recent sentimental revival of a newly discovered
Victorian “style” the grandeur and purity of the great Victorian age seemed sud-
denly to be revealed. It was as though there emerged a noble outline previously
hidden in the chaos of preposterous detail. (1953: 77)
I prefer Gombrich’s view, which relates the process to a fluctuation in

psychic distance. It can also be applied to a process in music. Sometimes
certain fleeting notes or glissandi that are not part of the harmony are sub-
liminally sandwiched between notes sung by the human voice to secure a
smooth transition. These notes may have enormous emotional effect, as
long as they remain unconscious, especially when they seem to be choking
with tears, as it were. In some styles, as in certain Neapolitan songs and in
the singing of Ashkenazi Jewish cantors, such choking sounds are amplified
and lengthened so that they become consciously audible, as if in response
to strong emotion. Members of a typical audience find such “passing notes”
particularly pleasing. The typical German opera or classical music concert-
goer, by contrast, finds this particularly offensive and in fact “schmaltz.”
These embellishments constitute a direct presentation of strong emotions,
rather than having recourse to complex artistic structures. Listeners who
are sufficiently distanced from these musical styles, even lovers of classi-
cal music and German opera, may consider such vocal mannerisms less
offensive and simply characteristic of certain styles—briefly, they turn
them into style. In other words, it is crucial to ask, “In bad taste for whom?”
People who share traditions may find such choking “passing notes” par-
ticularly pleasing, people distanced to some extent find them particularly
offensive, whereas others who are sufficiently distanced may accept them
with equanimity as the characteristic of a legitimate style. I personally feel
them to be particularly offensive, though I understand that they may belong
to a legitimate style. But Gombrich’s notion of fluctuating psychic distance
makes it possible to discuss these issues without prejudging values.

9 1
CHAPTER 2
Some Implications of D’Andrade’s

Assumptions
I n chapter 1 I quoted a passage from a lecture by Roy D’Andrade from

which I derived one of the main assumptions of the theory elaborated
in this book:

larities. (1981: 182)
In this chapter I will look in greater detail into the implications of

D’Andrade’s position. I find his view very stimulating. But at a certain
point we need to ask for even more detailed information, such as: What
are the relevant natural capacities (and constraints) of the human brain?
What is a “good fit” to those capacities and constraints? What are the prin-
ciples of “the process of repeated social transmission” that bring about
this result? In what follows, I briefly explore each part of D’Andrade’s
formulation.
In c hapter 1 I contrasted two approaches to cultural processes. One relies
on influence, migration, and culture begetting culture. The other relies on
the natural capacities and constraints of the human brain and on fossiliza-
tion processes. Both approaches assume repeated social transmission, but
with different implications.
02
According to the migration view, poetic conventions migrate from one

cultural center to another such that older cultural traditions influence later
ones and more dominant cultural traditions influence the less dominant
ones. The conventions of Petrarch’s love poetry were imported in the six-
teenth century to England, France, and many other European countries. In
the seventeenth century, poets such as John Donne embraced Petrarchan
conventions only to deviate from them, thus generating an ironic effect.
The logical question is where Petrarch took these conventions from. It is not
difficult to find the answer: from Provençal troubadours. And from where
did the troubadours get these conventions? Here the answer is less readily
apparent. But most likely they got them from tenth- or eleventh-century
Muslim and Hebrew poets in Spain. And where did these tenth- and
eleventh-century poets get these conventions? Most probably from early
Arab desert poetry. The inquisitive soul asking these questions should be
satiated by now. Nevertheless, we have not solved the mystery but merely
transferred it from one place to another. By the same token, we have cre-
ated two new problems: How did the first audience of some convention
know how to respond to it, and how did certain linguistic devices used in
early poetry become conventions?
By contrast, I assume that poetry involves a creative rather than a con-
ditioning process and that poetic conventions originate in cognitive and
depth-psychological processes acquired for adaptation or coping purposes.
We all are familiar with the wicked stepmother convention in fairy tales.
According to Freud, it results from a coping mechanism, designed to cope
with distressing ambivalent emotions toward the mother by splitting—in
dreams, folktales, and fairy tales—the mother figure into a good mother
and a wicked stepmother, or witch, or alternatively into a grandmother and
a wolf disguised as a grandmother. In this sense, the initial coping or adap-
tive responses may be “knowledge untaught and unlearned,” and they may
have a tolerably good fit to the natural constraints of the human brain; that
initial response is further improved through repeated social transmission.1
In this section I draw on the example of one such convention: the cata-
log of paradoxes that spread from Petrarch’s poetry to France (e.g., Villon,
Charles d’Orléans, Ronsard) and England (e.g., Wyatt, Drayton). According
to the migration approach, this convention originated in Provençal riddles,
1. Note that in one sense the two approaches differ with respect to what they take for
granted: the influence-hunting approach takes for granted the existence of a full-blown
convention as its starting point, whereas the cognitive-fossils theory takes for granted
the existence of responses acquired for adaptation purposes.

21
whose solution was “love.” I start with the Freudian notion of an ambiva-
lent attitude toward the beloved. I do not claim that all love is ambivalent
or that there are no love poems that are not ambivalent. But some forms of
love are certainly ambivalent. Or to take one of the greatest literary mas-
terpieces, according to Ernest Jones (1949), Freud’s colleague and biogra-
pher, Hamlet’s inexplicable behavior is related in complex ways to splitting
the father figure into a chivalrous loving father and a “smiling damned
villain” (King Claudius). A straightforward wicked stepmother seems to
be less threatening than an ambivalent emotion. Such a response is thus
“untaught and unlearned.”
At this stage, if Freud is right in his supposition, the psychological pro-
cess is fossilized into a thematic convention. In poetry, ambivalent attitudes
can be sharpened into paradoxical statements that are less threatening
than ambivalent emotions but still hard enough to take. A catalog of para-
doxes may mitigate their disquieting effect. Here, the same psychological
process is fossilized into a structural convention. At this point, an additional
question arises: Why is it so obvious to Provençal riddle-makers that a cat-
alog of contradictions suggests “love”? Obviously, it is a convention. But
then we are back to the question of where conventions come from and, in
this case, to the answer that they arise from sharpening ambivalent atti-
tudes into contradictions.
According to the cognitive-constraints conception, we need not assume
an infinite recursion of influences. It is easy to grasp why the first poet
invented some verbal device, why the first audience could immediately
appreciate it, and how it could become a convention (through repeated
social transmission). In the next sections I briefly explore two crucial cog-
nitive constraints and their literary implications: (1) The cognitive system
has a limited channel capacity and therefore demands great parsimony;
(2) language is typically logical and conceptual, whereas some poetry at
least is supposed to convey emotional or mystical qualities that are non-
conceptual and illogical.
PARSIMONY
According to the “limited-channel-capacity” hypothesis (see, e.g., Neisser

1968), there is a rigid upper limit on the amount of information that an
organism can process at any given time. The limited channel capacity of
our cognitive system forces us to conserve mental processing space in all
our mental activities. The key term is parsimony for both long-term and
S o m e I m p l i c at i o n s of D ’A n dr a de ’ s A s s u m p t i o n s [ 21 ]
2
immediate mental activities both in theoretical models and in terms of per-

ception and language processing.
One of the fundamental principles in theology, philosophy, and scien-
tific thinking is “Occam’s razor,” that is, the notion that “entities should
not be multiplied beyond necessity” (entia non sunt multiplicanda praeter
necessitate). One example of the way in which Occam’s principle is active in
the history of ideas cuts right across theology and physics. Kenneth Burke
showed how “the monotheistic concept of an all-inclusive God was itself
an ambiguous preparation for naturalism”: “For if nature was deemed, as
it was by many of the devout, to be a perfect exemplification of God’s will,
then nature’s design would accurately represent the design of God. Hence,
reference to God as the locus of motives would involve an unnecessary
duplication of terms” (1962: 80). In this way, the Christian medieval world-
view evolved into a modern scientific worldview.
According to George Miller (1956), the capacity of our short-term
memory is limited to “the magical number seven, plus or minus two.” This
has far-reaching implications for the way we process language, on the one
hand, and rhythm in music and poetry, on the other. For instance, scholars
of versification have long observed that the longest verse line that has no
compulsory break is ten syllables long. This can be accounted for by the
magical number seven, plus or minus two. In Greek and Latin versification
this perceptual need became a formal convention, requiring a break called
a caesura in the middle of verse lines. This convention was handed down
to later corpora of poetry and versification systems. Thus, the caesura is a
fossilized perceptual need.
Gestalt theory is also governed by the Law of Parsimony. The fundamen-
tal law of perception, the Law of Prägnanz, or “Pithiness,” is commonly
defined by Gestalt psychologists as follows: “The psychological organiza-
tion of any stimulus pattern will always be as good as the prevailing con-
ditions allow” (Koffka 1935: 110). This does not imply that it will always
be very good, only as good as the prevailing conditions allow: “The effect
depends on the degree of simplicity of the whole as compared with the
degree of simplicity of the parts. Greater simplicity of the whole makes for
greater unity. The simpler the parts, the more clearly they tend to stand
out as independent entities” (Arnheim 1957: 61); “the general rule is that
to the extent that stimuli possess similar features they form groups and are
perceived as unified, coherent, and stable structures” (Herrnstein Smith
1968: 41).
I have already touched upon the problem of caesurae in c hapter 1. Here
I recapitulate some of my arguments in a Gestaltist perspective. Consider
the iambic tetrameter and hexameter. Both can be divided into two

3 2
symmetrical halves, that is, with an identical length and an identical struc-
ture. They can be divided into segments of 4 + 4 and 6 + 6 syllables, respec-
tively. Since the iambic meter consists of regularly alternating weak and
strong positions, both segments in both meters begin with a weak position
and end with a strong position. Since the two segments of the line possess
similar features, they tend to stand out as relatively independent entities,
enforcing a break in the precise middle. Since the hexameter is longer than
ten syllables, it has an additional reason to enforce a caesura (the magical
number 7 ± 2).
But the “prevailing conditions” do not allow for such simplicity in all
iambic lines. As suggested above, their psychological organization will be
only as good as the prevailing conditions allow. An iambic pentameter line
can be divided into 5 + 5 syllables, but in that case the two segments will
have different structures: the first segment will begin and end with a weak
position, and the second, with a strong position. It is possible to obtain seg-
ments that begin with a weak position and end with a strong position if we
divide the line into segments of unequal length, 4 + 6 or 6 + 4. Obviously,
we cannot have two segments of equal length and equal structure, so they
stand out less clearly, and the whole line is more integrated. Furthermore,
there is no single point of caesura but an area of caesura after the fourth,
fifth, or sixth position.
RELATIVE DURATION AND PREVALENCE
The foregoing model itself is parsimonious: it reaches down to increasingly

subtle issues in versification. In this section I briefly explore how these lev-
els of versification are also shaped and constrained by the limited channel
capacity of the cognitive system.
From what has been said so far, one might expect that it makes little
difference whether the caesura occurs after the fourth or the sixth posi-
tion. Nevertheless, in English, Hebrew, and Hungarian iambic pentameter
lines, the overwhelming majority of caesurae between segments of unequal
length occur after the fourth rather than the sixth position. Outside of
syllabotonic verse, in French syllabic verse, in decasyllabic lines, the over-
whelming majority of caesurae occur after the fourth position. According
to Jakobson (1960), in the decasyllabic lines of the Serbian folk epic there
is an obligatory caesura after the fourth position. To account for this per-
vasive phenomenon, one must restructure the problem. Attention must be
shifted from the placement of the boundary called “caesura” to the relative
length of the resulting segments; that is, the longer segment comes last.
42
Put this way, the principle applies to parallel simple phrases as well. As
Jakobson pointed out, “Joan and Marjory” sounds somehow better than
“Marjory and Joan.” Outside linguistic structures, according to Cooper
and Meyer (1960: 61), where parallel musical phrases occur, the longest
comes last. This is due to the constraints of the cognitive system. In all
these instances, the longest segment comes last so as to avoid jamming
short-term memory—which has a limited channel capacity—while infor-
mation is still coming in.
On the foot level as well there is a tendency to prefer feet in which the
longest constituent comes last. Tenth-century Hebrew poets in Spain
imported a complicated metrical system from contemporary Arab poets
based on short and long syllables and adjusted it to the constraints of
Hebrew. This was a deliberate, conscious act. In this system there is an
enormous number of possible meters. However, in a very short time, for
mysterious reasons, two of these emerged as the most frequent meters
(I claim that the reasons are not so mysterious, since they can be accounted
for by Gestalt theory). Of these two meters, one occurs several times more
frequently than the other. The longest-comes-last principle can easily
explain this pattern of occurrence as well. Both meters are based on feet
that consist of four units: a schwa mobile and three full vowels.2 The schwa
is considerably shorter than the full vowels. In the more frequent meter,
the schwa occurs in the first half of the foot, at the very beginning, so that
the longer part comes last. In the less frequent meter it occurs in the sec-
ond half, in the second to last position, so that the longer part comes first
(as would be in English “about painting” [əbaʊt peɪntɪŋ] and “this cutlery”
[ðɪs kʌtləri], respectively—the issue at stake being not the sequence of
stresses but the placement of the reduced vowel ə among the full vowels
and diphthongs). There is no trace of this preference in poets’ explicit poet-
ics; it is all intuitive (Tsur and Bentov 1996). There was a similar develop-
ment, mutatis mutandis, with the overwhelming dominance of the iambic
meter (where the longer syllable of the foot also comes last) in a number of
Western languages.
At this point I must acknowledge three difficulties and attempt to
account for them within this model. First, in Hungarian iambic pentameter
lines, when we compare caesurae after the fourth and sixth positions, the
former do, indeed, considerably outnumber the latter; however, the most
frequent placement of caesurae (unlike in English and Hebrew poetry)
is after the fifth position. This seems to be due to the constraints of the
2. Schwa mobile is a “reduced” vowel in Semitic languages, similar to the first vowel in
English amid (əmɪd).

5 2
Hungarian language: a much greater proportion of words have their stress

on their penultimate syllable in Hungarian than in English or Hebrew.
Moreover, a closer look shows that even the 5 + 5 iambic line complies with
the longest-comes-last principle: the first segment contains two strong and
three weak positions, and the second segment contains two weak and three
strong positions. Since strong positions are typically occupied by longer
syllables than weak positions (see below), the second segment is typically
somewhat longer.
Second, in French syllabic poetry and the Serbian epic there are no alter-
nating weak and strong positions, so a crucial part of the previous anal-
ysis cannot explain the preference for the 4 + 6 to the 6 + 4 division in
these poetic texts. Such examples support the assumption that the basic
dynamic exists on the segmental level, whereas the alternating weak and
strong positions, if present, greatly enhance it. Arnheim’s observation cited
above—that the simpler the parts, the more clearly they tend to stand out
as independent entities—may provide a clue to this mystery as well. On
the whole, he says, “the simplicity of any part must be modified or weak-
ened sufficiently to make the part dependent on, and therefore integrated
with, its context” (1957: 65). Segmentation into unequal parts would count
as weakening. In other words, unequal segmentation renders the decasyl-
labic line more integrated than a symmetrical segmentation, in which the
similar parts tend to stand out at the expense of the whole. And when the
longer segment comes last, the decasyllabic line is more natural than when
the shorter segment comes last.3
Third, “weak and strong positions” are not the same perceptual phenom-
ena in all versification systems. They constitute abstract patterns that may
be filled with different kinds of immediately observable constituents within
different metric systems. In Greek and Latin versification, for instance,
they are filled with short and long syllables, whereas in English, Hebrew,
and Hungarian syllabotonic verse they are completed with unstressed and
stressed syllables. Thus only the names and abstract schemata of the feet
(e.g., “iambic” or “dactylic”) are identical in the two systems. This inconsis-
tency, however, is only apparent. Linguistic stress is signaled by a mixture
of phonetic cues, among which duration is one of the most effective; thus,
3. In the twelve-syllable-long, symmetrical alexandrine dominating French poetry

the demand for a caesura precisely after the sixth position is so strong that segments
of unequal length can be generated only at the cost of having the verse line fall apart.
By the nineteenth century such recurrent structures became so tiresome that Paul
Verlaine identified musicality in verse with odd-numbered syllables in a verse line: “De
la musique avant toute chose, / Et pour cela préfère l’Impair” [Music before all, / And
for this prefer odd numbers].
62
stressed syllables are typically longer than unstressed ones. This is why the
most natural sequence of positions in the feet is “the strong position comes
last” in syllabotonic verse too, even in Hungarian, where stress comes typi-
cally on the first syllable of the word.
The theory I develop here requires that a wide range of plausible and
implausible forms should arise randomly that, in the course of repeated
social transmission, assume a good fit to the natural capacities and con-
straints of the human brain. In a study of Arabic metrics, Golston and Riad
(2010) found that end-accented meters make up about 80–90 percent of
Arabic verse (they call them “iambic meters”).4 If they mean their terms in
the same way as I do, this would suggest that in classical Arab meter a simi-
lar process took place to what we have found in medieval Hebrew poetry.
One of the most dramatic pieces of evidence for the cognitive-constraints
approach, where cognitive constraints override even the stress rules of a
language, is to be found in the continued existence, even dominance, of
iambic lines in Hungarian poetry. “Influence” could at best ensure that the
iambic would continue to exist in Hungarian but not that it would be felt to
be the most natural meter (I discuss this question in chapter 9). My argu-
ment is mainly based on two sets of experiments. One set is H. Woodrow’s
tick-tack experiments back in the 1920s, which found that in a series of
tick-tacks, “with equal temporal spacing, a regularly recurring, relatively
greater intensity exerts a group-beginning effect, and a regularly recurring,
relatively greater duration a group-ending effect” (1951: 1233): “Intensity
has a group-beginning effect: duration, a group-ending effect: pitch, nei-
ther a group-ending nor a group-beginning effect” (1911: 77). In L. B.
Meyer’s paraphrase, “Durational differences tend to result in ‘end-accented
rhythms,’ and intensity differences tend to result in ‘beginning-accented
rhythms.’ . . . Pitch has neither group-beginning nor group-ending effect”
(1956: 106–107). A recent experiment (with more advanced technical facil-
ities) indicated that variations in pitch lead to a significant shift toward
iambic groupings (Rice 1992).
The other set is D. B. Fry’s experiments on stress perception, which dem-
onstrated that linguistic stress is signaled by a mixture of acoustic cues
including pitch, duration, and amplitude, in decreasing order of effective-
ness. If pitch differences are irrelevant to grouping direction and duration
differences are more effective in stress perception than amplitude differ-
ences, end-accented meters should be more natural in poetry in a variety of
languages. If variations in pitch also lead to a significant shift toward iambic
groupings, they should reinforce this effect. Significantly, even in Hungarian
4. A different presentation of their findings is available in Golston and Riad 1997.

7 2
poetry, where stress is invariably on the first syllable of the word, the iam-
bic meter is far more natural and widespread than the trochaic. The great
Hungarian poet Miklós Radnóti (1943: 167) observes in the epilogue to his
volume of poetry translations that the iambic is the most natural meter for
modern Hungarian poets.
VERSIFICATION SYSTEMS AS CULTURAL ARTIFACTS
In an interview for the Literature, Cognition and the Brain website concern-
ing the origins and implications of cognitive poetics, Beth Bradburn asked
me: “Cognitive Poetics emphasizes the cognitive processes of individual read-
ers and poets while also assuming the universality of those processes. Does
this tend to set a limit on the extent to which literary texts may be read as
cultural artifacts?” (Bradburn 1998). My answer included the following com-
ments. Cultural artifacts do not spring like Pallas Athena, fully armed, from
Zeus’s head. There is a process of development, until the convention reaches
its optimal fit to the natural capacities of the human brain. But, curiously
enough, these conventions do not become uniform, even when shaped and
constrained by the same cognitive mechanisms. D’Andrade accounted for this
process by repeated social transmission. Thus, when similar cultural programs
are found in most societies around the world, there is reason to search for psy-
chological factors that can account for these similarities. On the other hand,
one might add, if these “cultural programs” contain features that do not con-
flict with the natural capacities and constraints of the human brain, or there
are some good culture-specific (e.g., religious)5 reasons to preserve them, huge
differences or even conflicting patterns may evolve between the various cul-
tures on more concrete levels.
5. An illuminating example of such religious reasons was put forward by John
Skoyles (1997). In ancient times there seem to have been good physical and perhaps
neurological and cognitive reasons for writing from right to left. When vowels were
inserted between the consonants in Greek writing, the cognitive reading process is
hypothesized to have changed as well. Skoyles quotes a wide range of cognitive experi-
ments that support this. Current models of reading suggest that adults use several
kinds of processes involving at least two different routes by which words can be identi-
fied and pronounced. For instance, there are separate routes for reading whole lexical
items (in, e.g., Semitic languages, where only the consonants are written) and reading
words assembled from their consonants and vowels. In Hebrew, for example, the con-
sonant sequence spr can be identified as a unit as sappar (barber), sepher (book), saphar
(counted), or sphar (border). By contrast, the words sappar, sepher, saphar, and sphar
written in Latin characters can be assembled from their parts. Thus, the insertion of
vowels into writing may have entailed the gradual change of writing direction first
into bidirectional (that is, one line from right to left, the next line from left to right)
82
My work on poetic meter in two different metrical systems belonging

to different cultural systems helps illustrate how the cognitive approach
may shed some light on the way poetic conventions come into being (see,
e.g., Tsur 2002a). The metric system of pegs and chords,6 for example, was
introduced into Hebrew poetry by the conscious effort of Dunash Ben
Labrat, and its reception process involved violent ideological and other
conflicts (Schirmann 1995: 132–139). But the overwhelming dominance of
one metrical structure within this system was the result of an uninten-
tional process of natural selection: the metrical pattern that had the best
fit to the natural constraints of cognitive economy was the one that had a
better chance to prevail and multiply.
Versification systems are cultural artifacts. The differences between them
can be accounted for by the interaction of three factors: (1) the attempt
to import foreign metrical models of authoritative status—various cul-
tural systems import different models, for example, English, French, and
Hungarian poetry imported classical Greek and Latin models, and medi-
eval Hebrew poetry imported classical Arabic models; (2) the constraints
of the importing languages; (3) the pressure of the aesthetic demands for
unity and complexity. Increases in the number of versification constraints
lead to increases in both the complexity and the unity of the verse lines.
In the syllabotonic system, for instance, there is a greater number of con-
straints than in either the syllabic or the tonic system; consequently, it dis-
plays greater complexity and greater unity at the same time.
In French, which is a syllable-timed language, attempts to import quan-
titative meters from classical literature were a notorious failure. According
to Schirmann, in Medieval Hebrew there is no distinction between long
and short syllables; therefore, Arabic quantitative meters based on long
and short syllables could not be imported without modifying their princi-
ples. According to the Arabic system, all metrical patterns consisted of long
and short speech units (roughly as in classical Greek and Latin poetry), but
such a distinction was unknown in Hebrew. To solve this problem, Dunash
decided that any consonant with a schwa mobile or one of its allophones
and then into left-to-right writing, finally reaching, in terms of the present book, the
best fit of the changed cognitive route to the natural capacities and constraints of the
human brain. In Hebrew and Arabic these changes did not take place, because of the
prohibition against altering the sacred texts of the Jews and Muslims. These religious
works are considered by believers to be the word of God and thus unalterable. For a
more recent neurological presentation of the two routes of reading, see Dehaene 2009.
6. In medieval Hebrew poetry in Spain, the meter, imported and adjusted from
contemporary Arab poetry, was based on full vowels (termed “chords”) alternating
in all sorts of combinations with units consisting of a schwa mobile plus a full vowel
(termed “pegs”).

9 2
should be considered short, whereas all the consonants with vowels proper
should be considered long (Schirmann 1995: 120). In this way, a radically
new system was invented. This system would be inconceivable in most (and
perhaps all) European languages but is germane to classical Hebrew. Thus,
the adoption of foreign models is quite conspicuously constrained by the
properties of the adopting languages.
Syllabic meter would be well suited to any language I know of, and in
quite a few of them there have been relatively short periods in which the
syllabic system was prevalent. Downright dominance of the syllabic ver-
sification system for extended periods, however, is surprisingly rare in
European languages and is found mainly in the (syllable-timed) Romance
languages. We may account for the relative frequency of meters other than
syllabic with the aesthetic demand for unity and complexity. These two
demands constrain each other and seem to exert some pressure on versifi-
cation systems to display greater complexity and unity than mere syllabic
meter. In French, where attempts to accommodate other versification sys-
tems failed, complexity was imposed by means compatible with the syl-
labic system: in the French alexandrine there is an obligatory caesura after
the sixth syllable and a systematic alternation of feminine and masculine
rhymes. This rhyming convention was also restricted by the constraints of
the French language. In English, German, Hebrew, Hungarian, and some
other languages, in some words the main stress is on the last syllable, and
in some words, on the penultimate or antepenultimate syllable. In French,
by contrast, there are no such contrasts of stress. The only way in which
French rhymes can be systematically varied is to contrast words that end
with an e muet (feminine rhyme) and those that do not (masculine rhyme).
This increases the complexity of the rhyme pattern; but at the same time
it groups together a masculine- and a feminine-ending line, thus dividing
a quatrain into two symmetrical halves. Rhymed words ending with an e
muet are not the same as rhymed words stressed on their second to last syl-
lable in some other European languages but are the nearest options avail-
able in French. They are, therefore, usually perceived as equivalent.
When, in Dante’s time, the pegs-and-chords meter was imported from
Spain to Italy, the Hebrew poets were exposed to Italian syllabic meter,
and their pegs-and-chords meter began to resemble the Italian syllabic
meter. Characteristically, in the Hebrew sonnets by Emanuel of Rome some
present-day scholars designate the meter as hendecasyllabic, whereas oth-
ers define it as belonging to the pegs-and-chords system. What appears to
be quite surprising is that this newborn syllabic meter gradually assumed
the characteristics of the syllabotonic iambic. It was the great Hebrew poet
and scholar the late Dan Pagis (1993) who described this development in
03
Hebrew versification in Italy in great detail. Obviously, the syllabotonic

pattern conformed better to the constraints of the Hebrew language than
to those of Italian, and thus the Hebrew versification system gradually
yielded to the pressure of aesthetic demands for unity and complexity. This
interaction of foreign models, the constraints of importing languages, and
the pressure of aesthetic demands for unity and complexity yielded the
culture-specific and language-specific metric conventions in which inter-
cultural principles are individuated.
Via the process of repeated social transmission, these intercultural prin-
ciples took forms that have a good fit to the natural capacities of the human
brain. As discussed above, the two most conspicuous natural capacities of
the human brain are the limitations and capacities of short-term memory
and the Gestalt rules of perception. The Gestalt rules of perception refer to
the conditions that maximize our tendency to perceive a stimulus pattern as
an integrated whole. The better, or simpler, the Gestalt of a stimulus pattern,
the less mental processing space it occupies, and therefore, the more likely
it is to be contained within the scope of short-term memory and perceived
as rhythmical. This Gestalt principle affects the placement of caesurae both
in syllabic and in syllabotonic verse and also accounts for the overwhelming
prevalence of the hammǝrubbɛ and the haššalem meters in medieval Hebrew
poetry. As to the limitations of short-term memory, they are among the
stringencies that require the application of Gestalt principles in rhythmic
organization and are responsible for the overwhelming prevalence of the
longest-unit-comes-last principle in a variety of languages, in versification
systems, and outside verbal communication, in music.
SCHEMATA
When the information to be processed exceeds the limit of its channel capac-
ity, the organism can have recourse to a variety of cognitive strategies and
devices. One possible way of handling excess information is to recode it in
a more efficient manner, so as to require less processing space.7 Suppose
you are asked to memorize the following series of numbers and then recall
them one week later: 5 8 1 2 1 5 1 9 2 2 2 6. In George Katona’s experiment
(reported by Roger Brown [1958]), none of the subjects could recall them.
Now, suppose you are to learn the same numbers along with their principle
7. In my work on the rhythmical performance of poetry (Tsur 2012a) I show that
in case of metrical deviation, the perceptual solution consists of recoding the sound
material into better gestalts, so as to keep it within the capacity of immediate memory
(see chapter 9).

31
5 8 12 15 19 22 26
3 4 3 4 3 4
Figure 2.1 A series of apparently random numbers is difficult to memorize. The same num-
bers are better memorized when their principle of organization is given prominence, allevi-
ating cognitive overload.
of organization: the difference between 5 and 8 is 3, between 8 and 12 is 4,

between 12 and 15 is 3, between 15 and 19 is 4, and so on. Suppose this
principle is demonstrated to you by showing the numbers as in Figure 2.1:
“Subjects learning the series in this fashion could grasp it very quickly and,
indeed, could quickly learn a much longer series organized according to
the same principle. . . . Katona, furthermore, found that this second group
retained the series after an interim of a week” (Brown 1958: 72). It has been
objected that the experiment relies on reconstruction rather than remem-
bering. This is true enough, but as we shall see, Bartlett (1932), for instance,
assumed that our memory, as well as our perception, is essentially recon-
structive rather than reproductive.
In terms of parsimony, a huge number of unrelated items is reduced to a
single system, which can therefore easily be remembered. These organiza-
tional principles may take the form of a category or a schema (the former
is based on the shared features of diverse entities; the latter, on an orga-
nized pattern of thought or behavior that organizes categories of informa-
tion and the relationships among them; David Rumelhart in psychology
[e.g., 1979] and Roger Schank in artificial intelligence [e.g., Schank and
Abelson 1977] used schema theory to account for our understanding of
narrative and stories). For present purposes the difference between them
is of little importance, and here I will discuss only schemata.
Schemata are instrumental in remembering, perceiving, inferring, and
many other cognitive activities (Bartlett 1932); they affect parsimony in two
different ways at least. First, they help to efficiently organize information in
memory since information for which we have no schemata is easily forgot-
ten. Second, our everyday speech is riddled with gaps, implied meanings,
and ellipses that render speech more parsimonious. Schemata may help us
to infer what has been suppressed. Relying on a schema, an artificial intel-
ligence program may infer, for instance, from the sentence “John went to
the movies yesterday” that John probably also saw a movie, not merely
went there. Or relying on a different schema, from a story such as “John
was sitting in a fancy restaurant; he gave a large tip and left,” the machine
may infer that John probably had been served food and that he paid before
23
leaving. The brain’s need for parsimony requires us to develop schemata for
remembering large quantities of information; once developed, they allow us
to infer suppressed information, which, in turn, allows further parsimony
in communication. In this sense, schemata are constrained by the natu-
ral capacities of the brain. In different cultures, different schemata can be
developed by the same constraints. The same constraints of the brain may
further shape the different schemata in different cultures.
We have an enormously rich lexicon stored in our long-term memory, in
a way that we can use it with amazing flexibility to communicate about all
possible topics in the world. Linguists and psychologists disagree as to how
these meanings are coded in our memory, but all agree that we have access
to them with amazing ease (for disruptions of this process, see chapter 10).
Suppose we adopt one of the semantic-primitives models, or one of the
semantic-networks models, or a combination thereof, to account for our
versatility in using words. Would it be sufficient to account for our ability
to communicate with words?
Today it is absolutely clear that a model of semantic representation would
be insufficient and that we also need a model for the representation of world
knowledge. In his early work on artificial intelligence, Roger Schank (1973)
included all the information required in the definition of words in a system
called “conceptual dependency.” He gradually discovered that for effective
language processing, he also needed such wider notions as scripts, plans, and
goals. One of the prevalent models is what Bartlett (1932) and Norman and
Rumelhart (1975) called “schemata” and Schank and Abelson (1977) called
“scripts.” I will give two brief examples to indicate how scripts and schemata
help us establish implied relationships between events and make infer-
ences as to what has not been explicitly stated in an utterance. Rumelhart
(1979: 85–86) illustrated the limitations of the semantic approach of com-
ponential analysis with the following sentence frame: “X raised his hand and
stopped the car.” If we substitute the phrase “the policeman” for X, the most
likely account for this event according to Rumelhart is
one involving a traffic cop who is signaling to a driver to stop his car. Note that
this brings under consideration a number of concepts that are not mentioned
in the sentence itself. For example, this interpretation requires that the car has
a driver and that the policeman stopped the car by signaling with his hand to
the driver, who then most likely puts his foot on the brake of the car causing it
finally to halt. (1979: 85–86)
If, however, we substitute “Superman” for X, we have to assume that the

car was stopped with no intentional cooperation on the driver’s part or that

3
there is no driver at all. There is no possibility to account for this difference

between the two sentences by relying on the different semantic makeup of
the two nouns. Rather, we need to assume that the phrase “the policeman”
instantiates the policeman script, whereas the word Superman instanti-
ates the actions regularly associated with superman.
Even the computer in the artificial intelligence mode can make infer-
ences that change with the context in this way. Consider a story such as
“John was sitting in a hotel room in Paris. He was exhausted and hungry.
From his bag he took out X.” If you substitute “a sandwich” for X, the com-
puter will infer that John was going to eat it. If you substitute “the Michelin
Guide,” the computer (as well as a human reader) will not infer that John
was going to eat it but, rather, that John was making a plan to go to a
restaurant (cf. Schank and Abelson 1977). The different objects instantiate
two different scripts.
Such schemata were initially developed for nonliterary purposes such as
adaptation to effective communication and the handling of extralinguistic
life situations. In literary narratives, schemata are deployed to aesthetic
ends. Gaps are more frequent and more sophisticated, and the reader/lis-
tener must make more sophisticated inferences than the computer, some-
times applying alternative or even conflicting schemata.
In the Eskimo story called “The War of the Ghosts” below, something appar-
ently went wrong, and no schema is readily available to the Western reader:
One night two young men from Egulac went down to the river to hunt seals,
and while they were there it became foggy and calm. Then they heard war cries
and they thought; “Maybe this is a war-party.” They escaped to the shore, and
hid behind a log. Now canoes came up, and they heard the noise of paddles and
saw one canoe coming up to them. There were five men in the canoe and they
said; “What do you think? We wish to take you along. We are going up the river
to make war on the people.”
One of the young men said; “I have no arrows.”
“Arrows are in the canoe,” they said.
“I will not go along. I might be killed. My relatives do not know where I have
gone. But you,” he said, turning to the other, “may go with them.”
So one of the young men went, but the other returned home. And the war-
riors went on up the river to a town on the other side of Kalama. The people
came down to the water and began to fight, and many were killed. But presently,
one of the young men heard one of the warriors say; “Quick let us go home. That
Indian has been hit.”
Now he thought; “Oh, they are ghosts.” He did not feel sick, but he had been
shot. So the canoes went back to Egulac, and the young man went back to his
43
house and made a fire. And he told everybody and said; “Behold, I accompanied
the ghosts, and we went to fight. Many of our fellows were killed and many of
those that attacked us were killed. They said I was hit, but I did not feel sick.” He
told it all, and then he became quiet. When the sun rose, he fell down. Something
black came out of his mouth. His face became contorted. The people jumped up
and cried. He was dead. (Bartlett 1920)
Which schema enables the reader to infer “Oh, they are ghosts” from the
observation “He did not feel sick, but he had been shot”? Similarly at the
end of the story, we have no schemata to account for “They said I was hit,
but I did not feel sick” and “Something black came out of his mouth.” But
for the original audience of the story this is precisely to be expected when
you are killed in a war against ghosts.
REPEATED SOCIAL TRANSMISSION
How can repeated social transmission cause a text or a verbal device to

take forms that have a good fit to the constraints of the cognitive system?
Bartlett (1920, 1932) devised a controlled experiment using the Eskimo
story to explore how schemata affect our memory and the reproduction of
folktales. He assumed that our memories, as well as our perceptions, are
essentially reconstructive rather than reproductive; that is, we recreate the
meaning of a memory or a percept by merging elements of what actually
occurred with knowledge from our existing schemata. We fit information
into our already existing schema. Bartlett tried to determine how a story for
which we have no mental schemata assumes, through repeated transmis-
sion, a good fit to the mental schemata we do entertain. He staged a game of
“Chinese whispers” or “broken telephone” where he had experimental sub-
jects read “The War of the Ghosts” and then, after some time (ranging from
several hours to several weeks), write it down from memory. These texts
were then read by other subjects, who wrote it down after some time, and so
forth. He found that the story changed from one transmission to the next
until it achieved a good fit to a mental schema prevalent in Western culture;
it then stopped changing in subsequent transmissions.
It is impossible to reproduce the chains of transmission and Bartlett’s illu-
minating stage-by-stage commentary, so I will only give a few highlights.8
For instance, Bartlett observed that subjects, acting almost always unwittingly,
8. The entire sequence of transmissions and commentary is available online in

Bartlett 1920.

5 3
supplied connecting links, whereas in “The War of the Ghosts” events follow
one another, but their connection is not explicitly stated (cf. 1932: 86):
Now in the original narration, although it is not put forward specifically as a

reason, the casual “they are ghosts” serves as a rationalising factor through-
out the whole story. With this inserted, all the rest is satisfactory. But I have
already shown how all mention of the ghosts dropped out of my reproductions.
(1920: 37–38)
And this in spite of the fact that ghosts appear in the original title. (1920: 35)
Or, consider again, at the end of the story, the phrase “Something black
came out of his mouth.” This complies with the original audience’s schema
of being killed in a war of ghosts but is utterly meaningless to a
Western audience. The following sequence of renderings of this phrase is
telling:
“he gave a cry, and as he opened his mouth a black thing rushed from it”
“a black thing rushed out of his mouth”
“a great black thing flew out of his mouth”
“his soul fled black from his mouth”
“his soul passed out from his mouth”
“his spirit fled”
“his spirit left the world”
The vague “something black” is superseded by a more concrete “black

thing,” which in due course is entirely superseded by the idea of the pas-
sage of the soul. The neutral verb “came out” is replaced by the more inten-
sive “rushed out” (presupposing a living agent as subject), which eventually
becomes “fled” and “left the world,” which is more suitable to “soul” or
“spirit” than to the vague, unidentified, “something black.” Over the course
of repeated transmissions, Bartlett says, a further troublesome element
disappeared. In spite of his serious wound, the Indian continues to live for
a fairly long time, up to a certain point in the transmissions. However, the
Indian’s long survival is a source of concern for all of Bartlett’s subjects. In
the versions toward the end, the wounded man dies immediately: “The very
common and conventional phrase: ‘his spirit fled’ is employed, and the idea
of a material soul also disappears” (Bartlett 1920: 42). Similarly, the phrase
“When the sun rose, he fell down” turned into “At sunset his soul fled black
from his mouth,” which is perfectly consistent with Western literary con-
ventions (where people die at sunset rather than at sunrise). This was liter-
ally repeated in the next transmission.
63
Bartlett’s “Chinese whispers” experiment explores, then, how Western

subjects handle a story through repeated social transmission for which they
have no appropriate schemata. The story keeps changing until it assumes a
good fit to some Western schema; then it stops changing.
INTERJECTIONS, VOCATIVES, AND REPETITIONS
One of the underlying motifs of this book is the natural capacities and con-
straints of the human brain. But in the previous examples I have made no
explicit reference to the brain. It would have been an unnecessary duplica-
tion of terms. I referred to the “cognitive system,” “limited channel capac-
ity,” “immediate memory,” and “schemata,” all of which have been shaped
and constrained by the human brain. In this section I discuss an issue
where an explanation cannot be made without explicit reference to the
brain. This issue concerns the handling of emotive discourse with concep-
tual language:
Claudius. And now, Laertes, what’s the news with you?

You told us of some suit. What is’t, Laertes?
You cannot speak of reason to the Dane
And lose your voice. What wouldst thou beg, Laertes,
That shall not be my offer, not thy asking?
The head is not more native to the heart,
The hand more instrumental to the mouth,
Than is the throne of Denmark to thy father.
What wouldst thou have, Laertes? (Hamlet I.ii.42–50)
Granville-Barker commented that this passage is “quite affectionately gra-

cious,” “the repeated name is almost a caress” (1957: 53). How does such a
passage mobilize the emotional capabilities of the brain? What is so affec-
tionate about a repeated name? Now consider “O God! God!” in the passage
below. The interjection “O” followed by a vocative (or repeated vocative)
became a poetic convention in Elizabethan drama for expressing intense
emotion:
O that this too too solid flesh would melt,

Thaw, and resolve itself into a dew!
Or that the Everlasting had not fix’d
His canon ’gainst self-slaughter! O God! God! (Hamlet I.ii.129–132)

3
7
Again, if D’Andrade is right that cultural programs, socially transmitted

over time, take forms with a good fit to the human brain, then we should
assume that this verbal construct became a convention not merely because
it arouses certain beliefs concerning the character’s emotions but because it
mobilizes certain specifically emotional capabilities of the brain. Thus, this
verbal construct occurs at a significant junction of stylistic development. It is
still emotionally active, but on the verge of fossilizing into a convention. The
more hackneyed a device, the less emotional activity it retains. In Claudius’s
speech to Laertes, or even in Hamlet’s soliloquy, the verbal construct may
directly affect the audience’s perceptions or subjective experiences. But in
much Elizabethan drama it affects the audience’s beliefs concerning the char-
acters’ emotions instead. One of the conditions in the definition of percep-
tion (as opposed to thought processes) is that responses must not be based
on complex and abstract inferences (Lloyd 1972: 19). In the case of voca-
tives in Elizabethan drama, the key term is infer. The audience is aware of an
inference—it infers from the interjections and vocatives that the character
has (or pretends to have) certain emotions. But in the response to Claudius’s
affectionate tone, inference may also be involved, since the all-important
question is whether the audience is aware of this inference.
In Claudius’s speech the repeated vocatives appear to be emotionally
active, even if the audience is inclined to doubt his sincerity. But beyond a
certain point some form of ironic incongruity may arise between the emotive
purport of this convention and its lack of emotional effect. This is because the
interjection “O” followed by a vocative (or repeated vocative) had become so
much of a convention that Shakespeare himself parodies it in A Midsummer
Night’s Dream and thus renders the ironic incongruity ludicrous:
Pyramus. O grim-look’d night! O night with hue so black!

O night, which ever art when day is not!
O night, O night! alack, alack, alack,
I fear my Thisbe’s promise is forgot!
And thou, O wall, O sweet, O lovely wall,
That stand’st between her father’s ground and mine!
Thou wall, O wall, O sweet and lovely wall,
Show me thy chink, to blink through with mine eyne! (V.i.168–175)
What “natural capacities and constraints of the human brain” can

account for the emotive effectiveness of the repetition of Laertes’s name or
such phrases as “O God! God!”? To understand this, we need to realize that
King Claudius does not just repeat the name but uses it as a vocative. In
83
Brain and Language, Jakobson (1980) pointed out that language and speech
are typically related to the left hemisphere of the brain. However, there
are verbal elements such as interjections and exclamations that permeate
speech yet act like immediate signals. These “stand outside the general syn-
tactic patterning of language, and they are neither words nor sentences. . . .
[S]‌
emantically they are reduced to stereotyped affective expressions”
(Jakobson 1980: 23) and are typically related to the right hemisphere.9
Jakobson quotes an 1874 article by John Hughlings Jackson: “The com-
munist orator did not really make a blunder when he began his oration
‘Thank God I’m an atheist,’ for the expression Thank God is used by careless,
vulgar people simply as an interjection” (Jakobson 1980: 23). Hamlet obvi-
ously uses “God” here as an interjection (not as “the Everlasting,” which
he uses as a full content word). Proper nouns have no meaning either but
are immediate signals used to designate a particular being or thing; that is,
semantically they are reduced to tags (when you say “Herbert” you don’t
have in mind “army + bright” [the name’s Germanic roots] but use it to tag
a person). Vocatives can also act as “immediate signals” and “stand out-
side the general syntactic patterning of language” and are typically related
to the right hemisphere. What is more, repetitions and analogies typically
involve the right hemisphere. The left hemisphere of the brain, the linguis-
tic hemisphere, is generally logical. Emotions and emotional processes are
typically related to the right hemisphere. Given the logical and conceptual
nature of language, it seems to be ill-suited to convey emotional experi-
ences in poetry. One way that poets attempt to overcome this limitation
of their medium is to increase the activation of the right hemisphere by
some verbal device.10 “Laertes” in the king’s speech is not merely a name
but a repeated vocative, a verbal device that activates the right hemisphere,
which in turn bestows affective power on it. Interjections, vocatives, and
repetitions (which are typically related to the right hemisphere) are some of
the most elementary devices. By now they have lost much of their effective-
ness because they are so hackneyed—precisely because they have become a
convention. There are other more sophisticated and poetically more effec-
tive devices as well related to space perception.11
We may better understand the nature of the fossilization process if
we realize that we are not facing an all-or-nothing distinction but, rather,
9. Lateralization (= being largely under the control of one side of the brain) has
become an important topic in brain research. A highly readable introduction to lateral-
ization can be found in Ornstein 1975. A recent summary of research on lateralization
with reference to poetic language can be found in Kane 2007.
10. Another device involves divergent structures (see, e.g., Tsur 1972, 2008a).
11. See, e.g., Tsur 2008a: 285–403, 595–622; Tsur and Benari 2002.

9 3
a gradual process involving differences of fine shadings. Consider again

Granville-Barker’s language describing Claudius’s speech as “quite affec-
tionately gracious,” “the repeated name is almost a caress.” Claudius may
be hypocritical, untruthful, “a smiling damned villain,” but the speech
may be perceived as genuinely “affectionate,” “almost a caress.” In Hamlet’s
“O God! God!” by contrast, one may be aware of a trace of theatricality. In both
instances we have repeated vocatives. But whereas in the former the affective
overtones of the vocatives are imperceptibly woven into a friendly speech, in
the latter the emotive tone is all too conspicuously foregrounded by the addi-
tion of the interjection “O” and the explicit syntactic function: exclamation.
Moreover, while the repeated vocatives are scattered at intervals in the for-
mer, in the latter they are piled up as two consecutive words. Consequently,
“O God! God!” is prone to effect the audience’s beliefs about, rather than per-
ceptions of, the speaker’s emotions. If the audience perceives, in spite of all,
something like a genuine emotion in this exclamation, it is because other
parts of this soliloquy arouse the impression of some towering passion.
The correspondences between poetic effects and the topography of
the brain have little to contribute to our understanding of how literature
works. In most instances dwelling on such correspondences merely restates
in “brain language” what has already been said in “psychology language” or
“poetry language.” A reference to brain topography can be justified only
if it solves problems that one cannot solve by having recourse to “poetry
language” alone. MRI and PET studies of literary responses may have great
scientific value for neurologists, but, in most instances, their main literary
significance resides in the hope that someone sometime will make insight-
ful use of their results (see Tsur 2012b).
My example of lateralization is a fairly unsophisticated instance of point-
ing out how poets attempt to overcome the logical and conceptual nature
of language to convey emotional, nonlogical and nonconceptual qualities.
I chose to use precisely this unsophisticated example because it conspicuously
illuminates a further stage in stylistic development, represented by the ballad
“Edward” (see chapter 4). In this ballad, interjections, vocatives, and repeti-
tion have fossilized into extremely rigid formulae, but as discussed below they
acquire exceptional expressive power by virtue of their poetic structure.
INTERIM SUMMARY
I began this chapter with the suggestion that “in the process of repeated
social transmission, cultural programs come to take forms which have a
good fit to the natural capacities and constraints of the human brain.” In
04
the preceding sections I briefly explored each constituent part of this sug-
gestion. Language is logical, conceptual, and typically related to the left
hemisphere of the brain. Poetry is sometimes assumed to convey illogical,
nonconceptual, highly emotional information. Thus, language is poorly fit-
ted to convey emotional qualities, and poets must find verbal devices that
can increase the activity of the right, emotional hemisphere of the brain.
Interjections, exclamations, and vocatives are typically related to the right
hemisphere, and so are repetitions of all kinds. The repetition of Laertes’s
name by King Claudius in Hamlet may be required by his character, his
“mellifluous excess of speech,” an individual choice to achieve an affection-
ate effect; whereas in Elizabethan drama interjections followed by repeated
vocatives became a conventional means to express strong passion. Hamlet’s
“O God! God!” would be, according to John Hughlings Jackson, an interjec-
tion rather than a vocative. Thus, the conventional device has a good fit
to the emotional capacities of the brain but becomes less effective when
hackneyed: a fossilized cognitive process.
The capacities and constraints of the brain may govern a process of
“natural selection.” Medieval Hebrew poets in Spain imported a wide
range of poetic meters from contemporary Andalusian Arab poetry. They
had no explicit preference for any of them. But in the process of repeated
social transmission one meter gained overwhelming predominance, owing
to its good fit to the Gestalt rules of perception, on the one hand, and to
the limited channel capacity of the cognitive system, on the other (the
longest-comes-last principle).
Bartlett’s ingenious experiment illustrated the process of repeated social
transmission through which the Eskimo story “The War of the Ghosts” was
detached from the original schemata underlying it and then gradually came
to take a form that has a good fit to the schemata entertained in Western
culture. When it achieved that good fit, the story stopped changing in sub-
sequent transmissions. Though “a good fit to schemata” is not quite the
same as “a good fit to the natural capacities of the human brain,” Bartlett’s
experiment reproduced the mental dynamics involved. As he himself said,
“The reproductions themselves illustrate the operation of principles which
undoubtedly help to determine the direction and character of convention-
alisation as it occurs in everyday experience” (1920: 31).
TROUBLE WITH THE FOREGOING MODEL
At this point, I wish to reproduce my dialogue with an anonymous

reviewer, who had some comments of great importance on my argument

41
(originally set out in Tsur 2010) and forced me to place it in a wider con-

text. The reviewer suggested a “counterproposal” to my constraints-seeking
approach, namely, “that the human brain is extremely plastic & nearly any
cultural program is a ‘good fit.’ ” He or she was also critical of my obser-
vation that a good fit to schemata is not quite the same as a good fit to
the natural capacities of the human brain: “It’s more significant than ‘not
quite the same.’ In fact, it makes the opposite point. Both the original and
the Westernized story are presumably good fits for the brain, which sug-
gests flexibility rather than constraint.” I agree that both the original and
the Westernized story are presumably good fits for the brain and in one
respect that does indeed “suggest flexibility” in the sense that the story can
be imported from one culture to another. As will be seen, further investi-
gation has revealed that the brain’s plasticity had a much more important
part in the process, but without detracting from the importance of its con-
straints. To be more precise, the all-important question is At which point
does plasticity come in: as an alternative to constraints or in turning exist-
ing constraints to new ends?
However, I do not agree with the reviewer’s conclusion, which makes the
opposite point to mine. I claim that en route the story is changed so as to
suit the constraints of one schema rather than another. Let me put it dif-
ferently. The fact that any wild invention can be invented reflects the enor-
mous flexibility of the human brain. The process by which these inventions
fossilize into conventions reflects the brain’s constraints. The assumption
that the process can start with all sorts of bizarre, haphazard, and arbi-
trary inventions and then fossilize in conformity with the constraints of
the brain may account for the wide variety of cultural conventions all over
the world.
When I say “it’s not quite the same,” I mean that in this case the con-
straints are not related to the brain (which seems, in fact, to be compatible
with both schemata) but, rather, to the schemata. Since I cannot set up a
controlled experiment (nor have I found one set up by others) in which
“repeated social transmission” results in “good fit to the constraints of
the brain,” I have to content myself with Bartlett’s controlled experiment,
which demonstrates how repeated social transmission results in good fit to
the constraints of culture-bound schemata. The constraints may be differ-
ent, but there are good reasons to suppose that the dynamics are similar.12
12. Initially I was working with D’Andrade’s 1980 lecture. Later I discovered that he
had also published it (D’Andrade 1981). In the abstract he uses the phrase “a good fit to
the natural capacities and constraints of the human information processing system,”
suggesting that he uses it interchangeably with “a good fit to the natural capacities and
24
At this point we arrived at a stalemate. I could not provide the missing

links to account for the possibility that there are rigorous constraints of the
brain that can, nevertheless, generate an indefinite number of widely dif-
ferent cultural programs around the world. Nor could the reviewer provide
any specific details on the plasticity of the human brain. This is the point
where the humanities must turn to brain science, to seek rescue. Three
years after the publication of my original essay in Style, I ran into an article
that provided just that.
THE BRAIN’S LETTERBOX AND NEURONAL RECYCLING
It was with reference to the illuminating article by Stanislas Dehaene

(2013) that I could come the nearest to considering how “repeated social
transmission” results in “good fit to the constraints of the brain” (rather
than merely to the constraints of the cognitive system). Moreover, this
article also demonstrates, through controlled experiments, how the brain’s
plasticity contributes to the process.
In 1979–1980, during a sabbatical in New Haven, I adopted two of the
basic assumptions of cognitive poetics. From Roy D’Andrade’s plenary
talk at the Second Conference of the Cognitive Science Association at Yale
I adopted the foregoing generalization; from the Haskins Laboratories
researchers I adopted the suggestion that in evolving language, old adap-
tation devices may have been turned to linguistic ends. I applied this
conception to cognitive poetics by arguing that “in literature, adaptation
devices (including linguistic devices) are turned to aesthetic ends” (Tsur
2008a: 43). These conceptions are massively reinforced on the cerebral
level in Dehaene’s article. In this article he poses the following question:
The brain of any educated adult contains a circuit specialized for reading. But how
is this possible, given that reading is an extremely recent and highly variable cul-
tural activity? The alphabet is only about 4,000 years old, and until recently, only a
very small fraction of humanity could read. Thus, there was no time for Darwinian
evolution to shape our genome and adapt our brain networks to the particularities
of reading. How is it, then, that we all possess a specialized letterbox area? (2013)
The answer lies in a principle I have adopted from researchers at Haskins

Laboratories. Cultural inventions always involve the recycling of older
constraints of the human brain.” Fortunately, I discovered this only after I was forced
to find Dehaene’s work to support my thesis in its original formulation.

3 4
cerebral structures that originally were selected by evolution to address

very different problems but manage, more or less successfully, to shift
toward a novel cultural use. More specifically, whenever we read—whether
our language is Japanese, Hebrew, English, or Italian—each of us relies
on very similar brain networks. In particular, a small region of the visual
cortex becomes active with remarkable reproducibility in the brains of all
readers. A brief localizer scan, during which images of brain activity are
collected as a person responds to written words, faces, objects, and other
visual stimuli, serves to identify this region. Written words never fail to
activate a small region at the base of the left hemisphere, always at the
same place. Experts call this region the visual word form area; Dehaene
dubbed it “the brain’s letterbox” because it concentrates much of our visual
knowledge of letters and their configurations:
Neurons in the ventral visual pathway often respond to simple shapes, such as
those formed by the intersections of contours of objects. Even in the macaque
monkey, the inferotemporal visual cortex already contains neurons sensitive to
letter-like combinations of lines such as T, L, X, and *. The ventral visual system
seems to favor those shapes because they signal salient properties of objects that
tend to be robustly invariant to a change in viewpoint. (2013)
We are thus dealing with natural capacities and constraints of the

brain, shared not only by humans but also by other primates. The capacity
involved seems to be crucial for cognitive constancy in visual perception.
D’Andrade suggested that in the course of repeated social transmission,
cultural programs came to take shapes that have a good fit to the natural
constraints and capacities of the human brain. Dehaene not only reinforces
this conception but also shows how the same constraints of the human (in
fact, the primate) brain may shape and constrain a multiplicity of writing
systems around the world:
In all of the world’s cultures, scribes in generation after generation progressively

selected their letters and written characters to closely match the set of shapes
that were already present in the brains of all primates and, as a result, were easy
to learn. This hypothesis is corroborated by a large-scale analysis of the world’s
writing systems. Writing systems do vary in their “grain size”: the linguistic units
that are marked in writing vary from phonemes (in our alphabet) to syllables (in
Japanese Kana notation) or even entire words or morphemes (as in Chinese).
However, visually speaking, they systematically make use of the same set
of shapes, precisely those that abound in natural visual scenes and tend to be
internalized in the ventral visual cortex. (2013)
4
In his 2009 book, Dehaene provides a much more detailed account of

the whole process. He describes the common features of the world’s writing
systems as follows: “In spite of their obvious diversity, all writing systems
have numerous visual features—highly contrasted contours, an average
number of about three strokes per character, and a reduced lexicon of
shapes that constantly recur, even in unrelated cultures” (2009: 174–175).
This allows him to include Egyptian hieroglyphs, too. Egyptian hieroglyphs,
however, have no good fit to the abstract shapes in the brain’s letterbox
but, rather, to objects in the physical world, processed in areas neighbor-
ing the letterbox or in its right-hemisphere counterpart. Thus, all writing
systems (except, perhaps, Egyptian hieroglyphs) seem to rely on the set of
shapes to which our primate brain is already highly attuned—living proof
that culture is constrained by brain biology.
In an exceptionally lucid essay, Orly Goldwasser (2010) traces how the
alphabet was born from hieroglyphs. Not surprisingly, the beginnings of
writing were motivated by a mimetic conception. Egyptian hieroglyphs
sought to achieve a good fit to the world of referents. This does not mean
that the symbols necessarily resembled their referents, but they resembled
something in the physical world, as in the following text:
From Millmore (online)
This text means “No limit may be set to art, neither is there any craftsman
that is fully master of his craft” (The Instruction of Ptahhotep) (Figure 2.2).
Later scribes, by contrast, sought to achieve a good fit to the shapes preex-
isting in the brain’s letterbox, at the expense of similarity to the physical
world. Let me illustrate the process with one highly illuminating example.
Semitic alphabets adapted Egyptian hieroglyphs to write consonantal val-
ues based on the first sound of the Semitic name for the object depicted
by the hieroglyph. ʔalp or ʔālep in Phoenician and in ancient Hebrew meant
“ox.” In the Phoenician script, the first letter of the word ʔalp was derived
from the Egyptian glyph for “ox,” yielding the letter aleph.
Dehaene (2009: 185) begins this string of shapes one step earlier, with
an ox head from the famous Lascaux cave paintings (approximately eigh-
teen thousand years ago; Figure 2.3). At the beginning of the process, the

5 4
Figure 2.2 How the alphabet was born from hieroglyphs. (From Goldwasser 2010;
Millmore, http://discoveringegypt.com/egyptian-hieroglyphic-writing/.)
Figure 2.3 Ox from the Lascaux cave.
smoothly rounded Egyptian glyph, though a simplified contour, did not con-
form to the abstract shapes in the brain’s letterbox but, rather, to the shape of
the ox head out in the world. The same holds for the first alphabetic writing,
Proto-Sinaitic. But Phoenician scribes “progressively selected their letters and
written characters to closely match the set of shapes that were already pres-
ent in the brains of all primates”—to use Dehaene’s words. Subsequently, in
Greek the aleph was rotated upward, to become A. Significantly for our pur-
pose, Goldwasser observes: “The iconic meaning of the hieroglyphs was so
important that even today, when the Hebrew letters have lost all iconic con-
nection to the old pictorial models (we can’t recognize what the letters are sup-
posed to picture), most letters are still named after the old pictures!” (2010).
In terms of the present argument, this means that here, too, a typical fos-
silization process took place. At the beginning of the process the characters
had a good fit to physical reality, facilitating immediate understanding of the
64
iconic meaning of the characters, but in the course of repeated social trans-
mission they assumed a good fit to the constraints of the brain. To be sure, the
set of iconic characters too had, in a sense, a good fit to the constraints of the
human brain. As Goldwasser suggests, “People learned the letters from one
another orally. For this kind of use, the pictorial nature of the signs was very
important. It was easy to learn the alphabet simply by memorizing the pic-
tures” (2010). But for some reason, in order to gain such universal currency,
writing systems had to assume good fit to “the set of shapes that were already
present in the brains of all primates” (Dehaene 2013)—even at the price of
losing “one of its greater assets: its mnemonic power” (Goldwasser 2010).
The Phoenician script differs, then, in two crucial respects from the hiero-
glyphs: It is alphabetic, not pictographic; and its symbols have a good fit to
the primitive shapes in the primate brain, not to the physical objects out in
the world. The Proto-Sinaitic script performed only one of these two revolu-
tions: Its symbols designate single consonants, not objects in physical real-
ity; but they still retained the shapes of the physical objects. However, once
the icons were used to designate the first sound of the Semitic name for the
object depicted by the hieroglyph and not the object itself, that cleared the
way to progressively match the primitive shapes to which some neurons in
the primate brain were sensitive, detaching themselves from the hieroglyphic
pictograms. It took about half a millennium to accomplish this process.
I have recently coauthored a paper on a cognitive approach to Dao de Jing
(Wagshal-Te’eni and Tsur forthcoming). So I venture to use my experience
gained there to illustrate characters we find in Chinese: 道德經第十四章
贊玄 (meaning “Dao de Jing, chapter 14: Praising the Mysterious” [http://
www.yellowbridge.com/onlinelit/daodejing14.php]). These figures consist
of complex combinations of primitive intersecting lines and angles.
As Dehaene notes,
In the end, the neuronal recycling hypothesis leads me to believe that we are
able to learn to read because within our preexisting circuits, there is one that
links the left ventral visual pathway to the left-hemispheric language areas. This
circuit is already capable of recognizing many letter-like shapes, and it possesses
enough plasticity, or adaptability, to reorient toward whichever shapes are used
in our alphabet. (2013)
At this point one may pinpoint where my constraints conception of the

brain meets the reviewer’s plasticity conception (see discussion above,
p. 43); namely, “the inferotemporal visual cortex [that] already contains
neurons sensitive to letter-like combinations of lines such as T, L, X, and *,”
shared by all primates, constrains the possible symbols of all writing

7 4
systems of the world. But the capability of recycling such “older cerebral
structures that originally were selected by evolution to address very differ-
ent problems” indicates the plasticity of the brain.
Now we have to face the question of how sensitivity to the same combina-
tions of lines in primate brains can shape and constrain such a wide variety
of writing systems around the world (with the exception of Egyptian hiero-
glyphs). I suggest that one may account for these differences in the world’s
writing systems by invoking Polányi’s (1967) “Principle of Marginal Control”:
namely, that each higher process makes use of the margins left indetermi-
nate by the lower-level processes. In the present instance, existing brain
structures may determine the graphic primitives of alphabets but leave inde-
terminate their specific combinations into characters, as well as the linguistic
units to which the characters may be attached. In the foregoing Chinese text,
for instance, they combine into very complex characters. And in various
writing systems the characters may get attached to phonemes in the Latin
alphabet, to consonants only in the Hebrew and Arab alphabets, to syllables
in Japanese kana notation, or to entire words or morphemes as in Chinese.
In Bartlett’s experiment repeated social transmission is discussed in
great detail, but there is no evidence regarding the constraints of the brain,
only the cultural cognitive schemata involved. In Dehaene’s case, on the
contrary, repeated social transmission through generations of scribes is
only hypothesized; but he provides specific information about the natu-
ral capacities and constraints of the primate brain, on the one hand, and
cultural artifacts (writing systems), on the other. To fill in this gap, in a
rather coarse-grained account of the standard historical facts, I have tried
to trace the gradual development of writing systems, from hieroglyphs that
had a good fit to the physical objects out there in the world, through the
Proto-Sinaitic alphabet, through Semitic alphabets, to the Greek and Latin
alphabets that have a good fit to the abstract shapes to which certain neu-
rons in primate brains have special sensitivity. In all this I relied on more or
less traditional, noncognitive scholarship.
Thus far we have considered the brain’s constraints that governed the
process. In the major part of his article, Dehaene explores the plasticity of
the human brain that enabled the exploitation of existing brain structures
for new cultural ends. He puts the predictions of the neuronal recycling
hypothesis to the test. One of the predictions has to do with the notion of a
cortical competition process. The model predicts that as cortical territories
dedicated to evolutionarily older functions are invaded by novel cultural
objects, their previous organization will slightly shift away from the origi-
nal function (though the original function is never entirely suppressed).
As a result, reading acquisition should displace and dislodge whichever
84
evolutionarily older function is implemented at the site of the visual word

form area.
Dehaene tested two groups of adults about age fifty who had not
attended school. Ten were illiterates, and twenty-one were ex-illiterates
(had attended alphabetization classes); they were compared with thirty-two
literate adults with the same socioeconomic background. He also tested
nine-year-old good and poor readers. He summarized the results as follows:
Exactly as predicted, we also observed a small but significant cortical competi-

tion effect, precisely at the site of the letterbox area. For the first time, our study
revealed which shapes triggered a response at this site prior to learning to read.
In illiterates, faces and objects caused intense activity in this region—and, strik-
ingly, the response to faces diminished with literacy. It was highest in illiter-
ates, and quickly dropped in ex-illiterates and literates. This cortical competition
effect, whereby word responses increased while face responses decreased, was
found only in the left hemisphere. In the symmetrical fusiform area of the right
hemisphere, face responses increased with reading. Thus, at least part of the
right-hemisphere specialization for faces, which has been repeatedly observed
in dozens of fMRI studies, arises from the fact that these neuroimaging studies
always involved educated adults. Obviously, the acquisition of reading involves
the reconversion of evolutionary older cortical territory, and text competes with
faces for a place in the cortex. (2013)
Dehaene and his colleagues also replicated this finding in children:
When scanning 9-year-olds who were good readers versus poor readers, we found
two interesting differences in the ventral visual pathway. . . . The poor readers
not only showed weaker responses to written words in the left-hemispheric
visual word form area, but also showed weaker responses to faces in the
right-hemispheric fusiform face area. Thus, the acquisition of reading seemed
to induce an important reorganization of the ventral visual pathway, which dis-
places the cortical responses to face away from the left hemisphere and more
toward the right. (2013)
Dehaene’s experiments illustrate both the constraints and the plasticity

aspects of the brain experimentally. He shows how preexisting brain struc-
tures have shaped and constrained a wide variety of writing systems in
the world’s cultures. He hypothesizes that this occurred through repeated
social transmission, over the course of which the writing systems assumed
a good fit to the graphic primitives shared by humans with other primates.
Dehaene also shows how literacy alters the brain functions associated with

9 4
various brain areas. This is, of course, evidence for the brain’s plasticity.
He further indicates the constraints that required such alterations: the
“cortical competition effect, whereby word responses increased while face
responses decreased” in the visual word form area.
Thus we do not face an either/or choice: the constraints conception and
the plasticity conception of the human brain complement each other in
generating such cultural artifacts as writing systems and poetic conven-
tions. Dehaene’s article has filled two gaps disputed by my reviewer and me.
MORE TROUBLE
I have discussed the principle that in the process of repeated social trans-
mission, poetic conventions come to take forms that have a good fit to the
natural capacities and constraints of the human brain. This does not mean,
however, that there are no exceptions or even poetic conventions that pre-
scribe precisely the opposite of what has a good fit to the natural capacities
and constraints of the human brain. I noted the perceptual principle that
in a series of units of equal importance, the most natural sequence is that
the longest comes last, on the phrase level, the verse-line level, and the foot
level, as well as in music. As we have seen, this is so even on the elementary
“tick-tack” level. A useful explanatory model must be able to account for
apparent counterexamples as well, such as, for instance, when convention
demands a series in which the longest item comes first. At the beginning
of this chapter I invoked an unintentional process of natural selection by
which the metrical pattern that has the best fit to the natural constraints of
cognitive economy is the one that has a better chance to prevail and multi-
ply. Accordingly, in medieval Hebrew poetry, a foot in which the longer part
comes last became the dominant one. In English, Hungarian, and Hebrew
syllabotonic verse the iambic foot became the dominant one, in which the
longer position comes last as well, and this is probably the case in classical
Arab prosody too. In the Serbian folk epic and in French syllabic poetry the
longer segment of decasyllabic lines comes last. This is the case in musical
phrasing too.
In this context, however, there is a disconcerting issue. The natural
selection discussed here works well in some poetic traditions, but other
traditions force rigid limitations upon poets and prevent them from pre-
ferring the metrical patterns that have a good fit to the natural constraints
of cognitive economy. For instance, the venerable Greek and Roman epic
tradition is dominated by the dactylic hexameter (“the heroic rhythm”).
The dactylic foot is the least natural of the most frequent feet, namely,
05
the iambic, the trochee, the anapest, and the dactyl. The trochaic and the
dactylic feet are relatively unnatural because the strong position (longer
syllable) precedes the weak one(s). The dactylic is less natural than the tro-
chaic because two weak positions lean back upon one strong position. The
appraisal that the trochee and the dactylic hexameter are the least natural
ones among the predominant meters is corroborated by Aristotle’s com-
ment in Rhetoric: “Of the various rhythms, the heroic . . . lacks the tones
of the spoken language,” and “the trochee is too much akin to wild danc-
ing” (1932). Why, then, should precisely that rhythm that has a poor fit to
the constraints of the cognitive system, that is, is the least natural of the
rhythms, prevail and multiply?
For years I had to content myself with the answer that rigid conven-
tion robbed poets of the freedom to follow their own intuitive preferences.
According to the present conception, however, this solution poses another
problem. We would expect rigid conventions to perpetuate the more natu-
ral rather than the less natural meters. I found the solution to this riddle
only recently, when I reread, for a different purpose, Aristotle’s paragraph:
Of the various rhythms, the heroic has dignity, but lacks the tones of the spoken
language. The iambic is the very language of ordinary people, so that in common
talk iambic lines occur oftener than any others: but in a speech we need dignity
and the power of taking the hearer out of his ordinary self. The trochee is too
much akin to wild dancing: we can see this in tetrameter verse, which is one of
the trochaic rhythms. (1932)
Aristotle claims that “of the various rhythms, the heroic has dignity” and
“the power of taking the hearer out of his ordinary self.” In other words,
heroic action does not demand the most natural rhythms but, rather, those
that take the hearer out of his or her ordinary self—that is, the least natu-
ral meters. This is how the most “dignified” rhythm is generated, so as to
support the intensive human quality of the epic. The “heroic rhythm” (the
dactylic hexameter) is the most dignified one because of its deviation from
the commonplace “tones of the spoken language.” In D’Andrade’s terms, the
most “dignified” meter turns out to be the one that has a less good fit to the
natural constraints of the human brain—precisely because it goes against
them.13 Hence the fact that some, or many, or most conventions acquire a
13. Elsewhere I discuss the question of what “dignified” may mean with reference to
metrical structures (Tsur 2012a: 408–409). The view I put forward there may account
for the “dignified” quality of the spondaic substitution in hexameter verse; the present
argument accounts for the “dignified” quality of the dactyl.

51
good fit to the natural capacities and constraints of the human brain does
not exclude the existence of conventions whose fit to those capacities and
constraints is poorer. Poor fit may have its own expressive value. One may
even assume that poets sometimes deliberately seek out the conventions
that have a poor fit, for effects that the Structuralists would call “marked.”
Finally, another example shows how poor fit has its own expressive
value. I argued above that in the iambic pentameter line a caesura after
the fourth position is the “unmarked,” more natural caesura: the nearer an
intruding event to the end of the verse line (that is, the shorter the seg-
ment that comes last), the more it increases the need for the missing seg-
ment, and the greater the satisfaction when it “clicks in” (Tsur 1972). Thus,
the caesura after the fourth position may have a natural, relatively relaxed
character, whereas a caesura after the sixth position may elicit a mild sense
of unease or frustration, followed by a mild sense of satisfaction when the
missing segment is supplied.
MUSIC AND EFFICIENT CODING
Another conspicuous counterexample is the so-called ballad stanza. Here

a tetrameter line is followed by a trimeter; that is, the shortest comes last.
My initial explanation for this was that ballads have been sung since the
dawn of their history, and music can facilitate the inversion of the natural
order, achieving greater complexity. It is this verbal structure that has been
fossilized, even after the disappearance of the music. This, of course, raises
again the question of why the unnatural form was fossilized. Elsewhere
(Tsur 2013), I answer this question very much in the spirit of the answer
given above with reference to the “heroic rhythm.” Poets may attain spe-
cific effects by having recourse to the marked forms. Thus, for instance, the
long-short-long-short form of the ballad stanza may be perceived as curt,
downright, and relentless. This psychological atmosphere may reinforce
the relentless, straightforward atmosphere generated by the ballad plot.
There is some evidence from widely different corpora that where music is
involved, the cognitive rule “the longest comes last” is sometimes radically
violated. As we have seen, in medieval Hebrew poetry in Spain, meter is
most often based on a foot in which the shortest unit, the schwa mobile
(see note 2 above), occurs at the beginning (followed by three long units),
whereas the second most frequent meter is based on the same units, but
with the schwa mobile toward the end, before the last long unit. Yehosheva
Bentov and I noted in a joint article that, in liturgical poetry written in
the classical meters, a third meter that puts the longest foot first slightly
25
exceeds the frequency of the meter that looms largest in the combined cor-
pus of secular and liturgical poetry (Tsur and Bentov 1996). This can be
explained in the following way: as we have seen above, there is an aesthetic
pressure for unity and complexity. The perceptual unity of a verse line is
preserved only if it is kept within the scope of short-term memory, which
functions in the acoustic mode like an echo box. The span of short-term
memory cannot be extended; only the verbal material can be recoded in a
more efficient manner. Poets want to increase the complexity of verse lines
without sacrificing their integrity. Liturgical poetry is more frequently
sung than secular poetry. Music may thus serve as an additional coding
device by alleviating the load on short-term memory. This allows the poet
to increase the complexity of the verse line without violating its integrity.
The same explanation applies to another phenomenon reported by David
Gil (see Tsur and Bentov 1996). Gil explored such noncanonical poetic
genres as cheers at soccer games, political demonstrations, and the chants
of peddlers and hucksters at the marketplace. His corpus of texts (from
Western Europe, the United States, the Middle East, and the Far East) offers
an impressive intercultural sample. Gil found that most of the cheers were
rhythmical, and only a minority of them were sung. In the vast majority of
them the longest unit came last. But he found it puzzling that all the texts
to which a melody was attached had their prosodic weight at the beginning
of the versification units. Here too, I contend, music is enrolled as an addi-
tional coding device to alleviate the burden on short-term memory when
the text increases the load. Briefly, music makes it possible to increase the
complexity of the prosodic unit without violating its integrity.
Perhaps the most intriguing issue is related to biblical cantillations,14
as reported by Ronit Shoshany (2009). She uses the terms iambic and tro-
chaic in a roughly figurative sense: namely, to define divisions in which the
shorter section precedes the longer one or follows it, respectively. She uti-
lizes my proposal that music may relieve the burden on short-term mem-
ory and make it easier for the longer section to precede the shorter one.
Here is her abstract:
This essay discusses the question of the original purpose of accents in bib-
lical text. Scholars differ as to whether the accents were introduced either as
14. “The Hebrew Bible text is annotated with a system of diacritic marks called ‘accents’
(Hebrew ‫ טעמים‬ṭeʕamim; singular ‫ טעם‬ṭaʕam ‘taste, sense, reason’). These accents,
assigned to every word in the Bible, parse each verse in minute detail. This complex
system of representation, developed in and around Tiberias over several generations
up until the 10th century . . . , serves several purposes: among other things, it marks
the position of stress, and guides the musical cantillation of the text” (Dresher 2013).

3 5
punctuation or for musical purposes. After reviewing the arguments by scholars

in favor of the two hypotheses, I conclude that in early stages, the accents did
not play any musical role. This role was added at a later stage. My conclusion is
based on facts not previously taken into account by scholars, i.e., the transition
in the division of the accents from iambic (a short section preceding a long sec-
tion) to trochaic (a long section preceding a short section) structure. This transi-
tion is demonstrated in the Babylonian system of accentuation. Under the early
Babylonian system, there is a tendency to prefer the iambic division, whereas
the late Babylonian system (which is identical to the Tiberian system in this
respect) shows a tendency to prefer the trochaic division. My explanation refers
to a universal correlation: iambic structures and spoken language, on the one
hand, trochaic structures and melody, on the other. (2009; my translation)
Shoshany provides ample documentation of what is known for cer-

tain: that today the cantillation marks indicate both punctuation (division
of the text into chunks) and instructions for musical performance; initially
the shorter chunks tended to precede the longer chunks, but at a certain
point in history this was reversed. One issue is still disputed: whether they
served as musical guidelines from the very beginning or from roughly the
time when the “trochaic reversal” occurred. Shoshany assumes that the
above considerations suggest that intonation was increasingly overarticu-
lated, until it became music, which, in turn, made the “trochaic reversal”
possible, increasing complexity.
CONCLUSION
My earlier work has been sometimes praised, sometimes censured for pur-
portedly elaborating a theory of poetic universals. I never intended to do
so and am not qualified to make claims regarding universals of any kind.
The most I can claim is that certain conventions and kinds of response
can be found in a number of cultures; but this is a far cry from univer-
sals. I have a more modest ambition: to contrast two approaches to literary
conventions, the migration (influence-hunting) and the cognitive-fossils
(constraints-seeking) approaches. Both approaches assume repeated social
transmission but different states of affairs as points of departure (full-blown
conventions or responses acquired for adaptation purposes) and usually
rely on different conceptions of mental operations—behavioristic and cog-
nitive, respectively.
Nor do I seek to imply that the constraints of the brain impose an abso-
lute determinism on possible literary forms. In fact, my argument is that
45
all imaginable oddities can or could have been invented at one time or
another. This process of “natural selection” through repeated social trans-
mission renders them into conventions by acquiring a good fit to the natural
capacities and constraints of the human brain. Moreover, as we have seen,
the same constraints of the brain may afford several alternative solutions.
This may be one reason for the extreme cultural variation observed. Being
shaped by the constraints of the brain does not necessarily imply “com-
monsense” or “logical.” I have argued elsewhere (Tsur 2008a: 414–422)
that even such “wild inventions” as the metaphysical conceit could result
from the constraints of the brain, of which Dr. Johnson says in his essay on
Abraham Cowley that the reader wonders “by what perverseness or indus-
try they were ever found” (1951 [1779]) (see, e.g., c hapter 6). I argue that
the same implausible conceit could have been invented more than once
through the same apparatus.
The migration approach is frequently associated with factualism. Its
exponents often object to the cognitive-constraints approach on the
grounds that it goes beyond the hard facts by applying to them cognitive
hypotheses; and like all hypotheses, these cannot be true but, rather, at
best, plausible. On closer inspection, however, migration-hunters cannot
make arguments either without having recourse to hypotheses; they only
disguise them as facts. What they can know for a fact is that a certain con-
vention did occur in places A, B, and C. They can also know the geographic
distance between these places. But that the convention migrated from place
A to B and from B to C in many cases is a very plausible hypothesis rather
than a fact. Apart from this, the migration approach can only point out the
occurrence of a convention at a certain place and time but cannot account
for how it acquired a meaning, an emotional import, or a perceptual quality
such as “dignified” or how these could have been communicated to its first
audience. It can, at best, point out that at that time it was a convention to
attribute such and such a meaning or emotional import to a metaphor or to
regard, for example, the dactylic hexameter as “dignified.”
Moreover, even where deliberate migration is explicitly intended, unex-
pected constraints encountered en route may bring about unexpected results.
When, for instance, English Renaissance poets thought that they were
importing the Greek and Latin versification systems into their stress-timed
language, they unwittingly generated the syllabotonic system. When French
poets attempted to import the Greek and Latin versification systems into
their syllable-timed language, they simply failed. When Hungarian poets
attempted to import the same, their results were even more inextricable.
Hungarian iambic is syllabotonic, whereas Hungarian trochees and dactyls
are basically “quantitative,” as in Greek and Latin poetry. Likewise, when

5
medieval Hebrew poets deliberately imported the Arabic versification sys-

tem, they too imported only the names and abstract patterns of feet and
meters but found themselves filling the weak and strong positions with a dif-
ferent kind of immediately observable constituent, generating a new system.
By contrast, the cognitive-fossils or constraints-seeking approach traces
fossilized conventions back to once-active cognitive processes; thus it can
easily account for how a certain convention acquired a certain meaning, emo-
tive import, or perceptual quality and how the constraints of the importing
language changed the imported goods. Bartlett’s “broken telephone” experi-
ment suggests how an individual poet’s idiosyncratic linguistic invention
could become a convention by gradually acquiring a good fit to the constraints
of the brain or prevalent schemata, through repeated social transmission.
The cognitive-constraints theory is compatible both with the migra-
tion theory and with the assumption that the same convention could be
invented more than once, in different places. In chapter 1 I cited evidence
that in the evolution of the species, as well, such evolutionary “devices” as
wings, for instance, were “invented” more than once, on different branches
of the evolutionary tree. Thus there is no reason to assume that cultural
conventions could not be invented more than once, in different places, at
different times, for similar purposes.
65
7 5
CHAPTER 3
Poetic Conventions as Fossilized

Cognitive Devices
The Case of Medieval and Renaissance Poetics
SOME FUNDAMENTAL QUESTIONS
In “Arabic Poet Al-Mutanabbi: A Maslovian Humanistic Approach,” Ratna

Roshida Abd Razak (2007) highlights the question of whether we can read
a poet’s personality through his works and, in particular, whether we can
do so with medieval or Renaissance poetry. I believe that Abd Razak’s
attempt to apply Maslow’s (1962) theory to medieval poetry is an impor-
tant event and has significant implications for literary research. Abd Razak
employs Maslovian theory to consider the great Arab poet al-Mutanabbi
(915–965) as a self-actualizing person. According to Maslow, human
beings are governed by a hierarchy of biologically based needs. He used
the word hierarchy because he felt that these needs formed a hierarchy of
propensity, since the lower needs must be satisfied in order to be able to
move on and gratify the upper ones. Abd Razak assumes that through the
various genres of his poetry, one may examine al-Mutanabbi’s strong and
forceful personality and the successive levels of his self-actualization in
Maslow’s sense.
The dominant (and sometimes militant) view among scholars of medi-
eval poetries is that one should not apply modern psychological theories to
medieval poets and poetry that were not shared by the poets themselves.
This view is not unlike the claim that one should not assume that blood
circulated in poets’ bodies before 1628, when William Harvey published his
85
study on the circulation of blood.1 Abd Razak’s violation of the prevalent

professional taboo is, therefore, more than welcome. I also believe, how-
ever, that we should not take for granted what is explained by the applica-
tion of a psychological theory and should scrutinize its implications in light
of research on poetic conventions. This is the aim of this chapter.
In what follows, I will make no statements about al-Mutanabbi’s poetry,
only ask questions, and then will present a generalized discussion on the
nature of poetic conventions in medieval and Renaissance poetry and
about the psychological foundations of poetic conventions in general. I will
proceed in four steps. I will raise some questions regarding Abd Razak’s
conception of al-Mutanabbi’s poetry, present a caveat against regarding
medieval and Renaissance poetry as a clue to a poet’s personality, develop
an alternative perspective on the application of psychological theories
to poetic conventions, and apply this model to one widespread poetic
convention.
Among other things, Abd Razak argues:
In order to please patrons, poets had to compose panegyrics (love poems) that
satisfied their patrons by showing mastery of the language. The poet would be
conferred prestige after giving praise for his patron’s accomplishments. . . .
The recital of a panegyric was an important formal occasion and provided
an opportunity for the sovereign to demonstrate his generosity publicly by
handsomely rewarding the poet, who, if he genuinely, but secretly admired his
patron, could be inspired to produce truly excellent work.
. . . Al-Mutanabbi was able to feel safe and secure in the company of
Sayf al-Dawlah, who became his intimate friend and comrade in arms. As a
result, al-Mutanabbi was better able to appraise the heroism of the prince
wholeheartedly. . . .
. . . Having gained intimate knowledge of Sayf al-Dawlah, we can presume
al-Mutanabbi then grew to love him as his patron. (2007)
Abd Razak uses indicative statements about the social and literary for-
malities related to panegyrics: “poets had to compose panegyrics,” and “the
recital of a panegyric was an important formal occasion.” As to the warm
1. I learned at school that William Harvey discovered the circulation of blood in the
human body. Wikipedia, however, tells us that the circulation of blood was discovered
as early as 1242 by the Arabian physician Ibn al-Nafis. So Arab poets and Hebrew poets
in Muslim Spain can be considered to have had circulating blood as of the thirteenth
century on. This, however, would still leave al-Mutanabbi and the Hebrew poets of the
golden age without circulation.

9 5
personal relationship between al-Mutanabbi and his patron, she uses the
conditional form “if he genuinely, but secretly admired his patron” and
carefully emphasizes affordances rather than facts, as in “was able to feel
safe and secure in the company of Sayf al-Dawlah” and “was better able
to appraise the heroism of the prince wholeheartedly.” She also uses the
hypothetical construction “we can presume al-Mutanabbi then grew to love
him as his patron.” Linguistically, at least, certain things are presented as
known, and other things emerge as conjectures. Briefly, the social constel-
lation as well as the requirements of the genre prescribed the praise and the
emotions to be expressed in these poems, irrespective of what the poets
felt. Is there any evidence that, in some instances at least, the poets did
indeed experience the emotions that the conventions of the occasion and
the genre prescribed for their poems? Furthermore, what kind of evidence
would be deemed satisfactory? Or is it merely a post-Romantic assumption
that poets attempt to express their genuine feelings?
GENRE CONVENTIONS AND THE BIOGRAPHICAL FALLACY
Tenth- and eleventh-century Hebrew poets in Spain explicitly adopted

the aesthetic conceptions as well as the conventions in Arabic poetry of
thematic genres (love poetry, panegyrics, garden descriptions, drinking
poems, boasting poems, etc.), figurative language, and even prosody. In
prosody, for instance, they imported the qassida (monorhymed verse) and
the muwwashaḥ (girdle poems) and adapted the Arab quantitative meter
(based on long and short syllables) to the constraints of the Hebrew lan-
guage (see chapter 2). Hebrew poets of the era put forward their poetics
explicitly in treatises written in Arabic(!).
One problem concerning the personal element in this poetry is that
even when a poet says “I,” in certain genres he does not necessarily refer to
his subjective self but, rather, what I would call an “exemplary I.” Take the
genre “Poems of Contemplation.” In this genre, poets contemplate Man’s
fate and place in the universe. In one of his contemplative poems, Moses
Ibn Ezra wrote,
?‫אֹותי ָח ְׁשבּו ַא ַחי לְ נָ כְ ִר‬

ִ ְ‫ ו‬,‫רֹובי‬ַ ‫וְ ֵאיְך מּוזָ ר נְ ָתנּונִ י ְק‬
!‫הֹורי‬
ִ ְ‫אֹותי יְ לָ ַדי– וְ כֵ ן זָ ְד ִּתי וְ ָׁשכַ ְח ִּתי ל‬
ִ ‫וְ כָ ֶהם זָ נְ חּו‬
[How did my relatives render me strange
and how did my brethren regard me a stranger
And just like them my children have forgotten me
And similarly I have willfully forgotten my parents]. (my translation)
P o e t i c C o n v e n t i o n s a s F o s s i l i z e d C o g n i t i v e De v i c e s [ 59 ]
06
Based on such lines, some critics (e.g., Shim’on Bernstein) came to grim
conclusions as to the personal relationships in the Ibn Ezra family. I call
this the “biographical fallacy.” Hebrew scholar and poet Dan Pagis (1970),
however, compared the use of the first-person singular in this genre (in
both Arabic and Hebrew poetry of the period) with its use in the genre of
“Poems of Complaints.” The latter uses a personal “I,” whereas the former
has an exemplary “I.” This is an example of the condition humaine—that is,
of how each newer generation forgets the older generation—and has noth-
ing to do with what happened in the Ibn Ezra family.
In elegies, as in the elegy on the death of al-Mutanabbi’s grandmother,
authentic autobiographical elements do occasionally occur. But one can-
not make inferences based on this genre regarding the authenticity of
experiences in the panegyric genre, for instance. Furthermore, the elegy
itself abounds in conventional hyperboles, even if we have good reasons
to suppose that the poet experienced a genuine shock at the death of the
deceased. Thus, for instance, the great eleventh-century Hebrew poet
Shlomo Ibn Gabirol wrote a long elegy on the murder (or execution?) of his
friend and benefactor Yekutiel Ibn Hassan, which begins as follows:
‫קּות ֵיאל ֲא ֶׁשר נִ גְ ָמרּו אֹות ּכִ י ְׁש ָח ִקים לַ ֲחֹלף יֻ ּצָ רּו‬
ִ ְ‫ִּב ֵימי י‬
[In the days of Yekutiel that are over
(It’s a sign that the skies are doomed to pass away)]
In panegyrics, poets were expected to express great praise. This is what

both the genre and the social occasion demanded, irrespective of what the
poet felt toward the object of praise. It would be misleading to conclude that
extreme expressions of love indicate a poet’s great love of the addressee.
Do such lines as the following by al-Mutanabbi sound like expressions of
heartfelt love or conventional hyperboles?
Every life you do not grace is death

every sun that you are not is darkness. (Abd Razak 2007)
Probably, the more exaggerated the praise and the expression of love,
the less credible it appears. Poets as well as the targets of praise were well
aware of a certain discrepancy between what the poet thought and what
the poet said in this genre. In his discussion on panegyrics in his trea-
tise on poetics (written in Arabic), Moses Ibn Ezra makes the following
remark: “Once a poet went into excesses in praising one of the high state
officials. . . . But the extolled one resented this and answered: I am less

61
than what you say but more than what you think of me” (1975). This leads
to the question of how we can determine whether a poet did or did not
feel the sentiments being expressed. Let me state at the outset that I don’t
know the answer.
In sixteenth- and seventeenth-century English poetry we encoun-
ter similar problems. Consider the following well-known couplet from
Shakespeare’s Sonnet XVIII:
As long as men can breathe, or eyes can see,

So long lives this, and this gives life to thee.
At one point in my literary education I was told that these lines indi-
cate (or prove?) that Shakespeare was already aware of the immortality
of his poetry. Such studies as J. W. Lever’s The Elizabethan Love Sonnet
(1966) have amply demonstrated that almost every image or idea found
in Shakespeare’s sonnets could be found in innumerable other sonnets
of the time, but Shakespeare simply wrote much better poetry than they
did.2 Countless poets promised immortality to their beloved through
their verses, but most of them have long been forgotten, and rightly
so. Medieval Arabic and Hebrew boasting poems differ from this only
in the degree of exaggeration, not in the degree of the authenticity of
experience.
Elizabethan poets and editors seem to have been aware that the pas-
sions conveyed in their lyric poems were exemplary rather than authen-
tic personal experience; poets use such linguistic devices as metaphors or
the first-person singular for rhetorical purposes. In his admirable study
of Elizabethan poetry, Hallett Smith (1968) collected some illuminat-
ing comments to that effect from the Mirrour for Magistrates, both by an
anonymous annotator of Thomas Watson’s poetry and by Watson him-
self: “Watson observes to the friendly reader in his preface that he hopes the
reader will excuse any faults escaped, ‘in respect of my trauaile in penning
these loue passions, or for pitie of my paines in suffering them (although
2. The same feature was shown for Villon by Italo Siciliano in his monumental
François Villon et les thèmes poétiques du Moyen Age (1934). About sixty-seven years
later this became a commonplace in Villon studies: “[The naïve reader] is captivated by
Georges Brassens singing La ballade des dames du temps jadis, only to find that Villon
wrote two other poems on precisely the same theme, and that every second poet of the
later Middle Ages had made use of the same well-tried cliché” (Taylor 2001: 6). Siciliano
demonstrates that La ballade des dames du temps jadis is simply more euphonious than
the other poems of this genre. To this one might add that the ubi sunt motive was wide-
spread in eleventh-century Hebrew poetry as well.
26
but supposed)’ ” (Smith 1968: 138). Elsewhere the annotator comments,

“In this Passion is effectualIy set downe, in how straunge a case he liueth
that is in loue, and in how contrary an estate to all other men, which are at
defiaunce with the like folIye. And this the Author expresseth here in his
owne person” (in Smith 1968: 138). Smith construes this as follows: “The
‘I’ in the sonnets is then a role for effectiveness; it is to make the sonnet
more ‘patheticall.’ ” The last two lines of No. 22 are added by Watson, says
the annotator, “to make the rest to seeme the more patheticall’ ”: “The com-
mentator points out that the interest of the poem lies in the fact that ‘cer-
taine contrarieties, whiche are incident to him that loueth extreemelye, are
liuely expressed by a Metaphore’ ” (Smith 1968: 136). Again, the contra-
rieties are not about the poet’s personal experiences but about “him that
loueth extreemelye.”
I have explored the nature of “I” in eleventh-century Hebrew and
sixteenth- and seventeenth-century English devotional poetry (Tsur
1974). In these traditions there are texts that can be read as poems,
prayers, or meditations depending on the circumstances. The relation-
ships change between the textual “I” and the flesh-and-blood person who
utters the “I.” Consider the following problem. When I utter an interjec-
tion, say, “ouch” or “alas,” I may be indicating that I have certain emotions
or feelings. If I do have these emotions, that is, if there is a corresponding
fact in the extralinguistic context, these emotions or feelings are genuine.
If there are no corresponding extralinguistic facts, the emotions and feel-
ings are feigned. When we say that a person is saying prayers with devo-
tion (or, in Hebrew, with kawwana), we mean (1) that he or she asserts
all its explicit and implicit statements, (2) that heightened mental activi-
ties are indicated, and (3) that these heightened activities are genuine.
Elsewhere I make an additional, related distinction: the text may or may
not give the illusion of an authentic experience, irrespective of whether the
flesh-and-blood poet did or did not undergo the suggested experience (Tsur
1969). This does not necessarily make one of the poems better; rather, it is a
stylistic difference.
When people utter a text as a prayer, they assert its statements and do
their best to live up to the emotions and attitudes indicated in it. When one
reads a poem, one need not assert its statements or live up to the emotions
and attitudes indicated in it. As Philip Sidney wrote in his Apology, the poet
“nothing affirmeth, and therefore never lieth” (Sydney). The reader need
not assert the statements of a poem; he or she only must accept that the
speaker in the poem asserts them. A reader of Donne’s Holy Sonnets need
not believe in Jesus, just as a person attending a performance of Oedipus
Rex need not believe in Zeus.

3 6
POETIC CONVENTIONS AS FOSSILIZED COGNITIVE DEVICES
We may, then, ask why the publication of an essay applying Maslovian theory
to al-Mutanabbi’s poetry is an important event. But we should also be care-
ful about what we have explained using the application of some psychologi-
cal theory. The psychological theory may explain something that has great
poetic significance, but not necessarily the poet’s subjective experience.
It should be recalled that this poetry is not merely highly conventional
but also ornamental. This book discusses the ways in which poetic conven-
tions and ornaments arise. My answer to this question is based on a psy-
choanalytic concept adopted from Ehrenzweig and a “Darwinian” concept
adopted from D’Andrade’s (1980) cognitive anthropology. Briefly, as dis-
cussed in chapter 1, the reasoning is that poetic conventions are psycho-
logical processes fossilized into verbal devices and that in the process of
repeated social transmission, cultural programs come to take forms that
have a good fit to the natural capacities of the human brain. So, what light
does this theoretical model throw on the application of Maslow’s psycho-
logical theory to al-Mutanabbi’s poetry? It will not necessarily explain the
individual poet’s actual experiences and personal emotional development.
But it may perhaps illuminate the psychological processes whose expres-
sions have fossilized in such highly ornamental poetic conventions as
boasting poetry, love poetry, the panegyric, and so forth.
Let us take a look at Abd Razak’s argument from this point of
view: “Maslow’s theory is primarily concerned with that part of human
motivation and behavior which is based on the higher needs.” Take, for
instance, “the need for self-esteem,” as applied to al-Mutanabbi’s poetry:
Assuming that at this point the first three levels of the hierarchy [the search for
safety, the need to belong, and the need for affection] of al-Mutanabbi’s needs
were adequately satisfied, we would expect him to be concerned with the need
for esteem. Maslow distinguishes two types of esteem needs. The first is esteem
from others. This involves the desire for reputation, status, recognition, fame
and a feeling of being useful and necessary. Individuals need to feel respected
and valued by others for their accomplishments and contribution. Self-esteem,
on the other hand, involves a personal desire for feelings of competence, mas-
tery, confidence and capability.
Self-esteem is therefore closely linked to the desire for superiority and
respect from others. (Abd Razak 2007)
Consider the following utterance by al-Mutanabbi: “If I am conceited, it

is the conceit of an amazing man who has / never found any surpassing
46
himself.” Is this indeed what he thinks of himself, or is he using a con-

ventional hyperbole? My point is that having recourse to the genre of
boasting poetry does not necessarily indicate that the first three lev-
els of the poet’s needs have been met as defined in Maslow’s model.
Rather, I strongly suspect that a very different dynamic is involved
in a poet’s choice of genres. However, one could make a convincing
case for the assumption that Maslow’s model accounts for the person-
ality dynamics that originally prompted certain boastful expressions,
which, in turn, through the process of repeated social transmission,
eventually fossilized in the conventions of boasting poetry. I propose
to illustrate the process propounded here by elaborating on another
poetic convention.
TEST CASE: CATALOGS OF CONTRADICTIONS

(a) Petrarch, Wyatt, and Ronsard
As we have seen, Watson’s commentator mentions “certaine contra-

rieties, whiche are incident to him that loueth extreemelye.” In this
context, let us consider the processes by which highly effective expres-
sive resources are turned into conventions or ornaments. Leveling and
sharpening may play an important role in these processes. A conspicu-
ous case in point is the venerable thematic convention of “the contrari-
ous passions in a lover.” The most famous example is Petrarch’s sonnet
“Pace non trovo”:
Canzoniere CXXXIV
Pace non trovo e non ho da far guerra
e temo, e spero; e ardo e sono un ghiaccio;
e volo sopra ‘l cielo, e giaccio in terra;
e nulla stringo, e tutto il mondo abbraccio.
Tal m’ha in pregion, che non m’apre nè sera,
nè per suo mi riten nè scioglie il laccio;
e non m’ancide Amore, e non mi sferra,
nè mi vuol vivo, nè mi trae d’impaccio.
Veggio senz’occhi, e non ho lingua, e grido;
e bramo di perire, e chieggio aita;
e ho in odio me stesso, e amo altrui.
Pascomi di dolor, piangendo rido;
egualmente mi spiace morte e vita:
in questo stato son, donna, per voi.

5 6
This sonnet was imitated and translated by poets throughout the

Renaissance, among others by Wyatt:
Description of the Contrarious Passions in a Lover

I find no peace, and all my war is done;
I fear and hope, I burn, and freeze like ice;
I fly aloft, yet can I not arise;
And nought I have, and all the world I seize on,
That locks nor loseth, holdeth me in prison,
And holds me not, yet can I scape no wise:
Nor lets me live, nor die, at my devise,
And yet of death it giveth me occasion.
Without eye I see; without tongue I plain:
I wish to perish, yet I ask for health;
I love another, and thus I hate myself;
I feed me in sorrow, and laugh in all my pain.
Lo, thus displeaseth me both death and life,
And my delight is causer of this strife.
This poetic convention has a long history and goes back at least to
Catullus (circa 84–54 bc):
Odi et amo. Quare id faciam, fortasse requiris?

nescio, sed fieri sentio et excrucior.
[I hate and I love. Why do I do it, perchance you might ask?
I don’t know, but it is a fact, I feel it, and it excruciates me.]
(my translation)
The same convention appears in Ronsard’s original Sonnet No. XII of his
Amours:
J’Espère et crains, je me tais et supplie,

Or’ je suis glace, et ores un feu chaud,
J’admire tout, et de rien ne me chaut,
Je me délace, et puis je me relie.
Rien ne me plaît sinon ce qui m’ennuie:
Je suis vaillant, et le cœur me défaut,
J’ai l’espoir bas, j’ai le courage haut
Je doute Amour, et si je le défie
Plus je me pique, et plus je suis rétif
J’aime être libre, et veux être captif
6
Cent fois je meurs, cent fois je prends naissance

Un Prométhée en passions je suis
Et pour aimer perdant toute puissance
Ne pouvant rien je fais ce que je puis.
[I hope and fear, I hold my tongue and supplicate.
Now I am ice, now a hot fire,
I admire everything and nothing matters to me,
I unloose myself and then tie myself again.
Nothing pleases me except what annoys me;
I am valiant and my heart defeats me;
my hopes are low and my courage high;
I fear love, and yet I defy it.
The more I spur myself the more stubbornly I stand still.
I love to be free and want to be a captive.
I die a hundred times and a hundred times am born.
A passionate Prometheus I am
and in order to love, losing all power,
unable to do anything, I do what I can.] (my translation)
This convention reflects two psychological phenomena: ambivalence and

defense mechanisms against expressive devices in art, as pointed out by
Ehrenzweig (1965). Ambivalence is a state of conflicting attitudes or feelings,
as when a mother causes pain to her child by vigorously pressing it to her
bosom. According to Freud, much love is characterized by such ambivalence.
The Fontana Dictionary of Modern Thought makes the following comment on
literature in its ambivalence entry: “Whereas, in general, ambivalence is a
potential source of undesirable stress . . . , in a writer it is widely regarded
as a source of strength and desirable tension, and in a fictional character
as evidence of subtlety in his or her creator.” An exquisite example of such
subtlety can be found in c hapter 4 (in the mother-son relationship in the
ballad “Edward”).
There are two conspicuous cognitive devices in these poems that tone
down the disturbing element of ambivalence. The process of sharpening
enhances in these poems the discordance between the conflicting atti-
tudes and presents them as clearly perceptible symmetrical opposites, thus
removing the potential source of undesirable stress: uncertainty and the
disquieting element from an unpleasantly ambiguous feeling. This process
is not unlike the process that, in dreams and fairy tales, eliminates ambiv-
alence toward one’s mother by splitting the mother figure into a good
mother and a wicked stepmother.

7 6
Since ambivalence is a potential source of undesirable stress, readers

of poetry may have sought out mental strategies that reduce undesirable
stress and may have discovered, independently from one another, that
in certain poetic situations of considerable emotional strain, sharpening
proves to be an exceptionally effective means to render these “contrarious
passions” harmless. In this way, one possible mental manipulation of cer-
tain highly emotional, disquieting, ambivalent texts may have fossilized
into a solid convention, reinforced through repeated social transmission.
My point is that in light of D’Andrade’s comment “that in the process of
repeated social transmission, cultural programs come to take forms which
have a good fit to the natural capacities and constraints of the human
brain” (1981: 182), ambivalence, leveling and sharpening, and the need for
defense against overly strong expressive resources seem to have a good fit
with the natural capacities and constraints of the human brain. Note, how-
ever, that by eliminating the disturbing element of ambivalence, we have
created another one: the baffling quality of asserted logical contradictions.
The need to eliminate may suggest an explanation for the typical presence
(in all these instances) of another technique—frequently implemented
to eliminate the disquieting quality from a unique emotion—known as
the “catalog technique.” Catalog verse is a term “to describe lists of per-
sons, places, things, or ideas which have a common denominator” (Green
2013: s.v. “catalog verse”). Lengthy catalogs are difficult to remember. One
way to alleviate the load on memory is to abstract a common denomina-
tor from all the items. When the mnemonic device outweighs the items
memorized, it may be turned, in certain circumstances, to an aesthetic
end: namely, into the emotional quality of a poem.
Thus, for instance, from the catalog of contradictions in Petrarch’s son-
net on the lover’s “contrarious passions” such “common denominators” as
“love” and “contrarious passions” can be abstracted. Since the processing
of parallel verbal structures typically involves the (emotional) right hemi-
sphere of the brain, whose output is global and diffuse, those abstractions
may be perceived as emotional qualities rather than compact concepts.
According to D’Andrade (1981), emotions are active in the background,
without preempting everything else. Experiments with the Stroop effect
would suggest that such (unmentioned) abstractions may also be highly
active in the background. The Stroop effect is a demonstration of interfer-
ence in the reaction time of a task. When the name of a color (e.g., “blue,”
“green,” or “red”) is printed in a color not denoted by the name (e.g., the
word red printed in blue ink instead of red ink), naming the color of the
word takes longer and is more prone to errors than when the color of the
86
ink matches the name of the color (see Stroop 1935). Further experiments
suggest that (unmentioned) category names too, extracted from parallel
items, may interfere with tasks with which they are incongruous. Thus,
“love” or “contrarious passions,” abstracted from the catalog of contradic-
tions, may be active in the background of Petrarch’s (or Wyatt’s) sonnet,
while the reader attends to the catalog of contradictions.
When one encounters some absurdity, one may laugh it off (the comic is,
indeed, a defense against threat) or may admit the authority of the threat
and try to resolve it. In the former case the effect will be frivolous and
delightful, and in the latter, disquieting. Alternatively, one may be able to
resolve it, in which case it will yield great satisfaction. Such satisfaction
would be particularly significant in light of Maslow’s model as quoted by
Abd Razak, with reference to “self-esteem”: it “involves a personal desire
for feelings of competence, mastery, confidence and capability” (2007).
When these absurdities occur in close succession, one is unlikely to be able
to work out the details of the resolutions. This leaves one with several pos-
sible strategies. If the succession of absurdities occurs in a situation that
affects one’s well-being, one will experience disorientation, discomfort,
and distress. If they occur in situations that are unrelated to concrete reali-
ties, specified objects, or actual instances, there are no clues to resolving
the absurdity; hence the most convenient way to cope with it is to laugh it
off. The third possibility is to direct one’s attention away from the contra-
dictions to the common elements of the items on the list.
There is a tendency in the human mind to impute coherence to what is
incoherent. If it can organize disconnected elements into a coherent story
or state of affairs (i.e., into some schema), it will do so. If not, it will search
for common elements (i.e., generate a category). This is what typically hap-
pens in catalogs: attention is directed away from the discontinuous items
(in this case the baffling logical contradictions) to a category that includes
them, thus reducing the force of bafflement.
Suppose the common denominator of the catalog is “The Contrarious
Passions in a Lover.” The reader can divert the baffling contradictions into
the unitary abstraction of love. Such an abstraction is active in the back of
one’s mind, without preempting everything else, just like emotions, thus
generating an emotional quality attributed to the poem. This tendency to
abstract a common denominator is enhanced by the rhythmic quality of the
poem, the recurrence of the verse line as a versification unit, and the rhyme
pattern, which reinforces a sense of parallelism between the items.
Scholars of the source-hunting tradition have claimed that Petrarch’s
sonnet grew out of Provençal riddles, the solution to which was “love.” One

9 6
of the conventions of the “riddles” genre is to propose an intellectual prob-

lem embodied in a paradoxical statement to which one must find a solu-
tion. The expected solution is a state of affairs in which the incompatible
elements can coexist, in this case, love (Smith 1968: 138). At some variance
with Smith I suggest that in certain circumstances a catalog of paradoxes
may be used as either a riddle or a poem with emotional overtones. Both
options are based on the need to keep the list of paradoxes in an active
state in one’s processing space. To facilitate this, the common denominator
(e.g., love) must be abstracted from the catalog. In the riddle, the common
denominator must be intensified (Freud would say “loaded with cathexis”),
so as to reach full consciousness and replace the paradoxes as a solution
with greater value and priority than the paradoxes themselves. In the poem,
the activity of the common denominator is increased, with the restriction
that it still remain in the background, while the paradoxes themselves are
active in the foreground.
Naturally these riddles can also be seen as sharpened and fossilized
versions of the ambivalence implicit in love. The riddle, however, requires
us to focus our attention on an intellectual solution to a logical prob-
lem involving such mental processes as thought and reasoning rather
than perceptions and emotions. In the sonnets by Petrarch, Wyatt, and
Ronsard, by contrast, the lack of explicit problem-setting, the descrip-
tive tone, and the first-person singular may induce the reader to treat the
catalog of paradoxes as a mental state rather than an intellectual assign-
ment (as shown in the examples below, personal or impersonal gram-
matical constructions may substantially affect the outcome). The catalog
technique has a triple function. It alleviates the load on the cognitive
system; lessens the disquieting feeling induced by the logical contradic-
tions, by directing attention away from the contradictions to their com-
mon denominator, love; and transfers some of the linguistic processing
to the right, emotional hemisphere of the brain. This seems to have a
good fit with the natural capacities and constraints of the human brain. It
is demanded by the natural constraints of memory. As mnemonic device,
these category names or intellectual abstractions are typically active in
the back of one’s mind, so as to prevent usurping the items to be remem-
bered. As a solution to a riddle, on the contrary, they are meant to replace
the items to be remembered. In a poem, the rhythmical recurrence of
ten- or eleven-syllable versification units affects the diffusion of the con-
ceptual abstraction “love” so as to seem to reflect the emotional quality
“love.” We have now come full circle. We started with a disquieting state
of emotional ambivalence, transitioned through sharpening into explicit
07
logical contradiction, and then toned down its baffling effect through the
catalog technique, to finally reach a poetic structure that evokes a less
disquieting emotional quality of love, one that is now perceived in the
poem rather than experienced.
As we have seen, the need to impute coherence may force the reader to
ignore individual contradictions to secure a common superordinate that
can resolve the conflicting statements and account for the co-occurrence of
various “contrarieties”—in this case, “love.” There are instances, however,
in which no such superordinates are available. In this case the human mind
searching for coherence may settle for a formal common denominator, the
category “contradictions.” In such cases, since the incongruities cannot be
resolved by meanings, the only way to cope with the succession of absurdities
is to “laugh it off” and treat them as frivolous trifling. At this juncture, having
recourse to personal or impersonal grammatical constructions may make a
substantial difference.
(b) Villon and Drayton
Now, what happens when the catalog of paradoxes cannot be traced back
to conspicuous ambivalent feelings that have been sharpened and fossil-
ized? When the meanings suggested by the logical contradictions can-
not be ascribed to a superordinate category or state of affairs, the brain,
eager to impose coherence upon the disconnected contraries, will not give
up but may abstract the formal category “contradictions,” which is then
repeated throughout the series. Such a poem would quite plausibly be per-
ceived as frivolous playing around. But sometimes something happens to
the frivolous trifling, and the poem may emerge as somehow more serious.
Consider, for instance, the following two poems:
Idea, by Michael Drayton LXII.

WHEN first I ended, then I first began,
The more I travelled, further from my rest,
Where most I lost, there most of all I wan,
Pinèd with hunger rising from a feast.
Methinks I fly, yet want I legs to go,
Wise in conceit, in act a very sot,
Ravished with joy amid a hell of woe;
What most I seem, that surest am I not.

71
I build my hopes a world above the sky,

Yet with the mole I creep into the earth,
In plenty I am starved with penury,
And yet I surfeit in the greatest dearth;
I have, I want, despair and yet desire,
Burned in a sea of ice and drowned amidst a fire.
Ballade du concours de Blois, by François Villon

Je meurs de seuf auprés de la fontaine,
Chault comme feu et tremble dent à dent,
En mon pays suis en terre loingtaine,
Lez ung brasier frisonne tout ardent,
Nu comme ung ver, vestu en président,
Je riz en pleurs et attens sans espoir,
Confort reprens en triste desespoir,
Je m’esjoys et n’ay plasir aucun,
Puissant je suis sans force et sans pouoir,
Bien recueully, debouté de chascun.
Riens ne m’est seur que la chose incertaine,
Obscur fors ce qui est tout evident,
Doubte ne fais fors en chose certaine,
Scïence tiens a soudain accident,
Je gaigne tout et demeure perdent,
Au point du jour diz «Dieu vous doint bon soir!»,
Gisant envers j’ay grand paeur de chëoir,
J’ay bien de quoy et si n’en ay pas ung,
Eschoicte actens et d’omme ne suis hoir,
De rien n’ay soing, si mectz toute m’atayne
D’acquerir biens et n’y suis pretendent,
Qui mieulx me dit, c’est cil qui plus m’actaine,
Et qui plus vray, lors plus me va bourdent,
Mon ami est qui me faict entendent
D’ung cigne blanc que c’est ung corbeau noir,
Et qui me nuyst, croy qu’i m’ayde a pourvoir,
Bourde, verté, au jour d’uy m’est tout ung,
Je retiens tout, rien ne sçay concepvoir,
27
Prince clement, or vous plaise sçavoir

Que j’entens moult et n’ay sens ne sçavoir;
Parcïal suis, a toutes loys commun.
Que sais je plus? Quoy! les gaiges ravoir,
I die of thirst beside the fountain
I’m hot as fire, I’m shaking tooth on tooth
In my own country I’m in a distant land
Beside the blaze I’m shivering in flames
Naked as a worm, dressed like a president
I laugh in tears and hope in despair
I cheer up in sad hopelessness
I’m joyful and no pleasure’s anywhere
I’m powerful and lack all force and strength
Warmly welcomed, always turned away.
I’m sure of nothing but what is uncertain
Find nothing obscure but the obvious
Doubt nothing but the certainties
Knowledge to me is mere accident
I keep winning and remain the loser
At dawn I say “I bid you good night”
Lying down I’m afraid of falling
I’m so rich I haven’t a penny
I await an inheritance and am no one’s heir
I never work and yet I labor
To acquire goods I don’t even want
Kind words irritate me most
He who speaks true deceives me worst
A friend is someone who makes me think
A white swan is a black crow
The people who harm me think they help
Lies and truth today I see they’re one
I remember everything, my mind’s a blank
Merciful Prince may it please you to know
I understand much and have no wit or learning
I’m biased against all laws impartially
What’s next to do? Redeem my pawned goods again!
Warmly welcomed, always turned away.] (trans. Galway Kinnell)

3 7
Figure 3.1 “Nu comme ung ver, vestu en president.” David Kleinman’s illustration of
Reuven Tsur’s Hebrew translation of François Villon’s poems Danse Macabre.
One conspicuous characteristic of Drayton’s catalog of paradoxes will

become obvious when contrasted to his most famous (and perhaps best)
sonnet:
Farewell to Love
Since there’s no help, come let us kiss and part;
Nay I have done, you get no more of me;
And I am glad, yea, glad, with all my heart,
That thus so cleanly I myself can free;
Shake hands for ever, cancel all our vows,
And when we meet at any time again,
Be it not seen in either of our brows
That we one jot of former love retain.
Now at the last gasp of loves latest breath,
When his pulse failing, passion speechless lies,
When faith is kneeling by his bed of death,
And innocence is closing up his eyes,
Now if thou would’st, when all have given him over,
From death to life thou might’st him yet recover.
Although “Farewell to Love” is placed in a concrete dramatic situation

defined in the here and now, Drayton’s catalog of paradoxes in the “Idea”
sonnet is enumerated “in the abstract,” detached from any concrete situ-
ation. Let us compare the first lines of Drayton’s and Villon’s catalogs of
47
paradoxes. Consider Drayton’s “When first I ended, then I first began.” This
line contains a straightforward logical contradiction between two incompat-
ible statements that are unrelated to any concrete realities, specified object,
or actual instance. Consequently, there are no clues for, or constraints on,
its interpretation. There is not sufficient information to decide whether they
are relevant to the speaker’s well-being. Villon’s line “I die of thirst beside
the fountain,” by contrast, contains an indirect or empirical opposition. In
fact, the inferences from being “beside the fountain” and from “dying of
thirst” conflict. The opposing terms are actually compatible, and they occur
within a concrete situation. It could be literally true as, for instance, in
Water, water every where,

Nor any drop to drink.
Or it may suggest another tragic situation, as when one reaches the foun-
tain but dies before being able to drink, and so forth. This situation is clearly
relevant to the speaker’s well-being and suggests a tragic flavor. The effect
is not dependent on Villon’s personal disposition. The line was given to him
by Charles, Duke of Orléans, himself a fine poet, at a poetry contest, where
eleven contestants began their entries with this line. Having set the tone, the
next three paradoxes can also be construed to conform to a tragic experience.
Charles d’Orléans’s own entry went as follows:
Je meurs de soif en couste la fontaine;

Tremblant de froit ou feu des amoureux;
Aveugle suis, et si les autres maine;
Povre de sens, entre saichans l’un d’eulx;
Trop negligent, en vain souvent songneux;
C’est de mon fait une chose faiee,
En bien et mal par Fortune menee.
[I die of thirst beside the fountain
Shaking from cold or the fire of lovers;
I am blind and yet guide the others;
I am weak of mind, among men of wisdom one of them;
Too negligent, often cautious in vain,
I have been made a spirit,
Led by fortune for better or for worse.] (my translation)
While Villon’s shivering with cold and the heat of the fire are both literal,
Charles d’Orléans has “Shaking from cold or the fire of lovers,” choosing

5 7
the Petrarchan solution for his contrarieties. This may bestow some figu-
rative meaning on the first line, for example, I am near the object of my
yearning, but I cannot reach her.
Goethe uses the dying-of-thirst-beside-the-fountain metaphor in a
very un-Petrarchan context. In his poem “Das Tagebuch” (The Diary) he
recounts how he found himself in bed with a beautiful virgin, but unex-
pectedly, presumably for moral scruples, he had no erection:
Ohnmächtig jener, dem sie nichts verwehrte.

Vom Schlangenbisse fällt zunächst der Quelle
Ein Wandrer so, den schon der Durst verzehrte.
[Powerless, he whom she refused nothing,
Snakebitten falls beside the fountain
As a wanderer already consumed by thirst.]
Though Robert Louis Stevenson regarded this as one of Villon’s poorest

poems, Charles d’Orléans is reported to have admired it very much. And
today it has exceptionally great appeal; some readers feel it to be particu-
larly modern. This is most remarkable in view of the relative obscurity of
Drayton’s sonnet. I myself have come under the spell of Villon’s “Ballade,”
although I feel that Drayton’s “Idea” sonnet is a successful exercise, if far from
a masterpiece. I said above that the first line of Drayton’s poem contains a
straightforward logical contradiction between two incompatible statements
that is removed from any concrete realities or actual instances. The subse-
quent lines, however, abound in visual imagery and concrete instances.
Though I have no way to demonstrate this, I have the impression that the
poet deliberately chose as diverse images as possible, so as to render his cata-
log as incoherent as possible. The unity of the poem comes from the “punch
line,” an exceptionally forceful closing couplet. A closer look, however, shows
that Villon’s poem also abounds in conspicuously disconnected images.
So, how do we make sense of these two poems? Both poems can be read
as unresolved paradoxes, which would yield a frivolous, trifling interpre-
tation. As such, they would be entirely satisfactory. It would, however,
be quite difficult to exclude a reading that imputes a certain kind of com-
plex personality into the series of contradictions. Again, I have the feeling
that it would be more difficult to exclude this for Villon’s ballade than for
Drayton’s sonnet. Consider Drayton’s line “The more I travelled, further
from my rest.” It is not really paradoxical, and it could be literally true. But
the reader, eager to render the speaker consistent, even if not meaningful,
readily assumes that the aim of the journey is to end one’s travails and
67
attain a state of tranquility. But even so, this paradoxical statement can be
construed as, for instance, being confronted with a particularly inquisitive
spirit. Thus the image of a complex personality begins to emerge. Then one
may attempt to relate such statements as “Wise in conceit, in act a very sot”
to this emerging personality, which, again, could be literally true.
Villon’s ballade begins with the line “I die of thirst beside the foun-
tain,” where the contradiction is not between two abstract statements but
between a scene term and the speaker’s action. This can be construed as sug-
gesting a tragic situation, followed by three lines in which the paradoxes can
be construed as suggesting some intense physical or psychological anguish.
This sets the scene for the emergence of an intense and coherent human
experience, suggesting, in turn, a complex human personality. Within this
framework, such lines as “I keep winning and remain the loser” suggest a
sophisticated but plausible character. To this one may add the grotesque
image of “Naked as a worm, dressed like a president” or the self-sarcasm of
the misfit: “At dawn I say ‘I bid you good night.’ ” The better the integration
of these features into a unified character, the more credible the character.
The more diversified the features suggested by the paradoxes—assuming
that unity is taken care of—the more complex the emerging character.
Is there any evidence that readers go such a long way to integrate such
a diversity of features? And if so, is there anything in the texts themselves
that would advance or obstruct such an integration? Meir Sternberg (1976)
quotes a series of experiments by Luchins and others on primacy and recency
effects that may clarify two aspects of this issue. The experimenters com-
posed two diametrically opposed passages describing a hypothetical per-
son, Jim. One passage described him as friendly and extroverted; another,
as unfriendly and introverted. Then they combined the two passages in two
different ways: the friendly paragraph first or the unfriendly paragraph
first. The subjects were asked to give their impression of the person and
were divided into two groups, one that read the friendly-paragraph-first
version and another that read the unfriendly-paragraph-first version. The
results showed a strong primacy effect. Subjects tended to interpret the
first part of the description as the real character of the person described.3
Surprisingly enough, in the interviews, subjects in both conditions reported
3. Recently I served as the chairperson on a promotion committee. One referee wrote

a two-paragraph letter. In the first paragraph he pointed out a long list of flaws in the
candidate’s work. The second paragraph began with “If, despite all this I recommend
the candidate for promotion, it is because . . . ,” and here came a long list of praise.
I phoned the referee and asked him to write exactly the same thing but in the reverse
order. He did so but added a short, innocent-looking clause, which he knew would fail
the candidate.

7
that they had not noticed the discrepancy in the description. Sternberg
accounts for this curious fact as follows:
Due to the successive order of presentation, the first block was read with an open
mind, while the interpretation of the second—in itself as weighty—was deci-
sively conditioned and colored by the anterior, homogeneous primacy effect. In
other words, the leading block established a perceptual set serving as a frame of
reference to which subsequent information was subordinated as far as possible.
In each case, accordingly, the leading block was taken to represent the “real”
Jim, the “essential nature” of Jim, while the second was taken to describe excep-
tional behavior, which is to be explained away in terms of temporary variations
in mood or circumstances—in short, as a mere qualification of the previously
established conception of character. (1976: 298)
These experiments suggest that readers will go a long way to impose

coherence not only to inconsistent but even to diametrically opposite infor-
mation. This would imply that Drayton’s “Idea” sonnet and Villon’s ballade
ought to be perceived as suggesting equally integrated characters. I can
think of three ways to point out factors that would be conducive to inte-
gration. First, one could meticulously analyze the specific contrarieties in
Villon’s and Drayton’s poems and see whether there is something in them
that would be more or less conducive to integration. Second, one might look
at the beginning of the poems for elements whose primacy effect would abet
integration of the traits or not. In the present case we saw that Drayton’s
first line contains a straightforward logical contradiction between two
incompatible statements divorced from any concrete reality, with no clues
for its interpretation. Nor is there sufficient information to decide whether
they are applicable to the speaker’s well-being. Villon’s poem, by contrast,
begins with four consecutive lines that may suggest an emotionally loaded
tragic situation that relates directly to the speaker’s well-being. Thus, the
primacy effect strongly favors the construal of a coherent situation of great
human interest in Villon’s ballade, while it seems to discourage it in the case
of Drayton’s sonnet.
Finally, euphony and other prosodic and syntactic features may be very
effective here. Consider the first two lines of the ballade for the Blois con-
test. In “Je meurs de seuf” the vowel -eu is repeated. Feu in the second line
contains the last two phonemes of seuf, in reverse order. Fontaine also con-
tains the consonant f. Repetition of the sounds of dent is entailed by the
repetition of the word, as demanded by the figurative language. The nasal
vowel of dent is foreshadowed by the first vowel of tremble and another
nasal vowel in the first syllable of fontaine.
87
A word must be said about Villon’s rhymes. As the Romantic poets argued,
rhyme is unity in diversity. The more similar the sounds, and the more dif-
ferent the meaning structure of the rhyming words, the more effective the
rhyme. The ballade form has a sophisticated rhyme pattern. In the present
instance, the a and d rhymes of the stanza have two members each; the b and
c rhymes have three. To make things more difficult, all the later stanzas rhyme
with the rhymes set by the first stanza. As a result, poets take the liberty of
rhyming similar grammatical structures (what Jakobson [1960] calls “gram-
matical rhymes”) or even have recourse to the same words with different pre-
fixes. A similar problem arises in the quassida of medieval Hebrew and Arabic
poetry, where the monorhyme needs to rhyme with an indefinite number of
similar-ended words. As to rhyme quality, medieval Hebrew poets explicitly dis-
tinguished between “passable,” “becoming,” and “excellent” rhymes, according
to the number of shared phonemes in the rhyme words. Consider the following
rhymes in Villon’s ballade: espoir–desespoir, aucun–chascun, incertaine–certaine.
In view of such monotony, the complexity of the a and b rhymes in the first
stanza will be all the more salient, for instance. Fontaine is a noun, whereas
loingtaine is an adjective, displaying a contrast on the part-of-speech level. At
the same time, the two words have four consecutive speech sounds in common.
I have already mentioned that the sequence dent à dent necessarily repeats the
same speech sounds and that the nasal feature of the vowels is foreshadowed
in tremble and fontaine. Moreover, the alternating a-b-a-b rhymes have a special
phonetic relationship, which Kenneth Burke called “colliteration” (1957: 369;
see chapter 10). The t of fountaine and the d of dent are cognates: the d is a
voiced t. Whereas fountaine ends with a full consonant, a (nasal) n, the nasal
vowel ẽ is generated by an attenuated n. Thus, in the a-b-a-b rhyme, two sets
of cognate but different speech sounds alternate, yielding a particularly eupho-
nious effect. The three b rhyme words, dent–ardent–président, rhyme a nasal
vowel preceded by a d. Dent is a unitary word, a noun; ardent and président
end with the formative -ent, which derives a noun or an adjective from a verb.
In this case, ardent is an adjective, and président is a noun. What is more, as
far as I can judge as a foreigner, and from another century, président is also
perceived as a unitary word. The adjective ardent intervenes between the two
nouns in the rhyme pattern. Furthermore, the rhymed syllables of both the a
and b rhymes (-taine and -dent) preserve all their speech sounds throughout
the three stanzas. This is not necessarily so in the poetry of the period.
Below I quote the opening stanza of Alain Chartier’s ballade of proverbs
and the opening stanza of Villon’s parody of that poem from Jane Taylor
(2001). Taylor points out significant similarities between them, including
the identical rhyme patterns. By identical rhyme patterns, I suppose she
also means identical specific rhyme endings. At this point, however, I want

9 7
to point out the difference in the quality of the rhymes. In both stanzas, the
a rhymes (villain–plain, fain–faing) are unitary, suffixless words (but Villon
happens to rhyme a pair of homonyms with very different meanings). The
d rhymes (including the refrain) are not very exciting: both poets rhyme
the grammatical suffix -eux. This is not very respectable or euphonious. But
while Chartier rhymes a single phoneme (d’amoureux → joieux), Villon adds
a common preceding consonant: r (peureux → qu’amoureux). This r persists
throughout Villon’s poem. The b rhymes in these two poems consist of mas-
culine /i/ rhymes; the c rhymes, of a pair of feminine endings (/i/ plus e muet).
But while in Chartier’s stanza a different consonant precedes the vowels in
each member (enrichi → amy → jalousie → seignori), in Villon’s stanza in all
members the vowels are preceded by a nasal consonant (in the b rhyme by
/m/, in the c rhyme by an /n/ and a palatalized /n/: d’ennemy → endormy →
félonnie → regnie). In the ensuing stanzas as well Villon assiduously inserts
pairs of supporting consonants before the final vowel. In the third stanza
and the “envoi” there are six c rhymes, in all of which there is a supporting
/d/. In the last three paragraphs subtle euphony plays a substantial part in
the appeal of some of Villon’s ballades, and this degree of euphony should by
no means be taken for granted in the poetry of the period.
This rhyme quality is probably not accidental in Villon. He seems to
have been fascinated by rhyme words that are exceptionally similar from
the phonetic point of view and exceptionally dissimilar from the semantic
point of view, as the following rondeau may testify:
Jenin l’Avenu,
Va-t-en aux estuves;
Et toy la venu,
Jenin l’Avenu,
Si te lave nu
Et tu baigne es cuves.
Jenin l’Avenu,
Va-t-en aux estuves.
[Jenin Avenue,
Go there to the heat chambers;
And you came there,
Jenin Avenue,
If you wash yourself naked
And you bathe in tubs,
Jenin Avenue,
Go there to the heat chambers.] (my translation)
08
Villon rhymes here three homonyms in different senses: l’Avenu (the

Avenue), la venu (came there), and lave nu (wash naked).
(c) Trif ling/Parody
Returning to meanings, we find a very different specimen in Villon’s

other catalog of contradictions, “Ballade des contre-vérités” (Ballade of
Countertruths). Here any possible personal element is drastically elimi-
nated. The statements are highly generalized in an “all-exclusive but
one” format. Not only is the first-person singular pronoun meticulously
avoided in this ballade, but there are only two finite verbs before the
envoi, both with the same impersonal construction of “Il n’est.” All the
rest is a lengthy elaboration of the negation of the form: “No a but b / No
c but d” and so forth. The contradictions should, conspicuously, not be
resolved. When the Bible, for instance, says, “He who spares his rod hates
his son” (that is, he who does not beat his son hates him), the resolution
of the paradox is almost self-evident. “The rod” is a metonymy for punish-
ment; hates refers to wishing or doing wrong to someone: that is, he who
does not punish his son when he deserves it, does him wrong in the long
run. If we take a comparable paradox from Villon’s ballade, for example,
“No service but from an enemy,” no such construal is readily available.
The paradox “There is no conceiving but in the bathtub” is conspicuously
untrue and makes fun of women who claim that they get pregnant from
sperm in the water. What is more, the rigid formula “No a but b” repeated
throughout the ballade emphasizes that the common denominator here
is not a notion that may resolve the paradox but a logical structure, focus-
ing on, rather than drawing attention away from, the baffling contradic-
tions. So, the only course left for the reader is to read the catalog for the
fun of it:
Ballade des contre-vérités

Il n’est soin que quand on a faim
Ne service que d’ennemi,
Ne mâcher qu’un botel de fain,
Ne fort guet que d’homme endormi,
Ne clémence que félonie,
N’assurance que de peureux,
Ne foi que d’homme qui renie,
Ne bien conseillé qu’amoureux.

81
Il n’est engendrement qu’en boin

Ne bon bruit que d’homme banni,
Ne ris qu’après un coup de poing,
Ne lotz que dettes mettre en ni,
Ne vraie amour qu’en flatterie,
N’encontre que de malheureux,
Ne vrai rapport que menterie,
Ne tel repos que vivre en soin,
N’honneur porter que dire: “Fi!,”
Ne soi vanter que de faux coin,
Ne santé que d’homme bouffi,
Ne haut vouloir que couardie,
Ne conseil que de furieux,
Ne douceur qu’en femme étourdie,
Voulez-vous que verté vous die?
Il n’est jouer qu’en maladie,
Lettre vraie qu’en tragédie,
Lâche homme que chevalereux,
Orrible son que mélodie,
[There is no care but hunger
No service but from an enemy
Nothing chewy but a bale of hay
No reliable watchman but a man asleep
No clemency but felony
No confidence but in the frightened
No good faith but in forswearing
No well-advised but the enamored.
There is no conceiving but in the bathtub
No good reputation but that of an exile
No laughter but after receiving a punch
No merit but denial of debts
No true love but flattery
No meetings but unhappy ones
No true rapport but in lies
28
No rest like a life of worry

No respect but saying “Fie!”
No showing off but with false coins
No good health but of a man with dropsy
No high resolve but cowardice
No good advice but from the furious
No sweetness as in a vociferous wife
Verity, are you ready to hear?
In sickness alone is there joy
Literal truth only in tragedy
Lack of courage only in braveness
Offensive sound only in melody
No well-advised but the enamored.] (my translation)
I used to think that this ballade demanded a frivolously trifling read-

ing. However, it turns out that this fun is subordinated to entertain-
ment of a higher order. Villon scholars claim that this ballade is a parody
of a ballade on catalogs of proverbs by Alain Chartier (see, e.g., Taylor
2001: 146; Pinkernell n.d.). By putting the two poems side by side, Taylor
demonstrates their unmistakable similarity (and, by the same token,
their difference):
Alain Chartier François Villon

Il n’est danger que de villain, Il n’est soing que quant on a fain
N’orgueil que de povre enrichi, Ne service que d’ennemy,
Ne si seur chemin que le plain, Ne mascher qu’un botel de faing,
Ne secours que de vray amy, Ne fort guet que d’homme endormy,
Ne desespoir que jalousie, Ne clémence que félonnie,
Ne hault vouloir que d’amoureux, N’asseurance que de peureux,
Ne paistre qu’en grant seignorie, Ne foy que d’homme qui regnie,
Ne chere que d’omme joieux. Ne bien conseillé qu’amoureux.
(Taylor 2001)
The identical rhyme pattern, per se, of the two poems is not very remark-
able in my mind, because this is one of the most widespread rhyme patterns
in the ballade form, but the identical specific rhyme endings, by contrast,
are. Likewise, the endlessly repeated formula “No a but b” is more than tell-
ing. Moreover, consider Chartier’s line “Ne hault vouloir que d’amoureux.”
Its second half recurs in Villon’s refrain “Ne bien conseillé qu’amoureux.”

3 8
Its first part occurs in Villon’s line “Ne haut vouloir que couardie,” amplify-
ing the ironic twist.
“Villon’s ‘proverbs,’ ” Taylor writes, “though just as lapidary, just as
respectable-looking as Alain’s, are pseudo-proverbs whose sense is, to a
moment’s thought, patently absurd: they replace the irritating predictabil-
ity of Chartier’s poem with proverbs which are not trite—but not ‘true.’
What he demonstrates, in other words, is the sheer fatuity of Alain’s exer-
cise.” “But a string of proverbs like Alain’s,” she continues, “by definition has
no function; a poem consisting solely of proverbs subordinates universal
wisdom to the adventitious constraints of rhyme and metre” (2001: 146).4
Taylor ignores the fact here that Villon himself wrote two ballades of prov-
erbs just as “irritatingly predictable” as Chartier’s, as well as the fact that
in his “Ballade du concours de Blois” he used the technique of catalogs of
contradictions to a very different purpose and effect. Nevertheless, her
conception of this parody is well taken. I would add that Villon does not
explicitly criticize Chartier’s platitudes and complacent values of “respect-
able” people, unlike what he does in his parody on the pastoral values of
the “Franc Gontier” poems. But if you put the two poems side by side, this
intertextual relationship casts an ironic light on Chartier’s ballade and
renders it ridiculous (very much in the manner in which Lewis Carroll’s
“Father William” renders Robert Southey’s “Father William” ridiculous).5
4. In a Marcello Mastroianni film (Stanno tutti bene) there is an anecdote in the fam-
ily about an old winemaker who, on his deathbed, reveals a professional secret to his
sons: wine can be made of grapes too, and then it has a special flavor. In the same vein
I want to reveal that sometimes people write poetry not to instruct but for fun. People
write and read ballades that contain catalogs of proverbs not in order to make readers/
listeners wiser but to cause them to have fun. Presumably, the reader knows all these
proverbs by heart. The ballade form has a complicated rhyme scheme but can despite all
this be very euphonious. Readers, for their part, may derive pleasure from the fact that
everything falls into place. The more rigorous the formal constraints, the greater the plea-
sure when they are overcome. That is how I propose to (mis)read Taylor’s comment: “a
poem consisting solely of proverbs subordinates universal wisdom to the adventitious
constraints of rhyme and metre.” Alain Chartier’s ballade is inferior to Villon’s not in that
he “subordinates universal wisdom to the adventitious constraints of rhyme and metre”
but in that “the adventitious constraints of rhyme” are less rigorous in it than in Villon’s
ballades and the solution of the problem is, therefore, less pleasurable (cf. c hapter 7).
5. In what appears to be a Villon entry in an online encyclopedia, Professor Gert
Pinkernell writes: “Besides, V. apparently had a love affair at that time, because the
refrain of the ballade repeats ‘No well-advised but the enamored’ ” (n.d.; my transla-
tion). This is a fine example of the biographical fallacy refuted above by Dan Pagis that
involves mistaking the exemplary nature of the genre for personal experience. This
refrain is based on the reversal of a commonplace of the time, is conspicuously in line
with the other countertruths of the list, and does not indicate an actual love affair any
more than his refrain “But where are the snows of yesteryear?” would indicate that
Villon went on a skiing trip the preceding year.
48
SUMMARY AND CONCLUSIONS
In this chapter I have focused on certain issues from the broad perspec-
tive of medieval and Renaissance poetry in general. My argument has
three stages. First, in a highly conventional kind of poetry it is dangerous
to make inferences from what the poet writes to what the poet feels (or
thinks). In this respect, I refer to the “biographical fallacy,” (mis)applied
to al-Mutanabbi’s, Moses Ibn Ezra’s, and François Villon’s poetry. I warn
against applying psychological hypotheses to a poet’s empirical personal-
ity through his or her use of poetic conventions. In this respect, in many
genres, when poets say “I,” they do not even necessarily mean themselves
but, rather, a certain exemplary “I,” Everyman. In this case, the first-person
singular is meant to induce the audience to activate certain processing
strategies, to generate certain kinds of effects. Impersonal constructions
serve to induce different processing strategies, with different effects.
This does not imply that psychological theories are irrelevant to lit-
erary research. We, however, need to carefully scrutinize what we have
explained using these psychological theories. In a highly conventional kind
of poetry one may learn very little about the empirical personality of the
poet. I strongly suspect that in this period, a poet’s recourse to this or that
genre reflects a fashion rather than stages in the development of his or
her personality. But we may learn a lot about the origin of conventions.
With reference to such genres as the poetry of boasting and the panegyric,
for instance, Maslow’s theory can describe the psychological processes that
have fossilized into these genre conventions.
Second, following Ehrenzweig and D’Andrade, I have tried to explore the
nature of poetic conventions and their evolution. As Laozi said, every jour-
ney, even the longest one, must begin with a single step. Likewise, every
poetic tradition, even the longest and widest one, must begin with one
poet composing one poem, or one verse line, for one particular audience or,
perhaps, several poets, each composing one poem for the same audience or
several different audiences. How, from such humble beginnings, did large
poetic traditions emerge? In a world in which there was no radio or televi-
sion, no Internet, no airplanes, only horses, camels, and sailboats? I am not
proposing an actual historic process but, rather, a hypothesis that must be
verified against a wide variety of actual processes.
On this issue I have adopted D’Andrade’s position: “In the process of
repeated social transmission, cultural programs come to take forms which
have a good fit to the natural capacities and constraints of the human
brain.” One promising beginning for such a study would be, with the neces-
sary changes, Sir Frederick Bartlett’s classical experiment with the Eskimo

5 8
story “The War of the Ghosts.” Bartlett’s purpose in this study was slightly
different. He tried to determine how a story for which we have no mental
schemata assumes, through repeated transmission, a good fit to the men-
tal schemata we entertain. When the story achieved a good fit to a mental
schema prevalent in Western culture, the story stopped changing in sub-
sequent transmissions (see chapter 2). I take up the process at a relatively
late stage, when defense mechanisms are used against expressive devices,
which are turned, according to Ehrenzweig, into style. In this respect,
I claim, artistic conventions and ornaments are fossilized psychological
processes.
Third, I illustrate this model with reference to a widespread convention
in medieval and Renaissance poetry: catalogs of incongruous or incom-
patible statements. I start at a point well known to psychoanalytic the-
ory: ambivalence. Ambivalence is a potential source of undesirable stress.
Dreams and fairy tales handle ambivalence by splitting the object of ambiv-
alent attitudes into two opposing figures, for example, a good mother and
a wicked stepmother. It would appear that a straightforward wicked step-
mother is less threatening than an ambivalent attitude. The poetic conven-
tion of contradictions originates not in a split of the ambivalent attitude
but in applying to it the cognitive mechanism of sharpening,6 thus generat-
ing explicit contrarieties. Again, baffling contradictions seem to be less dis-
tressing than ambivalent attitudes, but they are still threatening. One way
to defuse the disquieting element in contradictions is to generate a catalog
of contradictions. The constraints of human memory induce the mind to
abstract a common category from parallel items. At the same time, this
helps shift attention from the contradictions to their common element or
superordinate category. The most widespread variety of this convention is
that in which the common superordinate category is “love,” which is active
in the back of the mind, without preempting everything else. At the same
time, the repetition of parallel items generates a rhythm that amplifies and
diffuses the superordinate category “love,” and via interaction with other
poetic devices, such as the use of the first-person singular, the mnemonic
device is turned into an emotional quality of the poem. It is claimed that
the ambivalence of love and the need to remove the disquieting element has
a good fit to the capacities and constraints of the human brain. I suggest
that even catalogs of contradictions that historically do not originate in
the ambivalent emotions of love may eventually be accommodated in this
tradition and thus assume a good fit to the natural capacities of the human
6. See chapter 1.
68
brain. There are, however, instances in which the “love” solution simply
does not work. Then, to get rid of the disturbing element of the unresolved
contradictions, the human mind may try other strategies, including the
construal of some complex personality roughly compatible with the list
of paradoxes. And if this fails as well, we can still laugh it off and treat it
as frivolous trifling, making fun. We have seen one instance in which this
solution is subordinated to a higher degree of making fun, namely, parody.

7 8
CHAPTER 4
Frozen Formulae and Expressive Force

The Ballad “Edward”
CONVENTIONAL STYLE AND INDIVIDUAL EXPRESSION

IN “EDWARD”
The quintessence of fossilized cognitive devices turned into ornament that

yield a literary masterpiece is the Scottish ballad “Edward.” In what fol-
lows I take a close look at this work. This chapter is based on two basic
assumptions. One of the central theses of this book, derived from Anton
Ehrenzweig (1965), is the view that in order to avoid the more “dangerous”
expressive elements of an artistic work, we tend to suppress expressive
features by formalizing them—we turn them into characteristic features
of a style, into harmless ornament. The other is that pigeonholing gives
certainty but no insights or sensitive responses to a poem. This can conve-
niently be explained in terms of the mechanisms assumed by Structuralist
poetics to underlie the use of rhetorical figures:
The repertoire of rhetorical figures serves as a set of instructions which readers

can apply when they encounter a problem in the text, though in some cases it
is not so much the operations that are important as the reassurance that what
seems odd is perfectly acceptable since it is figurative expression of some kind
and therefore capable of being understood. If one knows that hyperbole, litotes,
zeugma, syllepsis, oxymoron, paradox and irony are possible, one will not be
surprised to find words or phrases that must be dealt with in the ways that these
figures suggest. (Culler 1975: 181)
8
Paradoxically enough, this passage gives us a clue not only to how “the rep-
ertoire of rhetorical figures” provides insight but also to how “pigeonholing
gives certainty, but no insight” (de Mourgues 1953). The crucial clause seems
to be “not so much the operations that are important as the reassurance.”
However, “instructions” and “reassurance” are not and should not be taken to
be two independent factors, as this clause implies. “Reassurance” should not be
considered as a goal in itself but, rather, as a psychological precondition under
which the instructions can be carried out. Pigeonholing occurs when readers
or critics are content with “the reassurance that what seems odd is perfectly
acceptable since it is figurative expression,” or a convention of a genre, or a
symptom of a psychological syndrome, but have no expectation of getting (or
executing) any further instructions. For these readers, carrying out instruc-
tions could at best lead to the final “reassurance that what seems odd is per-
fectly acceptable” and at worst might unnecessarily complicate things or even
lead to new ambiguities. So why bother? In George Klein’s (1970: 134–136)
terminology, they tend to level out differences between unique works. In such
cases, explication of the poem may be of decisive importance.
For instance, we are taught that the key characteristics of the ballad style
are repetition (often merely decorative) and ellipsis (often quite easily com-
pleted). When we ask, “What does this repetition or that ellipsis contribute
to the ballad?” we often are given the diagnosis, “It is a characteristic of bal-
lad style,” or, “It contributes to the balladic atmosphere.” Whenever there
is threefold repetition, one can expect to be told that “it contributes to the
popular-ballad style.” In other words, repetition and ellipsis are treated as
merely symptoms of a style.
I would by no means deny that repetition does contribute to the balladic
character of a ballad. I would even admit that these characteristics give a
certain charm and naïveté to the ballad. Nevertheless I would submit that
this naïveté is sometimes the consequence of repressing some utterly non-
naïve and expressive feature of the poem.
Let us have a look at “Edward”:
1
“Why dois your brand sae drap wi bluid,
Edward, Edward,
And why so sad gang yee O?”
“O I hae killed my hauke sae guid,
Mither, mither,
O I hae killed my hauke sae guid,
And I had nae mair bot hee O.”

9 8
2
“Your Haukis bluid was nevir sae reid,
Edward, Edward,
Your Haukis bluid was nevir sae reid,
My dear son I tell thee O.”
“O I hae killed my reid-roan steid,
Mither, mither,
O I hae killed my reid-roan steid,
That erst was sae fair and frie O.”
3
“Your steid was auld, and ye hae gat mair,
Edward, Edward,
Your steid was auld, and ye hae gat mair,
Sum other dule ye drie O.”
“O I hae killed my fadir deir,
Mither, mither,
O I hae killed my fadir deir,
Alas and wae is mee O!”
4
“And whatten penance wul ye drie for that,
Edward, Edward,
And whatten penance wul ye drie for that,
My deir son, now tell mee O.”
“Ile set my feit in yonder boat,
Mither, mither,
Ile set my feit in yonder boat,
And Ile fare ovir the sea O.”
5
“And what wul ye doe wi your towirs and your ha,
Edward, Edward,
And what wul ye doe wi your towirs and your ha,
That were sae fair to see O?”
“Ile let thame stand tul they doun fa,
Mither, mither,
Ile let thame stand tul they doun fa,
For here nevir mair maun I bee O.”
6
“And what wul ye leive to your bairns and your wife,
Edward, Edward,
F r o z e n F or m u l a e a n d E x p r e s s i v e F or c e [ 89 ]
09
And what wul ye leive to your bairns and your wife,

When ye gang ovir the sea O?”
“The warldis room, late them beg thrae life,
Mither, mither,
The warldis room, late them beg thrae life,
For thame nevir mair wul I see O.”
7
“And what wul ye leive to your ain mither deir,
Edward, Edward,
And what wul ye leive to your ain mither deir,
My deir son, now tell mee O.”
“The curse of hell frae me sall ye beir,
Mither, mither,
The curse of hell frae me sall ye beir,
Sic counsels ye gave to me O.” (in Kinsley 1969: 239)
All the repetitive features of this poem, by themselves, can be considered

markedly expressive. The mere repetition of a line can be very emphatic, as
is the doubled apostrophe, such as “Edward, Edward” and “Mither, mither,”
or the addition of the interjection “O” (in the middle and last line of each
stanza). All these features tend to lose their emphasis when they occur reg-
ularly, in their preassigned places. We tend not to assign thematic empha-
sis to them but, rather, see them as ballad-style features (since we do not
perceive the poem at once, like a picture, but, instead, in a chronological
sequence, we may attribute more emphatic value to them in the first stanza
than in the subsequent ones).
The final line of the stanza can have two opposite effects: it can either
round off the stanza in a soft ending or cut it off with a sharp, pointed,
epigrammatic ending. If we compare the last lines of the first three stanzas,
two of them (“And I had nae mair bot hee O” and “That erst was sae fair and
frie O”) have a rounding effect, in harmony with the other repeated lines,
whereas the third one (“Alas and wae is mee O!”) has a sharp contrasting
quality. The same may be said of the last three stanzas.
There are a number of small differences in the individual stanzas, which
together contribute to the impressions of rounding or contrasting effects:
* The shocking revelation at the end of stanza 3 and the brutal statement
at the end of stanza 7 contribute to the quality of sharpness.
* All the last lines, except for in stanzas 3 and 7, are syndetic; all the syndetic
phrases have a connecting character (that, for, and), never contrasting.

91
* The abrupt, asyndetic lines at the end of the third and the seventh
stanza are perceived as indicating a break. In fact, the second to the last
line of the stanza becomes “sharper” when it is end-stopped, in virtue of
the asyndetic new start in the next line.
* The last line of the ballad reveals a hitherto concealed, highly essential
piece of information. The repetitions in this stanza, although antici-
pated, are highly functional.
* “The curse of hell frae me sall ye beir” strikes the reader by its unex-
pected character. The apostrophe “Mither, mither” has a double func-
tion here: (a) In this unexpected context, it regains some of its original
emotive quality; (b) it defers the solution to this striking riddle. The rep-
etition, in spite of its expectedness, further stretches the reader’s antici-
pation, until the last line finally conveys the unexpected, long-awaited
solution.
* In the last line of the third stanza, “Alas and wae is mee O!” the formu-
laic “O” occurs after a sequence of two parallel interjections, reviving its
meaning and restoring its high emphasis.
This ballad has action in the Aristotelian sense: It has a peripeteia (rever-
sal) with an anagnorisis (revelation), or to be precise, two revelations (one
at the end of the third stanza and one at the end of the seventh); that is,
there is a shift from one extreme situation to its opposite, in this case from
ignorance to knowledge.
The ballad itself is seriously limited in its expressive resources. It consists
of a series of questions and answers, expressed in a rigidly formulaic way
that leaves virtually no room for dramatic manipulation. Nevertheless the
peripeteia is brought about with pointed theatricality. There is no continu-
ous action but, rather, a series of fragmentary, abrupt, formulaic dialogues;
it is up to the reader to complete the situation and generate hypotheses to
make a coherent whole out of the fragments.
Stanza 1, besides advancing the action through dialogue, sets the stage
for the action and provides the necessary information, simultaneously
serving as exposition. We learn that the dialogue occurs between mother
and son. Metonymies bring information to the foreground that lingers in
the background (in the immediate past or within the protagonist). From
“your brand sae drap wi bluid,” the reader infers that Edward has just
committed a murder. Sad in medieval usage frequently meant “heavy,
firm, steadfast.” Consequently, “so sad gang yee” is primarily a physical
metonymy, indicating Edward’s state of mind. Both together awaken the
reader’s curiosity and seem to be sufficient motivation for the mother’s
inquisitiveness.
29
As such, the rigid formula seriously impairs the possibilities of expres-

sion. The expressive elements stop being expressive and are turned into
mere symptoms of the ballad style. On the other hand, the use of formu-
lae provides a highly effective means of expression (if duly exploited)—
namely, significant variation: “Significant variation is a type of structure in
which the effect is secured by an alteration in a pattern of action which has
become familiar by repetition” (Brooks and Heilman 1966: 51).
In “Edward,” the pattern is set in the first two stanzas by the moth-
er’s questions, Edward’s answers, and again, the mother rejecting those
answers. The deviation from the pattern occurs in stanza 3, where the
mother does not reject the answer (prompting the reader to believe that
this time the truth has been told). The same holds true, in a more sophis-
ticated manner, for the last three stanzas. Here, it is more difficult “to
alter the pattern” since the reader has two contradictory expectations (so
it seems more difficult to disappoint both). Here, again, two stanzas set
the dominant pattern. The mother asks Edward, “And what wul ye doe wi
your towirs and your ha”? “And what wul ye leive to your bairns and your
wife”? The answer to both questions is, virtually, “I don’t care.” This pat-
tern requires the same answer to the third question too: “And what wul ye
leive to your ain mither deir”? However, by now a rival pattern has been
established: The third stanza has to end in another way; one expects that
his beloved mother would be of more concern to Edward, more than his
towers, wife, and children. The end of the ballad frustrates both expecta-
tions: this time, Edward does not answer, “I don’t care,” but, undermin-
ing all previous expectations, shows concern for his mother that is unlike
anything one might predict. Therefore, the revelation is both expected and
unexpected at one and the same time and is revealed to the reader as well
as to the mother (although the information the mother gets here is of quite
a different quality and weight than that offered to the reader).
One can, therefore, give two different answers—one “stylistic,” the
other “expressive”—to the question “Why are there exactly three parallel
questions and answers?” The first answer states that exactly three repeti-
tions is what the popular formula requires; the other suggests that three
repetitions are required since this is the smallest unit in which significant
variation may be effective, that is, in which a pattern can be established
and altered.
The differences between a “stylistic” solution and an “expressive” solu-
tion to questions of text interpretation can be highlighted by a closer
analysis of the ballad. Let us begin by asking once again why the mother
is so inquisitive. Obviously, from a stylistic point of view, this is her part
in the formula. But there is more to it. Let us assume that the readers do

3 9
not know the end of the ballad. They observe the mother’s steady ques-
tioning. During the first stanza, they could conclude that these are the
questions of a loving mother, troubled by her son’s unusual appearance.
During the second and third stanza, readers are led to modify this appre-
ciation somewhat: The mother makes statements showing that she doesn’t
believe Edward, and Edward produces one pretext after the other, but his
mother is cleverer than that; she presses him with his back to the wall.
Revelation I (as opposed to the preceding pretexts) throws new light on
this hide-and-seek. Obviously, Edward may be trying to conceal the terrible
truth from his mother. Revelation II changes the readers’ point of view
and creates a new puzzle: Why does Edward conceal the murder from his
mother, when he knows that she is bound to welcome it?
There is an obvious Oedipus-like situation in this ballad. The son kills his
father, in order to please his mother. Over the course of the dialogue, readers
gradually discover its meaning, as it emerges little by little, through their
more or less adequate assumptions and expectations, while the ballad
keeps refuting or confirming them. This emerging meaning is more com-
plex and rich and qualifies the basic Oedipal situation. Thus it cannot be
summed up but, rather, must be followed step by step. The mother is wor-
riedly inquisitive upon seeing her son so troubled. Although the readers’
perception of the events is changed by the two revelations, they are not
expected to abandon their interpretation but, rather, qualify it constantly.
At the beginning we hear that the son looks sad—or, rather, walks sad. Sad
may have a wide range of degrees, even when qualified by so. This vague-
ness renders the word ambiguous, and as we shall see, this ambiguity will
prove crucial.
Readers may be puzzled by the mother’s expertness in comparing blood
of different hues: “Your Haukis bluid was nevir sae reid.” One may even
wonder whether there is any real difference between the possible degrees
of redness. Therefore, readers are induced to find some metaphorical mean-
ing. Assuming that the mother does not know more at this stage than the
readers, one might conjecture that there is some metonymic connection
between this blood and Edward’s sorrow, and red being an intense color,
the mother’s question might imply something like “Your sorrow is too
intense to let me believe that it is only your hawk you have killed, so let
me question you further.” But since this is not explicitly stated, there is
some mysterious flavor about the mother’s illogical statement that may be
readily ascribed to the “balladic atmosphere.” Since the readers do not see
Edward, they may attribute several meanings to sad of varying intensities
for having slain his hawk or having murdered his father. When he reveals
the murder, redness acquires connotations of guilt (without lessening the
49
connotations of graveness and sorrow); it is as though the mother had said,

“You look more guilty than that.” One may even be reminded of the bibli-
cal verse: “Though your sins are like scarlet, they shall be as white as snow;
though they are red like crimson, they shall become like wool” (Isaiah 1:8).
The mother, in rejecting Edward’s evasive answers (e.g., by “Your steid was
auld, and ye hae gat mair”), draws our attention to Edward’s eagerness
to conceal the truth from her (which is apparently inconsistent with his
mother’s share in his guilt revealed at the end).
Why does Edward specifically mention the hawk and the steed? Once
again there is a “stylistic” answer and an “expressive” answer. The hawk
and the steed are the most akin to the medieval knight; if the formula
allowed for a third animal, no doubt a hound would have been mentioned.
The knight and his hunting animals are bound by mutual loyalty, according
to the medieval worldview. Killing one’s hawk or steed, although a lesser
crime than parricide, would be considered a far greater evil in the knightly
era than today. So there is a gradual accumulation of a sense of guilt, before
the real crime is revealed. In addition, there is an ironic overtone in this
preparation for the real guilt. While one might right ask, “If you loved your
hawk and steed so much, why have you slain them?” without getting a rea-
sonable answer, Edward’s ambivalence toward his father is almost explicit
in “O I hae killed my fadir deir” and seems to be satisfactory as motivation
at least to a post-Freudian mind.
Revelation II compels readers to change mental sets and reevaluate every-
thing they have encountered so far in the ballad, including Revelation I.
“And what wul ye leive to your ain mither deir” has to be interpreted, on
first reading, as a rather conventional expectation that the son might per-
haps prefer his mother to his wife and children. The final revelation puts
quite a different emphasis upon “your ain mither deir”—there is a par-
ticularly close relationship between mother and son—a sharing of guilt,
at least, and perhaps more. Readers may become aware of the ambiguous
silence after stanza 3. Assuming that the mother was ignorant, one might
well expect her to react in some violent way. Instead of this, she goes on,
rather quietly, inquiring: “And whatten penance wul ye drie for that”?
Readers need not really be too troubled with the mother’s silence: they may
readily attribute it to the ballad’s formulaic style (i.e., there was no place in
the formula for the expression of the mother’s shock). Only the final line
reveals that, presumably, she had no reason to feel thunderstruck at all.
We see how the balladic formula becomes, time and again, structurally and
expressively significant in the course of “decoding” the poem.
The final revelation shows readers a bit of Edward’s ambivalence toward
his mother. Now readers can discover that Edward, in trying to hide what

5 9
he has done, was not afraid that his mother would be too hard upon him.
One might even be tempted to interpret her “And why so sad gang yee O?”
as an invitation, “Let us now make merry.” Nevertheless, Edward tries to
conceal what he has done, as if he were rather unwilling to let his mother
feel satisfaction, let alone express her affection for him. This shameful
confession of his crime becomes a shameful confession of his crime (that
is, both mother and son know all the elements of the situation very well,
but Edward now shies away from her participation in his father’s murder).
This throws some new light on the mother’s stubborn inquisitiveness: not
being satisfied with the mere effects of the murder, she wants to have her
share of her son’s affection (emphasized, again, by “your ain mither deir” in
the last stanza). One might then reinterpret their disguised discussion of
the hawk’s blood and its hue as though each of the two had said: “I know
exactly what you mean, but I want you to speak out first.”
This reading of “Edward” captures the hypothetical nature of our inter-
pretation. Throughout the first three stanzas, readers can project a con-
sistent situation on the abrupt dialogue. They imagine an initial set of
conditions, compatible with the questions and answers presented in the
text, as well as with known human behavior and mental processing. This
initial set of conditions can be considered highly plausible in view of the
available evidence but not strictly true:
The characteristic feature of critical interpretation that is philosophically most

interesting is its tolerance of alternative and seemingly contrary hypotheses. . . .
Given the goal of interpretation, we do not understand that an admissible account
necessarily precludes all others incompatible with itself. (Margolis 1962: 116)
Where the statements “P is true” and “Q is true” are contraries, the state-
ments “P is plausible” and “Q is plausible” are not. (Margolis 1962: 117)
Our set of hypotheses, which were viewed as highly plausible in the

course of reading the first three stanzas, appear inconsistent with the final
revelation. Since, however, hypotheses are plausible rather than strictly
true, readers can change their mental set and make a different hypothesis,
seemingly contrary to the initial one. As we have seen, even the mother’s
ambiguous silence after the third stanza, which is highly important to an
understanding of the poem, only becomes possible by mediating two con-
trary hypotheses: she may or may not have been shocked by Edward’s first
revelation. This interpretation, however, requires readers to assume an
attitude toward the merely possible while continuing to be highly tolerant
of the emotionally loaded, ambiguous situation.
69
A final word about the mother’s reasons for having her husband mur-
dered by her son. Saying that she wanted to enact her part in the Oedipal
situation would be a platitude. It seems to me that an illuminating, though
disquieting, answer can be found in Aristotle’s Poetics, chapter 9. When
dealing with the difference between history and poetry, he remarks that
the latter is more philosophical because history deals with particulars such
as “what Alcibiades did or suffered,” whereas poetry deals with universals
such as “how a person of a certain type will on occasion speak or act, accord-
ing to the law of probability or necessity” (1951: 37–38). Following this line
of thought, the mother’s immediate motives seem of no relevance to the
ballad. The murder provides the opportunity for people of certain types to
act, while its lack of explicit motivation contributes to the mysterious bal-
ladic atmosphere—an atmosphere that is easily eliminated when we supply
a reason for her action such as in the case of the Oedipus complex, which
thus enables a rapid closure on the readers’ part.
Elsewhere (Tsur 1975, 2006c: 11–77) I have criticized the psychological
syndromes of rigidity and intolerance of ambiguity derived from research
on perception and personality. One of their most notorious symptoms is
the inability to be “in uncertainty, mysteries, doubts,” without any “irrita-
ble reaching after fact and reason”—in Keats’s famous phrase. It is readily
apparent that the Aristotelian answer to the question “Why did the mother
want to have her husband murdered by her son?” is bound to frustrate any
“irritable reaching after fact and reason.” Another possible symptom of this
syndrome is the concern with concrete details and their functioning—in
the present case, namely, a preference for dealing with such particulars as
“what Alcibiades did or suffered” rather than with such universals as “how
a person of a certain type will on occasion speak or act, according to the law
of probability or necessity.” The law of probability in Aristotle’s proposition
poses another difficulty to rigid individuals, since one of the symptoms of
rigidity is an unwillingness to assume an attitude toward the merely pos-
sible (Goldstein and Scheerer 1941; see also Frenkel-Brunswik 1968: 136).
I began this chapter by making a distinction between pigeonholing and a
set of instructions that readers can apply. This metacritical distinction is by
no means recent in critical theory. Apparently pigeonholing in literary criti-
cism may thus satisfy some personal need rather than the requirements of
any discipline or critical school. There is probably no critical art so readily
associated with labeling as classical rhetoric. Nevertheless, the following
observation can be found in one of the classical authorities on rhetoric:
For it makes no difference by which name either is called, so long as its stylis-
tic value is apparent, since the meaning of things is not altered by a change of

7 9
name. For just as men remain the same, even though they adopt a new name,
so these artifices will produce exactly the same effect, whether they are styled
tropes or figures, since their values lie not in their names, but in their effect.
(Quintilian n.d.: IX, 7–8)
These pairs of terms contrast two different critical activities, which, in turn,
are frequently behavioral corollaries of different psychological attitudes.
Pigeonholing, instead of bringing out the complexity of a given experience,
tends to reduce experience to a single item that falls into acknowledged
and well-classified categories (the Oedipus complex, the ballad genre).
Classifying moods, feelings, or human situations is hence a means of tak-
ing the disquieting element out of them.
'Why does your sword so drip with blood,

Edward, Edward?
Why does your sword so drip with blood?
And why so sad are ye, O?'
'O, I have killed my hawk so good,
Mother, mother:
O I have killed my hawk so good:
And I had no more but he, O.'
'Your hawk's blood was never so red,
Edward, Edward:
Your hawk’s blood was never so red,
My dear son I tell thee, O.'
'O, I have killed my red-roan steed,
Mother, mother:
O, I have killed my red-roan steed,
That once was so fair and free, O.'
'Your steed was old, and we have got more,
Edward, Edward:
Your steed was old, and we have got more,
Some other evil ye fear, O.'
'O, I have killed my father dear,
Mother, mother:
O, I have killed my father dear,
Alas! and woe is me, O!'
'And what penance will ye suffer for that,
Edward, Edward?
89
And what penance will ye suffer for that?

My dear son, now tell me, O.'
'I'll set my feet in yonder boat,
Mother, mother:
I’ll set my feet in yonder boat,
And I’ll fare over the sea, O.'
'And what will ye do with your towers and your halls,
Edward, Edward?
And what will ye do with your towers and your halls,
That were sae fair to see, O?'
'I’ll let them stand till they down fall,
Mother, mother:
I’ll let them stand till they down fall,
For here never more may I be, O.'
'And what will ye leave to your children and your wife,
Edward, Edward?
And what will ye leave to your children and your wife
When ye go over the sea, O?'
'The world is large, let them beg through life,
Mother, mother:
The world is large, let them beg through life,
For them never more will I see, O.'
'And what will ye leave to your own mother dear,
Edward, Edward?
And what will ye leave to your own mother dear?
My dear son, now tell me, O.'
'The curse of hell from me shall you bear,
Mother, mother:
The curse of hell from me shall you bear,
Such counsels you gave to me, O.'
(http://www.poemhunter.com/poem/edward-edward-
a-scottish-ballad/)

9
CHAPTER 5
Artistic Devices and Mystical Qualities

in Hebrew Devotional Poems
T his chapter explores a tradition of Hebrew devotional poems said to

have mystical-ecstatic qualities.1 The first text is a Merkabah hymn;
the rest are from the golden age in Spain. The discussion is based on a the-
ory expounded in several writings by Tsur (1992b, 2003, 2008a: chap. 13,
2008b, 2010) regarding the fossilization of poetic expressions and the
structures that can generate mystical-ecstatic qualities in poetry. Thus
this theory is presented first at some length. The poems to be discussed
(and others similar to them) deal, from the point of view of their content,
with the sublime (God and his higher worlds). They abound in mystical
symbols and notions such as “the heavenly throne,” “the curtain,” “the
merkabah” (chariot), the various symbols of light and fire, creatures of
the Merkabah, angels and Holy Living Creatures (Hayyoth ha-Kodesh),
and their heavenly songs of praise and glory. Do these mystical symbols
generate the mystical qualities perceived in a poem, or, rather, are all the
symbols more or less equal, and something else is the source of these
mystical qualities? The basic assumption behind this chapter is that the
source is something else; the mystical quality is derived from certain
poetic structures rather than from the specific symbols inserted into the
text. In other words, the mystical quality is an emergent regional quality of
1. Coauthored by Idit Einat-Nov. The theoretical framework and the discussion of
Merkabah mysticism are reproduced here from Tsur 2003. The analysis of poems from
the Spanish period is taken from Einat-Nov 1995.
01
poetic structures and cannot be reduced to the meanings of some estab-

lished set of symbols.2
We will endeavor to show how two very different texts, both abounding
in mystical symbols of the same kind, can arouse very different impres-
sions: one of “heightened expressive power,” the other of “fossilized poetic
conventions.” The kind of poetic structure that can generate a mystical qual-
ity is characterized by what Gestalt psychologists call “weak” gestalt: that
is, ambiguous, indistinct, with blurry contours. This assumption is based
on work by Anton Ehrenzweig, who associates “metaphysical intuition”
with thing-free and gestalt-free qualities.3 Ehrenzweig, in turn, draws upon
Henri Bergson, who characterizes “metaphysical intuition” as a continuous
flux “of states each of which announces that which follows and contains
that which precedes it. In reality no one begins or ends, but all extend into
each other” (in Ehrenzweig 1965: 34–35). According to Gestalt psycholo-
gists, such weak gestalts evoke a sense of uncertainty, typically display-
ing some emotional character, as opposed to solid, “good” gestalts, which
typically tend to be marked by intellectual control and evoke a feeling of
certainty and security. Weak linguistic shapes may thus arouse an impres-
sion of dissolution and suspension of boundaries that may be perceived as
isomorphic with some mystical experience.
However, a look at the poems discussed here shows that they are char-
acterized, on the contrary, by exceptionally rigid, polished, and stylized
verbal formulae and structures. Such a conclusion applies to both the
Merkabah hymn and the group of Spanish poems that follow, but not to the
same extent. In the liturgical poems from Spain, formal and verbal details
are meticulously worked out in extremely rigid formulae, whereas mysti-
cal experience—and ecstatic experience even more so—involves an over-
whelming sweep of fluid emotions. So the paradoxical question is, How can
the rigid formulae of these poems be related to mystical or ecstatic experi-
ence? We will explore the problems involved in this phenomenon, relying
on traditional scholarship of Merkabah hymns (Gershom Scholem, Rachel
2. Regional quality is a Gestaltist term from Beardsley’s Aesthetics. It refers to a per-

ceptual property of a whole that is not a property of its parts (Beardsley 1958: 83–88).
3. Ehrenzweig explains the term thing-free quality with reference to music and visual
arts. To take a literary example, the word gentleness denotes an intellectual abstrac-
tion. In the phrase “gentle heaven” it denotes a property of a compact concept. In “The
gentleness of heav’n broods o’er the sea,” Wordsworth loosens the relationship of the
property to the compact concept (in which it is “grown together” with other proper-
ties) and manipulates it into the referring position. The abstraction becomes a spatial
entity, typically processed by the right, emotional hemisphere of the brain (whose out-
put is diffuse, emotional). Broods charges gentleness with mental energy. Thus “gentle-
ness” becomes an intensely perceived, supersensuous presence in the landscape—a
thing-free quality.

1
0
Elior), on the one hand, and on cognitive poetics and Ehrenzweig’s studies
in the psychology of art, on the other. We will utilize two further illumi-
nating insights of Ehrenzweig’s (1965‫)‏‬, which he applies to music and the
visual arts.
The first insight is the one underlying this book: we apply to poetic
language the concept Ehrenzweig developed with reference to the visual
arts and music regarding the generation of ornaments. He elaborates at
great length, as we have seen, on the defense mechanisms human soci-
ety invokes to protect itself from the excessive expressive force of artistic
devices and turn them into style, that is, harmless ornaments. As we saw
in chapter 1, Ehrenzweig distinguishes three stages in the development of
artistic devices. In what he defined as Stage 1, artistic devices are perceived
subliminally and affect what he termed the “depth mind.” As these devices
become more emphatic, their emotional appeal increases, so long as they
do not become consciously perceptible. If they do become somewhat per-
ceptible, they are considered in bad taste or cheaply emotional. Ehrenzweig
called this the second stage. In the third stage, these devices turn into orna-
ments, with drastically reduced emotional appeal. In c hapter 4 we added a
fourth one, when artistic structure revives a frozen formula. We are dealing
here with some more extreme cases of formulaic poetry than in c hapter 4.
As to Ehrenzweig’s second insight, following Freud, he speaks of “oce-
anic dedifferentiation” in relation to mystical experience. William James’s
Varieties of Religious Experience indicates that this concept is consistent with
the subjective experiences reported by various individuals. Ehrenzweig
translated these psychoanalytic notions into a set of cognitive-Gestaltist
terms that can be directly applied to the subtle details of works of art. He
speaks of “a creative ego rhythm that swings between focused Gestalt and
an oceanic undifferentiation”:
The London psychoanalysts D. W. Winnicott and Marion Milner stressed the
importance for a creative ego to be able to suspend the boundaries between self
and not-self to become more at home in a world of reality where objects and self
are clearly held apart. . . . Seen in this way, the oceanic experience of fusion, of
a “return to the womb,” represents the minimum content of all art; Freud saw
in it only the basic religious experience. But it seems now that it belongs to all
creativity. (1970: 135)
Thus, oceanic experience has a cognitive and a psychoanalytic facet.

Being opposed to focused gestalt is a cognitive description of the phenome-
non; return to the womb is a psychoanalytic notion. We are more interested
in the former. As to the creative ego rhythm that swings between focused
A r t i s t i c De v i c e s a n d M y s t i c a l Q ua l i t i e s [ 101 ]
021
gestalt and an oceanic undifferentiation and the suspension of boundaries

between self and not-self, we reproduce two brief excerpts from William
James’s book. James quotes an account by Tennyson, “Individuality seems
to dissolve and fade away into boundless being,” and also comments, “This
overcoming of all the usual barriers between the individual and the Absolute
is the great mystic achievement” (James 1902: 384n, 419). In terms of the
Merkabah hymn discussed below, we argue that its forceful effect derives
from the fact that in an important sense it occurs at the point where the
solid gestalts dissolve and the boundaries between them are suspended.
This chapter focuses on what can be described as the third and fourth
stage in the development of ornaments. We elaborate on texts from two
periods. In the Merkabah hymn we point out two rigid formulae; but we
also suggest that the occurrence of the second rigid formula after the
first one generates a perceptual dynamics perceived as oceanic experi-
ence, the verbal imitation of an altered state of consciousness. Such for-
mulaic discourse is further formalized in texts from the Spanish period.
Some traditional critics claim that when writing such formulaic poems,
Ibn Gabirol actually underwent a mystical experience.4 We argue that we
cannot know what the poet experienced when writing the poem. At any
rate, some other critics assume that the formulaic nature of such poems
eliminates the expressive force of the words and fossilizes the discourse.
Regardless, it would appear that there are ways to use such formulaic pat-
terns to revive their expressive force and impart to them an emotional,
even ecstatic character. In this respect, we draw attention to the elusive
but highly significant differences between very similar poems by Shlomo
Ibn Gabirol and Itshak Ibn Ghiyat.5 We claim that Ibn Ghiyat’s text exem-
plifies Ehrenzweig’s third stage, whereas Ibn Gabirol’s text revives to a
considerable extent Ibn Ghiyat’s fossilized formulae. We are not concerned
here with the chronological order of these two poets but, rather, with their
different use of the same kinds of ornaments within one poetic tradition.
ANCIENT MERKABAH HYMNS
Gershom Scholem writes in Jewish Gnosticism, Merkabah Mysticism, and

Talmudic Tradition (1960) that the first phase in the development of Jewish
mysticism was also the longest. Its literary remains are traceable over a
4. Rabbi Solomon Ibn Gabirol (1021–1058) was the greatest Hebrew poet and a phi-
losopher in Arabic in the Spanish golden age.
5. Rabbi Itshak Ibn Ghiyat (1038–1089) was a minor Hebrew poet in the Spanish
golden age.

031
period of almost a thousand years, from the first century bc to the tenth
ad (Scholem 1961: 40). Most of the tracts are called “Hekhalot books,”
that is, descriptions of the hekhaloth, the heavenly halls or palaces through
which the visionary passes and in the seventh and last of which there rises
the throne of divine glory (Scholem 1961: 45). Scholem writes, “There is
little hope that we shall ever learn the true identity of the men who were
the first to make an attempt, still recognizable and describable, to invest
Judaism with the glory of mystical splendor” (1961: 41):
The Greater Hekhaloth presents us with a large number of Hebrew hymns, which
it treats in an unusual manner: the very same hymns are characterized by the
text as representing two different types of songs. On the one hand, the hymns
are addressed to the throne and to Him who sits upon it, and are described as
celestial songs of praise sung by “the Holy Living Creatures” (Hayyoth ha-Kodesh)
who, in Ezekiel 1:5 ff., are the bearers of the throne. On the other hand, these
same hymns are the ones the mystic is instructed to recite before and during his
ecstatic ascent to heaven (which, in a very curious and so far unexplained change
of phraseology, is always referred to in this text as a descent unto the Merkabah).
The hymns describe, in a plethora of solemn phrases, the spirit of majesty and
solemnity that permeates the heavenly realm, “the Palaces of Silence” in which
God’s Shekhinah dwells. (1960: 20)
Before turning to the poetic text itself, we propose two short semantic
exercises concerning two phrases in the last sentence of the foregoing quo-
tation: “the Palaces of Silence” and “God’s Shekhinah.” A palace is a large,
impressive, and majestic building of imposing magnificence; it is usually
the official residence of a sovereign, bishop, or other exalted personage.
The orientation mechanism collects information from the environment
that specifies one’s place in and relationship to the surrounding space. The
huge hall of a palace casts self-specifying information on the perceiving
self that makes the individual feel extremely small (which appears to be the
reason why ancient sovereigns chose such buildings for a residence—and
elevated thrones for seats—to inspire their subjects’ awe). Great silence
deprives the person of information about the surroundings and thus may
reinforce such a feeling. “Palaces of Silence” may suggest real palaces domi-
nated by great silence, casting self-specifying information of smallness.
One of the assumptions underlying the present work is that it is the orien-
tation mechanism that handles such self-specifying information, involving
a ready integration of many inputs at once and producing a diffuse output.
This has an intense nonrational effect that, when heightened, generates a
quality similar to altered states of consciousness, such as meditation and
0
41
mystic experience. However, in a figurative construal, Silence may be the

referring expression, and Palaces, its metaphoric modifier. In this case, the
conflicting terms of the metaphor delete such physical and visual features
of “palaces” as [+ stable shape], transferring to “silence” such features as
<+ impressive + majestic + imposing magnificence + self-specifying
spatial information>. The result is the intense but diffuse presence of an
engulfing thing-free and gestalt-free quality.
Scholem sometimes uses the Hebrew term Shekhinah in the English
text. But in one of the hymns, for instance, he translates Shekhinat
ha-Shekhinah as the “presence of Presence” (1960: 21). Such a tautological
construction bestows an enigmatic character on the text. Using Presence
to refer to God can be accounted for by two transformations of such an
expression as “God is present”: the adjective is turned into an abstract
noun and manipulated into the referring position of the phrase. Tsur
(e.g., 2008a: 461–468, 610–616, 2012c: 64–67, 76–77) has called this sty-
listic device a thematized predicate or a topicalized attribute. The phrase
“presence of Presence” reflects a double transformation of this kind. This
device is most widespread when conceptual language is required to convey
some nonconceptual experience such as altered states of consciousness,
whether meditative, mystical, or ecstatic. In poetry, thing-destruction
and thing-free or gestalt-free qualities readily contribute to the impres-
sion of such an experience.
Such constructions can be found in seventeenth-century meditative
texts, as well as in English Romantic poets such as Wordsworth and Keats
and French Symbolists such as Baudelaire and Rimbaud. Tsur has elsewhere
shown that this is one of the most effective stylistic devices to distinguish
Whitman’s “meditative catalogue” from his “illustrative catalogue” (Tsur
2008a: 456–468). In c hapter 4 of his book, Scholem provides three short
texts that he claims are outstanding paradigms of what Rudolf Otto called
“Numinous Hymns” (Otto 1959: 21–22). We will reproduce and briefly dis-
cuss one of them here:
Excerpt 1.
‫וָ יוֹם‬ ‫‏‬מ ֶּׁש ַבח וְ ִשׁ ָירה ֶשׁל יוֹם‬ ִ
‫ִמּגִ ילָ ה וְ ִרּנָ ה ֶשׁל עִ ִּתים וְ עִ ִּתים‬
‫דוֹשׁים‬ ִ ‫וּמ ִהּגָ יוֹן ַהּיוֹצֵ א ִמ ִּפי ְק‬ ִ
‫‬מ ִּפי ְמ ָשׁ ְר ִתים‬ ִ ‫וּמּנִ ּגוּן ַה ִּמ ְתּגַ ֵּב ‏ר‬ִ
‬‫ָה ֵרי ֵאשׁ וְ גִ ְבעוֹת לֶ ָה ָבה‏‬
‫נִ צְ ָּברוֹת וְ נִ גְ נָ זוֹת וְ נִ ָּתכוֹת ְּבכָ ל־יוֹם‬
‬'‫וכו‏‬ ‫כדבר שנאמר קדוש קדוש קדוש‬

051
[From the praise and song of each day,

From the jubilation and exultation of each hour,
And from the utterances that proceed6 out of the mouth of the holy ones,
And from the melody that welleth out of the mouth of the Servants,
Mountains of fire and hills of flame
Are piled up and hidden and poured out each day.] (Scholem’s
translation)
As Scholem insisted, nearly nothing is known about the origins of this

poetry. One remarkable thing about it is that from its very beginnings it
was frozen into rigid formulae. There are, for instance, long catalogs about
songs and praises, and most visible things are described as made of fire
and flames. We know almost nothing about the people who wrote these
poems and even less about their psychodynamic or cognitive processes.
In her essay on the unique characteristics of the religious phenomenon in
Hekhalot literature, Rachel Elior concludes that
the Hekhaloth literature does not deal with ecstatic outbursts and their inciden-
tal individual interpretation. . . . The unified conception and the routine descrip-
tions of the contents of the visions in the Hekhaloth literature indicate that they
do not leave speculative or descriptive space for the interpretation of the vision
of the descender unto the merkabah. Pluralism of conception, view, observation
or interpretation is not allowed; the meaning, order and conditions of the vision
are firmly determined. (1987: 93)
In other words, the descriptions of the heavenly visions were already fro-
zen in ancient mysticism into “meaningless formulae.” They do not express
individual experience but become “an objective fact and pattern of refer-
ence for metaphysical reality” (Elior 1987: 13). Thus, for instance, the
unique, sublime biblical event, “suddenly a chariot of fire appeared with
horses of fire . . . ; and Elijah went up by a whirlwind into heaven” (2 Kings
2:11), is turned into a routine practice in the Merkabah (Chariot) literature.
This could epitomize the phenomenon pointed out by Ehrenzweig.
Later in this chapter we shall see that the frozen formulae of the
Merkabah hymns become even more rigid in medieval Hebrew poetry in
Spain. In light of Ehrenzweig’s hypothesis, the rigid formulae of this poetry
may still have had too much expressive power for the later poets.
6. The grammatical form of the Hebrew verbs for “proceed” and “welleth forth” is
more like the present participle than a finite verb. Unless otherwise noted, all transla-
tions are ours.
061
The reiteration of rigid formulae is meant to have a strong hypnotic

effect.7 That, indeed, may be the case for the initiated. However, as Tsur
(1974) has argued elsewhere, the uninvolved outsider, the reader of poetry,
may need much more for detecting the kind of devotional experience expe-
rienced by the initiated in the poem. As for this specific hymn, we will argue
that when the poem switches from one rigid formula to another, from the
reiteration of “praise and song” to the reiteration of words that denote
“mountains of fire and hills of flame,” suddenly an overwhelming feeling of
oceanic dedifferentiation may be detected.
The first four lines are different at the immediately observable level but
are synonymous on a higher level: they all contain expressions of glorifica-
tion. This structure directs attention to the common elements, away from
the individuating elements of each line. Syntactically, all are parallel prepo-
sitional phrases, introduced by from. There is also some extra parallelism
within each pair of lines. This excessive repetition has a magic, incanta-
tory effect. The first two lines offer smaller-scale repetitions as well. The
first line offers two parallel nouns that suggest glorification: “praise and
song.” The second line as well offers two parallel nouns that suggest ecstatic
joy: “jubilation and exultation.” The Hebrew phrase translated as “each day”
literally reads “day and day”; the phrase translated as “each hour” literally
says “times and times.” Such an enigmatic incantatory device may have
some limited quasi-hypnotic effect, but in itself it would be insufficient for
evoking some ecstatic quality. The “descenders unto the Merkabah” claim
that in their adventures they encounter great dangers. This straightfor-
ward, obtrusive rhythm gives some security to the Platonic censor in us.
But there is a syntactic device in this hymn (shared, mutatis mutandis, with
the first few lines of Milton’s Paradise Lost, Baudelaire’s “Hymne,” and quite
a few other poems) that turns this rhythm less straightforward, render-
ing the security false. The preposition from reiterated four times generates
suspense in two different senses of the word: holding in an undetermined
or undecided state awaiting further information and mental uncertainty,
anxiety. The parallel prepositional phrases do generate a straightforward
rhythm, but by the same token, they impede the straightforward move-
ment of the sentence by delaying the predicate required for clarifying its
structure. The subject phrase “Mountains of fire and hills of flame” causes
an additional delay, and the predicate “Are piled up and hidden and poured
out” (syntactically but not semantically predicted by the prepositional
phrases) occurs only in the last line.
7. For hypnotic poetry, see, e.g., Snyder 1930; Tsur 2008a: chap. 19.

1
07
When one of us (Tsur) first read Scholem’s book forty years ago, he had
a strong intuition that after the four incantatory lines there was a switch
to some overwhelming quality that powerfully engulfed him. At that time
he could not account for it. Today we know that it suggests some struc-
tural resemblance to oceanic dedifferentiation. The last two lines, contain-
ing the subject and the predicate phrases, also shift the poetic strategy of
the hymn. Fire and flame in the second-to-last line continue the pattern of
obtrusive formulaic reiteration. But by the same token they introduce a
gestalt-free, unstable, threatening entity, radiating excessive energy. From
the figurative point of view, the genitive phrase “Mountains of fire and
hills of flame” may be construed similarly to the genitive phrase “Palaces
of Silence.” When fire is the referring member of the phrase, Mountains
suggests “enormous mass”; the feature [+ stability] is deleted. However,
Mountains and hills reinforce in each other their literal meaning; when
they are the referring members, they indicate a vast landscape to which
the perceiving self may relate, activating the orientation mechanism. The
three parallel verb phrases “Are piled up and hidden and poured out” not
only recapitulate the incantatory effect, in that the contrast between the
absence of finite verbs in the first five lines and the sudden pileup of three
consecutive finite verbs in the last line introduces much movement; by
the same token, the meaning of the verbs indicates the intense fluctua-
tion of stable scenery. “Poured out” is associated with liquids (in fact, the
Hebrew verb suggests both pouring and alteration from a solid to a liquid
state, usually by heat), and the sequence as a whole seems to subvert the
surrounding landscape with its self-specifying information. The landscape
itself is unstable: it comes into existence, disappears, and reappears.
Finally, the whole presentation has an interesting existential aspect. The
syntactic agents (“holy ones,” “Servants”) are conspicuously inactive. It is
the praise, the song, the jubilation, the exultation, the utterances, and the
melody that are active, displaying voluntary or spontaneous impulses to
act by employing “linguistic devices that shift the agent to a non-volitional
role” (Balaban 1999: 259). Balaban found empirically that people who report
overwhelming religious experiences “tend to manipulate their own self into
a non-volitional syntactic role” (1999: 259). The praise, the song, the jubi-
lation, the exultation, the utterances, and the melody occur, as it were, of
themselves. Such a consistent syntactic strategy indicates a suspension of
voluntary control. As we have observed, “Mountains of fire” and “hills of
flame” serve as referential terms for the orientation of the self, but at the
same time they have no objective existence at all. They issue from, or their
pileup is caused by, the praise, the song, the jubilation, the exultation, the
utterances, and the melody, which, in turn, issue from the mouth of the agents.
0
81
HEKHALOT HERITAGE IN THE GOLDEN AGE
In this section we shall consider a group of poems in which fossilization is

more relentless than in any other corpus of poems we know and the psycho-
logical processes underlying fossilization are perhaps more palpable than
in any other example in this book. In medieval Hebrew liturgical poetry
there are poems that are similar to the Hekhalot hymn discussed above, in
both content and linguistic shaping. These involve certain liturgical genres
(ophanim, rahitim, silukim) that, according to recent scholars, display traces
of content, form, and style derived from ancient Hebrew piyutim (Fleischer
1975) and Hekhalot hymns (Levin 1986). These devotional poems are about
the grandeur of God and his higher worlds. Chariots and objects of fire
loom large in them. In terms of verbal shaping, they are characterized by
the repetition of verbal formulae, rigid verbal and stylistic elements, acros-
tics, various symmetric structures, anadiplosis (concatenation), and rep-
etition of stock phrases. The organizing principles in piyutim of this kind
are very prominent, and this fact usually imparts to them a thought-out
quality, a stylized and polished character, the impression of calculated and
fastidiously controlled discourse. A rigid order that allows for no devia-
tion turns these poems into significantly predictable discourse and, by the
same token, decreases their informativeness. It tends to evoke a sense of
certainty and security in the reader, greatly obstructing any experiencing
of the threat involved in what Rudolf Otto called mysterium tremendum.
Consequently, in terms of the four-stage development of artistic devices,
most of these poems belong to the third stage, where the expressive power
of artistic devices is eliminated and they are perceived as mere ornament.
Let us consider some instances, all by Shlomo Ibn Gabirol:
Excerpt 2.
‫‬אוּרים וּלְ ַמ ָּטה ִּב ְבנֵ י‏ ְ‬ּב ִח ִירים‬ ִ ‫לְ ַמעְ לָ ה ְּב ֵאלֵ י‏‬ ‫ֵ‬אל נִ ְק ָּדשׁ‬
‬‫ וּלְ ַמ ָּטה ְּב ַד ַעת‏ ְ‬ּדגָ לִ ים‬ ‫לְ ַמעְ לָ ה ְּבגַ עַ שׁ‏‬ּגַ לְ ּגַ ּלִ ים‬
‫ת ַ‬ה ְּבזָ ִקים וּלְ ַמ ָּטה ְּבוַ ַעד‏‬וְ ִת ִיקים‬ ‫לְ ַמעְ לָ ה ְּב ִהּלוֹ ‏‬
‫שׁוּקים‬ִ ‫ וּלְ ַמ ָּטה ְּבח ֶֹסן‏ ֲ‬ח‬ ‬‫לְ ַמעְ לָ ה ְּבזִ ְמ ַרת‏‬זִ‏ּק‬ִים‏‬
‫ וּלְ ַמ ָּטה ְּביִ ְק ֵרי‏‬יְ ָח ִסים‬ ‬‫לְ ַמעְ לָ ה ְּבט ַֹה ‏ר ָ‬ט ִסים‏‬
‫בוּבים‬ ִ ְ‫ וּלְ ַמ ָּטה ְּבלֶ ַקח‏‬ל‬ ‬‫רוּבים‏‬ ִ ְ‫לְ ַמעְ לָ ה ְּבכֶ ֶתר‏‬ּכ‬
‫ וּלְ ַמ ָּטה ְּבנֵ צַ ח‏‬נְ ִסיכִ ים‬ ‬‫לְ ַמעְ לָ ה ְּב ַמעֲ נֵ ה ְמלָ כִ ים‏‬
‫טוּפים‬
ִ ‫ וּלְ ַמ ָּטה ְּב ֶע ֶתר‏ ֲ‬ע‬ ‬‫‬שׂ ָר ִפים‏‬ ְ ‫לְ ַמעְ לָ ה ְּב ִשׂיחַ‏‬
‫רוּפים‬ ִ ְ‫ וּלְ ַמ ָּטה ְּבצֶ ַרח צ‬ ‬‫רוּפים‏‬ ִ ‫לְ ַמעְ לָ ה ְּב ֶפצַ ח ְּפ‬
‫אשׁים‬ ִ ‫ וּלְ ַמ ָּטה ְּב ִרּנַ ת‏ ָ‬ר‬ ‬‫דוֹשׁים‏‬ ִ ‫לְ ַמעְ לָ ה‏ ִ‬ּב ְק ַהל‏ ְ‬ק‬
‫קוּפים‬ ִ ‫קוּפים‏‬ וּלְ ַמ ָּטה ְּבתֹכֶ ן‏ ְ‬ּת‬ ִ ‫‬שׁ‬ ְ ‫לְ ַמעְ לָ ה‏ ְ‬ּב ִשׁ ַירת‏‬

091
El niqdaš lǝmaʕla bǝʔeley ʔurim ulǝmata bibhne bǝḥirim

lǝmaʕla bǝgaʕaš galgalim ulǝmata bǝdaʕat dǝgulim
lǝma la bǝhilot habǝzakim ulǝmata bǝvaʕad vǝtikim
ʕ
lǝmaʕla bǝzimrat ziqim ulǝmata bǝḥosen ḥašuqim

lǝmaʕla bǝtohar tasim ulǝmata bǝjiqre jǝḥasim
lǝmaʕla bǝkhɛtɛr kǝruvim ulǝmata bǝlɛqaḥ lǝvuvim
lǝmaʕla bǝmaʕane mǝlakhim ulǝmata bǝneṣaḥ nǝsikhim
lǝmaʕla bǝsiaḥ sǝraphim ulǝmata bǝʕɛtɛr ʕatuphim
lǝmaʕla bǝphɛṣaḥ pǝruphim ulǝmata bǝṣɛraḥ ṣǝruphim
lǝmaʕla biqǝhal qǝdošim ulǝmata bǝrinat raʔšim
lǝmaʕla bǝširat šǝquphim ulǝmata bǝtokhɛn tǝquphim
(1971: 139–140)
[God is glorified above with angels of fire
below with the children of the chosen ones
above with the storm of the wheels [spheres]
below with the wisdom of the tribes [of Israel]
above with the flash of the lightning Creatures
below in the company of the righteous
above with the song of Seraphim
below with the adoration of the beloved [people]
above with [Seraphim] flying around in purity
below with those of precious ancestry
above with the crowns of Cherubim
below with the fine words of the warm-hearted
above with the responses of Kings
below with the eternity of Princes
above with the discourse of Seraphim
below with the prayers of those covered [with a prayer shawl]
above with the song of those with touching [wings]
below with the outcry of the purified
above with the company of Saints
below with the rejoicing of keystones [= children of Israel]
above with the song of the transparent [angels]
below with the homage of those firm in their belief]
Before discussing this poem, we must briefly elucidate a unique tech-

nique prevalent in the vast corpus of piyutim: namely, the sophisticated use
of epithets. According to Ezra Fleischer, this is “one of the most character-
istic marks of the style of the paytanim in all times and in all poetic schools”
0
1
(1975: 107), especially in the ancient piyut of Eretz Yisrael. The epithet is

a simple replacement of the name of a concept: “A paytan has recourse to
an epithet whenever he needs to refer to a well-known concept which he
does not want to directly designate by its proper and usual name, because
it is too explicit and trite” (Fleischer 1975: 105). Note that new words or
new meanings had to be invented continually to meet the relentless formal
requirements of this genre. Various techniques can be used to create “the
esoteric language of epithets”; the epithet may be a “brief periphrasis of
the notion referred to” or “derived by way of metonymy or allusion to some
event mentioned in the Bible” or by resorting to “some word or phrase that
refers to that notion in the Bible or in the sayings of the Talmudic sages”
(Fleischer 1975: 266, 106–107). In Excerpt 2, each verse line contains two
parallel phrases introduced by above and below, respectively, followed by an
epithet for singing and praising and an epithet for higher creatures versus
the children of Israel.
To give just two examples from Excerpt 2, in line 4 the poet refers
to the children of Israel with the epithet ‫שׁוּקים‬ִ ‫ =( ֲח‬beloved ones), alluding
to the biblical verse “The Lord did not set His love on you . . . but because the
Lord loves you” (Deuteronomy 7:7–8). Another, wildly sophisticated but
not uncommon kind of epithet is found in ‫רוּפים‬ ִ ‫ּב ֶפצַ ח ְּפ‬.ְ In his commen-
tary, Yarden paraphrases this passage as “with the song of the Holy Living
Creatures, whose wings touch one another” (in Ibn Gabirol 1971). The
nonexistent noun pɛṣaḥ is derived from the Hebrew verb paṣaḥ, meaning
“open, begin.” The meaning of “song” is attributed to it by back-formation
from the idiom paṣaḥ bǝšir (= began to sing). Pǝruphim means “the touching
[ones].” This is an allusion to Ezekiel’s description of the “living creatures”
of his vision: “Their wings touched one another” (Ezekiel 1:9). However, the
verb in this biblical verse does not fit the needs of the rhyme pattern and
acrostic in this poem. So the poet went one step further and substituted a
possible synonym for it from the Talmudic tractate Shabbath (1.1), which
refers to a completely different issue.
Such piyutim as Excerpts 1–3 (and many others like them; see Excerpt 3
below) are characterized by syntactic-semantic patterns reiterated
with rigorous precision and other rigid stylistic characteristics such as
acrostics, anadiplosis (concatenation; see below), and intensive formu-
laic repetitions of stock phrases. They pile up expressions of praise to
God. The first two lines of Excerpt 2 state that the whole universe sings
songs of glory: the angels in their higher worlds and humans “below.”
This notion is central to the Hekhalot literature in that it connects

1
celestial worship to human prayers (cf. Elior 1994: 33). On the pho-
nological level, three formal devices are prominent in Excerpt 2: the
acrostic from aleph to tav, each speech sound in the acrostic is repeated
(alliteration), and in each verse line the first hemistich rhymes with
the second.
These piyutim and the Hekhalot hymn share an insistence on songs of
glory conveyed by a pattern of repeated identical verbal structures. There
is, however, a conspicuous difference in this respect regarding the length
of the catalogs, which in piyutim are, as a rule, far more extensive. The
longish catalogs are one of the factors that contribute to the artificial
and frivolous impression of these Spanish piyutim. Another difference
between the Hekhalot hymn and the Spanish piyutim concerns the degree
of similarity between the verbal patterns. In the hymn it is not a verba-
tim repetition but, rather, the repetition of a syntactic-semantic scheme;
moreover, this abstract scheme also changes from lines 1–2 to lines 3–4.
Consequently, “unity in variety” is perceived, resulting in relative com-
plexity. In the majority of the Spanish piyutim discussed here, by con-
trast, verbatim repetition is prominent in the verbal schemes, evoking an
impression of “tediousness” or “lack of progress.” Owing to the acrostic,
the fixed length of verse lines, and the exact repetitions in these piyutim,
rigid order and thought-out organization become their most prominent
features. The single words that change in accordance with the needs of
the acrostic, occurring in strictly reiterated syntactic patterns, reinforce
the common elements in each other. Thus they obliterate each other’s
unique semantic features and contribute to the impression of artificial-
ness and technical frivolity. An additional difference between repetition in
the hymn and repetition in the piyutim concerns syntactic suspension: in
the former, the meaning unit is completed only after the catalogs in lines
1–4—in fact, only in line 6. In the latter, by contrast, each one of the
analogous verse lines is, at the same time, a complete syntactic-semantic
unit. In the catalogs consisting of such syntactically complete units, the
impression of monotony and the sense of “saturation” are amplified: there
is no development from one verse line to another, and the completed
units arouse no expectations for new information, unlike the incomplete
units of the hymn, which require their completion in a verbal unit beyond
the opening catalog. As we have said, the verse lines “Mountains of fire
and hills of flame / Are piled up and hidden and poured out each day” do
indeed contain a new syntactic structure, but its relationship to the pre-
ceding units is not sufficiently clear.
2
1
ACROSTIC AND SEMANTIC FOSSILIZATION
Excerpt 3.
‫‬שׁירוֹת יַ ְשׁ ִּביצוּ‬ ִ ‫ ְ‬ּב ֶשׁ ֶקד ְּב ֶשׁ ַבח‏‬,‫קוּפים ְ ּב ֶשׁנֶ ן‏‬
ִ ‫ְּת‬
‫ְּבלַ ַהב ְּבלֶ ֶמד ְּבלֶ ַקח‏‬לְ ָבבוֹת יָ לִ יצוּ‬
‫‬מּלוֹת יַ ְמ ִריצוּ‬ ִ ‫ְּב ַמעַ ן ְּב ֶמלֶ ץ ְּב ֶמ ֶר ‏ץ‬
‫ְּב ֶהגֶ ה ְּבהוֹד ְּב ָה ָדר‏ ָ‬הגוּת יְ ַה ְר ִּביצוּ‬
‫הוּרים יָ צִ יצוּ‬ ִ ְ‫ְּבצַ ַהל ְּבצֶ וַ ח ְּבצֶ ַר ‏ח‬צ‬
‫ְּבעֶ ֶתר ְּבעֶ ֶרץ ְּבעֶ לֶ ץ‏‬עֲ נָ וִ ים יַ ֲעלִ יצוּ‬
‫ְּבי ֶֹשׁר ְּבי ֶֹתר ְּבי ֶֹק ‏ר‬יְ ֵשׁנִ ים‏‬יָ ִקיצוּ‬
‫ְּב ַר ַחשׁ ְּב ַרעַ שׁ ְּב ֶרגֶ שׁ‏ ַ‬רגְ לָ ם יָ ִריצוּ‬
‫ֹלהי יִ ְשׂ ָר ֵאל יַ ֲע ִריצוּ‬
ֵ ‫וְ ִה ְק ִּדישׁוּ ֶאת ְקדוֹשׁ יַ עֲ קֹב וְ ֶאת ֱא‬
tǝquphim bǝšenen bǝšɛqɛd bǝšɛvaḥ širot jašbiṣu
bǝlahav bǝlɛmɛd bǝlɛqaḥ lǝvavot jaliṣu
bǝmaʕan bǝmɛlɛṣ bǝmɛrɛṣ milot jamriṣu
bǝhɛgɛ bǝhod bǝhadar hagut jǝharbiṣu
bǝṣahal bǝṣɛvaḥ bǝṣɛraḥ ṣǝhurim jaṣiṣu
bǝʕɛtɛr bǝʕɛrɛṣ bǝʕɛlɛṣ ʕanavim jaʕaliṣu
bǝjošɛr bǝjotɛr bǝjoqɛr jǝšenim jaqiṣu
bǝraḥaš bǝraʕaš bǝrɛgɛš raglajim jariṣu
vǝhiqdišu ʔɛt qǝdoš Jaʕqov vǝʔɛt ʔɛlohe Jisrael jaʕariṣu
(Ibn Gabirol 1971: 140)
A poem like the one in Excerpt 3 cannot be translated even literally; only its
semantic structure can be explicated. It is a sequel to Excerpt 2. It begins
with the last word of the earlier poem, ‫קּופים‬
ִ ‫ּת‬,ְ that is, “the firm ones” (= the
children of Israel). This is the subject phrase for all the ensuing predicates. In
the rest of the poem, each verse line consists of three consecutive adverbials
of mode or instrument that suggest “talking” or “devotion” of one kind or
another. The last two words of each line contain a direct object followed by a
verb. Some of the words do not exist in the dictionary, and those that do are
used here in a slightly or radically different sense. The nonexistent words can
be shown to be derived from an existing word of the same root. The adverbi-
als have one semantic ingredient in common: “talking” or “adoration.” The
semantic features that distinguish them from one another are effaced.
Let us take a closer look at the first three lines of this poem:
bǝšenen bǝšɛqɛd bǝšɛvaḥ širot jašbiṣu

bǝlahav bǝlɛmɛd bǝlɛqaḥ lǝvavot jaliṣu
bǝmaʕan bǝmɛlɛṣ bǝmɛrɛṣ milot jamriṣu

1
3
The first three words of each line begin with the prefix bǝ-, whose func-
tion is to turn a noun into an adverbial of mode or instrument. Bǝšenen
is derived from šinnun (repetition, learning by rote); Yarden in his com-
mentary construes it as “speech.” Bǝšɛqɛd is derived from šǝqida, meaning
“industriousness.” Bǝšɛvaḥ means “in praise.” Širot jašbiṣu means “inlay
songs [with words].” Bǝlahav means “with flame, ardently.” Bǝlɛmɛd is
derived from lǝmida (learning). Lɛqaḥ means “moral, lesson.” Lǝvavot jaliṣu
means “cheer up hearts.” Bǝmaʕan is derived, according to Yarden, from
maʕane lašon (response, reply). Bǝmɛlɛṣ is derived from mǝliṣa, “flowery
speech.” Bǝmɛrɛṣ means “vigorously.” Jamriṣu, derived from the same root,
means “urge, encourage, stimulate.” Thus, milot jamriṣu means “stimulate
words.” And so forth.
The graphemic-phonemic dimension of the poem is remarkable.
Ignoring the repeated prefix, in every line each word begins with the same
consonant (alliteration). The letters designating the alliterating speech
sounds yield, in the consecutive lines, the acrostic ‫( שלמה צעיר‬that is, “young
Šǝlomo”). We use the cumbersome phrase “the letters designating the allit-
erating speech sounds” because the letter ‫ה‬, for instance, designates the
consonant [h]‌in the alliterating sequence but serves as a special character
designating a vowel in the acrostic “Šǝlomo.” Such a shift emphasizes that
the same graphemes change their very nature from text to acrostic.
Many of the consecutive adverbials are near-homonyms (that is, they
differ in only one phoneme), such as bǝmɛlɛṣ bǝmɛrɛṣ, bǝṣɛvaḥ bǝṣɛraḥ, bǝʕɛrɛṣ
bǝʕɛlɛṣ, bǝjošɛr bǝjotɛr bǝjoqɛr, and bǝraḥaš bǝraʕaš. Note that bǝṣɛvaḥ bǝṣɛraḥ
became stock epithets and occur in Excerpt 3 as well. All nine verse lines
rhyme on the same monorhyme ending with -iṣu, with the stress on the
penultimate vowel.
Excerpt 3 is similar to Excerpt 1 in that it too contains obsessive repeti-
tions of words suggesting song and adoration. It presents a rich catalog of
words, most of which contain the semantic component “talking” or “voic-
ing.” These words display an enormous variety of meanings. However, here
too the words change according to the requirements of the acrostic and
occur in reiterated syntactic patterns. They reinforce their shared semantic
components and lose their distinguishing features. Thus they contribute to
the effect of monotonous repetition, with no novel information, and ulti-
mately give the impression of artificiality and technical trifling. This prin-
ciple is so effective here that readers “understand” even the words whose
meaning they do not know and construe them as “some kind of saying” or
“some kind of adoration.”
As we have pointed out above, the letters/phonemes in this text are
treated in a special way (which is quite characteristic of this corpus). Apart
4
1
from their function in ordinary language, they are organized into three
additional patterns, all arbitrary from the point of view of ordinary com-
munication. First, in every line each word begins with the same letter (after
omitting the prepositional and conjugational prefixes). Second, these let-
ters add up to the poet’s “signature,” his name and nickname, ‫שלמה צעיר‬
(young Šǝlomo). Third, the last words of the lines make up an apparently
“innocent” monorhyme. However, this monorhyme is part of a fourth pat-
tern: When the last line appears, the monorhyme takes on new significance
after the event; it heralds, as it were, the biblical verse “They will hallow the
Holy One of Jacob, / And fear the God of Israel” (Isaiah 29:23). The occur-
rence of this biblical verse is expected and surprising at the same time and
ends the poem with a powerful closure. Likewise, Excerpt 4 (below) termi-
nates with the verse “‫ֹלהינוּ ַה ַּמגְ ִּב ִיהי לָ ֶשׁ ֶבת‬
ֵ ‫”מי ּכַ יְ יָ ֱא‬
ִ [who is like the Lord our God
who dwells on high], quoted from Psalms 113:5. Such a closure is a conven-
tion in this corpus, and as such, it is expected. But readers do not know
which verse they should expect or at what point of the sequence. Only after
the event, when the verse occurs, do they feel that the sequence has ended,
that they have “reached home.”
The signature of the poet’s name is widespread in liturgical poetry. It
is usually regarded as a device to indicate the poet’s identity and nothing
more. Alphabetic acrostics are found in the Bible. In this poem, however,
the fact that the letters occur in three different patterns beyond their
ordinary linguistic use foregrounds most emphatically the arbitrary, non-
linguistic combination of letters. A comparison to Psalm 34 highlights
the unique treatment of letters in this poem. In ordinary language, as
well as in poetic language, the letters are “transparent” signs: we do not
linger on them, since their function is to point to the phonological sig-
nifieds (speech sounds), which, in turn, serve as signifiers of semantic
units. Though in Psalm 34 the syntactic coherence is less smooth than
in other psalms, the semantic-thematic richness of the verses is not
impaired. Conversely, the blurring of the semantic load of the words in
this poem and the increase in the number of patterns in which the letters
of the rhyme and the acrostic participate result in the disruption of the
chain of signifiers and signifieds, thus directing the reader’s attention to
the combination of letters.8
In light of these differences, the “innocent” signature of the author’s
name takes on special significance in the context of the processes discussed
8. It should be noted that when the early mystics talk of “letters,” it is not always clear
whether they mean the written signs or the speech sounds; but this does not affect our
essential argument.

1
5
here. We argue that it can be regarded as a fossilized remnant of power-

ful processes of a magic-ecstatic nature. This will be better understood if
we situate these poems between the creation of the world by fiat, as in
the book of Genesis, and the combination of letters in the Book of Yetsira
and Abraham Abulafia.9 According to Yehuda Liebes, the Book of Yetsira
expresses the view that “the world was created by words, and it is, so to
speak, a work of literature” (2001: 16). The letters of the alphabet are foun-
dations: “foundations of both language and reality signified by language,
or, in another formulation, a foundation of reality that was created by
speech” (Liebes 2001: 16).
Steven T. Katz views the theurgical and contemplative manipulation of
letters found in the Sefer Yetzira (an early Jewish mystical text) as paradig-
matic.10 Gershom Scholem points out that “these theurgical doctrines form
a kind of meeting-place for magic and ecstaticism” (1961: 78). According to
this view,
the letters of the Hebrew alphabet, out of which words are composed, are the
fundamental building blocks of creation. The letters have ontic capacity and
can be—indeed, have been—employed by God to create the world and every-
thing within it. In this reading, the creation accounts wherein “God speaks” are
taken with extreme, if original, literalness. The process of creation resides in the
manipulation of the alphabetical ciphers. (Katz 1992: 16)11
In this passage, the creative decree (where “God speaks”) undergoes a

curious metamorphosis. Such phrases as “Let there be light” or Milton’s
“Silence ye troubled waves, and thou deep, peace! / Said then th’Omnific
Word, Your discord end!” (P.L. VII ll. 216–217) exploit the semantic compo-
nent of the Word to perform the act of creation, in a possible world where
the forces of chaos obey imperative verbs. In the magic-ecstatic use of mys-
ticism, the creative power of the Word is displaced from the semantic to the
9. Book of Yetsira is the title of the earliest extant book on Jewish esotericism.
10. Theurgy is the art or technique of compelling or persuading a god or beneficent or
supernatural power to do or refrain from doing something.
11. “Writing has a very magical quality—not because of anything divine about
its origins, but because it greatly increased our brain’s capacities. It is close to
miraculous that Homo sapiens, a mere primate, was able to dramatically increase
its memory by making a few marks on paper,” says the neuropsychologist Stanislas
Dehaene (2009: 172). But notice this. When mystics believe in such magic, it is one
thing; when critics insist on it, it is another. A lot of nonsense has been written,
for instance, Kabbalistic speculations on symbolic meanings in Coleridge’s “Alph the
sacred river” (e.g., Beer 1959: 209–212; Tsur, 2006c: 25–26).
6
1
graphemic component. Katz quotes a short passage from Sefer Yetzira that
describes such “meaningless” manipulation of letters:
How did He combine them, weigh them and set them at opposites?
Aleph with all of them, and all of them with Aleph,
Bet with all of them, and all of them with Bet.
It rotates in turn, and thus they are in two hundred and thirty-one gates.
And everything that is formed and everything that is spoken goes
out from one term. (in Katz 1992: 17)
Scholem speaks of the “science of combination of letters” in Abraham

Abulafia’s teaching as of an “adequate methodical guide to meditation with
the aid of letters and their configurations. The individual letters of their
combinations need have no ‘meaning’ in the ordinary sense; it is even an
advantage if they are meaningless, as in that case they are less likely to
distract us” (Scholem 1961: 133‫)‏‬. Newberg, D’Aquili, and Rause suggest a
wider context for such a meditational practice:
Active types of meditation begin not with the intention to clear the mind of
thoughts, but instead, to focus it intensely upon some thought or object of
attention. A Buddhist might chant a mantra, or focus upon a glowing candle or
a small bowl of water, for example, while a Christian might pray with the mind
trained upon God, or a saint, or the symbol of a cross. (2001: 120)
Similarly, a Jew may focus on the heavenly halls or the combination of the
letters of God’s names.
An anonymous disciple of Abulafia’s gave an extended, masterly descrip-
tion of such an experience; a short passage is reproduced below:
When I came to the night in which this power was conferred on me, and
midnight—when this power especially expands and gains strength whereas the
body weakens—had passed, I set out to take up the Great Name of God, consisting
of seventy-two names, permuting and combining it. But when I had done this for
a little while, behold, the letters took on in my eyes the shape of great mountains,
strong trembling seized me and I could summon no strength, my hair stood on end,
and it was as if I were not in this world. At once I fell down, for I no longer felt the least
strength in any of my limbs. And behold, something resembling speech emerged from
my heart and came to my lips and forced them to move. (in Scholem 1961: 150–151)12
12. We quote Abulafia’s disciple not as someone who influenced Ibn Gabirol (who
lived 150 years before him) but as someone who gave an exceptionally lively descrip-
tion of the experience whose fossilized remnants we encounter in our corpus.

71
Figure 5.1 Electroencephalogram showing epileptic waveforms. (Reproduced from

Sacks 2012).
What is the relationship among God experiences, talking gibberish, and

seeing letters as mountains? Persinger (1987: 16) pointed out the neuro-
psychological correlates of such overwhelming experiences in which they
co-occur. God experiences are associated with temporal lobe transients,
which are electrical perturbations of the temporal lobe in the human
brain (see Figure 5.1). During such experiences, some people sometimes
speak in gibberish: “The gibberish is often called ‘speaking in tongues’ and
resembles the sounds emitted from young infants during the babbling
stage” (Persinger 1987: 30). Such perturbations of the temporal lobe may
also cause perceptual alterations, which “are usually limited to the sudden
expansion of visual and auditory images. Objects in the room may suddenly
seem to grow very small and then increase in size again. Sounds may fluctu-
ate from very faint and distant to very loud and near” (Persinger 1987: 18).
A conjunction of such responses would render the process particularly rel-
evant to the experience of Abulafia’s disciple. It may perhaps also account
for experiences echoed in the verse lines “Mountains of fire and hills of
flame / Are piled up and hidden and poured out each day” in Excerpt 1.
The emphatic foregrounding of letters via the stylistic devices of the
poems in the corpus under discussion seems to contain a remote echo of
such an overwhelming experience—more remote than the closing lines of
Excerpt 1. If we are right in our assumption, the alphabetic acrostic and the
conventional signature of the poet’s name represent the ecstatic experience
of the “combination of letters” fossilized into stylistic devices (fossil, as sug-
gested in c hapter 1, refers to the remnant or trace of something constituted
to carry on the activities of life, but only some of its outward shape has
been preserved, not the activities of life it was constituted to carry on). The
disciple’s account also gives an indication of the “dangers” involved in such
8
1
an overwhelming experience: one loses voluntary control of one’s body and

mental processes. The acrostic, by contrast, does not presuppose electrical
perturbations in the brain—it is “harmless” ornament. In other words, it
is the “combination of letters” minus electrical perturbations in the brain.
ANADIPLOSIS: THE “BLOCKING” AND “FLUX” EFFECTS
Another widespread form of repetition is the pattern of anadiplosis (= con-

catenation): namely, a pattern in which every new line must begin with
the last word of the preceding one. Anadiplosis is a highly “technical” and
rigid formal pattern. Nevertheless, in different verbal conditions the same
pattern can produce radically different perceived effects. To see this better
we scrutinize anadiplosis and its implications for the perceived effect in a
piyut by Shlomo Ibn Gabirol (Excerpt 4, below), and in order to sharpen
our distinctions, we analyze two brief excerpts from a piyut by Itshak Ibn
Ghiyat (Excerpt 5, below), another poem with an anadiplosis pattern but
with radically different effects.13 In terms of the model we offered above
regarding the four stages of the development of ornaments, Ibn Ghiyat’s
poem represents the third stage and Ibn Gabirol’s, the fourth (though, as
we shall see, Ibn Ghiyat also exploits certain semantic subtleties, but for a
rather different poetic effect).
Here is the section “your ceiling, Mighty” from Ibn Gabirol’s composite
poem “‫”א ֲא ִמיר ַא ֲא ִדיר‬
ַ (I extol I praise):
Excerpt 4.
‬‫א ִּדיר ַּב ָּמרוֹם‏‬נִ ּצֶ ֶב ‏‬ ‬
‫ת‬ ַ ‫קר ְתָך‬
ְָ ִּ‫‬ת‬
‫‬יוֹשׁ ֶבת‬ֶ ‫נִ ּצֶ ֶבת ֶמ ְמ ַשׁלְ ּתְ‏ָך‬וְ לָ נֶ צַ ח‏‬
‬‫‬שׁלְ ֶה ֶבת‏‬ ַ ‫יוֹשׁ ֶבת‏ ִ‬ס ְת ָרתְ‏ָך ַ‬ּב ֲח ֻתּלַ ת‏‬
ֶ
‫וּמ ֹּפה‏‬לַ ֶה ֶבת‬ ִ ‫ַשׁלְ ֶה ֶבת ִמּפֹה‬
‫‬מ ְרּכֶ ֶבת‬ִ ‫ת‬ ‫לַ ֶה ֶבת ְתלַ ֵהט לְ עֻ ַּמ ‏‬
‫ִמ ְרּכֶ ֶבת עֻ ּזְ‏ָך ִ‬היא‏ ַ‬הּנִ ּצֶ ֶבת‬
‫יוֹשׁב‏ ַ‬ּב ֶּשׁ ֶבת‬
ֵ ‫ַהּנִ ּצֶ ֶבת מוּל ִט ְפ ָסר‬
‫ַּב ֶּשׁ ֶבת ִמּיָ ִמין ְשׁ ָח ִקי ‏ם‬רוֹכֶ ֶבת‬
‫רוֹכֶ ֶבת ֲא ֻפ ַּדת ְשׁ ִביעִ י ְּברֹאשׁוֹ‏ ְ‬ּב ַמּצֶ ֶבת‬
‫‬יוֹשׁ ֶבת‬ֶ ‫ְּב ַמּצֶ ֶבת ִשׁ ְבעָ ה ְּב ַפ ְרּגוֹ ‏ד‬
13. Ibn Gabirol’s, but not Ibn Ghiyat’s, poem has the extra formal constraint: an elab-
orate acrostic. If you combine the first letters of each line, you obtain “‫אני שלמה ברבי יהודה‬
‫אבן גבירול‬,” that is, “I Šǝlomo son of Rabbi Yǝhuda Ibn Gabirol.”

1
9
‬ ‫‬יוֹב ֶבת‏‬
ֶ ‫יוֹשׁ ֶבת ִמּלְ ָפנִ ים נִ ְשׁ ְק ָפה‏‬ ֶ
‫יוֹב ֶבת מוּל ִמ ְרּכַ ְב ְּתָך ַ‬הּנִ ְשׁלֶ ֶבת‬ ֶ
‫רוּח וְ עָ נָ ן‏‬וְ צָ ֶר ֶבת‬
ַ ‫ַהּנִ ְשׁלֶ ֶבת ְּב‬
‫‬ּדוֹא ֶבת‬
ֶ ‫וְ צָ ֶר ֶבת ּתוֹ‏ְך ֶ‬ק ַרח ֵמ ַאיִ ן‏‬
‫רוֹפ ֶפת עֲ לִ ּיָ ה‏ ַ‬הּנֶ ְחצֶ ֶבת‬ֶ ְ‫ּדוֹא ֶבת ו‬ ֶ
‫‬אוֹה ֶבת‬
ֶ ‫ַהּנֶ ְחצֶ ֶבת לְ קוֹל ֲה ֻמּלָ ה אוֹתְ‏ָך‬
‫אוֹה ֶבת עֻ ּזְ‏ָך‬ּגַ ם ְּבקוֹל ּגַ ם‏ ְ‬ּב ַמ ֲח ֶשׁ ֶבת‬ ֶ
‫‬נוֹב ֶבת‬
ֶ ‫ְּב ַמ ֲח ֶשׁ ֶבת ְוּבקוֹל ָרּנֵ י זְ ִמירוֹת‏‬
‫‬ּגוֹב ֶבת‬
ֶ ‫נוֹב ֶבת ִרגְ ַשׁת ָה ַא ְר ַּבע ָס ִביב‏‬ ֶ
‫לשׁים וָ ֵשׁשׁ נֶ גְ ּד‏ָך ְ‬ּב ִמ ְקצֶ ֶבת‬ ִ ‫ּגוֹב ֶבת ְשׁ‬ ֶ
‫‬יוֹה ֶבת‬
ֶ ‫וּשׂמֹאל לְ‏ָך‬ ֹעז‏‬ ְ ‫ְּב ִמ ְקצֶ ֶבת יָ ִמין‬
‫יוֹה ֶבת ֶק ֶדם וְ ָאחוֹר ְּב ַאלְ ֵפי ְר ָבבוֹת‏‬רוֹכֶ ֶבת‬ ֶ
‫סוֹב ֶבת‬ ֶ ְ‫קוֹמָך‬הוֹלֶ כֶ ת‏‬ו‬ ְ ‫רוֹכֶ ֶבת עַ ל ְמ‬
‫סוֹב ֶבת לִ ְראוֹת ְּבזִ יו‏ ַ‬ה ִמ ְרּכֶ ֶבת‬ ֶ ְ‫ו‬
‫ֹלהינוּ ַה ַּמגְ ִּב ִיהי לָ ֶשׁ ֶבת‬ ֵ ‫ִמי ּכַ יְ יָ ֱא‬
tikratkha ʔadir bamarom niṣɛvɛt
niṣɛvɛt mɛmšaltǝkha wǝlanɛṣaḥ jošɛvɛt
jošɛvɛt sitratkha baḥatulat šalhɛvɛt
šalhɛvɛt mipo umipo lahɛvɛt
lahɛvɛt tǝlahet lǝʕumat mirkɛvɛt
mirkɛvɛt ʕuzkha hiʔ haniṣɛvɛt
haniṣɛvɛt mul tifsar jošev bašɛvɛt
bašɛvɛt mijamin šǝḥaqim rokhɛvɛt
rokhɛvɛt ʔafudat šǝviʕi bǝroʔšo bǝmaṣɛvɛt
bǝmaṣɛvɛt šivʕa bǝfargod joṣɛvɛt
joṣɛvɛt milfanim niṣqǝfa jovɛvɛt
jovɛvɛt mul mirkavtǝkha hanišlɛvɛt
hanišlɛvɛt bǝruaḥ bǝʕanan wǝṣarɛvɛt
wǝṣarɛvɛt tokh qɛraḥ meʔajin doʔɛvɛt
doʔɛvɛt wǝrofɛfɛt ʕalija hanɛḥṣɛvɛt
hanɛḥṣɛvɛt lǝqol hamula ʔotkha ʔohɛvɛt
ʔ
ohɛvɛt ʕuzkha gam bǝqol gam bǝmaḥṣɛvɛt
bǝmaḥṣɛvɛt uvǝqol rane zǝmira novɛvɛt
novɛvɛt rigṣat haʔarbaʕ saviv govɛvɛt
govɛvɛt šǝlošim wašeš nɛgdǝkha bǝmiqṣɛvɛt
bǝmiqṣɛvɛt jamin usmol lǝkha ʕoz johɛvɛt
johɛvɛt qɛdɛm wǝʔaḥor bǝʔalfe rǝvavot rokhɛvɛt
rokhɛvɛt ʕal mǝqomkha holɛkhɛt vǝsovɛvɛt
vǝsovɛvɛt lirʔot bǝziw hamirkɛvɛt
mi kaʔadonaj ʔelohenu hamagbihi laṣɛvɛt. (1971: 244–246)
201
[Your ceiling, Mighty, in the sky [is] positioned

positioned [is] your sovereignty and forever seated
seated [is] your secret place wrapped in flame
flame on this side and on this side blaze
blaze burns fiercely against the chariot
[it is the] chariot of your might that stands upright
that stands [upright] in front of the Commander who is sitting
sitting on the right on the skies [is] riding
riding on the seventh vest [sphere] with its head in heaven[?]‌
in heaven seven behind the curtain sits
sits in the front and is seen sobbing
sobbing in front of your chariot that [is] intertwined
that [is] intertwined with the wind and the cloud and the fire
and the fire within the ice from nothingness [emerges] painful
painful and unstable [is] the heaven that is carved
that is carved to the voice of tumult that loves you
loves your might both in voice and in thought
in thought and in voice of the singers of songs it talks
talks the excitement of the four [groups of angels] around crowding
crowding together in thirty-six [thousand times ten thousand miles]
rhythmically
rhythmically right and left honor to you give
give before and behind by thousands of thousands it chariots
chariots to your place goes and turns around
turns around to see the splendor of the chariot
who is like the Lord our God who dwells on high.]
Anadiplosis foregrounds both the initial and final boundaries of the verse
lines. Thus it may generate an effect of blocking or disjoining the lines.
On the other hand, repetition of the same word at the end and begin-
ning of two consecutive lines can have the effect of conjoining them, blur-
ring the distinction between the separate lines, and blending them into
one semantic sequence. Anadiplosis thus has two potentially opposite
effects (as the great modernist Hebrew poet Yehuda Amichai wrote of
the sea: “perhaps connects, perhaps separates”). We argue that one of
the texts realizes the former and the other, the latter option. We label the
former option “the blocking effect” and the latter “the flux effect.” The
flux effect connects the linguistic units in the consecutive verse lines and
can also be labeled “the mystic effect” because it blurs the boundaries and
differentiation of the various verse lines, generating weak, ambiguous,
indistinct verbal structures. In Ibn Ghiyat’s poem (see below) anadiplosis

1
2
has a perceived effect of blocking, whereas in Ibn Gabirol’s poem it gener-

ates the flux effect, imparting to it a perceived mystic quality. Below we
discuss the verbal elements that achieve these effects.
Examining the syntactic-semantic status of the line-final concatenating
words, we can see that in Ibn Ghiyat’s poem (below) most are indispensable
for completing their verse line as a syntactic-semantic unit. The majority of
the concatenated verse lines in this poem end with words that must belong,
syntactically, to the verse line in which they occur, either because they con-
stitute the prepositional phrase of a genitive construct or because they are
conjoined nouns in “contracted sentences” or nouns internally required by
the verb as adverbials of place (we cannot say, “I put”; we must add what
and where). These three syntactic functions determine the status of these
words as the last link in their verse lines, needed to complete their syn-
tactic structure. This foregrounds the blocking effect of anadiplosis, which
emphasizes the separateness of the verse lines.
A considerable number of verse lines in Ibn Gabirol’s text, by contrast,
have syntactic structures that generate the opposite effect of anadip-
losis, the flux effect. In these verse lines, Ibn Gabirol bestows a complex
syntactic-semantic status on the line-final concatenating words, such
that they may be perceived as gratuitous but required at the same time
for the completion of the utterance in the verse line. This double status of
the line-final concatenating words in Ibn Gabirol’s poem generates tension
and a sense of uncertainty. This is achieved, mainly, in three ways: (1) plac-
ing the adverbial of place before the verb, (2) substituting an optional for
an inherently required complement, (3) taking advantage of the fact that
the present tense of a Hebrew verb may also serve as an adjective or a
noun—the ambiguous prefix ha- may serve as a relative pronoun before a
verb or as a definite article, turning the word into a noun or an adjective.
Adverbial of Place Preceding the Verb
Let us consider the first verse line: “‫”ּת ְק ָרתְ‏ ‬ָך ַא ִּדיר ַּב ָּמרוֹם נִ ּצֶ ֶבת‬
ִ [Your ceiling,
Mighty, in the sky [is] positioned]. In Hebrew, both the adverbial “in the
sky” and the adjective positioned may serve as predicates without the
copula is. Consequently, in a first reading the adverbial may be perceived,
momentarily, as the predicate completing the utterance: “Your ceiling,
Mighty, [is] in the sky.” In such a reading, the concatenating word posi-
tioned will be perceived as unnecessary for the utterance of the line and
freed to link with the utterance in the next line, which, in turn, begins
with the same word. In this reading, readers can treat the two tokens of
21
positioned as reduplication and thus join the two lines into one seman-
tic sequence. However, while this is a plausible reading, there is a more
obvious organization of the syntactic-semantic information in this verse
line: when readers reach the line-final positioned, they reorganize the ver-
bal elements to suggest “Your ceiling, Mighty, [is] positioned in the sky.”
In such an organization of the verbal information in the line, “[is] posi-
tioned” becomes the nucleus of the predicate and is no longer perceived as
superfluous for the utterance in the line. At this point it becomes clear that
the two performances—the one that assigns a central role to positioned in
the predicate and the one that relegates it to the next line—generate two
sentences that are near-synonymous: “Your ceiling, Mighty, in the sky [is]
positioned” = “Your ceiling, Mighty, [is] in the sky.” What William Empson
(1955) called “double syntax” conveys here no additional meaning. So, what
is their effect? They blur each other, enhancing the fluidity of the passage,
and thus contribute to the mystical effect of the sequence.
A more striking example of the effect of double syntax generated by the
inversion of the adverbial of place and the concatenating word is found
in line 4: “‫ּומּפֹה לַ ֶה ֶבת‏‬
ִ ‫”שׁלְ ֶה ֶבת ִמּפֹה‬
ַ [flame on this side and on this side blaze].
Here as well there are two possible readings: one that sees the concatenat-
ing word blaze as redundant in its line, thus freeing it to link with the
next line, and another that regards it as required by the preceding utter-
ance. Idiomatic Hebrew would require saying “flame on this side and on
this side” for “flame on each side” or “flame all around.” The present-day
Hebrew speaker has difficulty reading this sequence of words otherwise,
even after becoming overfamiliar with this verse line. But readers then
realize that the last word enables them to reorganize the line into two par-
allel sentences, that is, into a symmetrical closed shape. Again, the double
syntax conveys no added meaning; rather, the two possible sentences blur
each other. Thus, the sequence becomes well articulated and fluid at the
same time. The conjunction and changes its function from one reading to
another. In the former it serves to conjoin two parallel adverbials of place
in an idiomatic expression, whereas in the latter it conjoins two sentences,
thus enhancing the sense of uncertainty, instability, and insecurity.
Let us briefly present some additional examples of double syntax gen-
erated by such verb-adverbial inversion at the concatenation junction.
In line 19 we find: “‬‫ּגוֹב ֶבת‏‬ ֶ ‫”נוֹב ֶבת‏ ִ‬רגְ ַשׁת ָה ַא ְר ַּבע ָס ִביב‬
ֶ [talks the excitement of
the four around crowding]. Here again “talks the excitement of the four
around” can be read as a self-contained, complete, though slightly obscure
sentence (and one that is utterly awkward in English): it has a straightfor-
ward verbal predicate (talks), a subject (the excitement of the four), and an
adverbial of place (around), meaning “the excitement of the four is talking

231
all around” (in Hebrew, transposition of the verb before the noun is per-
fectly legitimate). In this way, the concatenating word, crowding, is made
redundant and linked to the utterance in the next line. In this reading,
the syntactic-semantic status of the line-final concatenating word becomes
obscure; by coupling with the word at the beginning of the next line, it is
perceived in the sweep of reading as an irrational, gratuitous repetition,
reinforcing the emotional import displayed by the already obscured pat-
tern of anadiplosis. On second reading, however, the same constituents
of the sentence allow for an alternative organization that renders the
line-final concatenating word highly required by the preceding phrases in
the line: “talks the excitement of the four—all around crowding together.”
Likewise, line 22 may be read as “‫”יוֹה ֶבת‏ ֶ‬ק ֶדם וְ ָאחוֹר ְּב ַאלְ ֵפי ְר ָבבוֹת‏—‬רוֹכֶ ֶבת‏‬
ֶ [give
before and behind by thousands of thousands—it chariots]14 or “‫יוֹה ֶבת ֶק ֶדם‬ ֶ
‫[ ”וְ ָאחוֹר‏— ְ‬ּב ַאלְ ֵפי ְר ָבבוֹת רוֹכֶ ֶבת‏‬give before and behind—by thousands of thou-
sands it chariots].
Optional and Internally Required Complements
Another way in which Ibn Gabirol imputes a double syntactic-semantic sta-

tus to line-final concatenating words, a sense of requiredness versus non-
requiredness, is the substitution of optional complements for inherently
required ones. This device can be found in lines 21–22:
‫יוֹה ֶב ‏‬
‫ת‬ ֶ ‫וּשׂמֹאל לְ‏ָך‬עֹז‬ ְ ‫ְ‬ּב ִמ ְקצֶ ֶבת יָ ִמין‬
‬‫‬יוֹה ֶבת ֶק ֶדם וְ ָאחוֹר ְּב ַאלְ ֵפי ְר ָבבוֹת רוֹכֶ ֶבת‏‬
ֶ
[rhythmically right and left honor to you give
give before and behind by thousands of thousands it chariots]
There are verbs that cannot stand alone and inherently require comple-
tion. “He closed” cannot be considered a complete utterance unless we can
reconstruct the missing object from the context. “He closed the door,” by
contrast, is a complete utterance. The objects are indispensable for, inher-
ently required by, such verbs. If we say, “He closed the door yesterday,”
the adverbial of time is not inherently required but an optional addition,
not mandatory. Ibn Gabirol proceeds in these lines as follows: At first
(line 21), he presents the verb gives with its two required object phrases
14. This is one of the typical obscurities of this poem. It’s a verb we took from Shelley’s
“Ode to the West Wind” (“Oh thou, / Who chariotest”).
2
41
(“[give] honor to you”); the second time (line 22), he does not provide
the internally required complements but adduces, as an afterthought, an
optional one (the adverbial “before and behind”) that becomes natural—if
at all—only after the internally required objects have already been sup-
plied. In this way, in order to perceive a complete utterance, the reader may
tend to accept the reduplication of gives as suggesting an afterthought, so
as to perceive the two verse lines as complementing each other to consti-
tute one meaning unit: “rhythmically right and left—honor to you give /
(give) before and behind.” The line boundaries are obscured and ruptured,
and the verbal units (verse lines) extend into each other such that the flux
effect is generated.
A similar phenomenon occurs in lines 12–13:
‫יוֹב ֶבת מוּל ִמ ְרּכַ ְב ְּתָך ַ‬הּנִ ְשׁלֶ ֶבת‏‬ ֶ

‫רוּח וְ עָ נָ ן וְ צָ ֶר ֶבת‏‬
ַ ‫ַ‬הּנִ ְשׁלֶ ֶבת ְּב‬
[sobbing in front of your chariot that [is] intertwined
that [is] intertwined with the wind and the cloud and the fire]
The verb intertwine inherently requires a complement. An object cannot be

intertwined; it must be intertwined with something. This complement is
absent from line 12 but is amply supplied by three(!) parallel objects (the
wind and the cloud and the fire) after the second token of the word, in
line 13. Consequently, in order to satisfy the demand of intertwined for
an internally required complement, here again the reduplication of the
concatenating word obscures the line boundary, just as in lines 21–22,
and the two lines tend to extend into each other, so as to constitute one
syntactic-semantic unit.
An additional example can be found in lines 6–7:
‫ִמ ְרּכֶ ֶבת עֻ ּזְ‏ָך ִ‬היא‏ ַ‬הּנִ ּצֶ ֶבת‏‬

‫יוֹשׁב ַּב ֶּשׁ ֶבת‏‬ֵ ‫ַהּנִ ּצֶ ֶבת מוּל ִט ְפ ָסר‬
[[it is the] chariot of your might that stands upright
that stands [upright] in front of the Commander who is sitting]
According to the Even-Shoshan Dictionary, the Hebrew word ‫ ניצב‏‬has

several meanings, including “stand upright” and “stand before.” It has
“stand” as its basic sense; “upright” is only a semantic feature in it, not a
separate adverb; in the latter meaning, it internally requires an adverbial
of place. When passing from line 6 to line 7, the verb internally changes

251
its semantic makeup, shifting some of the emphasis from “upright”

to the required adverbial of place “in front of the Commander.” At the
same time, it refers back to its syntactic subject, “chariot of your might.”
Such a structure induces readers to blend their perception of lines
6–7 into one syntactic-semantic unit. Thus the flux effect is realized
and supported.
The requiredness of “in front of the Commander” acts to connect lines
6–7 in the reader’s perception. The ambiguity of the prefix ha- serves to dis-
connect these two lines. But such a reading will syntactically impair line 7:
haniṣɛvɛt will be construed as a noun, while the relative pronoun ha- is
changed into a definite article, and the whole line is perceived as one sub-
ject clause—“the one standing [upright] in front of the Commander who
is sitting”—with the predicate missing. Consequently, in order to complete
the syntactic structure in line 7, the reader will be compelled to connect
line 7 to line 8, thus: “‫יוֹשׁב ַּב ֶּשׁ ֶבת‏— ַ‬ּב ֶּשׁ ֶבת ִמּיָ ִמין ְשׁ ָח ִקים רוֹכֶ ֶבת‬
ֵ ‫”הּנִ ּצֶ ֶבת מוּל ִט ְפ ָסר‬
ַ [[the
one] standing [upright] in front of the Commander who is sitting / sitting
on the right, on the skies [is] riding]. Thus, in one place or another (lines
6–7 or 7–8) readers will be compelled to obscure or rupture their percep-
tion of the line boundaries.
Relative Particle and Anadiplosis
The third way Ibn Gabirol renders the boundary words ambiguous
is to attach the relative particle to the concatenating word. Consider
lines 15–16:
‫רוֹפ ֶפת עֲ לִ ּיָ ה ַהּנֶ ְחצֶ ֶבת‏‬

ֶ ְ‫ּדוֹא ֶבת ו‬
ֶ
‫‬אוֹה ֶבת‏‬
ֶ ‫ַ‬הּנֶ ְחצֶ ֶבת לְ קוֹל ֲה ֻמּלָ ה אוֹתְ‏ָך‬
[painful and unstable [is] the heaven that is carved
that is carved to the voice of tumult that loves you]
How should this verbal sequence be realized? Again there are two possibili
ties. One possibility is to foreground the verse lines as closed, complete units,
reading as follows: “‬‫‬אוֹה ֶבת‏‬
ֶ ‫ ַ‬הּנֶ ְחצֶ ֶבת לְ קוֹל ֲה ֻמּלָ ה‏—‬אוֹתְ‏ָך‬/ ‫רוֹפ ֶפת עֲ לִ ּיָ ה ַהּנֶ ְחצֶ ֶבת‏‬
ֶ ְ‫”ּדוֹא ֶבת ו‬
ֶ
[painful and unstable [is] the heaven that is carved / that is carved to the
voice of tumult—that loves you]. Line 15 means, roughly, that heaven
was carved from an unstable state and the process was painful. (“Carved”
inherently requires some hard material as its object, suggesting here a
paradox.) Line 16 means that the carving of heaven was accompanied by
261
the loving voices of a tumult (of angels?). With some goodwill we can say
that line 15 is complete and self-sufficient, closed, and that line 16 is a
separate unit providing some additional information that the painful carv-
ing of heaven was accompanied by some loving voices. This stable, closed
structure is reinforced by a symmetric chiastic pattern: line 15 ends with
the relative clause preceded by all the rest, whereas line 16 begins with a
relative clause, followed by all the rest. One cannot ignore, however, that
the relative clause in line 16 refers back to the subject (“heaven”) in line 15.
Hence neither of the two verse lines “begins or ends, but . . . extend into
each other” (Ehrenzweig 1965: 35). The two tokens of “that is carved” are
perceived as reduplication for emotive purposes, or the sake of obscuring
the juncture between the two lines, or both. The two lines will be read as
“painful and unstable are the heavens that are carved (that are carved) to
the voice of tumult / that loves you.”
ANADIPLOSIS IN IBN GHIYAT
To demonstrate the unique nature of Ibn Gabirol’s use of the anadiplosis

technique, we will briefly compare it with lines 1 and 17–22 of another
poem of the same genre, by Itshak Ibn Ghiyat:
Excerpt 5.
‬‫מוֹשׁ ָביו‏‬
ָ ‫ַא ָּתה ֵאל ִמ ְס ַּת ֵּתר ְּב ֶח ְביוֹן‬
. . .
‫עוֹב ָריו וְ ָשׁ ָביו‏‬
ְ ְ‫נוֹדעוּ ל‬ ְ ‫בוֹתימוֹ ל ֹא‬ ֵ ‫נְ ִת ָיביו עִ ְק‬
‫טוֹביו‬
ָ ְ‫נוֹה ִרים ל‬ ֲ ְ‫ּפוֹח ִדים ֵאלָ יו ּכַ ּיוֹם ו‬ֲ ‫וְ ָשׁ ָביו‬
‫רוֹביו‬
ָ ‫טוֹביו צָ ֲהלוּ ֵמהוֹדוֹ ְּבשׂוּמוֹ ְרחוֹקוֹ ִּב ְק‬ ָ ְ‫ל‬
‫דוֹשׁיו נְ ִד ָיביו‬ ָ ‫דוֹשׁיו ְוּק‬
ָ ‫רוֹביו ְק‬ ָ ‫ְק‬
‫נְ ִד ָיביו ְרצוֹנוֹ לְ ָה ִפיק ְּב ָק ָריו וַ ֲע ָר ָביו‬
‫אוֹריו ְּב ַמ ֲע ָר ָביו‬ ָ ‫וַ עֲ ָר ָביו ָשׁ ֲחרוּ ּכִ ְשׁקֹע‬
ʔ
ata ʔel mistater bǝḥɛvjon mošavav
. . .
nǝtivav ʕiqvotemo loʔ nodʕu lǝʕovrav vǝšavav
vǝšavav poḥadim ʔelav kajom vǝnoharim lǝtovav
lǝtovav ṣahalu mehodo bǝsumo rǝḥoqo biqǝrovav
qǝrovav qǝdošav uqǝdošav nǝdivav
nǝdivav rǝṣono lǝhafiq bǝqarav vaʕaravav
vaʕaravav kǝšaḥar kišqoʕa ʔorav bǝmaʕaravav (in David 1987)

271
[you are a God who hides in the secret place of his residence
. . .
the footsteps of his routes are unknown to his passers and his returners
[= passersby]
and his returners dread to him today and throng to his goods
to his goods they rejoice in his majesty turning remote ones into those
close to him
those close to him are his saints his saints are his nobles
his nobles attempt to please him in his mornings and his evenings
[= always]
and his evenings are black while his sun is setting in his West]
Ibn Ghiyat follows verbal strategies that are very different from Ibn
Gabirol’s. Owing to the syntactic function of the majority of the line-final
concatenating words in Ibn Ghiyat’s poem, they are highly required in
order to complete the line as a syntactic-semantic unit. Most of the con-
catenated verse lines in this poem close with words that, because of their
syntactic-semantic function, must belong to the line in which they occur,
either because they are the prepositional phrase of a genitive construct, as
in “the secret place of his residence”; or because they are nouns preceded
by the conjunction and, creating a pair of parallel nouns in a contracted
clause, as in “in his mornings and his evenings” (their symmetry is rein-
forced by the facts that they are antonyms and that the expression is an
idiomatic stock phrase in medieval devotional poetry, meaning “always”);
or because they are nouns that are inherently required by the verb, as in
“turning remote ones into those close to him” (the verb turn into inher-
ently requires a complement; the same holds true of throng to, requiring an
adverbial of direction). In some cases the second token of the concatenat-
ing word is used with a radically different meaning, indicating a completely
new start. The collocation “his mornings and his evenings” is an interesting
case. The conjunction and conjoins an idiomatic pair of parallel (antonymic)
nouns. In the next token of “and his evenings” it conjoins two independent
sentences, dismantling the idiom, where the repeated word refers to an
entirely different aspect of “evening” and starts a statement that is con-
spicuously irrelevant to the preceding utterance; moreover, quite untypi-
cally in this corpus, it gives a description of nature: black evening as the
background to the setting sun. The same holds true of “the footsteps of his
routes are unknown to his passers and his returners,” where and conjoins
an idiomatic pair of parallel (near-synonymic) nouns meaning his (casual)
“passers-by.” The next token of “returners” suggests, literally, return to a
2
8
1
place (from exile, as in the book of Jeremiah?) or, possibly, religious repen-
tance; both construals are conspicuously irrelevant to the “passers-by” of
the preceding utterance.
These three syntactic functions and the semantic change in the last
member of an idiomatic phrase determine the status of the line-final con-
catenating word as the last link demanded to complete the utterance and
close the verse line. This emphasizes the “blocking effect” of anadiplosis,
whose task is to foreground a clear-cut boundary between verse lines.
We are confronted with two different aesthetic conceptions. The “block-
ing” and “flux” effects are associated with different semantic conceptions
of concatenated words. Two meanings of the word ambiguous highlight
these different conceptions. Merriam-Webster’s Collegiate Dictionary defines
ambiguous as “doubtful or uncertain especially from obscurity or indistinct-
ness” and “capable of being understood in two or more possible senses.”
The former characterizes Ibn Gabirol’s usage, and the latter, Ibn Ghiyat’s.
Ibn Gabirol typically obscures the difference between the two tokens of the
concatenated words (as we have pointed out with reference to cases of “dou-
ble syntax”), whereas Ibn Ghiyat typically sharpens it (as in “in his morn-
ings and his evenings [= always] / and his evenings are black”). Versification
units with clear-cut boundaries have “good gestalt” and suggest strong
intellectual control, inspiring certainty, whereas units with blurred bound-
aries have “weak gestalt,” inspire uncertainty, are regarded as affectively
charged, and indicate emotional excitability. This view of Gestalt theory is
supported by the Rorschach inkblot test. Furthermore, the abrupt change
of meanings in the repeated words introduces considerable movement into
the frozen formula in Ibn Ghiyat’s poem as well.
SUMMARY
At the beginning of this chapter we raised the issue of the relationship

between an overwhelming, highly emotional mystic experience and the
rigid, thought-out formulae in the texts discussed here. We responded in
two stages. First, such formulaic poetry may have its origins in a poetry
that reflects powerful mystical experiences, but it reached us when it was
already a fossilized tradition. We have argued, following Ehrenzweig, that
elusive, powerful artistic devices may arouse anxiety resulting from uncer-
tainty; consequently, they are prone to become “style,” ornaments, and
rigid formulae. Thus we assume that the acrostic in general and, in par-
ticular, the line-initial signatures of the poets’ names in the poems under
discussion are the fossilized remnants of the “science of the combination of

291
letters” in powerful mystical experiences. Thus such poems may represent

a mystical tradition, even if they do not reflect the actual mystical experi-
ences of their authors. This view is supported by traditional research, for
example, Rachel Elior’s work quoted above regarding the frozen, formulaic,
nonexperiential character of Hekhalot hymns, which were cast, even in the
mystical literature, into “meaningless formulae” and fixed and solid “pat-
terns of reference” to metaphysical reality.
Second, stylistic devices occur in poetic structures that determine their
character. A brief comparison of two poems of anadiplosis, for instance,
showed that the same stylistic device serves in one of them to clearly sepa-
rate the perceptual units and articulate the boundaries between them,
whereas in the other it creates a continuous flux “of states each of which
announces that which follows and contains that which precedes it.” “In real-
ity no one begins or ends, but all extend into each other”—to use Bergson’s
terms. Our analysis of these poems suggests that the rigid patterns may
evoke not only a sense of certainty and security but its opposite as well.
In certain uses they may become irrational patterns charged with tension
that evoke in the reader a sense of uncertainty, generating an increased
emotional effect. Concurrently, the sense of security achieved through the
obtrusive formulaic repetition of rigid patterns may become false security
through syntactical uncertainties, which, when combined with certain
kinds of content (e.g., what Rudolf Otto called mysterium tremendum), may
generate a mystical poetic quality.
We compared such formulaic poems with the description of an over-
whelming God experience (by an anonymous disciple of Abulafia), trig-
gered by an intensive practice of the “combination of letters,” in which he
lost control of his physical and mental faculties. This is doubtless a state
of mind that arouses both exaltation and anxiety. Recent research on the
neurobiology of God beliefs has revealed that the neural correlates of such
experiences are electrical perturbations of the temporal lobe. The various
kinds of acrostics in such formulaic poetry involve the manipulation of let-
ters, with no need to be exposed to electrical perturbations or loss of volun-
tary control. This may epitomize the dynamics that govern the process of
fossilization from threatening experiences to harmless ornaments.
301
1
3
CHAPTER 6
Figurative Language and Sociocultural

Background
Hebrew Poetry as a Test Case
A COGNITIVE APPROACH: POETIC LANGUAGE, MANNERISM,

AND SIGN RELATIONSHIPS
The term Mannerism is used in three different but related senses: (1) exces-

sive use of some distinctive, often affected style in art, literature, speech,
or behavior; (2) the name of the cultural period between the Renaissance
and the Baroque, characterized by complexity, ingenuity, and refine-
ment; (3) other cultural periods of a similar nature (e.g., some trends in
medieval art and literature and modernism). Some scholars (e.g., Wylie
Sypher) claim that Mannerist styles tend to recur in certain types of cul-
tural periods. Toward the end of this chapter I will attempt to account for
this association. This chapter explores the nature of figurative language
in some medieval and modernist Hebrew poetry in relation to a wider
historical-cultural background and makes the assumption that the cog-
nitive system of poets and readers mediates between them. To make my
argument more reader-friendly to an English-reading audience, I will occa-
sionally draw on English literature and literary theory. The view adopted
here does not regard the cognitive mechanisms underlying aesthetic
responses as autonomous but, rather, as adaptation devices turned to
aesthetic ends.
In typically Mannerist periods, people live in a world in which they are
forced to adapt very quickly to situations that have very different internal
321
logics. People usually do this with amazing ease. But, en route, something
happens to them, as the following joke confirms. A mental asylum inmate
is pulling a shoe box on a long string and says: “Come on, come on, come
on!” A doctor who is passing by says, “What a lovely dog you have there,
what is its name?” “Are you crazy, doctor, don’t you see this is merely a
shoe box?” “Oh yes, now I realize, it is a shoe box. Why are you pulling
it on such a long string?” “Because I am afraid it will bite me.” This joke
involves several reversals of mental sets and a readiness to enter, one after
the other, into several situations, each of which has an internal logic that
differs from the preceding one. In each such reversal, the reader receives a
shock but is able to readjust immediately to the new situation. The shock
and the ensuing readjustment are perceived as wit. Although in real life
attention is directed toward the shifting situation appraisals that affect
one’s well-being, in the joke the audience’s attention is directed away from
them to their phenomenological quality. This is what I mean throughout
this book by the statement that “in the response to literature, cognitive
devices evolved for adaptive purposes are turned to aesthetic ends.”
Metaphysical conceit deals with grave issues in a witty mode, such as
the pangs of love or the existential problems of facing Divinity. When
experiencing the processes involved, readers may feel reassured that their
adaptation mechanisms function properly and derive pleasure from this
reassurance. It should be noted that in Excerpts 3 and 4 below the conceits
develop, simultaneously, in two different situations, one in the physical
and the other in the spiritual domain, following radically different logics.
Nevertheless, the two situations are closely related, owing to the fact that
they consistently develop two respective aspects of one image. Thus the
mechanism underlying thought, the image, becomes the center of aware-
ness. A similar story, with minor changes, fits Excerpts 7–12 (below).
This approach sees language as a hierarchical system of signifiers
and signifieds: from the point of view of reading, for instance, a system
of graphemes (letters) on paper signifies a system of phonemes (speech
sounds). Combinations of phonemes designate semantic representation
units; semantic representation units, in turn, are used to refer to entities in
extralinguistic reality. Individuals, as sign-using animals, are programmed
to get as quickly as possible from the signifiers (which have little survival
value) to the signifieds (which may have high survival value). When we read
the instructions for using some piece of electrical equipment, we remember
the instructions rather than the type font of the letters. After some time we
may not even remember what language the instructions were in. According
to Roman Jakobson (1960), poetic language typically directs attention back
to the various signifiers and forces us to linger at them. Meter and rhyme

31
induce readers to attend to the phonological system; this is perceived as

musicality of one kind or another or as witticism. This difference between
musicality and witticism is, I will argue, a matter of style. At another level,
figurative language complicates semantic information processing and pro-
longs a reader’s dwelling on the meaning units.
Here I propose to modify Jakobson’s approach. Attending back to lan-
guage is not a unitary phenomenon. The issue at stake is to what extent
the signified becomes subordinated to the signifier when attending back
to it. As long as the signified preserves its dominance to some extent, the
emphasis on the signifier is more natural, less “marked.” When the rela-
tive weights of the signifier and the signified are balanced or the former
overrides the latter, the process is perceived as less natural, more artificial,
more “marked” (see, e.g., Excerpt 3 below).
This reasoning is not the invention of modern semiotics or cognitive
poetics. It is inherent, in a less fine-grained version, to neoclassical aesthet-
ics. The neoclassical theoretician Joseph Addison gives an account of the
relative naturalness of poetic devices in the following terms:
As true Wit generally consists in this Resemblance and Congruity of Ideas, false Wit
chiefly consists in the Resemblance and Congruity of single Letters, as in Anagrams,
Chronograms, Lipograms, and Acrosticks: Sometimes of Syllable, as in Ecchos and
Doggerel Rhymes: Sometimes of Words, as in Punns and Quibbles; and sometimes
of whole Sentences or Poems, cast into Figures of Eggs, Axes or Altars. . . . As true
Wit consists in the Resemblance of Ideas, and false Wit in the Resemblance of
Words, according to the foregoing Instances; there is another kind of Wit which con-
sists Partly in the resemblance of Ideas, and partly in the Resemblance of Words;
which for Distinction Sake I shall call mixt Wit. This Kind of Wit abounds in Cowley,
more than in any Author that ever wrote. Mixt Wit is therefore a Composition of
Punn and true Wit, and is more or less perfect as the Resemblance lies in the Ideas
or in the Words. (1951 [1711–1712]: The Spectator, no. 62)
The Acrostick was probably invented about the same Time with the Anagram,
tho’ it is impossible to decide whether the Inventor of the one or the other were
the greater Blockhead. (1951 [1711–1712]: The Spectator, no. 60)
Neoclassical poetic theory detested mannerisms but was well aware of

the qualitative difference between poetic devices that consist of the pat-
terning of signifiers and those that consist of the patterning of signifieds.
From Addison’s account, a hierarchy of signs can be abstracted that goes
from letters (graphemic signifiers), to syllables (phonological signifiers),
to the ideas conveyed by words. Poetic devices consist of the patterning
of signs at different levels of this hierarchy. The placement of patterning
F i g u r at i v e L a n g ua g e a n d S o c i o c u lt u r a l B ac k g r o u n d [ 133 ]
341
in this hierarchy of signs determines its stylistic nature: false wit consists

of the patterning of graphemic and phonological signifiers (“letters” and
“syllables,” respectively), and true wit consists of the patterning of “ideas”
signified by them. Mixed wit is located in the middle: it consists partly of
the resemblance of ideas and partly of the resemblance of words. Thus neo-
classical poetic theory derived a normative principle from this hierarchy
of signs, which can be translated into quantitative descriptive terms: The
greater the proportion of signifieds in the poetic device structure, the more
perfect it is; the smaller, the more objectionable. The device is more or less
perfect depending on whether the resemblance lies in the ideas or in the
words. Picture poems (such as Herbert’s “Easter Wings”) thus were prob-
ably the worst of all. In terms of this model one cannot distinguish, for
example, between legitimate rhymes or alliterations, on the one hand, and
“doggerel rhymes,” on the other, but the principles appear to be quite clear.
I believe that bad classicism (“false wit”) can make excellent mannerisms
and that these evaluative terms can be translated into descriptive terms with
great profit. Instead of “and is more or less perfect” we could read “and its
focus is more or less integrated” or “conforms with neoclassical or Mannerist
taste depending on whether the resemblance lies in the ideas or in the
words,” so as to make Addison’s formulation fit perfectly into our scheme.
In this way, one can make illuminating generalizations on mannerisms at
their best, based on the theoretical writings of their fiercest opponents, the
neoclassicists.
Such an appraisal thus is not purely subjective, since the difference can
be described in terms of structures and relative distances. For instance,
attention to the semantic or phonological units by artistic means is per-
ceived as more natural than attention to the string of graphemes on paper.
In fact, the semantic signifier is one step removed from the signified in the
extralinguistic world, the phonological signifier is two steps removed, and
the graphemic signifier is three steps removed.
Furthermore, two kinds of strategies are involved in manipulating sig-
nifiers. In the sound stratum of poetry, when John Donne writes, “When
thou hast done, thou hast not done [Donne],” the ambiguity of the pho-
neme sequence /dʌn/ is crucial for the meaning and is perceived as meta-
physical wit. In neoclassic or Romantic poetry the patterning of sound
clusters is backgrounded to some extent and is perceived as musicality (cf.
Excerpts 1–2, below). Likewise there are two kinds of strategies in manipu-
lating graphemic signifiers: the segmentation of lines in poetry and calli-
grams (picture poetry). One of the differences between poetry and prose is
that in prose, the end of the printed line is determined by the margin of the
page, whereas in poetry it is determined by the poet. This graphic device is

351
subordinated to the rhythmic organization. Its task is to give the reader

instructions on how to perform the text in terms of intonation contours
and, in the last resort, to draw attention to the verse line as a perceptual
unit. Calligrams are picture poems in which strings of letters are arranged
in designs that may or may not be related to the content of the poem. In
this way, the arrangement of the graphemes on the page may take on a
higher degree of independence in the reader’s perception. Consequently,
calligrams are usually considered particularly artificial poetic forms (see
Addison 1951 [1711–1712]; Pagis 1977; Tsur 1997, 2000a, 2003: 199–230).
Thus in various styles the signifiers and signifieds may be integrated into
a hierarchy, or the signifier may acquire some independence, causing the
work of art to disintegrate to some extent or another.
In art history there are periods in which works of art integrate their
material to a relatively high degree, effecting a harmonic disposition of
parts. In other periods, the harmonic dispositions disintegrate, and in
poetry the images have a considerable degree of independence, even domi-
nance. Such styles are frequently tagged “manneristic,” and the application
of the semantic and phonological material is perceived as witty.
Sypher (1955) claims that seventeenth-century Mannerism had its ori-
gins in the disintegration of Renaissance style, and there is a view that,
likewise, some varieties of modernism have their origins in the disinte-
gration of the Romantic style. Within the Mannerist style there are two
subcategories: Metaphysical style, in which the poetic devices refer to a
complex, problematic existence full of tensions and preserve their expres-
sive force, and the Précieux style, in which the same devices have lost their
expressive force to a considerable extent and have become playful orna-
ments referring to an existence from which complexities and tensions have
been practically purged. Medieval Hebrew poetry has ample recourse to
both kinds of Mannerism.
MANNERIST AND “PLATONIC” POETRY
To elucidate the relationship between the relative integration of meanings

and their disintegration, let us take a look at the following two excerpts:
Excerpt 1.
‫ן–א ִבינ ַֹעם ְּב ַמיִ ם ּכַ ִּב ִירים‬

ֲ ‫ם ּובא ֶּב‬
ָ ַ‫וְ נָ ַפל ָה ַרע‬
vǝnafal haraʕam uvaʔ bεn-avinoʕam bǝmayim kabirim
And the thunder fell and the son of Avinoam came with mighty waters
361
Excerpt 2.
‫בֹורה‬ ָ ‫עֲ ֵׂשה לָ ֶהם ּכְ ִס ְיס ָרא וַ עֲ ֵׂשה לִ י ּכְ ַמ ְע ֶׂשיָך לְ ָב ָרק ְּוד‬
‫ֹאבדּו ִמּגְ ָע ָרה‬
ְ ‫ֹלהי וַ ֲה ִפיצֵ ם ּגְ ַער ָּב ֶהם וְ י‬ ַ ‫ְּברֹק ָּב ָרק ֱא‬
ʕ
ase lahεm kǝsisraʔ vaʕase li kǝmaʕsεkha lǝvaraq udǝvora
bǝroq baraq εlohay vahafiṣem gǝʕar bahem vǝyʔovdu migǝʕara
ʔ
Do to them as to Sis’era, and to me do as to Baraq and Debora

Send lightning to them, my God, and disperse them,
Vituperate them, and let them perish by vituperation. (my translation)
One of the characteristic stylistic devices of medieval Hebrew poetry is

what is known as “inlay language,” which generates a kind of intertextu-
ality. Poets derive whole phrases from a biblical verse, frequently in an
altered sense. This results, in most instances, in interesting semantic inter-
actions between the original and acquired senses of the phrases or between
their earlier and new context. In Excerpts 1 and 2, the phonological string
[baraq] refers to a pair of accidental homonyms: “lightning” and a proper
name in the Bible. Both excerpts allude to the victory of Debora and Baraq
the son of Avinoam over Sis’era, recounted in Judges 4–5. The first excerpt
is part of a poem on rain after a long drought by Shlomo Ibn Gabirol; the
other is quoted from a war poem by Shmuel Hanagid. This poet was the
commander in chief of the army in a Muslim emirate in Spain and wrote
this poem on the eve of one of his crucial battles. Both excerpts are based
on the fact that the phonological string /baraq/ has two unrelated signi-
fieds referring to two very different entities in the respective projected
worlds of the poems.
In Excerpt 2, the proper names Debora and Sis’era direct attention to
the biblical story in which “Baraq” occurs as well as a proper name: Baraq
the son of Avinoam. The string of graphemes “Baraq” signifies the pho-
neme sequence /baraq/, perceived as a proper name that designates the
hero of the battle against Sis’era, inspired by Debora the prophetess. In
this context, readers are inclined to abstract some human significance from
the story such as “great victory.” The poem centers on the speaker’s hopes
of achieving a great victory on the battlefield. In the background, readers
may perceive that the text directs attention to the phonological sequence
/baraq/, beyond what is needed for comprehension of the human signifi-
cance. The same phonological sequence recurs at the beginning of the next
verse. Since readers are focusing their attention on the human significance,
the similarity of sounds reverberates as musicality only in the back of the
mind. Not so in Excerpt 1. The phonological sequence [baraq] does not occur

371
in it at all, only the patronymic of the person called “Baraq.” However, the
ear trained on parallelisms and medieval inlay language will find the paral-
lelism of “the thunder fell” and the explicit “the son of Avinoam came” less
than satisfactory. Indeed, we are confronted with a school example of what
in medieval Hebrew poetics is called a “hidden pun.” Without an aware-
ness that the unmentioned word Baraq has two unrelated meanings, the
verse line cannot be understood. So, readers are compelled to suspend their
“communicative competence” and exit the chain of signifiers and signifieds.
They must pass sidewise from the explicit patronymic “the son of Avinoam”
to the unmentioned part of the proper name, “Baraq,” and become aware,
first, that this name is the signified of the phonological sequence [baraq];
second, that the phonological sequence has another semantic signified as
well, unrelated to the previous signified, meaning “lightning”; and, third,
that this semantic signified has a good fit with a nature description. In
Excerpt 2 readers may comprehend the meaning of the two verse lines even
in translation into a language in which the similarity of sounds does not
exist. In the Hebrew original, the phonological sequence will reverberate
in the back of readers’ minds as musicality: At most they will regard it as
paronomasia (a play on words, pun); at worst, they will miss the sound
patterns altogether but will get the message. In Excerpt 1, by contrast,
one may not attend away from the phonological sequence [baraq] and its
two sign functions: if one attends away from them, the entire verse line
becomes meaningless. Furthermore, one cannot give rules as to at which
point one must exit the ordinary semiotic chain of signifiers and signifieds
to construct a meaningful utterance. Consequently, the expression is more
similar to a riddle than to a metaphor. The description is received as wit or,
rather, as Mannerist witticism.
So far I have made two crucial distinctions. One concerns poetic and
nonpoetic language. Both are based on a hierarchy of signs. In the latter,
however, readers or listeners attempt to reach the final referent as fast
as possible by attending away from the signifiers. In the former, they are
forced, by various poetic devices, to attend back to the signifiers. The sec-
ond distinction concerns types of poetic style. In Excerpt 1 (the Mannerist
style type), the signifiers and poetic devices are more obtrusive than in
Excerpt 2. The difference can be described in structural terms or in terms
of the duration of delay. Assuming that in both style types the reader or
listener must linger longer on the chain of signifiers than in nonpoetic
language, in Excerpt 2 the reader may, eventually, attend away from the
verbal expressions to an abstraction of human significance. The concrete
images are perceived as illustrating some idea. John Crowe Ransom (1951)
dubbed such poetry “Platonic poetry.” In Excerpt 1, the structure of the text
381
prevents the reader from attending away from the concrete verbal images
to an abstraction. This is Mannerism of the Précieux type. “Romantic” and
“classic” styles are usually regarded as opposites. But in this respect, in con-
trast to Mannerism, they are similar in that both are typically “Platonic.”
In the two preceding excerpts, the text forces the reader to attend back
to the phonological signifier to some extent. In the following two excerpts,
both of the “Metaphysical” kind, the reader is forced to attend back to the
semantic component:
Excerpt 3.
‫י ּד ְמ ִעי וְ ִּת ְׁש ָט ֵחם ְּב ֶׁש ֶמׁש זָ ֳה ָרּה‬
ִ ‫עָ ְפ ָרה ְּתכַ ֵּבס ְּבגָ ֶד ָיה ְּב ֵמ‬
ʕ
Ofra tǝkhabbes ʔεt bǝgadεha bǝmey dimʕi vǝtištaḥem bǝšεmεš zohora
Ofra [the Gazelle] washes her clothes in the waters of my tears
and spreads them against the sun of her splendor (Yehuda Halevy;
my translation)
Excerpt 4.
‫הֹויֶ ה‬ ‫לְ ָך ַה ְּמצִ יאּות ֲא ֶׁשר ִמצֵ ל ְמאֹורֹו נִ ְהיֶ ה ּכָ ל‬

‫ֲא ֶׁשר ָא ַמ ְרנּו ְּבצִ ּלֹו נִ ְחיֶ ה‬
lǝkha hamǝṣiʔut ašεr miṣel mǝʔoro nihyε kol hoyε
ʔ
ašεr ʔamarnu bǝṣilo niḥyε
Yours is the reality whose light’s shadow generated all existing things
of whom we said: under its shadow we shall live. (Shlomo Ibn Gabirol;
my translation)
In these excerpts as well, the reader is forced by similar means to attend

to the verbal signifier, but the weight of the process is shifted from the
phonological to the semantic component. In the word tears or shadow
the phonological sequence [dimʕa] or [ṣel] has only one semantic signi-
fied, but it is developed in a unique way. “Tear” is a physical entity that
represents, by way of metonymy, an intense human emotion, usually
sorrow. In medieval Hebrew (and Arabic) poetry, “Ofra” (the gazelle) is a
conventional epithet for the female object of the lover’s yearning (when
the lover yearns for a male object, he is called “Ofer”). In such a context,
“tears” suggest the pangs of love (not unlike in Petrarchan poetry). The
physical properties of tears, composed of water and salt, are irrelevant
to the representation of emotions. Yehuda Halevy, however, exploits

391
precisely one of these properties: “Ofra washes her clothes in the waters

of my tears.” By doing so he emphasizes the abundance of the tears,
that there are enough tears for washing one’s clothes, suggesting a very
intense emotion. By the same token, he merges two different realities in
the same physical object: love and clothes washing. Moreover, these two
realities display opposite emotional tendencies: noble sentiment and a
baser activity. This stylistic phenomenon has an interesting ontological
concomitant in that the two realities cannot coexist except on the verbal
level. One cannot, therefore, attend away from the linguistic structure
to the crucial human significance. In Addison’s terms, this is a similarity
of words rather than similarity of ideas, and therefore we cannot attend
away from the words.
The image “and spreads them against the sun of her splendor” is
similarly treated. Here again, the point of departure is a cliché of love
poetry: Ofra is as beautiful as the sun. Two different properties of the
sun are exploited here—the splendor and the heat. The splendor serves as
a metaphor for Ofra’s beauty; the fact that the sun radiates heat is irrel-
evant to this metaphor. But lo, Yehuda Halevy again exploits precisely this
aspect of the sun and, again, in the clothes-washing context, since one can
dry wet clothes in the sun. Once again we are confronted with two “het-
erogeneous ideas yoked together by violence,” to use Dr. Johnson’s (1951
[1779]) definition of metaphysical conceit: namely, conflicting emotional
tendencies, again, a combination that can exist on the verbal level only but
not in extralinguistic reality. The tear image and Ofra’s splendor reinforce
each other in the “love” universe of discourse, as do the clothes-washing
and clothes-drying images in the “laundry” universe of discourse. In
both images, the universe of discourse related to clothes washing is no
less real than the one related to love. Moreover, the combination of the
two images in a single description may enhance both the “love” universe
and the “clothes-washing” universe in our perception. The two universes
vie for readers’ attention and reassert themselves in their perception. As
a result, both are perceived as more vivid and take on a witty effect. In
James Smith’s formulation, “The elements of the conceit must be such
that they can enter into a solid union and at the same time preserve their
warring identity” (1933: 234).
Some readers attempt to subordinate the universe of clothes washing
to the universe of love and its pangs. The washing image serves merely as
a hyperbole, to demonstrate the enormous amount of tears shed and the
immense splendor of Ofra’s beauty. Such an attitude regards the washing
image and the sun image as merely more or less appropriate illustrations
4
01
of the enormous amount of tears and Ofra’s exceptional beauty. It allows

the reader to escape the excruciating experience of conflicting emotional
tendencies and the need to stay in two different figurative universes with
conflicting logics at the same time. Such a “solution,” however, ignores the
unique stylistic character of these images and many other similar ones.
I do not know how reliable the Arabic glosses in the diwans (collections of
secular poetry) of the period are; at any rate, the gloss of this poem refers
to the laundry situation as the actual situation: “Improvisation composed
when he passed the river, where the washerwomen were washing clothes;
and it is one of his most beautiful matters” (my translation).
The first segment of Excerpt 4 alludes to the Neoplatonic conception
of creation by way of light emanation. The universe was created from
a stream of light emanating from a source. The farther away from the
source, the more shadow-like the stream becomes. The nearer the exis-
tence to the light, the more spiritual, the nobler it is; the nearer to the
shadow, the baser, the more material it is. The second segment is a literal
quotation from Lamentations 4:20. In the present context “its shadow”
suggests something like “protection” and ultimately “Divine providence”
(in Lamentations it refers to the protection of “God’s anointed”). Whereas
in the first segment “shadow” is regarded as less valuable than “light,” in
the second segment another physical potential of shadow is exploited: it
provides protection from the destructive forces of light and heat. Thus,
again, different realities displaying conflicting emotional tendencies are
embodied in one image. Moreover, the shadow image yokes two incom-
patible conceptions of the creation together by violence: the Neoplatonic
view of creation by emanation and the biblical view of creation by a per-
sonal creator.
The figurative expressions discussed in Excerpts 3 and 4 are fine and
typical examples of what in seventeenth-century English poetry is called
“metaphysical conceit.” Now let us take a look at a complete poem by
Shlomo Ibn Gabirol:
Excerpt 5.
‫זָ נ ַֹח‬ ‫אתיָך ֲא ִבי‬
ִ ‫ֹד עד ּכִ י ְק ָר‬
ַ ‫ָאכֵ ן ְמיֻ ָּדעִ י זְ נַ ְח ַּתנִ י ְמא‬
‫נֹוח‬
ַ ‫ׁשֹוטט ָּב ֱאנֹוׁש ל ֹא ָמצְ ָאה ּכִ י ִאם ְּבָך ָמ‬ ֵ ְ‫יֹום ֻׁשּלְ ָחה נַ ְפ ִׁשי ל‬
‫נ ַֹח‬ ‫ַּב ֶּמה ְּתכַ ֶּבה ַא ֲה ַבת לֵ ָבב?—וְ ֵאל נִ ְׁש ַּבע ְּברּוחֹו ֵמ ֲעבֹר ֵמי‬
ʔ
akhen mǝyudaʕi zǝnaḥtani mǝʔod ʕad ki qǝraʔtikha ʔavi Zanoaḥ
yom šulǝḥa nafši lǝšotet baʔɛnoš loʔ maṣǝʔa ki ʔim bǝkha manoaḥ
bamɛ tǝkhabbɛ ʔahavat levav?—vǝʔel nišbaʕ bǝruḥo meʕavor mey Noaḥ

1
4
Indeed, my friend, you have neglected me very much

So that I called you father of Zanóaḥ [father of neglect].
Since the day my soul was sent forth to wander among humans,
She found no place of rest but in you.
With what will you quench the love of the heart?—as God
Swore by His Soul that the waters of Noah should no more go over [the
Earth]. (my translation)
The phrase “father of Zanóaḥ [father of neglect]” is the most complex

and the most effective. It involves a complex version of the riddle-like
device encountered in Excerpt 1, and at the same time, it requires a
reading that activates some human significance of great importance,
as encountered in Excerpt 2. In the “human significance” reading, the
noun phrase “father of” suggests the first and foremost prototype, one
that originates; the infinitive zanóaḥ suggests the abstraction “neglect.”
The “riddle” reading relies on another (in fact, main) meaning of father
as “male parent” and the fact that the phonological sequence /zanóaḥ/
occurs in the Bible as a proper name: “Yekutiʔel, the father of ‘Zanóaḥ’ ”
(1 Chronicles 4:18). Only the son’s name is mentioned for the sake of
its homonym for the human significance: neglect. However, the omitted
member of the name also has a crucial function in that it indicates the
name of the poem’s addressee. From Ibn Gabirol’s poetry we know that he
had a close friend and benefactor called Yekutiʔel Ibn Ḥassan. Eventually
Ibn Ḥassan was murdered or executed, and Ibn Gabirol lamented his
death in his poetry. This poem seems to concern some minor com-
plaint of the poet. Thus, Yekutiʔel Ibn Ḥassan the person becomes the
father of Zanóaḥ in both ways: according to the “human significance”
reading, he is the very incarnation of neglect, and with the help of the
riddle meaning of the inlay, the poet indicates the addressee’s name
of Yekutiʔel.
The second verse line alludes to the Platonic conception of friendship.
According to Plato, friendship is halfway between absolute good and abso-
lute bad. In an ideal good world there is no need for friendship. In a world
in which man is a wolf to man, friendship is impossible. Friendship gives
a feeling of stability and permanence in a stormy, unstable world. Now
consider verse 2: “Since the day my soul was sent forth to wander among
humans, / She [It] has found no place of rest but in you.” These lines pres-
ent the soul of the speaker in a world of turmoil, obtaining rest and stabil-
ity only in friendship. The words “sent forth . . . found no place of rest” are
derived from the story of Noah’s ark: “Then he sent forth a dove from him,
421
to see if the waters had subsided from the face of the ground; but the dove
found no place [of rest] to set her foot” (Genesis 8:8–9). This allusion gen-
erates a metaphorical interplay between the Flood story and the verse line.
Noah’s ark, the Flood, and the waters are deleted, thus foregrounding and
amplifying the feeling of stormy unrest and instability in the “wanderings”
of the soul among humans.
But what has been deleted—the many waters, Noah, and the
Flood—is restored in two additional powerful metaphysical conceits in
verse line 3: “With what will you quench the love of the heart?—as God /
Swore by His Soul that the waters of Noah should no more go over.” The
first hemistich alludes to the verse “Many waters cannot quench love, /
neither can floods drown it” (Song of Songs 8:7). In fact, the hemistich
contains a rhetorical question, to which an obvious answer would be the
biblical verse. The verb quench transfers to “love” the transfer feature
<fire>, thus rendering love intense like a fire that consumes the lover.
The phrase “Many waters cannot quench” suggests two different possible
construals: (1) love is a kind of fire that water cannot affect, or (2) love
is a fire that can be affected by water, but in this case it is so power-
ful that no amount of water can extinguish it. The verse in the Song of
Songs seems to suggest the first possibility. The question “With what?”
in the poem apparently proceeds in the same direction. Then, however,
the sequel creates a surprise: Many waters may perhaps quench love,
but in the present instance you do not have that amount of water; my
love for you is so powerful that a flood is needed to extinguish it. This,
however, is impossible, owing to God’s covenant that “never again shall
there be a flood to destroy the earth” (Genesis 9:11). In this instance, Ibn
Gabirol uses a verse from Isaiah as an inlay that, in turn, alludes to the
verse in Genesis:
For this is like the days of Noah to me:

as I swore that the waters of Noah
should no more go over the earth. (Isaiah 54:9)
This suggests yet another twist: In Ibn Gabirol’s poem the flood is not
merely a means to destroy the Earth; it is beneficial, and God’s covenant
is harmful.
The most outstanding English metaphysical poet of the seventeenth
century, John Donne, cannot be suspected of ever having heard of Ibn
Gabirol or read his poems. But in one of his Holy Sonnets he develops the
same metaphysical conceit and in a very similar manner:

1
43
Excerpt 6.
You which beyond that heaven which was most high
Have found new spheres, and of new lands can write,
Pour new seas in mine eyes that so I might
Drown my own world by my weeping earnestly,
Or wash it if it must be drowned no more.
Here as well many tears serve as a metonymy for intense emotion; all the seas
would not be enough to have the required amount of water for the tears. Then
Donne alludes to a passage from 2 Peter in the New Testament referring to
the destruction of the world by water and fire and to God’s covenant to Noah.
MANNERISM AND SOCIOCULTURAL BACKGROUND
We have seen several characteristics of the Mannerist-Metaphysical style in

medieval Hebrew poetry. As Ernst Robert Curtius (1953) pointed out, there
have been several periods of Mannerist style in the history of European liter-
ature from the Middle Ages to our times. Some scholars (e.g., Wylie Sypher)
claim that the Mannerist style, especially in its Metaphysical variety, tends to
occur in sociocultural periods marked by strife and social instability, in which
the world picture has been undermined or is dominated simultaneously by
more than one scale of values, with no firm foundation to decide between
them. The most conspicuous periods from this point of view are the golden
age of Hebrew poetry in Spain, the seventeenth century in Western Europe,
and the twentieth century. In the seventeenth century, the great events that
shook the spiritual world of the sixteenth century permeated into the con-
sciousness of the wider intellectual strata. From a religious point of view,
this was the period when the hegemony of the Roman Catholic Church was
considerably weakened. The Council of Trent, which continued intermit-
tently between 1545 and 1563, declared the Counter-Reformation. In terms
of cosmology, Copernicus (1473–1545) and Galileo (1564–1642) shook the
Ptolemaic world picture that had dominated the world for centuries. In the
poetry of the great seventeenth-century English poets such as John Donne,
for instance (in, e.g., Excerpt 6 above or in such lines as “The new philosophy
calls all in doubt,” “’Tis all in peeces, all cohaerance gone”), we hear echoes of
the great spiritual shock caused by the enormous geographic and astronomi-
cal discoveries of the era, beginning with the discovery of America. In Donne
and Milton we encounter the Copernican and the Ptolemaic world pictures
side by side in the same texts (cf. Milton, Paradise Lost, VIII).
41
In eleventh-century Spain things are less clear. But the encounter of the
three great monotheistic religions in Spain generated tensions, with no solid
ground for preferring one religion over another. Yehuda Halevy’s treatise
Hakuzari (The King of Khazars) is an illuminating document from this point
of view. The Khazars were a nation who founded an empire in the seventh
century in the northern Caucasus along the Caspian Sea, where over time
Judaism became the state religion. Halevy’s treatise is a Platonic dialogue
that tells the fictional story preceding the king’s conversion and discusses a
wide range of controversial issues, from religious to cultural. The king invites
a Christian, a Muslim, and a Jewish scholar and questions them about their
religion to find out which one is the true religion. The mere need to write
such a treatise highlights the problems facing a society in which not one but
three religions vied for primacy. For our purposes it makes little difference
that in the treatise of the Jewish poet-philosopher the Jewish position wins
the day. It is the very need to convince a neutral outsider that is significant.
From a literary point of view it is important to note that Yehuda Halevy
had an attitude of cultural ambivalence even toward poetic prosody. As
is well known, he was one of the greatest poets (if not the greatest) who
wrote Hebrew poetry in the pegs-and-chords meter imported from Arabic
poetry.1 But he has his Jewish scholar say harsh things about the inferiority
of the pegs-and-chords meter, as compared with the prosody of the book
of Psalms (cf. Rosen 1994).2 Furthermore, as Excerpt 4 suggests above
(as well as some other poems by Ibn Gabirol), poets and educated readers
of the time had to cope with a perplexing problem in theology: namely,
the need to integrate the biblical concept of a personal Creator with the
Neoplatonic conception of creation as an emanation of light.
This approach to the sociocultural background of Hebrew poetry in
eleventh-century Spain gets massive support from Ross Brann’s 1991 book
The Compunctious Poet: Cultural Ambiguity and Hebrew Poetry in Muslim Spain:
As Raymond Scheindlin has put it, the courtier-rabbis simultaneously lived

according to “two opposing principles of life.” On the one hand, they were enam-
ored of the world of material culture and its indulgent pleasures and beauty; and
on the other, they were devoted to the world of religion with its emphasis on
1. In medieval Hebrew poetry in Spain, meter, imported and adjusted from con-
temporary Arab poetry, was based on full vowels (termed “chords”) alternating in all
sorts of combinations with units consisting of a schwa mobile plus a full vowel (termed
“pegs”). The schwa mobile is, in Semitic languages, a reduced vowel, similar to the first
vowel in English against.
2. The critical sources for evaluating Halevy’s ambivalence toward Hebrew verse are
the two complementary texts: book 2 of The Kuzari (64–78) and the Judeo-Arabic
“Treatise on Hebrew Meters” (Brann 1991: 96).

1
45
divine judgment and life in the hereafter. The ideal man lived in both worlds and
found the ambiguity most attractive. (1991: 11)
On one point, however, Brann expresses a view opposite to the one underly-
ing the present work: “As liturgical poets, the courtier-rabbis of the Golden
Age therefore strike us as ‘poets of high seriousness’ seeking truth, serving
the community of Israel, and giving voice to its spiritual yearnings. But as
secular poets they present themselves to us as ‘poets of wit and whimsy’ in
the service of courtly ideals” (1991: 10).
As Excerpt 4 above indicates, some of the best samples of medieval devo-
tional poems are examples of metaphysical wit. Elsewhere (Tsur 2008b)
I have compared a devotional poem by Shlomo Ibn Gabirol with a poem by
Levy Ibn Altaban offered as a tribute to Ibn Gabirol. They have the same
monorhyme, the same meter, and similar phrases for opening and clos-
ing. Yet the two poems differ in two crucial respects: Ibn Altaban’s poem is
far inferior to its model, and it is typical of “Platonic” poetry, whereas Ibn
Gabirol’s is an exquisite specimen of metaphysical poetry. Perhaps, how-
ever, all the disagreement between Brann and me is a matter of terminology
rather than essence. In the terminology of the present chapter, some of the
best specimens of liturgical poetry “seek truth” in the style of Metaphysical
wit, whereas the majority of secular poetry of love, wine, and garden descrip-
tions are in the Précieux line. Brann does not go into such stylistic details,
and perhaps he does not use “high seriousness” in Matthew Arnold’s sense.
THE “ARBITRARY GENITIVE”
We will need to account for the relationship between the kind of poetics
discussed here and the kinds of sociocultural periods in which it tends to
occur. But before doing so, I would like to point out a figurative construc-
tion for which Moses Ibn Ezra had a particular predilection—genitive
phrases of the “concrete of the abstract” form:
Excerpt 7.
‫ר ּפ ְרסּו ֲעלֵ י ֹֻשלְ ַחן ְמזִ ָּמה ָע ְרכּו‬
ָ ‫וַ ּיעֲ ִבירּו קֹול לְ לֶ ֶחם ָּדת ֲא ֶֹש‬
‫עּודה ָמ ְסכּו‬ָ ‫ שֹכֶ ל וְ ֶאל יֵ ינֵ י ְת‬
ֵ ‫לְ כּו ֶאל נַ ֲהלֵ י‬ :‫ולִ כָ ל צְ ֵמ ֵאי ִּבין‬
They passed the word to the bread of justice which
they sliced on a table of judgment they set
And to all those thirsting for wisdom: go to the rivers of reason
and to the wines of moral obligation they poured. (my translation)
46
1
Excerpt 8.
‫לֹוח ִמים ָּפ ָרסּו‬
ֲ ‫ֻׁשלְ ַחן ּכְ ִזָבים עָ ְרכּו בֹו ַמ ֲאכָ ל ֶ ק ֶׁשר וְ לֶ ֶחם‬
They set a table of lies with food of conspiracy
on which they sliced bread of animosity. (my translation)
In her analysis of metaphoric constructions, Christine Brooke-Rose (1958)

called such genitive phrases “Pure Attribution.” She comments that “the
really interesting thing about Pure Attribution is that it expresses a some-
what artificial split of one idea into two terms which are basically identi-
cal” (1958: 162). In my Hebrew writings I call them the “arbitrary genitive,”
yoking together the most heterogeneous ideas by violence, to utilize
Dr. Johnson’s (1951 [1779]) phrase.
In Excerpt 7 there is a series of genitive phrases: “bread of justice,”
“table of judgment,” “thirsting for wisdom,” “rivers of reason,” “wines
of moral obligation.” In these concrete of the abstract phrases the
concrete member belongs to the semantic field of eating and drink-
ing: bread, table, thirsting, rivers, wines; whereas the abstract member
belongs to the semantic field of spiritual and moral abstraction: justice,
judgment, wisdom, reason, moral obligation. The concrete words do not
illustrate the abstract words; they are assigned to them in an arbitrary
manner. Still, a solid unity is achieved because all the concrete terms
are taken from one semantic field and all the abstract terms are taken
from another. The terms taken from one semantic field reinforce each
other so as to preserve their “warring identity.” Each one of the rival
semantic fields attempts to establish itself in the reader’s perception
so that the concrete is perceived as more concrete and the abstract is
perceived as more abstract, generating wit. The poet resorts to another
verbal device to prevent the metaphors from being rendered “Platonic”
by subordinating the concrete images to the abstractions: the relative
clauses “they sliced,” “they set,” and “they poured” are conspicuously
applicable solely to the concrete objects to reinforce their actual exis-
tence. However, the eating and drinking images are “yoked together”
once with positive moral abstractions and once with negative ones.
They do not illustrate these abstractions; rather, after the event the
abstractions cause them to assume their respective healthy, life-giving
quality or some negative quality “as a swallowed bait on purpose laid to
make the taker mad.”
Such genitive phrases also abound in the twentieth century in the early
poetry of the Imagist poet Abraham Shlonsky:

471
Excerpt 9.
:‫מֹושיט לִ י ּכֹוס־לְ ַחּיִ ים‬ ִ ‫ּגֹורלֹות‬ ָ ‫ֶמלְ צַ ר ַה‬

.'' ‫ ִמּיֶ ֶקב ַהּנֵ כַ ר‬,‫ ֲאדֹונִ י‬,‫ִ''ילְ ּגֹם נָ א‬
9. The Waiter of Fate offers me a glass “to your health,”
Do sip, sir, from the winery of Strangers’ Land.
Excerpt 10.
,‫ ֵּבית ַהעֶ צֶ ב‬,‫ּגּומיָך‬
ֶ ‫ּכְ ַב ָּנ ִאים עַ ל ִּפ‬
.‫ִמּיָ ד ֶאל יָ ד נִ ְמסֹר ֶאת ֶא ֶבן ַה ֵּטרּוף‬
Like masons on your scaffolding, House of Sorrow,
We pass on, hand to hand, the stone of Insanity.
Excerpt 11.
‫ָשֹמונִ י ּפֹה ּכַ ָּתף לָ ֵׂשאת ַא ְבנֵ י ַהּבֹהּו‬
.‫אֹובים‬
ִ ְ‫ּגומי ַה ַּמכ‬ֵ ‫ִּב ִפ‬
11. I have been made a bearer, to shoulder the stones of Chaos
On the scaffoldings of sufferings.
“Waiter” does not illustrate “Fate,” nor does “winery” illustrate “Strangers’
Land.” Likewise, “stone” was not chosen as an apt representative of
“Insanity” or of “Chaos,” nor is “scaffolding” an apt representation of
“sufferings.” “Stone” was chosen because it fits into a narrative including
masons, bearers, and scaffolding, just as “Waiter” and “winery” make up
a coherent narrative. It is only when the abstract terms (which also rein-
force each other) are inserted that two incompatible universes of discourse
emerge, each striving to establish itself in the reader’s perception. No
“Platonic” representation of ideas can occur.
Ḥayyim Naḥman Bialik was, indisputably, the greatest figure in
so-called modern Hebrew poetry. His early poetry grew out of what is
termed “Enlightenment” poetry, but its main bulk was conspicuously
Romantic. Toward the end of his career he adopted some of the prosodic
features of modernism, as well as some features of Mannerist figurative
language. In the next two excerpts I compare two poems in which the
same image occurs in a “Platonic” structure and an “arbitrary genitive,”
respectively:
4
81
Excerpt 12.
‫י־מעִ י ִמ ַזְּבח לְ ַב ְבכֶ ם ֶֹש ָֹּש ַמם‬
ְ ֵ‫וַ עֲ ל‬
.‫יְ ִֵילִ יל וִ ַיפ ֵהק ֲהתּול ַה ִֹּש ָּממֹון‬
And upon the desolate3 wreck of the altar of your hearts
The cat of desolation will be wailing and yawning. (my translation)
Excerpt 13.
,‫ִּב ְהיֹות יֵ ֵמי ַסגְ ִריר ְּוב ֶנ ֶֹשף ַהּלַ יְ לָ ה‬
,‫ִּב ְֹשלֹוט ְּב ֵבית ָא ִבי ְּד ִמי ַק ְדרּות ּגְ דֹולָ ה‬
‫דֹומם‬ ֵ ְ‫וַ ֲחלָ לֹו ַה ֶמ ֻדּכָ א ּכְ מֹו הֹזֶ ה ו‬
.‫ֹשֹומם‬
ֵ ‫וְ חֹולֵ ם ֲחלֹום עַ גּום ִּבכְ נַ ף ִֹשּקּוץ‬
,‫ ְּד ִמי ַדּלּות ְמנֻ ּוֶ לֶ ת‬,‫זֹה עֲ ַקת ַה ַּד ֲהקּות‬
‎—‫ֹשֹואלֶ ת‬ ֶ ‫ִּבנְ ׂשא ֶֹש ַבע נְ ָפֹשֹות עַ יִ ן‬
;‫ ִּד ְמעָ ה נֶ עֱ צֶ ְרת‬,‫ֹשֹוממֹות‬ ֵ ‫זָ וִ ּיֹות‬
,‫ַעל־ּגַ ֵּבי ַהּכִ ָירה ֲחתּול ְמיַ ּלֶ לֶ ת‬
,‫אֹור—ּב ִמ ֶׁש ֶא ֶרת‬
ַ ‫ ְׂש‬,‫ַּב ַּסל ֵאין פַּ ת־לֶ ֶחס‬
—‫ ּוגְ ִר ִיסין ַּב ִּסיר‬,‫לַ ֲה ַמם ֵאין ּגַ ֶתלֶ ת‬
‫ָאז ֵהצִ יץ ַהּצְ ָרצַ ר ִמּנִ ְק ַרת ַה ִּקיר‬
,‫ ָה ֵר ָיקה‬,‫וִ ינַ ֵּסר ִֹש ָירתֹו ַה ֵיֵב ָֹשה‬
.‫ּבֹוק ָקה‬ ֵ ‫ ִנ ְֹש ָמ ִתי‬,‫קֹוס ָסה לִ ִּבי‬ ְ ‫ּכָ עָ ׁש‬
,‫ ָּבכָ ָתה‬,‫ ל ֹא־נִ ֲח ָמה‬,‫ל ֹא–זָ עֲ ָפה ִֹש ָירתֹו‬
;‫ֹשֹומ ָמה ָהיָ ָתה‬ ֵ —‫גַּ ם־קֹב ל ֹא יָ ָדעָ ה‬
,‫ ּכַ ֲה ֵבל ַחּיַ י ְּת ֵפלָ ה‬,‫שֹומ ָמה ּכַ ַּמוֶ ת‬ ֵ
.‫ ְּבלִ י ַא ֲה ִרית וְ ִתכְ לָ ה ֲא ֵבלָ ה‬,‫וַ ֲא ֵבלָ ה‬
In the days of rainstorm, and in the darkness of night,
With the silence of great blackness, reigning in my father’s house,
Its depressed void like a daydreamer, silent—
Dreaming a grievous dream, under the wings of a desolate abomination
It is the hardship of need, the silence of degenerate poverty,
With seven souls raising questioning eyes—
The corners desolate, the tears stopping
Upon the stove the cat is wailing,
No bread in the basket, no leaven in the kneading trough,
No ember for heating, no groats in the pot,
Then the cricket peeked from the crack in the wall,
3. In Hebrew, “desolate” is expressed by a line-final relative clause that can qualify
either “the wreck,” or “the altar,” or “your hearts” but most likely the whole package.

1
49
And grated his empty, dry song,

Like a moth it gnawed my heart, depleted my soul.
His song was not enraged, or comforting [regretful], it whimpered,
Nor did it know how to curse, it was desolate;
Desolate as death, insipid as the absurdity of my life,
And was mournful—endlessly and boundlessly mournful. (my translation)
Excerpt 12 is constructed exactly like Excerpts 7–11. The simple sequence of

a literal narrative, “And upon the . . . wreck of the altar . . . / The cat . . . will
be wailing and yawning,” is interrupted by two prepositional phrases: “of
your heart” and “of desolation.” They generate a metaphoric contradiction
and two rival levels of reality, each striving to establish itself in the reader’s
perception. In the italicized line in Excerpt 13 we encounter exactly the
same literal narrative, only the “altar” has been shifted from the sacred to
the profane domain.
According to Wimsatt, typical Romantic nature imagery is based on lit-
eral expressions describing the surrounding natural scene, but from these
objects certain qualities can be abstracted that contribute to the emotional
atmosphere of the landscape (1954). In Excerpt 13, desolation and depri-
vation may be abstracted from the concrete description of the scene, in a
typical “Platonic” use of imagery. This is not the case in Excerpt 12, where
two “arbitrary genitive” phrases generate conspicuous logical contradic-
tions; in Excerpt 13, too, the root desolate occurs four times, but in a way
that does not generate logical contradictions as in “the cat of desolation”
or logical and visual contradictions as in “the desolate wreck of the altar of
your hearts.” A considerable part of the emotional atmosphere in the scen-
ery described in Excerpt 13 is due to thing-free and gestalt-free entities
qualified by emotionally charged epithets, as in the lines:
Its depressed void like a daydreamer, silent—

Dreaming a grievous dream, under the wings of a desolate abomination
It is the hardship of need, the silence of degenerate poverty.
Such thing-free and gestalt-free entities do not conflict visually with the
solid objects described. Furthermore, the inner space of a room is absence
that looks as though it were something. When the void, the inner space,
“dreams,” its activity is imperceptible to the senses, but its mental energy
bestows an intense presence on this absence. Thus, desolateness and a cat
on an altar/stove may contribute to metaphysical wit in one poem but to
an emotional atmosphere in another. Significantly enough for the present
argument, Bialik had recourse to such metaphysical figures in a group of
501
poems called “Poems of Anger and Admonition,” in which he responded to

what he regarded as a national crisis of values.
FIGURATIVE LANGUAGE AS AN ADAPTATION DEVICE
At the beginning of this chapter I said that the aesthetic response is based
on adaptation devices turned to aesthetic ends. I also claimed that the
Mannerist style—especially of the Metaphysical variety—tends to occur in
eras and societies marked by strife, radical changes, and disorientation in
which cultural disintegration exceeds a certain limit, there is more than one
scale of values, and there is no agreed-upon basis to decide between them.
Such a description, however, would suit Romanticism as well, for example.
In fact, the Romantic era, on the one hand, and the seventeenth and twen-
tieth centuries, on the other, differ mostly in their relative degree of dis-
integration. Romantic poetry handles disintegration with the help of the
orientation mechanism. The greater the disintegration, the more pointed
the effect of the mechanism of integration and orientation—but only up to
a certain point. Beyond that point, the disintegrating environment escapes
from the orienting mechanism’s control, and a different kind of coping pro-
cedure must be instantiated. Orientation is the ability to locate oneself in
one’s environment with reference to time, place, and people. This process is
marked by speed and relative imprecision. One gathers much information
via all the senses about one’s environment as well as about oneself, inte-
grating it as fast as possible. This is the source of the intuitive and impre-
cise nature of orientation. This makes it possible to only distinguish a few
solid objects; its essence is to find relationships that enable the individual
to make fast decisions regarding overall directions.
When disintegration exceeds a certain point, this mode of adaptation
becomes inadequate. In such cases we must find alternative modes of adap-
tation and orientation. Information “about oneself, like all other informa-
tion, can only be picked up by an appropriately tuned schema,” says the
cognitive psychologist Ulrich Neisser (1976: 116). When, in circumstances
of extreme disintegration, something suddenly seems to go wrong, one has
to check the tuning of one’s own schemata: “Consciousness, according to
Bartlett, enables an organism ‘to turn around its own schemata’ ” (Miller and
Johnson-Laird 1976: 150). I have called such awareness “meta-awareness.”
It should be pointed out that critical philosophy is characterized by similar
terms. Not unlike the poetry of disorientation, it tends to be prevalent in
societies dominated by more than one set of values, where there appear
to be no unquestionable truths, and when philosophy cannot take its own

1
5
assumptions for granted. In critical philosophy, “thought turns around and

examines itself instead of examining its own shadows in the void” (Pears
1971: 30):
When human thought turns around and examines itself, where does the inves-
tigation start? . . . The short answer . . . is that there are two forms in which the
data to be investigated may be presented. They may be presented in a psycho-
logical form, as ideas, thoughts and modes of thought, or they may be presented
in a linguistic form, as words, sentences and types of discourse. Kant’s critique
starts from data of the first kind, and the second wave of critical philosophy, the
logico-analytic movement of this century, starts from data of the second kind.
(Pears 1971: 27–28)
It is noteworthy that Mannerism also “has two modes, technical and

psychological” (Sypher 1955: 116). Sypher speaks of “Donne’s false and ver-
bal (perhaps false? perhaps verbal?) resolutions—his incapacity to commit
himself wholly to any one world or view”: “The resolution is gained, if at all,
only rhetorically, not [through] reason” (1955: 122–123).
The metaphysical conceit deals readers a shock through the deviant
use of language: they cannot pass from the phonological signifier to the
semantic signified and from the semantic signifier to the signified in
extralinguistic reality, as would be demanded by their cognitive apparatus
developed for survival in human society. Instead, the multiplicity of signi-
fieds require them to keep the linguistic signifier in consciousness. When
settling on one signified readers are compelled to abandon the other signi-
fied. This process shocks them out of their linguistic tuning to reality. This,
in turn, compels them to find ways of readjustment. Between shock and
readjustment, their cognitive apparatus is exposed to immediate percep-
tion. In this way, we can account for three kinds of human activities: the
understanding of one’s own conceptual system yields critical philosophy,
the understanding of the perceptual system yields cognitive psychology,
and the immediate perception of one’s own perceptual and/or conceptual
system yields metaphysical poetry.
At the beginning of this chapter I indicated the incompatible worlds and
incompatible logics in which the Mannerist poet lives, with the help of a
joke about a mental asylum inmate who treated a shoe box as a little dog.
But to the doctor, who pretended to treat it, too, as a pet, he said, “Are you
crazy, doctor, don’t you see this is merely a shoe box?” “Why are you pulling
it on such a long string?” “Because I am afraid it will bite me.” The meta-
physical conceit poses more strenuous demands. Rather than switching
between universes of discourse with incompatible logics, it uses one image
521
to convey two or more incompatible universes of discourse. Metaphysical

poetry applies such “metaphysical wit” to such weighty experiences as love
and God experiences, dealing the reader shock after shock.
Alternatively, the same issue can be approached from the direction of
Thomson’s paragraph on Christian Morgenstern:
Morgenstern’s playfulness . . . has a serious side to it. He is on record as claim-

ing that man’s basically unsatisfactory relationship to his fellows, his society
and the world in general stems from his being imprisoned by language, which
is a most unreliable, false and dangerous thing, and that one must “smash” lan-
guage, destroy man’s naive trust in this most familiar and unquestioned part
of his life, before he can learn to think properly. Morgenstern’s brilliantly witty
games with words are thus, seen from this point of view, devious devices of
alienation, and at their most radical succeed in producing in the reader a strange
sensation—making one suddenly doubt one’s comfortable relationship with
language—not unlike the sense of disorientation and confusion associated with
the grotesque. (Thomson 1972: 165)
When Ibn Gabirol in Excerpt 5 has recourse to the phrase “‫”א ִבי זָ נ ַֹח‬
ֲ (father
of Zanóaḥ [father of neglect]) in two different senses that follow two dif-
ferent kinds of logic, the common communicative logic and the “riddling”
logic, or in some of the conceits discussed above, precisely such smashing
of language may occur that deal the reader a shock, followed by recovery
from the shock and readjustment to the universe of the poem.
This view may provide a clue to solving another riddle. Why should a reli-
gious or mystical poet ever want to have recourse to such ingenious, artifi-
cial modes of expression? As I have just argued, the “smashing of language”
deals a shock to readers, making them suddenly doubt their comfortable
relationship with language. This shock is not unlike the sense of disorienta-
tion and confusion associated with mystical paradoxes, as argued by Steven
T. Katz:
Such linguistic ploys exist in many places throughout the world, usually con-
nected with the conscious construction of paradoxes whose necessary violation
of the laws of logic are intended to shock, even shatter, the standard epistemic
security of “disciples,” thereby allowing them to move to new and higher forms
of insight or knowledge. That is, mystics in certain circumstances know that
they are uttering nonsensical propositions, but in so doing they intend, among
other things, to force the hearers of such propositions to consider who they
are—to locate themselves vis-à-vis normal versus transcendental “reality.”
(1992: 7–8)

531
This seems to be one reason why, far from “high seriousness,”

eleventh-century Hebrew poets and seventeenth-century English poets so
frequently resort to the varieties of metaphysical ingenuities, to what the
neoclassic critics called “false wit,” in voicing their deeply felt religious zeal.
SUMMARY AND CONCLUSION
This chapter seeks to establish a systematic relationship between certain

types of figurative language and the social-cultural backgrounds in which
they tend to occur. It focuses on three typical metaphysical devices—the
metaphysical conceit, the arbitrary genitive (“Pure Attribution”), and the
riddle—as they occur in three typical Mannerist periods—namely, eleventh-
and twentieth-century Hebrew poetry and seventeenth-century English
poetry. It conceives of stylistic devices as adaptation mechanisms turned
to aesthetic ends. This may explain why Mannerist styles, especially in
their Metaphysical variety, tend to occur in social-historical-cultural peri-
ods marked by strife and social instability, in which the world picture has
been undermined or is dominated simultaneously by more than one scale
of values, with no firm grounds to decide between them.
We have briefly observed three Mannerist periods. The analysis of
Excerpts 12 and 13 may serve as an epitome of the argument propounded
in this chapter. In Excerpt 13 we saw the image of a literal cat on a literal
stove, perceived as a good example of the abstraction “desolation.” There
is no logical or visual conflict between the two sets. As in much Romantic
poetry, the actual, physical scenery described in literal language suggests a
metaphorical meaning as well. In Excerpt 12, by contrast, the “concrete of
the abstract” image has recourse to the same elements, but the abstract
set interrupts the continuous concrete description time and again, gener-
ating a metaphorical contradiction. This combines, in an arbitrary manner,
two different sustained sets but does not necessarily generate new mean-
ings. In fact, Excerpts 12 and 13 map the transition from Romanticism to
Mannerism in Bialik’s poetry. As I noted above, Excerpt 12 was extracted
from a “poem of anger and admonition,” written at the time of a national
crisis of values. It is quite significant for the foregoing argument that Bialik
had recourse to typical Mannerist poetics precisely in this and some other
poems of this genre.
Cultural studies frequently assume that “culture begets culture.” If you
find certain metaphorical conventions in eleventh-century Hebrew love
poetry in Spain, you should look for their source in Andalusian Arabic
love poetry. These poets, in turn, may have derived these conventions
541
from earlier Arabic desert poetry. This is sufficiently remote to exempt

us from asking further questions. Then, Andalusian Arab and Hebrew
poetry influenced Provençal poets, who, in turn, transmitted the conven-
tions to Petrarch in Italy, from whence they spread throughout Western
Renaissance culture. Finally, when such conventions were so hackneyed
that they could no longer evoke interest, Shakespeare (“My mistress’ eyes
are nothing like the sun”), John Donne (“What merchant’s ships have my
sighs drowned? / Who says my tears have overflowed his ground?”), and
other seventeenth-century metaphysical poets gave the conventions an
ironic or parodistic twist.
In this context I would like to raise several questions that are very rarely
asked, namely, whether the same convention could not be invented, inde-
pendently, more than once; how the first inventor “hit” upon the meta-
phor and how its first audience understood a new metaphor; and how
the new metaphor became a convention. The present assumption is that
poetic conventions originate in adaptation devices turned to aesthetic
ends. D’Andrade’s passage quoted throughout this book suggests how the
cognitive-evolutionary view and the cultural-transmission view can be inte-
grated. This would suggest a reasonable answer to all of these questions.

larities. (1981: 182)
As to the metaphysical poets’ ingenuity, Dr. Johnson’s comment in his

essay on Abraham Cowley seems to be well taken: “If Wit be well described
by Pope, as being ‘that which has been often thought, but was never before
so well expressed,’ they [the metaphysical poets] certainly never attained,
or ever sought it; for they endeavored to be singular in their thoughts,
and were careless of their diction” (1951 [1779]: 460); and “the reader, far
from wondering that he missed them, wonders more frequently by what
perverseness of industry they were ever found” (1951 [1779]: 461)—
sometimes even more than once. So, did John Donne need Ibn Gabirol’s or
Yehuda Halevi’s model for his tears conceit, or did he invent it through his
own “perverseness of industry”?
Just as in the evolution of the species the wings of bees, birds, and bats
were “invented” independently on different branches of the evolutionary
tree, it is possible that the same “unlikely” literary devices were invented

51
more than once in different cultural periods and centers, under similar
psychological and cultural pressures. Shlomo Ibn Gabirol and John Donne
could take the metonymic relationship between tears and certain emotions
from their everyday experience, whereas they certainly took the story of
the Flood and God’s covenant never to destroy the Earth again by flood
from the same source. But they may have independently generated their
respective conceits of joining tears with flood via violence by applying to
them the same “metaphysical” or psychological logic. This chapter has
explored the raison d’être for such “unlikely” conceits.
561
571
CHAPTER 7
The Translated Poem as an

Aesthetic Object
How Conventions Constrain One Another in a Poem
T his chapter attacks the issue of conventions from a different vantage

point than the other chapters. The whole book is about how poetic
conventions are shaped and constrained by the natural capacities and
constraints of the human brain and cognitive system. This chapter, by
contrast, explores how poetic conventions constrain each other in gen-
erating an aesthetic object. The translated poem was chosen as perhaps
the only case where the question “What was the poet’s intention?” can
be legitimately answered with reference to internal evidence—by compar-
ing the target text to the source text. I am going to contrast two polar
approaches to translation, Clive Scott’s and mine. The two approaches
handle the individual conventions in opposite ways. Both aim to get an
insight into the aesthetic nature of the poem. In Scott’s view, in order
to obtain that insight, one must get as accurate information as possible
about the individual conventions that will, in turn, help the reader access
the subtleties of the poem in the source language. I, by contrast, conceive
of poetry translation as an art in its own right. The translator seeks an
elegant solution to the problem of integrating the conflicting conventions
into a target-language poem that has aesthetic merit in its own right.
Such an approach entails compromises at the expense of the individual
conventions. Briefly, while the earlier chapters of this book explore how
poetic conventions with their respective aesthetic potentials are shaped
and constrained by the natural capacities and limitations of the human
581
brain, the present chapter focuses on a later stage: how conflicting poetic

conventions constrain one another, so as to generate a whole of consider-
able aesthetic merit. This conception applies to all poetry but is brought
out most conveniently with respect to the art of poetry translation, where
alternative solutions can be compared.
The great Hungarian poet and translator Mihály Babits was asked
which in his opinion was the most beautiful Hungarian poem. He
answered: “Shelley’s ‘Ode to the West Wind’ in Árpád Tóth’s translation.”
My point is that this poem is not only a Hungarian masterpiece in its own
right but also flies in the face of the male chauvinist aphorism: “A transla-
tion is like a woman: if she is beautiful, she is not faithful, and if she is
faithful, she is not beautiful.” It provides a reasonable equivalent to the
English source text. I will elucidate what “equivalence” might mean in this
context below. Even a brief analysis that does justice to this magnificent
translation would require a separate essay and might be less than meaning-
ful to an English reader, but I will refer at some length to another transla-
tion by this translator.
This chapter is, then, about how the various poetic conventions interact
to generate an aesthetic whole; it was born as a counterproposal to Clive
Scott’s article “Free Verse and the Translation of Rhythm” (2011). The con-
currences and differences between our approaches are epitomized by the
following issue. Scott objects to
the assumption that constitutive elements are the same from language to lan-
guage, if adjusted by a touch of equivalence. The iambic pentameter is metrical
in the same way that the alexandrine is metrical (though this is clearly untrue).
Rhyme in English is the same as rhyme in French, even though rhyme in French
recognizes different degrees of rhyme, makes alternating rhyme-gender a prin-
ciple of construction, and rhymes on endings and suffixes, which necessitates
a certain practice of avoidances, all features unknown to English rhyming. To
encourage readers to think of French and English rhymes as equivalent is seri-
ously and irresponsibly to mislead them. (2011: 72–73)
I agree with this objection, but we draw different conclusions. Scott

assumes that nothing is as similar to a source text as the source text itself,
and therefore a translation’s function is to help readers work through the
source text and expose themselves to its meanings and sound patterns. He
proposes that translation, irrespective of what kind of verse it is translat-
ing, should always opt for free verse in the translated text. My counterpro-
posal is that a translated poem can be an aesthetic object in its own right
and aspire to a different kind of equivalence. In fact, while some scholars

591
efface fine distinctions like the ones suggested by Scott’s paragraph above,
those distinctions can serve as a basis for translators’ decisions: they must
choose the nearest option available in the target language to the device in
the source poem. I even tend to agree with Scott’s next assertion, though,
as we shall see, in a less categorical version and with different implica-
tions: “Literary translation makes no sense to me if the reader of the tar-
get text (TT) is ignorant of the source language and source text (ST). My
approach to translation always presupposes that the reader of the TT is
familiar with the ST. This presupposition makes the act of translation a
linguistically dialectical act, and an act of textual comparison” (2011: 67).
In my work I have had recourse, in my own way, to both kinds of solu-
tions. I was brought up in the Hungarian language and literature. When,
at the age of sixteen, I mastered Hebrew, I devoted myself to a realization
in Hebrew of the conception of translation I adopted from Babits, Tóth,
and other great Hungarian translators: that is, a rhymed and metered
poem in the target language that makes as good a poem as possible and
chooses those options from the target language that are closest to the ones
in the source language. The precision of the translation depends on how
fine-grained the sign units of the target system are. If the target system
is sufficiently fine-grained and its nearest alternatives are chosen to rep-
resent a source phenomenon, the translation may evoke a perception that
the two are “equivalent.” I virtually gave up these activities while writing
my D.Phil. dissertation, when I drifted in the direction of Scott’s concep-
tion and beyond, for practical rather than theoretical reasons. In my disser-
tation and many of my ensuing publications I frequently analyzed poems
in languages other than the one I was writing in. I quoted the original
poem, providing a literal translation (and, if necessary, a transliteration).
Then I presented a close reading, pointing out the subtleties in the poem’s
sound stratum, units of meaning stratum, and projected world stratum
(the extralinguistic possible world to which the poem refers) and even tried
to integrate all of them (see the discussion below of the Hungarian and
Hebrew translations of Verlaine’s “Chanson d’Automne” and the Hebrew
translations of the line from Hamlet). I felt that in this way I could con-
vey more of the subtleness of the source text. In my artistic translations
I had the satisfaction of conveying as many subtleties of the original poem
as possible; later the suppressed subtleties had their way, paralyzing my
translation activity.
Readers can derive two different kinds of experience from these two
types of output. They can have a more direct, imaginative experience of an
integrated aesthetic object that is equivalent in some sense to the original;
or, alternatively, they can have a more purely rational understanding of a
T h e T r a n s l at e d P o e m a s a n A e s t h e t i c O b j e c t [ 159 ]
6
01
greater number of the source poem’s subtleties, which may or may not lead,
eventually, to a direct imaginative experience of it. Clive Scott’s solution is
somewhere in the middle.
Which solution is right? This question makes no sense to me. Each one is
the right solution relative to the kind of experience a translation is meant
to evoke. Translators and literary scholars are perfectly free to choose one
conception or another; the same person may have recourse to different con-
ceptions at different times. This is not an argument of the anything-goes
type. The final result will be judged not by whether one kind of solution
or another is selected but by how well the chosen solution instantiates its
respective conception. Each kind of solution does something better than
the other kinds do and fails to do things that the other kinds may do very
well. Good reasons can be given to prefer precisely this kind of solution to
the other ones—with reference to each kind. To use Morris Weitz’s (1962)
term, a “crucial recommendation” determines which package of gains and
losses to prefer. From this point of view, Clive Scott’s article does not pro-
pose the best kind of solution but is a crucial recommendation for one kind
and enumerates excellent reasons to support it. The present chapter makes
a crucial recommendation for a different kind of solution and supports it
with the best reasons I can mobilize. As we shall see, Douglas Hofstadter
states, no less categorically, a position that collides head-on with Scott’s.
THE CUCKOO, THE FLUTE, AND THE PIGLET: THE ELEGANT

SOLUTION TO A PROBLEM
There is a parable by Izmailov about the cuckoo who tells her neighbors
in the provinces about the wonderful song of the nightingale she heard
in a faraway country. She learned this song and is willing to reproduce it
for the benefit of her neighbors. They are all eager to hear that marvelous
song, so the cuckoo starts singing: “kukuk, kukuk, kukuk.” The moral of
the parable is that this is what happens to bad translators of poetry. My
thesis is that Izmailov does an injustice to the cuckoo (not to some transla-
tors). When you translate from one semiotic system to another, you are
constrained by the options of the target system. The cuckoo had no choice
but to use cuckoo language for the translation. The question is whether she
utilized those options of cuckoo language that were nearest to the nightin-
gale’s song. After all, Izmailov himself committed exactly the same kind of
inadequacy he attributes to the cuckoo. The bird emits neither the speech
sound [k]‌nor [u]; it uses no speech sounds at all. But every poet in human
language is constrained by the phoneme system of their language; they can

1
6
translate the cuckoo’s song only to those speech sounds. Their translation
will be judged adequate if they choose those speech sounds that are most
similar in their effect to the cuckoo’s call.
How do systems of music sounds and verbal signs take on perceptual
qualities endemic to other systems, such as human emotions or ani-
mal calls? At this stage of my argument I only want to point out that the
resources available in the target systems impose severe strictures on the
process. Usually only very few features or configurations thereof are avail-
able in the target systems that can be shared with the source phenomena.
So, the best one can do is to choose the nearest options available in the
target system. Minute differences may suffice to transform the perceived
character of a complex whole. As Krueger observed, the overall perceived
qualities of “total complexes” are determined by minute differences: “It
has been observed over and over that the smallest changes in experi-
ence are felt emotionally long before the change can be exactly described”
(1968: 100–101).
In onomatopoeia, the phonological system of a language cannot repro-
duce the actual sounds of, for instance, the cuckoo’s call: neither the
minor-third interval, nor the sound quality, nor the abrupt onset. The bird
says neither [k]‌nor [u]. The only thing one can do is to choose the speech
sounds with the nearest formant structure (see Figure 7.1). A symphony
Figure 7.1 Wave plot, the first and second formants of the cardinal vowels i-a-u, and the
European cuckoo’s call. (Formants are concentrations of overtones that determine vowels
and sound color.) Note that the formants of the bird’s call are most similar to, but not iden-
tical with, the vowel [u]‌. (Produced on SoundScope.)
621
Figure 7.2 Wave plot and pitch abstract of the European cuckoo’s call and the cardinal vow-
els read by a professional reader.
Figure 7.3 Sound waves and pitch extract of the imitation of the cuckoo’s call in Leopold
Mozart’s “Toy Symphony.” (Produced on Praat.)
631
orchestra, by contrast, can reproduce the minor-third interval but not the
formant structure of the call.
The nearest option available to codify the abrupt onset of the call in
human speech is the abrupt consonant [k]‌—all the other features of [k] are
irrelevant. This use of voiceless stops to indicate abrupt onset appears to
have some intercultural validity. The Chinese word for “cuckoo” is pu-kû. In
Japanese we have the semantically based hatodokei = “dove” + “clock” but
also poppoo, kokyu, and kakkou. In orchestral versions, the abrupt onset is
indicated more directly (see Figure 7.3). Thus, the voiceless plosive [k] is a
bundle of perceptual features, a subset of which is frequently exploited by
the context to suggest some abrupt metallic noises such as “ticktack” or
“click”; but in the case of “cuckoo” only the perceptual feature [+ abrupt] is
utilized. Thus, the same elements or configurations in a target system can
serve as the “nearest option” for a wide range of source phenomena.
The philosopher John Dewey (1980) and others following him conceive
of an aesthetic object as an elegant solution to a problem. Such an elegant
solution is relevant to a work of art only if both the problem and its solu-
tion can be discerned in it at the same time. Flute players are frequently
praised as follows: “The flute sounds exactly like a soprano singer.” If the
soprano’s voice is so much more beautiful than the flute’s sound, why not
give the part to a singer? But no less frequently we hear the opposite praise
as well: “The voice of this singer sounds like a flute.” The praise does not
indicate preference for the soprano’s voice or the flute’s sound but, rather,
refers to the artistic achievement that one produces a certain sound qual-
ity using a basically different sound quality and that this can be perceived
even if one cannot see the source of the sound (on a record, for instance).
In such a case, both sound qualities are perceptible. Then one may say that
the problem of producing the sound quality of a flute using a human voice
has been elegantly solved.
There is an old Greek parable about a swineherd who entertained his
audience by imitating a piglet’s shriek. Another swineherd, who envied his
colleague’s success, hid a real piglet under his coat and whenever asked to
imitate a piglet’s shriek would pull its tail. The trick failed because the real
piglet was received with scorn. This parable throws an unfavorable light on
the audience’s taste, since it could not differentiate an imitation from the
real thing, giving, by the same token, a good lesson to the envious swine-
herd. The great Hungarian poet János Arany, in his “Vojtina’s Ars Poetica,”
gave a different interpretation of this parable that is in line with certain
nineteenth-century aesthetic views: “Not the truth, but its heavenly image”
is the essence of art. The imitator shrieked like the piglet shrieks in general,
whereas the real piglet may have shrieked as it had never shrieked before.
641
I will interpret this parable in the light of my foregoing argument. The real
piglet will not please the audience more than the imitator even with its most
habitual shrieking, because it provides no solution to the problem. The imita-
tor’s voice, by contrast, remains basically human but weaves characteristics of
the piglet’s voice quality into it, thus creating the illusion of a piglet’s shriek.
With these two parables in mind, let us return to the question of how
a poetic translation can offer some equivalent in the target text of the
sounds and sound patterns of the source text. Speech sounds are per-
ceived as unitary events. If you replace a French word with its Hungarian
equivalent, you are bound to get different sequences of speech sounds: “To
translate alliteration by alliteration, or assonance by assonance usually
entails two second-bests: . . . one does not alliterate the same sounds”
(Scott 2011: 73). This is one way to view the issue. Structuralist phonology
since Jakobson, Fant, and Halle (1952), by contrast, views speech sounds
as bundles of distinctive features on which a closed system of binary oppo-
sitions is constructed. These distinctive features are systematically related
to certain nonlinguistic perceptual qualities. Thus, for instance, the front
vowel [i]‌is perceived as higher and brighter than its back vowel counter-
part [u] pronounced at the same pitch. Similarly, front vowels, in general,
are perceived as higher and brighter than their back vowel counterparts,
as, for instance, [e] and [o] as well. This approach is more fine-grained: it
gives the translator the opportunity to choose distinctive features and
nonlinguistic perceptual qualities in the target language that, if not the
same speech sounds as in the source language, are nearer to, or more
remote from, the ones in the source language. Let me give an example.
The French word for “violin” is violon; the Hungarian word is hegedű. The
Hungarian word is conspicuously unlike its French counterpart. The for-
mer approach can only acknowledge this conspicuous unlikeness. The lat-
ter approach can assign a structural description to it. Thus, for instance,
in the French word, a voiced, continuous, aperiodic and a voiced continu-
ous, periodic consonant as well as two back vowels (oral and nasal, respec-
tively) are present.1 In the Hungarian word, two abrupt, voiced stops [g,
d], the front vowel [ɛ] (twice), and a middle vowel [ű] are dominant. The
oral back vowel [o] is perceived as relatively dark, and the nasal back vowel
[õ] is perceived as even darker, whereas the front vowel [ɛ] is relatively
bright, while [ű] has a particular sheen. Thus, we can say that they are not
1. In periodic speech sounds the same wave shape is repeated. Voicing is periodic.
The consonants [l, m, n, r, w, j] and all the vowels are periodic. The other consonants
are aperiodic. In voiced stops and fricatives, such as [b, v], the [p, f] ingredients are
aperiodic, whereas the voicing ingredient is periodic. Periodic speech sounds are near
to musical tones; aperiodic speech sounds are near to noises.

651
merely as different as any other two words but are contrasted on a great
number of distinctive and perceptual features.
Below I compare Verlaine’s “Chanson d’Automne” with two Hungarian
translations and a Hebrew one. The Hungarian translator Lőrinc Szabó
uses the straightforward translation hegedű, generating a conspicuously
inappropriate sound effect. Árpád Tóth, by contrast, has recourse to a
metonymy of the violin, húr (string), which contains a dark long back vowel
and a voiced continuous periodic consonant. Owing to the vowel harmony
in Hungarian, the possessive suffixes hegedűje and húrja also contrast a
front vowel with a back vowel. Thus, we can say that Tóth chose a nearer
available alternative in Hungarian than Szabó, even though, from a seman-
tic point of view, Szabó used a straightforward translation, whereas Tóth
had recourse to a conspicuous metonymy. In the word húr no nasal vowel
is available; even the best translator cannot remedy such a state of affairs.
But, as we shall see, Tóth amply compensates for this in other words.
Problem solving implies “constraints” or “stringencies.” In poetry, the
grammar and vocabulary of the language as well as the various kinds of
poetic conventions are the most notable ones. These stringencies are fre-
quently incompatible. Syntax may demand one word order; meter, a differ-
ent order of the same words; and rhetorical emphasis, yet another, while
manipulating a certain word into the rhyme may demand still another
word order. The poet must find an elegant way to satisfy the demands of
one solution without infringing on the demands of another.
Thus, for instance, a poetic style demanding more unpredictable figura-
tive language than usual may facilitate the manipulation of a word required
by rhyme for the line ending, without violating word order, rhetorical
emphasis may justify certain deviations from word order required by syn-
tax for the sake of meter or rhyme, and so forth. Poets can be quite inven-
tive and unpredictable in finding such elegant solutions. This holds true
for all poetry. As to translations, the translator of poetry must meet one
more all-important stringency: the translated poem must be as similar as
possible to the original poem in reproducing all these stringencies. Another
imperative is, of course, that the final result must make a good poem in the
target language. This is the sense in which I embrace Scott’s statement that
“literary translation makes no sense to me if the reader of the target text
. . . is ignorant of the source language and source text” (2011: 67), which is,
of course, not the meaning he intended.
One can, of course, compare a translated poem with its original and
point out all the figures of speech and sound patterns and other poetic
devices of the source text that were lost in the target text. Likewise, one
can point out all the poetic devices in the target text that do not occur in
61
the source text. No translation in the world can survive criticism based on
such distinctions. But if the objective is to discover whether the translator
chose the closest options in the target language to the poetic devices of the
source text and whether he or she found a way to bring together as many of
them as possible without one violating the other, one may find that some
translations are more successful than others. Then, indeed, some target
texts may be reasonably akin to the source text in their sound patterns,
figurative language, syntactic structures, and idea content; some of them
may even be masterpieces in the target language. The translation will fail to
be identical with the source text but may present a reasonable equivalent
in the reader’s mind. The reader may derive aesthetic pleasure not from
the point-by-point resemblance between the source text and the target
text but from experiencing the elegant solution of a problem posed by a
great number of stringencies. In this context, equivalence with the source
text (in the sense discussed here) is just another stringency, and complying
with it may increase the reader’s aesthetic pleasure.
TEST ISSUE 1: TRANSLATING VERLAINE’S EUPHONY
I propose to illustrate the foregoing conception with two examples taken

from poetry translation, one concerning periodic sound patterns and the
other involving metrics. I have discussed the former issue in Tsur 1992a
and the latter in Tsur 1977b (see also chapter 9). The first stanza of Paul
Verlaine’s “Chanson d’Automne” is notorious for its exceptional euphony.
We will examine the first stanza of this poem, with two Hungarian transla-
tions and a Hebrew one:
Les sanglots longs
Des violons
De l’Automne,
Blessent mon coeur
D’une langueur
Monotonne. (Paul Verlaine)2
Ősz húrja zsong,
Jajong, búsong
A tájon,
2. Listen to three readings of this stanza, by two native speakers of French and one
probably nonnative. https://www.tau.ac.il/~tsurxx/Soundfiles_counterproposal/
Soundfiles_counterproposal.html

671
S ont monoton
Bút konokon
És fájón. (trans. Árpád Tóth)3
Zokog, zokog
Az ősz konok
Hegedűje,
Zordúl szivem
S fordúl szivem,
Keserűre. (trans. Lőrinc Szabó)4
‫ִּבנְ ִהי ַמ ְמרֹור‬
‫הֹומה ּכִ ּנֹור‬
ֶ
, ַ‫ֵט ֵבת ָּפרּוע‬
‫הּלֵ ב‬ ‫ל‬
ַ ‫וְ ֶא‬
‫ּכְ ֵאב‬ ‫חֹודר‬ ֵ
‫וְ גַ עֲ גּוע‬
Binhi mamror
Homɛ kinor
Tevet paruʕa,
Vǝʔɛl halev
Ḥoder kǝʔev
Vǝgaʕaguʕa. (trans. Zǝʔev Jabotinsky)5
One of the prime sources for this poem’s striking musicality is derived
from its nasal vowels and the sound sequence -eur. This sound sequence
and the nasal vowels have two characteristics in common: from an acous-
tic point of view both are continuous and periodic, and from the point of
view of the infant’s phonological development, both are late acquisitions.
Following Jakobson (1968), I have argued elsewhere (Tsur 1992a) that the
infant’s latest phonological acquisitions have the greatest emotional and
aesthetic potential in adult language, for better or for worse. Among the
late acquisitions, such abrupt consonants as the affricates [pf] and [ts]
are “ugly” or express unpleasant emotions; the speech sounds that are
continuous and periodic from the acoustic point of view are musical and
“beautiful” and express pleasant emotions. The “beautiful” sounds abound
3. Listen to a reading of Tóth’s Hungarian translation. https://www.tau.ac.il/~tsurxx/

Soundfiles_counterproposal/Soundfiles_counterproposal.html
4. Listen to a reading of Szabó’s Hungarian translation. https://www.tau.ac.il/~tsurxx/
5. Listen to Jabotinsky’s Hebrew translation. https://www.tau.ac.il/~tsurxx/
6
81
in French Impressionistic-Symbolistic poetry relative to French Classicist

poetry, for instance, and occur very frequently in rhyme words. This is one
of the key sources of its musicality and beauty.
In the phonological system of Hebrew there are no nasal vowels and no
vowel eu. In his Hebrew translation, therefore, Jabotinsky had to find pho-
nological equivalents that had a similar emotional character. Among the
relatively late acquisitions in Hebrew is the consonant [r]‌, which is continu-
ous and periodic. Jabotinsky attempted to compensate for the absence of
nasal vowels and eu by using as many sonorants and back vowels as pos-
sible and by manipulating them into as prominent places as possible. He
founded his first rhyme on the continuous and periodic [r], which occurs in
the first two short lines three(!) times, and on the back vowel [o]. He also
had recourse in these two lines to other periodic consonants: [m] three
times, [n] twice. This is a striking instance of an attempt to generate a simi-
lar atmosphere in the Hebrew translation by choosing the periodic and late
consonants available in the language.
The perceived quality of [r]‌requires additional elucidation. This speech
sound is double-edged. On the one hand, it is periodic; on the other hand, it
is multiply interrupted. Being both periodic and interrupted, it frequently
serves as an imitation of noises and creaking sounds; but in the context
of periodic consonants, as, for example, in “Lili Marleen,” the reader may
attend away from the interrupted to the periodic aspect, and the [r] inte-
grates well into the pleasant, euphonic, sonorant context. This is the case
in the present phonetic and thematic context as well.
It is impossible within the Hebrew phonological system to come nearer
to the resources of musicality in French. Hungarian translators of this poem
can consider themselves luckier. First of all, [ø:] is available to them; what
is more, they can hardly avoid it, because it occurs in the Hungarian word
for “autumn”: ősz. Nasalized vowels are available in Hungarian, although
less readily than in French, in a few root words and most notably in the fre-
quentative suffix -ong. In fact, it occurs three times in the first two lines of
Tóth’s translation (zsong, jajong, búsong [= reverberates, whines, bewails]).
Consequently, when a Hungarian poet such as Szabó has no nasal vowels in
the whole stanza, while he emphatically repeats the velar plosives in zokog,
the translation is perceived as exceptionally unmusical. Obviously, he did
not have recourse to the nearest alternatives in the Hungarian phonologi-
cal system. In this respect, the translation is not to be judged by the cri-
terion of whether the phonemes that determine the poem’s character in
the source language do or do not exist in the target language but, rather,
whether the translator did or did not exploit the possibilities inherent in

691
the closed phonological system of the target language and, in the present
instance, whether he chose as many alternatives as possible involving the
distinctive features [+ continuous + periodic + nasal] that contribute to
the poem’s effect in the source language.
Tóth’s preference for nasal vowels and consonants in this translation
is obvious. Recall that the Hungarian word for “its violin” (hegedűje) was
retained in Szabó’s translation, whereas in Tóth’s it is replaced by a meton-
ymy, húrja (its string), introducing a dark back vowel ([u:]) and a sonorant
([r]‌), as well as eliminating, by the same token, such plosives as the /g/ and
/d/ of the proper term. The ű is a rather late developmental acquisition but,
according to Jakobson (1968), has a certain sheen that would be foreign to
the atmosphere of this poem (Hungarian ő and ű are brighter than their
French equivalents).
Encouraged by the Hungarian translator’s feast of nasality, I attempted
many years ago to render Verlaine’s poem in Hebrew; lines 4–6 run as
follows:
‫יָ גֹון‬ ‫יַ נְ עִ ים‬

‫עֲ מּום ַהּגֹון‬
. ַ‫וְ גַ עֲ גּוע‬
Yanʕim yagon
ʕAmum hagon
Vǝgaʕaguʕa.
[Sounds monotonous woe and yearning.]
These words make ample use of dark back vowels and nasal consonants, as
well as of the voiced velar stop [g]‌in close vicinity. Nonetheless, the poem
refused to assume a musical quality comparable to the French original
or Tóth’s Hungarian translation. This was most clear-cut precisely where
I expected the greatest similarity, in the rhyme words yagon–hagon. At
that time I could only describe the difference in an intuitive fashion: the
gon sequence in the Hebrew rhyme sounded somehow too decisive, too
conclusive, too assertive, too solid, as compared with the correspond-
ing sequences in French and Hungarian. Phonetically and phonologically
speaking, in French we are dealing with nasal vowels proper, whereas in
Hungarian we are confronted with an allophone of the oral vowel, strongly
nasalized by coarticulation with the subsequent nasal consonant, which,
in turn, is attenuated by the ensuing voiced velar stop. In Hebrew, by con-
trast, though both [g] and [n] are present, there is no coarticulation, no
nasalizing effect.
701
Haruko Kawasaki’s work in experimental phonology throws an inter-

esting light on the aesthetic repercussions of this difference between
Hebrew and Hungarian: “It has been well documented that . . . nondistinc-
tive [i.e., nonphonemic] nasalization . . . has a physiological cause: lower-
ing of the velum adjacent to nasal consonants” (1986: 86). Historically,
many instances of nasal vowels are derived from such allophonic nasal-
ization, when the nasal consonant is dropped, as, for instance, French /
sã/ (cent = hundred) from Latin centum. This, Kawasaki claims, is a phono-
logical universal, and she cites supporting evidence from a large variety of
languages and a wide range of types of phonological constraints in them.
She (1986: 94) provides experimental evidence showing that the degree
of perceived nasality of a vowel is enhanced by the attenuation of adja-
cent nasal consonants or, conversely, is reduced by the presence of adjacent
nasal consonants.
Compare the name of the great German philosopher Kant with the
British pronunciation of the contraction can’t. They can be treated as prac-
tically a minimal pair. One of the most obtrusive differences between them
concerns precisely the issue discussed here: In Kant the nasal consonant
[n]‌has its full solid body; accordingly, no or little nasalization is perceived
in the preceding vowel. In can’t, by contrast, the [n] is strongly attenuated
(by coarticulation with the [t]); accordingly, a strong nasal quality is per-
ceived in the preceding vowel. Likewise, in Hungarian zsong (or, for that
matter, in English song) the voiced velar stop [g] drastically attenuates the
adjacent [n] (both are produced by manipulation of the velum); by the same
token, the nasal perturbation in the preceding oral vowel becomes palpable.
In Hebrew yagon and hagon, by contrast, the preceding [g] causes no such
attenuation in the nasal consonant; hence, no nasal perturbation becomes
perceptible in the preceding vowel. The nasal perturbation in Hungarian
zsong (or in English song) is perceived as diffuse but rich precategorial sen-
sory information that increases chaotic overdifferentiation, whereas the
weak, residual [n] is perceived as a diffuse, vague, and evasive rather than
a compact phonetic category. In this way, the perceptual contrast between
the vowel and the consonant is reduced as well. All this reinforces the emo-
tional atmosphere typically cherished by Symbolist poetry. In Hebrew no
such attenuation of nasal consonants takes place; so, the Hebrew translator
of Verlaine must be content with fully realized, relatively compact nasals.
I used phonetic equivalence to exemplify the problem of translation
from one semiotic code to another, but a similar principle seems to gov-
ern metric, semantic, and syntactic equivalence as well. It should be noted,
however, that the foregoing arguments regarding Jabotinsky’s selection
of speech sounds do not enable us to predict whether the equivalents will

7 1
generate an effect similar to the source language in the target language.

Rather, if readers hear a certain impression of the translation’s musicality,
they may use such arguments to account for it after the event. The musical-
ity of speech sounds can only be heard not inferred.
TEST ISSUE 2: TRANSLATING ALEXANDRINES

INTO SYLLABOTONIC VERSE
In what follows, I give a further example, this time from the domain of met-
rical equivalence, when different kinds of semiotic systems are involved. In
chapter 9, I use this same example to illustrate the perceptual difference
between the iambic and the ternary meters; here I use it to demonstrate
the issue of equivalences between versification structures in syllabic and
syllabotonic verses.
What is the equivalent of a poem’s meter when we translate it into a
language in which a different metrical system is prevalent? More specifi-
cally, I explore the problem of translating French alexandrines (based on
the syllabic metrical system) into languages in which the syllabotonic is
the dominant metrical system. I look at the problem of meter in translating
Baudelaire’s “Correspondances” into such languages as English, Hebrew,
and Hungarian. In an epilogue to my Hebrew volume of poetry transla-
tions I put forward a cognitive theory of translation similar to the pres-
ent one. Among other things, I compare the metrical organizations of six
Hebrew translations of this poem. Two have no recognizable meter, one
is in a mixture of ternary meters, and three are in the iambic hexameter.
Which one of these solutions best suits the spirit of the French meter? And
what explanation can be given for this choice?
The metrical system dominant in English, Hebrew, Hungarian, and
some other modern literatures is the syllabotonic system: that is, the sys-
tem that determines, ideally, the number of syllables as well as the number
of stresses and their placement in the verse line. In iambic pentameter, for
instance, there are ten syllables in a line, and every even-numbered syl-
lable ought to be stressed. In the preceding sentences I used the phrases
“determines,” “ideally,” and “ought to be stressed” because in fact there is
such a discrepancy between the ideal and the real stresses that the most
fruitful way to speak of poetic rhythm is to define the metrical pattern
and the stress pattern separately and identify where the two converge
and where they diverge. In French poetry (and in certain other Romance
languages), by contrast, the syllabic system is the dominant metrical sys-
tem: that is, the system in which the syllables are counted, whereas the
721
number of stresses and their placement are ignored (Scott would qual-
ify this statement: they are not ignored, only unpredictable). In French
poetry, usually two more organizing prosodic principles are added: the
caesura at a fixed point (the middle) of the verse line and a predictable
arrangement of “masculine” and “feminine” rhymes, according to certain
principles of symmetry. When translating a poem from French to English,
Hungarian, or Hebrew, the question necessarily arises of what syllabo-
tonic meter will be equivalent to the French poem’s meter from the point
of view of perceived effect.
In English, Hebrew, and Hungarian, French alexandrines are usually
translated into iambic hexameter. My argument will rest on a Hebrew
translation that deviates from the prevalent practice. So, unfortunately,
I will have to keep my discussion at a highly general level.
In all my work in metrics I distinguish metrical pattern, stress pattern,
and pattern of performance. When stress pattern and metrical pattern
converge, they yield strong prosodic gestalts with a psychological atmo-
sphere of certainty and patent purpose. When they diverge, the verse is in
danger of falling into chaos. The coherence of the metrical foot depends,
to a considerable extent, on the downbeat. In binary meters (the iambic
and trochaic) only one upbeat “leans on” each downbeat for support; in
ternary meters two upbeats lean on one downbeat. Consequently, binary
meters are more stable than ternary meters and more resistant to disin-
tegration. In ternary meters, to prevent disintegration, performers are
inclined to subordinate the prose rhythm to the regular metrical beat. The
iambic foot, with its stronger gestalt, seems to tolerate greater deviance
and complexity. In extreme cases of deviation, to prevent chaos the per-
former needs to accommodate the divergent patterns in a strong gestalt
of additional grouping. This underlying strong gestalt, whether in the pat-
tern of meter or performance, makes rhythmicality possible when diver-
gent elements are present. As can be inferred from Woodrow’s tick-tack
experiments (see chapter 9), end-accented meters (such as the iambic and
the anapest) allow for greater flexibility in manipulating the time fac-
tor than beginning-accented meters (such as the trochaic and dactylic).
Hence the relatively greater rigidity of trochaic verse, which is manifest
in its “compelling” nature, observed by so many critics from Aristotle
to Zhirmunsky, Chatman, and the generative metrists. Consequently,
the iambic is more tolerant of deviations than the trochaic and the
ternary meters.
At this point I will refer once again to Scott’s objection to the claim that
“the iambic pentameter is metrical in the same way that the alexandrine is
metrical (though this is clearly untrue)” (2011: 72). My attitude toward this

731
claim depends very much on what is meant by “is metrical in the same way
that.” If it implies that the iambic pentameter is the most frequent meter
in English, Hungarian, and Hebrew syllabotonic systems, just as the alex-
andrine is the dominant meter in the French syllabic system, I wholeheart-
edly agree. But there are other possibilities as well. Whereas in English
and Hungarian syllabotonic verse the iambic is, indeed, the most frequent
meter, in Hebrew it is not; for reasons relating to average word length in
Hebrew, the anapest and amphibrach are more frequent. Yet, as we shall
see, in Hebrew as well, the iambic is felt to be more appropriate for the
translation of the alexandrine than some ternary meter.
The alexandrine, the most widespread meter in French, is syllabic. It
consists of twelve syllables (thirteen in “feminine”-ended lines), with a
compulsory caesura after the sixth. It makes no systematic use of contrast
between prominent and nonprominent syllables. Alexandrines by Racine,
Baudelaire, and other French poets are usually translated into the iambic in
languages as dissimilar as Hebrew and Hungarian. In my corpus of Hebrew
translations of Baudelaire’s “Correspondances,” however, one translation is
in a ternary meter, a mixture of the amphibrach and the anapest. Though
it is remarkably polished, the rhythm sounds strikingly unsuitable. It has
a vigorous, straightforward quality, quite unlike the vague impression the
poem makes in French.
One’s first response to such an encounter is surprise. Then it comes to
mind that surely it is only a matter of habit that the ear feels the iamb
rather than the amphibrach as the syllabotonic equivalent of the alexan-
drine. From a purely arithmetic point of view, both possibilities are equally
plausible. After all, 4 × 3 = 12 just as 6 × 2 = 12. The caesura after the sixth
syllable will coincide with a foot boundary, whether binary or ternary.
Something perceptual, however, resists this arithmetic equivalence. As
I have said, compared with the iambic meter, ternary meters are inflexible
and generate “a psychological atmosphere of certainty, security, and patent
purpose, in which the listener feels a sense of control and power as well as
a sense of specific tendency and definite direction” (Meyer 1956: 160) in
the poem, conflicting with the vague, ambiguous atmosphere, the elusive
apprehension of another reality, characteristic of the French original (cf.
Hofstadter’s translation of Armand Silvestre’s poem below).
At first sight, there appears to be no reason why Baudelaire or Racine
should not be translated into the anapest or the amphibrach. A closer look
at the French poem “Correspondances” even reveals that the “stress pat-
tern” of the first hemistich confirms the ternary anapest. But the second
hemistich confirms the binary iambic. The third one nearly confirms the
iambic (except for the “inverted first foot”), while the fourth confirms
7
41
anapest, and so forth. Consider the first two lines, indicating the position
in which the “stress” occurs:

3
6
4
6
La Nature est un temple // où de vivants piliers

1 4 6 3 6
Laissent parfois sortir // de confuses paroles
In the light of the discussion above, this irregularity becomes significant.

The alexandrine is similar to the iambic in one important respect. It tol-
erates greater irregularity of the stress pattern than trochaic or ternary
meters such as the anapest, the dactyl, or the amphibrach would tolerate
(see Fowler 1966). This is the reason why the iambic meter lacks the “psy-
chological atmosphere of certainty, security, and patent purpose” char-
acteristic of the trochaic and ternary meters. Here, of course, one could
ask whether this goal could not be better attained by relinquishing meter
entirely. The answer to this question seems to be that the relinquishing
of meter may detract from the aesthetic nature of a poem. In view of the
“elegant-solution” conception of the aesthetic object, irregularity in spite of
an established meter has a more aesthetic effect than irregularity devoid
of meter. This conclusion is borne out by the two Hebrew translations
that have no recognizable meter. The English reader will have no intu-
itions regarding a Hebrew translation. But I will illustrate the issue below,
in a comparison of three English texts, all versions of the same speech by
Shakespeare’s Gloucester (in Richard III), by three different authors. Meter
is one of the constraints that constitute the problem for which the poet or
the translator must find an elegant solution. It would be quite safe to con-
clude that the iambic hexameter is the syllabotonic system’s nearest option
to the alexandrine, not because they are the most widespread meters in
their respective systems but because they are similar in tolerating irregu-
larity in the linguistic dimension.
THE ESSENTIAL TENSION
Here we find an interesting overlap and difference between Scott’s

approach and mine. Scott describes the rhythmic nature of French alexan-
drines in terms of an unpredictable grouping of syllables in the linguistic
stream and a predictable grouping of syllables in the versification struc-
ture, “suggesting that sequence is not a process of addition but of meta-
morphosis (including morphing into something numerically the same

751
but constitutionally different)” (2011: 80), but expects, for example, the

English translator to accurately reproduce the unpredictable word group
patterns of each French verse line and give up any systematic meter.
I would translate Scott’s position into my terms as follows. In French
there is tension between the predictable number of syllables in the verse
line and the unpredictable number of groups of syllables in the linguis-
tic stream. This tension only occurs when the predictable stream and the
unpredictable stream have the same number of syllables and a coincid-
ing end. The unpredictable grouping of linguistic units suspends the sense
of certainty; when the unpredictable linguistic pattern and the predict-
able metric pattern have a coinciding end and turn out to have the same
number of syllables, certainty is restored and a sense of relief is achieved.
The suspense itself might be pleasurable because the reader is assured
from the beginning that coincidence will be restored on the sixth and the
twelfth syllables of the verse line.
This presentation of Scott’s position can be illustrated by the above
quotation from Baudelaire’s “Correspondances.” This would be in perfect
harmony with the present proposal. Now what Scott proposes is that in
translation one should reproduce the unpredictable stream accurately,
at the expense of the predictable stream, relinquishing the essential ten-
sion. This is the theory. In practice, Scott himself translates the same lines
thus: “Nature is a temple: from time to time, its living pillars sibylline, let
slip bewildering words” (in Culler 2014: 7). There is no way to recover from
this text the essential tension between language and meter. Culler com-
ments on Scott’s rhythm: “Scott’s version is more engaging [than Keith
Waldrop’s], doubtless . . . because of the more clearly articulated rhythm
(Scott says he is trying to create a text with ‘a firm bearing rhythm’), and
he boldly marks the confuses paroles as ‘sibylline,’ unclear yet enigmatic and
potentially prophetic, rather than simply disjointed” (2014). (It seems to
me that Culler describes the semantic rather than the rhythmic structure
of the translation.)
My proposal, by contrast, implies that in the target text, too, the essen-
tial tension should be preserved by observing both the predictable stream
of (syllabotonic) meter and the unpredictable grouping of linguistic units,
but not necessarily the exact unpredictable sequence of stresses. Besides,
in the syllabotonic system another alternative is available “Où l’Indécis au
Précis se joint” [Where the Uncertain joins the Precise], to use Verlaine’s
words. This system not only counts syllables but also observes a regular
sequence of downbeats and upbeats. There may be an irregular stream of
downbeats and upbeats in the linguistic dimension and a regular stream of
downbeats and upbeats in the metrical dimension, which, at certain crucial
761
points, may have coinciding downbeats. In English poetry, for instance,

one crucial difference between Alexander Pope’s metrical style, on the one
hand, and Milton’s and Shelley’s metrical styles, on the other, is that in the
latter the stream of linguistic downbeats is less regular than in the former,
while regular meter is somehow still perceived. Consequently, Pope’s poetry
is perceived as witty; Milton’s and Shelley’s, as emotional. What Scott gains
is the accurate reproduction of the linguistic patterns in one dimension of
poetic rhythm, while he loses the interplay between two dimensions. My
proposal gains the interplay between two dimensions but loses the exact
reproduction of the linguistic patterns; it preserves only a general notion of
unpredictability. In the final resort, one must choose which set of gains and
losses one prefers: the essential tension or the exact sequence of groups in
the unpredictable linguistic stream.
THREE VERSIONS OF GLOUCESTER’S SPEECH
I illustrate this essential tension and the lack of it by comparing two metered
passages with a text where there is no versification pattern (reproduced,
with the necessary changes, from Tsur 2008a: 145–146). The three excerpts
below are versions, by three different authors, of the same speech. One
of the most memorable instances of Gloucester’s villainy (in Shakespeare’s
Richard III) is the scene in which he gets rid of Lord Hastings:
If? thou Protector of this damned Strumpet,

Talk’st thou to me of Ifs: Thou art a Traytor,
Off with his head; now, by Saint Paul I sweare,
I will not dine, untill I see the same. (Richard III III.iv.73–76)
The situation has its particular wit. It is reinforced by the “sharpness” of

Richard’s speech, which, in turn, is a corollary of the rapid shift of diverse
phrases, without violating the integrity of the line. Here we witness not
merely a sequence of irregular groupings of syllables but group boundar-
ies enhanced by shifting levels of speech. Note the isolated tense “If?” at
the beginning of the speech, quoted from Hastings’s discourse, whereas
the rest is Gloucester’s direct speech. The rapid shift of levels of discourse
appears here in the shift from “If” to direct speech and back to the quoted
“Ifs,” as well as in the quick shift from the second person to the third per-
son, varying the direction of his address (“thou Protector,” “Off with his
head,” “now, by Saint Paul,” etc.). In this speech, the conflicting linguis-
tic pattern and versification pattern compete to establish themselves in

71
the reader’s perception, thus enhancing each other. The group boundaries
intrude on the line, but the line strives to retain its integrity.
Shakespeare actually amplifies a technique that he found in Dolman’s
poem in The Mirrour for Magistrates: “Yf, traytor quod he? playest thou
with yfs and ands? / Ile on thy body avowe it with these hands.” Notice
the isolated Yf followed by a short vocative (traytor), followed by a short
parenthesis (“quod he”). The quoted yfs and ands not only diversify but also
intensify the subdivision of the line. They give rise, at the same time, to a
bold “antigrammatical” rhyme, by rhyming a noun with a conjunction (in
the plural!), involving a leap from first-order language to metalanguage and
back. It strengthens the closure of the rhymed couplet, thus heightening
its sense of unity.
It is revealing to compare this couplet of Dolman’s to Sir Thomas More’s
prose account of the same incident: “What quod the protectour thou seru-
est me I wene wt iffs & with andes, I tell the thei have so done, & that
I will make good on thy body traitour” (1963: 48). Verbally, More’s account
does not differ greatly from Dolman’s. It displays the same (or even more)
emphatic segmentation of the syntactic stream. Nevertheless, there seems
to be a perceptual difference between them, in that the segmentation
stands out less clearly in the unmetered version. The present comparison
highlights an illuminating aspect of the issue. The first two excerpts clearly
demonstrate how shifting phrases can enhance the prosodic unit. The third
excerpt foregrounds an unexpected, obverse, aspect: namely, how the pro-
sodic structure imposed on the phrases in Dolman’s poem renders their
shift far more emphatic than in More’s prose account.
PROBLEM WITHIN THE SYLLABOTONIC SYSTEM
There is a notorious problem regarding translation within the syllabotonic

system as well. The average word length is different in various languages.
Consequently, there is a pernicious if largely implicit belief that translators
are free to translate iambic pentameter into iambic hexameter or the other
way around, according to the requirements of the average word length in
the target language, as long as the basic iambic pattern is observed.
So far, I have encountered only one rather militant and explicit state-
ment of this position.6 But practice based on this assumption is virtually
ubiquitous. English pentameter verse is frequently translated into Hebrew
6. It was at a conference at Bar-Ilan University on Hebrew translations of Shakespeare’s

sonnets. A Hebrew translator of the sonnets attacked those translators who insisted
on the pentameter for their translations.
78
1
hexameter, and French alexandrine is very frequently translated into

English decasyllabic verse. Hungarian translators tend not to commit this
sin. Most notably, the great Hebrew poet and translator Abraham Shlonsky
translated Hamlet into iambic hexameter, and his work is usually regarded
as a masterpiece of translation. Shlonsky himself seems to have felt that
something was wrong with this, because later he translated King Lear into
pentameter. Indeed, it is extremely difficult to render all the information
in an English pentameter line in a Hebrew pentameter. The average word in
Hebrew is two and a half times longer than in English. The reverse happens
when translating French alexandrines into English. The average French
word is much longer than the average English word.
The most outrageous instance of this tendency is Shlonsky’s translation
of Hamlet’s line “It is not, nor it cannot come to good.” This line occurs at
the peak of one of Hamlet’s “hypomanic” outbursts and is uncontrolled,
within controlled prosodic constraints. Syntactically, Hamlet “corrects”
himself in mid-sentence, leaving the first clause incomplete. Stylistically,
it has that double negation “nor it cannot.” Prosodically, this is one of the
very few iambic pentameter lines in major English poetry in which you can-
not observe a reasonable caesura after the fourth, fifth, or sixth position;
the major syntactic juncture occurs after the third position (as we shall see
in a moment, the caesura in iambic tetrameter and hexameter is rigidly
fixed at the middle, for good perceptual reasons, whereas pentameter is
much more flexible).
Shlonsky renders this verse line in an elegant hexameter line:
‫סֹופֹו‬ ‫ וְ ל ֹא יִ ַיטב‬,‫ל ֹא טֹוב ַה ַּמ ֲע ֶׂשה‬
lo tov hamaʕasɛ, vǝlo ji:tav sofo

[This deed is wrong, and it will not come to good]
Syntactically, this is a coordinate sentence with both clauses completed.

Prosodically, the major syntactic juncture occurs after the sixth position
and produces a perfectly symmetrical parallelism, yielding a psychological
atmosphere of a rational analysis of the situation, in the neoclassic vein.
In an undergraduate seminar paper many years ago I criticized Shlonsky’s
translation and offered an alternative translation of this line:
‫ ַאף ל ֹא יִ ּגָ ֵמר ְּבטֹוב‬,‫ּכִ י ֵאין זֶ ה‬

ki eyn zɛ, ʔaf lo jiggamer bǝtov
[for it is not, nor will it come to good]

1
79
Here the first clause is interrupted, leaving a meaningless stump (“it is

not”), and the major syntactic juncture occurs after the third position.
The belief that translators are free to translate iambic pentameter into
iambic hexameter and the other way around ignores the really important
thing, the gestalt qualities generated by the various meters. Iambic hexam-
eter and iambic tetrameter divide the verse line into two segments of equal
length and equal structure. Each segment has the same number of sylla-
bles, begins with a weak position, and ends with a strong position. Iambic
pentameter, by contrast, can be divided into segments of equal length
and dissimilar structure or similar structure and unequal length. It can be
divided into 5 + 5 positions, but then the first segment begins and ends
with a weak position, whereas the second segment begins and ends with a
strong position. To obtain segments that both begin with a weak position
and end with a strong position, the verse line must be divided into 4 + 6
or 6 + 4 positions. Owing to these respective structures, hexameter and
tetrameter enforce a caesura exactly in the middle, are symmetrical and
stable, and frequently have a rational or simplifying effect. The pentameter
may have a caesura after the fourth, fifth, or sixth position and is necessar-
ily asymmetrical and more flexible. This difference can crucially affect the
perceived aesthetic quality of a poem. Thus, for instance, in Verlaine’s view,
to render a verse line musical, first of all one must liberate it from the tyr-
anny of symmetry: “De la musique avant toute chose, / Et pour cela préfère
l’Impair” [Music before everything / And for this prefer odd numbers]. In
iambic pentameter we have an exceptionally sophisticated version of asym-
metry within an even-numbered structure.
People who entertain a belief in the equivalence of pentameter and hex-
ameter sometimes do not even realize that by shifting from the one to the
other the rules of the game have changed. While the placement of the cae-
sura is flexible in pentameter, in hexameter it is rigidly fixed after the sixth
position, for good perceptual reasons. First, as George Miller (1956) showed,
the span of short-term memory is limited to the magical number 7 ± 2; lon-
ger verse lines have a compulsory break. Second, there is a gestalt rule that
the organization of a perceptual object cannot be equally good at all levels.
Greater simplicity of the parts makes them stand out at the expense of the
whole; the parts must be weakened to some extent to make them dependent
on, and integrated within, the whole. The symmetry of the hexameter line
makes the segments divided by the caesura stand out, whereas the asymme-
try of the pentameter line renders them integrated within the whole.
Let me illustrate the dynamics involved by quoting from one of my pre-
vious books (Tsur 2012a: 116–117) the following exercise: “Invoke thy aid
8
01
to my advent’rous song” (Milton, Paradise Lost, I, 13). In this pentameter

line, the caesura occurs after aid in position IV. Suppose, however, that we
add two more syllables to the verse line, turning it into an iambic hexam-
eter, thus: “Invoke thy aid to my advent’rous song of praise.” If one contin-
ues to observe a caesura after aid, the line is liable to fall apart. Here the
caesura, in harmony with the perceptual needs of the iambic hexameter, is
automatically shifted to after my in position VI, even though this happens
in mid-phrase.
HOFSTADTER ON TRANSLATION
At this point I turn to the essay “Different Music, Same Condition: Hofstadter

and Lyotard,” by Peter Dayan (2012). Hofstadter, Dayan says, “has shown
why, for him, only translations that take account of the formal patterns of
the original poem can be received as themselves poetic; why regular verse
can only be translated as regular verse” (2012: 11), leading to the conclusion
that “the essence of the act of writing poetry is the indissoluble fusion of
a medium with a message, the unsunderable wedding of form to content
as equal partners” (2012: 11, quoting Hofstadter 1997: 524). Hofstadter
adopts this position in an extreme version:
He finds poetry that eschews the constraints of regular verse, like atonal music,
aesthetically objectionable because in it, he cannot see content wedded to form.
Instead, he sees a purely intellectual art in which “form is seen as the dog’s tail,
content as its body” (527), and the artist refuses to let the formal tail wag the
dog of content. (Dayan 2012: 12, quoting Hofstadter 1997)
All this, as I said, is in head-on collision with Clive Scott’s position.

It will be noted that this argument contains elements of mine, but
there are two substantial differences. First, while Hofstadter speaks of
“the unsunderable wedding of form to content” and “the constraints of
regular verse,” I refer to mutually constraining stringencies. This is not ver-
bal hair-splitting but a distinction of substance. When I claim that Scott
“loses the interplay between two dimensions,” I do not mean between
form and content but, rather, between two sound patterns—that is, the
sound patterns of language and meter, both within what for Hofstadter
would count as form. In other words, though both of us defend the rights
of regular verse, the “mutually-constraining-stringencies” conception is
more fine-grained than the “form-and-content” conception. Second, while
Hofstadter adopts an uncompromising all-or-nothing stance, I speak of

1
8
“crucial recommendations,” each of which is justifiable in view of the kind

of response it is meant to evoke.
Now let us see how Hofstadter’s practice bears out his generalization.
In the conclusion chapter of Hofstadter’s book, Dayan writes, “at last,
Hofstadter answers a question which, for anyone interested in the condi-
tion of music in verse, will have been hanging in the air for five hundred
pages: what, exactly, is the music of language, in praise of which, according
to the subtitle, the book has been written?” (2012: 11). So I consulted this
chapter. Here, among other things, Hofstadter presents his own transla-
tions of two French poems by Armand Silvestre. I will refer only to the first
one. Consider the first stanza of “Aurore” (Dawn), in French and in English:
Des jardins de la nuit s’envole les étoiles,

Abeilles d’or qu’attire un invisible miel,
Et l’aube, au loin tendant la candeur de ses toiles,
Trame de fils d’argent le manteau bleu du ciel.
From the gardens of night the bright stars are in flight,
Golden bees subtly lured by a nectar unseen,
And the dawn, far off spreading its canvases white,
Shoots its silvery threads o’er the sky’s azure screen. (Hofstadter 1997)
On the whole he does a beautiful job: the English poems are good poems in
their own right and fairly convey the meaning. The versification is highly
polished. In this stanza, the meaning comes through, and so does the
rather moderate figurative language. There is one interesting exception. In
the English version, the dawn spreads “its canvases white,” whereas in the
French original, it spreads “the whiteness of the canvases.” There is a huge
difference between the two. In “canvases white” a physical action predicate
is applied to a physical object, resulting in physical movement in space. In
“the whiteness of the canvases” the physical action is applied to an abstrac-
tion, a thing-free quality, generating an intense, thing-free presence. To
be sure, we are speaking of the dawn’s canvases, which are immaterial and
have no stable visual shape, but, still, “the whiteness [or “candor”] of the
canvases” is an abstract quality of that immaterial entity. In other words, in
a concrete noun many abstract properties have “grown together”; the geni-
tive construction loosens the relationship between the abstract quality and
the object whose quality it is; the application of a predicate to the abstract
quality as an object suggests that it has some independent existence.
This change is not forced on the translator by the constraints of ver-
sification; he could solve the problem with a simple apostrophe, by writ-
ing “spreading its canvases’ white.” “Canvases white” suggests relative
821
stability, whereas “the whiteness of the canvases” as the object of the verb
spread suggests massive presence, diffuseness, elusiveness, and intangibil-
ity. Thus “canvases white” generates a psychological atmosphere of relative
certainty, security, and purpose.7
As to versification, there seems to be a problem here relating to meter.
Hofstadter translates the French alexandrine into English anapests. A closer
look at the French stress pattern reveals that, just as in “Correspondances,”
the first hemistich is clearly anapestic. The second, third, fourth, and fifth
hemistichs, however, are clearly iambic. The sixth hemistich, again, is
clearly anapestic. The seventh hemistich is iambic with an “inverted first
foot,” while the last one contains two consecutive stressed syllables—
“manteau bleu”—rendering the pattern indecisive.
As I suggested earlier, the iambic is more tolerant of deviant stresses,
without falling apart; the anapest more rigidly exerts its will and tends
to suppress linguistic stress that happens to occur in a weak position
(and there are quite a few of them in the translation of this stanza). The
effect of the anapest here is by far less devastating than in Baudelaire’s
“Correspondances” because the content is less elusive; but still, it rein-
forces the sense of control and the atmosphere of relative certainty,
security, and patent purpose generated by manipulating the solid object
“canvas” rather than its thing-free quality into the referring position. To
use Hofstadter’s own words, “the indissoluble fusion of a medium with a
message, the unsunderable wedding of form to content as equal partners,”
has here the wrong effect—the unsunderable wedding of the ternary meter
to the manipulation of a relatively solid object into the referring position
of the phrase. Their indissoluble fusion suggests here a psychological atmo-
sphere of stronger-than-appropriate certainty and security (reinforced by
the internal rhyme at the caesura and the end of the first line, which articu-
lates the verse line into two clear-cut, terse segments). In other words, it is
not enough that form and content are unsunderably wedded; their inter-
play must generate the right kind of atmosphere or perceived quality.
NOTE ON TERMINOLOGY
Finally, I want to emphasize that Scott and I use two of our crucial terms,
rhythm and equivalence, in different ways. From my vantage point, Scott
7. As I have argued elsewhere (and see chapter 6), such constructions of “the
abstract of the concrete” abound in Baroque, Romantic, and Symbolist poetry and
in Whitman’s “meditative catalogue,” with similar effect.

31
8
does not give an account of a rhythmic experience, only of one important

ingredient, which by itself does not make a rhythmic experience.
Scott handles equivalence in an all-or-nothing manner: equivalence for
him is near-synonymous with identity. For me, “equivalence” is based on
structural resemblance (not identity), which allows for varying degrees of
similarity. Equivalence is achieved when the most similar options avail-
able in the target system are chosen. The guiding principle for this is the
similar perceived effects to which the structures in question contribute. In
Structuralist phonology, speech sounds are conceived of as bundles of dis-
tinctive features. For Scott, different speech sounds cannot be equivalent,
only different. For me, different speech sounds may share similar distinc-
tive features that can contribute to similar perceived effects. In this case
I refer to varying degrees of equivalence.
As to metrical equivalence, French meter is based on syllable count,
whereas English meter is syllabotonic, that is, based on syllable count and
a regular sequence of upbeats and downbeats. We agree that equivalence
between, for example, the French alexandrine and English iambic cannot
be based on their similar traditional status. For Scott, however, English
meter is opposed to French meter, wholesale, by virtue of the “regular
sequence of upbeats and downbeats” in English meter. For me, English
iambic is equivalent to French meter in a way in which the trochaic, anapes-
tic, and dactylic meters are not: by virtue of their similar perceived effects,
both are more tolerant than the other meters of irregular stress patterns.
As to rhythm, for Scott it is contrasted to meter; for me, it results from
an interplay between prose rhythm and meter. I adopted this view from
Wellek and Warren, who say that poetic rhythm can only be accounted
for by relying on three dimensions: prose rhythm, meter, and perfor-
mance. Wellek and Warren need “performance” to make clear that when
you measure relationships in a recording, you measure an accidental per-
formance, not poetic structure (this is, indeed, what happens with many
measurements from the 1920s to the 1960s or even later). In his brilliant
essay “ ‘Prose Rhythm’ and Metre,” Roger Fowler (1966) elaborates on
the first two of Wellek and Warren’s notions, mentioning “performance”
only for the sake of completeness. Generative metrists have reinvented
Wellek and Warren’s “prose rhythm” and “meter” but ignore performance
(some of them even object to it). I go one step further: poetry reading is
a problem-solving activity on both the semantic and the prosodic levels.
In understanding a metaphor you accommodate the conflicting meanings
in a semantic interpretation, whereas in a vocal (or subvocal) reading you
accommodate the conflicting patterns of prose rhythm and meter in a
rhythmical performance. Thus, the semantic and rhythmic processings of
8
4
1
a poem are governed by a homogeneous set of principles. According to this

view, we are not measuring accidental vocal features but, rather, vocal fea-
tures that serve to accommodate the conflicting patterns of prose rhythm
and versification in a rhythmical performance constrained by them.8
Accordingly, Scott and I mean very different things by rhythm. If I under-
stand him correctly, he means something very similar to prose rhythm,
whereas I mean the accommodation of conflicting patterns of prose rhythm
and meter. Briefly, one of the main points in my counterproposal is to sug-
gest the conception of rhythm-as-problem-solving instead of Scott’s con-
ception of rhythm-as-prose-rhythm.
SUMMARY AND CONCLUSION
This chapter is a counterproposal to Clive Scott’s claim that translation,

irrespective of what kind of verse it is translating, should always opt for
free verse in the translated text. It reviews three possible approaches to the
issue: literal translation plus close reading, opting for free verse, and an ele-
gant solution to a problem, resulting in an aesthetic object in its own right.
The third is the solution expounded here. None of these approaches is right
or wrong per se. Each one offers its own set of gains and losses, which can
offer the reader different kinds of experience. Thus, each approach should
be assessed in view of the kind of experience it provides.
The basic assumption of the proposal propounded here is that the
aesthetic object is an elegant solution to a problem. This solution is con-
strained by conflicting stringencies. The elegant solution emerges when a
text accommodates incompatible stringencies in an overall structure. The
grammar and vocabulary of a language as well as poetic conventions are
such stringencies. The requirement that the translated poem should be as
similar to the original as possible is just one more stringency. Not all the
options in one semiotic system are available in another. Equivalence in
translation will be judged according to whether the nearest options avail-
able in the target system are chosen.
I agree with Scott’s misgiving that “it is easy to forget that translation
is not a translation of the signifier into the signified, but of the signifier
into another signifier” (2011: 73). Consequently, this chapter explores the
8. Gerry Knowles (personal communication) objected to my use of prose rhythm. He

said that this was too fuzzy a notion. So we agreed to use linguistic patterns instead.
Likewise, musicologists and some generative metrists use meter in a different sense,
and this has led to frequent misunderstandings. So, I added versification, which by the
same token also covers rhyme, alliteration, etc.

851
problems of equivalence in relation to two aspects of the sound patterns of

poetry: speech sounds and meter. In this respect I noted that versification
patterns and linguistic stress patterns are conveyed by the same noises;
but they conflict, nevertheless. The pattern of performance that accommo-
dates them is also conveyed by the very same noises. Of course, the whole
array of patterns such as alliteration, consonance, pararhyme, and so on
are conveyed by the same noises.
The Structuralist view of speech sounds as bundles of distinctive fea-
tures and perceptual qualities offers a wealth of fine-grained options for
oppositions and similarities to choose from. Poetic meter is based on regu-
larity and predictability. The French alexandrine is based on the predict-
ability of the number of syllables, the placement of the caesura, and the
order of “feminine” and “masculine” rhymes. The order of stresses is unpre-
dictable. In syllabotonic verse, stress is more predictable. Even when stress
pattern is irregular to some extent, it typically confirms the metrical pat-
tern at certain crucial points. In translating French poetry into languages
with syllabotonic meter the option most tolerant of unpredictable stresses
should be selected within this metric system, which is the iambic meter.
These considerations were explored with reference to French poems trans-
lated into Hebrew, Hungarian, and English (as well as one line from Hamlet
and its Hebrew translations).
The present chapter contributes to the debate on what in my mind is
a crucial variable: the perceived effect generated by the interaction of the
other variables. Equivalence is not based on a point-by-point correspon-
dence but, rather, on a device’s potential to help achieve a certain per-
ceived effect.
8
61
71
8
CHAPTER 8
More Is Up—Some of the Time
T here is a widespread convention in our culture, that rapid sound vibra-

tions are called “high” whereas slow vibrations are called “low.” That
this is a convention can be inferred from the fact that in some other cul-
tures different names are applied to the same distinction, “the crocodile”
and “those who follow the crocodile,” for instance, in the culture of the
Shona people in Zimbabwe. The sounds are played on an instrument called
the mbira (Figure 8.1), consisting of metal tiles on a wooden board (Berliner
1978). “Crocodile” refers to a low sound, and “those who follow the croco-
dile,” to a series of high sounds that “pursue” it.1
In this chapter I will compare and contrast two cognitive attempts to
account for the terminology used in our culture: the “mediated-association”
approach and the present “cognitive-constraints” approach. The latter
assumes that the spatial perception of sound is not an arbitrary conven-
tion but a cognitive fossil (in the weaker sense, as explained in chapter 1),
originating in a solution to a problem posed by cognitive constraints, in
this case the inconstancy of sound and the limitation on the resolving
power of the ear. The discussion below demonstrates in slow motion how
a fluid perceptual experience may solidify into a stable conventional usage.
Toward the end, however, I will consider evidence that in this particular
field fossilization may not be complete and conflicting intuitions may arise
in interesting ways.
Let us start with Zohar Eitan and Renee Timmers’s magnifi-
cent intercultural study of pairs of antonymic adjectives used to
describe music: “Beethoven’s Last Piano Sonata and Those Who Follow
1. I am grateful to Zohar Eitan for this information.

81
Figure 8.1 Mbira.
Crocodiles: Cross-Domain Mappings of Auditory Pitch in a Musical

Context” (2010). Though auditory pitch is customarily mapped in Western
cultures onto spatial verticality (high~low), both anthropological reports
and cognitive studies suggest that pitch may be mapped onto a wide variety
of other domains. These researchers collected a total number of thirty-five
pitch mappings from a wide variety of cultures and investigated in four
experiments how these mappings are used and structured. All experiments
included pitch metaphors of non-Western origins, of which participants
had no prior knowledge. Western participants’ matching of these meta-
phors with high or low pitch was congruent with their original applica-
tion. This congruence pertained not only to broad characteristics, such as
small~large and young~old, but also to seemingly idiosyncratic metaphors
such as sharp~heavy, grandmother~daughter, and crocodile~those who
follow the crocodile.
Eitan and Timmers work within the Lakoff (1993) model of spatial met-
aphors, but they repeatedly encountered some unexpected, even puzzling
findings. They found that there is a problem with the notion of more is up.
These findings crop up again and again:
Importantly, the conceptual mappings that serve as the very foundation of

other verticality metaphors do not apply, or apply weakly, to auditory pitch. In
particular, the “high is more” mapping, which (according to Lakoff and Johnson
[1980]) directly or indirectly serves as the most important foundation of verti-
cality metaphors (and, as mentioned, has also been proposed as the most impor-
tant basis of the pitch-verticality mapping) does not apply to pitch. Rather, for

91
8
pitch, “high is less” in several important ways: while spatially high objects are
large and higher in quantity, high pitch is small, empty and little in quantity. . . .
Similarly, there is a very high consensus that higher pitches are thinner and
sharper (i.e., of a smaller mass). (Eitan and Timmers 2010: 415)
The consensus in choosing “thin” to represent “low” is high for mass/size
(.93), but the consensus to choose thin to represent “low” in pitch is low (zero).
(Eitan and Timmers 2010: 417)
Higher pitch is positively correlated with higher intensity, and higher
intensity is positively correlated with larger size, but larger size is negatively
correlated with higher pitch. Such intransitive relationships indicate that the
cross-domain mappings of pitch are underlined by several, sometimes conflict-
ing conceptual metaphors. (Eitan and Timmers 2010: 420)
The present chapter explores only this and related issues. I claim that these
findings are puzzling only as long as you view them within the Lakoffian
model of spatial metaphors.2
MEDIATED ASSOCIATION VERSUS DIRECT PERCEPTION
I propounded the theoretical model used here in Tsur 1992a. There I also
warn that the behaviorist notion of “mediated association” has penetrated
some cognitive discussions, usurping the place of other, more adequate
cognitive explanations. In my later work I elaborated on these notions in a
criticism of mediated association and of Lakoff. In my article “Size-Sound
Symbolism Revisited” (Tsur 2006a) I contrast my position with the
mediated-association conception; in Tsur 2000b, 2002b, I contrast my
position with both mediated association and the Lakoffian conception. The
following discussion draws liberally on these works.
Roger Brown, the great psycholinguist who crucially influenced my early
thoughts on cognitive poetics, puts forward a conception of mediated asso-
ciation with which I disagree: “If the subject is required to guess, he will
call the loud and resonant voice ‘thick.’ This need not be because the voice
shares some inter-sensory quality with the visual or tactile apprehension
of thickness. It could be because the voice is loud and creatures who have
loud voices are usually thick, a mediated association” (1958: 152–153). In
Tsur 2012c I reproduce the photos of a thick soprano and a thin bass singer.
2. I have written on this and related issues in Tsur 1992a: chap. 3. At that time I did
not think of stating my position with reference to Lakoff. So it was not a response to
Lakoff but, rather, mainly to my own perceptions.
M or e I s Up — S o m e of t h e T i m e [ 189 ]
9
01
More importantly, I argue that bass voices are perceived as thicker than
soprano voices not because creatures that have bass voices are usually thick
and heavy but precisely because they share “some inter-sensory quality
with the visual or tactile apprehension of thickness.” Whereas the relation-
ship between thick people and bass voices appears to be quite incidental,
the relationship between thick violin strings and “thick” and “low” sounds
seems to have good physical reasons. Sounds are vibrations of the air or
some other material medium. The thicker the string, other things being
equal, the slower the vibrations and the greater the wavelength. To this one
may add that the “lower” the tones, the greater the number of overtones
within the audible range. Thus, the lower tones have a “thicker” envelope of
overtones than the higher tones.
The notion of “mediated association” is the same as Lakoff’s notion of
“grounding”: “Contemporary theory [of metaphor] postulates that the
more is up metaphor is grounded in experience—in the common experi-
ence of pouring more fluid into a container and seeing the level go up, or
adding more things to a pile and seeing the pile get higher” (1993: 240).
More recently, Lakoff makes an even more extravagant claim: “In a ther-
mometer oriented vertically, the mercury goes up physically as the temper-
ature increases (metaphorically goes up)” (2012: 31). Here one must even
manipulate the physical object in order to make it literally “go up” rather
than, for example, “from left to right.” It would appear that “grounding” is
merely a more cognitive phrasing for what Brown calls “mediated associa-
tion.” To be more precise, it disguises the behaviorist origin of the notion.
It relies, however, more on lifelong conditioning than on cognitive pro-
cesses proper.
The source of Eitan and Timmers’s problem is not so much “that the
cross-domain mappings of pitch are underlined by several, sometimes con-
flicting conceptual metaphors”; rather, it results from the nature of physics,
on the one hand, and our perceptual coping with it, on the other. Briefly,
frequency and wavelength co-vary, but in opposite directions: when fre-
quency is greater, wavelength is smaller, and when wavelength is greater,
frequency is smaller. Thus, faster vibrations are up, as predicted by Lakoff,
but this entails that greater wavelength is down. Here the doctrinaire slo-
gan “more is up” is helpless in the face of the irrevocable facts of phys-
ics. Frequency is a temporal notion, and length, a spatial notion; that is,
in cross-modal mapping the spatial template of pitch is matched with the
temporal, not the spatial aspect of the sound stimulus. In other words, if
you conceive of the phenomenon as perceptual rather than conceptual, you
must perceive (in terms of Eitan and Timmers’s second experiment) or think
(in terms of their first experiment) that the higher the pitch, the thinner

1
9
and smaller the sound. Such complexities are beyond the scope of the sys-
tem of conceptual metaphors.
The reasoning underlying this claim is that the sequence of sounds
has a certain phenomenological character that is conveyed by an acoustic
sequence with a different structure (for instance, instead of relatively fast
vibrations and relatively short waves we hear “high” and “thin” sounds).
The minute time periods between vibrations exceed the resolving power
of the human ear. The ear, therefore, fuses the sequence into one continu-
ous whole with a unique phenomenological quality: we become conscious
of the resulting sound as “higher” or “lower.” Such a phenomenological
quality has greater survival value than perceiving and counting the minute
time intervals per second, even if this were possible. In the same way, the
overtones of a sound exceed the resolving power of the ear, so that it is
forced to fuse them into “tone color.” Again, hearing a wooden or metal-
lic noise versus the smooth thump of some jumping feline may make
all the difference and may have greater survival value than perceiving a
multiplicity of overtones. Thus, both tone color and pitch result from the
limitations of the human ear’s resolving power. We attend to the phenom-
enological quality, but at the same time we perceive, subliminally, some of
the acoustic structure.
The foregoing discussion throws some new light on Eitan and Timmers’s
following argument: “Cross-cultural studies of sound symbolism in lan-
guage suggest an association of physical magnitude with pitch, such that
larger size and mass, as well as secondary qualities associated with bodily
magnitude, like slower pace and dominance, are related to lower-pitched
tones in speech and in other vocal utterances” (2010: 407). “Secondary
qualities associated with bodily magnitude, like slower pace,” are, again,
typical mediated associations. In light of the foregoing we need no “medi-
ated association” here: “slower pace” is not among the “secondary qualities
associated with bodily magnitude.” “Slower pace” and “bodily magnitude”
are directly perceived in the stimulus, in the relative rate of vibration and
the relative wavelength.
This conception can also account for the following findings mentioned
in Eitan and Timmers’s general discussion: “ ‘Higher’ loudness and pitch
interact with other dimensions in contrasting ways. Increased loudness, for
instance, is ‘bigger,’ while increased (higher) pitch is ‘smaller’ ” (2010: 419).
These findings are in perfect harmony with my foregoing argument. Both
findings are correlated with wave size: increased pitch with smaller wave-
length, increased loudness with greater wave width.
A similar view underlies at least one cross-cultural study of size-sound
symbolism in language (Ultan 1978). By examining a total of 136 languages,
921
Russell Ultan found that proximal distance is symbolized overwhelm-

ingly by front or high vowels. Let me add to Ultan’s sample a language not
included in it, my native Hungarian, in which itt means “here,” ott means
“there,” ez means “this,” az means “that,” így means “in this fashion,” úgy
means “in that fashion,” ilyen means “of this kind,” olyan means “of that
kind,” and so forth: “Since high front vowels reflect proportionately higher
second formant frequencies, and the higher the tone the higher the natu-
ral frequency, there appears a correspondence between a feature of high
frequency (= short wavelength in physical terms) and the category of small
size” (Ultan 1978: 545). Thus, Ultan’s findings also suggest that beyond
the phenomenological quality (vowel, thinness), we subliminally per-
ceive some of the physical processes of different structures excluded from
consciousness.
It would appear that more is up is true in all cases, except the ones in
which the opposite is true, as in “deep~shallow,” for instance, where more
is down and less is up. The rule seems to be this: when we measure upward
directions, more is up, and when we measure downward directions, more
is down. If this is true, Lakoff’s theory is based in this respect on a hidden
tautology. It would be more accurate to say that “up” is the unmarked verti-
cal direction (which we prefer when there are no explicit reasons to prefer
the downward direction).
ANTONYMOUS ADJECTIVES AND MARKEDNESS
That which is longer is less short, that which is higher is less low, that
which is faster is less slow, and so forth. The two terms of each pair,
however, are not symmetrical at all, as we shall see. In Tsur 1992a I tried
to solve some of the problems concerning such antonymous adjective
pairs by invoking the cognitive research of Clark and Clark and the
Structuralist notion of “markedness.” At that time I was not concerned
with arguing against Lakoff. Only years later did I compare my analyses
with his theory.
According to the Clarks, in each pair of antonyms small children
first learn the term that denotes the pole that has more (long, high, fast,
more, etc.):
When three year olds are asked about two toy apple trees Does one tree have more
apples on it? or Does one tree have less apples on it?, they can correctly answer yes,
just as adults do. But what happens if they are then asked which tree has more or

931
which has less? . . . In response to both questions, the children tended to choose
the tree with more apples on it, treating questions with less as if they contained
more. (Clark and Clark 1977: 505)
The preference for the dimension where there is more has good cognitive
reasons behind it. It is easier to see which has more. This seems to be
the source of two facets of the antonymous pairs of adjectives in adult
language:
1. The pole that has more is the unmarked pole, which is selected when
there is no explicit reason to select one or the other pole. We ask, “How
long was the movie?” if we don’t know whether it was long or short for a
movie. “How short was the movie?” would suggest that we already know
that it was very short.
2. The unmarked term is also used to denote the whole scale. We usually
speak of “measures of length”; only in very special cases would we speak
of “measures of shortness.”
Accordingly, we may say, “The child has a temperature” (where temperature

refers to the unmarked pole), but also, “Today temperatures are low” (where
it denotes the whole scale). The Clarks’ developmental finding that small
children use the unmarked term for both meanings can easily account for
this ambiguity; Lakoff’s anecdotal model of piles and liquids cannot, not
even the thermometer oriented vertically.
It would appear that the rule for cross-modal mapping is that the
unmarked pole of one scale is matched with the unmarked pole of the
other. Thus, when mapping the vertical scale onto the scale of vibra-
tions (fast~slow), the unmarked (upward) direction is used, and then
faster vibrations are matched with higher pitch. Temperature “goes up”
not because the mercury in the thermometer goes up (it may go in any
direction) but because the unmarked pole of the unmarked vertical scale
is matched with the unmarked pole of the temperature scale. In the case
of “deep~shallow,” by contrast, deep is the unmarked term; thus, more is
down, and less is up. Unmarked in “unmarked pole” means here “that
which has more”; in “unmarked vertical scale” it means “that which is pre-
ferred when there are no good reasons for preferring the opposite option.”
In fast vibrations are up there is no good reason to prefer the oppo-
site direction. In the case of deep is down and shallow is up or thin
sounds are up and thick sounds are down there are good reasons to
prefer the opposite option.
941
NONCATEGORIAL SOUNDS AND RECODING
Here two crucial questions arise: First, why do we need at all to recode the
sequence of sounds into a spatial template? Second, as we have seen, there
is an inverse relationship between frequency and wavelength: greater fre-
quency goes, necessarily, with shorter wavelength, and vice versa. So why
do we match the unmarked pole of the spatial template precisely with the
unmarked pole of sound frequencies and not with the unmarked pole of
wavelength? In other words, why do we say fast vibration is up rather
than thick sound is up?
To the first question Lakoff would answer “the embodied mind.” I, by
contrast, offer the following answer. The spatial template serves as a facil-
ity for handling noncategorial sensory information. Neisser has noted
that “sounds inform us about events. While vision and touch enable us to
explore stationary environments, hearing tells us only about movement
and change” (1976: 155). Sound is a rapidly changing stream of informa-
tion, consisting of minute stimuli that most accurately signal change. We
can discriminate such rich sensory sound information only as long as it
reverberates in echoic memory, that is, for a few seconds only. In order to
store auditory information for longer periods of time, it must be recoded
in some more “stationary” form that depends less on the niceties of unique
sensory information and is more easily managed by memory. For present
purposes, there appear to be two such ways of recoding: categorization into
a phonetic code and translation into a code of spatial relationships (of a
more stable nature).
The first of these codes uses a system of abstract linguistic categories
that lend themselves to storing for relatively long periods of time, but at
the price of excluding most of the acoustic information. Suppose we ask,
“What did the man say?” and get the answer, “The man said ‘ba.’ ” We are
able to remember this syllable for an hour, or a day, or a month, or even
a year but will be incapable of even reliably noticing the acoustic cues by
which it differs from, for example, “da” or “ga.” Had we received only the
acoustic cues, we could reliably remember them for a few seconds at best.
Consider Figure 8.2. The density of the curves represents frequency. The
distance between the curves represents wavelength, whereas the devia-
tion of the curves from the midline represents amplitude. Obviously, the
distance between the curves (wavelength) co-varies with their density
(frequency): the shorter the wavelength, the greater the density. In the
nonspeech mode we are able to perceive the physical shape of such a sound
wave, but not in the speech mode. If it were possible to perceive such a
sound shape in the speech mode too, we would be hard put to recognize

951
Figure 8.2 Wave plot of [ba].
the syllable ba or even to retrieve the sound sequence after several seconds.
The sound wave in Figure 8.2 gives information about the speaker’s artic-
ulatory gestures, which the listener’s articulatory system decodes as the
intended syllable, “ba.” Subliminally, at the same time, one may sometimes
also perceive some of the sound wave’s physical shape.
The second of these codes exploits the fact that spatial organization is
of a more stable nature (not necessarily determined by our bodily expe-
riences) than the sequence of sounds: it conceives of the relationship
between sounds as spatial relationships. Thus, if we cannot remember the
exact sensory information about a sound, we can still remember the exact
relationship between several sounds—that is, a scale or melody—and it
will be the same even if transposed from one scale to another.
The spatial structural template not only enables us to remember relation-
ships between sounds but also enables us to differentiate better between
them—but at the expense of losing valuable sound information. The
same is true, mutatis mutandis, of phonetic recoding. In both, boundar-
ies between categories can more reliably be discerned than within-category
information. One way to demonstrate this is via so-called categorial per-
ception. It has been experimentally established that listeners discriminate
the same intervals less reliably within categories than at their boundaries,
961
in both speech and music perception. In other words, those differences that
are required to distinguish between categories are more reliably discerned
than those that are not required. What is more, contrary to commonsense
expectation, professional musicians are less good at within-category dis-
tinctions than some musically naïve listeners. In natural noises there is no
categorial perception.
Categorial perception was long thought to belong only to speech
perception. In his Ph.D. dissertation, however, Mark J. Blechner (1977)
explored categorial perception in music. He used the categories “major”
and “minor chords.” A chord is a group of (typically three or more) notes
sounded together, as a basis of harmony. In a three-note group, the
two kinds of chord differ only in the pitch of the middle note. Blechner
divided the one interval between the two middle notes into nine equal
steps. In each chord, the high and low pitches were always 392 and 262
Hz, respectively. The central tone varied in discrete steps of 2.32 Hz,
from 311 to 329.6 Hz. (The complete set of stimuli is displayed graphi-
cally in Figure 8.3.) Blechner conducted two experiments with three
groups of subjects: professional musicians, nonprofessionals highly
skilled in music (NP-H), and nonprofessionals with low skill (NP-L). In
400
380
360
340
Hz
320
300
280
260
1 2 3 4 5 6 7 8 9
(prototype (prototype
minor) major)
Chord Number
Figure 8.3 Schematic representation of chords used in Experiment I. (From Blechner 1977.)

971
one experiment two consecutive triads were played, and then a third
one repeated one of them. The subjects had to identify which one was
repeated. In the other experiment only two triads were played, either
two consecutive triads or one triad repeated. Subjects had to answer
“same” or “different.” The performance of NP-L subjects was random.
The performance of professional musicians was accurate at the bound-
ary between the two categories; otherwise it was quite poor—they were
subject to categorial perception. Only the results for the NP-H subjects
did not indicate categorial perception.
Of course, one could argue that all our perceptions and categories are
determined by our body or our brain. In other words, the “embodied mind”
may cover the total range of mental possibilities. This may be a sweeping
and valid generalization, but it is of doubtful usefulness. In this case, we
must distinguish different kinds of “embodied mind,” such that they not
only acknowledge an overall notion of the body’s interactions with the envi-
ronment and the ontological assumptions about the world that are built
into the body and the brain but also account for, for example, three different
modes of cognition that prevail one at the expense of the others, according
to immediate needs. If the minute changes in our environment are impor-
tant, we attend to the noises changing from second to second; if we wish
to remember certain stable relationships between sounds, as in a scale or a
melody, we translate them into a stable spatial template, at the expense of
the changing minute sounds that indicate moment-to-moment changes in
the environment. Alternatively, we might wish to translate them into stable
speech sounds. To demonstrate how the same noise can be heard as unde-
fined noise in one context and as a well-defined speech sound in another,
I created a sound file containing five stimuli: a series of noises produced
by clapping my hands, the word suck recorded from the Merriam-Webster
Collegiate Dictionary—Audio Edition, the same word with the [k]‌replaced by
a clapping sound, an amplified token of the genuine [k], and a token of the
clap excised from the doctored word.3 The same noise is heard as a click with
an elusive echo-like aura in the clap series and as an abrupt [k] in the word.
The present chapter considers evidence that behind the stable catego-
ries of speech and music we subliminally perceive inconstant sound infor-
mation regarding the physical stimulus that escapes categorization. It is
treated as a byproduct of the categories directly used for communication
and is perceived as the thickness, size, color, relative brightness, or pace of
the sound stimulus. As I said, these aspects do not serve communication
directly but are perceptual qualities of the stimulus.
3. Listen to the sound file from the companion website.
9
81
CROSS-MODAL MAPPING
As to the latter question—namely, why we match the unmarked pole of

the spatial template precisely with the unmarked pole of sound frequencies
and not with the unmarked pole of wavelength—one way to answer is to
invoke the principle of maximum contrast: frequency is a temporal notion,
whereas wavelength is a spatial notion. In cross-modal mapping the spatial
template is matched with the temporal, not the spatial dimension of the
sound stimulus. Music is a time art, whereas a scale is a spatial notion. In
music one may distinguish event time and micro-time. Event time refers to
the duration of notes and pauses; micro-time refers to the duration between
vibrations. Why do we say that “event time and micro-time are indispens-
able attributes of sound” (to use Michael Kubovy’s [1981] term) rather
than “event time and wavelength are indispensable attributes of sound”?
Because “time is an indispensable attribute of sound” is more parsimoni-
ous than “time and space are indispensable attributes of sound” and more
clearly contrasts sound perception to visual perception (according to the
scientific principle of parsimony, things are usually connected or behave
in the simplest or most economical way). Therefore, perceiving micro-time
(frequency) as the primary quality of sounds (mapped on a spatial scale)
and wavelength (relative thickness) as the secondary quality seems to be
the most parsimonious organization of sound perception.
For some reason, the spatial template in music serves to convey the
main information: frequency (micro-time), the indispensable attribute of
sound. Melodies consist of sequences of “higher” and “lower” sounds. At
the same time, wavelength is perceived as thickness, which is treated as a
perceptual byproduct. Timbre (sound color, the structure of the overtone
envelope) as well is treated as a secondary quality of music. This is not nec-
essarily so in all cultures. In some Tibetan music, for instance, singers keep
pitch constant and vary only the sound color (which yields beautiful results
for the Western ear as well). But timbre, too, is generated by micro-time: it
is the perceived quality of a configuration of overtones determined by
frequencies.
Why does the spatial template in which we recode pitch go precisely
upward rather than, for example, from left to right or from back to front?
After all, keyboard instruments would suggest a left-to-right position,
whereas the violin would suggest a front-to-back position; what is more,
the violoncello would suggest a scale in an upside-down position. Past expe-
rience with the most familiar musical instruments can hardly account for
the “upward” position of the scale, just as holding the thermometer in an
upward position cannot (pace Lakoff) explain why temperature “goes up.”

91
John Lyons extracts from localist thinkers an illuminating idea that can
easily account for this but, alas, without giving references (I strongly sus-
pect that these are his own conclusions from reading the localist linguists):
We live and move, normally, on the surface of the earth (rather than in water
or in the air); and we do so, again normally, in an upright position. This gives
us the means of identifying one of the dimensions in a three-dimensional
space; it also gives us a fixed zero-point at ground-level. Directionality in the
vertical dimension—i.e. the difference between upwards and downwards—is
established by our experience of the effects of the force of gravity, by the fact
that, normally, the sky is above us and the ground beneath us and by the asym-
metry of the human body in the vertical dimension. For these, and other rea-
sons, verticality is physically and psychologically the most salient of the spatial
dimensions. . . .
There are two horizontal dimensions, neither of which is fixed, in the way
that verticality is, by the force of gravity or anything comparable. . . . But [man]
is asymmetrical in one of the two horizontal dimensions, and symmetrical in the
other: i.e. he has a front and a back, and two symmetrical sides. He has his prin-
cipal organs of perception directed towards the region in front of him. . . . The
asymmetrical front-back dimension is less salient than the vertical direction,
but more salient than the symmetrical right-left dimension. (1977: 690–691)
The “fixed zero-point at ground-level” may explain why the upward direc-
tion is the unmarked vertical direction, whereas the downward direction,
beneath ground level (as in deep~shallow), is the marked one. What is
upward from ground level is more easily accessible to our senses than what
is downward from it. So, we should rephrase Lakoff’s slogan as more is up
in the unmarked vertical direction and more is down in the marked one.
When there is no explicit reason for preferring any one of them, we have
recourse to the unmarked one.
Thus Lyons creates a markedness scale of the directions of spatial tem-
plates in cross-modal mapping: vertical, front~back, and left~right. When
there is no specific reason to prefer any one of these directions, the least
marked (vertical direction) is chosen, as with music and temperature. Each
later item in this scale is chosen when there is good reason not to prefer the
preceding one(s). The left~right direction is sometimes simply arbitrary, as
in left-wing and right-wing politics. To be sure, the usage “left-wing and
right-wing politics” has good historical reasons, but it is arbitrary in the
sense that it “does not contain in itself the reason why it is so, and not
otherwise”—to use Coleridge’s phrase. If you ask whether the adjectives
in “high and low pitch” or “advanced and backward education policies”
02
could be reversed, you will most probably get a negative answer. If you ask
whether in “left-wing and right-wing politics” they could be reversed, you
will most probably get a positive answer.
Note, however, that such an analysis explains only why the vertical
direction is preferred as spatial template in cross-modal mapping. But to
account for the matching of the “faster” pole with precisely the “height”
pole in music, or the “more” pole with the “down” pole with reference to
the depth of the sea, you need some cognitive-developmental model like
the Clarks’, combined with the Structuralist notion of “markedness.” At the
same time, to understand why “high” sounds are “thin” and “low” sounds
are “thick” rather than the other way around, one must know a thing or two
about the structure of the physical stimulus.
There seems to be some deep affinity between Lakoff’s ideas and those
of the localists who preceded him by decades or centuries (depending on
whom you ask). Jean-Michel Fortis suggests that some “cognitive linguists
had little knowledge of the past of their own discipline (that is why authors
like Lakoff advertise some of their ideas as new and even as breaking away
from the bonds of tradition)” (2011: 10). But his localist origins did not
prevent him from accounting for the more is up principle using our expe-
rience with liquids in a container or for temperature “going up” using the
mercury in a thermometer oriented vertically.
LAKOFF’S NEURAL THEORY OF METAPHOR
When I first read Lakoff’s 2008 essay “The Neural Theory of Metaphor”
(see 2012), I hoped that this fresh angle on metaphor would help him over-
come some of the shortcomings of his theory. Later I discovered an ear-
lier article, “The Brain’s Concepts: The Role of the Sensory-Motor System
in Conceptual Knowledge,” by Vittorio Gallese and George Lakoff (2005).
Gallese is an eminent brain scientist, one of the group of scientists who
discovered mirror neurons.
In the first part of his 2008 essay Lakoff gives an outline of his Neural
Theory of Language developed with Jerome Feldman, based on Gallese’s
findings. This in itself is a masterpiece of conciseness and elaborateness
at the same time. In the second part of the essay Lakoff presents his “old
theory.” I will not go into the details of this essay, but only dwell on the
one point that interests us here. As I mentioned above, in this later essay
Lakoff falls back on his earlier theory of grounding: “In a thermometer
oriented vertically, the mercury goes up physically as the temperature
increases (metaphorically goes up).” This he explains as “More Is Up: Our

2
01
bodies are constantly monitoring physical height more than computing

abstract quantity” (2012: 31). In turn, he explains that “those neurons that
fire more tend to develop greater firing capacities. And those involved in
physical bodily functioning tend to fire more” (2012: 28).
I find that this argument is problematic in two ways. First, with refer-
ence to the same neurons, nodes, and brain circuits, Lakoff could explain
not only why “more is up” in the thermometer for him but also why “more
is right” for me and “more is left” for my mother. The argument is fallacious
not on the neural level but on the level of the assumption that “more is
up” is grounded on conditioning by such everyday experiences as piles and
liquids becoming higher or the mercury in thermometers going up. Second,
both Gallese and Lakoff use embodied as their central term, but they use it
in radically different senses. Curiously, in the Neural Theory of Language
part of his essay and the “old theory” part of his essay Lakoff uses the term
in two incompatible senses without giving evidence of being aware of it. The
sentence “Those [neurons] involved in physical bodily functioning tend to
fire more” would make no sense in Gallese’s frame of reference. According
to Gallese’s conception (including their joint article), the same neuron fires
when you grasp an object, or see someone else grasping an object, or imag-
ine grasping an object, or use the abstract concept of grasping. In this case,
at least, rather than revising his old conception in light of new neurological
findings, Lakoff extrapolates his old conception to neurological structures.
INCOMPLETE FOSSILIZATION?
I have suggested that the spatial conception of sound (“high” and “low”
sounds) is a convention of our culture but not an arbitrary one: It resulted
from a set of cognitive processes that fossilized into an automatic usage, a
pair of unqualified dictionary entries; when we use them, we don’t neces-
sarily have in mind the cognitive processes that generated them. In a later,
mind-expanding study, “Which Way Is More? Pitch Height, Parametric
Scales, and the Intricacies of Cross-Domain Magnitude Relationships”
(2013), Eitan points out a much more intricate state of affairs in this field,
almost a muddle: there are intuitions conflicting in an interesting way, and,
still, there seems to be method in this madness—to use Polonius’s words.
This would suggest that the cognitive processes involved have not entirely
fossilized.
In this essay, Eitan reviews literature on research on cross-domain map-
pings of pitch. He presents a host of intriguing findings, which he sum-
marizes as follows: “The deceptively simple world of registral pitch thus
0
2
reveals a truly Carrollian structure: one where ‘more’ and ‘less,’ growing
or shrinking, keep shifting and reversing places and roles, depending on
whether one stands still or moves, on whether one is going up or downhill,
and on what objects of comparison one may happen to observe on the way”
(2013).
In what follows, I will attempt to sort out just a few of those muddles,
hoping that the discussion will light the way to making sense of the other
muddles as well. I draw on the principle underlying this and the preceding
chapter: that both in verbal metaphors and in cross-domain mappings of
pitch, visual and auditory stimuli are treated as bundles of features, each
of which may suggest a mapping, sometimes incompatible with the oth-
ers. Changing circumstances may require us to attend to different features
of the same stimuli or even have recourse to different modes of cognitive
processing. Thus, for reasons suggested in the foregoing discussion, greater
frequency of sounds is matched with the “high” end of the vertical spa-
tial template, so that in this case more is up. Greater frequency entails a
smaller wavelength and a poorer overtone envelope; consequently, “higher”
sounds are perceived as thinner and smaller than “lower” sounds, so that
in this case less is up. At the same time, in the intensity dimension of
the same stimulus, louder sound is associated with greater visual size. In
both cases, the conflicting “visual” sizes are correlated with the size of the
wave: thinner, high sounds are correlated with shorter waves, and louder
sounds are perceived as visually larger—with wider waves. “Higher” (more
frequent) sounds suggest greater speed; this can be accounted for by the
faster vibrations involved. The ingredient shared by all these accounts is
that we perceive a phenomenological quality conveyed by sound informa-
tion of a different structure; it is excluded from consciousness, but we per-
ceive some of it subliminally.
So far I have summarized the various aspects of the sound stimulus
as I have discussed them above. But Eitan points out a radical difference
between static and dynamic sound stimuli, where the mapping of more and
less may be reversed:
Dynamic and static pitch stimuli associate with other domains in different,
sometimes contradictory ways. Consider again, for instance, the ubiquitous
association of pitch with physical size. As noted above, the association of higher
pitch registers with smaller size crosses age and cultural boundaries, affecting
music, speech, and expressive vocalization, and was even associated with the
vocal behavior of diverse non-human species. However, when subjects (of differ-
ent cultures) associate dynamic pitch patterns—ascending or descending melo-
dies or pitch glides, rather than high and low pitch register—with visual stimuli,

0
32
they map pitch rise into growing size, and pitch fall into shrinking size—just the
opposite of what static pitch mappings would imply. In a recent study by Kim
and Iwamiya, Japanese participants judged rising pitch patterns as congruent
with expanding (rather than shrinking) animated visual shapes. (2013: 9)
I submit that these differences between static and dynamic stimuli stem
from different modes of cognitive processing. Shape perception and direc-
tion perception are related to different brain centers. In static shape per-
ception we attend to the (unchanging) relationship between the parts of
an object. Regarding moving objects, spatial location becomes salient: we
ignore the relationship between its parts and attend to the changing rela-
tionship between the object and surrounding space. Movement and direc-
tion direct attention away from the internal structure to the relationship
with surrounding space. A sustained “high” pitch suggests a certain physi-
cal structure: short wavelength and relatively few overtones—perceived
as “small” or “thin.” Regarding “ascending,” we attend away from the
structure of the object and focus on its changing relationship to the
surrounding space.
The experiments use explicit demonstrations (in audiovisual matching
or metaphor rating for music and speech sounds) and implicit demonstra-
tions, through a Stroop-like effect in which processing is faster for congru-
ent pitch and visual movement;4 for example, participants judged rising
pitch patterns as congruent with expanding (rather than shrinking) ani-
mated visual shapes, or, when presented with an octave leap, they tended to
term the higher pitch “big” and the lower pitch “small,” in contrast to results
for isolated pitches. In such experiments the experimental design deter-
mines whether pitch change is associated with one moving point or with a
continuously changing shape. What is common to both is that attention is
shifted away from the structure of the wave to movement through space. As
I have argued all along, musical pitch has two coinciding physical correlates,
frequency and wavelength. Accordingly, the same sound may be described
as “high” (spatial location), suggesting a greater extent than a “low” sound,
or as “thin” (perceived size), suggesting a smaller extent than “thick.” When
switching from static to dynamic percepts, we change cognitive processing
modes: we shift attention from perceived size to spatial location.
Eitan explains the finding that in a dynamic context subjects map pitch
rise to growing size and pitch fall to shrinking size as follows: “For dynamic
stimuli the more abstract (and perhaps language-based) analogy between
4. The Stroop effect is a demonstration of interference in the reaction time of a task

(see Stroop 1935). See chapter 3, p. 67–68.
4
0
2
pitch ‘rise’ and physical growth may take hold” (2013: 10). If we assume
that we are dealing with sound recoded into a spatial template, as suggested
above, we do not need an “abstract (and perhaps language-based) analogy.”
We simply shift from one mode of cognitive processing to another and
attend to different kinds of relationships in the situation.
As we have seen, Lyons suggests “a fixed zero-point at ground-level” in
the vertical dimension in spatial relationships. In a dynamic context we
shift attention from wave structure to motion away from the ground level;
this, in turn, suggests expanding, whereas moving back to ground level
suggests shrinking distance. Again, the conflicting mappings of the static
and dynamic stimuli result from switching from one aspect to another. The
switch is prompted by the changing needs of the task. This explanation dif-
fers from Eitan’s, since it does not assume linguistic mediation but, rather,
attending to different aspects of the same physical stimulus.
Or consider the following: “Pitch space is directionally asymmetri-
cal: pitch descent is not the opposite of pitch ascent. The two directions of
pitch change may associate with different, not opposite dimensions. While
pitch descent, for instance, implies spatial descent, pitch ascent does not
imply spatial ascent, but rather acceleration” (Eitan 2013: 15). To some read-
ers this sounds puzzling, but Eitan relies on solid experimental evidence.
In my opinion, this may be the truth but not the whole truth—which I will
try to sort out in what follows.
Pitch ascent implying acceleration would be fully consistent with the
foregoing analysis. As we have seen, high pitches are perceived as fast
because vibrations in the higher registers are faster than in the lower regis-
ters. Now, when you have a gradually rising pitch, vibrations become gradu-
ally faster—hence the sense of acceleration.
On the surface, this conforms to my analysis above. In “acceleration” we
focus on (changing) perceived speed; in “spatial descent” we attend away,
for some reason, to spatial location. However, ascending or descending
pitch may imply many things, among them acceleration or spatial ascent
for ascending pitch and spatial descent or coming to a rest for descend-
ing pitch. Moreover, a falling pitch may suggest stability or confidence,
whereas a rising pitch may suggest instability, lack of confidence, or, when
loaded with energy, defiance. In an interrogative sentence it may simply
imply uncertainty. Thus, rising pitch can imply various degrees of instabil-
ity. In music, pitch falling to the tonic, or in speech, intonation falling to
the zero point (at the “ground level”), indicates “coming home,” coming to
a rest. In John Ohala’s (1994) experiments with stripped speech (where
only the intonation contour can be heard, and the words cannot be dis-
cerned), a long-falling intonation contour suggests dominance. Again in

0
52
the macro-structure, gradually slowing down (ritardando) suggests com-

ing to an end, in both music and speech. Note that in the micro-structure,
gradually falling pitch consists of vibrations gradually slowing down. Thus,
falling pitch suggests, in both speech and music, coming to a rest on the
macro- and micro-structure levels (“coming home” and slowing down) at
the same time.
All this strengthens the basic conception underlying this chapter that
both verbal metaphor and cross-domain mapping of pitch treat visual and
auditory stimuli as bundles of features. Humans move with amazing ease
from one feature to another, generating a wide range of sometimes con-
flicting mappings. In certain experimental conditions fairly unambigu-
ous results may be obtained; in fact, scientific neatness demands highly
“significant” results. But in real-life situations our activities are marked by
exceptional flexibility that involves jumping from one mapping to another
or activating several features that indicate the same mapping.
Certain cross-domain mappings gain priority in experiments and yield
intriguingly asymmetrical results. We must assume that for some reason
they override other mappings, but when needed, all mappings are available.
Interestingly enough, in a paragraph quoted earlier, Eitan himself speaks
of symmetrical mappings of rising and falling pitch: “When subjects (of dif-
ferent cultures) associate dynamic pitch patterns—ascending or descend-
ing melodies or pitch glides, rather than high and low pitch register—with
visual stimuli, they map pitch rise into growing size, and pitch fall into
shrinking size.” I strongly suspect that it depends on the experimental task
whether pitch space is directionally symmetrical or asymmetrical.
I have argued that the terms high and low regarding sounds are conven-
tions in our culture, resulting from a set of cognitive processes fossilized
into a rigid dictionary entry. In this section I have added a new perspective
to my overall argument: the possibility that in some instances there may be
incomplete fossilization, leaving much room for creativity. Such creativity
is comparable to the way we treat live metaphors with unforeseen mean-
ings. Accordingly, sounds and words are treated as bundles of features
that may give rise to sometimes conflicting meanings (with metaphors)
or conflicting perceptual qualities (with sounds). In unforeseen situations,
human beings move, with amazing ease, from one feature to another, so as
to generate some significant (if unforeseen) meanings or perceptual quali-
ties by deleting certain features and retaining others, depending on the new
context. In Arabic and medieval Hebrew poetry the beloved is frequently
compared with a gazelle. Consider a metaphor such as “She has the legs
of a gazelle,” for instance; it may suggest “graceful,” “nimble,” “slender,” or
“hairy” legs in the appropriate contexts. With the first three suggestions,
62
0
nature would be “to advantage dress’d”; with the last one, to disadvan-
tage. You may also create a witty metaphor such as “She has the legs of a
gazelle—not as graceful, not as nimble, not as slender, but as hairy.” Such
a conception may account for the incompatible results produced by the
experiments reported by Eitan and may also suggest that the perception
of sounds and metaphors is governed by a homogeneous set of activities.
SUMMARY
In our culture, rapid sound vibrations are called “high”; slow vibrations,
“low.” In other cultures very different terms are applied to the same distinc-
tion. In this sense, this usage is a convention. It is, however, not entirely
arbitrary; it seems to originate in some active cognitive processes that have
fossilized into rigid dictionary entries. Lakoff attempts to locate these
cognitive processes with the help of his theory of “conceptual metaphor”
and principle of “grounding,” what Roger Brown calls “mediated associa-
tion.” In their impressive intercultural study of this dichotomy, Eitan and
Timmers invoked Lakoff’s theory, epitomized as more is up; but time and
again they came up against the same puzzling findings. “More is up” when
physical space is concerned but not necessarily when pitch is concerned.
Thus, for instance, sounds of “higher” pitch are “thinner,” not “thicker.”
This finding, they think, does not conform to the conceptual-metaphor
theory they espouse. They assume that there is some hidden conflict
between two or more conceptual metaphors. I argued here that “more is
up” only when you discuss upward directions, but in such pairs of adjec-
tives as deep~shallow, more is down and less is up. More importantly,
Eitan and Timmers’s puzzling findings cease to be puzzling if you do not
try to account for them within a theory of conceptual metaphor. Far from
being the result of a hidden conflict between two conceptual metaphors,
they conspicuously conform to the “cognitive-fossils-and-constraints”
approach proposed here, which relies on the feature-deletion theory of
metaphor, cognitive-developmental evidence, the limitations of the
human ear, alternative modes of cognitive processing, and an accurate
physical description of the auditory stimulus. Admittedly, this is less par-
simonious than simply more is up, but it has the advantage of account-
ing for the phenomena discussed. The conceptual-metaphor theory allows
you to point out that many metaphors may mean the same thing; the
feature-deletion theory allows you to point out that one metaphor may
mean many, sometimes unforeseen, things.

0
72
In a more recent essay, Lakoff (2012) expounds a neural theory of meta-

phor. I had hoped that such an advanced approach would resolve the prob-
lems pointed out by Eitan and Timmers and by me. On closer examination,
however, it turns out to be an extreme instance of the “neurological fallacy”
(see Tsur 2012b), where what is said in brain language at best repeats, or
is merely extrapolated from, what has already been said in the language
of other disciplines. As Lakoff himself says, “The fundamental outlines of
what we discovered remain as valid today as they were then” (2012: 17);
at variance with Lakoff, I construe this to mean that his “neural theory”
cannot account for the “anomalies” of more is up. Coming back to Eitan,
in a more recent essay he (2013) gives a wide survey of experiments in
cross-domain mappings of pitch. He reveals a veritable muddle of incom-
patible results. This adds a new possible perspective to the theory of cogni-
tive fossils: that in this domain fossilization is incomplete and the relevant
cognitive processes are very much alive and leave room for considerable
creativity in responding. One may introduce reason into this welter of
incompatible results by assuming that these responses to unforeseen map-
pings of spatial templates onto sounds are comparable to human beings’
ability to understand metaphors to which they have never been exposed.
8
0
2

9
0
2
CHAPTER 9
Some Remarks on the Nature

of Trochees and Iambs and Their
Relationship to Other Meters
The iambic is the characteristic rhythm of people as they talk. . . . The trochaic rhythm,
again, is too much akin to the comic dance, as may be seen in tetrameter verse, for the
rhythm of tetrameters is light and tripping.
—Aristotle, Rhetoric, 1408b
His rage armed Archilochus with his iambic: comedy and tragedy have adopted it, as being
natural for dialogue, able to drown out the noise of the audience and suited to action.
—Horace, The Art of Poetry
I n chapter 1 I said that this book is devoted to the question of how cogni-
tive processes shape and constrain cultural and literary forms and that
conventions can be used without any awareness of the cognitive processes
that shaped and constrained them or of the perceptions and experiences
related to those processes. In this chapter I will discuss a pair of metrical
conventions of major importance, the iambic and trochaic meters, and the
cognitive processes that shape and constrain their effect.1
1. Much of this chapter was written in 1971 and was published as c hapter 3 of my
1977 A Perception-Oriented Theory of Metre. Recently I realized that metrists are still
puzzled as to the difference between the iambic and the trochaic and the special status
of the iambic, which was pointed out as early as Aristotle and Horace. Since Tsur 1977b
is inaccessible today, I decided to republish this chapter as it appeared in that book,
with some updating and minor corrections.
021
My work in prosody assumes, following Wellek and Warren (1956),

that poetic rhythm in the syllabotonic system current in English poetry
can be best accounted for by three sets of patterns. First is an abstract
versification pattern, and second is the pattern of linguistic units. Some
aspects of the linguistic stress pattern confirm the abstract schema (one
could even say that the reader abstracts the metrical pattern from them);
some aspects deviate from it and produce tension. These deviations, far
from being signs of imperfection, or “unmetricality,” are major prosodic
and expressive assets. When readers encounter some deviation from the
abstract metrical pattern, they make adjustments in its performance, so as
to preserve both their metrical set, that is, their feeling of regularly alternat-
ing strong and weak positions, and, at the same time, the stress pattern of
their spoken language. This adjustment frequently consists in overarticula-
tion, overstressing, and additional grouping of stresses. This phenomenon
constitutes the third pattern, the pattern of performance. The greater the
deviation, the greater the adjustment required and the tighter the addi-
tional grouping. One of the basic assumptions of this chapter is that the
rhythm of a poem is accessible only through some kind of performance; an
adequate account of a poem’s rhythm can be given only by considering the
interplay of these three patterns: those of versification, linguistic stress,
and performance.
A lexically stressed syllable confirms meter in an s position; an unstressed
one, in a w position. Consider, for instance, Donne’s line:
1 2 3 4 5 6 7 8 9 10
(1) shall behóld Gód, and néver tást déaths wóe.
w s w s w s w s w s
The expected pattern of alternating w and s positions is confirmed in

positions I, IV, V, VI, VII, VIII, and X. In positions II, III, and IX it is
disconfirmed. Overarticulation, overstressing, and additional grouping
by the performer, however, may render the line acceptable to the ear
as iambic meter. By grouping together syllables 8, 9, and 10, we get a
sequence of three stresses, two of which coincide with s positions. In
this group, “deaths” can be pronounced as stressed, though relatively
less heavily than the adjacent syllables, thus preserving the basic iambic
cadence of the line. The deviant stress can be tolerated only so far as the
metrical pattern is emphatically reinstated in the two strong positions
of the group. Alternatively, one may perform the sequence as a group
of equally stressed syllables. In this case, the reader’s “metrical set” has

21
to rely on the mind’s inclination to memorize and anticipate the repeti-

tion of an initial pattern, creating counterpoint (and tension) between
the two sets. Overarticulation of the three stressed syllables is needed
to make mental processing space available for the parallel processing
of conflicting stress and meter. Thus, overarticulation and additional
grouping are a means of avoiding chaos while preserving the stress pat-
tern of ordinary speech.
In positions II and III, it is more difficult to preserve the stress pat-
tern and avoid chaos. Evidently, -hold cannot be performed so that both
adjacent syllables are more heavily stressed while preserving, at the same
time, the stress pattern of ordinary speech. The most obvious performance
here would be to overstress and overarticulate -hold and to tightly group
together the first four syllables, which would yield a symmetrical, fairly
“simple” gestalt: two unstressed syllables balanced against two stressed
ones. There is an impetuous drive to reach the fourth position, where the
stress pattern and metrical pattern have “coinciding downbeats.” Here,
the metric pattern is “reconditioned” and becomes “fresh and new,” to use
Leonard B. Meyer’s (1956: 118) phrases. I suggest that the counterpoint
in positions I–IV builds up greater tension than in positions VIII–X and,
altogether, they produce an uncommonly deviant (and tense) line that can
still be perceived as rhythmical.
THE TROCHAIC
The peculiar nature of the trochaic meter has frequently attracted atten-
tion. Time and time again, critics have felt the need to account for this
peculiarity. In what follows, I add yet another, cognitive item to the list of
would-be explanations and hope to demonstrate its explanatory force. “It
is interesting to note,” Chatman writes, “that lighter syllables seem more
readily ictic in trochaic than in iambic feet”: “The trochaic mode more easily
violates normal prose accentual patterns; it quite insists on dominating the
rhythm. Iambic verse seems not to exert its will so rigidly” (1965: 140–141).
As an explanation for this peculiar character, Chatman quotes Halpern and
then offers to substitute his own explanation:
Halpern’s thesis is that trochaic verse, along with anapestic and dactylic,
is a subspecies of the native Germanic strong-stress verse, which he feels is
both isoaccentual and isochronic. . . . I think his observations about the rela-
tive inflexibility of the trochaic verse are correct, but would suggest another
T r o c h e e s a n d I a m b s a n d T h e i r R e l at i o n s h i p t o O t h e r M e t e r s [ 211 ]
21
cause, namely the comparatively short history of the mode. The sophisticated
smoothness of the iambic verse has been long in developing; trochaic verse,
however, was not taken very seriously in England until the nineteenth century.
(1965: 141n)
Both explanations have the same weakness: neither the Germanic

strong-stress verse nor the history of English meter is a perceptual feature of
an actual piece of trochaic poetry. As Coleridge put it, “Nothing can perma-
nently please which does not contain in itself the reason why it is so, and not
otherwise” (Biographia Literaria, chap. 14). A child who attends to “Tackle,
tackle, Mother Goose / Have you any feather loose?” responds to the peculiar
trochaic quality in the poem with no particular instruction in the history
of English and Germanic verse. Furthermore, nursery rhymes in Hebrew
are preferably written in the trochaic meter. Russian poets and critics feel
that the trochaic meter has a particularly “encouraging, optimistic, vigorous”
quality. I take this to be a manifestation of the “dominating” character of the
trochaic, its tendency to “exert its will.” At any rate, a satisfactory explana-
tion of this peculiar trochaic quality will also have to account for the greater
rigidity of ternary meter and the greater flexibility of the iambic.
Jespersen (2010) has also pointed out an asymmetry between tro-
chaic and iambic lines. Whereas one may “invert” the first foot of
the latter by “substituting” a trochaic for an iambic foot, the former
seems not to tolerate the “substitution” of an iambic for a trochaic one
(reported in Halle and Keyser 1966: 199; cf. Beaver 1968: 314). Halle
and Keyser’s theory of “stress maxima” (see also 1971) offers an expla-
nation of why the first foot of an iambic line tolerates “inversion” more
than the subsequent feet. They, however, also try to apply this theory
to the alleged intolerance of inversion in the trochaic line. Their argu-
ment is supplemented by Beaver. The first and last syllables of a line (or
of a syntactic unit) cannot be a stress maximum: they do not have “two
adjacent syllables” to bear less “linguistically determined stress.” In such
iambic lines as Shakespeare’s “Proving his beauty by succession thine”
(Beaver 1968: 318) and Keats’s “Silent upon a peak in Darien” (Halle and
Keyser 1966: 189), the “inversion” is usually accepted as perfectly regu-
lar because, they say, it does not result in a stress maximum in a weak
(odd-numbered) position.
As for the trochaic, Halle and Keyser observe that Longfellow’s
“Life is but an empty dream” is rendered unmetrical if the initial trochee is

replaced by an iamb as in . . . “To live’s but an empty dream.” The reason for

3 21
this asymmetry between iambic and trochaic lines becomes clear once it is
realized that trochaic verses have stress maxima only at odd positions in the
verse and that an initial iambic foot locates a stress maximum on the second
(i.e., on the even) position in the line, in direct violation of the trochaic prin-
ciple. (1966:199)
This explanation, at last, refers to perceptual features of actual verses.

Beaver adds his own example. He changes Longfellow’s trochaic line
“Straight between them ran the pathway” to read “Between them straight
ran the pathway.” “We have created,” he says, “a stress maximum in an
even position, thus violating the rule of trochaic meter” (1968: 319). This
seems to account for the asymmetry of the two meters. Unfortunately,
however, the examples are asymmetrical in their own right, so that if
there is any asymmetry of the meters, the explanation must be sought
elsewhere.
The first foot of an iambic line can be “inverted” if the iambic pattern is
reasserted, or at least not violated, in the second foot. If the second foot is
“inverted” too, it creates a stress maximum in the third (weak) position, as
in the construct
(2) sílent viéw of a péak in Darién.

w s w s w s w s ws
There is no asymmetry in this respect between iambs and trochees. By the

way, such instances do occur in English poetry but are explained away by
Halle and Keyser as “the Italian influence of double trochee,” which, again,
“does not contain in itself the reason why it is so, and not otherwise.” In
terms of the present book, double trochee is legitimized not because we
hear it as rhythmical but because it is an influence.
What happened, then, to the Longfellow examples and their transcrip-
tions? In Longfellow’s trochaic lines the reader has a feeling that they “vio-
late normal prose accentual patterns” to some extent, in Chatman’s terms,
of artificiality. Consider the following two lines:
(3) Straight between them ran the pathway,

Never grew the grass upon it. (Beaver’s italics, to mark what he regards as stress
maxima)
This kind of artificiality is absent from Blake’s “The Tyger,” for instance,
“What the anvil, what the chain,” or any other line. This is only one of
the many sources of difference. In actual speech (and according to the
4
21
definition promoted by Halle and Keyser and by Beaver), conjunctions and

prepositions have no syllables that bear lexical stress and thus cannot con-
stitute a stress maximum. In the Longfellow lines, two strong positions are
occupied by the prepositions between and upon. Because of the compelling
nature of the trochaic meter (whatever its cause), Beaver marked them as
stress maxima (contrary to his definition), and this is, indeed, the way we
are disposed to recite these verses, although this is not linguistically justi-
fied. In Blake’s poem, Beaver could not assign a single stress maximum to
any preposition or conjunction.
In the line “Life is but an empty dream” (metrically identical with Blake’s
line: in both, the abstract versification pattern is swswsws) the second strong
position is occupied by but. If you treat it as an (unstressed) conjunction,
the rewritten line has not only an inverted first foot but two consecutive
feet where the stress pattern does not confirm the trochaic pattern, so that
in a “trochaic” line of seven syllables, only one unambiguous realization of
a trochaic foot is left (“émpty ”), and the example is not valid. If you treat
s w
it as an adverbial (in the sense of “no more than”), it bears lexical stress
and must be duly stressed in performance, neutralizing the alleged stress
maximum in the preceding w position (“To líve ′s bút an émpty dréam”). In
s w s w s w s
positions II and III, back to back, we get two stressed syllables, and the
line becomes similar to some perfectly legitimate trochaic lines, as we shall
see. At any rate, the line will be “metrical” under the stress maxima theory
itself. The question is thus not whether a trochaic line can bear “inversion”
of its first foot but, rather, whether but can bear full lexical stress. Hence
this example does not prove that trochaic lines do not tolerate inversion of
their first foot.
Beaver’s transcription “Between them straight ran the pathway” is not
much more felicitous. Between them does not exactly yield a stress maxi-
mum, as Beaver would like to indicate, nor does them straight realize exactly
a trochaic foot (if anything, it has an iambic rise). As a matter of fact, the
first foot of a trochaic line can be “inverted.” To be sure, it is not as frequent
or as easy as in an iambic line. It demands the overarticulation and over-
stressing of the deviant stressed syllable and very emphatic grouping of
the first two feet in order to make it acceptable. But the decisive condition
is, as in the iambic line, that the metric pattern should not be violated in
the second foot. If the first foot of a trochaic line is “inverted” and in the
second the trochaic is genuinely confirmed, there are, back to back, two
stressed syllables, and no question of a stress maximum in the second posi-
tion will arise.

5 21
Consider the following three examples from two poems notorious for
their regular rhythms:
(4a) In whát ds ístant

w
ds éeps or skies ("The Tyger")
s w w s
(4b) In whát fúrnace was thy bráin (ibid)

s w s w s w s
(4c) And éach séparate dýing émber ("The Raven")

s w s w s w s w
The reader may perform these lines rhythmically by grouping the first three
syllables tightly together. In the third position, the stresses of the trochaic
pattern and the linguistic stress pattern emphatically coincide. Thus, the
vigorous, dominant character of the trochee is not the result of its resis-
tance to the “inversion” of its first foot but, rather, its cause. Similarly, Halle
and Keyser are victims of their own mistaken assumption when they sug-
gest the phrase “On first looking into Chapman’s Homer” as an example
of an unmetrical utterance. Suppose we encounter it in an emphatically
trochaic context:
(5) Hów inténse are Kéats’s fámous sónnets

s w s w s w s s s w
"Áilsa Róck , " " When I have féars, " "To Fánny,"
s w s w s w s w s w
On Fírst Lóoking into Chápman’s Hómer"

s w s w s w s w s w
According to their definition, a trochaic line is one in which there are no

stress maxima in even positions. This is precisely the case here. According
to the assumptions in this chapter, the recognition of a line’s “metricality”
presupposes a correct hypothesis and/or proper performance. The “inver-
sion” of the first foot in the third line does not result in a stress maximum
in a weak position, as they say it would, because the second foot (“Looking”)
is regular and neutralizes the stress maximum.
Beaver puts forward an alternative explanation for the peculiar charac-
ter of trochaic meter:
Since most trochaic verse in English is in short lines, and since our impressions
of iambic verse are derived almost entirely from pentameter, it would seem
entirely possible that the issue has been falsely formulated—that the differ-
ences of rhythm encountered are attributed not to the type of foot but to the
length of line in which the foot characteristically appears. And it will be argued
621
below that the more regular beat of short-lined verse is accounted for by the
fact that a much higher percentage of positions available for stress maxima are
occupied than is the case in decasyllabic verse. (1968: 314)
One suddenly realizes that it is quite difficult to find poems written in

decasyllabic trochaic lines. Nevertheless, there are some. What is more,
Halpern, as quoted by Chatman, illustrates his observation quoted above
with a decasyllabic line:
And even where reversals are indicated the reader has strong desires to “wrench”
the stress or accent, as in Browning’s:
What there’s nothing in the/mōon note-/wōrthy? (Chatman 1965: 141)
Obviously, whatever the source of this impression in this particular

instance, it cannot be accounted for by the peculiarities of shorter-than-
decasyllabic lines.
Further, although I am not quite happy with the way Beaver subse-
quently documents his account of the rhythmic quality of short-lined
verse, I am inclined to acknowledge the rhythm he attributes to it (but
I prefer Aristotle’s view that the two rhythms are akin rather than that
the trochee’s effect is, in fact, to be attributed to the tetrameter’s—see
the epigraph above). As for the occupancy of “a much higher percentage
of positions available for stress maxima . . . than is the case in deca-
syllabic verse,” it seems to be a result rather than a cause. For simple
arithmetic reasons, the shorter the line, the less it tolerates deviating
feet. Whereas two deviating feet constitute more than 66.6 percent
of a trimeter line and 50 percent of a tetrameter line, they constitute
only 40 percent of a pentameter line. This, however, only means that
here, as in trochaic meter, poets can afford fewer deviations in absolute
numbers if they want to preserve an impression of rhythmicality. The
cooperative reader “has strong desires to ‘wrench’ the stress or accent”
whenever needed and possible. If this indicates anything, it suggests
that trochaic meter and short-lined verse, on independent grounds,
have something in common: a compelling quality. This may explain why
the trochaic is so scarce in decasyllabic lines and “prefers” to dwell in
shortened verse.
An alternative explanation would be that this compelling nature of the
tetrameter is due to its gestalt qualities. As I have argued in the foregoing,
a ten-syllable-long line cannot be divided into two symmetrical halves of
equal length and equal structure. If it is divided into 5 + 5, in the iambic

21
7
meter the first segment will begin and end with a weak position and the
second segment, with a strong position. In the trochaic meter the converse
will be the case. In the eight-syllable-long line, the caesura will divide it
into segments of equal length (4 + 4) and equal structure: In the iambic
meter both segments begin with a weak position and end with a strong
position; in the trochaic, vice versa. As opposed to the hexameter, in which
the second hemistich also exactly repeats the structure of the first one,
the iambic or trochaic tetrameter does not exceed the limits of short-term
memory: namely, the magical number 7 ± 2. Thus the tetrameter has a
stronger, more rigidly symmetrical shape than the pentameter or the hex-
ameter; that is, it is of a more compelling nature, which more forcefully
determines the character of its parts.
Trochaic meter and short-lined verse reinforce this compelling quality
in each other. But we still have no explanation why the trochaic should
be more compelling than the iambic. Curiously enough, the clue is quoted
in Chatman’s (1965: 26–27) book as well, but he does not connect it to
this specific problem. One would imagine that in an endless series of
equidistant tick-tacks no preference would be given to iambic or trochaic
rhythms; the only distinction would be whether the series began with an
upbeat or a downbeat. Experimental psychology, however, shows that
this is not so. H. Woodrow found in his tick-tack experiments, back in
the 1920s, that in a series of tick-tacks, “with equal temporal spacing, a
regularly recurring, relatively greater intensity exerts a group-beginning
effect, and a regularly recurring, relatively greater duration a group-ending
effect” (1951: 1233): “Intensity has a group-beginning effect: duration, a
group-ending effect: pitch, neither a group-ending nor a group-beginning
effect” (1911: 77).
Both Chatman and Meyer invoke Woodrow’s experiments; the latter
also indicates at some length their implications for a variety of metric feet:
When time intervals are equal, and every second sound is accented, the rhythm
will appear to be trochaic. If intervals are equal and every third sound is accented,
the rhythm will appear as a dactyl. Thus the trochee and dactyl may be grouped
together in the sense that both are primarily products of intensity differences
rather than durational differences.
Just the opposite is the case with iambic and anapestic rhythms. They are
basically products of durational differences. If we start with a trochaic rhythm
and gradually increase the interval after the louder sound, we arrive at an iambic
rhythm. Similarly, if we begin with a dactylic rhythm and gradually lengthen the
interval after the louder sound, the rhythm tends to become an anapest. Thus,
8
21
the greater the relative duration of one tone or beat of a group, the greater the
tendency for it to complete the group, while the greater the relative intensity
of a beat, the greater the tendency for it to begin the rhythmic group. In other
words, durational differences tend to result in “end-accented rhythms,” and
intensity differences tend to result in “beginning-accented rhythms.” (Meyer
1956: 106–107)
Woodrow also found that pitch has neither a group-beginning nor a

group-ending effect (Meyer 1956: 106–107).
Since the 1980s and 1990s, Woodrow’s work on nonlinguistic tick-tacks
has had an enormous impact on linguistic research, but from very dif-
ferent perspectives. A few are mentioned below. The generative linguist
Bruce Hayes (1985, 1995) applied Woodrow’s extralinguistic principle of
rhythmic grouping to iambic and trochaic feet in a wider linguistic per-
spective and showed that (a) elements contrasting in intensity naturally
form groupings with initial prominence and (b) elements contrasting in
duration naturally form groupings with final prominence. He was more
interested in generalizing to a cross-linguistic theory of stress assignment
than in versification, and his work is motivated by in-depth analyses of
stress patterns in a large number of languages (for a critique of Hayes, see
Revithiadou 2004).
In his doctoral dissertation, Curt Rice replicated Woodrow’s experi-
ments: “The technological resources for conducting this research are
dramatically more sophisticated than those which Woodrow had avail-
able” (1992: 198). At variance with Woodrow, Rice showed that varia-
tions in pitch do lead to a significant shift toward iambic groupings, a
result that Revithiadou (2004: 38) claims to disconfirm Hayes’s Iambic/
Trochaic Law. But it lends additional support to the view expressed here.
The reason for this difference appears to be that we have different pur-
poses: Hayes applies the extralinguistic principles of rhythmic grouping
to phonological stress rules, while I apply them to versification patterns
determined independently from linguistic patterns, as suggested by
Halle and Keyser.
Now, coming back to verbal rhythms, spoken language consists of sylla-
bles of varying duration. Schramm (1935), Fry (1958), and Chatman (1965)
demonstrated that the acoustic cue to linguistic stress is a complex of pitch,
duration, and loudness, in that order of decreasing effectiveness. Usually,
at least two of these three components are present. If pitch differences are
irrelevant to grouping direction and duration differences are more effec-
tive in stress perception than amplitude differences, end-accented meters
should be more natural in poetry in a variety of languages. If variations

9 21
in pitch also lead to a significant shift toward iambic groupings, it should

reinforce this effect. Significantly, even in Hungarian poetry, where stress
is invariably on the first syllable, the iambic meter is far more natural and
widespread than the trochaic.
In “stress-timed” languages, pitch obtrusion seems to be predomi-
nant; it is usually reinforced by duration or loudness or both. Thus, in an
“end-accented” meter such as the iambic, there is room for a complex inter-
play among pitch, loudness, and duration. In the trochaic, on the contrary,
the inherent properties of “beginning-accented” meters demand the sup-
pression of duration differences and the bestowal of priority upon loudness
in producing stress. Characteristically enough, Chatman reports: “Brown
also discovered an interesting difference between iambic and trochaic non-
sense lines; ictus in iambic meter was from 2.1 to 2.9 times longer than
non-ictus, whereas in trochaic meter, the ratio was profoundly differ-
ent, the ictus ranging only .46 to 1.04 times the length of the non-ictus”
(1965: 79). Consequently, if one wishes to preserve the trochaic character
of a verse line, one must overemphasize the loudness cue for stress in any
instance of an unstressed syllable in a strong position. That is why trochaic
lines like those in excerpt (3) sound relatively unnatural and why “lighter
syllables seem more readily ictic in trochaic than in iambic feet” (Chatman
1965: 140–141).
Thus, when performing a trochaic poem, the reader tends to suppress
quantity, which in English is merely an optional diversifying element. In
this manner, the contrast between stress and no stress is cued by ampli-
tude differences, and the contrast between longer and shorter syllables is
suppressed as much as possible. If the verbal material permits, the trochaic
will sound simple, single-minded, straightforward, vigorous, optimistic,
encouraging, and the like. Consider:
(6) When the wind is in the east,

Then the fishes do bite least;
When the wind is in the west,
Then the fishes bite the best.
Bite in line 4 falls in a strong (odd) position, and it is duly stressed. In line 2
it falls in a weak (even) position and should create metrical complexity. But
the reader prefers to suppress it and, rather, stress the preceding auxiliary
verb (do), which is unemphatic and is introduced mainly for the meter’s
sake. This tendency toward regularity can be reinforced by the repetition
of some notoriously trochaic word or sound pattern in the first line, for
example, “Handy spandy Jack-a-Dandy,” “Tackle, tackle, Mother Goose,” or
0
2
“Tyger, tyger, burning bright.” Such a sound pattern will tend to perpetuate
its trochaic cadence in the subsequent lines.
A glance at Henry Carey’s parody on Ambrose Philips’s overly smooth
style (in The Oxford Book of Nursery Rhyme) is revealing:
(7) Nauty Pauty Jack-a-Dandy

Stole a piece of Sugar-Candy
From the Grocer’s Shoppy-shop,
And away did Hoppy-hop. (The Oxford Book of Nursery Rhyme)
Compare with this the genuine, less smooth nursery rhyme:
(8) Handy spandy, Jack-a-Dandy,

Loves plum cake and sugar candy;
He bought some at the grocer’s shop,
And out he came, hop, hop, hop, hop. (The Oxford Book of Nursery Rhyme)
After having established in the first line a conspicuous trochaic lilt,

we feel that the trochaic rhythm is more disturbed by plum in a weak
position than by the “extrametrical” syllables (He, And) at the begin-
ning of lines 3 and 4. Plum is not only a stressed syllable (“neutraliz-
ing” the contrast between stressed and unstressed syllables), but it also
seems to be too long for squeezing between two downbeats. It upsets
the time scheme of the performance (it should be remembered that tim-
ing is a part of the performance pattern and not the poem itself). On
the other hand, the poem would not bear a “weighty” performance that
could anticipate a stressed long syllable in a weak position. The word
begins with a “jam” of four consonants (vzpl) and ends with a jam of
two (mk). Similarly, the hops in weak positions “insist” on their equal
length and stress with the adjacent hops and so “resist” being turned into
upbeats. The parody avoids such “inconveniences” by seeing to it that all
the unstressed syllables allow isochronic and isoaccentual performance,
with no undue slowing down of tempo. Consider one more point regard-
ing the extrametrical words at the beginning of lines 3 and 4 (He and
And): they ought to turn the trochaic verse into iambic instantly, but if
you suppress duration differences as much as possible, the trochaic char-
acter will be largely preserved.
This is also why Halle and Keyser and Beaver may feel that the first
foot of a trochaic line cannot be “inverted” like that of an iambic line. The
trochaic meter resists complexity, and “the reader has strong desires to
‘wrench’ the stress or accent” when the foot is inverted. Performance cannot

21
“compensate” by bringing in, for example, an interplay between the inten-

sity factor and the time factor. When, finally, the first foot of a trochaic line
is inverted, it demands stronger overstressing, overarticulation, and addi-
tional grouping than an iambic line, as well as the suppression of duration dif-
ferences. If this occurs in an intense and complex poem, the reader feels that
the rhythmic complexity is sufficiently motivated, as is the case with
(9) In whát dístant déeps or skies

s w s w s w s
Or consider the first four lines of Browning’s “Soliloquy of the Spanish

Cloister.” They ingeniously arouse two incompatible, equally strong
desires: “to wrench” the prose stress pattern in favor of the trochee and, at
the same time, to preserve it:
(10) GR-R-R  there gó, my héart ′ s abhórrence!

s w s w s w s w
Wáter your dámned flówer-póts, dó!

s w s w s w s
If háte kílled mén, Bróther Láwrence,

s w s w s w s w
Gód ′ s blóod, would not míne kíll yóu!

s w s w s w s
The pace for both desires is set in the first line. The meter of this line is as
regular as one could expect in a trochaic (with, probably, the need to sup-
press a possible stress on there, used interjectionally). The metrical pattern
is reinforced by the stress pattern coinciding with it, which, in turn, is rein-
forced by a recurring sound pattern: not the simple kind such as “Tackle,
tackle” or “Tyger, tyger” but, rather, a very sophisticated version of the type
“burning bright.” The contrast between the s and w position is heightened
by a sophisticated string of alliterations, emphasizing the s positions, cul-
minating in horr-. It repeats the rs of the first s position, the o of the second,
and the h of the third. Go (preceded by an r) repeats the sounds of GR-R-R,
which is an important keyword. As a nonsense word, it fulfills the function
of nonsense words at the beginning of nursery rhymes: to draw attention
to the sound pattern. As an emotive or onomatopoeic word, it arouses a
strong desire to preserve the rhythms of colloquial language. The first word
of the second line (Water) confirms the trochaic pattern. The next stress,
however, is displaced to a w position. The strong desire “to wrench” the
prose stress here is vigorously counteracted by the colloquial character of
damned. The only option left for readers is to accommodate both patterns
2
in an exceptionally tight additional grouping, while both seek to establish

themselves as strongly as possible in their perception.2 After all this, the
reader is ready for any grouping and overarticulation required to perform
the deviating feet “If háte kílled mén .” The rest of the line reasserts the
s w s w
trochaic. If emphatic stress is assigned to the pronouns mine and you,

the fourth line becomes “metrical” and may even corroborate the metri-
cal set, despite the six heavily stressed syllables in a line of seven (how
acceptable this is here, as contrasted with “Loves plum cake” in the nurs-
ery rhyme!).
A word should be said about the meter of Shelley’s “To a Skylark.” This
poem for some reason has puzzled some of our leading prosodists. The
problem is whether its meter is iambic or trochaic. To tell the truth, this
question appears to me similar to the question “Is the zebra white with
black stripes or black with white stripes?” Or to put it more mildly, “Is ‘The
Rime of the Ancient Mariner’ tetrametric or trimetric verse?” Unless our
purpose is bookkeeping or classification, none of these questions should
trouble us. The problem arises not only because the fifth line of each stanza
is iambic whereas the preceding ones are trochaic but also because the sec-
ond and fourth lines of each stanza begin and end in a strong position.
Therefore, they are said to be “ambiguous”: one cannot know whether they
are iambic with the first position unoccupied or trochaic with the last posi-
tion unoccupied. In light of Woodrow’s experiments with extralinguistic
tick-tacks and Brown’s experiments with nonsense lines, if we suppress
duration differences, the lines will be perceived as trochaic, but if we care-
fully preserve them, they will be perceived as iambic. In view of the gestalt
rule of Good Continuation,3 there are good reasons to suppose that per-
formers will tend to preserve the trochaic character of the first four lines of
2. In Tsur 1972 I pointed out that one of Browning’s favorite prosodic devices is to
insert a marked syntactic break before the last syllable of the line. This arouses a strong
desire to complete the verse. So, the last word is highly required and bestows a more
than usually closed shape on the line. This is precisely what happens here at the end of
the second line, with do, gathering momentum for the reader’s (frustrated) desire to
maintain the set of regularly alternating s and w positions. Another keyword is abhor-
rence. It not only contains, as we have seen, the phonetically crucial syllable -horr-; it
also forms an unusually witty “feminine” rhyme with Lawrence, having five(!) sounds
in common, with virtually no common ingredients of meaning, thus emphatically rein-
forcing the quality generated by the trochee (the “antigrammatical” rhyme do–you has
a similar effect).
3. According to the law of Good Continuation, the arrangement of the stimulus pat-
tern that makes the fewest changes or interruptions in it will be preferred as “good
gestalt.”

3
2
each stanza. So, we must decide which is more important: the name of the
metric structure or its perceived character.
Beaver writes: “As a matter of fact, many poets have capitalized on this
ambiguity. In ‘To a Skylark,’ Shelley chooses a stanza form which maintains
the ambiguity through the first four lines of each stanza, resolving it in
the iambic hexameter fifth line” (1968: 317). Let us have a look at the first
stanza of the poem:
(11) Hail to thee, blithe Spirit!

Bird thou never wert,
That from Heaven, or near it,
Pourest thy full heart
In profuse strains of unpremeditated art.
In fact, only two of the lines (2, 4) can be said to be “ambiguous.” Lines 1 and
3 are unambiguously trochaic, and line 5 is unambiguously iambic. Putting
aside the problem of the fifth line for the time being, one may observe that
the reader of French, German, Russian, Hungarian, and Hebrew poetry
(and possibly that of many other languages) is familiar with the alternation
of “feminine” and “masculine” rhymes that varies rhythm and governs line
grouping without basically changing the meter (see Tsur 2013). In English,
it is more difficult to do this because of the scarcity of “feminine” rhymes.
Everything that happens here is thus an extension of the grouping
principle of performance when irregularities occur. The dissimilarity of
the odd-numbered and even-numbered lines upsets their balance to some
extent, and they demand some further grouping. By grouping line 1 with 2,
and line 3 with 4, the four-line unit is divided into two corresponding
halves, reinforcing the grouping effect of the rhyme scheme. (This is, pre-
cisely, with the necessary changes, what happens in the ballad stanza too.
See discussion of the ballad “Edward,” c hapter 4.) The asymmetry of the
tetrameter and trimeter lines requires grouping them into two symmetri-
cal groups, thus enhancing the unity and simplicity of the stanza.
I am inclined, then, to disagree with Beaver’s more general “solution” to
poems such as Tennyson’s “Lockley Hall”: “An obvious solution to this prob-
lem is to view such poems or portions of poems as metrically ambiguous
in their surface structure, and postulate that they are, in their deep struc-
ture, either iambic with initial position always unoccupied, or trochaic,
with final position always unoccupied” (1968: 317). The problem will not
arise at all if we realize that the obligatory caesura divides the lines of this
poem, or “A Toccata of Galuppi’s,” so that their metric structure is exactly
the same as that of the Browning poem discussed above. The typographical
4
2
rearrangement of these lines only lends support to the grouping arising

from the asymmetry of feminine and masculine endings:
(12a) Comrades, leave me here a little, while as yet ’tis early morn:
Leave me here, and when you want me, sound upon the bugle-horn.
(12b) Comrades, leave me here a little,
While as yet ’tis early morn:
Leave me here, and when you want me,
Sound upon the bugle-horn.
(13a) Oh Galuppi, Baldassaro, this is very sad to find!

I can hardly misconceive you; it would prove me deaf and blind;
But although I take your meaning, ’tis with such a heavy mind!
(13b) Oh Galuppi, Baldassaro,
This is very sad to find!
I can hardly misconceive you;
It would prove me deaf and blind . . .
But although I take your meaning,
’Tis with such a heavy mind!
When preparing this chapter, I felt that I had to modify in one impor-
tant respect earlier positions I took in Tsur 1977b. While the alternation
of feminine and masculine endings enhances grouping and articulateness
in both (13a) and (13b), we do experience the two excerpts differently. In
vocal or subvocal performance, pause and falling intonation are the most
conspicuous cues for discontinuity. In (13b), these cues are equally arrest-
ing at the end of each line. As a result, one typically tends to experience fast
alternations of short perceptual units. In (13a), when some of these dis-
continuities occur in midline at the caesura, the cues are toned down, so as
to subordinate the first hemistich to the whole line. As a result, a broader,
slower rhythm is obtained, suggesting a more serious or dignified attitude.
The “unexpected” appearance of the iambic hexameter at the end of
Shelley’s stanza (11) resolves, in fact, nothing. Rather, it complicates things
further. It does not reveal the “deep structure” of the preceding lines; it devi-
ates from them. According to Barbara Herrnstein Smith (1968), a closural
effect can be achieved by deviating from a previously established structural
principle. This would imply a hierarchy of grouping principles operative
in the stanza: (1) the line, (2) two pairs of lines parallel to each other, and
(3) a fifth line deviating in length and foot, emphatically sealing the stanza.
Whether it actually does so depends on whether the reader realizes the hier-
archy of grouping principles or loses sight of this third, widest one.

5
2
Wimsatt and Beardsley seem to be after a far more deeply seated quality
when they remark on the first line of this Shelley poem:
The unquestionably iambic movement following the very strong first syllable
might, if we were desperate, be accounted for by saying that the word “Hail”
breaks into two syllables, “Hay-ul,” with a resultant needed extra syllable and
the familiar opening pattern of iambic inversion. But a much more energetic and
irrefutable assertion of the iamb appears to be in the progressive rise or stress
increase of the three syllables “thée, blit̋he, spír̋it.” (1959: 593–594)
This argument seems invalid to me. Even if we take for granted, for the sake
of the present argument, four degrees of stress in “to thée, blit̋he spír̋it,”
the only thing we need to do to refute this conception of “rising” rhythm
is to say that the energetic iambic rise is followed by a sudden trochaic
fall and that in every trochaic line the final falling movement is preceded
by a rising movement. In fact, the term iambic movement is used here in a
figurative sense: it does not refer to the recurring metrical pattern of feet
but, rather, to the overall rising pattern of performance. In a note, Wimsatt
and Beardsley add: “But many such lines, like the one from Shelley’s
‘Skylark’ . . . , can be shown in one way or another to be in fact iambic. The
shape of the phrases is likely to have much to do with it” (1959: 594).
The line has been suggested, but not shown, to be iambic. Formally, the
line is nothing but trochaic. But what is it that impresses the critics as so
“untrochaic” in this line? It is its complexity. Blithe consists of one stressed
syllable with a diphthong as its nucleus. Neither its stress nor its length
can be sufficiently reduced and fit into a weak position so as to equal to in
the preceding weak position. Thee in the preceding strong position bears no
lexical stress. The regular contrast of intensity, as well as the equal duration
of feet, is upset; this makes the line more like iambic. The stressed syllable
in a weak position can be accommodated only by additional grouping; this
was, indeed, done by Wimsatt and Beardsley, and this makes the line too
complex for an ordinary trochee. But now that we know what is so “untro-
chaic” about the line, we need worry no more, and there seems to be no
reason why we should not continue calling it “trochaic.”
IAMBIC TENSION
The iambic is usually thought to be the most natural meter, the nearest
to ordinary speech rhythms. A striking feature of this measure is that in
languages as different from English as Hebrew and Hungarian, poets use it
6
2
in a manner very similar to English blank verse. Note that this is particu-
larly noteworthy in Hungarian, where the end-accented nature of the iamb
generated by duration differences overrides the fact that in Hungarian the
first syllable of a word is invariably stressed. This seems to hold true even
outside the syllabotonic system, as Aristotle and Horace may bear witness,
for whom iambic meant a “quantitative” measure, alternating longer and
shorter syllables:
The iambic is the characteristic rhythm of people as they talk. . . . The tro-
chaic rhythm, again, is too much akin to the comic dance, as may be seen in
tetrameter verse, for the rhythm of tetrameters is light and tripping. (Aristotle
1932: 3.8, 1408b)
His rage armed Archilochus with his iambic: comedy and tragedy have
adopted it, as being natural for dialogue, able to drown out the noise of the audi-
ence and suited to action. (Horace 1951: 117)
It is difficult to know now exactly what Horace meant by the last two
phrases (“able to drown out the noise of the audience and suited to action”),
but the fact that the iambic was found to be “natural for dialogue” remains.
As we have seen, Aristotle agrees: “The iambic is the characteristic rhythm
of people as they talk.” Our explanation of its nature will have to take this
fact into account.
At the end of his “ ‘Prose Rhythm’ and Metre,” Roger Fowler comes to
the following conclusion: “This, paradoxically, may help to explain why the
iambic measure is felt to be suited to English: not because its pattern cor-
responds to the prose rhythms of language, for it does not; but because
it necessitates a constant syncopation of prose rhythm against its own
rhythm, inviting poets to be metrically complex, not to jog along with
simple regularity” (1966: 99). Surprising as this may be, the great vari-
ety of languages to which the iambic measure seems to be “suited” (even
Hungarian!) supports this conclusion. I would, however, prefer tolerates to
necessitates. Verbal necessity alone cannot account for syncopation; it can-
not explain why Pope should resort to syncopation less frequently than
Milton (as Fowler himself asserts) or why Shakespeare, in his earlier work,
should resort to it less frequently than in his later work. It seems, rather,
that syncopation is related to wider issues and that it takes deliberate dar-
ing to abandon the security of established strong shapes. Far from being
a necessity, syncopation is a powerful achievement; it is a delicate balance
between prose rhythm and meter. As Fowler states, “There can be a situa-
tion (Hopkins’s ‘counterpoint’) where the prose rhythm makes itself felt as

7
2
something playing against the ostensible meter. But extreme lack of fit of
words with feet results in the total assertion of prose rhythm” (1966: 94);
and “in Donne, who, according to Chatman, has a much higher propor-
tion of reversed feet [than Pope], the tension is less because the metrical
pattern is obscured” (1966: 93). In trying to achieve ambiguous, complex
rhythms, there is always a possibility that the meter will be lost sight of,
which is a risk some poets seem to be unwilling to run.
I submit, then, the following two explanations for Fowler’s generaliza-
tion about the iambic. First, as we have seen, end-accented meters allow
for greater flexibility in manipulating the time factor in performance than
beginning-accented meters. Second, bisyllabic feet have “stronger gestalt”
than trisyllabic feet; they can be more effectively maintained when con-
flicting with prose rhythm. When performing trisyllabic feet, the reader
is inclined to entirely suppress prose rhythm when it conflicts with prose
rhythm; otherwise the meter of the poem is in danger of disintegration.
(This may probably account for the fact that in “quantitative” measures as
well, the iambic was found to be most “natural for dialogue.”)
TERNARY METERS
We have seen that critics are inclined to link the general character of ter-
nary meters with that of the trochaic. We are now in a position to point
out the source of their common character. They resemble each other in that
they are all unlike the iambic. We have seen that the iambic is “natural for
dialogue,” not in that its pattern is similar to the stress patterns of ordinary
speech but in that it allows the greatest tension between prose rhythm and
meter. It has a stronger gestalt than the ternary meters, and therefore, the
feel of it can be preserved even when the lexical stress pattern strongly
deviates from it. In order to preserve the integrity of the larger feet of three
syllables, one has to suppress prose rhythm whenever it deviates from a
ternary foot. This is the common factor to ternary meters and the trochaic.
We have seen that in the trochaic too, although for different reasons, in
order to preserve its peculiar character, one feels inclined to suppress in
prose rhythm whatever deviates from perfect regularity. The greater regu-
larity of ternary meters and the trochaic renders them of a more compel-
ling, though less complex nature.
Finally, I wish to clarify the negative relationship between the iambic
and the ternary meters by briefly commenting upon prevailing practice
in some languages other than English. Unlike in English, where ternary
8
2
meters are very rare, in modem Hebrew poetry syllabotonic amphibrach

and anapest are common. The end-stressed character of the language ren-
ders the dactylic so unnatural that it is virtually nonexistent (except for
translations of classical poetry, where it creates almost insurmountable
difficulties). The correspondence of stressed syllables with strong posi-
tions is usually observed in the anapestic and amphibrachic measures, and
they are frequently used in highly convergent poems, with the psycho-
logical atmosphere of certainty and patent purpose (such as journalistic
poetry4). In order to prevent monotony and “saturation” in their verse,
the poets vary their rhythms by occasionally leaving a weak position unoc-
cupied rather than by displacing a stress from a strong to a weak position;
the same poets in the iambic do resort to stress displacement (though less
than English poets).
Hungarian poets have free access to both “quantitative” and syllabotonic
meters. In quantitative measures they usually resort to ternary feet, result-
ing in rather unnatural cadences. Curiously enough, the same poets write
iambic verse in syllabotonic meter involving the “syncopation” of metric
stress and prose rhythm. The reason seems to be as follows. Quantitative
meter, a relatively recent import to Hungarian, is at variance with the natu-
ral rhythms of the Hungarian language. As ternary measures subordinate
stress pattern to meter, this means that quantitative and ternary measures
are alike, in that they both demand the suppression of the natural stress
pattern of the Hungarian language and thus, for negative reasons, consti-
tute quite natural bedfellows.
Another observation on the negative relationship between the iamb and
ternary meters concerns the problem of translation from French into lan-
guages dominated by syllabotonic meter. The French alexandrine is usually
translated into the iambic meter in English, Hebrew, Hungarian, and some
other languages. When one encounters a French alexandrine translated
into syllabotonic ternary meters there is an uneasy feeling that something
has gone wrong. The case study that triggered a solution to this problem,
comparing Baudelaire’s “Correspondances” to its Hebrew translations, is
reproduced in c hapter 7, and the reader is referred to it. I will repeat here
only the upshot of that discussion. The alexandrine is similar to the iambic
in one important respect. It tolerates greater irregularity of the stress pat-
tern than the trochaic or ternary meters such as the anapest, the dactyl,
and the amphibrach would tolerate.
4. During World War II and several decades afterward, one of the greatest Hebrew
modernist poets, Nathan Alterman, published a weekly column commenting on cur-
rent events. He used the prosody of his serious poetry.

9
2
CONCLUSION
Poetic meters can be regarded as the archetypal neutral conventions. Poets

use them freely, without any awareness of the underlying psychological
processes. However, already Aristotle discerned certain typical effects asso-
ciated with the various meters. Prevalent practice treats those effects as
inherent in the convention, and there’s an end on it. In the twentieth cen-
tury there were quite a few attempts to account for the difference between
the effects of the iambic and the effects of the trochaic. Most of them
invoke external evidence that conflicts with Coleridge’s principle: “Nothing
can permanently please which does not contain in itself the reason why it
is so, and not otherwise” (1951: 267). Halle and Keyser, by contrast, offer
an explanation according to which metric structures contain in themselves
the reasons for their different effects; but I have shown that their argu-
ment is erroneous, and their solution is problematic even with reference
to their own example. This chapter has proposed a solution by invoking
three cognitive principles that govern the process. First, end-accented
meters are generated by duration differences (and perhaps also pitch differ-
ences), and beginning-accented meters, by amplitude differences. Second,
linguistic stress perception is governed by a mixture of pitch, duration, and
loudness, in this decreasing order of effectiveness. Third, the gestalt rules
of perception affect the perceived qualities of metric structures. Trochaic
meter is perceived as less natural and less flexible than the iambic because
in it stress is signaled by the least effective acoustic cue.
This chapter has distinguished three patterns in poetic rhythm: versifi-
cation (metrical) pattern, linguistic stress pattern, and pattern of perfor-
mance. When the linguistic stress pattern and metrical pattern converge,
they yield strong prosodic gestalts displaying a psychological atmosphere
of certainty and purpose. When they diverge, the verse is in danger of
falling into chaos. In ternary meters, performers are inclined to subordi-
nate the prose rhythm to the regular metrical beat. The iambic foot, with
its stronger gestalt, seems to tolerate greater deviations and complexity.
In extreme cases of deviation, the performer needs to accommodate the
divergent patterns in a strong gestalt of additional grouping to prevent
chaos. Some underlying strong gestalt, whether in the pattern of meter
or in performance, makes rhythmicality possible when divergent elements
are present.
0
32
231
CHAPTER 10
Poetic Language and

the Psychopathology of Everyday Life
T he psychological mechanisms Freud described in Psychopathology of

Everyday Life (1953) have fascinated researchers and laypeople alike,
not only after Freud but long before him as well. The reason seems to be
that they yield a direct glimpse into the hidden cognitive and psychody-
namic mechanisms underlying much of our mental activities. More impor-
tant for my purpose, they expose the cognitive mechanisms underlying
some of the most common poetic conventions.
In what follows, I will consider three of the Freudian psychopatholo-
gies: the first two are the tip-of-the-tongue (tot) phenomenon, when one
has a word on the tip of one’s tongue but is unable to recall it, and the
slip-of-the-tongue phenomenon, when one intends to say one word but, to
one’s own surprise, another, unintended one slips out. The third psycho-
pathology concerns a case study that, unlike the other cases, is ab ovo of
literary interest: a poem by Keats is misquoted in order to suppress painful
memories. By invoking Freudian analysis, phonetic and semantic analysis
of words, and computer simulation of language use in artificial intelligence,
I demonstrate, at an exceptionally fine-grained level, how the same kinds
of organized violence against language underlie poetic language, on the
one hand, and, on the other hand, misquotation, slips of the tongue, and
the tot phenomenon. In other words, by smashing the linguistic fossils
I expose the underlying life activities responsible for poetic language and
the psychopathologies of everyday life alike.
Freud’s older contemporary, William James, commented on the tot
phenomenon as follows: “And the gap of one word does not feel like the gap
2
3
of another, all empty of content as both might seem necessarily to be when

described as gaps. . . . But the feeling of an absence is toto coelo other than
the absence of a feeling. . . . The rhythm of a lost word may be there without
a sound to clothe it” (quoted in Ehrenzweig 1965: 10). Unlike Freud, James
does not discuss repressed unconscious wishes but describes the phenom-
enon as it appears to consciousness.
Otto Jespersen (1960: 19) wrote that every speech activity involves
three aspects: expression, suppression, and impression. Expression is
what speakers actually say. Suppression is what they might have said
but did not say. Impression is what affects the hearer: meaning or the
perceived quality emerging from the interaction of expression and sup-
pression. According to the semantic definition of literature, “a literary
work is a discourse in which an important part of the meaning is implicit”
(Beardsley 1958: 126). Here we simply must repeat William James’s “And
the gap of one word does not feel like the gap of another.” In certain cir-
cumstances, such suppressed meanings may give rise to the perception of
vague, elusive, but quite intense perceptual qualities. In extreme cases,
some Romantic and Symbolist poems are characterized by such vague,
elusive, but intense perceptual qualities. But where stable visual shapes
are involved, the suppressed meanings are “grown together” with the
properties of the stable object qualified and affect the emotional atti-
tudes toward it.
Consider such a hackneyed metaphor as “The roses of her cheeks”
and a more explicit paraphrase of it, “The pinkness, softness, and beauty
of her cheeks.” The perceived difference between them brings out the
impression generated by the interaction of expression and suppression
in the metaphor. In other words, in the metaphor the properties “pink-
ness, softness, and beauty” are suppressed but active in the background.
The tot experiment below may reveal the cognitive mechanism underly-
ing this process.
Psychoanalysts are usually surprised to hear that slip-of-the-tongue
and tip-of-the-tongue phenomena have both a depth-psychological
and a linguistic facet. For years I have had a strong intuition regarding
the close relationship between these psychopathological processes and
poetic language. If I am right, it may be additional evidence that cer-
tain poetic conventions came to take forms that have a good fit to the
constraints and natural capacities of the human brain. Over the years
I have returned time and again to explore this issue; in this chapter
I bring together some of these explorations and also present some new
insights.

32
“FREUDIAN” SLIPS AND SPEECH ERRORS
As a concise introduction to the issue, I present two stories in which I was

personally involved, as paradigmatic instances. A computer student used
to provide computer services to professors of demography at the Hebrew
University, among them Professor Roberto Bacchi. When the student fell
critically ill, she could not do her work and had terrible pangs of conscience.
On one occasion she intended to say something about Professor Bacchi but
instead she inadvertently said “Pagis.” At that time, the Israeli intellec-
tual community as a whole was in shock following the untimely death of
the great Hebrew poet and scholar of medieval Hebrew poetry Dan Pagis.
Knowing that I had some professional interest in such slips of the tongue,
she asked me to interpret her slip. Though my professional interest was only
in the linguistic aspects of the phenomenon, the case seemed to be so obvi-
ous that I ventured a more comprehensive interpretation. I said that there
was a linguistic and a depth-psychological side to the issue. Phonetically,
[b]‌is a voiced [p], whereas [g] is a voiced [k]. When mixing up the names,
she preserved the phonetic structure strictly and only transferred the fea-
ture [+ voiced] from the bilabial [b] to the velar [k] consonant, substitut-
ing Pagis for Bacchi. But by the same token, the mix-up of names expressed
a forbidden (unconscious?) wish. Had Professor Pagis been her employer
rather than Professor Bacchi, she would have been released from her
commitment.
A similar case can be made for the following “accident.” In a gradu-
ate seminar in Hebrew literature, a promising young poet intended to
say “mǝvakrim” (critics) but inadvertently said “mǝfagrim” (mentally
retarded). These two words contain the bilabial fricatives [f]‌and [v] and
the velar stops [k] and [g]. Phonetically, the speaker preserved this struc-
ture but inverted the values of the feature [± voiced] of the [v] and [k] in
mǝvakrim, yielding mǝfagrim. The “Freudian” message is obvious.
Briefly, my thesis is as follows. We have an enormous reservoir of words
in our long-term memory, and we retrieve these words with amazing ease.
We are usually aware of the retrieval of whole words, but there are indica-
tions that information is stored in a different way. There are several models
that purport to account for this process. I will opt for a somewhat antiquated
one that nevertheless has the advantage of providing a reasonable account
of the phenomena discussed here and is also supported by the empirical
literature. According to this model (see Brown 1970 below), semantic and
phonological components are stored separately, and the semantic fea-
tures of the meaning component as well as the distinctive features of the
P o e t i c L a n g ua g e a n d t h e P s yc h opat h ol o g y of E v e r y day L i f e [ 233 ]

4
3
2
phonological component have considerable psychological reality. As we

shall see, there is evidence suggesting that even the order of phonemes is
stored separately. Over the course of the retrieval process, these compo-
nents and features come together, and we experience the result as a uni-
tary word. Normally, the retrieval process is fast and smooth. Occasionally,
however, something goes wrong. We may have a word on the tip of our
tongue but are unable to recall it. Or we may intend to say one word, but to
our surprise, another unintended one slips out. This word-retrieval mecha-
nism thus may be disrupted for a variety of reasons: unconscious forbid-
den or conflicting wishes, inexperience with unfamiliar words, but also
the exigencies of poetic language. In other words, the same disruptions
of the word-retrieval process may be exploited by the unconscious for psy-
chopathological purposes and by poetic language for aesthetic purposes.
One (but not the only) reason these disruptions are so fascinating is that
through them the retrieval process can be observed in slow motion, which
otherwise would be inaccessible to observation.
A similar proposal to the suggestion articulated in the foregoing dis-
cussion was set out by the great psychoanalytic theorist of music Pinchas
Noy. In a very similar proposal to mine, Noy (2009) suggests that the same
“Him” who causes unintended utterances to slip from our tongue also
invents the melodies that “come to us” (from the gods or the muses, as it
were). Noy treats this as a representative case of all creative insight. To be
sure, owing to our different vantage points, we emphasize different aspects
of the process. Noy emphasizes that the source of creative insights evades
all conscious control and attributes them to the (metaphoric!) little green
man in us, who also causes us to have slips of the tongue. I, by contrast,
assume, following Köhler, that in the solution of a problem we suddenly
become aware of new relations: “The sudden correct organization of the
situation, and with it the solution, tend to occur at moments of extreme
mental passivity” (Köhler 1972: 160), in the bath, in bed, or on the bus.
We both agree on the suddenness element in creative insight. But I treat it
as a phenomenological quality of the emergence of the new organization,
whereas Noy regards it as another way of saying that the source of creative
insight evades conscious control.
Let us go back to my first example. The speaker, inexplicably to herself,
substituted the name of “Professor Pagis” for “Professor Bacchi.” A Freudian
psychoanalyst would point out that the shift of referents of these names
expresses an unconscious forbidden wish to get rid of an employer toward
whom she had an unfulfilled obligation. A phonologist would point out
that the two names are essentially the same, except that the [± voiced]
feature of the oral stops has been inverted; probably he or she would also

5
3
2
regard this as evidence for the psychological reality of distinctive features.

A literary theorist might say that the shift of meaning and the transfer of
the [+ voiced] feature from the bilabial to the velar stop direct attention
away from extralinguistic reality to both the sound and the meaning of
the words. The subtle changes that occur in this specific context in both
the semantic and the phonetic components of the words may serve as the
basis for certain literary effects (see below). Finally, the psychoanalyst and
the literary theorist are likely to make different recommendations as to
what to look for in the utterance and how to look at it. The former will
mainly dwell on what the unconscious mind tells the mind; the latter will
be mainly interested in the subtle changes that occur in the semantic and
phonological components of the words in that context.
Suppose we encounter the names “Pagis” and “Bacchi” in the same verse
line. We would obtain sound patterns in which some distinctive features
are repeated and some are contrasted. In his essay “Musicality in Verse,”
Kenneth Burke described these patterns in Coleridge’s work as “passages
that seemed to have a marked consistency of texture; yet this effect was
not got by some obvious identity of sound, as in alliteration” (1957: 369).
He calls such sound patterns “colliteration.” Burke (1957: 297) starts with
a model of consonants taken from traditional phonology, as shown in
Figure 10.1. In Structuralist terms, the two sets are contrasted at the locus
of articulation: namely, as labial versus alveolar consonants. Within these
sets, the speech sounds are contrasted in voicing and in manner of articu-
lation: they are nasals, oral stops (plosives), and fricatives. This model has
considerable psychological reality. Synchronically, it reflects the distinc-
tive features by which we recognize consonants. Diachronically, it reflects,
according to Roman Jakobson (1968), the splits that govern the acquisition
of the phonological system by infants all over the world and, in reverse
order, its loss by aphasic patients.
According to Victoria Fromkin (1973), the speech errors discussed here
provide additional evidence for their psychological reality. In a linguistic
study, she explored what kinds of linguistic units are transposed in real-life
“slips of the tongue.” She brought to light a wealth of evidence concerning
(a) v voiced
b
m
p
f unvoiced
(b) voiced
n d
t θ unvoiced
Figure 10.1 Model of consonants. (From Burke 1957: 297.)

6
3
2
the psychological reality of a wide variety of linguistic rules. Consider,

for instance, the following telltale example of transposing morphemes.
Instead of “Rosa always dated shrinks” someone said, “Rosa always dates
shrank.” Shrank was obtained by transposing the past tense from dated
to (the noun!) shrink and transferring the /s/ of shrinks to dates. To draw
attention to the grammatical sophistication of this unconscious process,
note that shrank belongs to a different paradigm of the past tense than
dated and that /s/ signifies the plural for nouns in its original place and
the present tense third person in its new place. An interesting aspect of
the more than six thousand speech errors Fromkin collected involving the
substitution and permutation of various sound segments concerned the
type of speech units affected. Besides such consciously accessible items as
words and clusters of speech sounds, speech units of which the speaker is
unaware may also be involved. For the time being I will mention only one
such unit—the distinctive features of speech sounds as in “glear plue sky”
for “clear blue sky,” where the feature [+ voiced] is transferred from the
initial phoneme of blue to the initial phoneme of clear: “When a person says
‘cedars of Lemadon’ instead of cedars of Lebanon, the nasality features of
the [b]‌and the [n] are reversed. The intended oral labial [b] becomes a nasal
labial [m] and the intended nasal alveolar [n] an oral alveolar [d]” (Fromkin
1973: 114).1
It is important for our purposes to note that the process itself is invol-
untary and inaccessible to conscious introspection. There are very few lan-
guage users who would be aware of the fact that the pairs of consonants
b/m and d/n, respectively, share the same place of articulation but are con-
trasted in the features [± nasal] [± continuous]; their disruption mech-
anism, however, knows this very well. This is exactly what happened in
Bacchi and Pagis, where the voiced labial [b]‌becomes [p] and the unvoiced
velar [k] becomes [g], or in mǝvakrim and mǝfagrim, where the voiced labial
[v] becomes [f] and the unvoiced velar [k] becomes [g]. There is, however,
a crucial difference. Whereas the shift from Lebanon to Lemadon results in
a meaningless utterance, the shift from Bacchi to Pagis or from mǝvakrim
to mǝfagrim not only is meaningful but even has deeper significance.
Fromkin’s point is that such errors demonstrate that units of language as
1. Kenneth Burke notes the same linguistic mechanism in the following para-
graph: “ ‘In Xanadu did Kubla Khan’ is found, by reason of the cognate relationship
between n and d, to be much more closely knit, on the phonetic basis, than would oth-
erwise be supposed. One might make this apparent by imagining himself pronouncing
the line with a head cold, thus: ‘Id Xadadu did Kubla Khad’ ” (1957: 300). The speech
error “Lemadon” would suggest that Burke is underscoring a phenomenon that does
have psychological reality.

7
3
2
defined by linguists are not merely theoretical constructs but real units
with an independent existence in mental grammar. In my view this is the
first step in a much more complex process.
Fromkin collected a great variety of speech errors. I will concentrate only
on those errors in which some symmetrical exchange of linguistic units
takes place. Some of these errors are called spoonerisms, that is, the trans-
position of initial or other sounds of words, usually by accident, attributed
to the Oxford don the Reverend W. A. Spooner, as in “a blushing crow” for
“a crushing blow,” “our queer old dean” for “our dear old queen,” or, chiding
one of his students, “You have hissed all my mystery lectures. I saw you
fight a liar in the back quad; indeed, you have tasted the whole worm.” As
we shall see below, there may also be switches of whole words at the one
end and, as we have seen above, even symmetrical switches of distinctive
features at the other.
Most genuine speech errors of this kind result in nonsense words.
Meaningless speech errors are, no doubt, funny. Bergson defined the
comic as “something mechanical encrusted on the living” (Bergson n.d.).
The ensuing nonsensical speech errors collected by Fromkin are, in fact,
a highly extreme case of this: the initial speech sounds of words are
mechanically switched, irrespective of meaning. Since the shift results
in a meaningless utterance, there is little more in it than “something
mechanical encrusted on the living.” The observer, in turn, may have—in
Muecke’s (1970: 36–37) words—a feeling of superiority, freedom, and
amusement in perceiving the failure of the other. And when the switch
does yield a meaningful expression, the meaning may be anomalous, as in
“a blushing crow.”
Molière, for instance, employed mechanical reversal of phrases to comic
effect. Consider the following dialogue in his L’Avare (The Miser):
Valère : . . . [A]‌ccording to the saying of one of the ancients, “We

must eat to live, and not live to eat.”
Harpagon : . . . It is the finest sentence that I have ever heard in my
life: “We must live to eat, and not eat to live.” No; that isn’t it. How
do you say it? (trans. Charles Heron Wall) (Molière n.d.)
Let us have a look briefly at a sample of speech errors collected by Fromkin:
keep a tape → teep a cape

the zipper is narrow → the nipper is zarrow
for far more → for mar fore
turn the corner → torn the kerner

8
32
I have suggested a few reasons why meaningless speech errors seem

funny to us. But in instances when speech errors result in meaningful
utterances, they may have very strong effects, witty or otherwise. That is
why such verbal devices can so readily be used to achieve literary effects.
Let us compare such spoonerisms and slips of the tongue with the above
random harvest from the genuine speech errors Fromkin collected. The
conspicuous difference is that the latter are nonsense, gibberish, whereas
the former have substantial meaning. In spoonerisms and slips of the
tongue a shift of consistent meaning is generated. This may yield much
more sophisticated effects.
JIGSAW PUZZLES AND INSIGHTS
This process is governed by the principle of the jigsaw puzzle. The shift of
letters or distinctive features undermines the meaning and generates cha-
otic utterances just as in the above speech errors. But then, in a split sec-
ond, the changed parts are restructured, so as to fit into a new whole. In
jigsaw puzzles there are sometimes thousands of pieces forming a chaotic
jumble; with sustained, patient work one can put the pieces together into
a coherent whole. This gives a sense of satisfaction caused by the transi-
tion from chaos to order and coherence. In spoonerisms and slips of the
tongue the transition from chaos to coherence occurs in split seconds
and appears to consciousness as an insight. Köhler’s discussion of insight
seems apropos here:
Those European psychologists, myself once included, sometimes went a bit too
far. Very much impressed by the essential rôle of insight in productive thinking,
they often said that the solution of problems is brought about by insight—as
though nothing else counted. Now this statement is not entirely correct for
the following reason. Insight is insight into relations that emerge when cer-
tain parts of a situation are inspected. . . . In the solution of a problem . . . we
suddenly become aware of new relations, but these new relations appear only
after we have mentally changed, amplified, or restructured the given material.
(1972: 152–153)
According to this analysis, what we usually call insight is the unique con-
scious quality that suddenly emerges when the objects of mentation are
restructured. In terms employed in cognitive poetics, it is not the cause but
the perceived quality of this sudden emergence. The jigsaw-puzzle effect
thus takes an interesting twist. Köhler (1972: 163) refers to the three Bs,

9
3
2
namely, “the Bus, the Bath, and the Bed,” where some of the greatest sci-
entific discoveries have been made (recall Archimedes and Kékulé!2). As for
the various insights reached in this way, “they all agree on one point. After
periods during which one has actively tried to solve a problem, but has not
succeeded, the sudden right organization of the situation, and with it the
solution, tend to occur at moments of extreme mental passivity” (Köhler
1972: 160).
The solution suddenly occurs at a time when, in Lowes’s phrase, “all con-
scious imaginative control is for some reason in abeyance” (Lowes 1927).
This is when a restructuring of the situation can take place. In piecing
together a jigsaw puzzle the element of suddenness is missing. Insight is
also preceded by a long period of sustained effort, but the sudden occurrence
of the solution renders it more dramatic and creates something new. In an
important way, this is exactly what happens in spoonerisms and slips of
the tongue as well. They suddenly restructure aspects of both the phonetic
component and the meaning of the utterance into a new whole. That is how
spoonerism and related phenomena become verbal imitations of insight.
One might add that in Freudian slips of the tongue, at least, long-repressed
or conflicting wishes emerge when “conscious . . . control is for some reason
in abeyance.” If such switches of linguistic units (ranging from distinctive
features, through phonemes, through clusters of phonemes, through mor-
phemes, to whole words) are embedded in a poetic text, they can amplify
and sometimes generate the poetic effects of the text.
FEATURE REVERSAL IN JOKES AND RHYMES
As a succinct instance of literary work, let us consider the subgenre of

political jokes in which a broadcaster “revises” a news item where cer-
tain linguistic units have been switched by mistake to deliver some biting
political message. In the early 1970s (before the Yom Kippur War) a popu-
lar joke in Egypt criticized the Egyptian army’s ineffectiveness and bogus
2. “There I sat and wrote my Lehrbuch, but it did not proceed well, my mind was else-
where. I turned the chair to the fireplace and fell half asleep. Again the atoms gamboled
before my eyes. Smaller groups this time kept modestly to the background. My mind’s
eyes, trained by visions of a similar kind, now distinguished larger formations of vari-
ous shapes. Long rows, in many ways more densely joined; everything in movement,
winding and turning like snakes. And look, what was that? One snake grabbed its own
tail, and mockingly the shape whirled before my eyes. As if struck by lightning I awoke.
This time again I spent the rest of the night working out the consequences” (http://
www.famousscientists.org/friedrich-august-kekule/).

0
42
claims: “Here is a rectification of a news item broadcast earlier: instead of

‘Two Egyptian frogmen sank an Israeli boat,’ it should read ‘Two Egyptian
boatmen sank an Israeli frog.’ ”
Wit is typically associated with sudden reversal. Here the phonological
sequences /bəʊt/ and /frɒg/ change places. This change qualifies as “some-
thing mechanical encrusted on the living” and is perceived as funny. By the
same token, the whole meaning structure is subverted. On closer examina-
tion, however, something more sophisticated takes place. A series of pho-
netic segments and semantic features are dislocated but “miraculously” fall
into place and yield a coherent new meaning. Syntactically, in the origi-
nal version boat is part of the object phrase of the verb sank, and frog is
the modifier of the noun men in the subject phrase. This syntactic rever-
sal causes some semantic features of the nouns involved to be reversed
as well. In the original version the meaning of boat includes the feature
[− animate], whereas in the “corrected” version it is part of a phrase whose
meaning includes [+ animate] and [+ human]; the word frog in the original
version is part of a phrase whose meaning includes the features [+ ani-
mate] [+ human], whereas in the corrected version it includes [+ animate]
[− human]. Normally, the reversal of all these syntactic and semantic fea-
tures would result in chaos; but here they result, as though by magic, in a
coherent meaningful utterance.
The reversal of such semantic features as [± verb] or [± concrete] or [±
animate] goes much deeper in poetic language than this example might sug-
gest. Consider the rhymes in the following three couplets from My Fair Lady:
This is what the British population

Calls an elementary education.
Why can’t the English teach their children how to speak?
This verbal class distinction by now should be antique.
Why can’t the English teach their children how to speak?
Norwegians have Norwegian, the Greek have got their Greek.
The first couplet rhymes [+ noun] with [+ noun] involving the same affix,
whereas the latter two rhyme [± verb]. Jakobson termed such these rhymes
“grammatical” and “antigrammatical,” respectively. Wimsatt described
them as “tame” and “vigorous” rhymes. I have discussed this issue elsewhere
(see, e.g., Tsur 2008a: 277–278). Wimsatt only refers to different parts of
speech, but I have shown that the reversal of the features [± concrete]
and [± animate] can be very effective in generating vigorous rhymes. Much
of this musical’s wit can be attributed to the vigorous, antigrammatical

2
41
rhymes of its lyrics. By contrast, in the first stanza of Verlaine’s poem

below, the vigorous, antigrammatical rhymes yield extremely strong
musicality and an emotional quality: the adjective longs rhymes with the
noun violons, and the noun Automne rhymes with the adjective mono-
tonne, whereas the concrete noun coeur rhymes with the abstract noun
langueur.
Some philosophers of art use the term multivalence (Philip Wheelwright)
or multiple relationship (Monroe C. Beardsley) between parts of an aesthetic
object. Multivalence (or multiple relationships) affects aesthetic appreci-
ation in an important way. This principle seems to apply to slips of the
tongue as well. The greater the number of inverted features in the inverted
units, the more fascinating the slip. As we have seen, in the phonological
dimension, distinctive features, phonemes, clusters of phonemes, or whole
words may be interchanged. We have treated the semantic component as
bundles of semantic features where each feature can be inverted. And as
we have seen, syntactic parts can also be inverted. This is one reason why
the frogmen joke is so effective; another reason is the extreme switch from
the bold and daring to the exceptionally petty. However, syntactic and
semantic reversals alone, even an anticlimax, would not do the job. The
effect requires a phonological reversal as well. Suppose the revision had
gone from “Two Egyptian divers sank an Israeli ship” to “Two Egyptian
sailors sank an Israeli toad.” The semantic information would be roughly
preserved, but there would be no joke.
The correction turns the bold and daring feat into a ridiculously trivial
exploit, generating biting irony. Ironists pretend to know nothing, not
even that they are ironic. Here the irony results from the juxtaposition of
two statements with incompatible overtones. The juxtaposition is not a
conscious, deliberate action of the ironist, as it were, but the working of an
unconscious verbal mechanism.
I claim that speech errors and poetic language both derive their effects
from the disruption of the smooth working of the same linguistic mecha-
nism. In the frogmen joke the witty effect was generated by the mechanical
reversal of the phoneme sequences /bəʊt/ and /frɒg/ as well as their syn-
tactic placement and a range of semantic features. In poetry this is also an
effective means to achieve witty effects.
However, when the conspicuity of such reversals is toned down by a
variety of linguistic devices, the result may be musical and emotional
rather than witty. Elsewhere I have quoted and discussed a unique example
from Verlaine’s “Chanson d’Automne” (Tsur forthcoming). Here I repro-
duce some of the key features related to the musical texture, which con-
tributes to an atmosphere that has been described as “full of omen, lilting,

4
2
sorrowful, deathly” (Szerb 1943: III, 143). Note the extreme musicality of

the first stanza of this poem:
Les sanglots longs
Des violons
De l’Automne,
Blessent mon coeur
D’une langueur
Monotonne.
[The long sobbings
of Autumn’s
violins
wound my heart
with a languor
monotone.]. (my translation)
Here I draw attention to one point that rarely receives attention, which
makes use of a dynamic discussed above in relation to speech errors. As
I have pointed out (Tsur 1992a), the vowel of the first syllable of sanglots
([sã]) in the first line is nasal, and the second syllable is oral. The last word
of the line repeats the preceding syllable ([lõ]), but the [+ nasal] feature
from the second-to-last syllable is transferred to it.
In the last stanza there is another (rather complex) poetic spoonerism:
Et je m’en vais
Au vent mauvais
[And I am walking
In the bad wind]
The word m’en in the foregoing excerpt is involved in three sound patterns.
First, in the sequence “m’en v–” the two labials [m]‌and [v] colliterate: they
are contrasted by the features [± nasal] and [± fricative]. Second,
“vent m–” in the next line repeats this pattern, except that the features
[+ nasal] and [+ fricative] are reversed. Furthermore, m’en and vent are
near-homonyms, in that only the same two distinctive features have been
reversed in them. Third, the two rhyming pairs consist of the sequence
“m—vais.” In the first member of the rhyme pair, m’en vais, the dash stands
for a nasal vowel, whereas in the second rhyme fellow, mauvais, the feature
[+ nasal] is deleted. In this way, the activity of the sound texture is ampli-
fied at a more fine-grained level than on the phoneme level; by the same
token, its diffuse quality is enhanced, so that not only the sound structure

4
32
of words is divergent, but the cohesion of the distinctive features within

the compact phonemes is also loosened.
What makes Verlaine’s spoonerism so musical and emotional, as
opposed to the witty effect of “crushing blow” → “blushing crow,” for
instance? First, in the latter example, the permutation of sound clusters
produces an extreme shift of meaning, where the shift of meaning per se is
perceived as witty. In Verlaine’s spoonerism, by contrast, the two phrases
complement each other to form one meaningful utterance. In addition, “a
blushing crow” is interestingly anomalous. Second, as will be seen below,
objects with a stable characteristic visual shape (like a crow) inhibit the
smooth fusion of incompatible information. Finally, in Verlaine’s “Et je
m’en vais / Au vent mauvais” all the consonants are periodic, or at least
continuous and voiced (voicing is periodic), whereas the [kr] of crush-
ing and crow consists of [k]‌, an interrupted unvoiced consonant, and [r],
which is voiced, continuous, and periodic but multiply interrupted.3
THE tot PHENOMENON
In On Metaphoring (Tsur 1987; see also 2012c) I discuss the tip-of-

the-tongue phenomenon. Here I will repurpose that account for the pres-
ent discussion. In discussing poetic and figurative language, linguists and
analytic philosophers sometimes use terms such as suppressed (Jespersen)
or implicit (Beardsley) meaning. I suggest that “suppressed” or “implicit”
meanings may rely on the same underlying cognitive mechanism as the tot
phenomenon and may have similar perceived effects, with one all-important
difference: the tot experience may contain an element of frustration that
is absent from “mere” suppressed meanings characteristic of the language
of poetry. “It is a gap that is intensely active,” says James of tot (quoted in
Ehrenzweig 1965: 10). When I ask students whether they have ever experi-
enced the tot experience and what it feels like, they usually mention frustra-
tion but also being engulfed in some thick, intangible and invisible stuff. This
feeling is not unlike the sensation induced by some of the greatest Romantic
and Symbolist poetry. In both cases it is an invisible and intangible but
3. In traditional phonetics, consonants can be continuous, as in [f, l, s, z, m], or inter-

rupted, as in [p, t, k, b, d, g]. Continuous consonants can be periodic, as in [l, m, n,
j], or aperiodic, as in [f, s, z, v]. In periodic sounds the same sound shape is repeated
indefinitely, and they are nearer to music; aperiodic sounds are nearer to noise. Voiced
stops such as [b, d, g] have an interrupted component plus voicing that is continuous
and periodic. [r]‌is periodic but multiply interrupted. As a result, as I point out else-
where (Tsur 2012c; see also c hapter 7 above), [r] is double-edged: in a smooth, periodic
context, as in “Lili Marleen,” it is perceived as smooth and soft, whereas in the context
of unvoiced interrupted stops such as [k] it sounds harsh.

42
immensely active mass of condensed semantic and phonetic features that

failed to grow together into a solid unitary word. In the tot state, when the
forgotten word is recalled—when the diffuse semantic and phonetic features
come together into a solid unitary word—that amorphous condensed mass
suddenly disappears. In poetry, by contrast, there is no one specific forgotten
word, only suppressed semantic features and loosened phonetic features not
necessarily belonging to one word; one cannot, therefore, recall the “forgot-
ten” word, and the thick atmosphere need not disappear (see the analysis
below of Keats’s “And taste the music of that vision pale”).
The underlying mechanism is treated in Roger Brown’s (1970) seminal
essay on tot. In collaboration with David McNeill, Brown carried out an
ingenious experiment: “The idea of the experiment is that when examining
the words that come to mind when searching for one that does not come
we should be able to discover the principles governing the classification
system utilized in our memory” (1970: 274). Freud claimed to discover the
unconscious psychodynamic motives of his informants using essentially
the same technique. How can one possibly “neutralize” the personal ele-
ments in such an experiment? This is, perhaps, the simplest and most inge-
nious part of the Brown and McNeill experiment. They compiled a list of
words that occur at least once in four million but no more frequently than
once in a million. In order to induce a tot state in potential subjects, they
read out the definitions of these words to groups of Harvard and Radcliffe
undergraduates and asked them to produce words having a similar mean-
ing (SM) or a similar sound (SS). To analyze the data Brown and McNeill
devised a dictionary in which words are entered on cards instead of pages
and the cards are punched for various features of the words entered: “With
real cards, paper ones, it is possible to retrieve from the total deck any
subset punched for a common feature by putting a metal rod through the
proper hole. We will suppose that there is in the mind some speedier equiv-
alent of this retrieval technique” (Brown 1970: 292). For the word sextant,
for instance, the following definition was read: “A navigational instrument
used in measuring angular distances, especially the altitude of sun, moon
and stars at sea.” Brown noted, “The SM words included: astrolabe, com-
pass, dividers and protractor. The SS words included: secant, sextet, and sex-
ton”: “The problem begins with a definition rather than a word and so the
subject must enter his dictionary backward” (1970: 292). It is possible to
imagine that a set of semantic features can be extracted from the defini-
tion, such as “navigation, instrument, having to do with geometry”: “Metal
rods thrust into the holes for each of these features might fish up such a
collection of entries as astrolabe, compass, dividers and protractor. This first
retrieval, which is in response to the definition, must be semantically based

4
52
and will not, therefore, account for the appearance of such SS words as sex-
tet and sexton” (Brown 1970: 293).
In the tot case the first retrieval must include a card with the definition of
sextant entered on it but with the word itself incompletely entered. The card
might, for instance, have the following information about the word: two syl-
lables, initial s, final t. “The entry would be a punch card equivalent of s_ _t.
Perhaps an incomplete entry of this sort is James’s ‘singularly definite gap’
and the basis of generic recall.” Subjects with a correct definition matching the
input and an incomplete word entry will know that they know the word and
will feel that they almost have it, that it is on the tip of their tongue. If they
are asked to guess the number of syllables and the initial letter they should
be able to do so. They should also be able to produce SS words. Subjects are
frequently able to tell even which syllable of the missing word is stressed.
When SM and SS words are “matched” with the missing word, “the match
brings out the missing parts the way the heat brings out anything written
in lemon juice,” in both the semantic and the phonetic dimensions. Such
features seem to be available not only in the definition but in the missing
word as well, even when the word itself is not available. Had they been
extracted only from the definition, one could not produce SS words. Thus
the semantic features and the phonetic information seem to be intensely
active even when the word itself cannot be recalled. The psychological real-
ity of such semantic and phonetic information is further supported, as we
have seen, by Fromkin’s findings with speech errors.
I have conjectured that both in the tot state and in poetic language the sup-
pressed semantic and phonetic features constitute some intangible and invisible
condensed mass that induces a thick atmosphere. There is, however, a serious
problem of communication here: How can one convey such information to stu-
dents, for instance? As to the tot state, everybody has experienced it. Someone
who hasn’t will never understand what we are talking about. As to poetic lan-
guage, we can only recommend what to look for and how to look at it and sug-
gest that what we are looking for is, in an important sense, something not
unlike the tot experience. Let us consider briefly stanza XLIX of Keats’s Isabella:
Ah! wherefore all this wormy circumstance?

Why linger at the yawning tomb so long?
O for the gentleness of old Romance,
The simple plaining of a minstrel’s song!
Fair reader, at the old tale take a glance,
For here, in truth, it doth not well belong
To speak:—O turn thee to the very tale,
And taste the music of that vision pale.

4
62
Such expressions as “wormy circumstance,” “yawning tomb,” and “the gen-

tleness of old Romance” set the emotional tone of the passage. But it is
the last line that bestows a condensed, invisible, inaudible, intangible, but
highly active presence. Let us see how it works.
The line contains a double intersense transfer, both of which are in the
expected direction (see below). Both are upward transfers, that is, the line
speaks of vision in terms of music and of music, in turn, in terms of taste.
Closer scrutiny of the interaction of the semantic features in these transfers
may account for the impassioned, uncanny atmosphere perceived by many
readers as perceptible in this line. The word vision denotes an abstraction
that has no stable visual shape. In contradistinction to sight, it suggests not
only the thing seen but also an impassioned state of mind with supernatu-
ral connotations. Also, the “visioner” contributes to the “vision” in a way
that the “seer” does not contribute to the “sight.” The paleness of the vision
may be associated with that of the dead, or Isabella, or the moonlit sight,
but none is explicitly mentioned, nor can they usurp each other’s place. So,
we have only a vague, indistinct, diffuse quality of paleness. The interac-
tion of the two terms of the genitive phrase deletes the feature [+ audi-
tory] in music, which in turn implicitly turns the gestalt-free vision into a
perceptual object that is a pleasant fusion of something full of energy and
expanding toward the perceiving self. In the sensory domain of taste no
shapes are possible at all. The interaction of taste with music deletes the fea-
ture [+ gustatory], foregrounding such meaning components as “directly
perceiving reality, or undergoing experience, or perceiving some fine tex-
ture or elusive quality.” The upward transfer from the less differentiated
sense taste enhances the indistinctness of the fused sensations. The pow-
erful fusion of the discordant senses heightens the discharge of emotions,
deleting the contradictory sensuous ingredients, leaving the reader with
the feel of a supersensuous, uncanny atmosphere. This line does not refer
to three different referents in the gustatory, auditory, and visual domains
but, rather, to one referent in the visual domain.
The notion of intersense transfer above alludes to Ullmann’s work on
panchronistic tendencies in synesthesia mainly in Romantic poetry. He
(1957: 280) works within a framework of a hierarchy of senses, where
touch and heat are at the lower end of the sensorium, hearing and seeing
are at the higher end, and taste and smell are in the middle (touch → taste
→ smell → sound → sight). Out of some two hundred transfers, only a
little more than one-sixth go downward (Ullmann 1957: 282). According
to Ullmann’s findings, the vast majority of transfers go from the less to the
more differentiated senses. In other words, contrary to what one might
expect, poets speak of the senses that have a richer vocabulary in terms of

7
4
2
the senses that have a poorer vocabulary, rather than vice versa. Ullmann
has no satisfactory explanation for this. I argue that language is a highly
differentiated tool and is, apparently, unsuitable to convey these less differ-
entiated, elusive, mysterious meanings it is sometimes supposed to convey
in poetry. Speaking of the more differentiated sense in terms of the less
differentiated one is one of the ways to do this. The above stanza is a fine
example.
“An abstraction that has no stable visual shape” alludes to my concep-
tion derived from Gestalt theory that the smooth fusion of sensory infor-
mation is inhibited across clear-cut contours (for further discussion, see
Tsur 2012c). Ullmann is puzzled by the following line: “The same bright
face I tasted in my sleep” (Keats, Endymion, I, line 895). He (1957: 287) says
that it is a strange phrase but cannot account for this strangeness. Here,
too, the poet speaks of the visual sense in terms of taste, that is, an upward
transfer. However, based on the above reasoning, the stable characteristic
visual shape of “face” inhibits the smooth fusion of the senses.
Thus, the slip-of-the-tongue and tip-of-the-tongue phenomena provide
empirical evidence that words are bundles of semantic and phonetic fea-
tures. The feature-deletion theory of metaphor, too, conceives of words
as bundles of semantic features. Metaphoric contradiction deletes incom-
patible semantic features and foregrounds others, thus performing what
Ehrenzweig called “thing-destruction” and generating “thing-free” quali-
ties (by the same token it can account for humans’ ability to understand
novel metaphors to which they have not been exposed before). In poetry,
just as in the tot state, a mass of unexpressed features generates a thick,
intangible, invisible, and inaudible undifferentiated atmosphere. This is
because suppressed meanings in poetry do not mean deleted features but,
rather, that the remaining undeleted features that convey crucial informa-
tion are not explicitly mentioned.
Thus I take issue with Brown and McNeill’s suggestion that “the entry
would be a punch card equivalent of s_ _t. Perhaps an incomplete entry of
this sort is James’s ‘singularly definite gap’ and the basis of generic recall.”
The “singularly definite gap” is generated not only by the incomplete pho-
netic entry but by the suppressed mass of phonetic and semantic features.
Moreover, while this model provides an exceptionally illuminating expla-
nation of the process underlying the phenomenon, the solid object, the
punch card, grossly interferes with its phenomenological quality, which we
tend to experience as a condensed mass of engulfing diffuse particles sus-
pended in the air.
I have said that subjects in the tot state sometimes have exact informa-
tion on the number of syllables and the placement of stress in the missing

8
42
word. It would appear at first sight that this piece of information has little
relevance to our inquiry. In poetry the thick atmosphere is not generated
by the semantic features of one word that failed to grow together but is
abstracted from several words in a context, and there is no uniquely identi-
fiable word to be retrieved. One can thus have no knowledge of the number
of syllables or the location of primary stress. On second thought, however,
it is precisely these features of language that are systematized into syllabo-
tonic meter, independently of the specific words.
SCRAMBLING AND ANAGRAM
One of Brown’s intriguing findings is this:
A subject trying to find the word ambergris thinks of Seagram. The usual sorts of
resemblance that constitute the main effects of the experiment (short strings of
identical letters at the beginning and ends of the words) are absent. Still there
is a resemblance that can hardly be accidental. All the letters [= speech sounds]
of Seagram are contained in ambergris. The word found seems to utilize the same
letter-stock or sound as the word sought without regard for order. This is a fasci-
nating outcome because it corresponds with one kind of rather common reading
mistake and, together with the reading mistakes, it suggests that order may be a
feature of a word that is stored independently of letters. (1970: 275)4
In “Musicality in Verse” Kenneth Burke (1957) refers to scrambled con-

sonants in Coleridge’s poetry and calls this an “anagram” or “consonan-
tal acrostic.” He collected a great number of verse lines from Coleridge’s
4. As to reading mistakes, more recently, drawing on a long series of rigorous neu-
ropsychological experiments, Dehaene gives a meticulously detailed account of the
reading process in the brain: “The adaptation process in the letterbox area reveals sev-
eral successive levels of letter coding, organized hierarchically from the back toward
the front of the occipito-temporal cortex, with an increasing degree of abstraction”
(2009: 91). At one stage, the system treats anagrams such as anger and range or cen-
ter and recent as the same; at the next stage it treats them as different (Dehaene
2009: 91–92; cf. chapter 2 above). This suggests that the reading mistake results from
stopping short in the cognitive chain. Sudden awareness of these subliminal stages
may account for pleasure in the scrambling of letters and speech sounds in witty or
emotional poetry. According to Freudian theory, one of the possible causes of pleasure
is regression to some earlier stage of functioning: “A part of the pleasure derives . . .
from the relation to infantile life”; “in all play with words, in puns as well as nonsense
talk, there is a renewal of the child’s pleasure when it just learns to master language”
(Kris 1965: 174). Thus, anagrams may serve as a publicly sanctioned mode of regression
to an earlier stage of information processing.

4
92
poetry in which sound patterns are carefully backgrounded by a wide

range of manipulations. Below I discuss a few in which backgrounding is
achieved by the reordering of consonants in words in proximity. Consider,
for instance, that in “green light that lingers” the g-r-n-l of green light is
acrostically reordered in the l-ng-r of lingers. In “tyrannous and strong,”
the t-r-n-s of the former adjective becomes str-ng in the latter. In “so
fierce a foe to frenzy” the consonants f-rs in fierce become fr-z in frenzy,
with the addition of the feature [+ voiced]. Burke writes, “Perhaps the
most beautiful example of a consonantal acrostic in Coleridge is a line
from ‘Kubla Khan’: ‘A damsel with a dulcimer,’ where you match d-m-
z-l with d-l-s-m (plus r)” (1957). Such verse lines considerably amplify
the musical texture, without drawing too much attention to the intense
sound repetitions.
In Tsur 2003 I suggest that the psycholinguistic mechanisms underlying
the permutation of speech sounds in Fromkin’s and Brown’s speech errors
underlie some mystic poetry as well. Take the anagram by George Herbert
in Figure 10.2. Here, too, the words Mary and Army utilize, in Brown’s
words, the same letter-stock or sound without regard for order. What is the
mystical significance of a permutation of letters, such as mary → army? In
Gershom Scholem’s (1961) analysis of the doctrine of Abraham Abulafia,
the great thirteenth-century Jewish mystic, he notes that Abulafia utilized
a peculiar discipline that he calls the “science of the combination of let-
ters.” This is described as a “methodical guide to meditation with the aid
of letters and their configurations. The individual letters of their combina-
tions need have no ‘meaning’ in the ordinary sense; it is even an advantage
if they are meaningless, as in that case they are less likely to distract us”
(Scholem 1961: 133; see also c hapter 5). One might object that when read-
ing such an anagram one undergoes no meditative or mystic experience.
This is, precisely, the point of the present book. We are facing a fossilized
remnant of such an experience.
When mystics talk about the permutation of letters, it is not always clear
whether they mean the written marks on the paper or the speech sounds
they signify. Mystical experiences of great force have been induced by both
the manipulation of the speech sounds and letters of God’s name and
{
Ana- MARY gram
ARMY }
How well her name an Army doth present,
In whom the Lord of Hosts did pitch his tent!
Figure 10.2 Anagram. (By George Herbert.)

0
52
some meaningless speech sound sequences. Whereas in Herbert’s ana-

gram it is the letters that are manipulated, in Hopkins’s poetry, for exam-
ple, clusters of speech sounds are exceptionally foregrounded, far beyond
what is acceptable in ordinary alliteration. Consider such notorious
densely packed phrases as “Heaven-Haven,” and “I caught this morning
morning’s minion, kingdom of daylight’s dauphin, dapple-dawn-drawn
Falcon” (“The Windhover”), and the sonnet “The Sea and the Skylark”:
ON ear and ear two noises too old to end

Trench—right, the tide that ramps against the shore;
With a flood or a fall, low lull-off or all roar,
Frequenting there while moon shall wear and wend.
Left hand, off land, I hear the lark ascend,
His rash-fresh re-winded new-skeinèd score
In crisps of curl off wild winch whirl, and pour
And pelt music, till none’s to spill nor spend.
How these two shame this shallow and frail town!
How ring right out our sordid turbid time,
Being pure! We, life’s pride and cared-for crown,
Have lost that cheer and charm of earth’s past prime:
Our make and making break, are breaking, down
To man’s last dust, drain fast towards man’s first slime.
To mention only the most conspicuous sound repetitions and scramblings

in stanza 2, for example, lark and curl share the same consonants; curl, in
turn, differs from whirl only in the first consonant. All the consonants of
score are repeated in crisps, and two of them are repeated in skeinèd and in
some other words. The speech sounds l-f-and of left hand are scrambled as
f-land in off land. All the consonants of rash are included in the same order
in fresh (and even the vowels are quite similar). Most interesting from
this point of view is the phrase “till none’s to spill nor spend,” where the word
spill contains the last two sounds of till and the first two of spend.5
I claim that when invested with enormous cathexis (that is, mental or
emotional energy), the unconscious processes involved may be experienced
as highly emotional, even ecstatic, rather than witty. One of Abulafia’s dis-
ciples reports that “the power of meditation became so strong in me that
5. For discussion of the relationship between tongue-twisters and ecstatic quality in

Hopkins’s “The Windhover,” see Tsur 2010.

251
I could not manage to write down the combinations of letters [which auto-
matically spurted out of my pen], and if there had been ten people present
they would not have been able to write down so many combinations as
came to me during the influx” (in Scholem 1961: 150–151) (see chapter 5).
In Hellenistic or Renaissance anagrams, the letters rather than the speech
sounds are manipulated, and the enormous charge of emotional energy
is eliminated. This process is nearer to wit than to ecstasy: the expressive
means are frozen into frivolous word games. I argue in Tsur 2003 that in
Herbert’s anagram this is taken one step further, where the frozen literary
device is revived by literary manipulations.
It is illuminating to compare the perceived effect of scrambling in
Coleridge’s “A damsel with a dulcimer” and Herbert’s “Mary” and “Army.”
The former sounds musical, smooth, emotional, “Romantic”; the latter,
rather sharp and witty, “Mannerist.” According to Jakobson (1960), the
distinctive characteristic of poetic language is that it compels the reader to
attend back to the linguistic signifier from the extralinguistic signified. In
chapter 6 I go one step further: I conceive of Mannerist and non-Mannerist
styles in terms of different sign relationships. As long as the signified pre-
serves its dominance to some extent, the emphasis on the signifier is more
natural, less “marked.” When the relative weights of the signifier and the
signified are balanced or the former overrides the latter, the process is per-
ceived as less natural, more artificial, more “marked.”
Jakobson (1968) distinguishes two stages in child language: the bab-
bling period, when the infant explores the similarity of speech sounds, and
the stage when the child learns to use speech sounds referentially. In this
stage, the organizing principle is contiguity. Similarity and contiguity are
organizing principles on all levels of language (Jakobson 1956). The dif-
ferentia of poetic language is that similarity is superimposed on contiguity
(Jakobson 1960). In our case, on the phonological level, in both “A damsel
with a dulcimer” and Mary and Army the contiguous speech sounds add
up to words and larger syntactic units. At the same time, the same speech
sounds form pairs of similar phoneme clusters. Syntactically, “with a dul-
cimer” is a restrictive attribute of “A damsel” and as such contiguous with
it. Thus, the referential function of language based on contiguity is in the
foreground; the similarity of the sounds is perceived as a smooth musical
background. The relationship between Mary and Army is neither restrictive
nor contiguous syntactically, but they are parallel in both their collitera-
tive and figurative structure, foregrounding the linguistic signifier at the
expense of the referent. This is what Joseph Addison (quoted in chapter 6)
called similarity of words rather than similarity of ideas. To be precise, both
instances display the same similarity of words, but in Herbert’s case the

2
5
two words refer to two dissimilar ideas, whereas Coleridge’s two phrases
join forces to designate one idea.
There is an additional aspect that may illuminate the difference
between the two examples, precisely where they deviate from regularity.
Both examples involve two similar clusters of signs: the same units, in
a different order, except one deviation. We have in dulcimer an /s/ and
in damsel a /z/, which is a voiced /s/. In Mary and Army the letters m,
a, r, and y are repeated, but the letter a refers to different vowels in the
two words. Consequently, the component shared by /s/ and /z/ may
fuse smoothly in perception, whereas the vowels signified by the letter
a in Mary and Army cannot. It is the deviant item that resolves whether
we are dealing with the patterning of speech sounds or graphemes: in
Coleridge the patterning of speech sounds, in Herbert the patterning of
graphemes. In chapter 6 (as well as previously in this chapter) I argued
that stable visual shapes interfere with the smooth fusion of elements
in poetry, yielding a sharp, witty, split focus, whereas sound gives infor-
mation about change and may fuse smoothly, yielding a soft integrated
focus. Consequently, the patterning of graphemes in Herbert, reinforced
by the genre “picture poetry,” results in a witty, “Mannerist” quality; the
patterning of speech sounds in Coleridge, in a smooth musical, emotional,
“Romantic” quality. The upshot of this discussion is as follows: the same
psycholinguistic mechanism underlies scrambling in certain speech and
reading errors, on the one hand, and in certain poetic devices, on the
other. What is more, in the poetic devices the same mechanism yields very
different stylistic effects, determined by aesthetic principles discussed
in chapter 6.
MISQUOTING POETRY: THE UNCONSCIOUS MIND

AND GENERATION BY RULE
In Tsur 1992a I discuss another issue from the psychopathology of daily

life that concerns a case of misquotation of a stanza from a poem by Keats,
thus revealing a repressed painful experience. In addition to the “Freudian”
aspect, the story also has an intriguing linguistic/cognitive aspect: it
reveals in great detail how the unconscious mind has its way in exploiting
the cognitive resources of linguistic competence in this and the foregoing
examples as well.
In Psychopathology of Everyday Life Freud (1953: 18–20) provides an
intriguing example, originally reported by Brill. It concerns the relation-
ship between cognitive and “psychopathological” processes, but because

3
5
2
it deals with a piece of poetry, it also provides insights into the mutual
interferences between psychopathological processes and poetic language.
It is illuminating to note that Keats himself consciously relied, as we shall
see, on the same rules in revising his poem “La Belle Dame sans Merci” for
stylistic purposes as the unconscious relied on in misquoting Keats’s “Ode
to Apollo” for psychopathological purposes.
A young woman misquoted four lines from Keats’s “Ode to Apollo” as
In thy Western house of gold

Where thou livest in thy state,
Bards, that once sublimely told
Prosaic truths that came too late.
The correct lines read as follows (the words forgotten and replaced by oth-
ers being italicized):
In thy western halls of gold

When thou sittest in thy state,
Bards, that erst sublimely told
Heroic deeds and sang of fate.
Significantly, the young woman initially could not recall when she memo-
rized these lines. But it came back after Brill suggested: “Judging by the
conversation, it would seem that this poem is intimately associated with
the idea of over-estimation of personality of one in love. Have you per-
haps memorized this poem when you were in such a state?” The upshot
of the ensuing story is as follows. Everything went well for a few months
until the woman suddenly received word that her Apollo, for whom she had
memorized these lines, had eloped with and married a very wealthy young
woman. A few years later she heard that he was living in a Western city,
where he was taking care of his father-in-law’s interests.
The misquoted lines are now quite plain. Brill’s comment about the
overestimation of personality among lovers unconsciously reminded the
woman of a disagreeable experience, when she herself overestimated
the personality of the man she loved. She thought that he was a god, but
he turned out to be even worse than the average mortal. The episode could
not come to the surface, because it was accompanied by very disagreeable
and painful thoughts, but the unconscious variations in the poem plainly
showed her present mental state. The poetic expressions not only were
changed to prosaic ones but clearly alluded to the whole episode (Freud
1953: 19–20).

4
52
In this case, unconscious processes made efficient use of some of the

commonest cognitive processes (as documented more recently by cogni-
tive science). Let us consider, first, some of the more trivial replacements
of the misquotation: house for halls, livest for sittest, and once for erst. How
would a computer program (which cannot be credited with an unconscious
mind) handle such a passage? In Roger Schank’s version of artificial intel-
ligence, it would recode it, first of all, into a representation using semantic
primitives—what Schank would call an “interlingua,” based on his system
of “conceptual dependency” (see, e.g., 1973). After this conceptual process-
ing, the program could answer, in the question-answering mode, such ques-
tions as not only “What tales did bards tell?” but also “What did ancient
poets sing about?” In the translating mode, the program would not directly
search for the target-language equivalent of Keats’s words but would gener-
ate a text from this interlingua, which is roughly equivalent to the original,
in French, or Spanish, or Chinese (in practice, the program—like so many
intelligent readers—may have difficulties with the extraordinary complex-
ity of the syntax, but in principle, this is how it works with news items, for
instance).
In the paraphrasing mode, with regard to the more trivial replacements
mentioned above, we might expect the program to do exactly what the
young woman did: namely, substitute house for halls, livest for sittest, and
once for erst. Here, too, the program would have recourse to the interlin-
gua representation of the text and generate another English text, much
the way it generated foreign-language texts. Faced with semantic primi-
tives constituting highly specific words, we might expect it to pick out their
more general synonyms. Now this is exactly how flesh-and-blood people
understand texts and store and retrieve them from long-term memory. We
can only remember the exact words in which we received a certain piece of
information for a very short time. After a few seconds we “recode” it into
an “interlingua” representation, more suitable for long-term memory stor-
age (cf. Tsur 1983: 9–12, 2008a: 6–7). After some time, we can no longer
tell in what language we were given a specific piece of information. When
asked “What did the man tell you a week ago?” we usually generate “by
rule” a new text from our own interlingua representation, unless we con-
sidered the message so important that we memorized it word for word.
An intermediary possibility between word-for-word memorizing and the
generation by rule of a new text from the interlingua would be to add to the
interlingua representation of words such as erst such indices as [+ archaic]
or [+ poetic]. So, when generating a new text by rule from the long-term
memory representation, a word is picked out that has these features, per-
haps the word erst itself or perhaps the phrase of yore.

52
In order to account for the performance of Brill’s young woman, we

must consider one more issue derived from Liberman, Mattingly, and
Turvey (1972), which I have discussed elsewhere (e.g., Tsur 1992a: 151).
Liberman and his collaborators see systems of versification as a kind of
secondary codes:
For a literate society the function of verse is primarily esthetic, but for preliter-
ate societies, verse is a means of transmitting verbal information of cultural
importance with a minimum of paraphrase. The rules of verse are, in effect, an
addition to phonology which requires that recalled material not only should pre-
serve the semantic values of the original, but should also conform to a specific,
rule-determined phonetic pattern. (1972: 325)
In the case of Brill’s young woman, the psychopathological process con-

sisted of an interference with recoding, which partially relinquished the
restrictions on versification. In this way, her cognitive functioning con-
formed more closely to the normal course of communication and remem-
bering. Instead of remembering the prosodic organization as embodied in
this specific piece of poetry, she encoded it as an additional rule [+ tro-
chaic meter] and [+ rhyme scheme: gold, state, told, -ate] to generate
the text from long-term memory. In this way, the unconscious fooled the
traditional purpose of versification, devised to ensure reproduction with a
minimum of paraphrase: in the last line it generated a completely new text
that, nevertheless, “conformed to the specific, rule-determined phonetic
pattern” of the original poem.
At the same time, the young woman systematically omitted indices such
as [+ archaic] and [+ poetic]. Thus the mechanism responsible for the
misquotation of Keats’s lines by Brill’s young woman has a cognitive and a
psychopathological facet. The cognitive facet is the commonest possible in
the normal course of communication. It is most typically simulated by arti-
ficial intelligence programs. Here, however, the normalization is a result
of the relaxation of the versification rules restricting the normal recoding
of the surface-structure representation into an interlingua representation.
What is different in this example is that the unconscious mind exploited
the normal cognitive processes for its psychopathological purposes. Here
the fact that the poetic expressions were changed to prosaic ones becomes
significant in the woman’s unconscious attempt to give expression to a
repressed episode and, at the same time, to avoid disagreeable and painful
thoughts.
There is a tendency to use as general terms as possible when generating
a surface representation from an interlingua representation by rule. This

6
5
2
tendency suited the purposes of the young woman’s unconscious mind,

which was interested for its own “psychopathological” reasons in turning a
poetic text into a more prosaic one.
These kinds of substitutions can occur for stylistic reasons as well.
When Keats revised his “La Belle Dame sans Merci,” he relied—albeit more
consciously—on the same kinds of indices, but in the opposite direction,
when changing the first line from “O what can ail thee, Knight at arms . . .”
to “Ah, what can ail thee, wretched wight. . . .” We rarely encounter wight
outside Spenser’s Fairie Queene, and it seems to have been introduced here
to increase the archaic effect of the poem; even the substitution of Ah for
O heightens the poetic diction. All this, however, has nothing to do with
motives such as those attributed to Brill’s acquaintance, which should not
be dragged into the poem’s context. This was a conscious stylistic decision,
consistent throughout the revised version, and not necessarily to the bet-
ter from an aesthetic point of view. The woman’s misquotation of the last
line (the one most telling from the psychoanalytic point of view) cannot
be accounted for in this way. However, it can be accounted for by assum-
ing that the unconscious psychopathological process, having removed the
restrictions of versification on recoding, exploited an ongoing cognitive
process at a different level of information processing.
In order to explain this, we have to turn to Bartlett’s discussion of sche-
mata involved in the perception and remembering of situations:
An individual does not normally take such a situation detail by detail and metic-
ulously build up the whole. In all ordinary instances he has an overmastering
tendency simply to get a general impression of the whole; and, on the basis of
this, he constructs the probable detail. . . . The construction that is effected is the
sort of construction that would justify the observer’s “attitude.” (1932: 206–207;
cf. Tsur 2008a: 17–18)
Here the unconscious mind had its way by interfering with the cognitive
process at the “attitude” level. By modifying the young woman’s attitude
toward the poem, the construction of the probable detail resulted in an
utterly different fourth line. “Prosaic truths that came too late” expresses
the young woman’s attitude toward the occasion when she memorized
Keats’s lines, but here it also affected the attitude that served the recon-
struction of the lines themselves from memory. In other words, the two
kinds of attitude were not kept properly separated. Here as well the line is
constructed in accordance with the prosodic rule memorized: The fourth
line deviates from the preceding lines in being iambic; it begins with a tri-
syllabic word and ends with the sounds -ate. The substitution of where for

7
5
2
when before this can only be explained by the assumption that once her
unconscious removed the poetic restrictions on the recoding of the text,
the woman needed to reconstruct the probable detail by rule.
CONCLUSION
People are fascinated by phenomena related to the psychopathology of

everyday life such as slip-of-the-tongue and tip-of-the-tongue phenomena
because they provide a glimpse into two kinds of unconscious processes to
which we have no direct access: unconscious “Freudian” processes and the
organization of semantic and phonetic units in mental grammar. The per-
mutation of letters and phonetic units for psychopathological purposes or
witty effects or to induce a mystical state of mind testifies to their intimate
relationship with the unconscious mind.
I have argued that speech errors may have a Freudian and a linguistic
facet and that the same psycholinguistic processes may be disrupted for
psychopathological and aesthetic purposes. Fromkin explored the transpo-
sition of linguistic units in speech errors ranging from phonetic features,
through phoneme clusters and morphemes, to whole words. A subclass of
such speech errors is called spoonerism, the transposition of correspond-
ing speech sounds of words, usually by accident, yielding a radically differ-
ent meaning. I have treated spoonerism as a verbal imitation of insight.
According to Köhler, insight occurs when “in the solution of a problem
. . . we suddenly become aware of new relations, but these new relations
appear only after we have mentally changed, amplified, or restructured the
given material” (1972: 153). That is one of several reasons for the enjoy-
ment prompted by spoonerisms. Transpositions of speech sounds or whole
words in their simplest forms are usually perceived as witty, but Burke,
for instance, collected instances of scrambling of consonants that yield
powerful emotive textures in Coleridge’s poetry. In Hopkins’s poetry the
scrambled consonants are exploited for ecstatic effects. In the tot state, an
invisible and intangible but immensely active condensed mass of seman-
tic and phonetic components is experienced, which is not unlike the thick
emotional atmosphere generated in the greatest Romantic and Symbolist
poetry. Finally, we saw that Freud quotes from Brill the case of a young
woman who misquoted a stanza from Keats’s “Ode to Apollo” as a result
of an emotional trauma. I pointed out that the substitutions make use of
the mechanisms of ordinary linguistic competence. Such mechanisms have
been successfully imitated by artificial intelligence programs that cannot
be credited with an unconscious mind.

8
52
This chapter may serve, in an important sense, as an epitome of the

thesis underlying this book. We have conceived of language in general and
poetic conventions in particular as powerful cognitive processes fossil-
ized into neutral tags. Poetic language is organized violence against lan-
guage and poetic conventions. By smashing language and conventions,
poetic language gains access to the effects of certain cognitive processes.
Thus the response to these effects is knowledge unlearned and untaught.
Such a conception solves the notorious problem that bedevils the
culture-begets-culture conception, which accounts for effects as being con-
ventions, but this merely transfers the mystery from one place to another.
It cannot explain how those effects became interpersonally accessible to
their first audience or how readers of our age know how to respond to
poetic structures to which we have not been exposed before. The Freudian
psychopathologies of everyday life can dramatically illuminate how all this
works. They too are organized violence against language (in the misquota-
tion of Keats’s poem, against poetic conventions, too), which renders the
workings of the relevant linguistic and cognitive mechanisms and their
perceived effects accessible to conscious introspection.
* * *
It would appear that the artwork as cognitive fossil has been with us for the
past eighteen thousand years or so. The great neuropsychologist Stanislas
Dehaene begins his book on consciousness and the brain with the follow-
ing example:
Deep inside the Lascaux cave, past the world-renowned Great Hall of the Bulls,
where Paleolithic artists painted a colorful menagerie of horses, deer, and bulls,
starts a lesser-known corridor known as the Apse. There, at the bottom of a
sixteen-foot pit, next to fine drawings of a wounded bison and a rhinoceros, lies
one of the rare depictions of a human being in prehistoric art [Figure 10.3]. The
man is lying flat on his back, palms up and arms extended. Next to him stands
a bird perched on a stick. Nearby lies a broken spear that was probably used to
disembowel the bison, whose intestines are hanging out.
The person is clearly a man, for his penis is fully erect. And this, according
to the sleep researcher Michel Jouvet, illuminates the drawing’s meaning: it
depicts a dreamer and his dream. As Jouvet and his team discovered, dreaming
occurs primarily during a specific phase of sleep, which they dubbed “paradoxi-
cal” because it does not look like sleep; during this period, the brain is almost
as active as it is in wakefulness, and the eyes ceaselessly move around. In males,
this phase is invariably accompanied by a strong erection (even when the dream
is devoid of sexual content). Although this weird physiological fact became

9
5
2
Figure 10.3 A dreamer and his dream: painting in the Lascaux cave, circa eighteen thou-
sand years old.
known to science only in the twentieth century, Jouvet wittily remarks that
our ancestors would easily have noticed it. And the bird seems the most natural
metaphor for the dreamer’s soul: during dreams, the mind flies to distant places
and ancient times, free as a sparrow. (2014: 1–2)
It is difficult to know, with such a singular remnant from the distant

prehistoric past, whether the erect penis as a metonymy for dreaming ever
became a convention and how much repeated social transmission, if any,
preceded it. But one thing appears to be clear: the erection reflects cogni-
tive constraints of dreaming, as pointed out by twentieth-century brain
scientists. And it acts as a symbol communicating the meaning of the paint-
ing, namely, that what you see is a sleeping man, dreaming of the spoils
of a hunt. In this sense the understanding of the painting’s symbolism is,
again, knowledge unlearned and untaught. While erection here is a natu-
ral metonymy suggesting a dreamer, if—and this is a great if—Dehaene is
right about the interpretation of the bird, this metaphor must be a pow-
erful symbol that is publicly accessible to members of the contemporary
interpretative community and later generations as well.

0
62
261
REFERENCES
Abd Razak, Ratna Roshida. 2007. “Arabic Poet Al-Mutanabbi: A Maslovian Humanistic
Approach.” PsyArt. Available online: http://psyartjournal.com/article/show/
roshida_abd_razak-arabic_poet_al_mutanabbi_a_maslovian_hum.
Addison, Joseph. 1951 [1711–1712]. “Spectator Papers.” In The Great Critics, ed. James
Harry Smith and Edd Winfield Parks, pp. 819–825. New York: Norton.
Aristotle. 1932. The “Rhetoric” of Aristotle. Trans. Lane Cooper. New York: Appleton
Century.
Aristotle. 1951. Aristotle, “Poetics.” Trans. S. H. Butcher. In The Great Critics, ed. James
Arnheim, Rudolf. 1957. Art and Visual Perception. London: Faber and Faber.
Balaban, Victor. 1999. “Religious Metaphor and Cognitive Linguistics: Perceptual
Metaphors for Knowledge at a Marian Apparition Site.” In Metaphor and
God-Talk, ed. Lieven Boeve and Kurt Feyaerts, pp. 253–271. Bern: Peter Lang.
Bartlett, F. C. 1920. “Some Experiments on the Reproduction of Folk Stories.”
Folk-Lore 31: 30–47. Available online: http://home.iscte-iul.pt/~fgvs/
Bartlett,%20Experiments.pdf.
Bartlett, Frederic C. 1932. Remembering. Cambridge: Cambridge University Press.
Beardsley, Monroe C. 1958. Aesthetics: Problems in the Philosophy of Criticism.
New York: Harcourt, Brace and World.
Beaver, Joseph C. 1968. “A Grammar of Prosody.” College English 29: 310–321.
Beer, J. B. 1959. Coleridge the Visionary. London: Chatto and Windus.
Bergson, Henri. N.d. Laughter: an Essay on the Meaning of the Comic. The Project
Gutenberg EBook.
Berliner, Paul. 1978. The Soul of Mbira: Music and Traditions of the Shona People of
Zimbabwe. Berkeley: University of California Press.
Blechner, Mark J. 1977. “Musical Skill and the Categorical Perception of Harmonic
Mode.” Haskins Laboratories: Status Report on Speech Research SR-51/52.
Available online: http://www.haskins.yale.edu/sr/SR051/SR051_12.pdf.
Bownds, M. Deric. 1999. The Biology of Mind—Origins and Structures of Mind, Brain,
and Consciousness. Bethesda, MD: Fitzgerald Science Press.
Bradburn, Beth. “Interview with Reuven Tsur.” Literature, Cognition & the Brain.
Available online: http://www2.bc.edu/~richarad/lcb/fea/tsurin/tsurmain.html.
Brann, Ross. 1991. The Compunctious Poet: Cultural Ambiguity and Hebrew Poetry in
Muslim Spain. Baltimore: Johns Hopkins University Press.
Brooke-Rose, Christine. 1958. A Grammar of Metaphor. London: Secker and Warburg.
2
6
Brooks, Cleanth, and Robert B. Heilman. 1966. Understanding Drama. New York: Holt,

Rinehart and Winston.
Brown, Roger. 1958. Words and Things. New York: Free Press.
Brown, Roger. 1970. “The ‘Tip of the Tongue’ Phenomenon.” In
Psycholinguistics: Selected Papers, pp. 274–303. New York: Free Press.
Bruner, Jerome S. 1973. Beyond the Information Given: Studies in the Psychology of
Knowing. New York: W. W. Norton.
Burke, Kenneth. 1957. The Philosophy of Literary Form. New York: Vintage.
Burke, Kenneth. 1962. “A Grammar of Motives” and “A Rhetoric of Motives.” Cleveland,
OH: Meridian Books.
Chatman, Seymour. 1965. A Theory of Meter. The Hague: Mouton (Janua Linguarum).
Child, Francis James. 2003 [1898]. English and Scottish Popular Ballads.
New York: Dover.
Clark, Herbert H., and Eve V. Clark. 1977. Psychology and Language: An Introduction to
Psycholinguistics. New York: Harcourt, Brace Jovanovich.
Coleridge, Samuel Taylor. 1951. “Biographia Literaria.” In Selected Poetry and Prose of
Coleridge, ed. Donald A. Stauffer, chapter IV, pp. 109–428. New York: Modern
Library.
The Complete Sefer Yetsira. 1972. Trans. and commentary by Rabbi Sa’adiah Gaon.
Jerusalem: Committee for the Publication of Rabbi Sa’adiah Gaon’s Writings.
Cooper, C. W., and L. B. Meyer. 1960. The Rhythmic Structure of Music.
Chicago: University of Chicago Press.
Cooper, William E., and John Robert Ross. 1975. “Word Order.” In Papers from
the Parasession on Functionalism, ed. Robin E. Grossman, L. James San, and
Timothy J. Vance, pp. 63–111. Chicago: Chicago Linguistic Society.
Culler, Jonathan. 1975. Structuralist Poetics. London: Routledge and Kegan Paul.
Culler, Jonathan. 2014. “Teaching Baudelaire through Translation.” Thinking Verse
4: 4–12.
Curtius, Ernst Robert. 1953. European Literature and the Latin Middle Ages. Trans.
Willard R. Trask. Princeton: Princeton University Press.
D’Andrade, Roy Goodwin. 1980. “The Cultural Part of Cognition.” Address given to
the 2nd Annual Cognitive Science Conference, New Haven.
D’Andrade, Roy Goodwin. 1981. “The Cultural Part of Cognition.” Cognitive Science
5: 179–195. Available online: https://people.ucsc.edu/~gwells/Files/Courses_
Folder/ED%20261%20Papers/D'Andrade%20Cultural%20Part%20of%20
Cognition.pdf.
Darwin, Charles. 1871. The Descent of Man. London: John Murray.
David, Yonah, ed. 1987. The Poems of Isaac Ibn Ghiyyat. Jerusalem: Akhshav.
Dayan, Peter. 2012. “Different Music, Same Condition: Hofstadter and Lyotard.”
Thinking Verse 2: 9–26.
Dehaene, Stanislas. 2009. Reading in the Brain: The New Science of How We Read.
New York: Penguin.
Dehaene, Stanislas. 2013. “Inside the Letterbox: How Literacy Transforms the
Human Brain.” Cerebrum, June 3. Available online: http://www.dana.org/news/
cerebrum/detail.aspx?id=43644.
Dehaene, Stanislas. 2014. Consciousness and the Brain: Deciphering How the Brain Codes
Our Thought. New York: Viking.
de Mourgues, Odette. 1953. Metaphysical, Baroque, and Précieux Poetry.
Oxford: Clarendon Press.
Dewey, John. 1980 [1934]. Art as Experience. New York: Perigee Books.
[ 262 ] References

3
6
2
Dresher, B. Elan. 2013. “Biblical Accents: Prosody.” In Encyclopedia of Hebrew Language

and Linguistics, ed. Geoffrey Khan. Leiden: Brill. Available online: http://dx.doi.
org/10.1163/2212-4241_ehll_EHLL_COM_00000248.
Ehrenzweig, Anton. 1965. The Psychoanalysis of Artistic Vision and Hearing: An
Introduction to a Theory of Unconscious Perception. New York: Braziller.
Ehrenzweig, Anton. 1970. The Hidden Order of Art. London: Paladin.
Einat-Nov, Idit. 1995. “Expressive Power and Fossilized Convention in Medieval
Hebrew Liturgical Poetry.” Unpublished master’s thesis, Tel Aviv University.
Eitan, Zohar. 2013. “Which Way Is More? Pitch Height, Parametric Scales, and
the Intricacies of Cross-Domain Magnitude Relationships.” In Musical
Implications: Essays in Honor of Eugene Narmour, ed. Lawrence F. Bernstein and
Alexander Rozin, pp. 131–148. Hillsdale, NY: Pendragon Press.
Eitan, Zohar, and Renee Timmers. 2010. “Beethoven’s Last Piano Sonata and Those
Who Follow Crocodiles: Cross-Domain Mappings of Auditory Pitch in a Musical
Context.” Cognition 114: 405–422.
Elior, Rachel. 1987. “The Concept of God in Hekhalot Mysticism.” In Early
Jewish Mysticism: Proceedings of the First International Conference on the
History of Jewish Mysticism. Jerusalem Studies in Jewish Thought 6: 1–2.
Jerusalem: Magnes.
Elior, Rachel. 1994. “Mysticism, Magic and Angelology in Hekhalot Literature.” In
Tribute to Sara—Studies in Jewish Philosophy and Kabbala Presented to Sara
O. Heller Wilensky, ed. Moshe Idel, Debora Dimant, and Shalom Rosenberg, pp.
3–53. Jerusalem: Magnes.
Empson, William. 1955. Seven Types of Ambiguity. New York: Meridian Books.
Even-Zohar, Itamar. 1978. Papers in Historical Poetics. Tel Aviv: Porter Institute for
Poetics and Semiotics.
Fleischer, Ezra. 1975. Hebrew Liturgical Poetry in the Middle Ages. Jerusalem: Keter.
Fortis, Jean-Michel. 2011. “On Localism in the History of Linguistics.” Workshop
on Localism—AFLiCo 2011. Available online: http://htl.linguist.
univ-paris-diderot.fr/fortis/Fortis_On%20localism.pdf.
Fowler, Roger. 1966. “‘Prose Rhythm’ and Metre.” In Essays on Style and Language, ed.
Roger Fowler, pp. 82–99. London: Routledge and Kegan Paul.
Frazer, Sir James. 1919. Folk-Lore in the Old Testament: Studies in Comparative
Religion, Legend and Law. London: Macmillan. Available online: https://www.
creationism.org/books/FrazerFolkloreOT/FrazerFolkloreOT_0.htm.
Frazer, Sir James. 1922. The Golden Bough, abridged, single-vol. ed.
New York: Macmillan. Available online: http://www.bartleby.com/196//.
Frenkel-Brunswik, Else. 1968 [1948]. “Intolerance of Ambiguity as an Emotional and
Perceptual Variable.” In Perception and Personality, ed. J. S. Bruner and D. Krech,
pp. 108–143. Westport: Greenwood Press.
Freud, Sigmund. 1953. Psychopathology of Everyday Life. New York: Mentor.
Fromkin, Victoria A. 1973. “Slips of the Tongue.” Scientific American 229 (6):
110–117.
Fry, D. B. 1958. “Experiments in the Perception of Speech.” Language and Speech
1: 126–151.
Gallese, Vittorio, and George Lakoff. 2005. “The Brain’s Concepts: The Role of the
Sensory-Motor System in Conceptual Knowledge.” Cognitive Neuropsychology
22: 455–479.
Gibbs, Raymond W., Jr., ed. 2012 [2008]. The Cambridge Handbook of Metaphor and
Thought. Cambridge: Cambridge University Press.
References [ 263 ]
4
6
2
Goldstein, K., and M. Scheerer. 1941. “Abstract and Concrete Behavior.” Psychological
Monographs 53 (2): 1–10.
Goldwasser, Orly. 2010. “How the Alphabet Was Born from Hieroglyphs.” Biblical
Archeological Review 36 (2): 40–53. Available online: http://members.bib-arch.
org/publication.asp?PubID=BSBA&Volume=36&Issue=02&ArticleID=06.
Golston, Chris, and Tomas Riad. 1997. “The Phonology of Classical Arabic Meter.”
Linguistics 35: 111–132. Available online: http://zimmer.csufresno.edu/~chrisg/
index_files/ArabicMeter.pdf.
Golston, Chris, and Tomas Riad. 2010. “Empirical Arabic Metrics: Implications for a
General Model.” Paper presented at “Metrical Studies Today: Topics, Tools and
Theories: An International Conference,” June 16–18, Halden, Norway.
Gombrich, E. H. 1954. Visual Metaphors of Value in Art. New York: Harper and Bros.
Gombrich, E. H. 1985. Meditations on a Hobby Horse. Chicago: University of
Chicago Press.
Granville-Barker, Harley. 1957 [1936]. Preface to Hamlet. New York: Hill and Wang.
Greene, Roland, ed. 2013. The Princeton Encyclopedia of Poetry and Poetics, 4th ed.
Princeton: Princeton University Press.
Halle, Morris, and Samuel Jay Keyser. 1966. “Chaucer and the Study of Prosody.”
College English 28: 187–219.
Halle, Morris, and Samuel Jay Keyser. 1971. English Stress: Its Form, Its Growth, and Its
Role in Verse. New York: Harper and Row.
Halpern, Martin. 1962. “On the Two Chief Metrical Modes in English.” PMLA
77: 177–186.
Halevy, Yehuda. 1994. “Treatise on Meter, appendix to Tova Rozen “The Strophic
movement in Yehuda Halevy’s Poetry—An Alternative Poetics?” In Essays
in Honour of Israel Levin, eds. Reuven Tsur and Tova Rozen, pp. 315–328. Tel
Aviv: Tel Aviv University (in Hebrew).
Halevy, Yehuda. 2009. The Kuzari: In Defense of the Despised Faith.
New York: Feldheim.
Hayes, Bruce. 1985. “Iambic and Trochaic Rhythm in Stress Rules.” In Proceedings of
BLS 11: Parasession on Poetics, Metrics, and Prosody, ed. N. Niepokuj, M. VanClay,
V. Nikiforidou, and D. Jeder, pp. 429–446. Berkeley: Berkeley Linguistic
Society.
Hayes, Bruce. 1995. Metrical Stress Theory: Principles and Case Studies.
Chicago: University of Chicago Press.
Herrnstein Smith, Barbara. 1968. Poetic Closure. Chicago: University of Chicago Press.
Hofstadter, Douglas R. 1997. Le Ton beau de Marot: In Praise of the Music of Language.
London: Bloomsbury.
Horace [Quintus Horatius Flaccus]. 1995. “The Art of Poetry (Ars Poetica).” Trans. Leon
Golden. In Horace for Students of Literature: The Ars Poetica and Its Tradition.
Gainesville: University Presses of Florida. Available online: http://www.baylor.
edu/theatre/index.php?id=89823.
Ibn Gabirol, Solomon. 1971. The Liturgical Poems. Ed. Dov Yarden. Jerusalem.
Jakobson, Roman. 1956. “Two Aspects of Language and Two Types of Aphasic
Disturbances.” In Fundamentals of Language, by Roman Jakobson and Morris
Halle, pp. 55–82. The Hague: Mouton.
Jakobson, Roman. 1960. “Closing Statement: Linguistics and Poetics.” In Style in
Language, ed. Thomas A. Sebeok, pp. 350–377. Cambridge: MIT Press.
Jakobson, Roman. 1968. Child Language, Aphasia, and Phonological Universals. The
Hague: Mouton.
[ 264 ] References

5
6
2
Jakobson, Roman. 1980. Brain and Language. Columbus, OH: Slavica Publishers.

Jakobson, Roman, C. Gunnar M. Fant, and Morris Halle. 1952. Preliminaries to Speech
Analysis: The Distinctive Features and Their Correlates. Cambridge: MIT Press.
James, William. 1902. The Varieties of Religious Experience: A Study in Human Nature.
New York: Longmans, Green.
Jespersen, Otto. 1960. Essentials of English Grammar. London: Allen and Unwin.
Jespersen, Otto. 2010 [1960]. “Notes on Metre.” In Selected Writings of Otto Jespersen,
pp. 344–363. Abington, UK: Routledge.
Johnson, Samuel. 1951 [1779]. “Abraham Cowley.” In The Great Critics, ed. James
Jones, Ernest. 1949. Hamlet and Oedipus. New York: Norton.
Kane, Julie. 2004. “Poetry as Right-Hemispheric Language.” Journal of Consciousness
Studies (Controversies in Science & the Humanities) 11/5–6: 21–59. Special
feature on “Consciousness and Literature,” ed. Roberta Tucker. Reproduced
in PsyArt (2007). Available online: http://psyartjournal.com/article/show/
kane-poetry_as_right_hemispheric_language.
Katz, Steven T. 1992. “Mystical Speech and Mystical Meaning.” In Mysticism and
Language, ed. Steven T. Katz, pp. 3–41. New York: Oxford University Press.
Kawasaki, Haruko. 1986. “Phonetic Explanation for Phonological Universals: The
Case of Distinctive Vowel Nasalization.” In Experimental Phonology, ed. John J.
Ohala and Jeri J. Jaeger, pp. 81–103. Orlando: Academic Press.
Kinnell, Galway. 1965. The Poems of François Villon. Boston: Houghton Mifflin.
Kinsley, James. 1969. The Oxford Book of Ballads. Oxford: Oxford University Press.
Klein, G. S. 1970. Perception, Motives and Personality. New York: Alfred A. Knopf.
Knight, G. Wilson. 1931. The Imperial Theme. London: Methuen.
Koffka, Kurt. 1935. Principles of Gestalt Psychology. New York: Harcourt, Brace and Co.
Köhler, Wolfgang. 1972. The Task of Gestalt Psychology. Princeton: Princeton
University Press.
Kris, Ernst. 1965. Psychoanalytic Explorations in Art. New York: Schocken.
Krueger, F. 1968 [1928]. “The Essence of Feeling.” In The Nature of Emotion, ed. Magda
B. Arnold, pp. 97–108. Harmondsworth, UK: Penguin.
Kubovy, Michael. 1981. “Concurrent Pitch-Segregation and the Theory of
Indispensable Attributes.” In Perceptual Organization, ed. M. Kubovy and J.
Pomerantz, pp. 55–98. Hillsdale, NJ: Lawrence Erlbaum and Associates.
Lakoff, George. 1993. “The Contemporary Theory of Metaphor.” In Thought and
Metaphor, 2nd ed., ed. Andrew Ortony, pp. 202–251. Cambridge: Cambridge
University Press.
Lakoff, George. 2012 [2008]. “The Neural Theory of Metaphor.” In The Cambridge
Handbook of Metaphor and Thought, ed. Raymond W. GibbsJr., pp. 17–39.
Cambridge: Cambridge University Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By. Chicago: University
of Chicago Press.
Leech, Geoffrey. 1974. Semantics. Harmondsworth, UK: Penguin.
Lever, J. W. 1966. The Elizabethan Love Sonnet. London: Methuen.
Levin, Israel. 1986. Mystical Trends in the Poetry of Solomon Ibn Gabirol.
Lod: Haberman Institute for Literary Research.
Liberman, Alvin M., Ignatius G. Mattingly, and Michael T. Turvey. 1972. “Language
Codes and Memory Codes.” In Coding Processes in Human Memory, ed. A. W.
Melton and E. Martin, pp. 307–344. New York: Winston.
Liebes, Yehuda. 2000. Ars Poetica in “Sefer Yetsira.” Tel Aviv: Schocken.
References [ 265 ]
62
Lloyd, Barbara B. 1972. Perception and Cognition. Harmondsworth, UK: Penguin.

Lowes, John Livingston. 1927. The Road to Xanadu: A Study in the Ways of Imagination.
Boston and New York: Houghton Mifflin Co., p. 343.
Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press.
Margolis, Joseph. 1962. “The Logic of Interpretation.” In Philosophy Looks at the
Arts: Contemporary Readings in Aesthetics, ed. Joseph Margolis, pp. 108–118.
New York: Charles Scribner’s Sons.
Maslow, Abraham H. 1948. “Higher and Lower Needs.” Journal of Psychology
25: 433–436.
Maslow, Abraham H. 1954. Motivation and Personality. New York: Harper.
Maslow, Abraham H. 1955. “Deficiency Motivation and Growth Motivation.”
In Nebraska Symposium on Motivation, ed. M. R. Jones, pp. 1–30.
Lincoln: University of Nebraska Press.
Maslow, Abraham H., ed. 1959. New Knowledge in Human Values. New York: Harper.
Maslow, Abraham H. 1962. Toward a Psychology of Being. New York: Wiley.
Maslow, Abraham H. 1969. The Healthy Personality: Readings. New York: Van
Nostrand.
Maslow, Abraham H. 1971. The Farther Reaches of Human Nature. New York: Viking.
Meyer, Leonard B. 1956. Emotion and Meaning in Music. Chicago: University of
Chicago Press.
Miller, George. 1956. “The Magical Number Seven, Plus or Minus Two: Some Limits
on Our Capacity for Processing Information.” Psychology Review 63: 81–97.
Available online: http://cogprints.org/730/.
Miller, George A., and Philip N. Johnson-Laird. 1976. Language and Perception.
Cambridge: Harvard University Press.
Millmore, Mark. N.d. “Egyptian Hieroglyphic Writing.” Available online: http://
discoveringegypt.com/egyptian-hieroglyphic-writing/.
Molière. N.d. The Miser (L'Avare). Charles Heron Wall (trans.) The Project
Gutenberg EBook.
More, Thomas. 1963. The Complete Works of St. Thomas More, vol. 2, The History of King
Richard III. Ed. Richard S. Sylvester. New Haven: Yale University Press.
Moshe ben Ya’akov ibn Ezra. 1975. Kitab al-Muhadara wal-Mudhakara (Liber
Discussionis et Commemorationis—Poetica Hebraica). Ed. and trans. A. S. Halkin.
Jerusalem: Sumptibus Societatis Mekize Nirdamim.
Muecke, D. C. 1970. Irony. London: Methuen.
Neisser, Ulric. 1968 [1963]. “The Multiplicity of Thought.” In Thinking and Reasoning,
ed. P. C. Wason and P. N. Johnson-Laird, pp. 307–323. Harmondsworth,
UK: Penguin.
Neisser, Ulric. 1976. Cognition and Reality. New York: Freeman.
Newberg, Andrew, Eugene D’Aquili, and Vince Rause. 2001. Why God Won’t Go
Away: Brain Science and the Biology of Belief. New York: Ballantine Books.
Norman, Donald A., and David E. Rumelhart. 1975. Explorations in Cognition. San
Francisco: W. H. Freeman.
Noy, Pinchas. 2009. “Where Do Melodies Come From?” PsyArt.
Available online: http://psyartjournal.com/article/show/
noy-where_do_melodies_come_from.
Ohala, John J. 1994. “The Frequency Code Underlies the Sound-Symbolic Use of Voice
Pitch.” In Sound Symbolism, ed. Leanne Hinton, Johanna Nichols, and John J.
Ohala, pp. 325–347. Cambridge: Cambridge University Press.
[ 266 ] References

7
6
2
Ornstein, Robert E. 1975. The Psychology of Consciousness. Harmondsworth,

UK: Penguin.
Otto, Rudolf. 1959. The Idea of the Holy. Trans. John W. Harvey. Harmondsworth,
UK: Penguin.
Pagis, Dan. 1970. Secular Poetry and Poetic Theory: Moses Ibn Ezra and His
Contemporaries (in Hebrew). Jerusalem: Bialik Institute.
Pagis, Dan. 1977. “Carmina Figurata in Pre-modern Hebrew Poetry” (in Hebrew).
Hasifrut 25: 13–27 (English summary: i–iii).
Pagis, Dan. 1993. “The Invention of the Hebrew Iambus, and Changes in Hebrew
Metrics in Italy” (in Hebrew). In Poetry Aptly Explained, pp. 166–255.
Jerusalem: Magnes Press, Hebrew University.
Pears, David. 1971. Wittgenstein. London: Fontana.
Persinger, Michael A. 1987. Neuropsychological Bases of God Beliefs. New York: Praeger.
Pinkernell, Gert. N.d. “Villon, François (1431–1463?): Sa vie et son œuvre.” Available
online: http://www.gert-pinkernell.de/villon/index.html.
Polányi, Michael. 1967. The Tacit Dimension. Garden City, NY: Anchor.
Quintilian. N.d. Institutio Oratoria. Trans. H. E. Butler. London: Loeb Classical
Library. Available online: http://penelope.uchicago.edu/Thayer/E/Roman/
Texts/ Quintilian/Institutio_atoria/home.html.
Radnóti, Miklós. 1943. Orpheus Nyomában. Budapest: Pharos.
Ransom, John Crowe. 1951. “Poetry: A Note in Ontology.” In The Great Critics, ed.
James Harry Smith and Edd Winfield Parks, pp. 769–787. New York: Norton.
Revithiadou, Anthi. 2004. “The Iambic/Trochaic Law Revisited: Lengthening and
Shortening in Trochaic Systems.” Leiden Papers in Linguistics 1 (1): 37–62.
Available online: http://www.lit.auth.gr/revithiadou/files/ITLRevisited.pdf.
Rice, Curt. 1992. “Binarity and Ternarity in Metrical Theory: Parametric Extensions.”
Unpublished Ph.D. diss., University of Texas, Austin..
Rosen, Tova. 1994. “The Strophic Trend in Yehuda Halevi’s Poetry—An Alternative
Poetics?” (in Hebrew, with English abstract). In Israel Levin Jubilee Volume I, ed.
Reuven Tsur and Tova Rosen, pp. 315–328. Tel Aviv: Katz Research Institute
for Hebrew Literature.
Rumelhart, David E. 1979. “Some Problems with the Notion of Literal Meaning.” In
Metaphor and Thought, ed. Andrew Ortony, pp. 78–90. Cambridge: Cambridge
University Press.
Sacks, Oliver. 2012. “Seeing God in the Third Millennium: How the Brain Creates
Out-of-Body Experiences and Religious Epiphanies.” The Atlantic, December
12. Available online: http://www.theatlantic.com/health/archive/2012/12/
seeing-god-in-the-third-millenium/266134/.
Schank, Roger C. 1973. “Identification of Conceptualizations Underlying Natural
Language.” In Computer Models of Thought and Language, ed. Roger C. Schank
and Kenneth Mark Colby, pp. 187–247. San Francisco: W. H. Freeman.
Schank, Roger C., and Robert Abelson. 1977. Scripts, Plans, Goals and
Understanding: An Inquiry into Human Knowledge Structures. Hillsdale,
NJ: Lawrence Erlbaum Associates.
Schirmann, Jefim. 1995. The History of Hebrew Poetry in Muslim Spain (in Hebrew).
Jerusalem: Magnes Press, Hebrew University, and the Ben-Zvi Institute.
Scholem, Gershom G. 1960. Jewish Gnosticism, Merkabah Mysticism, and Talmudic
Tradition. New York: Jewish Theological Seminary of America.
Scholem, Gershom G. 1961. Major Trends in Jewish Mysticism. New York:
Schocken Books.
References [ 267 ]
8
62
Schramm, Wilbur L. 1935. Approaches to the Science of English Verse. Iowa City: Iowa
University Press.
Scott, Clive. 2011. “Free Verse and the Translation of Rhythm.” Thinking Verse
1: 67–101.
Shoshany, Ronit. 2009. “The Original Purpose of Biblical Accentuation” (in Hebrew).
Masʾat Aharon: Linguistic Studies Presented to Aron Dotan, ed. Moshe Bar-Asher
and Chaim E. Cohen, pp. 469–486. Jerusalem: Bialik Institute.
Siciliano, Italo. 1934. François Villon et les thèmes poétiques du Moyen Age.
Paris: Librarie Armand Colin.
Skoyles, John. 1985. “Did Ancient People Read with Their Right Hemispheres?
A Study in Neuropalaeographology.” New Ideas in Psychology 3: 243–252.
Skoyles, John. 1988. “Right Hemisphere Literacy.” In The Alphabet and the Brain: The
Lateralisation of Writing, ed. Derrick de Kerchhove and Charles J. Lumsden,
pp. 362–380. Berlin: Springer-Verlag.
Skoyles, John. 1997. “Ancient Readers: Were They Right Hemisphere Literates?”
Available online: http://www.users.globalnet.co.uk/~skoyles/index.htm.
Smith, Hallett. 1968. Elizabethan Poetry: A Study in Conventions, Meaning, and
Expression. Ann Arbor: University of Michigan Press.
Smith, James. 1933. “Metaphysical Poetry.” Scrutiny 2: 222–238.
Snyder, E. D. 1930. Hypnotic Poetry: A Study of Trance-Inducing Techniques in
Certain Poems and Its Literary Significance. Philadelphia: University of
Pennsylvania Press.
Sternberg, Meir. 1976. “Temporal Ordering, Modes of Expositional Distribution, and 3
Models of Rhetorical Control in the Narrative Test—Faulkner, Balzac and Austen.”
PTL: A Journal for Descriptive Poetics and Theory of Literature 1 (2): 295–316.
Stevenson, Robert Louis. 1882. Familiar Studies of Men and Books. Available
online: http://www.fullbooks.com/Familiar-Studies-of-Men--Books.html.
Stroop, Ridley J. 1935. “Studies of Interference in Serial Verbal Reactions.” Journal of
Experimental Psychology 18: 643–662. Available online: http://psychclassics.
yorku.ca/Stroop/.
Sidney, Sir Philip. “The Defense of Poesy.” In English Essays: Sidney to Macaulay. The
Harvard Classics. 1909–14. http://www.bartleby.com/27/1.html.
Sypher, Wylie. 1955. Four Stages of Renaissance Style. Garden City, NY: Anchor.
Szerb, Antal. 1943. A Világirodalom Története. Budapest: Révai.
Taylor, Jane H. M. 2001. The Poetry of François Villon: Text and Context.
Cambridge: Cambridge University Press.
Thomson, Philip. 1972. The Grotesque. London: Methuen.
Tsur, Reuven. 1969. Studies in Mediaeval Hebrew Poetry (in Hebrew). Tel Aviv: Daga.
Tsur, Reuven. 1972. “Articulateness and Requiredness in Iambic Verse.” Style 6: 123–148.
Tsur, Reuven. 1974. “Poem, Prayer and Meditation: An Exercise in Literary
Semantics.” Style 8: 405–425.
Tsur, Reuven. 1975. “Two Critical Attitudes: Quest for Certitude and Negative
Capability.” College English 36: 776–788.
Tsur, Reuven. 1977a. A Perception-Oriented Theory of Metre. Tel Aviv: Porter Institute
for Poetics and Semiotics.
Tsur, Reuven. 1977b. “Some Remarks on the Nature of Trochees and Iambs and
Their Relationship to Other Metres.” In A Perception-Oriented Theory of Metre,
pp. 83–96. Tel Aviv: Porter Institute for Poetics and Semiotics. Available
online: http://www.tau.ac.il/~tsurxx/Trochee/bit_of_personal_history2.html.
[ 268 ] References

9
6
2
Tsur, Reuven. 1979. “Levels of Information Processing in Reading Poetry.” Critical

Inquiry 5: 751–759.
Tsur, Reuven. 1983. What Is Cognitive Poetics? Tel Aviv: Katz Research Institute for
Hebrew Literature.
Tsur, Reuven. 1987. On Metaphoring. Jerusalem: Israel Science Publishers.
Tsur, Reuven. 1992a. What Makes Sound Patterns Expressive? The Poetic Mode of Speech
Perception. Durham, NC: Duke University Press.
Tsur, Reuven. 1992b. Toward a Theory of Cognitive Poetics. Amsterdam: Elsevier (North
Holland) Science Publishers.
Tsur, Reuven. 1997. “Picture Poems: Some Cognitive and Aesthetic Principles.”
PsyArt. Available online: http://psyartjournal.com/article/show/
tsur-picture_poems_some_cognitive_and_aesthet.
Tsur, Reuven. 1998. Poetic Rhythm: Structure and Performance—An Empirical Study in
Cognitive Poetics. Bern: Peter Lang.
Tsur, Reuven. 2000a. “Picture Poetry, Mannerism, and Sign Relationships.” Poetics
Today 24: 751–781.
Tsur, Reuven. 2000b. “Lakoff’s Roads Not Taken.” Pragmatics and Cognition
7: 339–359.
Tsur, Reuven. 2002a. “Some Cognitive Foundations of ‘Cultural Programs.’” Poetics
Today 23 (1): 63–89. Available online: http://www.tau.ac.il/~tsurxx/all_
cohaerance_gone_3a.html.
Tsur, Reuven. 2002b. “Some Comments on the Lakoffean Conception of Spatial
Metaphor.” “Metaphor: A Multidisciplinary Approach,” special issue, ed.
Thomasz Komendzinski, Theoria et Historia Scientiarum 6: 245–267.
Tsur, Reuven. 2003. On the Shore of Nothingness: Space, Rhythm, and Semantic
Structure in Religious Poetry and Its Mystic-Secular Counterpart—A Study in
Cognitive Poetics. Exeter: Imprint Academic.
Tsur, Reuven. 2006a. “Size-Sound Symbolism Revisited.” Journal of Pragmatics
38: 905–924.
Tsur, Reuven. 2006b. “ ‘I Shall Be Right Indeed’ by Shlomo Ibn Gabirol: Rhythmic
and Grotesque Effects—Is the Past Homogeneous?” (in Hebrew). Criticism
and Interpretation. Journal for Interdisciplinary Studies in Literature and Culture
39: 43–72.
Tsur, Reuven. 2006c. “Kubla Khan”—Poetic Structure, Hypnotic Quality and Cognitive
Style: A Study in Mental, Vocal, and Critical Performance. Amsterdam: John
Benjamins.
Tsur, Reuven. 2008a. Toward a Theory of Cognitive Poetics, 2nd ed. Brighton,
U.K.: Sussex Academic Press.
Tsur, Reuven. 2008b. “Two Medieval Hebrew Devotional Poems—Convention,
Evaluation, and ‘Platonic’ vs. ‘Metaphysical’ Poetry.” Style 41: 434–444.
Tsur, Reuven. 2008c. “Poetic Conventions as Fossilized Cognitive
Devices; The Case of Mediaeval and Renaissance Poetics.”
PsyArt. Available online: http://psyartjournal.com/article/show/
tsur-poetic_conventions_as_fossilized_cogniti.
Tsur, Reuven. 2010. “Poetic Conventions as Cognitive Fossils.” Style 44: 496–523.
Tsur, Reuven. 2011. “Aesthetic Qualities as Structural Resemblance: Divergence
and Perceptual Forces in Poetry.” In Semblance and Signification, ed.
Pascal Michelucci, Olga Fischer, and Christina Ljungberg, pp. 233–250.
Amsterdam: John Benjamins.
References [ 269 ]
0
72
Tsur, Reuven. 2012a. Poetic Rhythm: Structure and Performance—An Empirical Study in

Cognitive Poetics, 2nd ed. Brighton, UK: Sussex Academic Press.
Tsur, Reuven. 2012b. “The Neurological Fallacy.” Pragmatics and Cognition
19: 420–446.
Tsur, Reuven. 2012c. Playing by Ear and the Tip of the Tongue: Precategorial Information
in Poetry. Amsterdam: John Benjamins.
Tsur, Reuven. 2013. “Masculine and Feminine Rhymes: Their Structural Effect.” Style
47: 1–24.
Tsur, Reuven. Forthcoming. “Elusive Qualities in Poetry, Receptivity, and Neural
Correlates.” In Expressive Minds and Artistic Creations: Studies in Cognitive
Poetics, Cognition and Poetics, ed. Szilvia Csábi. New York: Oxford
University Press.
Tsur, Reuven, and Motti Benari. 2002. “‘Composition of Place,’ Experiential Set,
and the Meditative Poem (A Cognitive-Pragmatic Approach).” Pragmatics and
Cognition 9: 203–237.
Tsur, Reuven, and Yehosheva Bentov. 1996. “Rhythmic and Strophic Organization
in Mediaeval Hebrew Poetry (A Cognitive Approach).” Empirical Studies of the
Arts 14: 183–206. Available online: http://www.tau.ac.il/~tsurxx/Tsur_Bentov_
abstract.html.
Ullmann, Stephen. 1957. The Principles of Semantics. Oxford: Blackwell.
Ultan, Russell. 1978. “Size-Sound Symbolism.” In Universals of Human Language, vol.
2, Phonology, ed. Joseph H. Greenberg, pp. 527–568. Stanford, CA: Stanford
University Press.
Wagshal-Te’eni, Sivan, and Reuven Tsur. Forthcoming. “The Dao Flickering through
Words.” In Mysticism: Twenty-First Century Perspectives, ed. Alex Kohav.
Weitz, Morris. 1962. “The Role of Theory in Aesthetics.” In Philosophy Looks at the
Arts, ed. Joseph Margolis, pp. 48–59. New York: Scribner.
Wellek, René, and Austin Warren. 1956. “Euphony, Rhythm, and Meter.” In Theory of
Literature, pp. 159–176. New York: Harcourt, Brace.
Wimsatt, W. K., Jr. 1954. “The Structure of Romantic Nature Imagery.” In The Verbal
Icon, pp. 103–116. New York: Noonday.
Wimsatt, W. K., Jr., and Monroe C. Beardsley. 1959. “The Concept of Meter: An
Exercise in Abstraction.” PMLA 74: 585–598.
Woodrow, Herbert. 1909. “A Quantitative Study of Rhythm.” Archives of Psychology
14: 1–66.
Woodrow, Herbert. 1911. “The Role of Pitch in Rhythm.” Psychological Review
18: 54–77.
Woodrow, Herbert. 1951. “Time Perception.” In Handbook of Experimental Psychology,
ed. S. S. Stevens, pp. 1224–1236. New York: Wiley.
Zhirmunsky, V. 1966. Introduction to Metrics. Trans. C. F. Brown. The Hague: Mouton.
[ 270 ] References
271
INDEX
Edward, 4, 6, 92, 223 anagnorisis, 91

grammatical and antigrammatical, 240 anagram, 133, 248–251
major and minor chords, 196 anapest, 50, 172–174, 182–183, 211,
217, 228
A Midsummer Night’s Dream, 4, 37 antigrammatical rhyme, 241
A Midsummer Night’s Dream, 57–58, 60, apostrophe, 90–91, 181
63, 68 Arany, János, 163
Abulafia, Abraham, 115–117, 129, arbitrary genitive, 145–147, 149, 153
249–250 Archilochus, 209, 226
acrostic, 108, 110–114, 117–118, 128–129, Aristotle, 9, 50, 96, 172, 209, 216,
248–249 226, 229
adaptation, 1, 6–7, 20, 33, 42, 53, Arnheim, Rudolf, 16–17, 22, 25
131–132, 150, 153–154, 248 Arnold, Matthew, 145
Addison, Joseph, 133–135, 139, 251 artificial intelligence, 31–33, 231,
(Abu at-Tayyib Ahmad ibn al-Husayn 254–255, 257
al-Mutanabbi al-Kindi), 57–60, Ashkenazi Jewish cantors, 18
63, 84
Alcibiades, 96 Babits, Mihály, 158
alexandrine, 25, 29, 158, 171–174, 178, Babylonian system, 53
182–183, 185, 228 Bacchi, Roberto, 233–236
alliteration, 11, 111, 113, 134, 164, Balaban, Victor, 107
184–185, 221, 235, 250 ballad, 4, 6, 15–16, 39, 51, 66, 87–88,
allusion, 110, 142 90–94, 96–98, 223
alphabet, 42–44, 46–47, 115 ballad stanza, 51, 223
altered state of consciousness, 102 ballade, 61, 71, 75–80, 82–83
Alterman, Nathan, 228 Bartlett, E.C., 31–32, 34–36, 40–41, 47,
ambiguity, 16–17, 88, 22, 66, 93–96, 55, 84–85, 150, 256
120–121, 125, 128, 173, 134, Baudelaire, Charles, 104, 106, 171, 173,
144–145, 193, 222–223, 227 175, 182, 228
ambivalence, 20–21, 66–67, 69–70, 85, Beardsley, Monroe C., 10, 100, 225, 232,
94, 144 241, 243
Amichai, Yehuda, 120 Beaver, Joseph, 212–216, 220, 223
amphibrach, 173–174, 228 Beer, J.B., 115
anadiplosis (concatenation), 16, 108, Beethoven, Ludwig van, 187
110, 118, 120–121, 123, 125–126, Bergson, Henri, 100, 129, 237
128–129 Berliner, Paul, 187
272
Bernstein, Shim’on, 60 cognitive anthropology, 5, 19, 63, 154

Bialik, Ḥayim Naḥman, 147, 149, 153 cognitive-fossils, 1–2, 20, 53, 55, 206
biblical cantillations, 52 coinciding downbeats, 176, 211
biographical fallacy, 59–60, 83–84 Coleridge, Samuel Taylor, 115, 199, 212,
Blechner, Mark, 196 229, 235, 248–249, 251–252, 257
boasting poems, 59, 61 colliteration, 78, 235
Book of Yetsira, 115 combination of letters, 114–118,
boundaries between self and not-self, 129, 249
101–102 communicative competence, 137
Bownds, Deric, 7 componential analysis, 32
Bradburn, Beth, 27 conceit, 54, 63, 70, 76, 132, 139–140,
Brain’s letterbox, 42–45, 48, 248 142, 151–155
Brann, Ross, 144–145 conceptual dependency, 32, 254
Brassens, Georges, 61 concrete of the abstract, 145–146, 153
Brill, Abraham A., 252–253, 255–257 constraints of the human brain, 5, 8,
broken telephone, 34, 55 19–20, 26–28, 36–37, 39, 42,
Brooke-Rose, Christine, 146 46, 49–51, 54, 67, 69, 84–85,
Brooks, Cleanth and Robert B. Heilman, 92 154, 157
Brown (iambic and trochaic nonsense constraints-seeking, 1–2, 41, 53, 55
lines), 219, 222 contrarieties, 62, 64, 70, 75, 77, 85
Brown, Roger and David McNeill, contrarious passions, 64–65, 67–68
244, 247 convergent and divergent poetry, 11
Brown, Roger, 30–31, 189–190, 206, Cooper, W. C. and L.B. Meyer, 13, 24
233, 244–249 Cooper William E. and John Robert
Browning, Robert, 216, 221–223 Ross, 12
bundles of features, 202, 205 cortical competition, 47–49
bundles of semantic features, 241, 247 counterpoint, 211, 226
Burke, Kenneth, 22, 78, 235–236, 248 Cowley, Abraham, 54, 133, 154
creative ego rhythm, 101
caesura, 9–14, 22–23, 22–25, 29–30, 51, critical philosophy, 150–151
172–173, 178–180, 182, 185, 217, cross-modal mapping, 190, 193,
223–224 198–200
calligrams, 134–135 crucial recommendation, 6, 160, 181
Carey, Henry, 220 Culler, Jonathan, 87, 175
Carroll, Lewis, 83 cultural artifacts, 27–28, 47, 49
catalog, 20–21, 64, 67–70, 73, 75, 80, cultural studies, 5, 7, 153, 191
82–83, 85, 105, 111, 113 culture begets culture, 1, 2, 5–6, 9–10,
catalog of contradiction, 83, 85 14, 153, 258
catalog of paradoxes, 20–21, 69–70, 73 Curtius, Ernst Robert, 143
categorial perception, 195–197
cathexis, 69, 250 D’Andrade, Roy, 5, 8, 19, 27, 37, 41–43,
Catullus, Gaius Valerius, 65 50, 63, 67, 84, 154
Charles D’Orléans, 20, 74–75 dactyl, 25, 49–50, 54, 172, 174, 183, 211,
Chartier, Alain, 78–79, 82–83 217, 228
Chatman, Seymour, 13, 172, 211, 213, Danse Macabre, 73
216–219, 227 Dante, Alighieri, 29
Chinese whispers, 34, 36 Dao de Jing, 46
Clark, Herbert and Eve V. Clark, Darwin, Charles, 3
192–193, 200 Dayan, Peter, 180–181
classicist, 9, 138, 168 de Mourgues, Odette, 88
[ 272 ] Index
273
defense mechanism, 14, 66, 85, 101 Fowler, Roger, 174, 183, 226–227
Dehaene, Stanislas, 28, 42–49, 115, 248, Frazer, Sir James, 7–8
258–259 Frenkel-Brunswik, Else, 96
depth-psychological, 2, 5, 20, 232–233 Freud, Sigmund, 20–21, 66, 69, 94, 101,
devotional poetry, 62, 127 231–234, 239, 244, 248, 252–253,
Dewey, John, 163 257–258
distinctive features, 164, 169, 183, 185, Fromkin, Victoria, 235–238, 245, 249, 257
233, 235–239, 241–243 Fry, D. B., 26, 218
Dolman, John, 177 fusiform, 48
Donne, John, 20, 62, 134, 142–143, 151,
154–155, 210, 227 Gallese, Vittorio, 200–201
double syntax, 122, 128 garden descriptions, 59, 145
Drayton, Michael, 20, 70, 73–75, 77 generation by rule, 254
Dresher, Elan B., 52 generative metrists, 172, 183–184
drinking poems, 59 generic recall, 245, 247
Dunash Ben Labrat, 28 genre, 16, 52, 57, 59, 64, 84, 108
Gestalt, 13, 22, 24, 30, 40, 100–102, 104,
echoic memory, 194 107, 128, 149, 172, 179, 211, 216,
Ehrenzweig, Anton, 3, 5, 14, 16–18, 63, 222, 227, 229, 246–247
66, 84–85, 87, 100–102, 105, 126, gestalt-free, 100, 104, 107, 149, 246
128, 232, 243, 247 Gestaltist, 10, 22, 100–101
Einat-Nov, Idit, 99 Gil, David, 52
Eitan,Zohar and Renée Timmers, glissandi, 16, 18
188–191, 206–207 Goethe, Johann Wolfgang von, 75
electrical perturbations, 117–118, 129 Goldstein K. and M. Scheerer, 96
elegant solution to a problem, 157, 160, Goldwasser, Orly, 44–46
163, 165–166, 174, 184 Golston, Krus and Tomas Riad, 26
Elior, Rachel, 101, 105, 111, 129 Gombrich, Ernst Hans, 3, 17–18
Elizabethan drama, 4, 36–37, 40 Good Continuation, 222
ellipsis, 15, 31, 88 Granville-Barker, Harley, 36, 39
Emanuel of Rome, 29 graphic primitives, 47–48
Empson, William, 122 grounding, 190, 200, 206
erotic symbolism, 14 grouping, 13, 26, 172, 174–176, 210–211,
Eskimo story, 33–34, 40 214–215, 218–219, 221–225, 229
Even-Shoshan, Abraham, 124
Even-Zohar, Itamar, 7 Halle, Morris and Samuel J. Keyser, 9,
event time and micro-time, 198 212–215, 218, 220, 229
evolution, 7, 14, 42–43, 47, 55, 84, 154 Halpern, Martin, 211, 216
evolutionary tree, 7, 55, 154 Hamlet, 21, 36–40, 159, 178, 185
Ezekiel, 103, 110 Harvey, William, 57–58
Hayes, Bruce, 218
fairy tales, 20, 66, 85 Hekhalot, 103, 105, 108, 110–111, 129
false wit, 133–134, 153 Herbert, George, 38, 134, 249–252
Father William, 83 Herrnstein Smith, Barbara, 22, 224
feature reversal, 239 hieroglyphs, 44–47
feature-deletion, 206, 247 Hofstadter, Douglas, 160, 173, 180–182
Feldman, Jerome, 200 Holy Living Creatures, 99, 103, 110
Fleischer, Ezra, 108–110 hominid, 7
Folk-Lore in the Old Testament, 7 Homo erectus, 7
Fortis, Jean-Michel, 200 Homo sapiens, 7, 115
I n de x [ 273 ]
274
Hopkins, John, 226, 250, 257 kana, 43, 47

Horace (Quintus Horatius Flaccus), Kane, Julie, 38
209, 226 Kant, Immanuel, 151, 170
hyperbole, 60, 64, 87, 139 Katona, George, 30–31
hypnotic poetry, 106 Katz, Steven T., 115–116, 152
Kawasaki, 170
iamb, 9–11, 13–14, 22–27, 29, 49–54, Keats, John, 96, 104, 212, 231, 244–245,
158, 171–174, 177–180, 182–183, 247, 252–258
185, 209–229, 256 Kékulé, August, 239
(Ala-al-din abu Al-Hassan Ali ibn Abi- Kim and Iwamiya, 203
Hazm al-Qarshi al-Dimashqi), 58 King Lear, 178
illustrative catalogue, 104 King of Khazars, 144
immediate memory, 30, 36 Kinnell, Galway, 72
immediate signals, 38 Kinsley, James, 90
immediately observable constituent, Klein, George, 15, 88
25, 55 Kleinman, David, 73
incomplete fossilization, 201, 205 knowledge untaught and unlearned, 3,
infinite regress, 5 20, 258–259
influence-hunting, 1–2, 7, 20, 53 Koffka, Kurt, 22
information processing, 4, 6, 41, 133, Köhler, Wolfgang, 234, 238–239, 257
248, 256 Krueger, E., 161
inlay language, 136–137 Kubovy, Michael, 198
insight, 8–9, 87–88, 101, 152, 157, 232,
234, 238–239, 253, 257 Lakoff, George, 188–190, 192–194,
interjections, 4, 15–16, 36–40, 91 198–201, 206–207
interlingua, 254–255 Lakoff, George and Mark Johnson, 188
intersense transfer, 246 Laozi, 84
intolerance of ambiguity, 96 Lascaux cave paintings, 45
irony, 87, 241 Lateralization, 38–39
Isaiah, 94, 114, 142 Law of Parsimony, 22
Itshak Ibn Ghiyat, 102, 118, 120–121, Leech, Geoffrey, 12
126–128 left hemisphere, 38, 40, 43, 48
Izmailov, Alexandr, 160 leveling, 15–17, 64, 67, 88
leveling and sharpening, 16, 64, 67
Jabotinsky, Ze’ev, 167–168, 170 Lever, J. W., 61
Jackson, John Hughlings, 38, 40 Levin, Israel, 108
Jakobson, Roman, 12, 15, 23–24, 38, Levy Ibn Altaban, 145
78, 132–133, 164, 167, 169, 235, Liberman, Alvin A., Ignatius G. Mattingly,
240, 251 and Michael T. Turvey, 255
Jakobson, Roman C., Gunar M. Fant, Liebes, Yehuda, 115
and Morris Halle, 164 light emanation, 140
James, William, 7, 101–102, 139, limited channel capacity, 21, 23–24, 36, 40
231–232, 243, 245, 247 litotes, 87
Jeremiah, 128 little green man, 234
Jespersen, Otto, 212, 232, 243 liturgical poetry, 16, 51–52, 100, 108,
jigsaw puzzle, 238–239 114, 145
Johnson, Dr. Samuel, 54, 139, 148, 154 Lloyd, Barbara B., 37
joke, 132, 151, 239, 241 localist, 199–200
Jones, Ernest, 21 long-term memory, 32, 233, 254–255
Jouvet, Michel, 258–259 Longfellow, Henry Wadsworth, 212–214
[ 274 ] Index
275
love poetry, 20, 59, 63, 139, 153 More, Sir Thomas, 177
Lowes, John Livingston, 239 Morgenstern, Christian, 152
Luchins, 76 Moses Ibn Ezra, 59–60, 84, 145
Lyons, John, 199, 204 Mozart, Leopold, 162
Lyotard, Jean-François, 180 Muecke, D. C., 237
multiple relationship, 241
magical number 7 ± 2, 13, 22–23, 179, 217 multivalence, 241
mannerism, 18, 131, 133–135, 138, 143, muwwashaḥ, 59
151, 153 mysterium tremendum, 108, 129
markedness, 9–12, 43, 51, 100, 133, 143,
150, 153, 192, 199–200, 205, 214, natural selection, 28, 40, 49, 54
222, 235, 251 Neanderthals, 7
Marranos, 4 Neapolitan songs, 18
Maslow, Abraham H., 57, 63–64, 68, 84 Neisser, Ulric, 21, 150, 194
Mastroiani, Marcello, 83 Neoplatonic, 140, 144
mbira, 187–188 Neural Theory of Metaphor, 200, 207
McNeill, David, 244, 247 neurological fallacy, 207
mediated association, 187, 189 neuronal recycling, 42, 46–47
meditative catalogue, 104, 182 neurons, 43, 46–47, 200–201
Merkabah, 99–100, 102–103, 105–106 Newberg, Andre, Eugene D’Aquili, and
Merkabah hymn, 99–100, 102, 105 Vincent Rause, 116
Merkabah mysticism, 99, 102 noncategorial sounds and recoding, 194
meta-awareness, 150 Norman, Donald. A., and David E.
metaphor, 54, 61, 75, 104, 137, 139, 146, Rumelhart, 32
154, 183, 188–191, 200, 202–203, Noy, Pinchas, 234
205–207, 232, 247, 259 Numinous Hymns, 104
Metaphysical, 54, 100, 105, 129, 132, nursery rhymes, 212, 221
134–135, 138–140, 142–143, 145,
149–155 Occam’s Razor, 22
Metaphysical conceit, 54, 132, 139–140, oceanic dedifferentiation
142, 151, 153 (undifferentiation), 101–102,
metaphysical intuition, 100 106–107
Metaphysical style, 135, 143 Oedipal situation, 93, 96
metaphysical wit, 134, 145, 149, 152 Oedipus complex, 96–97
metonymy, 80, 91, 93, 110, 138, 143, 155, Oedipus Rex, 62
165, 169, 259 Ohala, John, 204
metrical set, 210, 222 onomatopoeia, 161
Meyer, Leonard B., 13, 24, 26, 173, 211, ophanim, 108
217–218 organized violence against language,
migration, 1–2, 19–20, 53–55 231, 258
Miller, George A., and Philip N. orientation mechanism, 103, 107, 150
Johnson-Laird, 150 origins of language, 3
Miller, George, 13, 22, 179 ornament, 3, 5, 14, 16–18, 63–64,
Milner, Marion, 101 85, 87, 101–102, 108, 118,
Milton, John, 11–12, 106, 115, 143, 176, 128–129, 135
180, 226 Ornstein, Robert E., 38
mirror neurons, 200 Otto, Rudolf, 104, 108, 129, 232
Mirrour for Magistrates, 61, 177 overarticulation, 210–211, 214, 221–222
misquotation, 231, 252, 254–256, 258 overstressing, 210, 214, 221
Molière, Jean-Baptiste Poquelin, 237 oxymoron, 87
I n de x [ 275 ]
276
Pagis, Dan, 29, 60, 83, 135, 233–236 Psychopathology of Everyday Life, 231,
panchronistic tendencies in 252, 257
synesthesia, 246 Ptolemaic world picture, 143
panegyrics, 58–60 Pure Attribution, 146, 153
Paradise Lost, 11, 106, 143, 180
paradox, 20–21, 69–70, 73–76, 80, qassida, 59
86–87, 125, 152 Quintilian, Marcus Fabius, 97
parody, 37, 78, 80, 82–83, 86, 220
paronomasia, 137 Racine, Jean, 173
parsimony, 21–23, 31–32, 198, 206 Radnóti, Miklós, 27
passing notes, 18 rahitim, 108
Pears, David, 151 Ransom, John Crowe, 137
perception, 2, 4, 6, 16, 22, 26, 30–31, 34, rapid closure, 96
37–40, 43, 69, 93, 96, 125, 135, regional quality, 10, 99–100
139, 146–147, 149, 151, 159, 177, repeated social transmission, 1, 4–5, 8,
187, 189, 195–199, 203, 206, 209, 19–21, 26–27, 30, 34, 36, 39–43,
218, 222, 229, 232, 252, 256 45, 47–49, 53–55, 63–64, 67, 84,
peripeteia, 91 154, 259
Persinger, Michael, 117 retrieval, 233–234, 244–245
personal creator, 140, 144 Revithiadou, Anthi, 218
personality style, 6 rhyme, 29, 68, 78–80, 82–83, 110–111,
Petrarch, Francesco, 20, 64, 67–69, 75, 113–114, 132–134, 158, 165,
138, 154 168–169, 172, 177, 182, 184–185,
Philips, Ambrose, 220 212, 220–223, 239–242, 255
Phoenician script, 45–46 Rice, Curt, 26, 218
phonetic code, 194 Richard III, 174, 176
phonetic recoding, 195 riddles, 20, 68–69
pigeonholing, 87–88, 96–97 right hemisphere, 15, 38, 40, 48, 67
Pinkernell, Gert, 82–83 Rimbaud, Arthur, 104
pitch ascent, 204 Romanticism, 11, 59, 78, 104, 134–135,
pitch descent, 204 138, 147, 149–150, 153, 182, 232,
Platonic, 106, 135, 137–138, 141, 243, 246, 251–252, 257
144–147, 149 Ronsard, Pierre de, 20, 64–65, 69
Platonic dialogue, 144 Rorschach inkblot test, 128
Platonic poetry, 137 Rosen, Tova, 144
Poems of Complaints, 60 Rumelhart, David E., 31–32
Poems of Contemplation, 59 Russian school of verse study, 8
Polányi, Michael, 47
Pope, Alexander, 11–12, 154, 176, Sacks, Oliver., 117
226–227 saturation, 111, 228
possible world, 115, 159 Sayf al-Dawlah, 58–59
Précieux, 135, 138, 145 Schank, Roger, 31–33, 254
primacy and recency effects, 76 Schank and Abelson, 31–33
Principle of Marginal Control, 47 Scheindlin, Ray, 144
principle of maximum contrast, 198 Schema, 25, 30–36, 40–41, 47, 55, 68,
prosody, 9, 49, 59, 144, 210, 228 85, 150, 210, 256
Proto-Sinaitic, 45–47 Schirmann, Jefim, 28–29
proverbs, 78, 82–83 Scholem, Gershom G., 100, 102–105,
psychic distance, 18 107, 115–116, 249, 251
psychological reality, 234–236, 245 Schramm, Wilbur, 218
[ 276 ] Index
277
Scott, Clive, 157–160, 164–165, 172, static and dynamic sound stimuli, 202
174–176, 180, 182–184 Sternberg, Meir, 76–77
scrambling, 248, 250–252, 257 Stevenson, Robert Louis, 75
script, 32–33, 45–46 stock phrases, 108, 110
secondary elaboration, 5, 14 stress maxima in weak positions, 11
self-specifying information, 103, 107 stress maximum, 11, 212–215, 11, 212–216
semantic features, 111–112, 233, stress-timed language, 54
240–241, 244–248 Stroop, Ridley, 67–68, 203
semantic primitives, 254 Stroop effect, 67, 203
Serbian folk epic, 23, 49 sublime, 99, 105
Shakespeare, William, 4, 11, 37, 61, 154, Sumerian-Babylonian-biblical influences, 8
174, 176–177, 212, 226 syllable-timed language, 28, 54
sharpened, 16–17, 21, 69–70 syllepsis, 87
Shekhinah, 103–104 Symbolist, 104, 170, 182, 232, 243, 257
Shelley, Percy, Bysshe, 10, 123, 158, 176, syncopation, 226, 228
222–225 Sypher, Wylie, 131, 135, 143, 151
Shlomo Ibn Gabirol, 60, 102, 108, 110, Szabó, Lőrinc, 165, 167–169
112, 116, 118, 121, 123, 125–128, Szerb, Antal, 242
136, 138, 140–145, 152, 154–155
Shlonsky, Abraham, 146, 178 Taylor, Jane, 78, 82–83
Shmuel Hanagid, 136 temporal lobe transients, 117
short-term memory, 12–13, 22, 24, 30, Tennyson, Lord Alfred, 102, 223
52, 179, 217 The Golden Bough, 7
Shoshany, Ronit, 1, 52 thematized predicate, 104
Siciliano, Italo, 61 Theurgy, 115
Sidney, Sir Philip, 62 thing-destruction, 104, 247
sign relationships, 131, 251 thing-free, 100, 104, 149, 181–182, 247
significant variation, 92 Thomson, Philip, 152
signified, 114–115, 132–138, 151, 184, Tiberian system, 53
251–252 tip-of-the-tongue, 231–232, 247, 257
signifier, 114, 132–138, 151, 184, 251 topicalized attribute, 104
silukim, 108 tot experience, 243, 245
Silvestre, Armand, 173, 181 tot experiment, 232
slip-of-the-tongue, 231–232, 247, 257 tot phenomenon, 231, 243
Smith, Hallett, 22, 61–62, 69, 139, 224 Tóth, Árpád, 158–159, 165, 167–169
Smith, James, 139 translation, 12, 27, 53, 59, 65–66, 73–74,
Snyder, E.D., 106 79, 82–83, 105, 136–138, 140–141,
sonnet, 29, 61–62, 64–65, 67–69, 73, 75, 145–146, 148–149, 157–161,
77, 142, 177, 250 164–175, 177–178, 180–182,
Southey, Robert, 83 184–185, 194, 228, 242
spatial template, 190, 194–195, 197–200, trochee, 9, 27, 50, 52–54, 172, 174, 183,
202, 204, 207 209, 211–223, 225–229, 255
speaking in tongues, 117 troubadours, 20
speech errors, 235–238, 241–242, 245, true wit, 133–134
249, 257 Tsur, Reuven, 6, 9, 11–13, 24, 28, 30,
Spenser, Edmund, 256 38–39, 41–42, 46, 50–52, 54, 62,
Spooner, William Archibald, 237 73, 96, 99, 104, 106–107, 135,
spoonerism, 237–239, 242–243, 257 145, 166–167, 176, 179, 189, 192,
stages in the development of artistic 207, 209, 222–224, 240–243, 247,
devices, 16, 101 249–252, 254–256
I n de x [ 277 ]
278
Tsur, Reuven and Motti Benari, 38 Wheelwright, Philip, 241

Tsur, Reuven and Yehosheva Bentov, Whitman, Walt, 104, 182
24, 52 wicked stepmother, 20–21, 66, 85
Wimsatt, William Kurtz, Jr., 149,
Ultan, Russel, 191–192 225, 240
universals, 53, 96 Wimsatt, William Kurtz and Monroe
unmarked, 9–12, 51, 192–194, 198–199 C. Beardsley, 225
Winnicott, Donald, 101
Verlaine, Paul, 25, 159, 165–166, wit, 72, 132–134, 137, 145–146, 149,
169–170, 175, 179, 241, 243 152–154, 176, 240, 251
vibratos, 16 witticism, 133, 137
Victorian furniture, 17 Woodrow, Herbert, 13, 26, 172,
Villon, François, 20, 61, 70–71, 73–80, 217–218, 222
82–84 word-retrieval, 234
visual word form area, 43, 48 Wordsworth, Wiiliam, 100, 104
vocatives, 4–5, 15–16, 36–40 writing systems, 43–44, 46–49
Wyatt, Thomas, 20, 64–65, 68–69
Wagshal-Te’eni, Siwan and Reuven
Tsur, 46 Yarden, Dov, 110, 113
War of the Ghosts, 33–35, 40, 85 Yehuda Halevy, 138–139, 144
Watson, Thomas, 61–62, 64 Yekutiel Ibn Hassan, 60
Weitz, Morris, 6, 160
Wellek, René and Austin Warren, zeugma, 87
183, 210 Zhirmunsky, V., 172
[ 278 ] Index
279
280
281
282

Tsur, Reuven - Poetic Conventions As Cognitive Fossils

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tsur, Reuven - Poetic Conventions As Cognitive Fossils

Uploaded by

Copyright:

Available Formats

i

Poetic Conventions as Cognitive Fossils

COGNITION AND POETICS

Published in the United States of America by Oxford University Press

© Oxford University Press 2017

All rights reserved. No part of this publication may be reproduced, stored in

You must not circulate this work in any other form

Library of Congress Cataloging-in-Publication Data

1. Where Do Conventions Come From? 1

assumes that cultural—and poetic—conventions can be accounted for by

Biographical Note: Reuven Tsur was born in 1932 in Nagyvárad, Transylvania

[ xii ] Preface by the Author

transmission assumed a good fit to schemata entertained by Western

Chapter 5, written in collaboration with Idith Eynat-Nov, explores even

[ xiv ] Preface by the Author

In ­chapter 8 I explore the cognitive rationale of why we call sounds

tip-of-the-tongue phenomenon, etc.), on the other. Both are organized

[ xvi ] Preface by the Author

Parts of my 2010 article “Poetic Conventions as Cognitive Fossils” appear

ABOUT THE COMPANION WEBSITE

Poetic Conventions as Cognitive Fossils

Where Do Conventions Come From?

I n this book I contrast two approaches to cultural conventions in gen-

1. I am indebted to Ronit Shoshany for this citation.

The “migration” or “influence-hunting” approach sees a convention as

Fossil refers to the remnant or trace of something constituted to carry

[ 2 ] Poetic Conventions as Cognitive Fossils

As an epitome of “cognitive fossils” in the weaker sense, consider some

Let us complement Darwin’s passage with another one by Gombrich, from

Even the origin of language, that notorious problem of speculative history,

If Darwin and Gombrich are right, we have in the origins of language a

Where Do Conventions Come From? [ 3 ]

[ 4 ] Poetic Conventions as Cognitive Fossils

the introduction to the Literature and Cognition website, which is no lon-

An important assumption of cognitive anthropology is that in the process of

Where Do Conventions Come From? [ 5 ]

[ 6 ] Poetic Conventions as Cognitive Fossils

4. This seems to be true of biological evolution as well, as Deric Bownds claims:

During evolution there have been frequent reinventions of similar adaptations,

Where Do Conventions Come From? [ 7 ]

resemblances are to be explained by simple transmission, with more or less

COGNITIVE CONSTRAINTS VERSUS INFLUENCE HUNTING

There is a scholarly tradition, the so-called Russian school of verse study,

[ 8 ] Poetic Conventions as Cognitive Fossils

Where Do Conventions Come From? [ 9 ]

[ 10 ] Poetic Conventions as Cognitive Fossils

Where Do Conventions Come From? [ 11 ]

[ 12 ] Poetic Conventions as Cognitive Fossils

9. Woodrow’s experimental findings seem to be consistent with some of my other

Where Do Conventions Come From? [ 13 ]

TURNING EXPRESSIVE FEATURES INTO HARMLESS ORNAMENTS

Anton Ehrenzweig (1965), in his sometimes speculative but always illumi-

[ 14 ] Poetic Conventions as Cognitive Fossils

Why dois your brand sae drap wi bluid,

As Jakobson (1960, 1980) and others have noted, interjections and voca-

Where Do Conventions Come From? [ 15 ]

emotive emphasis to them and instead sees them as ballad-style features.

[ 16 ] Poetic Conventions as Cognitive Fossils

“Leveling” attempts to minimize or even eliminate the anomalous detail.

(a) (b) (c)

(d) (e) (f)

Where Do Conventions Come From? [ 17 ]

I prefer Gombrich’s view, which relates the process to a fluctuation in

In chapter 8 I explore the cognitive rationale of why we call sounds

I n chapter 1 I quoted a passage from a lecture by Roy D’Andrade from