Kubla Khan - Poetic Structure, Hypnotic Quality and Cognitive Style - A Study in Mental, Vocal, and Critical Performance

'Kubla Khan'
Poetic Structure, Hypnotic Quality and Cognitive Style

HUMAN COGNITIVE PROCESSING is a forum for interdisciplinary research on the
nature and organization of the cognitive systems and processes involved in
speaking and understanding natural language (including sign language), and
their relationship to other domains of human cognition, including general
conceptual or knowledge systems and processes (the language and thought
issue), and other perceptual or behavioral systems such as vision and non
verbal behavior (e.g. gesture). 'Cognition' should be taken broadly, not only
including the domain of rationality, but also dimensions such as emotion and
the unconscious. The series is open to any type of approach to the above
questions (methodologically and theoretically) and to research from any
discipline, including (but not restricted to) different branches of psychology,
artificial intelligence and computer science, cognitive anthropology, linguistics,
philosophy and neuroscience. It takes a special interest in research crossing the
boundaries of these disciplines.
Editors
Marcelo Dascal, Tel Aviv University
Raymond W. Gibbs, University of California at Santa Cruz
Jan Nuyts, University of Antwerp
Editorial address
Jan Nuyts, University of Antwerp, Dept. of Linguistics (GER),
Universiteitsplein 1, B 2610 Wilrijk, Belgium.
E-mail: jan.nuyts@ua.ac.be
Editorial Advisory Board

Melissa Bowerman, Nijmegen-, Wallace Chafe, Santa Barbara, CA;
Philip R. Cohen, Portland, OR; Antonio Damasio, Iowa City, IA;
Morton Ann Gernsbacher, Madison, WI; David McNeill, Chicago, IL;
Eric Pederson, Eugene, OR; François Recanati, Paris;
Sally Rice, Edmonton, Alberta; Benny Shanon, Jerusalem;
Lokendra Shastri, Berkeley, CA; Dan Slobin, Berkeley, CA;
Paul Thagard, Waterloo, Ontario
Volume 16
'Kubla Khan' - Poetic Structure, Hypnotic Quality and Cognitive Style
A study in mental, vocal and critical performance
by Reuven Tsur
'Kubla Khan'
Poetic Structure, Hypnotic Quality
and Cognitive Style
A study in mental, vocal and critical performance
Reuven Tsur
Tel Aviv University
John Benjamins Publishing Company

Amsterdam / Philadelphia
Library of Congress Cataloging-in-Publication Data
Tsur, Reuven
'Kubla Khan' - poetic structure, hypnotic quality and cognitive style : a
study in mental, vocal and critical performance / Reuven Tsur.
p. cm. (Human Cognitive Processing, ISSN 1387-6724 ; v. 16)
Includes bibliographical references and index.
1. Coleridge, Samuel Taylor, 1772-1834. Kubla Khan. 2. Poetics-
Psychological aspects. 3. Cognitive styles. I. Tsur, Reuven. II. Series.
PR4480.K83 T78 2006

821/.7-dc22 2005058865
ISBN 90 272 2369 6 (Hb; alk. paper)
© 2006 - John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or
any other means, without written permission from the publisher.
John Benjamins Publishing Co. • P.O. Box 36224 • 1020 ME Amsterdam • The Netherlands
John Benjamins North America • P.O. Box 27519 • Philadelphia PA 19118-0519 • USA
To the memory of my beloved daughter Inbal
With whom I shared many of my insights
Table of Contents
Kubla Khan: Or a Vision in a Dream, a Fragment

by Samuel Taylor Coleridge ix
Introduction 1
Part One: Criticism and Meta-Criticism

1. "Kubla Khan" and the Implied Critic's Decision Style 11
2. The Texture and Structure of "Kubla Khan" 79
Part Two: The Reader and Real Readers

3. Gestalt Qualities in Poetry and the Reader's Absorption Style. .115
Part Three: Vox Humana

4. Performing "Kubla Khan"—
An Instrumental Study of Four Readings 143
Afterword
Integration and Wider Perspectives 205
References 235
Index 245
Kubla Khan:
OR, A VISION IN A DREAM. A FRAGMENT.
By Samuel Taylor Coleridge
The following fragment is here published at the request of a poet of great and
deserved celebrity [Lord Byron], and, as far as the Author’s own opinions are
concerned, rather as a psychological curiosity, than on the ground of any
supposed poetic merits.
In the summer of the year 1797, the Author, then in ill health, had retired to
a lonely farm-house between Porlock and Linton, on the Exmoor confines of
Somerset and Devonshire. In consequence of a slight indisposition, an anodyne
had been prescribed, from the effects of which he fell asleep in his chair at the
moment that he was reading the following sentence, or words of the same
substance, in ‘Purchas’s Pilgrimage’: ‘Here the Khan Kubla commanded a
palace to be built, and a stately garden thereunto. And thus ten miles of fertile
ground were inclosed with a wall.’ The Author continued for about three hours
in a profound sleep, at least of the external senses, during which time he has
the most vivid confidence, that he could not have composed less than from two
to three hundred lines; if that indeed can be called composition in which all the
images rose up before him as things, with a parallel production of the
correspondent expressions, without any sensation or consciousness of effort.
On awaking he appeared to himself to have a distinct recollection of the whole,
and taking his pen, ink, and paper, instantly and eagerly wrote down the lines
that are here preserved. At this moment he was unfortunately called out by a
person on business from Porlock, and detained by him above an hour, and on
his return to his room, found, to his no small surprise and mortification, that
though he still retained some vague and dim recollection of the general purport
of the vision, yet, with the exception of some eight or ten scattered lines and
images, all the rest had passed away like the images on the surface of a stream
into which a stone has been cast, but, alas! without the after restoration of the
latter!
Then all the charm

Is broken—all that phantom-world so fair
Vanishes, and a thousand circlets spread,
And each mis-shape[s] the other. Stay awhile,
x “Kubla Khan”—Poetic Structure, Hypnotic Quality & Cognitive Style
Poor youth! who scarcely dar’st lift up thine eyes—

The stream will soon renew its smoothness, soon,
The visions will return! And lo, he stays,
And soon the fragments dim of lovely forms
Come trembling back, unite, and now once more
The pool becomes a mirror.
Yet from the still surviving recollections in his mind, the Author has frequently
purposed to finish for himself what had been originally, as it were, given to
him. Σαμερον αδιον ασω1 Αριον διον σω 1834: but the tomorrow
is yet to come. — Coleridge
Kubla Khan
In Xanadu did Kubla Khan
A stately pleasure-dome decree:
Where Alph, the sacred river, ran
Through caverns measureless to man
Down to a sunless sea.
So twice five miles of fertile ground
With walls and towers were girdled round:
And there were gardens bright with sinuous rills,
Where blossomed many an incense-bearing tree;
And here were forests, ancient as the hills,
Enfolding sunny spots of greenery.
But oh! that deep romantic chasm which slanted

Down the green hill athwart a cedarn cover!
A savage place! as holy and enchanted
As e’er beneath a waning moon was haunted
By woman wailing for her demon-lover!
And from this chasm, with ceaseless turmoil seething,
As if this earth in fast thick pants were breathing,
A mighty fountain momently was forced:
Amid whose swift half-intermitted burst
Huge fragments vaulted like rebounding hail,
Or chaffy grain beneath the thresher’s flail:
And ‘mid these dancing rocks, at once and ever
It flung up momently the sacred river.
Five miles meandering with a mazy motion

Through wood and dale the sacred river ran,
Then reached the caverns measureless to man,
1 To sing a sweeter song tomorrow.

“Kubla Khan”, or a Vision in a Dream xi
And sank in tumult to a lifeless ocean:

And ‘mid this tumult Kubla heard from far
Ancestral voices prohesying war!
The shadow of the dome of pleasure
Floated midway on the waves;
Where was heard the mingled measure
From the fountain and the caves.
It was a miracle of rare device,
A sunny pleasure-dome with caves of ice!
A damsel with a dulcimer

In a vision once I saw:
It was an Abyssinian maid,
And on her dulcimer she played,
Singing of Mount Abora.
Could I revive within me
Her symphony and song,
To such a deep delight ‘twould win me,
That with music loud and long,
I would build that dome in air,
That sunny dome! those caves of ice!
And all who heard should see them there,
And all should cry, Beware! Beware!
His flashing eyes, his floating hair!
Weave a circle round him thrice,
And close your eyes with holy dread,
For he on honey-dew hath fed,
And drunk the milk of Paradise.
Introduction
This book consists of three closely-related parts, but makes no claim for
organic unity. They were written at different stages of my professional career,
over a period of twenty years and used different research methods. While in
the first and third parts the theory is applied to the same poem ("Kubla
Khan"), in the second part it is applied to other texts. Nevertheless, the
research in all three parts has been guided by the same aesthetic and cognitive
conceptions, illuminating the same issues from different angles. The first part
(Chapters 1 and 2) was published in 1987 and discussed three aspects of a
complex aesthetic event: "Kubla Khan" as a hypnotic-ecstatic poem, validity
in interpretation, and the influence of the critic's cognitive or personality style
on his critical decisions in general, and when interpreting "Kubla Khan" in
particular. This part is entirely speculative, in the manner prevalent in literary
criticism. The other two parts are empirical. The empirical tests in these two
parts were conducted within different disciplines, and concern different
aspects of the aesthetic event. The second part (Chapter 3) explores gestalt
qualities of the text and "reader-response", with methodologies drawn from
Gestalt Theory and experimental Cognitive Psychology. The third part
(Chapter 4) employs Instrumental Phonetics, submitting four commercially-
available vocal performances of "Kubla Khan" by leading British actors to
instrumental investigation.
The second part of this book was published in the years 1990-1991, and is
based on research conducted with two younger colleagues, then PhD students
in Cognitive Psychology and Comparative Literature. We were testing my
conceptions of hypnotic poetry and of the reader's decision style. My
collaborators contributed the experimental design, and a preoccupation with
the personality trait "absorption". This research was guided by hypotheses
regarding response to hypnotic poetry in general, and varying rhyme patterns
required to account for hypnotic poetry, as expounded in the first part of this
book and my later work. It attempted to assess empirically whether people of
varying decision styles do indeed respond to pieces of poetry in ways pre
dicted in Chapter 1. Three variables were tested: the effect of text structure,
cognitive style, and professional training on reader response. In the first part,
psychological hypotheses had been drawn from personality variables consti-
2 "Kubla Khan"—Poetic Structure, Hypnotic Quality & Cognitive Style
tuting a wide range of dichotomic scales as expounded by psychologists of

perception-and-personality. In the second part we investigated the personality
variable "absorption", devised to predict hypnotic susceptibility, and of which
I had not been aware at the time of writing Part One. In our empirical study we
used a relatively unknown Hebrew poem rather than "Kubla Khan" as our
paradigmatic hypnotic poem, because the majority of our experimental
subjects were Israeli students, who would have no significant or reliable
intuitions concerning an English poem. Had we used native speakers of English
they would very likely have been biased by the reputation of this well-known
poem.
The idea for the third part grew out of the refereeing process of a projected
book consisting of the first two parts of the present book. An anonymous
reviewer suggested: "Because so much of the main argument of the book
centres on an analysis of prosodic structure (i.e. the claim that Kubla Khan is a
hypnotic poem and the proposed justification of that claim, as well as the
experimentally studied effects of rhyme patterns on real readers) the discus
sion could have benefited from more recent work on related issues, such as
analysis of the intonation unit and how it can be manipulated to influence the
flow of information". Since this suggestion accorded with my current research
interests, I was delighted to take up the challenge. However, for reasons to be
explained in due course, I decided to investigate actual readings rather than
theoretically analyse hypothetical intonation contours, as the reviewer probab
ly meant. Finally, both the first and the third part of this book are concerned
with interpretation and interpreters' decision-making. But these are actualised
in different modes: abstract metalinguistic discourse in Part One, and concre
tion of the poem in a vocal performance in Part Three. In the section "Sum
mary and Some Wider Perspectives" of Chapter 4 I make some generalisations,
based on the preceding discussion, on how vocal gestures can be perceived as
supporting one or another interpretation.
I am re-publishing, then, with minor additions, the two chapters that consti
tuted my 1987 book The Road to "Kubla Khan"—A Cognitive Approach,
adding a second part reporting an empirical study of reader response. A third
part of this book describes an instrumental investigation of recorded readings
of "Kubla Khan". It explores how poetic rhythm and emotional qualities are
displayed by the human voice. Here, for the first time in my critical career, I
am pointing out the vocal resources by which reciters may convey or suppress
the hypnotic quality of a poem. The book ends with an Afterword that ex
plores how the findings of the three parts affect or support each other.
Over the years I had little reason to change my views presented in the orig
inal book. Yet three important things have changed. First, North Holland Pub
lishers gave me an extraordinary opportunity to present an integrated view of
my previous sporadic discussions of cognitive poetics, in my 1992 book
Toward a Theory of Cognitive Poetics. Thus, this discussion became part of a
Introduction 3
more comprehensive theory. Second, I greatly extended the scope of my

discussion of hypnotic and ecstatic poetry as part of a wider conception of
"poetry and altered states of consciousness": in the afore-mentioned book
there is a group of chapters called "Poetry and Altered States of Conscious
ness". I have also enlarged the corpus: in these chapters additional instances of
English hypnotic and ecstatic poems are discussed; and I published a small He
brew book on Hebrew hypnotic poetry (Tsur, 1988), as well as a Hungarian
article on Hungarian hypnotic poetry (Tsur, 1994). In the Hebrew book I
explored some additional cognitive devices relevant to hypnotic poetry, among
them a detailed study of the gestalt qualities related to rhyme patterns. This I
have done through a thought experiment, by systematically manipulating the
rhyme pattern of one of Omar Khayyam's Rubáiyáts, then extending the
results of the analysis to a twenty-four line long Hebrew poem rhyming on
one, "monotonous" sound cluster. Third, in the study reported in the second
part of this book we submitted this theoretical analysis of gestalt qualities to
an empirical investigation, in a complex experimental set-up. We explored the
interaction of three variables: the perceived qualities of rhyme patterns as
predicted by gestalt theory, the readers' "absorption" style (a personality trait
assessed by a questionnaire developed to predict hypnotic susceptibility), and
readers' professional training. The results of this empirical study are presented
in the second part of this book (Chapter 3).
As will be seen in Chapter 1 of this book, some critics have been baffled by
the multiplicity of interpretations of "Kubla Khan" offered by outstanding
critics. I will argue that much (but not all) of this diversity is allowed by the
very nature of interpretation as analysed by such analytic philosophers as
Morris Weitzand Joseph Margolis; that criticism cannot offer a "true" inter
pretation, only what is "merely possible". This, however, does not mean that
"anything goes": within the boundaries of more or less legitimate interpreta
tions there are more plausible and less plausible ones, which can be discussed
in a principled manner. Now this poses an interesting problem: as Else
Frenkel-Brunswick argued, some personality styles are characterised by an
inability to assume an attitude toward the "merely possible". Such an inability
is hostile to what might be considered as a legitimate interpretation. The pre
sent book propounds one possible interpretation: "Kubla Khan" as a romantic
nature poem that assumes a hypnotic-ecstatic quality.
The present book adopts Kenneth Burke's notion of "the use of language as
a symbolic means of inducing co-operation in beings that by nature respond to
symbols". Such a formulation brings out the difference between being exposed
to a system of signs and being exposed to an electric wire. The former, but not
the latter, requires a certain kind of co-operation before a certain kind of effect
is produced. What is more, different kinds of co-operation may produce differ
ent effects. Many literary critics assume that certain verbal structures elicit
certain responses in the addressee. The present assumption is that the
responses depend, largely, on the reader's co-operation. Hence the multiplicity

of more or less legitimate interpretations to, e.g., "Kubla Khan".
Nor was the poem uniformly welcome as a masterpiece. John Spencer
Hillspeaks of a radical change in the critical evaluation of "Kubla Khan":
Throughout the nineteenth century and during the first quarter of the
twentieth century Kubla Khan was considered, almost universally, to be a
poem in which sound overwhelms sense. With a few exceptions (such as
Lamb and Leigh Hunt), Romantic critics—accustomed to poetry of
statement and antipathetic to any notion of ars gratia artis—summarily
dismissed Kubla Khan as a meaningless farrago of sonorous phrases
beneath the notice of serious criticism. It only demonstrated, according to
William Hazlitt, that "Mr. Coleridge can write better nonsense verses than
any man in England"—and then he added, proleptically, "It is not a poem,
but a musical composition". For Victorian and Early Modern readers, on
the other hand, Kubla Khan was a poem not below but beyond the reach
of criticism, and they adopted (without the irony) Hazlitt's perception
that it must properly be appreciated as verbalised music. "When it has
been said", wrote Swinburne of Kubla Khan, "that such melodies were
never heard, such dreams never dreamed, such speech never spoken, the
chief thing remains unsaid, and unspeakable. There is a charm upon [this
poem] which can only be felt in silent submission of wonder" (Hill, 1983:
93-94).
Some of the nature of this "charm" became better understood when Snyder
(1930) put forward his notion of "hypnotic", or "trance-inductive" poetry,
which I have further elaborated in the present book. Such poetry tends to
direct attention away from the contents to the sound of poetry. The new
"Afterword" explores, among other things, the cognitive mechanisms under
lying the ability to appreciate hypnotic-ecstatic poems as "verbalised music".
In our empirical study we found that low-absorption readers tend to effect
poetic closure wherever possible, high-absorptionreaders to leave shapes open.
I have speculated that such different inclinations to organise poetic texts into
stronger or weaker shapes may crucially affect the perception of the rich
precategorial auditory information that conveys the speech sounds: the weaker
the shapes, the more active and the more diffuse is the precategorial auditory
information. This may be one of several reasons for our finding that low-
absorption readers tend to judge hypnotic poems "boring" whereas high-
absorption readers find them "interesting". My exploration of this issue relies
both on our own experiments and on a wide range of empirical research on
speech processing and gestalt qualities conducted by others.
There is a vague suggestion in Hill's account that the different evaluations
are somehow determined by the taste of the time. While not denying the
possible importance of this temporal element, the present conception is that in
this paragraph we are not merely confronted with a string of words that elicits
Introduction 5
different verdicts from different readers, guided by a Zeitgeist. My assumption

is that the different verdicts are preceded by different kinds of co-operation,
different "mental performances" of the string of words on the reader's part.
Roughly speaking, Swinburne seems to have realised certain kinds of interac
tion between various aspects of the sound patterns, as well as between the
musical and the other elements of the text; Hazlitt seems to have failed to
realise them. Chapter 2 of the present book, "The Texture and Structure of
'Kubla Khan'", will be devoted to a detailed description of the type of
structural realisation of the text that is presupposed by the kind of impression
reflected in Swinburne's evaluation of the poem. It assumes that one should
offer a detailed description of the semantic and prosodic textures of the poem,
as well as the overall structure, before going outside the poem for an interpret
ation. From the metrical point of view, this is one of the most regular poems in
the English language. The paper takes up John Crowe Ransom's notion that a
fairly predictable metre may dispel anxiety in the presence of ambiguity—give
"false security to the Platonic censor in us". Here a distinction is made
between false and genuine security, that result in ecstatic and witty poetry
respectively, defining the semantic and prosodic conditions in which security
is "false" or "genuine".
Not only Zeitgeist may affect the various kinds of reader co-operation. The
present study attempts the principled investigation of another source of
influence, what might be called the reader's personality style, or cognitive
style; or, to avoid the need to choose between these two options, his "decision
style". "Style" in this phrase suggests that decisions made by a reader or a
critic are not mere "whims" of his taste, but display some significant (though,
perhaps, unintentional) consistency, governed by certain principles if not
"rules". In the late nineteen-forties two important symposia were held in
which leading psychologists attempted to integrate two domains of psycholo
gical studies in which research began independently, but gradually merged into
one field of study: perception and personality (Bruner and Kretch eds., 1948;
Blake and Ramsey eds., 1951). Their assumption was that one's personality
and emotional needs determine, to a large extent, what one perceives. This
makes it possible to handle individual differences in perception (and literary
interpretation) in a principled manner.
More recently the study of the role of emotions in decision-making received
a boost "when the focus of attention shifted from mathematical logic to
neuroscience and evolutionary biology: Rather than pursuing the (quixotic)
search for ideal rationality, scientists started enquiring how human agents
actually took decisions and were surprised to find that emotion was an in
dispensable part of the process. As recorded in his influential book Descartes '
Error (1994), the neurologist Antonio Damasio discovered that patients with
damage to the orbitofrontal cortex — the part of the brain that deals with the
social emotions —were unable to make the simplest of decisions and generally
acted in a decidedly irrational way" (Sutherland, 2001: 81). I began to explore

how "the social emotions" affect critical decisions in the mid-nineteen-sixties,
long before I was aware of such neuropsychological research. A careful reading
of Damasio's book (1994), however, makes it clear that his findings affect my
argument mainly on this very general level: rational decisions presuppose
proper emotional processes. If you refer to brain centres regarding the quest
for certitude in literary response you are, at best, re-stating the problem in
neurobiologicai terms in addition to depth-psychological or social-psychologi
cal terms. We need to handle the inhibition of proper emotional (and intellec
tual) responses not with reference to the suspension of the proper activity of
certain brain centres, but to an intolerance for ambiguity or uncertainty. I had
discovered this intolerance of ambiguity, as a secondary school teacher, long
before I ever heard of Else Frenkel-Brunswick. Certain students resented ambi
guity-hunting in poetry, and displayed all sorts of aggressive behaviour when
exposed to it. When I reported this at a literature workshop of angry young
teachers, educational psychologist Moshe Caspi said "There is a psychological
phenomenon called 'Intolerance of Ambiguity'". That is how I was first
introduced to the subject. The avoidance of certain emotional responses in such
instances need not be traced to the malfunctioning of brain centres responsible
for emotional responses, but rather to anxiety aroused by uncertainty.
The theory propounded here concerns a scale of critical attitudes, one end of
which I marked as "the quest for certitude", the other as "negative capability",
defined by Keats as "the ability to be in uncertainties, Mysteries, doubts,
without any irritable reaching after fact and reason", and "to let the mind be a
thoroughfare for all thoughts". The present work explores how research on
perception and personality can be utilised to account for certain critical
decisions concerning poetry. This scale of oppositions based on Keats's
definition corresponds to such scales of personality styles (established in
twentieth century psychology) as leveling and sharpening, concrete and
abstract personality, tolerance and intolerance of ambiguity, open and closed
mind, rigidity and flexibility, instrumental and experiential set, high and low
absorption, and so forth. Chapter 1, "'Kubla Khan' and the Implied Critic's
Decision Style", reviews a wide range of discussions of this poem. It assumes
that "Kubla Khan" is a hypnotic-ecstatic poem, and that "hypnotic" and
"ecstatic" are qualities that require a high degree of Negative Capability. Indeed,
as Chapter 3 will demonstrate in a controlled experiment, "high-absorption"
respondents are more prone to realise a possible "trance-inductive" quality in a
poem than "low-absorption" respondents. I assume that this finding does not
concern only responsiveness to "trance-inductive" quality in a poem, but,
presumably, to poetic qualities in general. However, "trance-inductive" quality
at one extreme and the absurd at the other make a greater demand on the
reader's Negative Capability than some other qualities. Chapter 1 argues that
some of the interpretations of "Kubla Khan" are not as illuminating as could be
Introduction 7
because the critics seem to be incapable "of being in uncertainties, Mysteries,

doubts, without any irritable reaching after fact and reason". It discusses such
methodological issues as how one interpretation of a poem may be preferred to
another, or the ways in which the application of the various psychological
models to a poem may be less or more valid from the aesthetic point of view.
At the same time, it also attempts to show how critical decisions may reflect a
critic's cognitive strategies, and how these strategies may result in more or less
adequate readings.
The fashionable think-aloud experiments may reveal the enormous wealth of
a reader's inner world; but this technique is too egalitarian. Not everything a
person utters when reading a piece of literature is equally relevant to its
realisation. Think-aloud experiments assume that whatever the reader utters a
propos a piece of literature is of the greatest importance. Such an approach
fails to distinguish between the reader's idiosyncratic responses and those
responses that reflect his underlying structural knowledge. The approach pro
pounded here, by contrast, gives a minute description of poetic structures.
This makes it possible to grade readers' responses according to the degree to
which they reflect structural knowledge. At the same time, it systematically
explores the (implied) reader's or critic's decision style, and offers a
psychological theory that may account for the preference for responses that
reflect more or less of one's underlying structural knowledge; and also for the
various ways of handling such structural knowledge. This is not merely a
matter of professional training. It is governed to a considerable degree by
personality style. As reported in Chapter 3, we have found that high- and low-
absorption readers with similar high academic qualifications in literature may
respond in systematically different ways to the same poetic texts.
One of the crucial terms used by New Criticism is "ambiguity". One of the
personality variables introduced by Else Frenkel-Brunswick into the study of
perception and personality is "intolerance of ambiguity". This would lead one
to expect that readers and critics may deploy strategies to avoid certain crucial
aspects of aesthetic structure. Thus some responses to a piece of literature
may be less adequate than others. Cognitive poetics assumes that response to
poetry exploits adaptive devices turned to aesthetic ends. Psychologists claim
that certain personality styles tend to be maladaptive in extra-literary reality.
Chapter 1 of this book argues at length that the same strategies may be
maladaptive also in responding to stimuli consisting of verbal structures. In the
second part of this book we have empirically assessed one personality variable
in real readers. When reading a piece of literary criticism, however, the meta-
critic usually doesn't have access to the empirical critic's personality style.
But sometimes we may notice that in a piece of criticism certain cognitive stra
tegies characteristic of certain personality or cognitive styles are consistently
deployed. This indicates "the implied critic's decision style", which need not
necessarily conform with the flesh-and-blood critic's personality style as-
sessed on a questionnaire (though I expect to find some affinity between the

two). In assessing the adequacy of a piece of criticism, we are interested in the
adequate or inadequate deployment of adaptive strategies, not in the flesh-and-
blood critic's empirical personality style.
This is not to say that the empirical testing of flesh-and-blood readers is
useless. It is most illuminating when one is able to isolate and establish several
independent variables on independent grounds and then observe some
significant interaction between them. In Chapter 3 of this book, "Gestalt
Qualities in Poetry and the Reader's Absorption Style" we have systematical
ly manipulated the rhyme structure of a quatrain and, based on gestalt theory,
analysed the resulting structures and made predictions as to their respective
perceptual qualities. This is "the reader's" response. We have, however,
recruited a panel of highly qualified real readers too, formally trained in litera
ture, with a control group of respondents with comparable training in other
disciplines in the humanities and social sciences. This is a second variable. We
obtained the third variable by administering a questionnaire to our respondents
in order to assess the personality variables "instrumental" and "experiential
set", or "low" and "high absorption", devised to predict hypnotic susceptibi
lity. Based on an analysis of these personality variables, we predicted that
low-absorption respondents, who are typically inclined to adopt an instru
mental set, would feel less at ease with weak, ambiguous structures than high-
absorption respondents, with a propensity to adopt an experiential set. The
latter may feel quite at ease with weak, ambiguous structures, and even enjoy
them. Accordingly, the former may resort to leveling or sharpening strategies
(constrained by the structure of the text) to resolve the ambiguity in one or the
other direction. The latter may actually seek out unresolved ambiguities. Con
sequently, where the "punch-line" quality of the last line requires the resolu
tion of ambiguity, as with the rhyme pattern of Omar Khayyám's Rubáiyát,
the low-absorption respondents' spontaneous response would be more
adequate; whereas where the poetic effect depends on exposure to some
"purposeless" intensive emotional quality, as in hypnotic poetry, the high-
absorption respondents' response would be more adequate. Kant characterised
the aesthetic object as "purposiveness without purpose". Other things being
equal, low-absorption respondents would prefer objects in relation to which
the first term must be slightly more emphasised in this phrase; high-absorption
respondents would prefer greater emphasis on the second term.
Our experimental results demanded that we insist on an additional distinc
tion, between cognitive strategy on the one hand and personality style and
professional training on the other. "Leveling and sharpening" are cognitive
strategies deployed by, e.g., low-absorption respondents to get rid of ambigu
ities, according to the constraints of the text. As to the "constraints of the
text", we have quite good reasons to suppose that low-absorption subjects
would prefer to have recourse to leveling rather than sharpening; but the rhyme
Introduction 9
structure of the Rubáiyát bars a leveling strategy, whereas sharpening may

yield exceptionally satisfying aesthetic results. George Klein used the "leveling
and sharpening" dichotomy to define a dichotomic scale of personality styles,
"levelers" and "sharpeners". We have found, however, that in poetic texts
students of psychology tended to level certain evasive differences, whereas
students of literature tended to sharpen them. It would be unreasonable to
suppose that students of psychology tend to be levelers, students of literature
sharpeners. It is more plausible that students of psychology would tend to
level those differences for whose handling they are not equipped, whereas
students of literature would tend to sharpen differences for whose handling
they have acquired effective academic tools. Presumably with other tasks,
requiring skills acquired by students of psychology, the reverse would be the
case.1 Thus, leveling and sharpening are alternative cognitive strategies that
may be utilised in the service of either personality style or professional
training.
The experiments reported in Chapter 3 are, then, an epitome of the complex
literary process. Aesthetic qualities are properties of verbal structures, but
only if performed in certain ways. Different people may perform the same
verbal structures in different ways. This, however, does not mean that "any
thing goes". The process is constrained by the structure of the aesthetic object
on the one hand, and the reader's cooperation on the other. The reader's coop
eration, in turn, is determined by such cognitive constraints as the gestalt rules
of perception and the properties and limitations of short-term memory; by his
past experience (professional training, general education, or command of
codes), decision style (personality or cognitive style), and the norms of his
interpretive community. Here we focus mainly on the gestalt rules of percep
tion, professional training, and decision style. There are verbal structures that
can be organised in one's mental performance in several alternative ways; these
alternative organisations display different perceptual qualities. A more than
usually regular rhythm can be perceived as witty, or childish-naive, or mono
tonous and boring, or hypnotic and exciting, depending on the reader's co
operation, on the relationship he establishes between the elements that induce
certainty and those that induce uncertainty. This, in turn, is determined by the
poem's contents on the one hand and the reader's tolerance or intolerance of
It is quite reasonable to suppose that students in different disciplines have

acquired propensity to adopt different cognitive strategies (perhaps their cog
nitive strategies have influenced their choices of disciplines). Recently I gave a
graduate seminar together with art historian Professor Nurith Kna'an-Kedar on
"the grotesque in art and literature". Our students came from both depart
ments, literature and art history. One of the outstanding art students noted
with great astonishment in our first meeting that the handling of literary texts
demanded from her the deployment of cognitive strategies she was unused to.
uncertainty on the other. In this way, contrary to general belief, individual dif
ferences in poetry reading can be handled in a principled, systematic manner.
Recently I received the reports on a submission of mine to a learned journal.
An anonymous reviewer who warmly recommended publication of the article
in question nevertheless made the following comment: "My point of difference
with [Tsur] is that he [...] assumes that readers will respond in such-and-such a
way as a result of this or that in the literary text. I believe on empirical grounds
that one cannot be so sure of such things". I too believe this. It depends on the
reader's cooperation.2 This is a rather complex and problematic issue. I believe
that empirical research on reader response is very important; but there you
may encounter a host of even greater problems, namely, whether everything a
reader says about a piece of poetry is equally relevant. You cannot escape the
necessity to find out which reported effects reflect the reader's underlying
structural knowledge and which do not. Furthermore, our experiments clearly
demonstrated that diversity of experimental results regarding perceived effects
need not reflect idiosyncratic responses: sometimes they can be accounted for
in a principled manner, with reference to personality style and/or professional
training. There may be additional variables. In this respect, some empirical
researchers make life all too easy for themselves. Empirical and speculative
research must complement each other in some way. The present book is an at
tempt to explore this kind of complementarity.
The reading of "Kubla Khan" in this book conceives of it as of a romantic
nature poem that displays a hypnotic-ecstatic quality. Such a quality can be
realised only in certain kinds of performance. This book proffers a detailed
structural analysis of the poem and considers the reading strategies (as
reflected in published criticism) that may be conducive to the realisation of its
emotional quality, and those that may impede it.
I have just completed another paper called "Delivery Style and Listener
Response in the Rhythmical Performance of Shakespeare's Sonnets". In this
paper I elicited responses of flesh-and-blood readers to two different delivery
instances of one line from the Sonnets, read by Sir John Gielgud, recorded
sixteen years apart. I obtained a welter of responses. After careful examination
I came to the conclusion that apparently incompatible responses to the same
delivery instance may result from the listeners' realisation of different subsets
of aspects of the same complex event. This presupposes a reasonably consis
tent and thorough description of the event by the critic.
"Kubla Khan"
and the Implied Critic's Decision Style
The only man who behaved sensibly was my tailor: he took my

measure anew every time he saw me, whilst all the rest went on
with their old measurements and expected them to fit me (Bernard
Shaw: Man and Superman).
Symbols and the Ineffable
Many readers who believe that "Kubla Khan" is a great poem feel that its
greatness may have to do with the irruption of the irrational and chaotic into
our rational and ordered world, with a force that is unprecedented in lyric
poetry. This irruption, with the enormous energy that infuses this poem,
generates what is frequently characterised as an "ecstatic quality". When we
say that '"Kubla Khan' is an ecstatic poem", we do not report the successful
arousal of an ecstatic experience in the reader, but the detection of an ecstatic
quality in the poem. The ecstatic quality is, then, a perceived quality of "Kubla
Khan"; it is also a "regional quality", that is, a quality that belongs to a whole,
but not to any of its constituent parts. Readers who consider "Kubla Khan" a
great poem usually feel that this ecstatic quality is present in the poem; readers
who tend to regard it to be less than a major poem usually have doubts as to
the presence of this ecstatic quality. Coleridge himself contributed to the
controversiality of his poem by adding the famous preface to it, in which he
claimed to have composed it in an opium-induced dream. Some readers believe
that being in direct contact with the unconscious mind is the source of real
greatness in poetry, and no poem can be credited with this virtue as much as
one composed in an opium dream. On the other hand, Coleridge himself
suggested that the poem remains "a psychological curiosity". Now when we
say ecstasy we denote a compact concept, no less conceptual than the words
logic or concept themselves; but the state of mind "ecstasy" appears to be
inaccessible to conceptual language. Since a literary discourse can hardly escape
the denotative use of language, the paradoxical conclusion seems to be that an
ecstatic poem is a contradiction in terms; which we know it is not.
It is sometimes suggested that this poetic dilemma is resolved through the
use of symbols, and that the symbol somehow partakes in, and "conjures up",
an unsayable reality. Precisely this is said to be the difference between symbol

and allegory: whereas the latter presents the reader with what can be expressed
in a clear conceptual language, the former gives us some mysterious insight into
an unsayable spiritual reality. Coleridge himself was one of the chief exponents
of this conception. Here I shall mention only one of his most frequently
quoted formulations of this distinction. An allegory merely translates abstract
ideas into a "picture-language". A symbol, on the other hand,
is characterised by a translucence of the special in the individual, or of the

general in the special, or of the universal in the general; above all, by the
translucence of the eternal through and in the temporal. It always partakes
of the reality which it renders intelligible; and while it enunciates the
whole, abides itself as a living part in that unity of which it is the represen
tative.
In a paper abounding in wise formulations, Peter Berek (1978:121) makes

the following observation, giving Coleridge's distinction a special twist:
Symbolism is a literary resource based on a metaphysical assumption: the

assumption that there exists an order of being inaccessible to the analytic
mind and inexpressible in discursive logical language. [...] Indeed, for the
symbolist the imagination is a synecdoche for the Transcendent.
And later, again,
Symbolism is perhaps a yearning after allegory in the absence of positive

ideas to allegorise, and as such it is a particularly valuable allegorical re
source for romantic and modernist writers whose intellectual subject is the
difficulty of the process of search, not the clarity of the thing found
(ibid.).
Coming back now to Coleridge's passage, it emphasises the revelation "of

the special in the individual, or of the general in the special, or of the universal
in the general; above all, of the eternal through and in the temporal". Now, how
can we distinguish a piece of literature which "merely translates abstract ideas
into a picture-language", from one which reveals the special through the
individual, etc., and above all the eternal through and in the temporal? In many
cases, I believe, it will be impossible; unless we consider the latter as a special
case of the former: symbols translate abstract ideas into a picture-language,
where the picture itself "is always a part of that, of the whole of which it is
the representative". But even this restriction is quite frequently applicable to
allegories. Allegoric images most typically have metonymic relationships to the
thing represented (the allegoric image of Summer bears flowers; the allegoric
image of death is a skeleton, and so forth).
The Implied Critic's Decision Style 13
It will be noted that I have substituted revelation for translucence when

quoting Coleridge's passage in the preceding paragraph. The key to the distinc
tion between symbol and allegory is to be found in the meaning of translu
cence. The Random House College Dictionary compares the adjectives trans-
parent and translucent. They "agree in describing material that light rays can
pass through. That which is transparent allows objects to be seen clearly
through it. That which is translucent allows light to pass through, diffusing it,
however, so that objects beyond are not distinctly seen". One important
feature that distinguishes symbol from allegory is that the spiritual reality
presented in and through the "picture-language" is perceived as more diffuse
and less distinct in the former than in the latter.
Precisely this would seem to explain why the realities represented by
symbols cannot be expressed in conceptual discourse, in "ordinary language".
Words refer to concepts, to categories, not to occurrences out there in the
external world, or even to subjective mental events. The resulting problem can
be explained by reference to "lateralisation", to the specialisations of the two
hemispheres of the brain. Language, logic, mathematics are "linear" activities,
and are typically associated with the left hemisphere of the brain.
If the left hemisphere is specialised for analysis, the right hemisphere [...]
seems specialised for holistic mentation. Its language ability is quite lim
ited. This hemisphere is primarily responsible for our orientation in space,
artistic endeavor, crafts, body image, recognition of faces. It processes in
formation more diffusely than does the left hemisphere, and its responsi
bilities demand a ready integration of many inputs at once. If the left
hemisphere can be termed predominantly analytic and sequential in oper
ation, then the right hemisphere is more holistic and relational, and more
simultaneous in its mode of operation (Ornstein, 1975: 67-68).
The right and the left hemispheres do not necessarily differ, then, in the kind
of information processed, but rather in the mode of processing. Words refer to
compact entities accessible to the analytic mind: categories or concepts; the
experiences associated with the right hemisphere, on the other hand, are typ
ically diffuse and global, accessible to "holistic mentation". Consequently,
words may capture the information associated with the right hemisphere. What
they cannot capture is its diffuse mode of processing. That is why it is so
often felt that information given about certain human experiences may be all
true, yet the experience itself may be "unsayable". States of consciousness
associated with mystic and ecstatic experiences are typically such experiences
related to the right hemisphere. In some styles at least, among them in romantic
poetry, poetic language typically has recourse to devices that tend to render
information as diffuse as possible and, at the same time, to integrate diffuse
inputs through simultaneous processing (cf. Ornstein, 1975: 95). Some of these
devices, at least, achieve this by activating the right hemisphere at the time
when the left hemisphere is involved in the processing of the linguistic input.
At this point I want to add a caveat. In this kind of inquiry one may appeal
to brain structure only in exceptional circumstances, as when, for instance, it is
used to account for a connection between apparently unconnected processes.
Ornstein's paragraph, for instance, may indicate that in certain circumstances
space description may increase the diffuse processing of information by the
right hemisphere, liberating some information from the tyranny of "compact"
conceptual language (see recently Tsur: 2003, 87-118; 2003b; Tsur and Benari
2002). Damasio's influential book Descartes' Error (1994) may contribute to
our discussion at such abstract and general levels as that emotion is an
indispensable part of the process of decision-making. What it has to say in this
context about the amygdala or the orbitofrontal cortex, by contrast, makes little
difference for our argument. It merely re-states in a parallel "brain" language
what for literary studies can best be described in an "emotion" or "attitude"
language. In fact, several years earlier I made similar generalisations concerning
critical decisions, based on psychoanalytic theory and social psychology,
which ignore all in all the specific brain centres that produce the responses
discussed. What is more, these disciplines allowed me not only to predict that
emotions and attitudes will be involved in critical decisions, but also to make
significant distinctions between, e.g., emotions based on the tolerance and
those based on the intolerance of uncertainty, which will prove essential for
my argument. More recently, Julie Kane (2004) published an illuminating
overview, "Poetry as Right-Hemispheric Language", of the various brain
functions underlying a wide range of poetic devices. This brain information too
may be used only in exceptional circumstances in accounting for poetic effects.
Otherwise it merely multiplies entities.1
The Implied Critic's Decision Style
When in a piece of criticism, or in the output of a critic, certain cognitive

devices are consistently deployed in a way that is characteristic of a certain
cognitive style, I call it "the implied critic's decision style". Paraphrasing
Booth (1961: 71-76) on "the implied author", the implied critic can be defined
as the person whose decisions are reflected in a given piece of criticism. "We
infer him as an ideal, literary, created version of the real man; he is the sum of
his choices" (74-75). In a paper devoted to the critic's possible decision styles
(Tsur, 1975, reprinted in Tsur 1992a: 471-489), I have elaborated on two
critical attitudes. They can be defined relative to each other as ranking higher or
1
Entia non est multiplicandum. I have discussed elsewhere at considerable
length the problem of reductionism and of re-stating the same facts in parallel
languages (e.g., Tsur, 1997).
lower on a scale, one end of which may be marked as what Keats called
negative capability, the other as positivism or factualism, or quest for certitude.
Various works of literature may be ranked as demanding various degrees of
Negative Capability (ecstatic and mystic poetry, for instance, require,
conspicuously, a higher degree than Pope's An Essay on Man). Moreover,
various readings of one and the same work may rank higher or lower in
Negative Capability (the demonstration of this will be one of the main objects
of the present chapter).
One end of the spectrum may be characterised, then, by Keats's description
of the quality
which Shakespeare possessed so enormously—I mean Negative Capability,

that is when man is capable of being in uncertainties, Mysteries, doubts,
without any irritable reaching after fact & reason—Coleridge, for instance,
would let go by a fine isolated verisimilitude caught from the Penetralium
of mystery, from being incapable of remaining content with half
knowledge.
In that paper I attempted to show that in the middle of that spectrum of

attitudes, within the boundaries of more or less legitimate literary criticism,
there are pieces of criticism that exhibit the capability of "being in uncertain
ties, Mysteries, doubts", an "ability to make up one's mind about nothing—to
let the mind be a thoroughfare for all thoughts", and some other pieces of
criticism with some "irritable reaching after fact and reason [...] from being
incapable of remaining content with half knowledge".2 Symbols, then, charac
terised above by Berek as "a yearning after allegory in the absence of positive
ideas to allegorise", seem to require a considerable degree of Negative Capa
bility. And we may expect critics craving for certitude "irritably to reach after
some positive ideas to allegorise".
This dichotomy, of Negative Capability and Quest for Certitude, based on
"literary" formulations in Keats's letters, is astonishingly similar to dicho
tomies formulated by psychologists about 80-100 years later, such as: liberal
vs. authoritarian personality; open vs. closed mind; flexibility vs. rigidity;
tolerance vs. intolerance of ambiguity; abstract vs. concrete personality; "level
er" vs. "sharpener", and so forth. The research reported below in Chapter 3
added two pairs to this list: experiential and instrumental set, and low and high
absorption. I have preferred not to resort to any of the pairs of terms used by
psychologists—in spite of my frequent recourse to their findings—for two
reasons. First, these dichotomies are similar to one another, but are not
synonymous; thus, for instance, the attitude of the Quest for Certitude
(supposed to be akin to the "leveler"), has recourse sometimes to "sharpening"
2 These are near-literal quotations from the above mentioned paper (Tsur,
1975).
tactics in cases where "leveling" would cause too coarse a distortion, or when
"sharpening" is an effective means for dispelling uncertainty (cf. below, when
Yarlott "sharpens" an unevaluated situation into an ambivalent situation which
he in turn "disambiguates" by settling on an unambiguous negative evaluation).
Second, the present chapter is concerned with the implied critic's attitudes as
they are manifest in his choices in critical works; I do not pretend to know
anything about any flesh-and-blood critic's psychology in extra-literary reali
ty. The above mentioned dichotomies, as treated by psychologists, reveal a
series of specific tactics that can be detected in critical writings. They seem to
be in the service of the dichotomy offered here {Negative Capability vs. Quest
for Certitude), which may be regarded as "general strategies". At the end of the
afore-mentioned paper I suggested that the attitudes of the flesh-and-blood
critic may relate to those of the implied critic as "competence" relates to
"performance". But as long as we have no further knowledge about this rela
tionship I shall continue to consider only the latter's attitudes; and for this, it
is advisable to use terms that may prevent mixing up these two notions.
"For art comes to you proposing frankly to give nothing but the highest
quality to your moments as they pass, and simply for those moments' sake"
says Walter Pater, (1951[18731]: 897). The poets, says the great Hebrew poet
Bialik, constantly chase those aspects of things that make them unique, the
fleeting moment that will never again return. These are precisely the least
tolerable things for someone on a Quest for Certitude.
The leveler is more anxious to categorise sensations and less willing to

give up a category once he has established it. Red is red, and there's an end
on't. He levels (suppresses) differences and emphasises similarities in the
interest of perceptual stability. For him the unique, unclassifiable sensation
is particularly offensive, while the sharpener at least tolerates such anoma
lies, and may actually seek out ambiguity and variability of classification
(Ohmann, 1970: 231).
As we shall see in this and the next chapter, some of the most notorious
effects of "Kubla Khan" are derived from an undermining of "perceptual stabil
ity"; and a considerable part of "Kubla Khan" criticism is directed toward the
elimination of "unique, unclassifiable sensations", and establishing in the poem
some "positive ideas to allegorise".
I have said above that when in a piece of criticism, or in the output of a
critic, certain cognitive devices are consistently employed in a way that is
characteristic of a certain cognitive style, I call it "the implied critic's decision
style". Let us consider briefly a short instance in which several such devices
are deployed in the service of one possible cognitive style. Brooke-Rose (1958:
32) quotes the following adverse comment on 19th century French Symbolism
by Jules Lemaître, back in 1888:
Un symbole est, en somme, une comparaison prolongée dont on ne nous

donne que le second terme, un système de métaphores suivies. Bref, le
symbole, c'est la vieille 'allégorie' de nos pères.
I do not pretend to know anything about Lemaître's cognitive style or, in

fact, anything beyond what is said in the above quotation. But, what he is
doing here, in the first place, is to level the difference between symbol and
allegory. Secondly, he not only levels the differences and emphasises the
similarities between allegory and symbol, he does so for a very obvious pur
pose: to deny the existence of a unique, unclassifiable sensation. One feature
that Impressionism has in common with nineteenth century French Symbolism
is that they both desired, in Weisstein's (1974) phrase "to capture the fleeting
impression at the very moment in which sensations are transformed into
feelings". When Lemaître resorts to the strategy of debunking, stating that the
poetic symbol is nothing but the good old allegory of our fathers, he denies in
fact that "the fleeting impression at the very moment in which sensations are
transformed into feelings" can be captured by poetry, or that such an experi
ence can exist in the first place. I have suggested above that the difference
between symbol and allegory does not necessarily reside in the kind of in
formation, but in the fact that information in the former is more diffusely
organised than in the latter. It is precisely this diffuse quality that is intolerable
for rigid persons, characterised by an "intolerance of ambiguity".
In a recent paper on "Kubla Khan" (Benzon, 2003, online) I encountered an
illuminating insight based on sharpening strategy, pointing out different
aspects of two terms referring to the same referent. Benzon elaborates on a
contrast between lines 1-11 and lines 12-30 of "Kubla Khan".
This contrast [...] is pointed up in the contrast between the phrases "sun
less sea" and "lifeless ocean". These phrases designate the same object; but
the concepts are different. As light and order dominate [the former
passage], the body of water is conceptualised in those terms; the sun does
not shine on the subterranean sea. But [the latter passage] is dominated by
life energy: sexual desire ["woman wailing"], water, the breath. Thus the
terminal sea is designated in those terms; it is said to be lifeless. [...] The
qualifying term is chosen to be consistent with the governing ontologicai
categories, light and order or life energy and disorder.
To end this section of the present chapter, I wish to dwell briefly on the
dichotomy concrete vs. abstract personality. Of the several ways in which
Harvey and his colleagues have found greater concreteness to be manifested in
contrast to greater abstractness, I wish to point out only five:
* A simpler cognitive structure, comprised of fewer differentiations and

more incomplete integrations within more central and ego-involving
domains but not within domains of low involvement.
* A greater tendency toward more extreme and more polarised evalua
tions, namely, good-bad, right-wrong, black-white.
* A greater intolerance of ambiguity expressed in higher scores on such
measures as the F-scale and Dogmatism Scale and in the tendency to
form judgments of novel situations more quickly.3
* A greater inability to change set and hence greater stereotypy in the
solution of more complex and changing problems.
* A greater insensitivity to subtle and minimal cues and hence a greater
susceptibility to false but obtrusive cues.
(Harvey, 1970: 316).
As to the last-but-one item in this list, Bernard Shaw's John Tanner (quoted
in the motto) would be a case in point. If images could complain, they would
join him in complaining that some critics take their measures from other texts
(Coleridge's other poems, or his real or alleged sources) and expect those mea
surements to fit them in "Kubla Khan" as well.
A word must be said here about the use of the terms concrete vs. abstract.
Readers familiar with Piaget's work on the concrete stage of child development
or with Goldstein's work on the concrete thinking of brain-damaged patients
may consider it inappropriate to apply the term concrete to the intellectual
functioning of professors of literature, or even of university students. When a
psychologist who read the present chapter made an objection to this effect, my
first reaction was to handle the problem by saying: I am using the term
concrete in Harvey's not in Goldstein's or Piaget's sense. Harvey gives a
detailed description of a psychological syndrome which can be transferred
verbatim to the description of the critical behaviour of a large number of
literary critics, scholars and students. Such a transference may yield consider
able insight into critical theory and practice, irrespective of whether the term is
used in a different sense by other researchers. One only should make it clear in
what sense he uses the terms.
On second thought, however, it seemed quite plausible that the phenomena
described by Goldstein in brain-damaged persons and by the present work in
the critical behaviour of some literary critics and scholars are instances of the
same principle. Consider the following classical description of the flexible
personality, tolerant of ambiguity:
The categorical or conceptual attitude is characterised by ability or readi

ness to assume a mental set voluntarily, to shift from one aspect of the
situation to another, to keep in mind, simultaneously, various aspects, to
grasp the essentials of a given whole, to break up a given whole into parts
3 The F-scale (F for fascism) and Dogmatism Scale are irrelevant to our present
inquiry. They are, however, very relevant to the question of the relationship
between the teaching of literature and the education for values (cf. Tsur,
1969, 1975; 1979).
and to isolate them voluntarily, to abstract common properties, to plan

ideationally, to assume an attitude toward "the merely possible", to think
and perform symbolically, and finally to detach our ego from the outer
world (Frenkel-Brunswick, 1968: 136).
This can be taken for a description of the smooth, undisturbed functioning

of the system. When Else Frenkel-Brunswick speaks of the intolerance of
ambiguity, or of the rigid personality, she implies that for some psycho-
dynamic reasons this smooth functioning has been impeded. But these
personality traits are very unlike the concrete functioning Goldstein found in
his brain-injured patients. Frenkel-Brunswick's "rigid" persons are perfectly
capable of highly sophisticated intellectual functioning, whereas Goldstein's
patients were sometimes incapable of the simplest cognitive performances,
such as repeating certain simple sentences after the experimenter. Yet notice
this: all the phrases that in the above passage characterise the categorical or
conceptual attitude of the flexible personality are, in fact, the headings of a
paper by Goldstein and Scheerer (1941) on "Abstract and Concrete Behavior".
In that paper the authors discuss concrete and abstract behaviour in general and
offer the behaviour of brain-injured persons as an instance of concrete
functioning in general that may throw light on the nature of abstract function
ing and the ways it may be disturbed. 4 Thus, the behaviour of Goldstein's and
his colleagues' patients is treated as instances or representatives of a wide
range of kinds and degrees of concrete behaviours. What seems to be common
to all these kinds of behaviours is that in all of them the smooth working of the
same cognitive processes has been disturbed, or impeded, to different degrees,
or at different levels. What is more, Goldstein and Scheerer claim that there are
various degrees of abstractness and also various degrees of concreteness. I
strongly suspect that these two scales meet somewhere in the middle and form
a dichotomic spectrum. Perhaps a less controversial way of speaking of
individuals of intellectual accomplishment would be to use the phrases "more
abstract" and "less abstract". 5 Having elucidated all this, I shall go on using
Harvey's terms concrete and abstract.
4 By the way, the authors point out that one of the signs of abstract thinking is
the capability of regarding an object or event as a representative or instance
of some wider category.
5 There arise some interesting analogies between these "less abstract" and
"more concrete" ways of functioning. Take, for instance, the difficulty to
report to oneself or to others what one is doing. Goldstein and Scheerer report
that certain patients who are perfectly capable of throwing a ball into three
boxes placed at various distances from them are incapable of reporting which
box is further away from them. Similarly, it has been widely observed in poetic
prosody classes that students are perfectly capable of correctly placing the
stress on words in their connected speech, or even in isolated words, but are
incapable of reporting where the stress is placed in a certain polysyllable.
"Positive Ideas"
Many critics who face the disturbing element in "Kubla Khan" and other
ecstatic poems go outside the poem in order to find evidence with the help of
which they may alleviate the discomfort caused by uncertainty. 6 Yarlott
(1967: 128) states the problem as he conceives of it, and enumerates a few
kinds of solutions offered hitherto:
If we restrict ourselves to what is 'given', appealing to the poem as a

'whole', we shall fail probably to resolve its various cruxes. Hence there is
a temptation to look for 'external' influences—to consider Kubla Khan in
relation to Coleridge's general reading, to relate it to his later critical the
ory, or to regard the whole thing as the product of an opium experience.
The trouble with all these approaches is, that they lead finally away from
the poem itself. Alternatively, we can examine Coleridge's handling of his
acknowledged source material (Purchas), rather as we might evaluate
Shakespeare's modifications of Holinshed and North, a technique which
has the virtue at least of grappling with the actual raw material. Or, again,
we can consider Kubla Khan in relation to the rest of Coleridge's poetry,
especially in the imagery it employs, which may be a less dangerous form
of influence-tracing.
Or take the issue of "detaching our ego from the outer world": the case of
the patient who cannot repeat the sentence "Snow is black", because "it is
false". When urged by the experimenter to repeat this meaningless sentence,
he does so but is compelled to mutter after it "No, snow is not black". I have
encountered some illuminating analogous behaviour at the university. A few
years ago there was a student in one of my graduate seminars whose intellec
tual performance puzzled the other students. She is highly intelligent, they
said, but there is something odd in her way of thinking. As a matter of fact,
she exhibited a wide range of extreme symptoms of the Quest for Certitude.
Later she handed in a proposal for a doctoral thesis to the Research Students'
Committee. The committee found that her hypotheses were stated at such a
high level of generalisation that they were completely uninformative. Since
she belonged to my department, I was asked to explain to her what was wrong
with her proposal. When the proposal was returned to her for the second time
and she was urged to formulate some more specific hypotheses, tentatively
though, she said "But I cannot write down those hypotheses, even tentatively,
before I make absolutely certain they are right".
I have discussed this issue at considerable length in relation to Rimbaud's
"Voyelles" in a paper published in my book What Makes Sound Patterns
Expressive?, 111-135.
I could not agree more than with the crucial statement "The trouble with all
these approaches is that they lead finally away from the poem itself'.
However, as I shall attempt to show, the approach advocated here by Yarlott
himself can also be accused of this. One of the serious problems with "Kubla
Khan"-criticism will become apparent precisely from Yarlott's above compari
son to Shakespearean criticism. Shakespeare critics have realised, at long last,
that if one wishes to appreciate Shakespeare's greatness as a poet, one should
look for the differences between Holinshed and the Shakespearean text; where
as Coleridge's critics tend to level (suppress) such differences.
Coleridge, in his notorious preface to this poem, indicated the exact page in
Purchas his Pilgrimage he was reading before sinking into that dream in which
he claimed to have composed "Kubla Khan". Yarlott (1967: 134), like so many
critics before and after him, quotes the relevant sentence from Purchas:
In Xamdu did Cublai Can build a stately Palace, encompassing sixteene

miles of plaine ground with a wall, wherein are fertile Meddowes, pleasant
Springs, delightfull Streames, and all sorts of beasts of chase and game, &
in the middest thereof a sumptuous house of pleasure, which may be
removed from place to place.
Yarlott comments on this passage: "Coleridge seems to have deliberately

modified the attractiveness implicit in Purchas's original description", one of
his examples of modification being this:
The substitution of "bright/sinuous" for "pleasant/delightful" produces sin

ister, almost reptilian, associations, recalling perhaps The Ancient Mariner
or this description of the "thing unblest" from Christabel, where snake
joins "bright" and "green" (the only colour details found in Kubla's
garden) in a cluster of positive malignancy:
When Lo! I saw a bright green snake
Coiled around its wings and neck
Green as the herbs on which it couched.
(Yarlott, 1967: 135).
I can only commend this critical technique of comparing a poem to its

purported source material, and point out the differences. But Yarlott seems to
have missed the point of the change in a curious way. Doubtless, Coleridge
diminished the attractiveness implicit in Purchas' description of the garden; he
eliminated the evaluative ingredient of the adjectives, increasing their descript
ive contents. What Yarlott does, instead, is to smuggle back, through the back
door, some evaluative ingredients, manifesting a "tendency toward more
extreme and more polarised evaluations, namely, good-bad". Thus, instead of
acknowledging the zero grade of evaluation, he replaces the positive evaluative
terms with negative ones. The reason for such a critical behaviour seems to be
quite clear. Purchas uses adjectives in which positive evaluation is unambig

uous. By eliminating the evaluative ingredient, Coleridge creates a state of
uncertainty, which is hardly tolerable for people with the critical attitude of the
Quest for Certitude. A unique, unclassifiable sensation is intolerable for people
with such an attitude; they must know, at least, whether it is good or bad. So,
Yarlott introduces the concrete image of a snake, which is obviously bad, into
the description. This evaluative acrobatics does not work even on Yarlott's
own ground, since the Ancient Mariner's redemption, for instance, began
precisely at the moment when he discovered the beauty of the water snakes
and of the other slimy, crawling creatures, and then "blessed them unaware".
It is worthwhile to notice the strategy that guided this manipulation, in the
light of Harvey's characterisation of the concrete personality. Consider the
technique of searching out another context in Coleridge's poetry where bright
and green occur together (bright green snake). In the first place, it enables the
critic to interpret the description while its integration with the text remains
"incomplete". In the second place, one characteristic of the aesthetic use of
images is that it has a great variety of meaning-potentials. These potentials are
frequently exploited in a variety of ways even within a single piece of
literature, sometimes yielding even conflicting meanings. This, of course,
requires great sensitivity to "subtle and minimal clues". On the other hand, the
ability to identify the occurrence of the words bright and green in two differ
ent contexts as the same words, and then to suggest that snakes have sinister,
reptilian associations, hardly require great sensitivity. These are, typically,
"false but obtrusive cues" in the present context. In the third place, the ex
ploitation of different, or even conflicting, potentials of the same image
requires frequent changes of mental sets (a mental set being a readiness to
respond in a certain way). The assumption that the interpretation of an image
need not be modified along a whole poem, or even across a poet's entire
output, enables the critic to persist in a single mental set on the one hand; on
the other hand, it conflicts with the very essence of aesthetic activities (again,
John Tanner's tailor comes to mind). Therefore, on the next page, Yarlott
returns to this issue, in a mitigated form:
Such parallels should not, however, be pressed too far. The "sunny spots
of greenery", for example, must on any unprejudiced reading of the poem
form an attractive feature of the garden even though, at the time the
poem was written, it seemed like a deliberate echo of Coleridge's
description of the gratification opium affords ("a spot of inchantment, a
green spot of fountains, & flowers & trees") (136).
According to the conception suggested here, neither expression should be

regarded as an echo of the other. "Spots of greenery" or "green spots" may
evoke an image of a landscape that is pleasant to look at or imagine. It has
nothing to do with the moral evaluation of this pleasantness, or the outcome of
one's enjoyment of it. Whether it is morally undesirable, or whether the

pleasure associated with it leads to undesirable results, one may only infer
from the context; that is, the pleasure suggested may be associated with a
literal landscape, or with a landscape conceived as a metaphor for the
gratification opium affords. In other words, one must be ready to change
mental sets, and realise various potentials of the landscape description (that is,
real or false pleasure).
One of the psychologists' dichotomies mentioned above was flexible vs.
rigid personalities. "Rigid" persons, who do not tolerate ambiguous situations,
"have in effect to relieve their anxiety by having rapid closure in cognitive and
perceptual reactions as well as in emotional and social spheres" (Miller, 1951:
263).
This is a way in which some organisms handle the problem of ignorance by

coming to a conclusion—any conclusion—in order to avoid the anxiety
that would otherwise arise (ibid.).
There are some fine examples of this kind of manipulation in Yarlott's book.
He rightly identifies some impersonal syntactic structures at the beginning of
the poem, as the passive voice ("So twice five miles ... were girdled round" and
"there were gardens"). From these syntactic structures, however, he jumps to
the conclusion that the Khan's "relationship with the slave force which,
presumably, enacts his decree for him is utterly impersonal" (ibid., 130). This,
in turn, is only a particular instance of a wider issue:
The Khan himself is peculiarly situated. Cut off from normal personal
relationships he inhabits a solitude almost like that of The Dungeon
prisoner or the Ancient Mariner. He hears only the ghostly voices of his
menacing ancestors and (possibly) that of the wailing woman (ibid., 129-
130).
Here, again, Yarlott is dispelling ignorance. The only evidence for his con
tentions is that there is nothing incompatible with them in the poem. But as a
matter of fact, what characterises the poem in this respect is complete un
certainty as to the Khan's motives, relations, attitudes, etc. What actually
happens in the opening stanzas is that in the first two lines we hear about the
Khan's decree, and then he vanishes into the background: the foreground is oc
cupied by a description of the landscape and of the artifices built as a result of
the decree. Now if we read the poem in a continuous sequence, the impersonal
passive voice for instance, in line 7, may be perceived as an indication of the
prompt execution of the decree: the Khan decreed and lo, the results are there.
Or, alternatively, we have here, again, an emphasis on those aspects of the
description that seem to be unique in a sense and unclassifiable: they are de
void of any statable purpose, evaluation, or the like. The unique aesthetic af-
fect of this uncertainty as to the values and motives inherent in the description
will become apparent when we discuss, toward the end of the next chapter, the
poem's rhythmic organisation.
With the dome-image, however, Yarlott seems to have a problem. It
appears both in positive and negative contexts in Coleridge's poetry. Thus,"as
associated with the religious feeling of 'deep heartfelt inward joy' the 'dome'
was wholly admirable"; on the other hand, "he used 'dome'-images to suggest
also moral laxity" (131). From what we have said above of the aesthetic use of
images, we should expect that in different contexts Coleridge should exploit
different potentials of the dome-image; one may not hope, therefore, to infer
the moral attitude associated with the image in one context, from the moral
attitude associated with it in other contexts. Yarlott, however, manages
eventually to come to a conclusion (we might suggest with Miller, to any
conclusion), in order to eliminate "ambivalence".
The bald ambivalence of "did Kubla Khan / A stately pleasure-dome

decree" scarcely permits of a positive conclusion one way or other about
the desirability of the dome, but we have both Purchas's authority and the
evidence of Coleridge's parallel usages to suggest that the hyphenating of
"pleasure-dome" may have implied strong moral disapproval (ibid., 132).
As a matter of fact, one cannot speak here of ambivalence proper, but rather
of "zero grade evaluation". Thus, Yarlott "sharpens" an unevaluated situation
into an ambivalent one. This ambivalence, in turn, he "disambiguates" by
settling on an unambiguous negative evaluation. This passage exhibits an addi
tional symptom, which we have already encountered and discussed above.
That the phrase pleasure-dome is "hyphenated" in the poem is perfectly true,
though rather trivial. From this graphic observation, however, Yarlott jumps,
again on unexplained grounds, to the conclusion that Coleridge "may have
implied strong moral disapproval".
We could, in this way, go through Yarlott's long chapter and see how he
"disambiguates" image after image, expression after expression. The details
vary, but the strategy of avoiding ignorance, uncertainty or unclassifiable
images is the same. Perhaps we should dwell briefly on one more, randomly
chosen example. "'Incense', for instance, suggests a manufactured perfume
rather than delicate, natural fragrance" says Yarlott (135), and as such it is
deplorable. Now, when the poem explicitly says "Where blossomed many an
incense-bearing tree", incense simply does not suggest manufactured perfume.
As a matter of fact, incense does not necessarily suggest "manufactured
perfume" in any context. But suppose it does. Even in that case, Yarlott's
handling of the expression would be very telling of his decision-style. In that
case, the procedure is this. Coleridge used incense, the artificial perfume, as a
"metaphor" for a very strong natural perfume. The reader is supposed to
abstract this component from incense and cancel all the irrelevant components
(i.e., [+ARTIFICIAL]), that are eliminated by blossomed and tree. One typical
symptom of a concrete personality is an inability to abstract a single quality
and to "forget" irrelevant details of the original context.
Lowes' (1927) enormous study of Coleridge's imagery in "The Ancient
Mariner" and "Kubla Khan" ushered in a tendency that was to become most
influential in Coleridge criticism. Rather than explaining the meaning of the
images, he points out their alleged or proved sources and presents a sort of
psychological theory, to explain how the "unchecked subliminal flow of
blending images" became poetry in Coleridge's poems. Since the publication of
his book an ever-growing corpus of highly erudite studies has attempted to
achieve similar aims, adding here and there more and more far-fetched putative
influences on the poems, occasionally using them also to "explain" the meaning
of its images. One such remarkable book is Beer's book on Coleridge the
Visionary (1959), two of whose chapters are devoted to "Kubla Khan". It is
set within a psychological framework that is much more satisfactory than
Lowes'. It also offers a beautiful hypothesis to explain the poem's unity
(which I shall discuss later). Yet its main bulk is devoted to the pursuit of
meaning-hunting activities very similar to those of Yarlott's study. It commits
all the fallacies we found in Yarlott's work, especially when considering
"Kubla Khan" in relation to the rest of Coleridge's poetry. But, in addition, it
indulges in an orgy of erudition concerning mythology, ancient history, prim
eval lore, etc. Thus, for instance, following Lowes, "Alph the sacred river" is
usually identified with the Alpheus, or a conflation of the Alpheus and the
Nile. Beer, however, believes that the shortening to Alph is not accidental. It
may have to do with the preoccupation in Coleridge's day with the history of
the alphabet. Here comes a three-page-long dissertation on the history of the
alphabet, Cabbalistic speculations on its symbolic meanings, the origin of
languages, etc., as understood in Coleridge's day. Thus, for instance, "If Alph
the sacred river is associated with Beth, the cavern, we do not need to go to the
cabbalistic writings to remind ourselves that the river and the cavern are
themselves male and female symbols, and were used with this significance in
such neoplatonic writings as Porphyry's exposition of the Cave of the
Nymph's in Homer" (209-210). Likewise, there is a long long discussion of
sun-worship and the cities built in the sun-god's honour. One really begins to
suspect a tacit assumption behind this study, that the more gods, or the more
primeval lore, the better the poem. All this and much more constitute solid
pieces of learning that are unlikely to change in response to fine-grained poetic
textures; and they also help to establish the enduring human values of the piece
of poetry. They cannot be integrated into the structure of the poem, but can be
easily isolated and contemplated without being disturbed by the ego's
involvement in the poem. On the whole, they help to reduce the piece of
poetry to something else, outside the poem, that is much less elusive than an
ecstatic poem. At the same time, insofar as they rely on cabbala and
neoplatonic mysticism, they pretend to account for the poem's ecstatic or

mystic quality.
Negative Capability and Symbolic Interpretation
One of the most illuminating studies of "Kubla Khan" is Schneider's (1975)

book. Regrettably, her reading is sketchy and suggestive. The major part of this
book, more than 360 pages long, is devoted to refutations and the establishing
of external evidence. Only its fifth chapter is devoted to a reading of the poem,
and only a very negligible part of this chapter is devoted to close textual
inspection. But Schneider's work gives, time and again, evidence of her high
degree of negative capability, in her dynamic and "oscillating" reading as well
as in her refutations of critical practices that are characteristic of the quest for
certitude. She works within an implicit theoretical framework that is quite in
adequate to account for her insights; but this makes her insights the more re
markable and highlights her sensitivity and negative capability. I am quoting
the following passage only to state and define the issue on a fairly general level;
the more specific discussions will be drawn from other sources.
This air of importance without visible foundation contributes to the sug

gestion of mystery about the poem that is part of a charm. [...] There are
moments in which this kind of charm, an air of meaning rather than
meaning itself, affects the reader (Schneider, 1975: 285).
This paragraph follows a long discussion of the foregrounding of dome in

the first stanza by phonetic means.7 I believe, however, that it can be extend
ed, with the necessary changes, to the whole description of Kubla's building
enterprise, which is presented in the poem in a way that appears to have a
much greater importance than what the facts enumerated in the poem would
seems to justify. Now, it will be noticed that it requires a considerable degree
of negative capability to attribute the mysterious power of the poem to what
is not in it.8 Most critics would handle such a situation in a different manner,
7 This phonetic discussion too is indicative of an attitude of negative capability.

As we will see below, critics characterised by a quest for certitude all too
frequently seek to associate certain sound patterns with specific "inherent"
meanings. This enables them to achieve rapid closure. Schneider, on the
contrary, after pointing out the elaborate sound patterns of the first stanza
suggests that dome is foregrounded by virtue of the fact that it is the only
word in the stanza that is not part of any of the repetitive sound patterns.
This critical strategy requires more than usually delayed closure.
8 It should be noted that in the sentences omitted in the quotation Schneider
discusses the dangers of and justification for this technique.
just like the ones we have discussed so far in this chapter. In order to account
for the mysterious air of the description, you have to supply the missing
information. So they search for positive pieces of information which the
imagination may seize upon, either from ancient myth and primeval lore, or
from other writings of the author. If necessary, they will follow a chain of
associations until they reach a point where the motive, the word or word-
cluster, is associated with some contents that might account for the mysterious
effect of the passage. These pieces of information are usually unavailable to the
ordinary reader, who in spite of this does feel quite frequently that "mysteri
ous" quality in the poem. Indeed, the only purpose of this kind of information-
mongering seems to be to dispel the anxiety of people incapable of accepting
uncertainties, mysteries and doubts, when compelled to face absence as
significant evidence.
To be sure, critics both before and after Schneider were alert to this kind of
source of mysterious qualities. Thus, for instance, in a footnote Yarlott (1967:
134) observes: "Professor J.T. Boulton points out to me that vagueness may
be an important contributory factor to the over-all effect of mystery here".
This, of course, is only an external and non-integral addition to Yarlott's
general source-mongering. What makes Schneider's position quite unique in
this respect in "Kubla Khan"-criticism is that she does not offer this
explanation in addition to, but at the expense of symbolic criticism:
Undoubtedly there are reasons why these things effect us as they do, but
the reasons are not usually furnished by the relatively simple equivalences
of symbolic criticism that has developed hitherto. They are probably too
subtle and too complex to be traceable, at least in the present state of
man's knowledge and consciousness; and they are certainly more inextric
ably bound up with the elements of form than we are in the habit of
supposing (Schneider, 1975: 285-286).
It might be quite impossible to find a piece of criticism that would meet

Schneider's maximal expectations. I wish, however, to quote from an
interpretation of this poem that does not refrain from source-mongering, nor
from symbolic criticism; but time and again it ventures into that finer texture of
poetry that is certainly much subtler than the one perceived by symbolic
criticism, and can be quite "inextricably bound up with elements of form". I
mean Maud Bodkin's interpretation in her Archetypal Patterns in Poetry. At
the present stage of my argument I propose to consider at some length
Bodkin's discussion of the emotional symbolism of caverns. The very notion
of emotional symbolism implies that she is not so much interested in what
images mean as in what images feel like. Caverns and abysses often symbolise
hell; and together with high mountains, they may serve in the archetype of Par
adise and Hades, or Heaven and Hell. Like so many critics of "Kubla Khan",
she begins her discussion of this issue with a collection of myths and geo-
graphic (or pseudo-geographic) descriptions of caves and subterranean streams

and seas by Plato, Purchas, Milton, Seneca, and others. From these myths and
descriptions she abstracts certain common qualities that are much subtler, and
more elementary as experiences, than the usual symbolic meaning. At the same
time, she indicates how the "gross" symbolic meanings may arise from those
subtle, elementary abstractions.
Here is the "eternal essence" gathered from experiences of cavern and

abyss—an essence of cold, darkness, and stagnant air, from which
imagination may fashion a place of punishment, the home of the Evil One
(Bodkin, 1963: 101).
One conspicuous aspect of the preceding passage is that it does not provide
a simple equation: CAVERN = HELL. Rather, it abstracts from "cavern and
abyss" elementary sensations such as "cold, darkness, and stagnant air"; these
sensations are unpleasant, and in extreme cases unfavourable to life. Such an
analysis has several advantages for the literary critic. For one thing, it explains
how a certain visual image may generate (under appropriate circumstances) a
"thing-free" atmosphere with a marked emotional direction. Second, as a result
it can fill with specific contents, applicable in changing circumstances, one of
Kenneth Burke's favourite ideas concerning "the principle whereby the scene
is a fit 'container' for the act, expressing in fixed properties the same quality
that the action expresses in terms of development" (Burke, 1962: 3). This is, in
fact, what may be meant by "a place of punishment": a place that expresses in
fixed properties the same qualities that punishment expresses in action (and
the Evil One in potential actions). Last but not least, it can explain a most cru
cial fact about readers of literature, namely that they can understand a poem
like "Kubla Khan" without ever having heard of Plato, Purchas, Milton,
Seneca, and all the rest. The only thing a reader needs, in this respect, is to
know what a cavern is; from that he abstracts, creatively, the relevant features.
A person who does not know what a cavern or an abyss is cannot understand
"Kubla Khan" in this respect; but ignorance of Plato's cave or Seneca's image
of the subterranean sea is no obstacle to understanding. It is this, and one more
all-important factor, that make it possible for even the naive reader to
appreciate the emotional significance of caverns and abysses, and for mythic
imagination to fashion from them the archetypal place of Hades:
In Plato's image, as in that of Milton, the character of abysmal depth is

made poignant to feeling by insistence upon headlong motion; just as when
standing on some precipice edge, amongst peaks and chasms, one feels
their lines overpowering and terrible through the suggested anguish of
falling. That horror overcome adds a kind of emotional exultation to the
sight of actual mountain chasms [...]. When Coleridge dreams of measure
less caverns, when Plato tells of rivers that pour their waters even to the
earth's centre, or Milton's rebel angels fall nine days through chaos down
to Hell, the imagination, seeking something enormous, ultimate, to
express what strove unexpressed within experience, is satisfied (Bodkin,
1963: 104).
This passage contains two items necessary for the issue in hand, namely,
items that render caverns and abysses emotionally suitable places for Hades as
an ultimate place of punishment, and that are beyond the coping-ability of a
person incapable of being in uncertainties. In the first place, it is precisely that
kind of anxiety, or anguish, aroused by the possibility of endless falling (that
renders an abyss so meet a place for ultimate punishment) against which the
quest for certitude is defending itself by clinging to hard facts. Second, such
negative concepts as measureless, infinite become, on the one hand, positive
when they are presented as enormous, as exceeding the scope of perception or
of imagination; on the other hand, they become emblems of what may be called
"absolute size", and mediately, of "ultimate" experiences. As for the attitude
of the quest for certitude, there is nothing so frustrating for it as facing this
negative entity that exceeds the scope of imagination.
In summarising the contribution of the cavern to the archetype of Paradise
and Hades, Bodkin writes:
[T]he cavern depth appears as the objectification of an imaginative fear—

an experience of fascination it may be, in which the pain of fear is lost in
the relief of expression; in other instances the horror of loss and
frustration symbolised in depth, darkness, and enclosing walls sounds its
intrinsic note of pain even through the opposing gain and triumph that
poetic expression achieves (114).
Before my last quotation from Bodkin on caverns, I must observe that she
made negative capability an explicit and deliberate part of her method, in the
sense of "to let the mind be a thoroughfare for all thoughts". And thus, though
using a technique adapted from psychoanalytic practice, she uses the insight
thus gained to voice what might seem a heretic view of the psychoanalytic
interpretation of caverns as symbols:
Examining my own response to the cavern image, as it occurs in Co

leridge's poem, I find a complex of reminiscence, including memories of
damp dark cellars and of a deep well, regarded with fearful interest in child
hood; also, fused with these, images of caverns and underground castle-
vaults, goblin-tenanted, which I gathered from an absorbed reading of
fairy-tales. These memories include no recognisable reference to the womb
(113-114).
Surely her free associations were not influenced by a theoretical position

which, in general, she accepts. But in the rest of the paragraph she offers a
plausible explanation that may be also illuminating of the poem's emotional

symbolism:
If, however, we accept the view that the earliest conscious apprehensions
are conditioned by yet earlier responses of the organism—unconscious
'prehensions', in Witehead's phrase, inherited by later conscious 'occa
sions'—we have a means for conceiving how earliest experiences of the
infant in relation to the mother's body, especially the violent adventure
of birth, may help to determine the first conscious reactions to dark en
closed places, and may contribute psycho-physiological echoes to dreams
and to the play of fancy (114).
Now consider this: the womb as the "meaning" of caverns directs attention
to their enclosing (solid) walls; whereas Bodkin's foregoing discussion
focussed attention on the enclosed vacuous space. Thus, her reluctance to
acknowledge the womb-associations may be regarded as another piece of
evidence for her negative capability. The supplemented explanation at the end
of the paragraph points to possible "psycho-physiological echoes" contributed
by the birth-trauma to the unpleasant qualities associated with caverns.
Consider this too: Bodkin's foregoing argument presupposes a capability for
more than usually delayed closure. Instead of making a rapid equation CAVERN
= WOMB, Bodkin had to let her mind be a thoroughfare for all thoughts, that
observed this stream of consciousness from a higher point of view, and decide
that "these memories include no recognisable reference to the womb"; and
consider, finally, what the womb-association can contribute, in spite of all, to
the felt quality of the cavern-image. Such a stream of thought is, indeed, a fairly
reliable indication of negative capability.
Bodkin is working within the theoretical framework of the Jungian version
of depth-psychology. We could add now, in terms of recent cognitive
psychology, that she resorts both to a "top-down" and a "bottom-up"
strategy. On the one hand, she does begin with a preconceived hypothesis of
Jungian archetypes; on the other hand, she also begins with the minute and
subtle aspects of the poem's imagery, so the two approaches control and
modify each other. As a result, Bodkin's work does not make the simplistic
impression of the "relatively simple equivalences of symbolic criticism".
Moreover, Bodkin's "bottom-up" analysis of the imagery can be meaningful
and even convincing even for a reader who does not accept the tenets of
Jungian criticism. And if Jung's theory of archetypes can be shown to be
sound, it will suggest that certain images carry additional emotional force,
beyond the appeal revealed by the "bottom-up" analysis.
Concrete and Abstract Functioning
It is said that abstraction is the greatest intellectual achievement of man. It has

been found, however, that primitive animals, too, do abstract; and so do small
children and drunkards. Brown (1968: 268) tells about the male stickleback,
which assumes its fighting posture when confronted with another male
stickleback; it will make the same response to wooden decoys so long as they
are marked with a blob of red paint resembling the red mark on the underside of
the male stickleback. Gombrich (1961: 115) tells an old music-hall joke describ
ing a drunkard
who politely lifts his hat to every lamppost he passes. Should we say that
the liquor has so increased his power of abstraction that he is now able to
isolate the formal quality of uprightness from both lamppost and the
human figure? Our mind, of course, works by differentiation rather than by
generalisation, and the child will for long call all four-footers of a certain
size "gee-gee" before it learns to differentiate breeds and forms.
These abstractions were made in the absence of differentiation. The fish does
not distinguish the various things that are red on their undersurface; the
drunkard does not distinguish between the various upright things, nor the child
between the various four-footers (see Brown's illuminating discussion, 1968:
264-297). From this, two generalisations are suggested, one in the area of
psychology, and one in the area of aesthetics. It has been mentioned above that
psychologists distinguish between abstract and concrete personalities. The
latter are notorious for their quest for certitude. "Different syndromes of
interpretive, affective and behavioral tendencies accompany or underlie
concrete and abstract functioning".
More concrete functioning is expressed at the behavioral level by high

stimulus-response requiredness, the extreme of which could be illustrated
by such one-to-one correspondence as that between the stimulus of a light
and the taxic response of a moth. More abstract functioning, on the one
hand, because of a more enriched and complex mediational system and a
greater ability to transcend and depart from the immediate and perceptual
characteristics of the impingements, results in less absolutism, that is,
greater relativism in thought and action (Harvey, 1970: 315).
I make no claim to diagnose the empirical personality of the various critics.

But we can compare their styles of cognitive functioning with respect to their
relative concreteness or abstractness. Concrete functioning manifests itself,
inter alia, in a difficulty to detach oneself from the concrete properties of
objects or situations from which the abstractions have been abstracted. A
person with concrete functioning does not necessarily have difficulties in
handling abstractions, but he may have difficulties in distinguishing at the same

time between the concrete items from which the abstraction has been ab
stracted; the concrete person seems to be unable to handle simultaneously both
the abstractions themselves and those aspects of the objects that are irrelevant
to the abstractions—especially in domains requiring high ego-involvement such
as response to works of literature. In the light of the foregoing discussion, it
will be easy to understand, without long explanations, Wimsatt's conception
of the Concrete Universal as a major principle in art in general, and in
literature and poetry in particular:
What distinguishes poetry from scientific or logical discourse is a degree of

irrelevant concreteness in descriptive details. [...] The fact is that all con
crete illustration has about it something of the irrelevant. An apple falling
from a tree illustrates gravity, but apple and tree are irrelevant to the pure
theory of gravity. It may be that what happens in a poem is that the apple
and the tree are somehow made more than usually relevant (Wimsatt,
1954: 76).
Viewed in the present context, this conception of the Concrete Universal

may be regarded as an unexpected confirmation of one of the basic
assumptions of cognitive poetics: response to art in general, and to literature in
particular, makes special uses of responses evolved for the purpose of adap
tation to man's physical and social environment. It is a major principle both in
literature and in scientific or logical thinking, and unlike in primitive ab
straction, it must be accompanied by differentiation. In literature, the "irrelev
ant" concrete properties from which the abstraction has been made have
special significance, since they may become "more than usually relevant".
What is most important in criticism is that the distinction between the concrete
and the abstract levels must be carefully maintained. The abstract personality
can examine them separately or jointly, as the requirements of circumstances
may be. 9 One more comment on Wimsatt's phrasing is required here. When he
says "the apple and the tree are somehow made more than usually relevant",
the italicised phrase is not replaceable by all irrelevant details, but by some
irrelevant details, or a considerable part of irrelevant details, or the like.
Concrete functioning is characterised by an inability to change "mental sets"

and a tendency to cling to a single level of thinking. We may find, therefore,
"concrete functioning" even among professors of philosophy. This may
account for phenomena that otherwise would have to be regarded as
incongruous. We sometimes encounter philosophers or literary critics (or,
with the necessary changes, linguists) who develop, on the abstract level,
brilliant theories of interpretation, but in their practical criticism display most
of the symptoms of the Quest for Certitude (or of concrete functioning, for
that matter).
Let us compare now two passages concerned with Coleridge's description

of the site of the building enterprise, one by Yarlott, the other by Harold
Bloom.
So twice five miles of fertile ground

With walls and towers were girdled round.
These lines are placed in close and obvious antithesis to the caverns
"measureless to man" through which the sacred river runs. This juxta
posing of infinite and finite is deliberate, intended as ironic comment
surely upon the precise and mathematical details of Kubla's fussy little
paradise. The "measureless" caverns of sacredness suggest mysterious
eternities which the oriental despot, intent upon the mechanical
trigonometry of his palisades, scarcely guesses at. One recalls Coleridge's
own dislike of mathematics:
... though Reason is feasted, Imagination is starved; whilst
Reason is luxuriating in its proper Paradise, Imagination is wearily
travelling in a dreary desert.
Kubla's paradise, one may feel, was a paradise for reason to luxuriate in.
By adding "towers" to these walls moreover (another detail superimposed
upon Purchas), Coleridge seemed intent on emphasising the garden's
exclusive differentiation from the larger, spiritual life of the universe
(Yarlott, 1967: 132).
Kubla picks his spot with precision. A sacred river runs into the ground at
just the point where the great dome is decreed. Beneath the dome is the
underground river, running in measureless caverns down to a sunless sea.
The dome rises above an artificial paradise, ten miles in diameter,
including both elaborate gardens and ancient forests. Amid these forests is
a chasm from which a fountain suddenly bursts, part earthquake, part
geyser (Bloom, 1963: 230-231).
Intuitively, Yarlott's passage seems somehow to miss the point of the

description. What is more, it does this in a way that is typical of the "quest for
certitude" syndrome. To substantiate this, let us consider briefly the sentence
"The 'measureless' caverns of sacredness suggest mysterious eternities which
the oriental despot, intent upon the mechanical trigonometry of his palisades,
scarcely guesses at", and contrast it with Bloom's summary. If I understand
Bloom correctly, he means that the sublime aspects of the "measureless
caverns" do have to do with Kubla's picking precisely that spot for building
his stately pleasure-dome. Let us sharpen this possible suggestion, thus (for
argument's sake): "Kubla picked this spot precisely because the measureless
caverns of sacredness suggest mysterious eternities". What arguments can one
bring in support of each of these two opposite claims? Bloom could argue (and
Yarlott could hardly deny it) that the measureless caverns somehow suggest
sacredness and mysterious eternities (that is, the measureless in space suggests
the measureless in time). Kubla's motives have been left tacit; but a
considerable part of the description is devoted to the sublime aspects of the
landscape. As long as there is no indication to the contrary, one must assume
that he intended to choose such a spot. It is true that for the area surrounded
by walls and towers exact measures are given. But as long as there is no
indication to the contrary, one may assume that Kubla intended to have the
sublime and measureless in conjunction with exact measures. In short, one may
assume that whoever chose the spot and whoever built the walls and towers,
carried out Kubla's "decree" (unless otherwise indicated). A cause-and-effect
relationship may be even suggested by So (the question being whether it refers
only to girdled round, or to all the ensuing description). Furthermore, the
sublime quality abstracted from the various items is experienced with the
immediacy of perception. Kubla Khan vanishes in the background, and only
the perceived quality of the landscape pervades the foreground of the poem.
The oriental despot's exclusion from this perception has nothing to rely on in
the poem. On the contrary, the lines "And 'mid this tumult, Kubla heard from
far / Ancestral voices prophesying war!" suggest that the Khan did perceive
some of the awe-inspiring aspect of the scene. Now, what arguments can
Yarlott give against such a position and in favour of his own? It would appear
that his contention is wholly gratuitous, relying mainly on what he knows
about Tartar khans in general. But notice the following passage by Elizabeth
Schneider (1975: 250-251):
The historical Cubla was an attractive subject [...]. Though he shared the
usual adventures of the successive Tartar Conquerors with their wars and
prophecies of wars, he was said to be distinguished above the others by a
breadth of mind and a tolerance foreign to most oriental rulers. Marco
Polo, in Purchas, reported that Cubla expected persons of all religions to
pray to their own gods; toward Christians especially he was well disposed.
Most significant of all for a poet was his good name among authors.,
and so forth. Coleridge himself made, in 1799, a note about a recent Tartar
khan with roughly similar interests who, in the year 1783 "set on foot a
Translation of the Great French Encyclopaedia into the Tartar Language"
(ibid., 251). I am far from suggesting that Kubla Khan of the poem is anything
like an enlightened monarch. Nothing to the effect of the above information (or,
for that matter, to the contrary) has entered the poem. The implication of these
quotations is twofold: first, the occurrence of Khan in the poem does not
warrant all the "facts" attributed by Yarlott to Kubla; there were khans of a
different disposition (and Cubla happened to be one of them). Second, Yarlott
supports his interpretation by referring to Coleridge's sources on the one hand,
and to Coleridge's preoccupations as manifest in his works and notebooks, on
the other. This he did, apparently, in a rather selective manner, guided by his
prejudice against khans and, more fundamentally, by his need to "handle the
problem of ignorance by coming to a conclusion—any conclusion—in order to

avoid the anxiety that would otherwise arise".
Now consider another item in Yarlott's discussion above: "By adding
'towers' to these walls moreover (another detail superimposed upon Purchas),
Coleridge seemed intent on emphasising the garden's exclusive differentiation
from the larger, spiritual life of the universe". It may be quite significant that
towers is a detail superimposed upon Purchas. The question is what this
significance is. It is after all a concrete detail, relevant to the concrete descrip
tion of the building enterprise, but not to such abstractions as "the larger,
spiritual life of the universe". Yarlott seems to have mixed up the concrete and
the abstract levels of Coleridge's poem. He does not grasp a whole situation in
its essentials and abstract from it one overall quality, but is seeking a one-to-
one correspondence between the concrete details and their spiritual im
plications, much in the manner of concrete functioning. The same can be said
about Yarlott's excursus on Coleridge's aversion to mathematics. The exact
measures ("twice five miles") are details relevant to the concrete description,
but not necessarily to the poem's spiritual implications. They are, however,
made more than usually relevant by the contrast to "measureless to man" on
the one hand, and to "But oh! that deep romantic chasm" etc. on the other.
Now Yarlott "sharpens" this difference of immediate impression into a con
ceptual contrast. And this he does, again, by isolating the measures from their
context and tearing them out of their proper proportion, creating a one-to-one
relationship between the isolated detail and its significance. This is another
instance of his failure "to transcend and depart from the immediate and
perceptual characteristics" of the poetic description, to utilise Harvey's phrase
again (1970: 315).
A word should be said about Yarlott's comment "This juxtaposing of
infinite and finite is deliberate, intended as ironic comment surely upon the
precise and mathematical details of Kubla's fussy little paradise." It is almost
impossible to refute such an interpretation of the description. Furthermore, the
phrase "ironic comment" seems to be defendable. Does irony not involve the
juxtaposition, balance or reconcilement of opposite or discordant elements?
But so does imagination too, according to Coleridge's famous definition; and so
do ambivalent or conflicting emotions. How can we tell, then, whether
Coleridge's poem contains irony or ambivalent emotions, or is a product of
imagination? First of all one should note that irony, as used by literary critics,
has two fairly distinct meanings. Both suggest the "balance or reconcilement of
opposite or discordant elements"; one sense is limited to acknowledging the
presence of these elements in a work of literature; the other also implies such
attitudes as emotional detachment, superiority, freedom and amusement. In the
first sense there need be no contradiction whatever between emotion, irony and
imagination. Yarlott, however, seems to imply the second sense. And "Kubla
Khan" seems to be lacking in the elements that distinguish the second sense of
irony. As far as the critic's attitude is concerned the question is, again, to what
extent is he sensitive to the fine-grained "minimal cues", and to what extent
does he respond to "obtrusive cues". Actually, as Bloom and other critics
imply, the opposite elements are not merely "juxtaposed": the measured has
been superimposed upon the sublime and "measureless to man" in a way that
is not unlike poetic imagination; or, perhaps, the infinite is revealed in and
through the finite. But on the whole, neither the juxtaposition hypothesis nor
the superimposition hypothesis can be dismissed and both remain quite plau
sible. But if the juxtaposition hypothesis entails irony in the second sense, it
becomes at once less plausible.
Finally, one should raise the question: "What is the proper degree of
abstraction from the concrete images?" This is one of the most important
questions that one can raise with respect to interpretation, and perhaps the
most difficult to answer. This is, perhaps, the point where most responsible
interpretations differ from one another. And we seem to lack a conceptual
system, or even a vocabulary, to distinguish various degrees of abstraction. So,
instead of attempting to define the proper measure in the abstract, let us
examine a specific instance and generalise from it. Many critics have remarked,
in one way or another, that "Kubla Khan" is a poem about poetry. What I
propose to investigate now is their different ways of making that remark. Let
us begin with Watson's (1973: 227-228) statement:
"Kubla Khan", then, is not just about poetry: it is about two kinds of
poem. One of them is there in the first thirty-six lines of the poem; and
though the other is nowhere to be found, we are told what it would do to
the reader and what it would do to the poet.
Watson does not offer this as a hypothesis, but as an indisputable fact:
What is "Kubla Khan" about? This is, or ought to be, an established fact of
criticism: "Kubla Khan" is a poem about poetry. [...] Anyone who objects
that there is not a word about poetry in it should be sent at once to the
conclusion and asked, even if he has never read any Plato, what in English
poetry this is like:
Weave a circle round him thrice,

And close your eyes with holy dread,
For he on honey-dew has fed,
And drunk the milk of Paradise.
There are dozens of parallels in Renaissance English to this account of po

etic inspiration, all based—though rarely at first hand—on Plato's view of
poetic madness in the Ion or the Phaedrus. Shakespeare's banter about
"the poet's eye, in a fine frenzy rolling" in A Midsummer Night's Dream
is perhaps the most famous. The "flashing eyes" and "floating hair" of
Coleridge's poem belong to a poet in the fury of creation. Verbal

resemblances to the text of Plato itself confirm that the last paragraph of
the poem is a prolonged Platonic allusion. Socrates, in the Ion, compares
lyric poets to "Bacchic maidens who draw milk and honey from the rivers
when under the influence of Dionysus" and adds that poets "gather their
strains from honeyed fountains out of the gardens and dells of the Muses .
. ." Ion himself, describing the effects of poetic recitation, confesses that
"when I speak of horrors, my hair stands on end. . . ." The very phrase
"holy dread" is Platonic (Laws 671 D) (Watson, 1973: 226).
The parallel between Plato and Coleridge's passage is convincing enough.

But does this prove, beyond Watson's rhetoric, that "Kubla Khan" is indeed a
poem about poetry? Let us have a look at Schneider's discussion of the same
issue (from whom the analogy with Plato apparently has been adopted):
Essentially, the picture is but the ancient conventional description of the

poet with his "eye in a fine frenzy rolling". This conception was old even
in Plato's day, and practically every detail used by Coleridge was a com
monplace in it. The description derived a good deal from the accounts of
persons possessed by the god in Dionysus worship and the Orphic cults—
flashing eyes and streaming hair, as well as honey, milk, magic, holiness,
and dread (Schneider, 1975: 245).
Then comes a long quotation from Ion, including the passages quoted by
Watson. Then Schneider concludes:
Coleridge's Inspiration, music, holiness inspiring awe, milk and honey, are
all explicitly here; and the flashing eyes and floating hair are implicit in
the "Corybantian revellers" and "Bacchic maidens" (246).
The differences between Schneider's and Watson's positions are very small,
but rather significant. Schneider extends the poet's fine frenzy to "persons
possessed by the god in Dionysus worship and the Orphic cults"; that is, she
regards it as a more specific instance of ecstasy, inspiration, possession. In her
concluding paragraph she mentions such abstractions as "Inspiration, music,
holiness inspiring awe", but no poetic inspiration. Both Schneider and Watson
go outside the text for the substantiation of their interpretation, and both go to
the same texts. But Schneider, firstly, widens the scope of her external sources,
and views within it Plato and Shakespeare in proportion. Secondly, from them
she abstracts a higher abstraction which, by the same token, does not contain
elements that conflict with what is explicitly stated in the poem. Watson, on
the contrary, fails "to transcend and depart from the immediate and perceptual
characteristics" of Plato's description of the poet, and insists on his presence in
the last lines of the poem. This is concrete functioning. In such allusions only
the most general abstractions should be imported into the poem; from the more
specific levels of abstraction, only those meaning components from the source-
text (Plato) should be imported (if at all) that do not conflict with the more
specific components in the target-text ("Kubla Khan").
It is most illuminating to see what happens in Humphrey House's paper in
this respect. He states: "For 'Kubla Khan' is a poem about the act of poetic
creation, about the 'ecstasy in imaginative fulfilment"' (House, 1973: 201).
This sentence is far from achieving the precision of academic writing; but
seems to aim at a precision required for capturing evasive intuitions. It suggests
one thing, and then adds, as a casual afterthought, a correction as it were, that
appears to be more precise. At any rate, the phrases "poetic creation" and "the
ecstasy in imaginative fulfilment" are far from synonymous. But it is perhaps
an indirect indication of House's great sensitivity to "subtle and minimal cues"
and to their balance and proportion, instances of which abound in his Clark
lectures on Coleridge, especially in his chapter on the "Ancient Mariner". In
this context, what can be regarded as "academic ineptness" is, in fact, a means
for capturing his evasive intuitions, and as such, a kind of indirect evidence of
his negative capability. Two further remarks are required here. First, "ecstasy
in imaginative fulfilment" is a more general term than "poetic creation", in the
sense that "poetic creation" is a specific case of "imaginative fulfilment". The
more general term can, under a certain interpretation, describe Kubla Khan's
building enterprise just as well (what is more, certain critics speak of the
creative and destructive energy of the fountain as well). Second, House's more
general phrase is an acknowledged quotation from Bodkin (1963: 95), supply
ing another instance of her abstract functioning.
A similar strategy of self-correction (in a different matter) is followed by
Knight (1960: 165): "As Kubla Khan himself, if we bring him within our
scheme, he becomes God; or at least one of those 'huge and mighty forms', or
other similar intuitions of gigantic mountainous power, in Wordsworth".
Knight's paper abounds with elements that appear to be arbitrarily imposed
upon Coleridge's poem from an external intellectual scheme; at the same time,
it abounds in more-or-less direct indications of abstract functioning and the
tolerance of uncertainty. The immediate sequel of the foregoing quotation is an
instance of this:
Or we can, provisionally—not finally, as I shall show—leave him out,

saying that the poet's genius, starting to describe an oriental monarch's
architectural exploits, finds itself automatically creating a symbolic
universal panorama of existence. This is a usual process, since the poet
continually starts with an ordinary tale but universalises as he proceeds
(165-166).
Notice Knight's continual self-qualifications as well as his capability of

committing himself "provisionally", as indications of his tolerance of uncer
tainty. Notice also his version of the Concrete Universal in this passage, as a
possible indication of his abstract functioning; though one might well question
his assertion about Kubla's "creating a symbolic universal panorama of
existence".
Finally, I would like to return for a moment to Bodkin's discussion of
caverns. In this respect, too, she appears to display an inclination for abstract
functioning, by having recourse to such ultimate abstractions as cold, dark, or
at least such thing-free qualities as stagnant air.
Alliteration, Onomatopoeia and Decision Style
There is general consent that the appeal of "Kubla Khan" is intimately related
to its music. This music consists in rhythmic structure, rhymes, and sound
patterns called alliteration and assonance. In Chapter 2 I will discuss at some
length the hypnotic effect of the poem's rhythm and its interaction with
rhyme scheme, alliteration and assonance. Here I will confine myself to alliter
ation and assonance in themselves. In this section I will briefly explore the
sound patterns of poetry from three vantage points: EXPERIMENTAL PHONE
TICS, GESTALT THEORY, and STRUCTURALIST MODELS of the relationship be
tween sound and meaning. Then I will compare, in light of the theory
expounded in the present chapter, the different ways in which two critics
handle the sound patterns in "Kubla Khan". I will also set forth a rudimentary
account of ALTERNATIVE MENTAL PERFORMANCES of sound patterns.
EXPERIMENTAL PHONETICS. The sound dimension of poetry is something of
an embarrassment for many critics who seek relief in rapid closure by shifting
attention from sound to meaning. In ordinary language we use words for a va
riety of functions most of which rely on their meaning, conveying information
that performs some speech act, or relieves ignorance. In most instances the
phonetic component is merely a vehicle for transmitting meanings. Conse
quently, language-users are programmed to move on from speech sounds to
meanings as fast as possible. In poetic language, by contrast, rhyme, metre, and
alliteration force us to linger for longer at the sound stratum than in ordinary
language. Psychologically, one "comes to a rest" only when reaching the
meaning. Any delay in the transition generates uncertainty.
Words have, then, a phonetic-phonological10 component (speech sounds),
and a semantic component (meaning). Words are transmitted by acoustic
information and immediately recoded into strings of phonetic units which, in
turn, are immediately recoded into semantic units. Nevertheless, there is
experimental evidence that some subliminal sensory information is available in
speech perception, facilitating certain verbal memory tasks as well as
10
For our purpose we need not distinguish between phonetics and phonology. But
M.A.K. Halliday's distinction may be illuminating: If speech is organised noise,
phonetics investigates the noise, phonology investigates its organisation.
generating expressive and imitating potentials of the speech sounds. Under

certain conditions this sensory information is amplified; and under others it is
inhibited. "Lateral inhibition" designates certain interactions between neural
responses in adjacent brain areas. In speech perception, there is lingering audi
tory information about the most recent arrival. Lateral inhibition occurs when
newer arrivals damage information that otherwise remains available in sensory
form, following auditory presentation.
Elsewhere I explore this aspect of the perception of alliteration at consider
able length, quoting inter alia personal communications from Bruno Repp of
the Haskins Laboratories and Robert G. Crowder of Yale University. Repp as
serts that "if a subsequent stimulus [...] is very similar to the preceding
stimulus, it may generate an enhanced response, because of integration with the
previous response; if it [...] is moderately similar, it will be reduced; if it falls
outside that area, it will be unaffected" (Tsur, 1992b: 37-38). Crowder suggests
that "the lateral inhibition model specifically includes the possibility that if the
two sounds residing together in auditory memory are close enough to one an
other, acoustically, their effect will combine rather than engage in inhibition.
There would be precedent for the assumption that the total effect would be the
larger for having had a repeated sound" (Tsur, 1992b: 38). This may account
for the enhanced effect of speech sounds in rhyme and alliteration.
GESTALT THEORY. Gestalt theory predicts that in poetic prosody such
amplification of the acoustic effect is "double-edged": it may have opposing
results, depending on its placement within the rhythmic gestalt. When the
repeated sounds occur in consecutive (weak and strong) positions (as in "five
miles" or "miles meandering", or "from far"), they tend to blur the contrast be
tween prominent and non-prominent events; the amplified acoustic impact
enhances this blurring activity. When they occur in stressed syllables in strong
positions (with one or more intervening positions), they drastically improve
the gestalt, by enhancing the contrast between prominent and non-prominent
events.
STRUCTURALIST MODELS. Structuralist criticism has explored the principles
that govern the meaning of sound patterns in poetry. In two classical papers
on the issue Benjamin Hrushovski (1968; 1980) claims that much of the dis
pute whether sound can or cannot be expressive comes to a dead end because
the issue is treated as if it were one phenomenon. "As a matter of fact, there
are several kinds of relations between sound and meaning, and in each kind the
problem is revealed in different forms" (1968: 412). He discusses four kinds of
such relations: a. Onomatopoeia; b. Expressive Sounds; c. Focusing Sound
Pattern; d. Neutral Sound Patterns (I have further elaborated Hrushovski's
conception at great length [Tsur, 1992b; 2001]). Onomatopoeia is the most
obtrusive of the four types, and also the most marginal one in poetry.
From the point of view of the issue discussed in the present chapter we may
observe that the four possibilities demand less or more delayed closure,
exerting lesser or greater pressure on one's tolerance of uncertainty. Ono

matopoeia allows the most rapid closure: it allows the most direct transition
from speech sounds to the "imitated" meanings. In onomatopoeia the language-
user detects some resemblance between recurring phonetic features in a stretch
of language and certain natural noises to which the words may refer. In an
expressive sound pattern closure must be delayed for considerably longer, for
two different reasons. First, because it requires more complex processing: "a
sound combination is grasped as expressive of the tone, mood or some general
quality of meaning. Here, an abstraction from the sound pattern (i.e. some kind
of tone or "quality" of the sounds) is parallel to an abstraction from the
meaning of the words (tone, mood etc.)" (Hrushovski, 1968: 444). Second,
while the meaning component of the words in onomatopoeia may designate
palpable objects or events, in expressive sound patterns one must abstract an
evasive mood or tone or quality from them. Eventually, in expressive sound
patterns one must settle on an evasive mood or tone or quality, not a solid
object or categorisable event. Likewise, in neutral sound patterns one must
remain with an aesthetic function as target, not anchored in some solid object
or substantial event; in fact, not anchored in any meaning at all. In what
follows, I will compare the handling of sound patterns by two critics, one who
displays great tolerance of meaningless sound patterns blurring each other, and
one who goes out of his way to force unwarranted meanings upon those sound
patterns.
Elizabeth Schneider treats the sound patterns of "Kubla Khan" as belonging
to the latter, "neutral" category. This she does in a way that exerts the greatest
demand on one's tolerance of ambiguity. Consider the following paragraph,
with the ensuing diagram: "The opening line, 'In Xa'nadu' did Ku'bla Kha'n',
receives its primary shape from the inclosed assonance of its four stresses, a—
u—u—a, which swings the sound as if in a shallow curve, the symmetry being
still further marked by the full rhyme of the inclosing syllables, Xan- and Khan
(Coleridge undoubtedly pronounced Khan as it was often spelled, Can) and the
embellishment of minor echoes, d's and short i's binding together the first part
and it's the end of the line" (Schneider, 1975: 274):
Figure 1 Sound patterns of the first line (re-drawn from Schneider)
Schneider attributes no meaning to these sound patterns. Rather, she finds

certain regularities in them and points at certain organising effects these
patterns have. I will not quote here all her examples; for brevity's sake, I will
just reproduce an adaptation of her second diagram, reflecting the rich and
interwoven network of sound patterns in the first stanza (ibid., 275):
Figure 2 Sound patterns of the first stanza (adapted from Schneider)
Now consider Schneider's following passage: "The tune and variations to be

played upon the œ-sound are established by the first sentence, in which three
of the five lines begin and end with stresses upon it. But this effect is overlaid
or interwoven with an elaboration so intricate that one could scarcely point to
its mate in English poetry if we except the more subtle harmonies of Milton
and Bridges" (Schneider, 1975: 275). She refers here to a poetic quality that
severely tests the critic's abilities. First, we have no appropriate metalanguage
to describe the quality (only the patterns that generate it). Second, it strains
the reader's tolerance of ambiguity to the utmost. Being arrested at the mean
ingless speech sounds ("Neutral Sound Patterns") of a poem makes, by itself,
great demands on one's tolerance of uncertainty. If this is not enough,
Schneider increases the strain on one's intolerance of ambiguity by treating
sound patterns as blurring each other. The perception of this quality presup
poses the perception of a thick texture of interwoven sound patterns which
blur each other, render each other "ambiguous"—in the sense of vague,
indistinct.11 Such a texture may be intolerable for a person for whom "the
unique, unclassifiable sensation is particularly offensive", to use Ohmann's
phrase above. As we shall see in the next chapter, such blurring may make a
decisive contribution to the hypnotic effect of the poem.
It is impossible to do justice here to all the subtleties of Schneider's
discussion. Let us just discuss one more passage in which she dwells on
another aspect of the phonetic texture in which, too, the distinct perception of
the speech sounds is considerably reduced, contributing to what she describes
as a "floating effect":
11
Roman Jakobson (1956: 371-372) explores the sophisticated network of
sound patterns in the last stanza of Edgar Allan Poe's "The Raven" but makes
no reference to their relative distinctness. Poe's "The Raven" and "Ulalume"
are paradigmatic instances of "hypnotic" or "spell-weaving" poetry. Jakobson
is interested in those sound patterns only as features that distinguish the poetic
function of language from other functions, not as features that distinguish
hypnotic from non-hypnotic poetry.
Often throughout the poem he repeats his old device of foreshadowing the
terminal rhyme by a preceding echo of assonance or alliteration—"sinuous
rills," "chasm which slanted," "ceaseless turmoil seething," "mazy mo
tion," "river ran" and "measureless to man" each used a second time,
"from far," "mingled measure," "loud and long." 12 This device, used
skilfully as it is here and partly concealed by the interlacing of other
patterns, contributes something to the floating effect of the whole, for the
assonance softens the impact of the rhyme and so lessens its tendency to
bring the line to earth at the close: the terminal rhyme does not settle so
heavily upon the mind when its emphasis has been partly stolen by its
preceding shadow. The forward movement is made to pause and "oscillate"
further at times by the considerable number of lines in which the meaning
looks forward while the rhyme looks back (Schneider, 1975: 276).
ALTERNATIVE MENTAL PERFORMANCES. Schneider is one of the few liter

ary critics who does not merely acknowledge the presence of rich sound
patterns, but also distinguishes between relatively foregrounded and relatively
backgrounded sound patterns. Furthermore, she does not do this impressionis-
tically, but gives a structural description of the phenomenon. Since there is no
appropriate metalanguage for this, she describes the various linguistic aspects
that point attention in conflicting directions. I have, however, doubts as to her
use of this specific type of alliteration. These concern the assertion that in the
above instances the "emphasis [of rhyme] has been partly stolen by its
preceding shadow". While I appreciate this kind of description very much, I
believe there is evidence that the "preceding shadow" enhances rather than
weakens the acoustic traces of the rhyme word. Obviously, we receive our
impressions from different mental performances.
In the above paragraph, Schneider provides us with a detailed description of
the elements of the mental performance in which "the terminal rhyme does not
settle so heavily upon the mind". I will argue that the suggested quality does
not arise universally but only in certain specifiable performances. Consequent
ly, this indicative sentence must not be understood as a factual statement, but
rather as a "crucial recommendation" how to perform the text. I opt for a
slightly different mental performance of these sound patterns, one in which the
12
All the sound patterns mentioned in this passage involve continuous (as
opposed to abrupt) speech sounds, all but one pair being periodic (vowels,
liquids and nasals). According to the received view, these are the most musical,
the most harmonious speech sounds. Elsewhere I argued (Tsur, 1992b) that
these speech sounds derive their musicality from their periodicity on the one
hand, and, on the other, from the fact that the greatest amount of pre-
categorial sensory information is available to consciousness in precisely these
speech sounds. According to Snyder (1930: 51-52), it is precisely these speech
sounds that are especially prominent in hypnotic poetry, so as "to satisfy and
soothe the ear".
repeated sound in such phrases as "mazy motion" or "mingled measure"

enhance each other's acoustic traces rather than blur them. This reflects my
intuitions and also follows from the models of "lateral inhibition" and Gestalt
Theory mentioned above. The repeated word-initial sounds in each phrase are
sufficiently similar to acoustically augment each other; and they occur in
stressed syllables in strong positions, with an intervening unstressed syllable
in a weak position, achieving a sense of considerable stability at the close.
I perfectly agree with a reading in which sound patterns have different
weights; I also commend Schneider's technique of pointing out the divided at
tention that involves forward-looking and backward-looking linguistic aspects;
but I rely on other devices by which attention is divided. One basic assump
tion of gestalt theory is that elements may change their nature when entering
into a wider context. Consider, for instance, "miles meandering . . . mazy
motion". In my reading, the last two words form an exceptionally strong
gestalt, "settling quite heavily upon the mind", or closing the line "with a
click"; but it is the preceding two words that mitigate this pattern, by
weakening the gestalt.
Let me spell out the perceptual dynamics involved. In such phrases as "river
ran", "measureless to man", "mingled measure," "loud and long", the repeated
word-initial sounds occur in stressed syllables in strong positions (with one or
more intervening positions), reinforcing, as I said, the contrast between
prominent and nonprominent events. It is the second token of the same
consonant that turns the first token into part of a sound pattern, after the
event: the two form a closed, symmetrical, "strong" unit. In a longer sequence
of repeated consonants, by contrast, as in "miles meandering . . . mazy
motion", the m's become part of an endlessly repeatable pattern which has no
self-generated end. Moreover, in "miles meandering" the repeated m occurs in
consecutive positions, strong and weak. Such a sequence has a considerably
weakened gestalt.
The very fact, however, that Schneider adopts a mental performance that
leads to the perception of some blurred sound texture gives us additional
evidence of her exceptionally high tolerance of ambiguity—irrespective of
whether I agree with her analysis or not.
Certain poetic effects (perhaps all) crucially depend on the "mental per
formance" in which they arise (more on this in Chapter 3 and the Afterword).
When there is unanimity as to the poetic qualities of a text, mental or vocal
performance usually go unnoticed; but when different poetic qualities are
reported by different readers, one should look for differences in vocal or mental
performances (Part Three of this book is devoted to vocal performance). I have
used, with reference to sound qualities, the near-vacuous phrase "Schneider's
mental performance". I have access to my own mental performance, but not to
Schneider's. As long as I don't specify the mental performance, the term is
vacuous. Now I propose to fill this phrase with some content. What kind of
mental performance may yield perceptions like the one reported by Schneider?
I wish to propose two possible kinds, based on three cognitive principles. One
principle (which we have already encountered in Repp's and Crowder's
communications) concerns lateral inhibition in the neural system, reducing
lingering auditory information ("if the subsequent stimulus is moderately
similar, it will be reduced"). The phrase "from far" (pointed out by Schneider)
would be an obvious example; in most performances the alliteration would not
be noticed at all: the fr sound cluster would be sufficiently dissimilar in the two
words to inhibit mutual enhancement. Accordingly, Schneider may have
moderated the similarity of some pairs of words in her performance so as to re
duce the resonance of their acoustic energy.
Since Schneider's discussion of the sound texture presupposes increased
rather than reduced response to resounding sensory information, I prefer an
alternative possibility, based on two further principles. One principle (which
we have also encountered already) is the gestalt assumption that elements
change their nature when entering into a wider context. Consider the phrases
"river ran" and "measureless to man" in their second occurrence. These
alliterations contribute to a clear-cut, enhanced articulation of the line ending,
typically marked by separate intonation contours. By the same token,
however, they direct attention away from the similarity of the rhyme words
"ran - man" to the similarity of their initial phonemes ("river ran" and
"measureless to man"). In Schneider's cogent formulation, "the terminal rhyme
does not settle so heavily upon the mind when its emphasis has been partly
stolen by its preceding shadow. The forward movement is made to pause and
'oscillate' further at times by the considerable number of lines in which the
meaning looks forward while the rhyme looks back". The rhyme words become
part of two different sound patterns. This is where the other principle comes
in. As will be argued in the Afterword, colour interaction in visual perception
and overtone interaction in music are enhanced within gestalt boundaries, but
inhibited across them. In speech perception there is experimental evidence for
a tendency to perceive certain rhythmic phenomena as active within intonation
contours, but not across contour boundaries. It seems quite plausible that the
perceptual corollaries of such split attention may effectively interact only in a
performance in which the line endings are not articulated too clearly by
intonation. The resulting interaction of precategorial sensory information
would generate an intense and diffuse gestalt-free texture. Such a performance
is deliberately called for in the first occurrence of "river, ran" and "measureless
to man"; but may, perhaps, be adapted to their second occurrence too. When
line endings are clearly articulated by intonation, such alliterative patterns are
likely to contribute a feeling to this articulation that the line has been closed
with a "click". Schneider seems to assume less clearly closed verse lines, and an
interaction of crisscrossing sensory information between the lines, which result
in sound patterns "partly concealed by the interlacing of other patterns".
When I was preparing the present, enlarged edition of this book, I stumbled
upon a recent book on "Kubla Khan" (Fleissner, 2000) that offers extensive
discussions of sound patterns. I have thus been able to compare the way he
and Schneider handle them, obtaining useful information and insight into the
implied critic's decision style. Fleissner was well acquainted with Schneider's
book quoting it fourteen times (he also lists the first edition of the present
book in his bibliography, but does not quote it). Not surprisingly, there is
some overlap between the two books on this issue. The more conspicuous,
therefore, becomes the fact that in most respects their approaches are
diametrically opposed.
Fleissner handles the sound patterns of "Kubla Khan" very differently. To
be sure, as I said, there are some overlaps between the treatments of these two
critics. First, it is not surprising that on the observational level they point out
roughly the same sound patterns. This is inevitable in describing the same
poem. Second, the following passage is undoubtedly compatible with Schnei
der's position, perhaps influenced by it (though less clearly articulated):
measureless to man. The most obvious effect of this pleasant, harmonious

phrase is the alliteration of the m sound, which then relates to the exten
sive alliteration in the first five lines of the poem itself, thus forming a
kind of structural unit:
Kubla Khan
-dome decree
river, ran
measureless . . . man
sunless sea (ibid., 14)
But then Fleissner brings the repeated sound pattern under the control of
meaning:
What is especially apropos is the use of the m sound for describing the
caverns onomatopoetically.
That is puzzling. I know what kind of sound the "murmuring of innumer

able bees", for instance, or "the silken, sad, uncertain / rustling of each purple
curtain" may have, but I don't know what kind of sound a cave has, or a
cabbage, for that matter. But Fleissner does know:
It suggests that something deep, remote, and sunken is in the caves—even

as the sound-effect BOUM describes the echo in Forster's Marabar Caves.
That sound-effect suggests not only what is incidentally non-human, but it
happens to connote the Hindu word for God, OM, even as measureless to
man conveys a similar anagogic effect. Phonologically, both expressions
point to the notion of infinity. The "infinite" phrase is echoed later in
the poem ("reached the caverns measureless to man"), thus serving to

stress the import of the sound effect. Comparable is the extended double
use of alliteration and onomatopoeia in the repetition of the m shortly
before: "miles meandering . . . mazy motion" (ibid., 14-15).
All this is ad hoc and the result of unbridled free association, nothing
principled. Fleissner does assign a proper description to some of Coleridge's
sound patterns, but cannot refrain from attributing a conspicuously arbitrary
meaning to them. Moreover, he cannot point out the principle according to
which he assigned those meanings. The word BOUM does not occur in, and is
not implied by, the poem, but is smuggled in through the backdoor, by way of
association with another text. Even if we grant the unwarranted association
with "the Hindu word for God, OM", this has nothing to do with onoma
topoeia. And I cannot figure out how can "phonologically, both expressions
point to the notion of infinity", nor how the phoneme m can imitate a "mazy
motion". What is the typical sound of infinity or of mazy motion? Pace his
deliberate intention, Fleissner's ensuing sentence points up, by way of
contrast, the absence of such a principle:
Comparable to the familiar Tennysonian expression "And murmuring of

innumerable bees," Coleridge's use of the m device was effected with intu
itive surety (ibid., 15).
As I have said before, I have an idea what sound the "murmuring of innu
merable bees" may have. I may also assume that there are principles by which
such murmuring can be related to speech sounds. The same principles must
also explain that, as I. A. Richards suggested, in the "murdering of innumerable
beeves" there is no onomatopoeia (I have elsewhere discussed those principles
at very great length; see Tsur, 1992b; 2001; in the latter, sound files, too, are
provided). Tennyson's line of verse contains the sound cluster m әr three
times: twice in murmuring, and once in innumerable. Richards' transcription
contains it only twice. That would still suffice for a quite effective onomato
poeia. Nevertheless, it disappears. The semantic component of the words
murdering and beeves does not activate the relevant features of the m әr
cluster, whereas that of murmuring and bees does. It would be illegitimate to
claim that the alliteration indicates that the beeves were mooing during the
hecatomb. The rule appears to be this: even in straightforward onomatopoeia
phonetic features cannot "conjure up" objects or events not mentioned in, or at
least strongly implied by, the text. Conversely rather, dormant phonetic fea
tures of speech sounds may be activated by events explicitly mentioned in the
text, events that contain auditory features that bear sufficient perceptual
similarity to them.
Or consider the following example from Fleissner:
woman wailing for her demon-lover. The alliterative use of the w here
(woman, wailing, waning) is noteworthy in that it is also onomatopoeic.
Let us infer that, among other ghoulish things, the phrase summons up
nothing less than the ululations of the she-wolf (ibid., 29).
Here, again, Fleissner anchors the meaningless sound patterns in an

arbitrarily introduced concrete image, "the ululations of the she-wolf". There
seem to be no constraints on Fleissner's stream of consciousness. Even
supposing that the repeated speech sound w does have the appropriate pho
netic features, these can be activated by wailing, but cannot conjure up a she-
wolf, or its ululation.
One of Agnon's characters says of her husband that "he is fond of making
assumptions, regarding them as certainties". One of the characteristics of con
crete functioning is an inclination to introduce unmentioned concrete elements
into a text and to regard them as certainly present. As we have seen, the
elements introduced into a text by way of interpretation must be of a consider
ably high level of abstraction, and must be regarded not as facts, but as "merely
possible". In this respect, Fleissner's "ululations of the she-wolf" and "the
sound-effect BOUM" are comparable to Fruman's "the woman wailing for her
demon-lover is calling for an incestuous lover", or Yarlott's assumption that
her wailing may reach the Khan's ears (see Chapter 2); or Watson's assertion
that "'Kubla Khan' is about two kinds of poem". This is the case even regard
ing Fleissner's comment on sunless sea. "Again an onomatopoeic meaning
dominates: continuity of sibilants suggests the surge of the sea". As I said, the
presence of certain phonetic features cannot conjure up an unmentioned object
or event. The sibilants are, indeed, there; and they may certainly serve as
imitation of all sorts of noises, as in Poe's "And the silken, sad, uncertain /
rustling". The only trouble is that Coleridge's phrase refers to a conspicuously
visual state, not to the noise made by the sea. The "surge of the sea" is
introduced here gratuitously. This is clearly indicated by the first stanza by
itself. But if anyone has doubts about this, the suggestion that it is a "lifeless
ocean" must convince him. This is an exquisite instance of what Harvey
described (see above) as "a greater insensitivity to subtle and minimal cues and
hence a greater susceptibility to false but obtrusive cues".
Throughout the present book I have pointed out evidence for Schneider's
high degree of Negative Capability; her handling of alliteration and assonance in
"Kubla Khan" is just one more instance of this. I have not devoted comparable
attention to Fleissner's book. But his discussion of onomatopoeia is embedded
in a book that displays a persistent attitude of a Quest for Certitude (cf.
Chapter 2, note 5). One important assumption of the "implied critic's decision
style" theory is that the avoidance of certain critical solutions does not
necessarily result from the critic's decision style; it may be due also to an
unsatisfactory conceptual system. If a critic avoids certain critical tools, he
may simply be unaware of their existence. In Fleissner's case, we have
evidence that he was exposed to Schneider's treatment of sound patterns. If he

offered a radically different one, he may have believed that Schneider's was not
good enough.
"Kubla Khan" and Depth-Psychological Interpretation
This poem is a depth-psychologist's paradise. A poem purported to have been

written in an opium dream (according to the 1816 preface), or in a state of
Reverie (according to the more recently discovered Crewe holograph manu
script) surely bears the signs of its origin in the unconscious. Does not
Coleridge himself testify in the preface that this fragment was published, "as
far as the Author's own opinions are concerned, rather as a psychological
curiosity, than on the ground of any supposed poetic merits"? And there are,
indeed, quite a few depth-psychological interpretations of this poem, of which
we shall consider and compare two (one Freudian and the other Jungian). As a
preliminary to this discussion, we shall consider two caveats, by two critics
whom we have already met. The first one is by Schneider (1975: 9-10):
Supposing Freud right in finding sexual symbolism to underlie almost all

human action, thought and dream—obviously, then, sexual symbolism
must underlie all poetry too. That, however, tells us little about any one
poem. The psychoanalytical critic's responsibility, in the interest of
clarity, is to make known his assumptions and tell his readers which of
several activities he is at the moment engaged in: whether he is using a
work of literature as a case history to teach us psychology or whether, on
the other hand, he is using his psychoanalytic theory to illumine a
particular literary work or literary history as a whole. The trap difficult to
avoid is that of mistaking what it is one has proved. If rounded mountains
always in human experience must mean breasts and caverns always wombs,
one might write an illuminating essay on infantilism and regression in
romantic poets, provided one can prove that they describe more mount
ains and caverns than other poets do. [...] In any case, it is difficult to see
how this kind of interpretation can throw light on any given poem unless
it can show something special in the use of caverns and mountains that is
not present in other cavern-mountain poems.
I have already suggested that Schneider gives ample evidence of her Negative
Capability, both in what she writes about the text itself, and in her criticism of
preceding critics. The above passage is obviously an instance of the latter. In
the first place, this is one more instance of Schneider's many (sometimes
sarcastic) remarks against symbol-mongering (such as "If one proceeds upon
the belief that one cannot open his mouth without being symbolic ...", 260). As
I suggested earlier, symbol-mongering may be a preferred means for dispelling
ignorance and uncertainty in the service of the Quest for Certitude. In the
second place, Schneider sharply criticises psychoanalytic (or perhaps pseudo-

psychoanalytic) attempts to use false but obtrusive cues that never change; in
short, she is also protesting against the refusal to change mental sets. Finally,
there is here an objection to the levelling of the difference between occurrences
of caverns and mountains in various poems. This, on closer inspection, turns
out to be a version of Wimsatt's Concrete Universal: if, on one level of
abstraction there is a universal sexual symbolism, on a more concrete level
there must be an "irrelevant texture" which is somehow made more than
usually relevant to the poem, and this relevance must be acknowledged by the
interpreter. We might add that if we equate mountains with breasts and caverns
with wombs, we turn a very partial identity into a complete identity; though
both breasts and mountains obtrude and are perhaps round, there are many
conflicting components in them, which we suppress (level out). caverns
The second caveat contains advice too, as to the appropriate measures to be
observed. Speaking of some critics' objection to "the Freudian reiterations
concerning sexual origins", Bodkin (1963: 113) says:
Enjoyment of the beauty of poetry is spoiled only if certain of these

psycho-physiological echoes are emphasised, as though they were some
how more real than all the other elements with which in a mature mind
they are fused—as though these elements that contribute to the actually
experienced response were a mere evasion or disguise of those few
primitive elements newly identified by the analyst.
In terms of our present inquiry, Bodkin's discussion may be regarded as a

"crucial recommendation" for the appropriate critical performance of poems.
First of all, the psycho-physiological elements are to be treated as echoes
rather than entities in the focus of attention, that is, as elements that give a
certain tint to the poetic text rather than determine the hard core of its meaning.
In other words, the psycho-physiological element is to be perceived according
to its proper dimensions in a complex whole rather than reducing the complex
to this single element. This suggests that separately each one of the elements
that constitute the complex is a classifiable, general norm, whereas the complex
whole may have a unique, unclassifiable quality (that is especially hard on the
attitude of the Quest for Certitude). The perceived ecstatic quality is, in
Sibley's (1962) term, an aesthetic quality; which in turn is, in Beardsley's
(1958: 83-88) term, a regional quality, that is, a quality that belongs to the
whole but not to any of its parts.
In the remainder of the present section of this chapter I am going to compare
two depth-psychological treatments of this poem, Fruman's (1971: 396-401)
Freudian reading, and Bodkin's Jungian reading. Such a comparison will enable
us to gain a double perspective on these pieces of criticism. On the one hand it
may show to what extent, and under what conditions, they may be illuminating
of the poem as an aesthetic object. On the other hand, it may illuminate some
of the intricacies of inferring the implied critic's decision style from the piece
of criticism discussed.
One important aspect of this comparison concerns the focus of interest of
the two critics. Whereas Fruman seems to be interested in concealed meanings,
Bodkin's main preoccupation is with emotional patterns. Fruman offers, then,
a Freudian reading of the poem. Since Coleridge himself says his poem is a
dream or reverie, one should ask, what dome-shapes are domes of pleasure in
dream symbolism?
Features of this kind, appearing in reveries or dreams, easily lend them

selves to interpretation as sexual symbols. But why confine the pleasure-
dome to a breast symbol? [...] a pleasure-dome in the vicinity of "caverns
measureless to man" suggests far more the mons veneris, especially as we
hear at once of "fertile" grounds and "forests", almost classic symbols for
primary sexual terrain (Fruman, 1972: 396).
It should be noticed, in the first place, that vicinity of is a well chosen

expression of Fruman's, rather than one used in the poem. If we prefer the
preposition above, "a pleasure-dome above the caverns measureless to man",
then we must return to the dome-as-breasts interpretation. Now, irrespective
of this, whether breasts or mons veneris, what does this indicate about the
poem as a poem? Very little indeed. The upshot of the sexual ingredient in the
"dream-poem" seems to be that:
The poem leaves us with a nameless crime. The dreamer is tabooed

because he has had a forbidden experience. The arrows all point in the
direction of incestuous love (401).
These arrows themselves are not very convincing. One of them is "the
woman wailing for her demon-lover is calling for an incestuous lover" (397).
The other one may be somewhat more convincing, in a context of dream
symbolism:
A black girl in the dream of a white man may of course mean many things,
depending on the dreamer's private beliefs and associations. Yet as a
generalisation it is true that powerful interracial fears or yearnings suggest
the problem of incest (400).
Now, what does such an interpretation contribute to the poem? First and
foremost, it isolates certain images and loads them with sexually "hot"
meanings. These meanings are specific, and have very little basis in the poem's
literal meaning. A very partial identity like the shapes of domes and female
breasts or, for that matter, of the mons veneris, is elaborated as a complete
identity, introducing a host of ingredients that have nothing to do with the
poem's theme, or with its emotional tone. In other words, one cannot defend
this kind of interpretation of pleasure-dome even on the grounds that the
hidden sexual symbolism may account for the emotional effect regularly
associated with the poem. 1 3 It also turns a poem that has a certain degree of
unity into discrete symbols, the interpretation of which can hardly yield a
coherent reading.
Apparently, the incest-theme has been introduced either in order to yield an
interesting reading, or to account for the horror (or magic rites) in facing the
"tabooed" person speaking in the last few lines. But the "holy dread" sounds
more like a response to the numinous than to incestuous love. 1 4 Curiously
enough, Fruman himself mentions the evidence that might suggest the
numinous powers attracted to the youth in the last stanza: "Against whom and
what is it an offense to feed upon honey-dew and drink the milk of paradise? "
(400, my italics); and again, in a negative context though, he mentions "food
and drink reserved for the gods" (401). But being trapped as he is in his
conception, he is looking for crimes and offenses. Now the "numinous"
element at the end of the poem can account not only for the "taboo", but also
for the ecstatic quality in the poem.
Now let us compare these specific meanings associated with the poem's
"sexual symbolism" to the last paragraph of Bodkin's discussion of this poem:
As in the preceding essay we traced a pattern of rising and sinking vitality,

a forward urge and backward swing of life, reflected in an imagery deployed
13
The case is different with Fruman's speculations on the relevance of the
Bacchus cult to his interpretation. The first and last sentences in a quotation
from Patricia M. Adair suggest: "The ecstasy which possessed the women who
followed Bacchus was a fierce and blind desire to be lost in the daemon-god.
[...] The original Crewe manuscript of Kubla Khan confirms this association
by spelling the 'demon-lover' of the 1816 version as 'Daemon Lover'" (397).
Unlike a putative "female-breast quality" or "mons-veneris quality", there is
an ecstatic quality in Coleridge's poem, regularly associated with it by most
readers. But, then, the evidence for the Bacchic association seems to be
twofold: the ecstatic quality and the reference to the "Daemon Lover". From
this point of view, the claim that the allusion to the Bacchus cult reinforces
the ecstatic quality in the poem turns out to be circular. Moreover, the
ecstatic quality in the poem is readily accessible to any reader ignorant of the
Crewe manuscript spelling or the identification of Bacchus as a daemon-god.
All this, however, may be relevant to Fruman's musings about Coleridge's self-
identification with Bacchus. But this brings us back to the issue raised by
Schneider: "The trap difficult to avoid is that of mistaking what it is one has
proved", something about the poem, or about its author's psychology.
14 It is Bodkin (1963: 102) who introduces the reference to Rudolf Otto's dis
cussion of the numinous in his The Idea of the Holy in relation to this poem,
but not necessarily this passage.
in time [...] so now we find an emotional pattern of somewhat similar

character presented statically, in imagery of fixed spatial relation—the
mountain standing high in storm and sunlight, the cavern unchanging,
dark, below, waters whose movement only emphasises these steadfast rela
tions of height and depth (Bodkin, 1963: 114-115).
This paragraph reveals no interest in specific meanings, but rather in

abstract emotional patterns expressed in abstract nouns and pairs of antony-
mous adjectives (e.g., "rising and sinking vitality, a forward urge and backward
swing of life"). These abstract patterns of opposing emotional tendencies can
be abstracted from a number of myths, poems and emotional states, with no
need to introduce into the poem concrete elements that conflict with elements
existing in the literal meaning. Moreover, these abstract patterns can account
for the emotional qualities regularly associated with the poems, as well as im
pute some degree of unity upon them, rather than divide it up into discontinu
ous images and meanings. This conception of Bodkin's is elaborated in the
preceding essay, on the Death and Rebirth archetype in "The Rime of the
Ancient Mariner", to which the above quotation refers:
I have compared, also, myth and the metaphor of religious confession and
of psychological exposition, selecting material in accordance with similar
ity of imagery, especially of form or pattern. Particular words and images,
such as those of wind, of storm-cloud, of slime, of red colour, have been
examined for their emotional symbolism, but mainly with reference to
their capacity to enter into an emotional sequence. Within the image-
sequences examined the pattern appears of a movement, downward, or
inward the earth's centre, or a cessation of movement—a physical change
which, as we urge a metaphor closer to the impalpable forces of life and
soul, appears also a transition toward severed relation with the outer world,
and, it may be, toward disintegration and death. This element in the
pattern is balanced by a movement upward and outward—an expansion or
outburst of activity, a transition toward reintegration and life-renewal
(54).
One important distinction this passage makes is between archetypal con

tents and archetypal patterns ("emotional symbolism" on the one hand, and
"capacity to enter into an emotional sequence" on the other). It is not
explicitly stated, but it appears to be the rule observed by Bodkin, that images
are not checked for their emotional symbolism unless she is satisfied with their
"capacity to enter into an emotional sequence". Whatever the other merits and
defects of Fruman's approach, he seems to be unaware of this issue of
emotional sequence. On a more specific level, Bodkin gives here a detailed
account both of the sequence and of its constituent elements, in as abstract and
general terms as possible. Oddly enough, Bodkin does not attempt to show in
the following chapter how this pattern applies in its details to the description
of rushing water in "Kubla Khan". The "movement, downward, or inward the

earth's centre, or a cessation of movement" is clearly indicated in such
passages as
1. Where Alph, the sacred river, ran

or
2. Five miles meandering with a mazy motion

And sank in tumult to a lifeless ocean.
The "movement upward and outward—an expansion or outburst of

activity" is manifest in the passage
3. And from this chasm, with ceaseless turmoil seething,

Amid whose swift, half-intermitted burst
Or chaffy grain beneath the thresher's flail:
And 'mid these dancing rocks at once and ever
It flung up momently the dancing river.
The fact that this third passage occurs in the poem between the first and
second quotations not only demonstrates the opposing movements but also
creates an indication of rhythmic alterations, one of the main ingredients in the
Jungian conception of emotion underlying Bodkin's work.15
At the end of this comparison between Fruman's and Bodkin's depth-
psychological interpretations I wish to emphasise that although the former
embraces Freudian whereas the latter Jungian theory, the difference we have
discovered is not the difference between the two theories, but between the two
critics' individual approaches. Bodkin herself emphasised the kinship between
Jung's conception of the Death and Rebirth archetype, and "Freud's theory of
a pair of opposite tendencies, termed by him life and death instincts" (70, my
italics). Likewise she points out that "Freud and his school are also aware,
naturally, of the fact of growth and readjustment attitude" (72). Moreover, in
15
Lines such as "As if this earth in fast thick pants were breathing", in which
several critics have identified the imagery of birth, may reinforce the pattern
of "Death and Rebirth".
my 1975 paper I discussed at length Eva Metman's Jungian archetypal inter

pretation of Waiting for Godot (cf. Tsur, 1992a: 477-487) and pointed out
tendencies in her work that are more similar to Fruman's than to Bodkin's; that
is, she is interested in archetypal contents rather than in archetypal patterns.
Finally, I wish briefly to consider an issue related to the implied critic's
decision style. Earlier I suggested, and the foregoing discussion has supported
the validity of my suggestion, that a high degree of negative capability can be
inferred from Bodkin's critical writing. On the other hand, the passages dis
cussed from Fruman's work seem to be typical of the attitude of Quest for
Certitude. They offer to reduce the poem to a series of unrelated specific
meanings based on "false but obtrusive" rather than "subtle and minimal cues".
All this appears to be meant to dispel ignorance and uncertainty. However,
surprisingly enough, the paragraph that concludes this discussion appears to
give some evidence of negative capability:
The dangers of analyses such as the foregoing should be self-evident.

Without the dreamer's own associations to serve both as check and guide,
one can easily veer off into private fantasy and association. And where
association is concerned, the possibilities seem to be endless in any
imaginative person, if he abandons his mind to it. The presence of pitfalls,
however, need not forbid the attempt to transverse hazardous territory
(401).
This passage appears to undermine the alleged certainty achieved with the
typical tools of the Quest for Certitude; at the same time it appears to
undermine the author's certainty in his own self-criticism. There appears, then,
to be a contradiction between the cognitive strategies employed in the inter
pretation of symbols and those used in undermining the certainties achieved.
This contradiction may be handled in several ways. For one thing, it may be
evidence for what may well be the case, that cognitive style is not a unitary
phenomenon: it is not an all-or-nothing choice, and rigidity in one respect may
go hand in hand with flexibility in other respects. Another possibility is that
the critic is, indeed, a hard-boiled adherent of the Quest for Certitude, but his
attention has been drawn in one way or other to the desirability of appending
some such reservations to his interpretation. It is difficult, with our limited
information, to get supporting evidence for such a hypothesis. I wish,
however, to opt for a third possibility. At the beginning of Fruman's psycho
analytic discussion there is another passage, with similar reservations and self-
refutations, that gives additional evidence of his Negative Capability, and
suggests a possible solution to the problem:
What follows might well be relegated to the comparative obscurity of a

note, introduced perhaps by an ironic gaiety so as to disarm criticism, the
whole suggesting extreme tentativeness. But it is doubtful that such strate-
gies would work, or are altogether honorable. Certainly, what follows is

not urged as the meaning of "Kubla Khan". Poems lend themselves—with
often fatal facility—to every manner of symbolic meaning. "Kubla Khan"
is a poem embodying some ideas about the power of art. It is also about a
pleasure-dome, an Oriental emperor, a wailing woman, an overflowing
fountain—and much else (395).
The last three sentences contain one of the most effective arguments against
reducing the poem to a series of concrete external meanings, attributed to it
with the certainty of facts, so much favoured by the adherents of the Quest for
Certitude. The passage as a whole admits of a multiplicity of approaches. At
the same time, the critic temporarily commits himself to one possible reading,
without urging that this is the meaning of "Kubla Khan". The capability of
doing this is, precisely, what characterises, according to Else Frenkel-
Brunswick, the flexible personality; and it is this that the rigid personality is
incapable of achieving. In short, the critic displays a high degree of flexibility,
in temporarily committing himself to a reading; the reading itself is one rather
typical of a rigid personality.
Gating and Closure
One important activity involved in interpretation is the attempt to abstract in

formation implicit in the text. Gating and closure are metaphors, incidentally
derived from the same kind of visual image, referring to two alternative
strategies for obtaining the information sought from the text. One may increase
the amount of information obtained by "opening the gate as wide as possible",
or by delaying "closure" for as long as possible. This latter strategy requires "a
capability of being in uncertainties, mysteries, doubts, without any irritable
reaching after fact and reason"; people who are incapable of this have
sometimes recourse to the alternative strategy, and open the gate as wide as
possible.
Gating is one of three terms I have adopted from a paper by Gombrich (to
gether with abstractive relevance and sign limit). Gombrich borrowed from
Karl Bühler the principle of "abstractive relevance". Bühler meant by this the
irrelevance of certain features within sign systems. There are infinitely many
ways of writing the letter a. They will signify the same as long as certain
invariant relationships are observed (Gombrich, 1969: 155). From Jerome
Bruner Gombrich borrows the principle of gating. We may close a book
without knowing in what letter type it is printed, even if our look has scanned
thousands of letters. "Where we cannot derive more information or do not need
it, we shut the gates and go on to business" (ibid., 164).
I suppose it could be argued that what we call the esthetic response in front
of works of art involves a certain refusal to gate. The image is open, as it
were, and we are free to look for further and further echoes of the sense in
an indeterminate level of sound or form. But of course this refusal is only
a relative one. We all can distinguish between sanity and insanity in critic
ism—or at least we hope we can. There are the constraints of tradition, of
medium, of genre, and of culture that apply reins to the historian's and
critic's fancy (Gombrich, 1969: 164).
The third restricting principle is called the principle of sign limit, which
Gombrich borrowed from an unpublished dissertation of Bühler's student,
Julius Klanfer. Sign limit is a crucial but rather problematic principle. All of us
have rather clear intuitions as for which interpretations are beyond the proper
sign limit; but we have got no unambiguous principles to determine where the
sign limit is. 16 We have got no rules for determining the sign limit in literature;
there is, however, a fair consensus that in the literary use of language it is
further away than in most other uses. Nevertheless, does "anything go" in the
literary decoding of signs? Many people doubt it. Gombrich implies the
solution that sane intuition may determine the sign limit. The present chapter
has foregrounded the difficulty facing such an assumption. It would appear
that Maud Bodkin's intuition eliminates what Fruman's intuition admits; and
Schneider's intuition seems to be even more "cautious". Alternatively, one
might suggest that one of the tasks of cognitive poetics is to give a description
of the cognitive structures that give rise to such intuitions.
The present chapter attempts to make a modest contribution to the
development of more precise tools for determining the "sign limit" or
"abstractive relevance", exploring certain personality variables that affect the
intuitions concerning them. It has been assumed that critical intuition is not
always determined by the legitimate needs of literature. One of the factors that
may affect the critic's intuition is the emotional needs of one or another
personality styles. The present work has not been concerned with the emo
tional needs or personality styles of critics as empirical human beings, but with
the decision style of the implied critic, consisting of the sum of his choices as
they are reflected in the piece of criticism; thus, they are in the public domain,
open to public inspection. One fundamental assumption of the present work is
that in literary discourse the relationships between linguistic and nonlinguistic
16 Gombrich quotes a charming example of an extreme violation of the sign

limit, from the first page of Dickens' Great Expectations:
As I never saw my father or my mother, and never saw any likeness of
either of them (for their days were long before the days of
photographs), my first fancies regarding what they were like, were
unreasonably derived from their tombstones. The shape of the letters
on my father's gave me an odd idea that he was a square, stout, dark
man, with curly black hair [...] (Gombrich, 1969: 164).
signs are more complex than in nonliterary discourse. Consequently, in literary

discourse longer endurance of uncertainty is required than in nonliterary
discourse, before the reader can come to a conclusion following the decoding of
the signs that constitute the text. The implied critic's choices are affected, to a
considerable degree, by his capacity for enduring uncertainty. In the course of
the present chapter we have attempted to isolate a variety of tactics for
achieving stability and certitude.
The cognitive mechanism underlying critics' conflicting intuitions is, per
haps, a gating strategy for obtaining information from poetic signs: one may
distinguish between duration and width of gating; that is, the question is
whether closure is delayed so as to enable the perception of as complex
relationships as possible, among as many signs as possible, or whether closure
is rapid. In the latter case, the critic may compensate with width of gating.
Since ignorance too induces uncertainty, the critic is "compelled" to compen
sate for rapid closure (that is, for the shortening of the duration of uncertainty)
by displacing the sign limit and violating the principle of "abstractive
relevance". In order to dispel anxiety arising from ignorance caused by rapid
closure, some critics apply their mind to those aspects of the signs that have
no "abstractive relevance" (frequently referred to as "noisy attributes"). As a
result, they have recourse to incomplete abstractions in interpretation, or look
for as concrete and specific "symbolic meanings" as possible.
An instance of a critical approach that presupposes as delayed closure as
possible is Bodkin's conception, as discussed above, of archetypal pattern (as
contrasted to archetypal contents), in the present case "a pattern of rising and
sinking vitality, a forward urge and backward swing of life, reflected in an
imagery deployed in time". Closure is further delayed owing to the consider
able number of aspects and elements in which this "imagery deployed in time"
is contrasted: "the pattern appears of a movement, downward, or inward the
earth's centre, or a cessation of movement—a physical change which, as we
urge a metaphor closer to the impalpable forces of life and soul, appears also as
a transition toward severed relation with the outer world, and, it may be,
toward disintegration and death. This element in the pattern is balanced by a
movement upward and outward—an expansion or outburst of activity, a trans
ition toward reintegration and life-renewal". It should be repeatedly empha
sised that "delayed closure" as such is nowhere explicitly mentioned in
Bodkin's discussion; but it seems to be implicitly recognised by critics
sufficiently sensitive to it, though, again, they may hesitate when trying to put
their fingers on what is it they are recognising. Let me quote a rather long
passage from a critic whom we have already credited with a considerable degree
of Negative Capability:
If one proceeds upon the belief that a man cannot open his mouth without
being symbolic, perhaps one can do no better than accept for Kubla Khan
some such interpretation as Professor Knight's or, a better choice, I think,

Miss Bodkin's [...] (Schneider, 1975: 260).
If we pause here for a moment, we may notice that Schneider objects to

"symbolic" criticism, including Bodkin's but, at the same time, she prefers her
work to other symbolic critics'. In the rest of this paragraph, Schneider
attempts to justify her distrust of symbolic criticism, as well as her sympathy
with Bodkin's work. As an abstract argument this passage is not very well
organised, but precisely this may give us a dim insight into the working of
Schneider's abstract thinking, with its divergent trends and tolerance of
discordant attitudes. Her next sentence can be understood as a discrediting of a
certain critical position: "To the critic who maintains that a poem has a
meaning of which the poet was unconscious there is no answer". On second
thought, however, and from a different point of view, she cannot dismiss this
kind of thinking either: "And indeed I do not doubt that this sometimes
occurs". The next sentence elaborates this idea, followed by a sentence that
makes reservations about it:
There is no question, I think, that unconscious and subconscious forces

lurk beneath the surface of our thought and feeling and that the appeal of
a work of art may derive partly from secret correspondence with these
forces. But that it always does so in the sense usually supposed is far from
being established; and that the conscious symbol-seeking of critics or
psychoanalysts often succeeds in salting the tail of this invisible bird is a
matter of very considerable doubt.
Clearly the considerable complexity of, especially, the second of these two
sentences, serves a tendency to avoid an all-inclusive or an all-exclusive state
ment. Notice how Schneider again makes an exception with Bodkin, as a
parenthesis so to speak, further elaborated in a parentheses-within-parentheses
structure:
If Kubla Khan naturally conveys to a reader one or another of the

meanings of sexual opposition or immortality or the life of man—or of
King Oedipus and incest—perhaps they are there. But to me—except for
Miss Bodkin's Paradise-Hell contrast, which very naturally come to mind
though Coleridge purposely diminished it by altering "Mount Amara", a
really traditional earthly Paradise, to "Mount Abora"—except for that
meaning, which I think of in this poem as rather associative than symbol
ic, such elaborate interpretations as have been made of Kubla Khan seem
more ingenious than compelling. They do not coalesce naturally with what
Coleridge actually wrote; on the contrary, the thought of these meanings
drives the poem itself out of one's mind (Schneider, 1975: 260-261).
It is not very clear what Schneider may have meant by "rather associative
than symbolic". One possible construal of this phrase may concern
Gombrich's supposition "that what we call the esthetic response in front of
works of art involves a certain refusal to gate". What we are up against is not a
"Symbolic meaning", but merely a small displacement of the sign limit, further
away from the centre, so as to admit some further information concerning
Mount Amara. One remarkable (and praiseworthy) feature of this discussion is
that even within her reservations Schneider makes a significant distinction
between the various degrees of the attenuation of this information as relevant
to the poem ("though Coleridge purposely diminished it"). It may well be that
all interpretation is symbolic (depending on how we define "symbolic"). So,
what seems to be wrong with some symbolic criticisms is not that they are
symbolic, but rather that they result from rapid closure and a wider than
acceptable opening of the "gate". It should be noticed, however, that "gating"
is not an all-or-nothing process. All aesthetic response seems to involve "a
certain refusal to gate". The question is whether the refusal to gate comes to
compensate for a certain refusal to delay closure.
Now the impatient reader may object: "I don't care about the critic's
motives or cognitive mechanisms; all I care about is whether his argument is
valid or not". I can only sympathise with such an objection. However, the
afore-mentioned cognitive mechanisms and the refusal to apply them may be
significantly correlated with certain conspicuous inadequacies in critical con
ception and practice. Consider the end of Schneider's passage quoted above:
"They do not coalesce naturally with what Coleridge actually wrote; on the
contrary, the thought of these meanings drives the poem itself out of one's
mind". In order for a meaning to coalesce naturally with what the poet actually
wrote a critic or reader must "bear in mind simultaneously various aspects" (in
Frenkel-Brunswick's phrase, v. infra). Above we quoted Harvey on concrete
functioning, that has "A simpler cognitive structure, comprised of fewer differ
entiations and more incomplete integrations within more central and ego-
involving domains but not within domains of low involvement". The reading and
criticism of poetry may require a high degree of ego-involvement. To discuss
poetic meanings that do coalesce naturally with what the poet wrote one must
delay closure with respect to the whole poem (or, at least, a whole poetic
passage) until all the issues concerning the symbolic meanings are weighted and
weighed. This, in turn, requires "a more complex cognitive structure,
comprised of more differentiations and more complete integrations". When a
critic encounters some evasive poetic quality and turns to establishing symbo
lic equivalences that need not coalesce naturally with what the poet actually
wrote, he is exempt from all this. He can pursue his source-hunting, parallel-
hunting or myth-mongering in the ways we encountered at the beginning of the
present chapter; all he must remember is that he is looking for items related to
domes (or caves, or the words bright and green, or the name Alph, or Abora).
He can safely achieve rapid closure and seek his fortune in ever-widening
circles. True, when he returns with his golden fleece, it does not naturally
coalesce with what there is explicitly in the poem; but he does not mind; in
fact, most probably he cannot mind. He is not equipped with the capability of
delayed closure, the touch-stone to test the purity of the gold of which the
fleece is made.
Psychological Models
When a critic has amassed many facts or alleged facts in an attempt to account
for some evasive quality in a poem, he sometimes faces the need to fuse those
facts into one unity. One favourite way to satisfy this need is to use words to
tune the reader's mind so that he performs this fusion for himself; or, at least,
so that he accepts these facts as relevant to the poetic quality in question and
as appropriately fused. This is one reason why hard-boiled factualism is so
frequently associated with what appears to be its very opposite: critical
impressionism.17 A more respectable way of facing the need to fuse the facts
is to offer a psychological process or model to indicate how those facts are
fused in the poet's or the reader's mind. One such model is offered by Lowes
in the paragraph that introduces his discussion of "Kubla Khan":
Suppose a subliminal reservoir thronged, as Coleridge's was thronged, with

images which had flashed on the inner eye from the pages of innumerable
books. Suppose these images to be fitted, as it were, with links which ren
der possible indefinite combination. Suppose some powerful suggestion in
the field of consciousness strikes down into this mass of images thus capa
ble of all manner of conjunctions. And suppose that this time, when in re
sponse to the summons the sleeping images flock up, with their potential
associations, from the deeps—suppose that this time all conscious imagi
native control is for some reason in abeyance. What, if all this were so,
would happen? That hypothetical question fairly covers, I think, the case
of "Kubla Khan" (Lowes, 1927: 343).
When we consider such a psychological explanation, a crucial distinction

must be made. When we use a seventeenth century psychological model to
account for, say, the behaviour of a Shakespearean character, we make a
17
There seems to be another reason, one that has a direct bearing on our present
inquiry. There are seemingly incompatible ways to achieve certainty in criti
cism, in the service of the Quest for Certitude. One way is by sticking to
irrefutable "objective" facts; another by clinging to one's own subjective feel
ings, equally irrefutable. Thus, the Quest for Certitude may admit factualism
and impressionism at one and the same time, in spite of their apparent
incompatibility.
judgment as to the consistency of the psychological model with the character's

behaviour. But when we use a psychological model to account for the poetic
quality of the poem with reference to the author's or the reader's
psychological processes, we also expect it to be adequate with reference to the
present state of the art. It should be remembered that Lowes's book is sub
titled A Study in the Ways of Imagination. And Schneider, for instance, quotes
the first paragraph in full, and criticises it for being inadequate in precisely this
respect. She takes up the "hypothetical question", and answers:
Psychoanalytical thought has one answer to this question: with conscious

control in abeyance, the images fall under the control of the unconscious
mind and become symbols of desires, fears, conflicts, expressing the will or
wish of the dreamer no less purposively than if they were conscious, but at
the deeper levels of unadmitted desire or conflict. Lowes's answer was not
this, for the dreamer in Kubla Khan, he thought, is "merely the detached
and unsolicitous spectator". And the only force determining the form and
sequence of the imagery is the "subtle potency of the associative links".
There is no plan, no "deliberate manipulation". It is in effect a poet's
exercise of free association without any implication that free is more than
chance association or than Hartley's juxtaposition of atoms; the
"bewildering hooks and eyes" of the unconscious memory alone were the
"irresponsible artificers of the dream". And so in the poem "the linked
and interweaving images irresponsibly and gloriously stream, like the
pulsing, fluctuating banners of the North. And their pageant is as aimless
as magnificent" (Schneider, 1975: 240-241).
Scientific inadequacy, however, is not the only defect Schneider finds with
this passage:
It is evident from Lowes's language that he was too dazzled to see quite
what was before him, his appreciation having outrun perception. [...] Co
leridge's preface and the music of Kubla Khan have so particularly encour
aged the impressionistic approach to that piece that we are apt to read it
with but half-conscious attention as a kind of glorified nursery rhyme even
while we call it the quintessence of poetry (ibid., 241).
Thus, Lowes's passage above turns out on closer inspection to be a piece of

impressionistic criticism under a psychologistic disguise. Now, is there a
logical connection between the two accusations, of relying on an insufficient
psychological theory, and of impressionism? In an attempt to answer this
question, let us consider another psychological model prefixed to a highly
erudite discussion of Coleridge's poem (Beer's). The two discussions are
opposed in almost all respects. In the first place, Beer does not credit
Coleridge's images with "irresponsibility" or "aimlessness", the lack of any
"deliberate manipulation". On the contrary, "however it was composed", he

says,
Kubla Khan the poem is not a meaningless reverie, but a poem so packed
with meaning as to render detailed elucidation extremely difficult. It will be
suggested that many of the images in the poem can be related to several
patterns of meaning which run parallel and are held together not by the
"story", but by a separate argument which runs through the poem, at times
explicitly stated, at times implicit in the imagery (Beer, 1959; 202).
In the second place, Beer's discussion is not carried away by his

enthusiastic language. His writing is plain scientific prose, with no extravagan
cies. In the third place, he does not embrace a vague, outdated psychological
theory that has not been effected by the Freudian revolution. He embraces the
idea "that very intricate mental processes can take place in states of imperfect
consciousness" and quotes at length from Dalbiez's book Psychoanalytic
Method and the Doctrine of Freud:
[T]he dreamer passes by association from one image to another; he uses

relations, but he does not isolate them; he does not understand them. On
the other hand the scientist, after the action of the unconscious relation
has caused the rise of a new idea, perceives the relation as a thing apart; he
understands. Discovery, in fact, is made in two stages. The first is compar
able to the evocation with unconsciousness of the relation in the case of
the animal or the dreamer, but with this difference that concepts of a
higher order of abstraction are involved. The second stage is strictly
rational, the intellection of relation considered simply in itself as a thing
apart. Sometimes discovery consists in the rising-up process and the
intellectual appreciation of a relation between two ideas already possessed;
the unconscious action of the relation is then confined to the appropriate
evocation. The accounts given by Poincaré of the origin of several of his
discoveries seem to show that their origins can be traced to a process of
this kind. The discovery of several structural schemata by Kékulé seems to
have been brought about by a process whereby relational unconsciousness
has developed into the creation of the new schema. He was dozing on the
top of an omnibus, and atoms were dancing before his eyes, first two and
two and then in groups of three or four. He spent most of the night
working out on paper the hypnagogic or oneiric images which had thus
appeared to him, and in the morning he had the result (Beer, 1959: 202-
203).
And so forth. This description and Beer's ensuing comments contain some
very sound information and speculation concerning the "kinship between
creative intellectual processes in different fields". It sounds serious, scientific
and respectable. And I do believe the unconscious processes suggested underlie
creative thinking in scientific discovery as well as in the writing and reading of

poetry. But what does it say, with respect to the fusion of images in "Kubla
Khan" into a poetic whole, beyond what is said by Lowes in "the sleeping
images flock up, with their potential associations, from the deeps"? Very little,
I think. In fact, in view of what follows, it does have one more implication:
"Since eminent atom physicists can literally 'dream up' their epoch-making
theories, whatever meanings I attribute to the poem cannot be refuted on the
grounds that it is too sophisticated or intellectual for a dream". But there is no
indication of how the structural properties of the psychological process
suggested may shape and constrain the fusion of images in the poem. Having
said this, there is a fundamental similarity between Lowes's and Beer's handl
ing of their respective psychological theories: each presents his psychological
model at the beginning , never to return to it in the course of his long discussion
of "Kubla Khan". Beer, like Lowes, presents a host of information supposed
to be thronged in Coleridge's subliminal reservoir, that in certain mysterious
circumstances may flock up with its potential associations from the deeps, but
he does not show what we can learn about its fusion from the processes dis
cussed by Dalbiez.
The upshot of this discussion seems to be this. Theoretical models or frame
works cannot be directly applied to works of literature; they must be applied
to texts via critical terms that are properly articulated and have a considerable
descriptive contents. Critical terms, on the other hand, with their descriptive
contents have little significance unless understood in relation to some theoret
ical framework or model. A critical statement can be true and remain trivial and
devoid of interest, unless some theoretical framework or model gives to it sig
nificance. Let me illustrate this briefly by an oversimplified example. A critical
statement like This poem contains twenty three lines may be perfectly true but
trivial from the aesthetic point of view. On the other hand, a statement like
This poem contains fourteen lines may be regarded rather meaningful from the
poetic point of view, when viewed within the theoretical model of the sonnet
form. In this simple example, fourteen lines and twenty three lines are terms
with precise descriptive contents; sonnet form is a simple "theoretical model".
Significant distinctions in a poem can be made only with the help of terms that
have precise descriptive contents, but they become significant only when they
can be related to some widely accepted theoretical model. Theoretical models,
in turn, can be fruitfully applied to a poem only via a term with a precise
descriptive contents. We cannot know whether a poem is a sonnet without
being aware of the number of lines and their grouping by the rhyme-scheme. A
good critical term is, then, one that has a well-articulated, precise descriptive
contents on the one hand, and can be related to some general model or
commonplace (in the rhetorical sense of the term) on the other. A good model,
in turn, is one that combines with terms on more specific levels and imputes
precise descriptive meanings to them, improves their articulation, or renders
the distinctions made with them significant18 (see also the Afterword, below).
Now, what seems to be wrong with both Lowes's and Beer's psychological
models, from the critical point of view, is that these critics do not apply terms
with articulate descriptive contents with which significant descriptions or
distinctions can be made in "Kubla Khan".
Before proceeding with this issue, a rather long comment seems appropriate
here, concerning Beer's psychological model adopted from Dalbiez. The
processes mentioned are certainly highly relevant to both the writing and the
reading of poetry. However, from the passage quoted from Dalbiez by Beer
one can easily get the impression that there is some mysterious, or at least
fuzzy relation between the unconscious process that goes on during sleep and
the final intellectual achievement to which it leads. There is no attempt to
describe the process in reasonably articulated descriptive terms. As a matter of
fact, the gestaltists have investigated this issue since the early nineteen-thirties,
and have come up with some quite interesting results. There are quite a few
cases similar to Kékulé's on record. Köhler (1972: 163) refers to the three B 's,
"the Bus, the Bath, and the Bed", where some of the greatest scientific dis
coveries have been made (remember Archimedes!). 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 organisa
tion of the situation, and with it the solution, tend to occur at moments of
extreme mental passivity (ibid., 160).
The solution occurs suddenly at a time when, in Lowes's phrase, "all

conscious imaginative control is for some reason in abeyance". That is
precisely the time when a restructuring of the situation may take place.
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 certain 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 (Köhler,
1972: 152-153).
What we usually call insight is the unique conscious quality of the sudden
emergence of the restructuring of mental processes. In the terms of cognitive
I have discussed these and related matters at greater length elsewhere (Tsur,
1983a: 28-36; 1983b; 1992a: 501-535).
poetics, it is the perceived quality of this sudden emergence. The last sentence
of the above quotation exactly describes what happens in a really insightful
reading of poetry. Now, from this presentation of the issue one could easily
derive very useful terms with quite well-articulated descriptive contents. One
only has to describe the given material before and after the new relationships
have been perceived, as well as the change that has taken place as a result of
amplification or restructuring. Did not Cleanth Brooks define his key-term
irony as "a general term for the kind of qualification which the various elements
in a context receive from the context"? (quoted by Wellek, 1963: 329). Now,
had Beer used these terms in dealing with "Kubla Khan", it could easily have
turned out devastating for his critical activity. From the foregoing we should
expect to find that all the material related to "Kubla Khan" discussed by him is
restructured so as to make it possible to discover new relations among old
items of material. Now, of the enormous amount of alleged source-material
amassed by Beer only a small percentage occurs in the poem. I seriously doubt
whether the rest can be regarded in anyway relevant. But even should it be so
regarded, it has not been restructured, and has not entered into new
relationships. That is why a depth-psychological model of unconscious pro
cesses is so useful for this kind of criticism. Anything can be claimed to be
relevant to the poem, and present in the poet's unconscious mind. Köhler puts
some structural constraints on what can be considered relevant in this sense.
I believe the best way to bring out the deficiency of Lowes's and Beer's
handling of their respective psychological models is to compare them to
another work (published only three years after Lowes's book). Snyder (1930)
observes a contrast between two groups of poems, one of "some well-known
spellweaving or hypnotic poems", and another of "some well-known
intellectualistic poems" (p. 8). Regrettably for the present comparison, though
he includes Kubla Khan and "much of the Ancient Mariner" in the list of
poems that constitute the first group, he makes only fleeting allusions to
Coleridge's poems, perhaps because they are so obvious examples for some of
his generalisations. But from the close textual examination to which he submits
Gray's elegy, for example, one may get a more than fair idea of the subtle
textual discriminations he makes with the descriptive terms derived from the
psychological model. I shall not follow him closely, only give a very brief
outline of his method. Viewed from the perspective of the present theory of
cognitive poetics, I should distinguish three dimensions in his theory and
practice. First, there is an observation that certain poems have a pervasive
"spellweaving" or "hypnotic" quality; this is what I call their perceived
quality. Second, Snyder offers a psychological model (the hypnotic process) to
account for the presence of this perceived quality in the text. Assuming that
Snyder is not merely toying around with metaphors but that the term hypnotic
is an apt term to refer to a genuine quality perceived in these poems, one may
also assume that it is possible to discover certain conspicuously similar ele-
ments in the hypnotic process and the structure of these poems. Third, with
the help of these elements Snyder derives a set of critical terms with clearly
articulated, precise descriptive contents. One should emphasise that this
precision does not derive only from the psychological model, but also from the
critic's keenness and readiness to face refutation. Thus, for instance, Snyder
found that most of these poems have excellent versifications or lulling
rhythms.
I repeatedly tried the theory that these poems gain their spellweaving
power because of the perfection of their versification [...]. It developed
that every poem of this sort is characterised by excellent versification,
and so far the explanation was satisfactory. But the flaw in the theory
appears as soon as it is applied negatively to the contrasting group
(Snyder, 1930: 14; my italics).
Snyder does not attempt to attune the reader's mind to a certain impression.
On the contrary, it is he himself who is looking for counterexamples to his own
generalisations. As a result, he is compelled to attend to subtle and minimal
cues, and is not misled by false but obtrusive ones. Thus, the counterexamples
lead him to the conclusion that
not beautiful rhythm alone, but a certain kind of rhythm combined with
other stimuli to put the listener into a light state of trance—a waking
trance in which aesthetic enjoyment is heightened until it may even reach
ecstasy (16-17).
The "other stimuli" too are referred to by terms that have clear-cut descript
ive contents, and some of these, too, may occur both in hypnotic and
intellectualist poems. Thus, only the co-occurrence of a considerable subset of
the "hypnotic" techniques can induce "spellweaving" poetic qualities. And the
presence or absence of some of these elements enable us to make further
distinctions within hypnotic poetry. "The particular form of trance with which
we are concerned in this study of poetry is one in which the emotional
sensitiveness of the subject grows more and more intense" (31). But "hyp
notic" poetry may exploit this emotional sensitiveness in two different ways:
Some hypnotic poems stop here: the listener is lulled by patterns of sound,
his attention is fixed without arousing of his mental faculty, and he falls
into whatever mood the poet "suggests". It is interesting to see how many
poets are thus content to stop without taking full advantage of the grip
they get on the listener's emotions. Such skillful artists as Poe, Swinburne,
the youthful Tennyson, and countless others persistently fail, or refuse to
galvanise the sensitive reader to action, determination, or even thought
(47).
That is, they ask the reader only "to let their mind be a thoroughfare for all
thought". From what we have seen of the interpretations of "Kubla Khan",
many critics are more than willing to supply what the poet "failed" or
"refused" to put into the poem, owing to their intolerance of situations in
which they are expected to respond to a unique, unclassifiable sensation while
their emotional sensitiveness grows more and more intense. In this way they
"level" the differences between these and other kinds of hypnotic poems,
while Snyder in contrast makes efforts to bring out the difference between
these hypnotic poems and those poems which "carry the parallel to
hypnotism still further by 'suggesting' an impulse to action, making a parallel
to the specific post-hypnotic suggestions" (47-48). These "hypnotic sugges
tions" draw upon the increased emotional sensitivity induced by hypnotic
techniques ("When there is not one iota of proof or argument stated or
suggested in Crossing the Bar, why do certain listeners get from it the over
whelming conviction of immortality?" [13; Snyder's italics]).
Now it may be objected that such "suggestions" are also commonly found
in non-hypnotic poems, and the objection, so stated, is valid. But there is
discoverable a difference in the nature of the suggestion and often in the
position of the suggestion [...]. Specifically, in a hypnotic poem the key
sentence "suggesting" an idea comes near the end, or at least only after
there has been a long preliminary soothing of the listener's senses by
monotonous rhythmic "passes". So in hypnosis. Also this key sentence
"suggesting" an idea carries conviction without argumentative support, or
with only the simplest of obvious arguments to support it. In the non-
hypnotic poem these conditions do not obtain (48).19
19 Snyder's uncompromising intellectual keenness and honesty are manifest in

the following passage where he presents a counterexample to this generalisa
tion too:
Pope, on the other hand, in his most brilliant failure, the Essay on
Man, argues throughout Epistle I in support of a thesis [...], and finally
states the thesis he has tried to establish: "Whatever is, is right".
Although the final position of this key sentence is in accord with the
hypnotic tendency, Pope's poem in most respects is of the other
type. In so far as it fails, the difficulty is probably due [...] to the
psychological discord between the soothing monotony of the
"rocking-horse couplet" and the intense mental activity demanded of
the listener (50).
Snyder treats Pope's poem as a hypnotic poem manqué. As I shall suggest in
the next chapter, further distinctions between, and combinations of,
categories are required. Regular metre is, in fact, "triple-edged", and Snyder
does not seem to be aware of the potential intellectual effect of good gestalts
as in regular metre and symmetrical couplets. But this kind of solution, too, is
in Snyder's spirit of "a certain kind of rhythm combined with other stimuli".
Finally, I propose to summarise Snyder's terms whose justification and

descriptive contents derive from the hypnotic process (as expounded on pages
39-51).
* Hypnotic poems in general give us heavy stresses falling regularly [...] and
so ornamented that the rhythmically inclined listener has his attention drawn to
the sound rather than to the sense.
* Another point in which practically all of these poems show an interesting
affinity to hypnotism is their freedom from any abrupt changes which would
be likely to break the spell, and especially freedom from ideas that compel
mental alertness.
* A certain vagueness of imagery.
* Paradoxical though it sounds, we may yet have to accept the view that in
the early stages of a hypnotic poem a foreign word, obscure phrase, or any
slight difficulty that causes fatigue from strain on the part of the listener may
actually promote the ultimate aesthetic effect at which the artist aims. [The
first line of our poem, "In Xanadu did Kubla Khan" more than satisfies this
condition—R.T.].
* The use of a refrain, or of frequent repetition.
* The use of suggestion on an entranced person.
I have dwelt at some length on Snyder's study of hypnotic poems, of which
"Kubla Khan" is one, in order to show the mutual relationship between
psychological model and descriptive terms, and their usefulness in criticism.
The subtlety, complexity and abstractness with which Snyder propounds and
applies this model and the terms derived from it are ample evidence of his
Negative Capability.20 But the main reason for introducing them here was to
show the methodological deficiency of Lowes's and Beer's psychological
models. Unlike Lowes and Beer, Snyder draws a set of principles from his
psychological model, with which meaningful distinctions can be made between
hypnotic and other poems, within the group of hypnotic poems among
subgroups, and within individual poems. We can now answer a question asked
above. Both Lowes and Beer make an impressionistic use of their respective
psychological models. By this we mean that irrespective of whether they are
adequate or inadequate in view of present day psychology, this judgment
depends, rather, on whether the psychological models are used merely "to tune
the reader's mind", or as a "rhetorical commonplace" and a source of clearly
articulated descriptive terms. Both critics seem to resort to this kind of
20
In spite of his subtlety and exceptional sensitivity to poetic effects, as well as
his great theoretical awareness, it is conspicuous that Snyder's book preceded
Wimsatt and Beardsley's seminal paper on "The Affective Fallacy" by nine
teen years. Indeed, the affective fallacy is the most obvious pitfall throughout
this remarkable little book. Thus, for instance, Snyder decides that Poe's
Ulalume is a "semihypnotic poem", because it "is a rather hypnotic poem in
its effect on some people, not on others" (71).
impressionism disguised as psychology in order to compensate for some kind

of dissatisfaction with the poetic significance of the "facts" accumulated by
them to account for the intuitive greatness of the poem.
Critical Theory and Negative Capability
In this last section of this chapter I am going to discuss "Kubla Khan" criticism
on three levels (as such, it will be representative of my conception underlying
the foregoing discussion). First, I shall briefly consider the problem of
incompatible or conflicting interpretations in relation to the poem. Second, I
shall consider at some length the theoretical and methodological issues involved
in handling such incompatibilities and offer a workable conception in view of
the analytical philosophy of criticism of the nineteen-'fifties and 'sixties.
Third, I shall consider the demands such a conception makes on the critic's
Negative Capability, and their implications for the implied critic's decision
style. During this discussion an all-important distinction must be borne in
mind. When a critic refrains from resorting to a methodological solution that
requires a more than usually high degree of Negative Capability, it may mean
one of two things at least: that the critic cannot cope with the emotional
demands of that kind of solution, or that he simply is not acquainted with the
theories of interpretation put forward by e.g. Morris Weitz or Joseph
Margolis. Indeed, I believe that most or all of the critics discussed in my paper
are unacquainted with their work. Thus, in some instances at least we may
assume that the critic was not aware of the complexity of the issue rather than
that he avoided its complexity on purpose in order to defend himself against its
threat. So, we must consider the negative evidence (i.e., that a critic did not
assume a certain critical position) against the background of his general
cognitive strategies and decision style.
When one is engaged in an overview of a considerable number of more-or-
less legitimate interpretations of one literary work of art, the question
inevitably arises how a single poem can mean all those things, or even some
part of them. Thus, for instance, Schneider devotes the first chapter of her
book to a consideration of the welter of interpretations to which Coleridge's
major poems have been submitted. I shall have to confine myself to a relatively
short quotation:
Mr. Warren, Mr. Burke, Mr. Knight, Miss Bodkin, and the others cannot
all be right (which does not, certainly, prove any one of them wrong).
Their various symbolic interpretations of Coleridge's poems not only are
not easily reconciled with one another on the basis of "different levels",
but also impute quite different moods or emotional tones to the same
poem. If Christabel is felt as the Inferno, it can scarcely also be felt as the
moment of balance between good and evil. If Kubla Khan is the Paradiso
of Dante, it is not easy to feel it also as exhibiting the conflict of heaven

and hell or Coleridge's somewhat less than heavenly domestic life. To the
confusion of these is added the voice of those other critics who maintain
that [...] Kubla Khan is wholly without meaning of any kind. Though
variety among critics is no doubt all to the good, one cannot help
wondering a trifle about the present state of criticism when we find as little
common ground as this among writers all very eminent, all brilliant and
persuasive in argument, and all engaged in describing the central effect of
the same poems (Schneider, 1975: 16).
After having agreed with Schneider so many times, this is one point on
which I disagree. As I have indicated more than once, I do agree completely
with Schneider's objection to the extravagancies of symbolic interpretation; but
I also believe that a variety of interpretations as enumerated by Schneider, and
much more, can all be right, though some clearly are wrong. The root of our
disagreement is certainly in Schneider's phrase "all engaged in describing the
central effect of the same poems", which is clearly an oversimplified position.
All these critics are, indeed, "engaged in describing the central effect of the
same poems"; but after an interpretation. As a matter of fact, these critics are
only marginally engaged in describing the central effect of these poems; what
they are doing primarily is trying to elaborate an interpretation.
The source of the problem is that although philosophers of the analytic
tradition have done exceedingly important work in the philosophy of criticism,
only a handful of critics seem to be aware of it. These philosophers have
formulated a long series of questions that ought to be asked, and have indicated
the directions in which the answers can be found. One of the most pernicious
sources of confusion is that many critics believe that all critical statements are
of one kind: factual statements, that are true or false. Philosophers like
Beardsley, Margolis or Weitz distinguish three kinds of critical statements.
Beardsley (1958: 11) speaks of description, interpretation and evaluation.
Weitz (1972: 228-284) speaks of description, explanation and evaluation (I
shall not go here into an explication of the difference between Beardsley's
interpretation and Weitz's explanation). In disagreement with Beardsley,
Margolis (1962: 116) in his discussion of the logic of interpretation suggests
that "the characteristic feature of critical interpretation that is philosophically
most interesting is its tolerance of alternative and seemingly contrary hypo
theses". The characteristic predicate in interpretations (as in scientific hypo
theses) is not true but plausible. Margolis (1962: 117) proposes a set of dis
tinctions between the "true" and the "plausible", one of his distinctions being:
"Where the statements 'P is true' and 'Q is true' are contraries, the statements
'P is plausible' and 'Q is plausible' are not contraries". Likewise, Weitz (1972:
258) suggests: "These explanations can only be adjudicated in vague terms as
for their adequacy but not in precise terms as to their truth or falsity".
In his 1972 book (first published in 1964), Morris Weitz undertook a

comprehensive critical study of the enormous number of Hamlet interpret
ations. What I have been doing here on a very modest scale with "Kubla Khan"
criticism does not pretend to compete with him, and has in fact very different
aims. Weitz's study brings out on a grand scale what made Schneider "wonder
a trifle about the present state of criticism when we find as little common
ground as this among writers". It manifests and argues that when a single work
of literature is submitted to a great number of interpretations, they will
necessarily include incompatible interpretations, though none of them need be
wrong. But it is mainly his theoretical considerations and conclusions upon
which we can liberally draw for our subject.
Let us consider, for instance, Schneider's misgivings: ''If Kubla Khan is the
Paradiso of Dante, it is not easy to feel it also as exhibiting the conflict of
heaven and hell". How can two such incompatible interpretations refer to the
same data in the poem? The point is that the various critics are not dealing
with the same data in "Kubla Khan", just as in Hamlet they are not,
where the data themselves are attributed ones, hypothesised by the critic.
That Hamlet delays is given in the play, but that his delay is central or
even that Hamlet is central is not; that Hamlet suffers is given in the play,
but that his suffering is his central trait and that this trait is most
important in the play are not given. What is central, primary, most
important, or what is the theme of Hamlet is not a datum but a hypo
thesis, which the critic defends by further hypotheses: from a specific hy
pothesis about Hamlet to a general one about drama (Weitz, 1972: 256).
The data accounted for in the various readings are determined by the relative
weight the critic bestows upon them. Consider this issue in "Kubla Khan".
That the caverns are there and that they are opposed to the "sunny pleasure-
dome" is given in the poem; but that they are also opposed to the mountain, or
that this opposition reflects the conflict of heaven and hell, is hypothesised by
the critic. What is the theme of "Kubla Khan" is not a datum but a hypothesis,
which the critic defends by further hypotheses: from a specific hypothesis
about "Kubla Khan" to a general one about some more general patterns, such
as the archetype of Paradise and Hades, or "Coleridge's Divine Comedy", or
romantic nature poetry, or the nature and structure of ecstatic poetry. Now, to
what degree are the caverns considered "central, primary, most important",
depends inter alia upon whether "Kubla Khan is the Paradiso of Dante", or
exhibits "the conflict of heaven and hell". Under the latter interpretation, all the
aspects of caverns we have discussed in relation to Bodkin's analysis receive
considerable weight, as part of "a pattern of rising and sinking vitality". Both
readings agree upon the centrality of Paradise in the overall pattern of the
poem, but in the former the poem is part of a wider pattern, together with two
more poems, constituting "Coleridge's Divine Comedy". Under this pattern,
some of the aspects of caverns suggested by Bodkin are necessarily toned

down, and the opposition "a sunny pleasure-dome and caves of ice" suggests
an overwhelming sense of wonder in the face of "a miracle of rare device",
enhancing the directly felt pleasure in the description of Paradise. This capab
ility of "switching" from one set of aspects to another of the opposition corre
sponds to the ability suggested by Wittgenstein (1976: 214e) to "understand
the request to pronounce the word 'till' and to mean it as a verb" or an adverb.
We could go on quoting endlessly from Weitz's illuminating book. But I
shall quote only one more passage crucial for our business:
Like the director of a production of Hamlet, the critic rehearses with his
readers the various possible ways of viewing the play, then invites them to
see it his way [...]. This recommendation to read critical interpretations as
rehearsals and performances and not as true or false statements must
nevertheless submit to criteria of adequacy. Even though there is no such
thing as the true or best production or performance of Hamlet, there are
better or worse ones (262).
The foregoing discussion in the philosophy of criticism has significant

implications for the implied critic's decision style. Thus, as I have suggested at
the beginning of this chapter, the Quest for Certitude is intimately associated
with factualism; whereas Weitz and Margolis deprive the critic of the certainty
that can be derived from a factual statement that can be, clearly, "true" or
"false". Instead, they offer the critic "plausible" hypotheses, and definitions
which are "perennially debatable" (Weitz, 1972: 307), what is characterised as
"an honorific redefinition [...] that restricts the use of the term to a selection
from its multiple criteria" (ibid., 309). In order to cope with the demands on
his Negative Capability, a critic needs certain personality traits, summarised in
a classical statement of the attitude of a flexible person:
The categorical or conceptual attitude is characterised by ability or

readiness to assume a mental set voluntarily, to shift from one aspect of
the situation to another, to keep in mind, simultaneously, various aspects,
to grasp the essentials of a given whole, to break up a given whole into
parts and to isolate them voluntarily, to abstract common properties, to
plan ideationally, to assume an attitude toward "the merely possible", to
think and perform symbolically, and finally to detach our ego from the
outer world (Frenkel-Brunswick, 1968: 136, quoted above in the section
"The Implied Critic's Decision Style").
It should be noticed, for instance, that Margolis' conception of a critical

hypothesis as what can be plausible rather than true does require "an ability to
assume an attitude toward the merely possible". Likewise, our discussion of
the data that are hypothesised, and the shift from one set of aspects of the
caverns to another, clearly require not only this ability, but also a "readiness to
assume a mental set voluntarily, to shift from one aspect of the situation to
another, to keep in mind, simultaneously, various aspects", and so forth.
I have been left with a certain debt to the reader. In one of the early sections
of the present chapter, I criticised Yarlott for stating, as a fact, what the
Khan's position was in relation to his slave-force, or what was his understand
ing of the sublime quality of the landscape he chose for his stately pleasure-
dome, and the like. It might be objected that I am unfair to Yarlott. Every
interpretation involves the addition of information ("myths", or "hypotheses",
in Margolis' terms, which are schemata of the imagination) that is not
explicitly stated in the text. Margolis gives the following account of the issue:
The imaginative schema (or "myth") that the critic uses [...] may merely
be a formulable conviction about life that the artist himself may be
supposed to have held, which, considered without regard to its own truth or
falsity, adequacy or inadequacy, may, in the hands of the critic, enable us
to impute a coherent design to a work otherwise defective or puzzling in
this respect (Margolis, 1962: 114).
I shall not argue that Yarlott's information concerning the Khan does not
satisfy Margolis' criteria for "schema of the imagination" (though it does not),
but shall rather point out that the first part of "Kubla Khan" is not "defective
or puzzling" or incoherent in this respect, unless we regard uncertainty as to
the value of things as a defect or as puzzling.21 Moreover, this information
does not help us to "impute a coherent design" to the poem. It merely helps to
impute an evaluative ingredient to the otherwise unevaluated description. Or
consider another issue in Yarlott's discussion. As we have seen, he finds the
same two adjectives in the description of Kubla's garden and in "Christabel,
where snake joins 'bright' and 'green' (the only colour details found in Kubla's
garden) in a cluster of positive malignancy: [...] bright green snake". This leads
him to the conclusion that the description of the garden "produces sinister,
almost reptilian, associations". Now, is this not an attitude that "the artist
himself may be supposed to have held"? that may well be the case. However,
two important issues arise in this connection. First, after having written
"Christabel", cannot Coleridge describe anything as bright and green without
producing sinister, almost reptilian, associations? Second, what we have here,
isn't exactly a formulable conviction about life, but an all too concrete, not
sufficiently general or abstract hypothesis. We may even add here a third
point: far from imputing a coherent design upon an otherwise defective poem,
it rather atomises a quite coherent poem. In this way, it encourages rapid
21
One must distinguish the possible defectiveness, puzzlingness, incoherence of
the poem as a whole, from the coherence and precision of the description of
the concrete landscape where the pleasure-dome was to be built (cf. e.g.
Bloom, 1963: 230).
rather than delayed closure. The best way to demonstrate, again, the flaws of
such a critical practice is to consider here a genuine case of an imaginative
schema that does, "in the hands of the critic, enable us to impute a coherent
design" to "Kubla Khan".
Throughout the present chapter I have been judging Beer's critical practice
with severity. It will be but fair, I think, to acknowledge also when I agree with
him and admire his work. Many prominent "textbook-examples" of interpret
ations take a ready-made myth or an "imaginative schema" that has a high
degree of unity by its very nature, and offer it as a hypothesis to impute
coherence on the whole work (one such exquisite example is Fergusson's
interpretation of Oedipus Rex, based on the Dionysus ritual; another might be
Bodkin's "death and rebirth archetype" in interpreting Coleridge's poems).
But this need not necessarily be the case. Beer offers us a fine example of
"synthesising" an imaginative schema that is "tailor-made" for the interpreta
tion of "Kubla Khan". From his wording it is not quite clear whether he offers
it as a "true" or rather as a "plausible" interpretation, but even if he means it as
a "true" one, the schema has all it needs to serve as an excellent example of
Margolis' theory. On pages 226-229 Beer discusses at some length Coleridge's
conception of the "absolute genius" and the "commanding genius". Here I shall
reproduce only one of his quotations, concerning the latter:
These in tranquil times are formed to exhibit a perfect poem in palace, or

temple, or landscape-garden; or a tale of romance in canals that join sea
with sea, or in walls of rock, which, shouldering back the billows, imitate
the power, and supply the benevolence of nature to sheltered navies; or in
aqueducts that, arching the wide vale from mountain to mountain, give
Palmyra to the desert. But alas! in times of tumult they are the men
destined to come forth as the shaping spirit of Ruin, to destroy the wisdom
of ages in order to substitute the fancies of a day, and to change kings and
kingdoms, as the wind shifts and shapes the clouds (Beer, 1959: 228).
Now, in the first place, this is doubtless "a formulable conviction about life
that the artist himself may be supposed to have held". Second, some critics have
held that the two lines "And 'mid this tumult Kubla heard from far / Ancestral
voices prophesying war" are puzzling; some even claim they are defective. If
we use Coleridge's conception of the "commanding genius" to impute a
coherent design to this poem, the incoherent elements are revealed as the two
sides of the commanding genius, the one dominant in times of peace, the other
dominant "in times of tumult". It should be noticed that such a hypothesis is
highly plausible not only because the poet himself may be supposed to have
held these views, but also because substantial parts of the first part of the
poem may be regarded as instances illustrating the acts of the commanding
genius in times of peace; whereas these two "puzzling" or "defective" lines
become hints at what may come of him. Later on Beer elaborates this concep
tion in a way that may illuminate the unity of the whole poem:
Kubla Khan, to sum up, is a poem with two major themes: genius and the
lost paradise. In the first stanza the man of commanding genius, the fallen
but daemonic man, strives to rebuild the lost paradise in a world which is,
like himself, fallen. In the second stanza, the other side of the daemonic
re-asserts itself: the mighty fountain in the savage place, the wailing
woman beneath the waning moon, the daemon-lover. The third stanza is a
moment of miraculous harmony between the contending forces: the sunny
dome and the caves of ice, the fountain and the caves, the dome and the
waves all being counterpoised in one harmony. Finally, in the last stanza,
there is a vision of paradise regained: of man re-visited by the absolute
genius which corresponds to his original, Unfallen state, of the honey-dew
fountain of immortality re-established in the garden, of complete
harmony between Apollo with his lyre and the damsel with the dulcimer,
of the established dome, and of the multitude, reconciled by the terrible
fascination of the genius into complete harmony (Beer, 1959: 266-267).
In view of this clear and illuminating imaginative schema it is hard to under

stand why Beer had to resort to so much erudition that directs attention away
from the poem. Even in this lucid "imaginative schema", the phrase "Apollo
with his lyre" gives away Beer's allegoristic temper. It is hard to tell how one
can reconcile these two kinds of critical activity of one and the same person. I
have my own guesses in this respect, but in the absence of satisfactory
arguments to support them, I prefer to leave the issue unresolved. The reader
may attempt to weigh the possible explanations for himself, in the vein of
arguments propounded in the course of the present chapter.
Summary
I have followed "Kubla Khan" criticism at great length on three levels. On the
first level I confronted pieces of criticism with the poetic text and examined
their adequacy. On the second I attempted to generalise from observations on
the preceding level to wider theoretical and methodological issues, taking into
account some of the theoretical generalisations of recent analytical philosophy
of criticism. On this level our inquiry seems to have contributed to the theory
of interpretation as well: it defined some of the evasive constraints upon
interpretation (such as the ones concerning the degree of abstraction). On the
third level I discussed the implied critic's decision style. We dwelt at length on
the dichotomous pair of attitudes Negative Capability and the Quest for
Certitude. The former is characterised by high tolerance of ambiguity and un
certainty, the latter by anxiousness to avoid them. Keats claimed that Negative
Capability is a prerequisite for literary accomplishment as a poet and, we
might add, as a reader as well. In this respect the critic is, in the first place, a
reader some of whose responses to literature are publicly accessible. We en
countered and isolated a considerable number of critical strategies and devices
in the writings of critics that seem to reflect an effort to avoid ignorance,
uncertainty and ambiguity, emphasising time and again that it is not necessarily
the adequacy of critical methods or the validity of specific statements that
determine the implied critic's decision style. Critical activities are determined,
in the first place, by prevalent critical approaches, academic climate, and
exposure to certain critical theories rather than others. Walter Weimer criticised
the name given to Popper's classical book The Logic of Scientific Discovery; it
should rather be called, he says, The Logic of the Completed Scientific Report.
Scientific discovery follows its own logic, says Weimer. Our business in the
present chapter has been to try to recover from behind the prevailing norms of
"the completed scientific report" the traces of the critic's process of discovery.
Certain deviations from these prevailing norms in either direction, and
consistency in these directions, may indicate the implied critic's decision style.
The term "decision style" suggests a more or less homogeneous set of critical
decisions. We found, indeed, that some critics do manifest a considerable
degree of consistency in their use of devices, whether in the service of the
Quest for Certitude or in the service of Negative Capability. But this cannot be
taken for granted either. We encountered instances in which opposite tenden
cies occurred in the writing of the same critic. In some cases such
inconsistencies can be explained. In others it might be wiser to keep silent and
wait for the emergence of further evidence.
Ecstatic states consist in a loosening of conscious control. As a result,
ecstatic poetry may arouse a sense of uncertainty or even anxiety. That is why
an ecstatic poem like "Kubla Khan" is more than usually apt to elicit
interpretations that manifest the syndrome of the Quest for Certitude. In a
Hebrew paper (Tsur, 1985b; 1996: 119-205), I examined a corpus of interpret
ations devoted to another genre that is more than usually apt to elicit
interpretations that manifest the syndrome of the Quest for Certitude: a
corpus of fiction described as "Literature of Extreme Situations", including
such works as Kafka's The Castle, the theatre of the absurd, and a large corpus
of short stories in Hebrew. A comparison of the findings of these two inquiries
are illuminating. The Quest for Certitude tends to have recourse, with regard to
both corpuses, to the same strategies of defence and avoidance.
The Texture and Structure
of "Kubla Khan"
Most discussions of "Kubla Khan" are concerned with one of two questions or
both: "What is the meaning of the poem?", and "What are the poem's source-
materials?". What is common to both approaches is a focus on a relationship
between the text and something outside it. There is nothing wrong with that, of
course. However, most such discussions dwell far more persistently on the
external member of this relationship (the meaning or the putative source) than
the internal member (the text) or the relationship itself. One reason for this is
presumably that it is easier to handle gross issues outside a poem than its own
complex structure, or the intricate relation between it and something outside.1
In the preceding chapter I discussed and criticised at great length some writings
of this kind on "Kubla Khan". In the present chapter I propose to focus more
attention on the text itself and to point out various aspects in it, attempting to
integrate them into a coherent reading of the poem and to foreground, by the
same token, its unique texture. Briefly, I will adopt Coleridge's own assertion
in a slightly different context that "nothing can permanently please which does
not contain in itself the reason why it is so, and not otherwise" (Biographia
Literara, Chapter 14).
Romantic Nature Description
In the preceding chapter I had occasion to point out several of the virtues of
Schneider's (1975) study of Coleridge's poem. Coming now to offer my own
reading of the poem, I again cannot ignore her work. She did not write, as she
claimed Lowes did, "as if Coleridge had existed in eternity but not in time"
(Schneider, 1975: 262). She seems, in fact, to be one of the few critics (if not
the only one), who treated "Kubla Khan" as a poem that is part of a poetic
tradition, in this case the romantic tradition, that may have belonged to a poetic
genre worth inquiring into, and even had a prosodic texture. Here, again, I find
myself setting out using one of her observations as my point of departure:
A recent paper (Benzon 2003) is devoted to the complex structure of the

poem and its relationship to neuropsychological processes.
for those who have felt it to be the quintessential romantic poem, some
thing of a point remains, for it lies squarely upon a crossroads where two
or three main romantic traditions meet2 (Schneider, 1975: 262).
Though I shall only occasionally allude (and have, indeed, occasionally

alluded in the preceding chapter) to her treatment of some of these traditions,
my approach to "Kubla Khan" may be characterised as regarding it, in many
respects, as "the quintessential romantic poem". Thus, to Schneider's "three
main romantic traditions" I wish to add two more: I propose to regard the first
part of the poem as a descendent of the main line of romantic nature
descriptions, and the whole as a hypnotic-ecstatic, secular-mystic poem.
The site chosen for the building of the stately pleasure-dome has many of
the characteristics of rather familiar landscapes of realistic plenitude, of human
proportions, such as in "Five miles meandering in a mazy motion,/ Through
wood and dale the sacred river ran", or in "And there were gardens bright with
sinuous rills,/ Where blossomed many an incense-bearing tree". This, however,
is inextricably interwoven with the immense forces of the infinite and the
sublime, as incarnate in the "caverns measureless to man" (twice), the "deep
romantic chasm", "lifeless ocean", or "sunless sea". We are here up against two
kinds of estimation of magnitude, according to the Kantian conception of the
sublime:
When we estimate magnitudes through numbers, that is, conceptually, the

imagination selects a unit, which it can then repeat indefinitely. But there
is a second kind of estimation of magnitudes, which Kant calls "aesthetic
estimation", in which the imagination tries to comprehend or encompass
the whole representation in one single intuition. There is an upper bound
to its capacity. An object whose apparent or conceived size strains this
capacity to the limit—threatens to exceed the imagination's power to
take it all in at once—has, subjectively speaking, an absolute magnitude: it
reaches the felt limit, and appears as if infinite (Beardsley, 1966: 218-
219).
In "Kubla Khan" we have both kinds of estimation, the indefinitely repeat-

able unit selected ("miles"), and what is "measureless to man", exceeding the
imagination's capacity to comprehend or encompass the whole in one single
2
By "romantic traditions" Schneider did not mean "traditions that distinguish
romanticism from other trends", but "a major line of poetry", a synthesis of
imagery that "had been created by the literary tradition itself, running back
through Milton to Spenser, Ariosto, and the other Renaissance romancers who
grafted these new trees upon medieval stock" (p. 262). "In Milton, as in
Ariosto, all the strands are united. Afterward they sometimes descend
separately, through Gray and Collins, through Addison and Johnson" (263);
and so forth.
The Texture and Structure of "Kubla Khan' 81
intuition. Psychologically speaking, Bodkin's (1963: 104) characterisation of

"the imagination, seeking something enormous, ultimate" seems to be relevant
here, "as when standing on some precipice edge, amongst peaks and chasms,
one feels their lines overpowering and terrible through the suggested anguish of
falling. That horror overcome adds a kind of emotional exultation to the sight of
actual mountain chasms" (as quoted in the preceding chapter). This "emotional
exultation" becomes a significant ingredient in the pervasive emotional qualities
of the poem, and will be discussed later.
In order to appreciate Coleridge's poem, a further Kantian distinction must
be made, "between the mathematical sublime, which is evoked by objects that
strike us as maximally huge, and the dynamic sublime, which is evoked by
objects that seem to have absolute power over us" (Beardsley, 1966: 218). In a
very important sense, "Kubla Khan" proceeds from the former to the latter
kind of the sublime. It is this feature that infuses the natural landscape with
tremendous energy: beginning with the maximally huge "caverns measureless to
man", through the dynamic sublime in the holy, enchanted and haunted
landscape, to the speaker's frenzy at the end of the poem, that seems to have
absolute power over his audience. Under further scrutiny, this sublime energy
undergoes a gradual transformation, from a vision of solid, stable objects of
nature, to a gradually increasing gestalt-free and thing-free vision. The "caverns
measureless to man" in the first stanza "strike us as maximally huge", and as
such tremendously powerful. This power, however, is static and inseparable in
our awareness from the stable physical thing itself. This may be usefully
contrasted with what happens in the first five lines of the second stanza.
"[The] woman wailing for her demon-lover" has attracted much attention in
criticism. A number of critics assume that there actually is such a woman
wailing in that place; some (e.g. Yarlott, 1967: 130) even guess that her wailing
may reach the Khan's ears. Fruman (1971: 397) goes so far as to claim that
"the woman wailing for her demon-lover is calling for an incestuous lover".
Few critics admit that the syntactic structure is ambiguous, and that there may
or may not be such a woman. In the latter case, the construction "as
ADJECTIVE as ever VERB PHRASE" is to be construed as a superlative, sug
gesting a quality in an extreme form that cannot be surpassed. In the present
case, it suggests an extreme instance of a certain kind of holy and enchanted
quality. The underlying conception suggests one of Kenneth Burke's favourite
ideas concerning "the principle whereby the scene is a fit 'container' for the
act, expressing in fixed properties the same quality that the action expresses in
terms of development" (Burke, 1962: 3). Assuming that there is no such a
woman there, the landscape becomes "a fit 'container' for the act" and the
actors that are not there. In other words, the landscape expresses in fixed
properties the same quality that the action had there been one would have
expressed in terms of development. This is, in fact, what is explicitly said in
the passage. In other words there is, in the scene, a sense of an extremely
powerful absence, indicating a supersensuous presence of a thing-free and
gestalt-free quality that can be characterised as holy, enchanted, demonic,

mournful, and the like. Some of these features are reinforced by various mean
ing components of the specific items of the description. Demon suggests, in
the first place, an in-dwelling spirit, reinforcing the thing-free and gestalt-free
quality suggested by the superlative construction and Burke's principle of "fit
container". The absence is, again, suggested, by the meaning component [LOSS]
implied by wailing. The sense of loss and gloom is reinforced by the waning
moon. Demon suggests great energy, as well as divine or evil nature;
likewise, savage suggests great energy and, possibly, destructive power.
Savage, as an epithet of place does not indicate actions, but potential violent
actions, expressed in fixed properties of the scenery. The wailing for the loss
of a lover, again, infuses the scenery with enormous emotional force. In short,
then, the description suggests a quality of intense, thing-free and supersen-
suous presence, loaded with immense emotional energy.3
On closer inspection, the construction "as ADJECTIVE as ever VERB
PHRASE" is not ambiguous at all. It serves as an intensifier for adjectives, as a
phrase of enforcement. Its apparent ambiguity is derived from the fact that it
may occur both in sentences that refer to actual conditions and in sentences
that refer to hypothetical or rejected conditions. But its main meaning compo
nents are [+EMPHATIC + EXCEEDING + in any possible case]; and unless the
presence of the "woman wailing" is explicitly specified as an antecedent to the
construction, it remains merely potential (as opposed to actual). In fact, even in
contexts where the actual presence is explicitly specified, the "superlative" is
achieved by likening the actual case to some potential extreme case. Again it is
Schneider (1975: 249) who makes the proper distinction: "the demon-lover
lines [are] not a new scene but only a comparison". A recent paper (Benzon,
2003), which acknowledges familiarity with the earlier version of the present
book, offers an explanation of the artistic context of this comparison: "Here
the external world of Xanadu is likened to subjective states (the chasm is like a
woman possessed by desire for her lover) and bodily processes (the eruption
of the fountain is like breathing)".
Why then do so many critics, native speakers of English and "well-nurtured
in their mother-tongue", misunderstand the construction? The reasons for this
seem to have little to do with the exact meaning of this idiomatic construction.
In the preceding chapter I adduced several instances of "interpretations" (that
had no such near-ambiguous constructions to rely on) where the only way to
account for the "missing information" supplied by the critic seemed to be
either his inability to assume an attitude toward the "merely possible", or his
reluctance to contemplate absence as a significant attribute of the poem. Here
3
Kipling commented on these lines: "In all the millions permitted there are no
more than five—five little lines—of which one can say 'These are magic.
These are the vision. The rest is only poetry'". The other two are Keats's
"Ode to the Nightingale", 11. 69-70.
The Texture and Structure of "Kubla Khan" 83
the relatively long description of a scene haunted by a woman wailing,

introduced by an idiomatic phrase that may or may not allude to a merely
hypothetical situation, makes it easier to introduce "missing" information as
facts. At any rate, critics who argue from the presence of the wailing woman
for the incoherence of the first part of the poem, seem to be killing a straw-man
of their own making. Thus, the gross misreading of the emphatic phrase may
be a useful device for introducing e.g. the motive of female inspiration into the
poem. 4 On the other hand, the foregoing analysis of the passage establishes it
as a remarkable piece of romantic nature description.
The ensuing description of the fountain and the river gave considerable
trouble to some critics. It forcefully impresses the imagination, but gives little
or no cue as to its "meaning". Thus we find such discussions as Watson
arguing with himself: "The vast power of the river is allowed to rise, but only
'momently', and then sinks back into silence, 'a lifeless ocean'. This is surely
not the River of Life. It is the river of the poetry of imagination" (Watson,
1973: 233). Though he seems quite confident in his preference of the river's
meaning, one thing seems to be absolutely certain for Watson, that the river is
The River of something.
Philology, too, may be invoked for removing such uncertainty. In a recent

book on "Kubla Khan" displaying a persistent attitude of the Quest for
Certitude, Robert F. Fleissner (2000: 53) takes issues with Charles I. Patterson,
who "tried to make a case for the 'demon-lover' in Samuel Taylor Coleridge's
dream-reverie as being a tutelary agent. His main point was that the term
demon- in print was spelled 'Daemon' in the Crewe MS. of the poem and
therefore suggested the Greek term for such a spirit". Fleissner provides solid
logical arguments to "prove" that Coleridge did not mean "daemon", but
rather "demon". Such a controversy can not only help us arrive at Coleridge's
"true" meaning, but may also help to get rid of the ambiguity of the word
"demon". There are, however, good reasons to suppose that the word demon
covers both meanings. The "DAEMON" entry of the Oxford English
Dictionary curtly says: "see DEMON1, etc". Sense 1. a. of "DEMON", in turn,
is defined as "In ancient Greek mythology: A supernatural being of a nature
intermediate between that of gods and men; an inferior divinity, spirit, genius
(including the souls or ghosts of deceased persons, esp. deified heroes). Often
written dæmon for distinction from sense 2". Merriam-Webster's Collegiate
Dictionary, and The Randon House College Dictionary give similar informa
tion. By the same token, the above strategy enables Fleissner to designate
Kubla Khan as "another major dictator" via a quotation from Goethe on
Napoleon. This is a stereotype not supported by the poem's text (from which
we know nothing about Kubla Khan except his building enterprise). As we saw
in Chapter 1, some historical sources at least suggest that Kubla was more
enlightened than one would expect from a Tartar khan. But such an intimation
provides the critic with something positive to cling to.
Yarlott speculates in greater detail upon the "meaning" of the description. In

the first place, he observes that "the fountain's inexhaustible energy signifies
the act of creativity" (Yarlott, 1967: 142). As for the position advocated here,
the fountain's inexhaustible energy "signifies" nothing. It only may be (and in
fact is) in some respects creative, and destructive in others.
After rising with difficulty it wanders 'mazily' through the pleasure-

garden, then sinks into a lifeless ocean. It appears to seek at first to
challenge and disrupt the ordered artificiality of the paradise, scattering
fragments of rocks like hail or chaffy grain. But amid such inimical
conditions nothing comes of the creative energy (Yarlott, 1967: 142).
While Yarlott attributes a specific purpose to every act of the fountain, the
present approach regards the description of the fountain as being characterised
by "purposiveness without purpose" (to use a Kantian phrase): these actions
only present the fountain in its most sublime aspect. In this respect, I side
with Humphrey House, who regards the outburst of the fountain as a uniquely
powerful, unclassifiable event, and does not attempt to classify it, but speaks
of "the sense of inexhaustible energy, now falling, now rising, but persisting
through its own pulse".
The whole passage is full of life because the verse has both the needed
energy and the needed control. The combination of energy and control in
the rhythm and sound is so great, as in
at once and ever

that we are in danger of missing the force of imagery, as in "rebounding

hail" and "dancing rocks". If we miss it, it is our fault, not Coleridge's; and
it sometimes appears as if readers are blaming or underestimating him
because they improperly allowed themselves, under the influence of the
rhythm, to be blind to the "huge fragments" and "dancing rocks" which
lay another kind of weight upon it, and to be blind to the construction of
the thought, which holds together the continuity and the intermission
(House, 1973: 204).
House's main object in this discussion seems to be to make sure that the im
agery of the passage is not misconstrued so as to mitigate its power, to "dom
esticate" the sublime. I propose, then, to consider the details of the description
of the fountain's outburst as meant to amplify the revelation of nature's
"inexhaustible energies" at its most sublime on the one hand and, on the other,
to add an "irrelevant concrete texture" so as to amplify the impression of
"purposiveness without purpose". Any specific purpose attributed to the
details, reduces the sublime or aesthetic quality of the description. According-
ly, while Yarlott credits the outburst of the fountain with such purpose as "to
seek at first to challenge and disrupt the ordered artificiality of the paradise,
scattering fragments of rocks like hail or chaffy grain", the present chapter
conceives of the same event as of a purposeless outburst, characterised as
sublime in several respects. It shows nature's hidden forces at work, with a
violence that seems to "exceed the imagination's power to take it all in at once"
(in creating, so to speak, the fountain). At the same time, it seems to "regress"
to a stage where the forces of chaos seem to be still active, toying around with
huge "primordial" fragments of rocks.5
I shall return to discuss the rhythmic character of this passage, and its
interaction with the description of the event or process of flinging up the
fountain, yielding immense energy and vigour. At present I wish to have a
closer look at the ensuing description of the river:

And sank in tumult to a lifeless ocean:
Here, even the purposelessness of the "mazy motion" of the river may have
a "moral" purpose for some critics: "The river issues at last only to meander
with purposeless 'mazy' motion, and 'mazy', likewise, was a characteristic
Coleridgean term for describing moral and spiritual uncertainty" (Yarlott, 1967:
143). According to my conception, however, one should attempt to abstract
the "unique, unclassifiable" perceptual quality of the river running in a mazy
motion, without translating it into a conceptual system of moral features and
purposes. The purposeless, mazy motion has a relatively relaxed quality about
it, especially after the highly tense and dramatic quality perceived in the "huge
fragments vaulting like rebounding hail", and the "dancing rocks". This
"relaxed" quality will be more apparent, if we compare the details of this
description to those of the essentially identical one in the first stanza:

5
Different critics may direct attention to different aspects of the image.
Consider William Benzon's (2003: online) remark: "In the image of the
dancing rocks the distance between these two semantic spaces vanishes; the
vaulting fragments are poetically tamed by the harvest image, threshing grain.
In line 23 the rocks are dancing; the dance comes from man, the rocks from
nature". In my opinion, on the contrary, the implication is that the enormous
forces toy around with the huge fragments as with such small things as hail and
grain.
In both descriptions we get essentially the same information about the

river's "behaviour". But a detailed comparison points up the greater tension of
the earlier passage, and the more "relaxed" quality of the later one. As for the
contents of the description, it pays more attention to the details of the
environment: "Five miles meandering with a mazy motion / Through wood and
dale" occur only in the later description. This suggests, perhaps, a certain
leisure in the describing consciousness. As to the syntactic structure, the earlier
passage contains a single finite verb, whereas the later passage contains three
finite verbs to indicate the same progression of the river: it "ran ... then reached
... and sank". This, again, may indicate some leisure in the observing con
sciousness, imputing a considerable degree of articulation to the description. If
I may borrow a pair of descriptions from Auerbach (1962: 21) who, in turn,
borrowed them from Goethe and Schiller, what the later and longer description
gives us is "simply the quiet existence and operation of things in accordance
with their natures", whereas the earlier description "robs us of our emotional
freedom". These contrasting qualities are reinforced by the respective prosodic
structures. The four lines of the later passage are organised in a stable,
symmetrical structure by an abba rhyme-scheme, whereas the three lines of
the earlier passage are, as I hope to show later, part of a much more complex
structure, one of whose typical effects is to upset any kind of focal stability.
A further element of restlessness can be found in the peculiar correspondence
(or rather lack of correspondence) of the syntactic structure with the prosodic
unit in "Where Alph the sacred river, ran / Through caverns". As I have
pointed out elsewhere (Tsur, 1972: 130), the nearer the syntactic break is to
the end of a verse line, the greater our relief when the missing part is supplied.
By the same token, the nearer the beginning of a run-on syntactic unit is to the
end of a verse-line, the greater the tension thus generated. Thus, the "re-
quiredness" of ran is very high in this instance, and so is the "momentum"
generated by the run-on sentence, underlining the speed-aspect of ran. In the
later passage, by contrast, the phrase "with a mazy motion" begins exactly in
the middle of the line, and so is not perceived as a run-on line, but almost as an
end-stopped one. Even the sequence "the sacred river ran", apparently
identical letter by letter with the corresponding sequence in the other passage,
differs from it in two important respects. First, ran in the earlier passage
begins an enjambment, whereas in the later passage it serves as the closure of
an end-stopped line. Second, punctuation enhances the break before the end of
the line in the earlier passage, and by the same token heightens the impetus of
the run-on sentence, whereas in the later passage the articulating commas are
omitted, and the "requiredness" of ran is somewhat toned down. As for the
length of lines, I have indicated in my 1972 paper as well as elsewhere (Tsur,
1977; 1998), that the iambic pentameter line possesses a peculiar kind of
flexibility, owing to the fact that it cannot be divided into two symmetrical
halves. This metre is therefore more suitable than any other to the cadences of
normal speech, whereas the iambic tetrameter line has a particular rigidity,
owing to the fact that it can be divided into two exactly identical halves. Now
the two passages can be contrasted in this respect too: the later passage
contains four lines of iambic pentameter, the earlier one contains two lines of
iambic tetrameter and one trimeter. The tension in the two lines of tetrameter is
heightened by the fact that in both the caesura (after the fourth position) is
overridden by language. Hence, again, the relative leisure of the later passage is
corroborated.
Now the relative leisure that emerges from this comparison, and is
intuitively perceived by the reader, is so meticulously established only to be
suddenly destroyed, in the fourth line of the passage. We have seen that as
part of the syntactic pattern of three finite verbs, sank reinforces rather than
disrupts this leisurely quality. At the same time, semantically, the verb phrase
sank in tumult introduces commotion into the "idyllic" description. It indicates
an outburst of violent energy and noise, and the sudden disintegration of the
linear (though meandering) shape of the river. In this context, "lifeless ocean" is
to be regarded as the amplification of "sunless sea". Thus, the leisurely quality
becomes functional in the poem, in its system of oppositions.
The river, then, cannot be regarded as the river of life, or the river of any
thing. It is a river brought into the focus of attention to such a degree that the
reader tends to abstract from its description certain qualities that appear to
have high emotional significance. If not the river of life, at any rate, water is
regarded as the source of life; and running water is perceived as living water. In
the description of the mazy motion of the meandering river a leisurely quality
has been pointed out. It is foregrounded by a comparison to a nearly identical
passage, and the contrast involves the stratum of described reality, as well as
the syntactic and the prosodic strata.
Things and Thing-Free Qualities
Another, more immediate foregrounding occurs when this description is con

trasted to the immediately preceding and immediately following stages of the
river's progression. In the preceding passage, the fountain is "flung up" with
"ceaseless turmoil seething"; and in the fourth line of the passage the river
sinks "in tumult to a lifeless ocean". Here Alph loses its identity in the lifeless
ocean. Thus, the opposition "life versus death" is expressed in both action-
language and static scene-language: on the one hand, the contrast between the
running, meandering river and its being lost in the ocean; on the other, the con
trast between "wood and dale" and "lifeless ocean" (or "sunless sea"). Thus
the line "And sank in tumult to a lifeless ocean" contains death or ceasing-to-be
in both languages. It should be noticed that these are not symbolic meanings of
any kind; we are dealing here with the imaginative perception of opposite
qualities, not with meanings. When we perceive the linear shape of a running
river, ending in the tumult of sinking into a lowly situated ocean, we directly
perceive disintegration and the release of hitherto contained violent energy, as
aspects of the event, but not as its meaning.
It is in this context of "thing-destruction" that we should consider the next
two lines:
3. And 'mid this tumult Kubla heard from far

Ancestral voices prophesying war!
Wilson Knight comments on these two lines:
... and sinks with first more tumult (i.e. death agony), to a "lifeless ocean",
that is, to eternal nothingness, death, the sea into which Timon's story
closes. This tumult is aptly associated with war: the principle of those con
flicting and destructive forces that drive man to his end. The "ancestral
voices" suggest that dark compulsion that binds the race to its habitual
conflicts and is related by some psychologists to unconscious ancestor-
worship, to parental and pre-parental authority (Knight, 1960: 165).
One may, I believe, accept such an interpretation of "Ancestral voices",

with some modification or other. The irrational and primordial elements are
conspicuous here. Our foregoing analysis, however, adds an all-important,
structural dimension. The perception of an irrational quality resides not only in
the sublime force of the tumult, but also (or, perhaps, foremost) in the diffuse
structure of the downpouring waters, closely ensuing after the linear run of the
river. As I indicated in the preceding chapter, the opposition between linear
and diffuse processes has a close structural resemblance to the opposition
rational versus irrational mental processes or, rather, the information-output of
the processing activities of the left and the right hemispheres of the brain.
Visually, the downpouring water and the "lifeless ocean" are perceived as
gestalt-free entities; the former belongs to the "dynamic", the latter to the
"mathematical" sublime. The "tumult" is inarticulate noise, that is, again, a
gestalt-free and thing-free quality; at the same time, it indicates enormous
energy. These serve as the perceptual medium of the "ancestral voices", which,
too, are thing-free qualities. This quality is reinforced by a grammatical
manipulation: in constructions like "the voices of ancestors" the ancestors, the
stable things, the stable source of the thing-free voices, still linger in the back
ground; by transforming the semantic information into an adjective (ancestral),
we have only the purely thing-free entity. It is this section, then, between the
"birth" and "death" of the river, where we get, in the most concentrated way,
the dissolution of solid things into thing-free and shape-free qualities: beginning
with the disintegration of the solid earth into "huge fragments" vaulting from
the depth of the earth, through the fluid river's losing its identity in the
"lifeless ocean" and ending with the thing-free entities of "ancestral voices". In
this sense, it is a pivotal passage.
The next few lines shift the focus of the visual image:
4. The shadow of the dome of pleasure

Floated midway on the waves;
Where was heard the mingled measure
From the fountain and the caves.
It was a miracle of rare device,
A sunny pleasure-dome and caves of ice!
But in an important sense, this image carries on perceptual processes that

began in the preceding lines. I mean the "dissolution" of the solid world into
thing-free or gestalt-free entities. First of all, we are confronted here not with
the solid dome of pleasure but with its shadow, which, though it may have a
clear-cut gestalt, can also be regarded as a most typical instance of thing-free
entity. However, being buoyant on the surface of the waves, it is continuously
modifying its stable shape, thus becoming a most typical image of an ever-
changing, shifting physical reality, that has—in this way—a structural
resemblance to emotional processes. The lightness of the visual percept is
corroborated by the sudden metrical shift from iambic pentameter to trochaic
tetrameter (with a light hypermetric syllable at the beginning of the first line).
The shadow of the dome floated midway—midway between the fountain and
the caves, I presume. This suggests a symmetrical disposition of the percep
tual space, that "counts toward" a strong gestalt. At the same time, this is the
spot where the inarticulate noises and tumult from the fountain and the caves
(mentioned in the two preceding sections of the stanza) meet and mingle. Now,
the relatively stable objects (the fountain and the caves) are far away, and only
a thing-free and gestalt-free entity, namely, the mingled sounds emitted by
them, is perceived. Upon this thing-free and gestalt-free entity a symmetrical
orientation-scheme (suggested by "midway") is superimposed.6
6
I have adapted to literary criticism the terms thing-free, gestalt-free, thing-
destruction, superimposition from Ehrenzweig (1965), who applies them to
music and the visual arts.
"Stability, constancy, consistency, differentiation" are among the key
words for the effects of cognitive organisation. In the visual mode, for
instance, from a stream of undifferentiated stimuli we differentiate a
stable, consistent world. [...] Whatever visual information can be
organised into clear-cut shapes or well defined objects, are emphasised,
promoted, organised as "figures" that stand out clearly against lowly-
differentiated "ground". All other visual information is relegated to the
mass of lowly-differentiated background. [...] Whenever we see a
person from a different angle, or in a different lighting, we receive
different visual information; were it not for perceptual constancy, we
Turning now to the last stanza of "Kubla Khan", the scene is radically
changed. We are in an environment where no trace of Kubla's pleasure-dome is
left. While the poem till now was almost exclusively devoted to a physical
scene, the last stanza (the "second part") of the poem "takes place" some
where detached from any physical background. We only know that in certain
circumstances "all" would react in a certain way to the speaker's behaviour.
We know nothing about who, or how many, or where, those "all" may be. We
have even very little knowledge about who "I" may be.7 It is the mental event
here that fills the entire present, but even this is mostly delivered in the
rejected conditional mode. All physical background has been removed. As for
the emotional mood, the last stanza seems to reach the peak of an emotional
experience, best described as ecstasy. The flashing eyes and floating hair
indicate violent mental agitation, wild excitement or enthusiasm. Schneider
(1975: 245-246) and several critics after her have pointed out that "The
description derived a good deal from the accounts of persons possessed by the
god in Dionysus worship and the Orphic cults—flashing eyes and streaming
hair, as well as honey, milk, magic, holiness, and dread. [...] Plato's Ion gives
what is probably the most famous passage" (I discussed this issue in some
detail in the preceding chapter). But, I believe, even readers who know little
about Plato or Dionysus worship, or Orphic cults, readily recognise here the
peak of an emotional experience. Now, what appears to be of great importance
here is, first, that the speaker (whoever he may be) arouses, when in the mental
ought to perceive each time a different person (if we could perceive, at

all, persons) (Tsur, 1992a: 16-17).
We have bought perceptual constancy at a considerable price. The capability
of responding to rich precategorial information is no less vital for survival.
We do perceive some of the inconstant, precategorial, inarticulate
information; the knowledge so gained is usually called intuitive.
Intuitive knowledge so gained is indispensable for quick orientation, or
for orientation in an ever-changing environment. In fact, as Bartlett
indicated back in 1932, most of the complex cognitive activities, such
as perceiving complex situations or remembering them, begin with the
awareness of some such precategorial information. He calls this
awareness "attitude", or "feeling", or "affect" (Tsur, 1992a: 17).
This state of affairs may explain the function of the phenomena denoted by
Ehrenzweig's terms in the artistic endeavour. In a work of art, where
communication is based on hard and fast categories, attitudes, feelings, or
affective and emotional qualities may be generated only by the partial
destruction of categories, things and gestalts. The enormous emotional force
perceived by some readers in "Kubla Khan" could not impress the readers of
words (which, as we know, denote stable categories) unless some kind of thing-
free and gestalt-free qualities were generated in the poem.
Lowes (1927), and some critics in his footsteps, speak—oddly enough—of a
Tartar youth.
state described, "holy dread" in his audience, not unlike the numinous: "For
man shall not see me and live" (Exodus, 33: 20). Second, this mental state is
somehow related to his ability to revive within himself the symphony and
song of an Abyssinian maid, and to rebuild with music loud and long the
impressive sight described in detail in all the preceding stanzas. Now, what
seems to be of even greater importance within the framework of the present
discussion is that music is, by its very nature, a preeminent instance of thing-
free quality. Thus, the peak of the emotional experience generated in the poem
occurs at a point where all physical background is removed, and there are only
mental experiences and thing-free qualities; even Kubla's building is said to be
rebuilt as a thing-free entity, of music (I shall return to this "peak-experience"
later).
There is an additional cluster of issues that deserves consideration. It may
appear significant that, as mentioned in the last two paragraphs, the verb
"float" occurs twice in this poem: "the shadow [...] floated", and "his floating
hair". This kind of imagery appears to add something essential to ecstatic
poems. Wordsworth's "Daffodils", too, begins with "I wondered lonely, as a
cloud / That floats on high, o'er vales and hills". Elsewhere (Tsur, 2003: 27-28)
I discussed Wordsworth's lines from two different perspectives: Maud
Bodkin's archetypal approach to poetry, and Michael A. Persinger's neuro
psychological study of God experiences. I will reproduce here some of the
discussion, and expand it in certain respects.
In her discussion of Dante's "Paradiso", Maud Bodkin speaks of "flight as
it is known in dreams". She comments on a very different image, characterising
its effect as "the absence of any sensation of effort, the wonder at effortless
attainment of a new sphere" (Bodkin, 1963: 143). The "cloud floating on high,
o'er vales and hills", or the "shadow of the dome of pleasure floating on the
waves" are natural symbols for precisely such effortless movement. The later
expression, "Beware! Beware! / His flashing eyes, his floating hair!" apparent
ly makes a quite different impression. Yet closer scrutiny may illuminate the
relationship between the two occurrences of the verb "float". Let us begin with
a couple of dictionary definitions of the obviously relevant meanings of this
verb: "to move or rest on the surface of a liquid without sinking"; "to move
slowly and lightly through the air". Some dictionaries connect this verb with
"suspend", defining its relevant meanings as "to hang so as to be free on all
sides except at the point of support"; "to keep from falling or sinking by some
invisible support (as buoyancy), as in 'dust suspended in the air'". In this
sense, "effortless attainment of a new sphere", the clouds floating on high and
the shadow floating on the waves become a metaphor for a physical (or
mental) state in which the resistance of the physical world is suspended
("suspended" in the sense of "made temporarily inoperative"). This, in turn,
may suggest some kind of ease, leisureliness or, in Dante's case, an inner peace
enjoyed by the soul in heaven. The flashing eyes and floating hair, by contrast,
indicate an inner turbulence, very much unlike the leisureliness or "inner peace"
associated with effortless movement. Notwithstanding, such enormous emo

tional force remains quite compatible with moving lightly through the air, or
with being kept from falling or sinking by some invisible support (as
buoyancy). In this respect, the "floating hair" retains the quality of moving
lightly through the air; at the same time, its emotional energy level is amplified,
"heightened, to any degree heightened"—thus contributing to the perceived
quality of an altered state of consciousness that involves enormous emotional
force (see below my discussion of "emotive crescendo").
Michael A. Persinger's (1987) neuro-psychological study may illuminate
this kind of imagery from a different aspect. He observes that God Experiences
(as well as some pathological conditions) are associated with temporal lobe
transients, electrical perturbations of the temporal lobe in the human brain
(16). I assume that, psychologically, a "God Experience involving temporal
lobe instability" (26) is relevant to ecstatic poetry too. I will explore the
relevance of two characteristics of such states: depersonalisation and vestibular
sensations. Later I will consider Persinger's claim that such states are "an alter
ation in the description of the self. Depersonalisation is typical" (Persinger,
1987: 18). As to the relevance of "floating"-imagery to "Daffodils" and "Kubla
Khan", the following observation may help to integrate it into the ecstatic
experience suggested by these poems: "Few people appear to acknowledge the
role of vestibular sensations in the God Experience. However, in light of the
temporal lobe's role in the sensation of balance and movement, these experi
ences are expected. [...] Literature concerned with the God Experiences are full
of metaphors describing essential vestibular inputs. Sensations of 'being lifted',
'feeling light', or even 'spinning, like being intoxicated', are common"
(Persinger, 1987: 26). After quoting an account of such an experience, he ob
serves: "Note the repeated references to vestibular sensations: 'floating,'
'lifted,' 'moving,' 'spinning'" (27). I appealed to the same mechanism in my
discussion of the phrase "a most dizzy pain" in Keats's "On Seeing the Elgin
Marbles" (Tsur, 2002: 305) as well.
I submit, therefore, that readers who are familiar with vestibular sensations,
effortless movement (from dreams, perhaps), or emotional turbulence, may
respond quite appropriately to Coleridge's verse, even if not acquainted with
Plato, or Dionysus worship, or Orphic cults. Indeed, such familiarities may
perhaps account for the fact that the original Greek audiences understood
Plato, or Dionysus worship, or Orphic cults in the first place. Speaking for
myself, I was captivated by this poem several decades before I became
acquainted with any of them, floating
Ecstasy, Insight and the Rebirth Archetype
There is only one issue on which all the critics of this poem seem to be in
fundamental agreement: that there is a considerable break in the poem before
the last stanza. I shall not discuss the explanations offered by the various
critics, but only offer my own version, against the background of a miniature
controversy between Wilson Knight and Elizabeth Schneider (who believes
that "the division of Kubla Khan into its two parts also seems fatal to the
unity of the poem if it must be regarded as a complete whole").
Professor Wilson Knight has ingeniously compared the form of Kubla

Khan to that of an enlarged Petrarchan sonnet. Read thus, however, it can
only be an imperfect "sonnet", for the requirement of that or any other
two-part poetic form, that the sestet must throw some transforming light
upon the octave, is not met in Coleridge's poem (Schneider, 1975: 249-
250).
Here I again disagree with Schneider.8 First of all, in many "real" sonnets the
transforming light thrown upon the octet by the sestet is rather slight. Second,
and more important, I submit that the "sestet" of "Kubla Khan" does "throw
some transforming light" upon its "octave", and in a way that is far from
trivial, more significant in fact than the way it happens even in some indisput
ably fine examples of sonnets. In order to demonstrate this, I propose to look
briefly at Schneider's conception of the first part of the poem.
On the whole, not only do the first thirty-six lines refuse to sound as if
they had been dreamed; they sound more than anything else like a fine
opening for a romantic narrative poem of some magnitude. [...] The
historical Cubla was an attractive subject for such a poem (Schneider,
1975: 250).
I shall not follow Schneider's brilliant discussion of this issue in its details (I
quoted more of it in the preceding chapter, though in a different context). What
is important for us is that the first thirty-six lines of the poem contain
something that is very much "like a fine opening for a romantic narrative poem
of some magnitude" (although, "the texture is exceedingly rich and concentrated
for the opening of a long poem"; ibid, 252). Whether we accept the "romantic
narrative" theory or not, certainly 'Kubla Khan has, throughout, a perfectly
normal meaning, one that is logical and, as far as one can tell, as conscious as
that of most deliberately composed poems" (Schneider, 1975: 241). The first
part of the poem attempts to present something like a solid piece of "epic
8
Although, in the context of the preceding chapter, I cannot refrain, again,
from praising her general approach to literature. As I claimed there more than
once, one of the crippling effects of the Quest for Certitude on critics is their
inability to handle precisely this issue that the context "must throw some
transforming light" upon the elements that have entered the poem. This is one
more example where Schneider gives evidence of her Negative Capability even
where her actual critical decision appears to be doubtful.
reality", what may be characterised as "realistic plenitude" and the result of

first-hand observation. The first thirty-six lines of the poem "are factual,
detailed, matter-of-fact" (Watson, 1973: 228). Or, at least, as he recalled in
Chapter XIV of his Biographia Literaria, about his part in the plan for the
Lyrical Ballads, Coleridge's endeavours were here, too, directed "so as to
transfer [...] a semblance of truth" to the description, whatever its "supernat
ural or romantic" aspects.
Now, whatever the reasons for the abrupt discontinuation of this descrip
tion, the last stanza does "throw some transforming light" upon both the de
scription and its discontinuation. This "transforming light" is manifest in four
respects at least. First, Kubla's building enterprise is reinterpreted in the light
of the last stanza, in a significant way. In the first part we have a pleasure-
dome decreed by an oriental despot, with no further implications. The second
part changes this to a considerable extent. Many critics have commented that
"Kubla Khan" is about poetry. As I pointed out in the preceding chapter, this
is not exact, and poetry is too concrete a term. Thus, for instance,
"Kubla Khan", then, is not just about poetry: it is about two kinds of
poem. One of them is there in the first thirty-six lines of the poem; and
though the other is nowhere to be found, we are told what it would do to
the reader and what it would do to the poet (Watson, 1973: 227-228).
As I implied in the preceding chapter, "Kubla Khan" is neither about one

kind of poem, nor about two kinds of poem. The first part is about the build
ing of a pleasure-dome, the second part is about something much more abstract
than poems, such as creative inspiration, or visionary frenzy. What the second
part may do to the first part of the poem is to promote the relative weight of
the possible inspiration related to the architectural or, more generally, to the
artistic ingredient of the building enterprise.
Second, and far more important, the "factual, detailed, matter-of-fact" pre
sentation of Kubla's earthly paradise, giving to it the semblance of truth, is
experienced in the perspective of the second part as a kind of Paradise Lost.
The earthly paradise with its realistic plenitude and matter-of-fact details
becomes a fleeting vision, very much like a prenatal or other-wordly experience
that the speaker is attempting in vain to recapture. It is turned into an inac
cessible reality, referred to, but beyond, direct apprehension. "Kubla Khan" is
one of the few poems, or perhaps the only one in the English language, that
attempts to present a direct vision of ecstasy that may be an overpowering
emotion or exaltation, or the frenzy of poetic inspiration, or something like
mental transport or rapture from the contemplation of divine things. Such
poems are so rare because ecstatic experiences are, as I indicated at the begin
ning of the preceding chapter, ineffable by their very nature.
What we have here, then, is very much like a prototypical mystic situation.
The mystic is yearning after the experiencing of an inaccessible reality which,
some mystics seem to believe, they experienced in a different kind of existence.

This reality, with its paradise-like attributes, has been haunting them ever
since, and mysticism is the supposed instrument to recapture it. In Coleridge's
poem the reader is intensely involved in the description of Kubla's world; but
it is suddenly left off, leaving the reader with a sense of wondering or even
frustration. I would even venture to say that he is left with a yearning to
rediscover that world and go on experiencing it. Whatever the genetic reasons
for interrupting the description at this point (perhaps Coleridge was disturbed
by a neighbour, or was just unable to finish a large-scale epic opening of such
intensity), this feeling of wondering or frustration seems to be the aesthetic
effect of the abrupt ending. In the last stanza the reader joins the speaker of the
poem in his attempt to re-create this lost reality.
There seem to be three all-important ingredients in ecstatic experiences:
overpowering emotion, insight into an inaccessible but highly significant
reality, and some kind of dissolution of the perceiving or contemplating
consciousness. In "Kubla Khan" we have the first two.9 Thus, far from failing
to satisfy the requirement "that the sestet must throw some transforming light
upon the octave", Kubla's world is transformed from a world perceived in a
direct vision, into the object of "mystic" yearnings.
Third10, these mystic yearnings are reinforced from an unexpected direction.
There is a grammatical shift from the description of Kubla's garden to the last
stanza: from third to the first person (it → I). 11 Taken at its face value, the
shift is from a more objective to a more subjective mode of description. In the
first part, the first person is lurking only behind such exclamations as "But
Oh!", or "A savage place!"—all exclamations are necessarily in the first
person. In the perspective of a sublime nature description it may be perceived
as a rhetorical device of amplification rather than as the presence of a personal
consciousness. In the perspective of an ecstatic poem, however, the "object
ive" mode may turn out to be "depersonalisation"—after the event. In this
respect, too, the second part throws some transforming light upon the first
9
In fact, we seem to have in "Kubla Khan" the third ingredient as well; but here
the process of dissolution does not affect the perceiving or contemplating
consciousness, but rather the solid reality perceived or contemplated.
10
Parts of this paragraph are reproduced here from Tsur, 2002: 315.
11
In the same vein, in a recent essay on "Kubla Khan" William Benzon (2003)
pointed out that "there is a very specific linguistic difference between the two
movements in the use of pronouns [...]. Movement 1 (lines 1-36) has only
four pronouns while Movement 2 (lines 37-54) has sixteen. This suggests that
the cognitive structures sustaining self-presence, which are linguistically
realised through personal pronouns, are strongly operative in Movement 2,
but not in Movement 1". Later he adds: "The first movement of 'Kubla
Khan' is allocentric in that the narrative voice is aware only of Xanadu and
events happening there. [...] The second movement of the poem is egocentric
in that the narrative voice is self-aware".
part. Recent neuropsychological studies of altered states of consciousness may

illuminate such depersonalisation. "Selfhood then seems to have evolved along
lines suggesting at least in shorthand the operations of a kind of 'I-Me-Mine'
complex. But what happens when this egocentric triad briefly dissolves? Novel
states of consciousness emerge" (Austin, 2000: 209). Above I quoted Michael
A. Persinger (1987) who claims that God Experiences (as well as some
pathological conditions) are associated with temporal lobe transients, which are
electrical perturbations of the temporal lobe in the human brain (16). A
characteristic of such states "is an alteration in the description of the self.
Depersonalisation is typical" (Persinger, 1987: 18). In some of my recent work
I have demonstrated that in meditative and mystic poetry (Tsur, 2003: 115;
243) 1 2 and in a Keats sonnet displaying some altered state of consciousness
(Tsur, 2002: 305) 1 3 deliberate grammatic manipulations are performed so as to
generate an effect of depersonalisation. In "Kubla Khan" such grammatical
manipulations are less conspicuous: they consist in a passive construction and
a pronoun used as an impersonal subject before were (to be discussed in a
moment). Here, as I have suggested, straightforward third-person description is
transformed into depersonalisation after the event.
Thus, for instance, analysing the sixteenth-century anonymous Spanish

masterpiece "No me mueve, mi Dios, para quererte", I write on the first nine
lines: "The subjects of all the verbs in these lines are external, whereas the
first person occurs only as a direct object, as a passive object of action: '(not)
heaven moves me to love You [...], You have promised me [...], (not) hell
moves me to cease sinning'. Perhaps the most notable instance is a nominal
construction that would be hardly acceptable in English: 'el infierno tan
temido', that is, 'Hell so feared by me'. In the second quatrain the root 'see'
occurs twice; this mental process is attributed to 'I'. However, it occurs in an
exceptionally complex syntactic construction, the conspicuous purpose of
which is to 'depersonalise' seeing, turning the agent into a passive object:
'seeing You nailed to that cross and despised moves me'. The phrase 'seeing
You nailed to that cross and despised' contains no overt subject or finite verb,
but is itself the subject of the finite verb 'moves me'. Thus even 'seeing' is
not a volitional act; it just 'happens' to the speaker".
In Keats's sonnet "On Seeing the Elgin Marbles" the octet is dominated by
first person singular pronouns; they disappear completely in the sestet. Most
conspicuous are the impersonal constructs "glories of the brain" and "round
the heart", in stead of "glories of my brain", and "round my heart". Pain (in
line 11) too is a psychological abstraction which, again, seems to be unrelated
to any individual consciousness. These phrases serve as a transition from the
"I", the enduring, conscious element that knows experience to a less conscious
state; that is, they serve as transition from a state of individual consciousness
to an altered state of consciousness. In this state there is an awareness of a
stream of images, but none of the self as thinking, feeling, and willing, and
distinguishing itself from the selves of others and from objects of its thought.
In the previous chapter I criticised Yarlott, who points out, and rightly,
some impersonal syntactic structures at the beginning of the poem as "So
twice five miles ... were girdled round" and "there were gardens". From these
syntactic structures he jumps to the conclusion that the Khan's "relationship
with the slave force which, presumably, enacts his decree for him is utterly
impersonal" (Yarlott, 1967: 130). I claim that Yarlott replaces here a direct
perception of immediacy by an inference of a state of affairs outside the poem,
of which we know nothing. I propose, instead, that the impersonal passive
voice in line 7 can be perceived as an indication of the prompt execution of the
decree: the Khan decreed, and lo, the results are there. In the present context,
these impersonal constructions change their nature: they become part of the
mystic-ecstatic mental process re-enacted in the poem.
Fourth but not least, what was presented in an "even daylight" becomes the
first stage of what can be best termed an "emotive crescendo". One way to
express immense emotional experience in such a non-representational art as
music is to use a fortissimo. However, when the listener gets used to the
fortissimo its overwhelming power gets devalued, becoming "bombastic" rather
than "powerful". One way to overcome this problem is to use a crescendo, as
if the music said "This is so! ... This is more so ... This is even more so", and
so forth, until reaching a peak, creating a sequence of gradually increasing
intensities. In this way, music may "contain in itself the reason why it is so".
The fortissimo is not just the composer's whim: it is determined as the peak of
a pattern of increasing loudness.
The same is true, mutatis mutandis, in verbal expression. A poem that uses
many superlatives to express overpowering emotions tends to become bom
bastic rather than powerful. And the same kind of "emotive crescendo" is one
of the preferred ways to overcome the problem. In Coleridge's poem, the
"This is so! ... This is more so ... This is even more so" pattern is created by
the gradual dissolution of the solid physical world into thing-free and gestalt-
free qualities, as well as the increase of energy level. As I indicated earlier, such
a gradual dissolution may underlie the gradual increase of a poem's emotional
force. This gradual pattern may be reinforced by additional elements on the
semantic, thematic and metric levels. But one thing should be noted: when
isolated, not all episodes, or stages, of the pattern are perceived as "more and
more" emotional, in a way that the overall pattern can be inferred from them.
There is, rather, a sketchy indication of the pattern, and when the reader
reaches the "peak", he retro-relates it to the preceding stages and superimposes
graduality upon them.14 Architecture is the most solid of the arts, whereas
music is the thing-free art par excellence. In the Nietzschean dichotomy, archi-
14
I have taken the terms "retro-relating" and "superimposition" from Anton
Ehrnezweig who, in turn, derived them from William James. I have elsewhere
discussed (Tsur, 1977: 213-214; 1992a: 466-470) the cognitive mechanisms
underlying such "retro-relating" and "superimposition".
tecture is the most Apollonian of arts, whereas music is the most Dionysian.
"In music, the paroxysms of Dionysian ecstasy are subjected to the
Apollonian order and measure" (Beardsley, 1966: 276). In this sense, the
rebuilding of Kubla's building with music should be an extreme instance of
"reconcilement of opposite elements", reinstating the paradox of imagination in
one of its extreme manifestations. Such a conception of music (and it makes
little difference that Coleridge preceded Nietzsche by a few decades) would
explain why the speaker needed the Dionysian intoxication of the exotic girl's
music to achieve the trance required for the rebuilding of the dome. Unfortu
nately, however, that music too resides in a reality that is not accessible at will
("Could I revive within me/ Her symphony and song"), and thus only ampli
fies the speaker's yearning and frustration.
As we saw in the preceding chapter, Schneider (1975: 245-246) traces the
description of the last lines in the poem back to Plato's "comparison of poetic
inspiration with the frenzy of the orgiastic cults", in the Ion. "This conception
was old even in Plato's day, and practically every detail used by Coleridge was
a commonplace in it" (245). I guess that even a reader unacquainted with
Plato's (or Nietzsche's) account would recognise here the "flashing eyes" and
"floating hair" as the body language of a certain kind of state of mind. The rest
is indicated by the awe aroused in the audience. The "commonplace" nature of
the description too seems to be quite significant here. The ecstatic effect is
achieved not by the ingenuity of details, but by "common language heightened,
to any degree heightened, but not an obsolete one".
At this point it would appear desirable to relate the foregoing analysis to
Bodkin's discussion of the Death-and-Rebirth archetypal pattern, as I extrapo
lated it to "Kubla Khan" in the preceding chapter. So, I shall reproduce that
discussion here with minor omissions.
Within the image-sequences examined the pattern appears of a

movement, downward, or inward the earth's centre, or a cessation of
movement—a physical change which, as we urge a metaphor closer to the
impalpable forces of life and soul, appears also a transition toward severed
relation with the outer world, and, it may be, toward disintegration and
death. This element in the pattern is balanced by a movement upward and
outward—an expansion or outburst of activity, a transition toward
reintegration and life-renewal (Bodkin, 1963: 54).
One important distinction this passage makes is between archetypal con-

tents and archetypal patterns ("emotional symbolism" on the one hand, and
"capacity to enter into an emotional sequence" on the other). Oddly enough,
Bodkin does not attempt to show how this pattern applies in its details to the
description of running waters in "Kubla Khan" (mentioned above). The
"movement, downward, or inward the earth's centre, or a cessation of move
ment" is clearly indicated in such passages as

or

And sank in tumult to a lifeless ocean.
The "movement upward and outward—an expansion or outburst of

activity" is manifest in the passage
7. And from this chasm, with ceaseless turmoil seething,

Amid whose swift, half-intermitted burst
Or chaffy grain beneath the thresher's flail:
And 'mid these dancing rocks at once and ever
The fact that this third passage occurs in the poem between the first and
second quotations not only demonstrates the opposing movements, but also
creates an indication of rhythmic alterations, one of the main ingredients in the
Jungian conception of emotion underlying Bodkin's work. It should be
remarked here that the ecstatic experience as discussed above is perhaps the
most extreme instance of what Bodkin describes as "an expansion or outburst
of activity". Thus we may assume that the ecstatic experience heightens the
Death-and-Rebirth archetype to its extreme. This archetype, in Jungian theory,
is an endless succession of rising and falling emotional sequences. The
"emotive crescendo", then, may be regarded in some instances as a relatively
small section of the Death-and-Rebirth archetypal pattern. Coleridge's poem
appears to have a minor peak in midpoem, with the outburst of the "dancing
rocks", and a major peak at its end.
In the preceding chapter I criticised Fruman's (1972: 395-402) Freudian
interpretation of the poem. What I found least acceptable in his discussion was
his claim that the pleasure-dome suggested either the female breast, or the
mons veneris (or both). I argued there that this introduces foreign elements into
the poem. By contrast, consider now the following line: "As if this earth in
fast thick pants were breathing". Some critics regard it as a metonymy for
birth; others as a metonymy for sexual intercourse. The latter interpretation
could serve as an illuminating example of how the Freudian insistence on

"finding sexual symbolism underlying almost all human action, thought, and
dream" (Schneider, 1975: 9) can be utilised for a legitimate description and
interpretation of a poem. Plutchik (1968) put forward a theory—very useful
for literary critics—of emotion rooted in instinct. The following summarises
basic prototypic dimensions: incorporation, rejection, destruction, protection,
reproduction, deprivation, orientation, exploration. We are interested here in
reproduction, which he characterises as follows: "Apparently at almost all
animal levels, sexual behavior is associated with some form of pulsatile or or
gastic behavior. [...] Pleasure is presumably associated with all forms of sexual
behavior" (Plutchik, 1968: 73-74). Plutchik presents a multidimensional
structural model of the emotions. "It shows the eight prototypic dimensions
arranged somewhat like the section of half an orange, with the emotion terms
which designate each emotion at maximum intensity at the top" (ibid., 76). He
asked experimental subjects to rate the emotional words in terms of the
intensity that they represent, using a scale of 1 to 11. Though the unpleasant
emotions usually had longer lists of differentiable terms, the longest list of all
was, nevertheless, that of the reproduction dimension, including Ecstasy
(10.00), Joy (8.10), Happiness (7.10), Pleasure (5.70), Serenity (4.36), Calm
ness (3.30). The sexual behaviour of the landscape is indicated by the imagery
of the line quoted at the beginning of this paragraph; and "some form of
pulsatile or orgastic behavior" associated with it is suggested and amplified by
the obtrusive rhythms of the poem, enhancing the psycho-physiological
echoes in the reader's response to the state of ecstasy. More recently
Newberg, D'Aquili and Rause (2001: 42) discussed this issue in their book on
brain science and the biology of belief: "The resulting trancelike state is
experienced as an ecstatic rush of orgasmiclike energy. This state can be
induced by intense and prolonged contemplation, during rapid ritualistic
dancing, and sometimes, briefly, during sexual climax". Further, in a footnote
they comment: "Essentially any repetitive stimulation, whether it be physical,
emotional, sensory, or cognitive, can potentially generate such states" (184).
This, of course, includes both the rhythmic structure of the poem, and the
dancing rocks depicted in it.
Prosodic Structure
Before I proceed to discuss the prosodic structure and texture of this poem, I
wish to briefly present the theoretical framework within which I propose to do
this. Let us consider the opening lines of the poem:
8. In Xánadu did Kúbla Khán

w s w s w s w s
A státely pléasure-dóme decrée:

W S W S W S W S
Where Álph, the sácred river, rán
W S W S W S W S
Through cáverns méasureless to mán
W S W S W S W S
Dówn to a súnless séa.
w s w s w s
The sequences of w and s under the vowels signify the alternating weak and
strong positions that constitute the iambic metre. This is the metric pattern.
The accents above the vowels signify lexical stress, assigned to the most
strongly stressed syllable of lexical words, i.e., nouns, verbs, adjectives and
nonclitic adverbs. In lines 2-3 stressed syllables occur only, and in all, strong
positions (this may be taken as an indication of high regularity of metre: in the
first 150 lines of Milton's Paradise Lost there are only two such lines). In
lines 1 and 4 there is one s position occupied by an unstressed syllable. These
are the least noticeable but most common deviations in this poem. Line 5
begins with a stress displaced to the left ("inverted foot"), one out of no more
than four in the poem. Only a negligible number of stressed syllables in w
positions occur (which I shall mention in due course). Such deviations abound
in English poetry, but in this poem are very scarce.15
Most readers feel that the rhythms of "Kubla Khan" have something to do
with its ecstatic effect, and most critics devote a sentence or two to this issue.
It is thus all the more surprising that there is so little metrical analysis in the
literature, or specific discussions of the poem's rhythmic structure. One reason
for this seems to be that where metre is so relentlessly regular little is left for
the prosodic analyst to say. At any rate, if regular metre can underlie the
ecstatic quality of "Kubla Khan", the rationalist and witty quality of Pope's
"An Essay on Man", and the naive quality of nursery rhymes and some of
Blake's "Songs of Innocence", what can be said about regular metre that can
account for the ecstatic quality of a poem? The present work is the second in a
long series of publications in which I explore the conflicting potentials of
regular rhythms (see, e.g., Tsur, 1985a; 1992a: 431-454; 2003: 24-29, 167-
197). I am reproducing here the relevant sections of these discussions, with
minor alterations, expanding them where necessary.
The effect of verse with a tendency for metric regularity is "double-edged".
On the one hand, regular metre implies clear contrast between prominent and
non-prominent syllables. In this sense, regular metre has a strong rational
quality. It has good shape (strong gestalt), "it creates a psychological atmo
sphere of certainty, security, and patent purpose"; it exhibits definite direc-
15
I borrowed the terminology and the graphic signs from Halle and Keyser
(1971), which I utilised in a perception-oriented theory of metre (Tsur, 1977;
1998; etc.).
tions and organises percepts into predictable orders. On the other hand, the
vigorous impact of regular metre may be very much like the beat of a primitive
drum, that may have the effect of heightening emotional responsiveness that
underlies ecstatic ceremonies of tribesmen. In other words, regular metre shares
some important properties with conscious control and the exercise of will; at
the same time, it is similar to fundamental involuntary physiological processes,
many of which consist of regularly recurring events. Intense physical and
emotional activities in humans and animals increasingly tend to possess regular
rhythm and to transcend voluntary control. Consequently, one factor that
differentiates between regular metre underlying a witty poem and that underly
ing an ecstatic poem is the energy level inherent in other layers of the poem.
Another factor we find at the root of this double-edged nature is the term
security. As the research of E. Frenkel-Brunswick (1968) has shown, intoler
ance of ambiguity may interfere with one's free emotional responses. J.C.
Ransom suggested that a fairly predictable metre can dispel anxiety in the
presence of ambiguity—give "false security to the Platonic censor in us"
(quoted by Chatman 1965: 212)—so that the reader feel free to attend to
ambiguities in the other layers of the poem. The crucial question seems to be
whether the psychological atmosphere generated by "good metric shape" is of
genuine or false "certainty, security, patent purpose", etc. In other words, if
other layers of the poem also possess a rational quality, then the psychological
atmosphere is one of genuine certainty, etc. If, however, some other layers of
the poem induce an intense psychological atmosphere of uncertainty—as, for
example, the "unreal" vision of "Kubla Khan" or "The Rime of the Ancient
Mariner"—regular metre will impart "false security", lull the vigilant "Platonic
censor in us" and make it accept the emotional quality of the poem. By the
same token, and at the same time, vigorous rhythms have a strong bodily ap
peal, amplifying whatever irrational qualities there may be.
It has been observed that the rhythm of some poems is more obtrusive than
that of others; there is a small number of poems whose rhythm thrusts itself,
so to speak, upon the reader or listener. It will be noticed that at least two of
Coleridge's masterpieces in which he caught a glimpse of the uncanny—"Kubla
Khan" and "The Ancient Mariner"—are notorious for their strong prosodic
shapes and convergent rhythms. Thus they provide a remarkable illustration of
rhythm that gives ''false security to the Platonic censor in us", opening the
way for imagination to roam on less firm ground. Keats has some illuminating
things to say about the "Platonic censor" in Coleridge in his famous passage on
"Negative Capability":
I mean Negative Capability, that is, when man is capable of being in un
certainties, Mysteries, doubts, without any irritable reaching after fact &
reason—Coleridge, for instance, would let go by a fine isolated verisimili
tude caught from the Penetralium of mystery, from being incapable of re
maining content with half knowledge.
Keats himself wrote quite a few unique sonnets said to contain an ecstatic
experience, with highly divergent metre, the intense peak of which is domin
ated by thing-free and gestalt-free negative entity as "death", or "nothingness",
or "the shadow of a magnitude". If Keats is right about Coleridge's lack of
Negative Capability we should not be surprised that his version of the
uncanny or the ecstatic has very obtrusive rhythms to give "false security to
the Platonic censor in us", and are of the few poems of this kind that give the
imagination some positive entities to seize upon at the end.
Having given false security to the Platonic censor in us, the following
passage from "Kubla Khan" appeals, at the same time, also to the most
primitive layers of our personality by realising the drum-beating quality of
metre. From the prosodic point of view, not only the stressed syllables
converge here with strong positions to an unusual extent, but these prominent
points are further emphasised by alliteration, sharpening the contrast between
prominence and non-prominence. These alliterations have additional functions:
they enhance the balance of the line (as in lines 18 & 22); or, occurring
"intermittently" (as in line 19), they enhance the obtrusive feeling of regular
alternation. As for the contents of the lines, they "depict" vehement physical
motion. Thus, contents and metre mutually actualise each other's vigorous
potentials, making a notable contribution to an ecstatic quality where other
conditions are appropriate:
9. 17And from this chásm, with céaseless túrmoil séething,

18 As if this éarth in fást thick pánts were bréathing,
19A mighty fóuntain momently was fórced:
2o Amid whose swift, hálf-intermítted búrst
2 1 Húge frágments váulted like rebóunding háil,
22 Or cháffy gráin beneath the thrésher's fláil:
23 And 'mid these dáncing rócks at ónce and éver
24 lt flúng up mómently the sácred river.
In the whole poem there are just two sequences of three consecutive
stressed syllables. One of them happens to be "fást thick pánts". It is hardly
meant to "slow down" the rhythm of the poem; on the contrary. While the
reader is inclined to maintain his "impetuous" reading, the neutralised contrasts
add weight and energy. Thus, the stressed syllable thick squeezed in a w
position is perhaps an iconic reinforcement of its meaning. A similar iconic
squeezing may be the case in "Húge frágments". The underlying iambic ca
dence is, nevertheless, preserved in both instances, owing to what Chomsky
and Halle call the "Nuclear Stress Rule". Another, unique, metric deviation is in
10. swift, hálf-intermítted búrst.

s w s w s w s
It is, so to speak, a metric icon of its contents; two of its aspects are vividly
perceptual. Half in a w position loads the line with tension, entailing swift
succession of the next two unstressed syllables. The "compensating" stress is
phonologically equal or subordinated to the "infirming" stress, so that metric
regularity is, precisely, half-intermitted. Finally, after the rather long sequence
of lighter syllables, metre is powerfully reinstated on burst, which is, at the
same time, the headword of the syntactic group, to which all preceding stresses
are subordinated.
Though the rhythm of "Kubla Khan" is vigorous and regular, it cannot as
sume so strong a shape as in, for instance, Tennyson's In Memoriam16 or
Pope's An Essay on Man. In both these poems there is a strong metric shape,
strong shape of lines and strong shape of stanza, all predictable to a large
extent. In "Kubla Khan" neither the length nor the grouping of lines is
predictable. Therefore, the psychological atmosphere of certainty associated
with the underlying strong metric shape comes up against an atmosphere of
uncertainty generated by the larger groups. The first five lines of the poem, for
instance, approximate two symmetrical structures of quatrains. Suppose the
poem began:
11. In Xanadu did Kubla Khan

A stately pleasure-dome decree:
Down to a deep and sunless sea.
If the reader can sufficiently overcome his familiarity with Coleridge's actual
rhythms, he will realise the following: first, the strong, symmetrical shape of
this transcript is softened in the actual poem by a shortening of the last line.
Secondly, the "interpolation" of a third a-rhyming line distorts this symmetry,
prolonging the expectations for a b-rhyming line which, "gratifying" as it is,
comes, at last, in a foreshortened form. The second strong shape which the
opening lines approximate is precisely Tennyson's a-b-b-a quatrain. Imagine
something like:
12. A stately pleasure-dome decree:

Down to a deep and sunless sea.
Even so, a highly resistant enjambment beginning in the last position of line
2 (blurring the symmetrical division of the stanza into two halves that are
See my discussion of this poem in this respect (Tsur, 2003: 172-175; 177;
186).
mirror-images of each other) is hardly like In Memoriam. Thus, the odd line is
not just another line in the group; it makes the shape of the whole "hope
lessly" ambiguous (the same holds true, mutatis mutandis, of the five-line-
section beginning with "But oh!").
As to the lines 17-24 (quoted above), they are grouped according to two
diverging grouping principles. In one respect they are simple couplets, grouped
by rhyme. Syntactically, however, a second pattern is superimposed. The line
"A mighty fountain momently was forced" (19) is grouped, rhyme-wise, with
the next line. Syntactically, however, it is grouped with line 17. From the
preposition from at the beginning of line 17 a verb is predicted; this prediction
is fulfilled only at the end of line 19, running on to the next couplet. Thus, the
interpolated simile "As if this earth in fast thick pants were breathing" not
only adds the figurative-mythological dimension, but also weakens the percep
tual shape of the whole passage by delaying the fulfilment of syntactic predic
tions and by upsetting, for a considerable stretch of lines, the convergence of
sentences and couplets. The next clause, two lines long, is, again, "straddled"
between two couplets. When it ends, in mid-couplet, another line is needed to
complete it; consequently, an "extra" simile is introduced after the fulfilment of
syntactic predictions. Only the last couplet of the passage—the "summary" of
the description—entirely converges with the couplet. Therefore, perceptually,
too, it has a"rounding-off"effect.
We find a similar structure of regularity and unpredictability on the level of
alliterative patterns. In an illuminating paper comparing the metrical styles of
Donne's Satyres and Pope's "reversification" of them, Chatman (1960) con
trasts their uses of alliterative patterns in two respects. In one respect,
Donne has (proportionately) twice as many occurrences of alliteration of

two syllables in immediate sequence as Pope, whereas Pope has almost
four times as many occurrences of an intervening unstress. This means
that Pope wants alliteration to cooperate with metre, not oppose it
(Chatman, 1960: 156).
In another respect, Pope's
alliteration of epithet-noun combinations is characteristic and carefully

done. [...] Donne, on the other hand, frequently alliterates words that have
little structural connection (giving the illusion of mere chance collocation)
(Chatman, 1960: 157).
The two poets created in different lines of wit. Now in "Kubla Khan",
which is a highly emotional poem, and perhaps less witty than any other
romantic poem, we may encounter both kinds of alliteration. Consider Kubla
Khan, measureless to man, sunless sea, woman wailing, mingled measure; or
in a more complex version, ceaseless turmoil seething. At the same time we
have got such alliterations as Xanadu did, dome decree, etc. The sequence river
ran is essentially of the first kind (though I am not sure that Chatman would
agree with me). At the same time, it is preceded by sacred and, in the first
occurrence of the sequence, the intervening comma and the beginning of the
run-on construction render it as near to "mere chance collocation" as can be.
Mazy motion, again, belongs doubtless to the first type of alliteration; but in
the same line we have two more tokens of m, miles meandering, that form an
alliterative pattern of the second kind, which also effects the character of the
other pattern. Girdled round appears to be an alliteration of the second type;
but it is also part of another kind of sound pattern. The first phoneme of
Girdled + round form a "perfect" homonym with ground. This pattern further
bifurcates to greenery. Thus, in fact, the alliterative patterns enhance the ten
dency of the metric organisation to give "false security to the Platonic censor
in us" by reinforcing both the psychological atmosphere of predictability and
of unpredictability in the poem. Likewise, Schneider's conception of prosodic
devices "partly concealed by the interlacing of other patterns" (discussed in
Chapter 1) gains its significance from the present conception of giving "false
security to the Platonic censor in us".
Three additional elements may affect the enhancement or weakening of
sound patterns: phrase structure, enjambment, and wider gestalt. Consider the
second occurrence of the words "river ran". Gestalt theory claims that strong
gestalts maximise our tendency to perceive a set of stimuli as a coherent whole.
The alliteration increases our tendency to group together the words of the
phrase. And conversely, the coherent phrase renders the alliterative pattern
more cohesive. Consequently, such phrases form an exceptionally strong
gestalt, "settling quite heavily upon the mind", or closing the line "with a
click". In their first occurrence ("river, ran"), by contrast, the two words
belong to different syntactic units, interfering with the cohesion of the repeated
sound pattern. Moreover, the second syntactic unit is run-on from one line to
another, effectively backgrounding the "clicking" rhyme. One basic assumption
of gestalt theory is that elements may change their nature when entering into a
wider context. The alliteration "mazy motion" is one instance which may
strongly reinforce the gestalt of the rhythm, according to the foregoing discus
sion. Viewed, however, in the context of the whole line, in one possible mental
performance at least, it becomes part of a random, weak gestalt: "Five miles
meandering with a mazy motion" (As will be demonstrated in Chapter 3, there
is experimental evidence for different mental performances in this respect). So,
I do agree with Schneider that certain wider patterns may soften the impact of
alliterations and rhymes, but the emerging picture is far more complex even
than what she suggests.
A word must be said about the structure of the last stanza:
13. 37 A damsel with a dulcimer x

38 ln a vision once I saw: x
The Texture and Structure of "Kubla Khan' 107
39 Ïtwas an Abyssinian maid, a

4o And on her dulcimer she played, a
41 Singing of Mount Abora. x
42 Could I revive within me b
43 Her symphony and song, c
44 T0 such a deep delight 'twould win me, b
45 That with music loud and long, c
46I would build that dome in air, d
47 That sunny dome! those caves of ice! e
48 And all who heard should see them there, d
49 And all should cry, Beware! Beware! d
50 His flashing eyes, his floating hair! d
51 Weave a circle round him thrice, e
52 And close your eyes with holy dread, f
53For he on honey-dew hath fed, f
54And drunk the milk of Paradise. e
As the graphic arrangement too may indicate, the stanza is made up of lines
of unequal length, and I shall not enumerate here the various possibilities. Lines
42-47 contain a single compound sentence that stretches over lines of several
structures. One possible purpose of this structure may be to shake the reader's
certainty, so as to render the certainty of regular metre in the ensuing lines
utterly false. The phrases in line 47 display a fundamental uncertainty in their
syntactic relationship to the preceding lines. Are they exclamations, or appos-
itive phrases to "that dome in air"? A similar ambiguity is displayed in lines
49-50. The noun phrases in line 50 could be the direct objects of the repeated
imperative verb in line 49 (as, for instance, in "Beware the ides of March").
However, the exclamation marks separate them from the phrases, turning them
into ostensive exclamations, as if the onlookers pointed at the speaker and ex
claimed with horror: "His flashing eyes, his floating hair!" Ostension is
associated with the right (emotional) hemisphere of the brain:
The inactivation of the right hemisphere leads to a deficit in ostensive

communication. In semiotic literature [...] this way of communication
might be defined as "placing something at the disposal of the cognitive
activity of a person". [...] Ostension merges with synecdoche: the voice of
a patient's wife, which he hears without seeing her, is her pars pro toto
(Jakobson, 1980: 27).
The speaker's flashing eyes and floating hair may be just such highly
significant, emotionally loaded synecdoches; and the potentially fluent
syntactic structure disintegrates into a series of just such ostensive warnings.
In line 47, too, the appositive phrases are transformed into ostensive phrases.
If Jakobson is right in relating ostension to the right hemisphere of the brain,
these phrases ought to have a more than usually direct emotive appeal,
reinforced by the emotionally loaded situation.
In the above quotation of the last stanza I indicated the rhyme-scheme in the
column on the right, x indicates unrhymed lines, or perhaps off-off-rhymes.
The discussions among the critics whether they can be regarded as rhymes or
at least off-rhymes (cf. "I do not suppose Coleridge thought of dulcimer as
even an off-rhyme to once I saw or Abora"—Schneider, 1975: 273) strongly
suggest that they introduce into the poem an element of uncertainty. Even
where the rhyme-scheme appears to organise the lines into a group of strong
shapes, an element of uncertainty enters. Thus, for instance, in line 51 a new
syntactic unit begins (metrically emphasised by the first w position being left
unoccupied). The last four lines are grouped by an e-f-f-e rhyme-scheme into a
symmetrical and closed quatrain ("thrice—dread—fed—Paradise"). This clear-
cut structure, however, is preceded by an e-rhyme in line 47, just enough to
make the reader doubt his own perceptual organisation. In other words, the
appearance of "caves of ice" four lines earlier bestows randomness upon what
otherwise would yield a symmetrical and closed shape. Likewise, lines 46, 48-
50 end in ¿/-rhymes. In this case too much grouping becomes no grouping at all:
there are four similar-ended lines, lumped together in random order. Thus, at
the peak of the emotional experience indicated by the poem, there is an intense
web of rhymes that on a lower level amplify the principle of rhythmic
recurrences so as to heighten emotional responsiveness; viewed from a higher
point of view, they are characterised by a considerable degree of uncertainty.
Thus, again, the certainty given to the Platonic censor in us turns out to be
rather doubtful. This interpretation of ddd rhymes is confirmed by our
empirical research (see Chapter 3).
There may nevertheless be here something of the kind described by Ehren
zweig as the secondary elaboration of a pattern superimposed upon the last
stanza, "retro-related" from the last four lines. There appears to be a pattern of
gradually increasing order and distinct shapes, beginning with a stage where
there is considerable uncertainty whether some lines are rhymed; then a series
of rhymed lines comes forth in which the order is unpredictable and the rhyme-
pattern (though intense) is indistinct. From this there emerges a symmetrical,
firmly closed quatrain, constituting a strong structural closure at the peak of
the emotional experience. But even this strong closure is relative: it is weak
ened by an antecedent e-rhyme in line 47. That is why one may say that this
last "quatrain" emerges from a jumble of randomly rhymed lines.
Schneider (1975: 273) suggests that Coleridge's "remark upon Milton's use
of an unrhymed line at the beginning of a verse paragraph might equally well
have referred to the first line of his own final paragraph, 'A damsel with a
dulcimer'".
Discussing a not very important question, whether Milton intended the

hill-rill couplet of lines 22-23 to close one paragraph in Lycidas or to
open the next, Coleridge argued for the first choice on grounds that, he
thought, must be "for a poet's ear convincing". The eighth line of the
preceding paragraph ("And bid fair peace be to my sable shrowd"), like the
first ("Begin then, Sisters of the sacred well"), "is rhymeless, and was left
so, because the concurring rhymes of the concluding distich were foreseen
as the compensations". In other words, Coleridge was arguing that Milton
had deliberately opened a paragraph with an unrhymed line but would not
have closed it so (Schneider, 1975: 272).
I believe that Coleridge was suggesting in these words more than that. This
idea of "compensation" implies a mutual dependence of a weakened beginning
and a strong ending: the weaker the prosodic organisation at the beginning, the
stronger the ending appears. And conversely, the stronger the ending, the more
justified, or functional, or acceptable, the weak beginning appears. In fact,
Coleridge is arguing here for the gestalt principle that the shape of a part must
be modified in order to make it dependent on the whole. If I am right, in
"Kubla Khan" the answer may be the "superimposed pattern" offered above.
Summary and Conclusions
The English language tolerates, for reasons that cannot be discussed here, more
metric deviation than most other European languages. The more remarkable is
the fact that the metric organisation of "Kubla Khan" is highly regular at the
foot-rank, perhaps the most regular among the major English poems. Neverthe
less, it does not sound like a nursery rhyme, or like the witty poetry of Pope. I
have suggested that predictable metre is "triple-edged": on the one hand it may
generate a psychological atmosphere of a simplified mastery of reality charac
teristic of naive attitudes, or of certainty, security and patent purpose of
rational activities; on the other hand it may give "false security to the Platonic
censor in us" and admit the irruption of the irrational, such as ecstatic states of
mind. Ecstatic states of mind occur at high levels of mental energy. In ecstatic
poetry we sometimes find that regular metre on the prosodic level is coupled
with rhythmical movement (e.g. dancing) of exceptional intensity on the
thematic level.17 The description of the dancing rocks flung up by the mighty
fountain in "Kubla Khan" is a typical case in point.
Whether security is "genuine" as in nursery rhymes or witty poetry, or
"false" as in ecstatic poetry, depends on whether in other layers of the poem
17
In my 1985 paper I pointed out the co-presence and mutual foregrounding of
three elements in Wordsworth's "Daffodils" (a poem frequently characterised
as ecstatic): more than usually regular metre, more than usually intense
dancing movement, and the pattern of emotive crescendo. In my later works I
further developed this notion with reference to this poem (Tsur, 1992a: 447-
450; 2003: 24-29).
certainty or uncertainty are engendered. [±CERTAINTY] may be engendered

either on the rank of prosodic superordinates, or on the various semantic and
thematic levels. On the prosodic level of "Kubla Khan" we found these
conflicting qualities in both metric and alliterative patterns. Whereas metre is
highly regular and predictable throughout this poem, the length of lines and the
order of their rhymes is hardly predictable and rather errant. Sometimes no
recognisable stanza shape emerges from the order of rhymes, sometimes they
approximate two kinds of symmetrical stanza forms, but remain highly
ambiguous. At the very end of the poem, at the peak of the suggested ecstatic
experience, a symmetrical closed stanza emerges from the jumble of randomly
rhyming lines, sealing the poem with an emphatic structural closure. I empha
sise structural because in the scene described there are no closural elements.
Although it is hard to imagine where the gradually heightened spiritual activity
could proceed to, there is nothing here to indicate a natural "cut-off' point,
such as Herrnstein-Smith (1968) found in poetic closure. It is this con
tinuously reverberating ecstatic quality at its peak that is forcefully reinforced
by the closural quality of the final "quatrain". Notice, however, that even in
this strong, symmetrical, closed rhyme pattern, closure is sabotaged to some
extent: the last word of the poem, Páradise, leaves the last strong position
occupied by an unstressed syllable, so that the last line remains insufficiently
closed.
I quoted Chatman who compared the metrical styles of Donne's Satyres and
Pope's "reversification" of them. In the latter's style he found a majority of
alliterative patterns that are focussed, and corroborate metrical regularity. In
the former's style he found a majority of alliterative patterns that go, typi
cally, against metre, and make an impression of diffusion and of chance
collocation. In "Kubla Khan" the majority of alliterations are like Pope's; but
there are also some of the other type. In the present framework of accounting
for ecstatic quality, alliteration contributes both to the psychological atmo
sphere of security and to that of uncertainty. By the same token, and at the
same time, the regularly recurring prosodic events heighten the emotional
responsiveness of readers.
In the preceding chapter I discussed at some length Snyder's stimulating
little book on hypnotic poetry. One of his generalisations was "that in the
early stages of a hypnotic poem a foreign word, an obscure phrase, or any
slight difficulty that causes fatigue from strain on the part of the listener may
actually promote the ultimate aesthetic effect at which the artist aims". This is,
perhaps, to strain the analogy to hypnosis too far, and the "meaningless"
words and phrases are only meant to achieve an "incantatory effect". Such an
incantatory effect may be achieved in the first line of our poem, "In Xanadu
did Kubla Khan", where only in and did are plain English words.
Among other relevant passages, I quoted the following two:
Some hypnotic poems stop here: the listener is lulled by patterns of sound,
his attention is fixed without arousing of his mental faculty, and he falls
into whatever mood the poet "suggests". It is interesting to see how many
poets are thus content to stop without taking full advantage of the grip
they get on the listener's emotions. Such skillful artists as Poe, Swinburne,
the youthful Tennyson, and countless others persistently fail, or refuse to
galvanize the sensitive reader to action, determination, or even thought
(Snyder, 1930: 47).
Some hypnotic poems, however, do "carry the parallel to hypnotism still

further by 'suggesting' an impulse to action, making a parallel to the specific
post-hypnotic suggestions" (Snyder, 1930: 47-48). In such instances,
in a hypnotic poem the key sentence "suggesting" an idea comes near the
end, or at least only after there has been a long preliminary soothing of
the listener's senses by monotonous rhythmic "passes". So in hypnosis.
Also this key sentence "suggesting" an idea carries conviction without
argumentative support, or with only the simplest of obvious arguments to
support it. In the non-hypnotic poem these conditions do not obtain
(Snyder, 1930: 48).
It is, then, the monotonous rhythms that heighten the reader's emotional
sensitivity which constitute, in a way, the raison d'être of "hypnotic" or
ecstatic poems. They "fail, or refuse to galvanize the sensitive reader to action,
determination, or even thought". Or, if they do suggest "an impulse to action",
or contain near the end a key sentence "suggesting" an idea, "this key sentence
'suggesting' an idea carries conviction without argumentative support, or with
only the simplest of obvious arguments to support it". 18 Thus, we should not
be surprised if an ecstatic poem carried no "message" at all, or if the "message"
it carried were a mere "simple idea". It is the intensity of experience rather than
the "idea" that counts.
This brings us to the semantic and thematic aspects of the poem that may
arouse uncertainties (so as to render the security given by metre to the Platonic
censor false). On the one hand, we discussed at length in the present chapter
18
I assume that one obvious case in point would be
Farewell, farewell, but this I tell

To thee, thou Wedding-Guest!
He prayeth well, who loveth well
Both man and bird and beast.
He prayeth best, who loveth best

All things both great and small;
For the dear God who loveth us,
He made and loveth all.
the thing-free qualities, negative entities, and irrational visions that constitute
this poem. All these enhance the psychological atmosphere of uncertainty. On
the other hand, I already mentioned above how the direct pointing (ostension)
at the speaker's "flashing eyes" and "floating hair" (synecdoches) may be
particularly loaded with emotion. I quoted Jakobson's semiotic definition of
ostension: "placing something at the disposal of the cognitive activity of a
person". In the preceding chapter I considered at some length the behaviour of
some critics when encountering a piece of unevaluated poetic information. This
too appears to be, in a somewhat different sense, the placing of something "at
the disposal of the cognitive activity of a person", without suggesting "an
impulse to action" or to an attitude. Some critics are reluctant to contemplate
such unevaluated things. Thus, for instance, I quoted Yarlott (1967: 134-135),
who compares Coleridge's description of Kubla's garden to Purchas' de
scription, Coleridge's source. Yarlott points out that Coleridge substituted the
adjectives "bright/sinuous" for "pleasant/ delightful" in Purchas' description.
Yarlott rightly realises that "Coleridge seems to have deliberately modified the
attractiveness implicit in Purchas's original description". But, instead of realis
ing that Coleridge eliminated the evaluative ingredient of the adjectives while
retaining some of their descriptive contents, Yarlott claims that Coleridge
produces sinister associations. He is reluctant to accept the zero-grade eval
uation merely placed at the disposal of his cognitive activity. He wants to
achieve a greater degree of certainty: to know whether the thing placed at the
disposal of his cognitive activity is beneficial or harmful. This short excursus
on this essentially secondary issue brings into the foreground a more central
one: there is a fundamental uncertain quality about the description of the site
of Kubla's building enterprise. What is missing here for a greater degree of
certainty is the suggestion of "an impulse to action" or to an attitude or, at
least, to some evaluation. This uncertainty is an additional ingredient that
enhances the false security given to the Platonic censor by the exceptionally
regular metre of this poem.
This appears to be quite essential for the hypnotic-ecstatic quality of the
poem. When I wrote the foregoing discussion, I was not yet familiar with
Tellegen's (1981) work on experiential and instrumental set. This distinction
will be introduced in Chapter 3. The relationship he pointed out between this
dichotomy and hypnotisability strongly supports the aesthetic position
advocated here. "By experiential set is meant a state of receptivity or openness
to undergo whatever experiential events, sensory or imaginai, that may occur,
with tendency to dwell on, rather than go beyond, the experiences themselves
and the objects they represent". This is, precisely, the point of my argument
against Yarlott, for instance, who notes that Coleridge substituted the adject
ives "bright/sinuous" for "pleasant/delightful" in Purchas' description, and
insists that by this substitution Coleridge produces sinister associations. I
claim, by contrast, that by this substitution Coleridge eliminated all evaluative
ingredients of the adjectives, placing the scene described at the disposal of the
cognitive activity of the reader, to dwell on, rather than to go beyond it to dis
cover whether it is beneficial or harmful Indeed, in our experiment we found
that readers disposed to adopt an experiential set are inclined to judge trance-
inductive poetry more favourably than those who are not (see Chapter 3).
I have dwelt on "Kubla Khan" at considerable length and in considerable
detail. Some readers may accuse me of trying to "squeeze" the poem dry. At
any rate, my reading was not longer than the majority of studies devoted to
this poem. The main difference between us appears to be the space devoted to
the elements inside the poem relative to the space devoted to the elements
outside. The thing to account for seems to be, in many studies, whether
explicitly stated or not, the peculiar emotional quality of the poem. Some
critics seem to believe that if you can show that the poem does not mean what
it means—that is, if you can show that the poem has got some "symbolic"
meaning—you can account for this peculiar quality. Others seem to believe
that if you can relate the poem to a great number of myths and primeval lore,
you can account for the peculiar quality of the poem. What appears to be
common to both approaches is that they both attempt to reduce the poem to
something outside it. The present approach opposes these tendencies in two
important respects. First, it attempts to account for as much as possible in
terms of internal elements and their various aspects. It assumes that much of
the external information loses its emotional significance once it is torn out from
its original context. Second, it attempts to account for the evasive emotional
qualities of the poem not by reducing them to something, but rather by
pointing at a complex interplay between a multiplicity of elements and aspects
inside the poem. Hence, myth or any other external information can help to
account for the emotional impact of a poem only if it helps to impute unity and
coherence to a work otherwise puzzling or defective in this respect (cf.
Margolis, 1962: 114).19 As a further step, it attempts to account for the
peculiar emotional (ecstatic) quality of the poem by pointing at structural re
semblances between the processes within the poem, and the particular kinds of
emotional processes. This is why it cannot ignore Maud Bodkin's conception
of this poem in terms of archetypal patterns.
By this I do not wish to imply that all symbolic or external criticism is
necessarily wrong or harmful, though I believe that much of it is not as
illuminating as it could be, or ought to be. What I do strongly imply is that one
cannot know what external meanings and sources are relevant to a poem before
one knows what is its internal structure, and how the internal elements are
organised by it. Though I don't pretend that the above reading is the interpret
ation of this poem, it is, certainly, a rather plausible one. The interpretation of
19
"We know a 'myth' to be 'objective' for criticism, though it may not be so
for science, when the habits of thought and perception and imagination of
normal persons are educable in its terms and when their responses to
appropriate stimuli are generally predictable" (Margolis, 1962: 113).
metaphors, says George Miller (1979: 241), "is not a search for a unique
paraphrase of the implicit comparison, but rather a search for grounds that will
constrain the basis of the comparison to a plausible set of alternatives" (my
italics). This seems to be perfectly true, with the necessary changes, of the
interpretation of whole poems as well. I claim that one cannot find grounds
that will constrain the basis of symbolic or other external meanings to a
plausible set of alternatives before carrying out some thorough and close
textual reading guided by reasoned principles—of the kind performed in the
course of the present chapter. One could carry this conception one step
further, though I shall not attempt it here. One might use the above reading, or
some equivalent of it, as a filter (in Max Black's sense) to filter out irrelevant
information from the mass of information amassed by scholars and critics as
possible sources and meanings of "Kubla Khan", and see what remains of it.
Some of the symbolic readings will still be valid and illuminating; others will
certainly undergo considerable modifications. But very many will not be ad
mitted at all. Let us adapt Black's filter-model of metaphor to the issue in our
hand.
Suppose I look at the night sky through a piece of heavily smoked glass on
which certain lines have been left clear. Then I shall see only the stars
that can be made to lie on the lines previously prepared upon the screen,
and the stars I do see will be seen as organised by the screen's structure
(Black, 1962: 230).
We think of the internal structure of the poem as such a screen, and the
system of crisscrossing small-scale and large-scale relationships among its
elements as the network of lines upon the screen. But perhaps something quite
unexpected may also happen then: when a critic has worked out an elaborate
internal structure for a poem like "Kubla Khan", he may find that his eagerness
to adduce external information has been drastically reduced.
Gestalt Qualities in Poetry
and the Reader's Absorption Style
At the outset1 I would like to locate our attempt within reader-response criti
cism in two respects. First, most such criticism is concerned with interpretive
strategies on the semantic and thematic level, whereas in what follows we make
an attempt to focus attention upon some perceived effects of prosodic
organisation. Second, our attempt may be located between two poles of critical
approaches. At one end, Gestalt qualities are frequently treated in art criticism
as properties of the stimulus-design; the agency of a human perceiver is only
tacitly assumed. In literary criticism "the reader", or "the qualified reader", or
"the competent reader" is frequently mentioned as the human context of the
text. In much such criticism, "qualified reader" means "a hypothetical reader
whose response conforms with my analysis". At the other end we find the
alternative approach, which is to induce real readers to produce protocols of
their responses; but in such studies there are frequently no clear criteria to tell
what aspects of the responses reflect the reader's structural knowledge, and
what aspects reflect personality variables that are incidental to poetic
structure, or some other accidental factor. In our study we have performed a
detailed theoretical analysis of the rhyme-structures of our poetic texts, within
a Gestalt-theoretical framework, and then submitted them to flesh-and-blood
highly-qualified readers, whose absorption-styles had been assessed on
Tellegen's absorption scale.
This chapter will investigate relatively simple instances of the relationship
between rhyme patterns and their perceived regional qualities by manipulating
the rhyme scheme of a short, four-line stanza. One of its basic assumptions is
that the realisation of such poetic effects requires certain kinds of poetic struc
tures and, at the same time, certain kinds of cooperation on the reader's part.
Another assumption will be that, in some cases at least, different kinds of
cooperation can be traced to different cognitive styles.
This chapter is based on research reported elsewhere (Tsur, Glicksohn, and

Goodblatt, 1990; Glicksohn, Tsur and Goodblatt, 1991; Tsur, Glicksohn, and
Goodblatt, 1991).
Rhyme Patterns and Stanza Structure
In the first part of the present chapter we are going to discuss the phonetic
aspects of linguistic signs, and their grouping into perceptual structures by
means of rhyming. There arises the question of what causes the phonetic as
pects of those linguistic signs to be perceived as coherent entities that possess
distinct shape and unity. It is Gestalt psychology that has systematically
investigated the possible answers to that question. "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). In a footnote, Herrnstein-Smith adds:
The "laws of organisation", as formulated by Max Wertheimer, designate

the conditions which maximise our tendency to respond to groups of indi
vidual stimuli as unified "percepts". These conditions include proximity
and similarity (1968: 41).
The couplet, for instance, tends to be perceived as a percept that has

considerable unity, owing to both the similarity of the rhyming line-endings,
and their proximity to each other. The quatrain with an alternating rhyme-
scheme (abab) contains two pairs of rhymes. The members of each pair are
similar, but are in less proximity to one another than in the couplet. In the
present section of our paper we shall explore some of the Gestalt principles
that seem to govern the variants of the quatrain. These principles have been
applied by Leonard B. Meyer and Rudolf Arnheim to music and the visual
arts, respectively. Meyer speaks of the "Law of Good Continuation", the
"Law of Return", and of "Saturation":
A shape or pattern will, other things being equal, tend to be continued in

its initial mode of operation. Thus, "to the factor of good continuation in
purely spatial organisation there corresponds the factor of the smooth
curve of motion and continuous velocity in spatio-temporal organisation".
[...] Actually, of course, a line or motion does not perpetuate itself. It is
only a series of lifeless stimuli. What happens is that the perception of a
line or motion initiates a mental process, and it is this mental process
which, following the mental line of least resistance, tends to be perpetu
ated and continued (Meyer, 1956: 92).
The "Law of Return", on the other hand, is "the law that, other things being
equal, it is better to return to any starting point whatsoever than not to return"
(1956: 151). The "Law of Good Continuation" seems to have precedence, and
only when its application fails, the law of return is applied. In other words, the
Law of Return is the marked one of the two. These two laws impose different
characteristics upon different strophic organisations, each of which will be
Gestalt Qualities and the Reader's Absorption Style 117
considered as a "good" perceptual organisation under the relevant law. Let us

take as an example one of Omar Khayyám's Rubáiyáts, in Edward Fitzgerald's
famous English version:2
1. Think, in this battered Caravanserai

Whose Portals are alternate Night and Day,
How Sultán after Sultán with his Pomp
Abode his destined Hour, and went his way.
Let us try to distort the rhyme-scheme of this quatrain, as follows:

How Sultán after Sultán did sojourn,
And went his way then — never to return.
Two different principles organise Omar Khayyam's Rubáiyát and our

transcription of it. The original Rubáiyát actualises the "Law of Return". The
third line deviates from the rhyme established in the first two lines; the fourth
line returns to it. The transcribed version actualises the "Law of Good
Continuation". The first two lines of this quatrain constitute a "strong" shape:
they are connected by a single rhyme, and constitute a symmetrical couplet
that may be described by a single structural principle (the second line rhymes
with the first one). The couplet-pattern "perpetuates" itself, it recurs in the
next two lines as well; and had the poem contained ten, or twenty, or one
hundred lines, the same "Good Continuation" could have gone on indefinitely,
as, indeed, happens in the poetry of Neo-Classicists such as Alexander Pope,
or in Racine's plays. But notice this: the b-rhyme of the third line is perceived
in the two versions very differently: in the first it is perceived as a deviation or
an intrusion, whereas in the second as part of a pattern that repeats the initial
couplet pattern. This difference is more poignant in the Hebrew texts, where
the third line is letter-by-letter identical in the two versions. According to the
"Law of Good Continuation" a couplet to be completed is expected after the
third line. In the isolated stanza, the Law of Return of the aaba rhyme-scheme
comes as a surprise, and is perceptually justified only after the event. It should
2
In our empirical tests some groups of subjects responded to this English text,
some to its Hebrew translation by Jabotinsky. Though this was a matter of
convenience rather than deliberate planning, it provided convincing evidence
that we were dealing with genuine Gestalt organisation and not with some
language-specific accidents. Thus, for instance, the possibility that some of our
English respondents treated "Caravanserai" and "Day" as rhymes for the ear
and some as rhymes merely for the eye did not affect our experimental results
as far as language differences were concerned.
be noted as well that in the original version the return is to a specific rhyme,
whereas in the transcribed version an abstract pattern is repeated; the second
couplet is based on a different specific rhyme.
It is also instructive to inspect the two versions from the point of view of
unity. The Law of Return in the original version generates a tightly-closed and
coherent unit. There is a feeling that the quatrain constitutes a single unit that
is closed with a sharp "click". In the transcribed version, the quatrain tends to
fall into two symmetrical sub-units, two couplets; and if it can be said to close
with a "click", it is each one of the two couplets that closes with such a
"click".
The rule that governs the process is evident. The effect depends on the
degree of simplicity of the whole as compared with the degree of simpli
city 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 indepen
dent entities (Arnheim, 1967: 61).
On the whole, however, the simplicity of any part must be modified or

weakened sufficiently to make the part dependent on, and therefore inte
grated with, its context (Arnheim, 1967: 65).
The simplicity of the couplets in the transcribed version is considerable;

what is more, they constitute a larger unit, divided into two parts with
identical structures; as a result, they clearly stand out in perception as two
semi-independent units. This is not so in the original version. Here, the
smallest unit that may recur with regularity is the whole stanza. The third line,
which is not part of any rhyme, is perceived in the quatrain as one that
weakens the simplicity of the parts, and so increases their dependence on the
whole.
How does the perceived effect differ in each of the two versions? Intuitive
ly, the original version is wittier, more conclusive, than the transcribed version.
The foregoing analysis may account for this perceptual difference. The key-
term is Poetic Closure.
Closure occurs when the concluding portion of a poem creates in the

reader a sense of appropriate cessation. It announces and justifies the
absence of further development; it reinforces the feeling of finality,
completion and composure which we value in all works of art; and it gives
ultimate unity and coherence to the reader's experience of the poem by
providing a point from which all the preceding elements may be viewed
comprehensively and their relations grasped as part of a significant design
(Herrnstein-Smith, 1968: 36).
Herrnstein-Smith distinguishes between structural and thematic closural

devices. We have just considered an effective structural closural device in the
original version, and we shall have recourse in this paper to only one thematic
closural device, closural allusion, explained in the following two quotations:
The most casual survey of the concluding lines of any group of poems will
reveal that in a considerable number of them there are words and phrases
such as "last", "finished", "end", "rest", "peace", or "no more", which,
while they do not refer to the conclusion of the poem itself, nevertheless
signify termination or stability (Herrnstein-Smith, 1968: 172).
There is another, similar form of closural allusion:
references not to termination, finality, repose, or stability as such, but to

events which, in our nonliterary experiences, are associated with these
qualities—events such as sleep, death, dusk, night, autumn, winter, de
scents, falls, leave-takings and home-comings (Herrnstein-Smith, 1968:
175-176).
If we now consider the two versions of Omar Khayyám's Rubáiyát, we find in

both two aspects of closural allusion: going away, that is a metaphor for
death. The source of the perceptual difference between the two versions is
that in the original version the recurrence, the return, causes the poem to
close with "a click", so as to impose maximal unity and coherence upon the
reader's experiencing of the poem. This quality joins forces with the thematic
element to render the closural allusion more effective. In the transcribed
version, by contrast, the rhyming end of the fourth line generates a structural
closure for the couplet only, disturbing, as a corollary, the perceptual integrity
of the whole quatrain. Hence the wittier and more conclusive character of the
original version of the quatrain.
Let us consider now a second transcription of the same Rubáiyát:

How Sultán after Sultán came to stay,
Abode his destined Hour, and went his way.
When one element or one pattern is repeated without sufficient variation, says
Meyer, there arises a feeling of saturation.
Our normal expectation is of progressive change and growth. A figure

which is repeated over and over again arouses a strong expectation of
change both because continuation is inhibited and because the figure is not
allowed to reach completion (Meyer, 1956: 135).
If repetition is fairly exact and persistent, change rather than further repe
tition is expected, i.e., saturation sets in (Meyer, 1956: 152).
Herrnstein-Smith relates the emotive effect of saturation to boredom and

fatigue (1968: 75).
When we contrast the impression made by the second transcription to that
made by the original version, there is a distinct feeling of saturation arising
from the four lines ending with the same rhyme. The source of unity in the two
versions is of different kinds. The key-terms for the distinction between them
are grouping and differentiation. The unchanging sequence of one rhyme in
four or more lines is perceived as a homogeneous sequence, without sufficient
differentiation. This kind of unity is relatively unstructured, its effect is pro
duced mainly through accumulation. The original Rubáiyát, with the initial
symmetrical couplet-pattern, followed by the third, "deviant" line, and the
fourth line "returning" to the rhyme established at the beginning, constitutes a
coherent structure, imposing unity upon the sequence. We may adapt to the
issue in hand Meyer's discussion of a slightly different issue:
It is this creation of a larger rhythmic unit that gives the total phrase its
over-all rhythmic form. For just as a series of beats which are equal both in
accent and duration will not give rise to an impression of rhythm (except
in so far as the mind imposes its own arbitrary differentiation upon the
stimuli) so, too, the smaller rhythmic groups will not give rise to larger
patterns unless differentiation of accent or duration is present (1956:
111).
There is no differentiation within the sequence of four lines ending in the same
rhyme; any organisation of the lines into groups originates either in the
syntactic structure, or in the listener's mind imposing its own arbitrary
differentiation upon them. In this respect, it is the third, "deviant" line in the
Rubáiyát that generates differences and distinctions, and through them—
grouping and structure.
We hypothesised that the aesthetic qualities of the text would be related to
the degree of perceived closure. Thus, the more "closed" the text is, it is
perceived as more tense, interesting and dynamic. In this respect, closure is a
dynamic principle. As Kreitler & Kreitler (1972: 91) suggest, "we may expect
that people will sometimes prefer gestalts which are not maximally good and
regular, precisely because they arouse tension". These three versions of the
Rubáiyát were employed in two empirical studies investigating the relationship
between aesthetic qualities and perceptual organisation. We assumed that the
perceived effects of poetry were a function of the degree of perceptual organ
isation that is inherent in, or can be imposed on, the poetic text.
The Empirical Studies
In our empirical studies we used, in addition to the English version, Jabotin-

sky's Hebrew translation of the Rubáiyát discussed above, and manipulated its
rhyme-scheme in a similar fashion. In the course of our discussion we shall
comment on the differences between the English and Hebrew versions where
these differences are significant.
We investigated reader response with reference to the personality variable
"absorption". Absorption may be defined as the propensity to adopt an
experiential set. Tellegen (1981: 222) described a dichotomy of experiential vs.
instrumental set.
By experiential set is meant a state of receptivity or openness to undergo

whatever experiential events, sensory or imaginai, that may occur, with
tendency to dwell on, rather than go beyond, the experiences themselves
and the objects they represent. In this set, experiences have a quality of
effortlessness, as if they happened by themselves, and in that sense, of
involuntariness.
Instrumental set, on the other hand, is defined as a state of readiness to
engage in active, realistic, voluntary, and relatively effortful planning, de
cision making, and goal-directed behavior. Input from receptors is not used
to enhance experiencing but to make the discrimination needed for guiding
instrumental acts and evaluating achievements against standards.
The trait of absorption, originally determined in the Tellegen and Atkinson

(1974) study and found there to be predictive of hypnotic susceptibility, has
subsequently been found to be predictive of subjective experiences in a number
of areas (e.g., Glicksohn, 1989; Pekala et al., 1985), and reports on the
relationship between absorption and synaesthesia (Rader and Tellegen, 1987)
and absorption and hypnotic susceptibility (Spanos, Brett, Menary and Cross,
1987) support its use in the present study. The very inclusion of these
experiments in the present volume suggests that they are part of a larger study
on hypnotic (or "trance-inducing") poetry, a notion established in literary
studies by Snyder (Snyder, 1930; Snyder & Shore, 1983), and further
developed in some of my own writings (Tsur, 1985a; 1988; 1992a: 431-454;
1994); thus, the absorption scale ties in more specifically with these larger
goals.
Study 1
Method
Subjects Eighteen subjects (8 males and 10 females) participated in this first

study. Nine were formally qualified readers of literature, as indicated by their
current teaching and literary scholarship, and of these 6 possessed a PhD

degree (including three tenured associate or full professors of literature) and 3
an MA. A comparison group of nine formally nonqualified readers of literature
was drawn from scholars in the fields of psychology (2), sociology (1),
education (1), linguistics (2), history (2) and Bible studies (1), 5 of whom
possessed a PhD (all tenured associate or full professors, in their respective
fields) and 4 an MA.
Questionnaire Each subject was given a questionnaire to complete in his/her

own free time. The three versions of the Rubáiyát comprised the first section
of the questionnaire. Subjects were asked to evaluate each version along twelve
7-point scales, anchored by the following terms: TENSE ~ RELAXED, BORING ~
INTERESTING, UNDEFINED ~ DEFINED, STATIC ~ DYNAMIC, REALISTIC ~
MYSTIC, UNEMOTIONAL ~ EMOTIONAL, DIVERSE ~ MONOTONOUS,
UNBALANCED ~ BALANCED, UNPLEASANT ~ PLEASANT, COMPLEX ~ SIMPLE,
OPEN ~ CLOSED, WITTY ~ NOT WITTY. Similar scales have previously been
employed by Hasenfus, Martindale and Birnbaum (1983) to rate poetry. In
addition, each subject was asked to note his/her response to the text in an
open-ended format, and further to note the number of times he/she had read the
text before answering the latter. Subjects were also requested to note whether
the text was familiar, and whether the source and/or author could be identified
(based on style, or familiarity with the text).
Order of presentation of the texts was counterbalanced within groups. The
two sets of open-ended questions together with the set of 12 scales were
presented in a counterbalanced order across subjects. Finally, order of present
ation of the 12 scales was counterbalanced across subjects and texts. The
subjects were requested to read the questionnaire and reply to the questions
therein in order, and to refrain from "correcting" or referring back to earlier
parts of the questionnaire.
In the second part of the questionnaire subjects were asked to write down
everything they considered relevant to a comparison of the three texts, in terms
of rhythm and rhyme, in an open-ended format.
Finally, the subject was required to complete the thirty-four items of the ab
sorption subscale of the Multidimensional Personality Questionnaire (MPQ)
(Tellegen, 1982).
Results
In addition to the grouping factor of formal qualification in literature (group),

the subjects were divided into "low" and "high" scorers along the absorption
scale, by median split (med = 18.5). A 2 x 2 table was thus formed by crossing
both factors, with an n of 4 or 5 in respective cells. The twelve scales were
then submitted to separate three-way analyses of variance (group x

absorption x version), with repeated measures for the last factor. Due to
missing data, total N varies in the following analyses.
Significant group X absorption interactions were found for two scales:
BORING ~ INTERESTING [F (1,12) = 6.10, p < 0.05], and UNEMOTIONAL ~
EMOTIONAL [F (1,13) = 10.69, p < 0.05]. A significant version x absorption
interaction was found for the scale of OPEN ~ CLOSED [F (2,26) = 4.36, p <
0.05]. All other two-way and three-way interactions were nonsignificant.
A significant main effect for group was found for 4 scales: BORING ~ IN
TERESTING [F (1,12) = 7.62, p < 0.05], STATIC ~ DYNAMIC [F (1,13) = 5.30, p
< 0.05], UNEMOTIONAL ~EMOTIONAL [F (1,13) = 9.18, p<0.05], and
UNPLEASANT ~ PLEASANT [F (1,12) = 5.72, p < 0.05]. A significant main
effect for absorption was found for the scale UNEMOTIONAL ~ EMOTIONAL [F
(1,13) = 6.86 p<0.05], and a significant main effect for version was found for
two scales: TENSE ~ RELAXED [F (2.24) = 4.21, p < 0.05], and REALISTIC ~
MYSTIC [F (2.24) = 4.01 p < 0.05]. All other effects were either non-significant
or marginally significant.
Figure 1 Group x Absorption interactions for EMOTIONAL~UNEMOTIONAL

ratings
For the scale REALISTIC ~ MYSTIC the following means were obtained for
versions A (aaaa), B (aabb) and C (aaba), respectively: 4.0, 3.2, 3.9. While a
significant main effect was found for this scale, the actual differences among
the three means (less than 1 point on a 7-point scale) indicates that in essence
such differences have no psychological meaning. Thus, no interpretation will

be given to this effect. For TENSE ~ RELAXED the respective means were 4.7,
4.6, 3.6, indicating that version C (the original) was viewed as relatively more
tense than the other two (as predicted by the theory), though again the size of
the effect was small.
Figure 2 Group x Absorption interactions for BORING~INTERESTING ratings
Turning to the group x absorption interactions (see figures 1-2), one notes
that the low-absorption "formally-qualified" group found the three versions to
be (overall) relatively more boring (M=2.1) and less emotional (M=1.3) than
did the other groups. This result is even clearer when post-hoc pairwise
comparisons are made, with corrections for alpha-level using the Bonferroni
partition of alpha. For the scale UNEMOTIONAL ~ EMOTIONAL all comparisons
with this group are significant, while none of the other pairwise comparisons
are. One can conclude that the low-absorption "formally qualified" group
views the texts (overall) to be quite unemotional. In contrast, all three other
groups rate the texts midway along this scale. For the scale BORING ~
INTERESTING, it is only within the low-absorption group that one can detect a
significant discrepancy. And again, this group gives an exceptionally low
rating—the texts are boring—whereas the other groups rate the texts midway.
When one looks at the correlations among scales, separately for high-absorp
tion and low-absorption groups, one finds a significant correlation of around
0.8 between the two scales for the low-absorption group, but no such
correlation for high-absorption subjects. Thus it is the low-absorption sub
jects, and especially low-absorption subjects formally qualified in literature,
who consistently give low ratings here. This may suggest that high-absorption
subjects, both professionals and nonprofessionals are equally responsive to
these texts; so are low-absorption subjects when not habituated by profession
al experience.
Turning to the version X absorption interaction (see figure 3), one notes
that relative to the other versions low-absorption subjects viewed version C as
being relatively closed (M=5.9), whereas high-absorption subjects viewed it as
being relatively open (M=3.0). While this discrepancy for the original version
is the major source of the interaction, one can also note the trends here: The
low-absorption subjects view the texts generally as more closed than do high-
absorption subjects—and in particular for the aaba scheme, but also for the
aaaa scheme.
Figure 3 Version x Absorption interaction for OPEN~CLOSED ratings
Study 2
Method
Subjects Our sample consisted of 83 subjects, studying introductory

psychology (n = 39), Hebrew literature (n = 14) and English composition or
literature (n = 30). The 53 students of psychology and Hebrew literature were
exposed to Hebrew versions of the English texts employed. However, 30% of
our subjects did not differentiate among the different versions along the critical
scale of OPEN~CLOSED (which tapped the effectiveness of the experimental
manipulation). When these subjects are removed, we are left with n = 54,
whose data we report.
Questionnaire Each subject was given a questionnaire to complete in his/her

own free time. The three versions of the Rubáiyát appeared side-by-side on
one page. The subject was instructed to read the texts carefully, to compare
them, and then to evaluate each version along the 7-point scales. Each set of
seven scales was printed immediately beneath each text, and a comparative
rating was required along each scale. Order of presentation of both the texts
and the scales was counterbalanced. The subject then completed the Absorp
tion scale.
Figure 4 Mean Ratings along the UNEMOTIONAL ~ EMOTIONAL Scale as a

Function of Group and Version
Results
In this study, absorption level (defined by median split) had neither a main nor
an interactive effect with language and/or version. Thus, a 2x3 (language by
version) analysis of variance (ANOVA), with repeated measures on the last
variable, was conducted for each of the seven scales. These analyses were
supplemented by trend analyses. The quadratic component of the main effect
for Version was found to be significant for the following scales: STATIC ~
DYNAMIC, OPEN ~ CLOSED, BORING ~ INTERESTING and TENSE ~ RELAXED.
The linear component of the main effect for version was found to be significant
for UNPLEASANT ~ PLEASANT.
Apart from the scale UNPLEASANT ~ PLEASANT, the same quadratic trend
was found for the other scales (where an effect existed). For all of these, the
version aabb is judged to be more active/energetic (i.e., interesting, or tense)
than the other two. Furthermore, language of the text does not make a differ
ence here. In order to ascertain whether subjects studying literature (n = 20)
differed from those not studying literature (n = 34) in their evaluation of the
texts, two groups were formed. A series of 2x3 ANOVAs with supplementary
trend analysis was conducted, as before. The quadratic component of the
version X group interaction was found to be significant for the scales
UNEMOTIONAL ~ EMOTIONAL and BORING ~ INTERESTING. In addition, all
previously found effects were robust. Figures 4-5 present the respective means
for the two version X group interactions revealed by this series of analyses.
As can be seen, when group interacts with version one obtains a more
pronounced quadratic trend for students of literature than for the other
subjects.
Figure 5 Mean Ratings along the BORING ~ INTERESTING Scale as a Function of

Group and Version
Overall, then, the version aabb is judged by these subjects to be more

dynamic, more closed, more interesting and more tense than the other versions.
In addition, students of literature judged this version to be relatively more

emotional. In contrast, when the three versions are judged for pleasantness, the
aabb version is rated midway between the versions aaaa and aaba.
Discussion
The results of the two studies significantly differ in two respects. In the first
study, but not in the second, we have found a significant absorption x version
interaction. And in the first study the aaba version was found to be most
TENSE, whereas in the second study the aabb version was found to be the
most active~energetic.
In study 2 we found distinct U~shaped curves relating aesthetic qualities to
regularity for 5 scales: STATIC ~ DYNAMIC, OPEN ~ CLOSED, BORING ~
INTERESTING, TENSE ~ RELAXED, UNEMOTIONAL ~ EMOTIONAL. The more
closed the text is perceived to be, the more dynamic, interesting, tense, and
emotional it is judged to be. The more open the text is perceived to be, the
more static, boring, relaxed and unemotional it is judged to be. Of course, these
are all relative judgments, and the size of the effect is low. The trends, how
ever, are consistent, and lend support to the Gestalt~approach to empirical
aesthetics that we have adopted here. Closure has an aesthetic effect on
readers, revealed in such dynamic~energetic scales. These relationships are not
affected by the language of the text or by the field of study of the reader, at
least as far as our student~subjects are concerned. This latter finding is impor
tant, because it highlights the structural factor of the rhyme~scheme (i.e.,
Gestalt) in determining aesthetic judgment.
Note that it is the aabb version which is viewed in this study to be more
closed than the (original) aaba version, rather than the reverse, as we had
hypothesised. Thus, a major contribution of the second study is in highlighting
the aesthetic qualities associated with such a good Gestalt in its literary use. It
is symmetrical and is well~articulated into two equal parts; and it displays
good continuation, that creates good closure. The aaba version, on the other
hand, is perceived by these subjects as a "bad" version of both the aabb and
the aaaa ones. Thus, within this intuitive scheme, the aaba version may
receive similar ratings to that of the aaaa one. Accordingly, differences in the
versions will be eliminated by either leveling or sharpening. Sharpening,
according to Arnheim (1957: 57), refers to the "changing ... [of] a figure in
which two structural patterns compete for dominance into another that shows
clear dominance of one of them". "Leveling" attempts to minimise or even
eliminate (under conditions that keep the stimulus control weak enough to
leave the observer with a margin of freedom) the unfitting detail. Thus, when
the reader notes the deviating "b"~line in the aaba version, its significance may
be minimised, in comparison with the aaaa version. The two versions will then
receive similar ratings. On the other hand, the weight of this deviant "b"-line
may be exaggerated (sharpened).
Thus, the aaba structure is double-edged. A literary device or structure is
double-edged if it has different, or even opposite, effects in varying contexts or
performances. There is a wide range of double-edged phenomena in literature
(cf. Tsur, 1985a; 1992a: 164-171, 186-187, 356-359, and passim). The aaba
version, for instance, may be perceived as having a weak shape in which two
structural principles compete for dominance. Thus it will be rated low (nearer
to the negative pole) on the PLEASANT ~ UNPLEASANT scale; or there may be a
(frustrated) attempt to "level out" the deviating line, in which case the rating
will be again low on the same scale; or there may be a (successful) attempt to
sharpen the disturbance of the deviating line, followed by a highly gratifying
return to the initial rhyme, in which case the rating of pleasantness will be high.
The ratings of closure, too, will tend to be low in the first two cases, and high
in the last. The fact that the aabb version was perceived in this study as more
closed than the aaba version indicates that the student group did not sharpen
the deviant line in the latter version as did our highly-qualified population. We
will have to account for this difference. As will be seen, leveling-and-
sharpening are significantly employed in the second study too, but at a
different stage of the process.
Figures 4-5 puzzled us to a considerable extent. Following such findings as
those reported in figure 3, we expected radically different findings for our dif
ferent groups. Instead, we obtained essentially the same graph shapes, with
only a small difference in their angles. In other words, the two groups had
similar responses, the difference was only quantitative, not qualitative. It was
not until we restructured the problem by a sudden insight that we discovered
that the findings were significant indeed. What we had considered so far as
essentially the same shape with slightly different angles was reinterpreted as
the application of opposite cognitive strategies: leveling and sharpening. On
the two dimensions in which the two groups of our student-population
differed from one another, the students of literature sharpened the contrast
between the preferred version and the rest, whereas the others appear to have
attempted to level it. This would suggest that professional specialisation may
affect the leveling or sharpening of the results. Students of literature, who in
the course of their training have acquired tools for handling subtle structural
differences tend to sharpen them, to emphasise their importance. Students who
have not acquired such tools tend to level such differences, to play them down
as much as possible. This may indicate that our students of literature did have
recourse to sharpening, but instead of using it to "impose" structure upon the
aaba form, they used it to "impose" structure upon the set of three versions
presented to them.
This may explain certain differences between the results of our two experi
ments. In the second experiment, subjects were explicitly instructed to com-
pare the versions before rating, whereas in the first one subjects were explicitly
requested not to refer back to preceding pages when answering the questions.
That is probably the reason why subjects in the second experiment tended to
regard the three versions as constituting one Gestalt, and thus certain effects
that were conspicuous in the first experiment may have been blocked in the
second.
In our first study we found that for this group of readers absorption
interacted with the critical scale of OPEN ~ CLOSED; low-absorption subjects
judged the aaaa and aaba versions to be much more closed than the aabb one,
whereas high-absorption subjects judged these versions to be much more open.
High-absorption persons are expected to be more tolerant of weak, ambiguous
shapes than low-absorption persons. In the aaaa version the stimulus control
is weak enough to leave the low-absorption observer with a margin of freedom
to impose arbitrary closure on the poem, whereas in the aaba version there are
legitimate cues for sharpening the deviating "b"-line, so as to achieve an excep
tionally strong closure.
When we began the evaluation of our findings, we first tended to interpret
the afore-mentioned experimental results as the failure of some of our respond
ents to recognise the "true" structure of the poem. In the course of the present
analysis, however, we had to restructure the problem, and to regard the weak
shape of the aaba version in which two structural principles compete for
dominance as the one that underlies the two kinds of responses recorded
hitherto; one that (vainly) attempts to level discrepancies and one that
(successfully) sharpens them into a good perceptual organisation. Paraphrasing
Arnheim (1957: 55), the foregoing analysis gives sufficient evidence that the
realisation of a poem involves the solution of a problem—namely, the creation
of an organised whole. "Organised whole", then, is not a given fact, but rather
an achievement gained by deploying certain cognitive strategies such as leveling
and sharpening. Thus, the cognitive mechanism leveling-and-sharpening
appears to be one of our major resources of literary performance, allowing us
to account, in a systematic and principled way, for significant differences in
literary response.
Absorption and Trance-Inductive Poetry
As I suggested in Chapters 1-2, "Kubla Khan" belongs to a class of poetry

characterised among other things by a great degree of metric regularity. Snyder
(1930) calls it "hypnotic poetry", suggesting that it has a "spell-weaving" or
"trance-inductive" effect on the reader. He argued for its ability to induce a
hypnotic or hypnoid trance in the reader. Valentine (1962: 12) notes that
Richards (1925) felt "that a certain handling of metre may produce in a slight
degree a hypnoidal state, and that we need not boggle at the word hypnosis".
More recently, Snyder and Shore (1983) re-addressed the issue of such trance-
inductive poetic effects, and suggested that more consideration be given to their
study. In what follows we present the results of our empirical investigation in
this domain.
I have distinguished (Tsur, 1988) between a trance-inductive effect and an
aesthetic appreciation of such "trance-inductive poetry". Following Hepburn
(1968) and my own paper (Tsur, 1978) I distinguish between "detecting" and
"experiencing" an emotion. The reader may detect an ecstatic or hypnotic
experience in a poem, without necessarily experiencing it. Snyder speaks of
devices shared by the hypnotic process and hypnotic poetry; I have argued
that the reader may detect the effects of some of these devices, and experience
the effects of some of them, but in an attenuated way. This experiencing is
sufficiently strong for the reader to get involved, but is sufficiently attenuated,
so that the emerging experience is displaced, away from the reader to the poem,
as a perceived emotional quality.
Our empirical findings suggest that readers respond in systematically differ
ent ways to poems said to be hypnotic. Presumably, both cognitive set and
trait (e.g., absorption) will codetermine subjective experience here (Glicksohn,
1987). In this respect it is worth noting that Fellows (1986: 42) cites a study
of his which compared reading involvement in a short story among subjects
scoring high and low on tests of hypnotic susceptibility. Highly hypnotisable
subjects were found to become more involved in the story. One can therefore
assume that such subjects, as opposed to others, upon reading trance-inductive
poetry will have a particular aesthetic experience, which may in fact contain
elements of or have a structure that resemble hypnotic experience. Rather than
limiting our discussion to hypnotic susceptibility, we felt that a prime candi
date for distinguishing individual differences in experience would be "ab
sorption" (e.g., Hilgard, 1979). We have experimented so far with the reader's
absorption style and his response to rhyme patterns. Now we will capitalise
on our findings to account for certain hypnotic effects in poetry.
We note, however, that while the relationship between hypnotic suscepti
bility and absorption is significantly positive, it is of low size (e.g., Kumar &
Pekala, 1988). In fact, the relationship between the two seem to be rather
complex. Monteiro, Macdonald and Hilgard (1980) conducted a factor-analytic
study that revealed the orthogonality of the absorption and post-hypnotic
amnesia factors, both of which underlie hypnotic susceptibility. Spanos, Brett,
Menary and Cross (1987) reported a nonlinear relationship between absorp
tion and hypnotic susceptibility that was influenced by the subject's attitude
toward hypnosis (e.g., to a statement such as "A deeply hypnotised person is
robotlike and goes along automatically with whatever the hypnotist suggests").
Both Hilgard (1981: 16-17) and Spanos (1987: 779) also noted this mismatch
between absorption and hypnotic susceptibility. Nevertheless, in relation to
"hypnotic or trance-inductive poetry" (Snyder, 1930; Snyder and Shore, 1983;
Tsur, 1988), we felt that it was the absorption factor that should be predictive
here of the reader's evaluation of the text, and it was this factor that was ex
plored in the present study.
We employed a Hebrew text, "Bircat Shoshanim" ("Greeting with Roses"),
by the poet Jonathan Ratosh. It comprises two parts, each of which has three
stanzas of four lines each, in very regular rhythmic form, and all of its lines end
with the same rhyme [et] or [et]. I have discussed this poem (Tsur, 1988), and
showed how it conforms to Snyder's (1930; Snyder and Shore, 1983) defini
tion and examples of trance~inductive poetry. Among other elements, it has a
monotonous, regular rhyme scheme that may draw the reader's attention to the
sound rather than to the sense. In fact, readers have spontaneously reported its
hypnotic~like quality—as a poem that transfixes (Tsur, 1988). The poem is
not well~known, allowing such a trance~inductive effect to have maximal
strength. In this study we explored individual differences in the aesthetic evalu
ation of the poem.
Method
Subjects
Forty~one subjects served in this study. These included seventeen of the eigh
teen subjects of the first study described above, and twenty~four students of
Hebrew Literature from one of my classes.
Questionnaire
Each subject was given a questionnaire to complete in his/her own free time.
On the first page appeared the poem by Ratosh, followed by seven 7~point
scales, anchored by the following terms: TENSE~RELAXED, BORING~INTEREST
ING, STATIC~DYNAMIC, UNEMOTIONAL~EMOTIONAL, UNPLEASANT~PLEAS
ANT, COMPLEX~SIMPLE, OPEN~CLOSED. Similar scales have previously been
employed in the rating of art and poetry (Hazenfus, Martindale & Birnbaum,
1983). Subjects were required to rate the text along these scales. The seven
scales were presented in a counterbalanced order across subjects (i.e., the order
was different for different subjects, with roughly an equal number of subjects
receiving one of 7 different orderings).
The second part of the questionnaire entailed, again, the thirty~four items
comprising the Absorption subscale of the Multidimensional Personality
Questionnaire (Tellegen, 1982). The Hebrew version of this has been found to
be reliable (Glicksohn, 1989).
Results
The median absorption level in this study (20) matched that found previously
(Glicksohn, 1989). A grouping factor of absorption level ("high" and "low")
was formed by median split and the seven rating scales were submitted to
independent t-test analyses, after a multivariate analysis revealed an overall
group difference (Hotelling's T-squared = 34.15, df= 7, 29.9, p <0. 0051>).
Table 1 summarises the results of these analyses.
Table 1 Differences between low- and high-absorption subjects along

the rating scales.
Rating Scale1 Low High t2 P3

1 7 Absorption Absorption
Complex-Simple 4.58 (.36) 3.55 (.41) 1.87 .069

Boring-Interesting 3.05 (.46) 4.38 (.41) -2.16 .038
Static-Dynamic 4.21 (.41) 4.86 (.39) -1.14 .261
Tense-Relaxed 3.53 (.29) 2.62 (.29) 2.21 .039
Unemotional-Emotional 4.50 (.49) 5.05 (.32) -0.94 .354
Open-Closed 5.21 (.46) 5.05 (.48) 0.24 .811
Unpleasant-Pleasant 2.84 (.37) 4.30 (.39) -2.70 .010

Notes. 1 — Means, S.E.M. in parentheses
2 — with df ranging from 36 to 38 (due to incomplete
responses)
3 — two-tailed
As can be seen from Table 1, the subjects found the poem to be relatively
dynamic (M = 4.55), emotional (M = 4.79) and closed (M = 5.13).
Furthermore, high-absorption subjects found the poem to be relatively more
interesting, more tense and more pleasant than did the low-absorption subjects.
A tendency by the high-absorption subjects to view the poem as being more
complex is also noted. Table 2 presents the Pearson intercorrelation matrix for
these seven scales, separately for high-absorption subjects and low-absorption
subjects.
Of primary interest in Table 2 are the following correlations observed for
high-absorption subjects: (1) between "complex" and "tense"; (2) between
complex and emotional; (3) between "interesting" and "dynamic"; between
"interesting" and "emotional"; between "dynamic" and "pleasant"; and (6)
between "tense" and "emotional". Thus, for these subjects the poem is
evaluated as being interesting, tense, pleasant and complex. In contrast, the
following correlations are observed for the low-absorption subjects: (1)
between "simple" and "static"; (2) between "boring" and "pleasant"; and (4)
"closed" and "unpleasant". Thus, for these subjects, the poem is evaluated as
being boring, relaxed, unpleasant and simple.
Table 2 Pearson intercorrelation matrix (with decimals omitted) for the

seven rating scales.
1 2 3 4 5 6 7
1. Complex-Simple -37 -28 51* -46* -13 31
2. Boring-Interesting -28 83* -12 46* 08 44
3. Static-Dynamic -54* 21 -09 25 -09 45*
4. Tense-Relaxed 21 33 -33 -51* -17 44
5. Unemotional-Emotional -11 23 44 00 32 08
6. Open-Closed 33 -22 -15 -21 -36 -12
7. Unpleasant-Pleasant 01 55* 04 62* 21 -55*
p <.05, two-tailed
Note. Correlations for high-absorption subjects appear above the diagonal;

those for low-absorption subjects appear below the diagonal.
Note that while all subjects agree that the poem is relatively dynamic,
emotional and closed, for high-absorption subjects these are associated with
the poem being interesting and pleasant; in contrast, the low-absorption
subjects associate these qualities with it being boring and unpleasant.
Discussion
The present study proceeds on the Gestaltistic assumption that the realisation
of a poem involves the solution of a problem—namely, the creation of an or
ganised whole. "Organised whole", as suggested above, is not a given fact, but
rather achieved by deploying certain cognitive strategies (cf. Arnheim, 1967:
55). The different qualities perceived by low-and-high absorption subjects in
Ratosh's poem are hypothesised as being the result of the application of
different cognitive strategies in organising the poem as a whole.
The alternative cognitive strategies exploit the fact that the effect of verse
with a tendency for metric regularity is "double-edged" (cf. Tsur, 1985a;
1992a: 164-171, 431-454; 2003: 168-197; cf. above, 131). On the one hand,
regular metre implies a clear contrast between prominent and non-prominent
syllables. In this sense, regular metre has a strong rational quality. It has good
shape (strong gestalt), "it creates a psychological atmosphere of certainty, se
curity, and patent purpose"; it exhibits definite directions and organises per
cepts into predictable orders. On the other hand, the vigorous impact of regular
metre may have the effect of heightening the emotional responsiveness that
underlies hypnotic or ecstatic experiences, as suggested in chapter 2. In other

words, regular metre shares some important properties with conscious control
and the exercise of will; at the same time, it is similar to some fundamental
involuntary physiological processes, many of which consist of regularly recur
ring events. Intense physical and emotional activities in humans and animals
tend to possess regular rhythm and in time to transcend voluntary control.
Consequently, one factor that differentiates between regular metre underlying a
witty poem and that underlying an ecstatic poem is the energy level inherent in
other layers of the poem.
Another factor we find at the root of this double-edged nature is the feeling
of security. As the research of E. Frenkel-Brunswick (1968) has shown, the
intolerance of ambiguity may interfere with one's free emotional responses.
J.C. Ransom suggested that a fairly predictable metre may dispel anxiety in the
presence of ambiguity—give "false security to the Platonic censor in us"
(quoted by Chatman 1965: 212)—so that the reader may feel freed to attend to
ambiguities in the other layers of the poem. The crucial question seems to be
whether the psychological atmosphere generated by "good metric shape" is of
genuine or false "certainty, security, patent purpose", etc. That is to say, if
other layers of the poem too have a rational quality, the psychological atmo
sphere is one of genuine certainty, etc. If, however, some other layers of the
poem induce an intense psychological atmosphere of uncertainty — as, for
example, the "unreal" vision of "Kubla Khan" or "The Ancient Mariner"—reg
ular metre will impart a "false security", it will lull the vigilant "Platonic censor
in us" and make it accept the emotional quality of the poem. By the same
token, and at the same time, vigorous rhythms have a strong bodily appeal,
amplifying whatever irrational qualities there may be.
We assume, therefore, that the different cognitive strategies used in organis
ing Ratosh's poem as a whole involve mainly the factors that determine
whether the rhythm is perceived as one of false or genuine security. On the
metric level, Ratosh's poem has an obtrusive, more than usually regular,
ternary metre. The length of the lines is regular, and thus strictly predictable.
On the grouping level, however, this poem is much less predictable than e. g.
"Kubla Khan". But it achieves this state of uncertainty by a device that is very
different from that which we encountered in "Kubla Khan": all twenty four
lines of this poem rhyme on one group of two speech sounds [et]. Neverthe
less, this succession of lines with their undifferentiated rhyme-pattern is
broken up into groups of four lines. This grouping, however, is not the result
of differentiation, of perceptual structure, but of an arbitrary graphic arrange
ment of lines on the page, imposed on the poem from the outside. According to
the Gestaltistic analysis of grouping, such an insufficiently differentiated
rhyme pattern is prone to arouse a feeling of saturation, strong cravings for
change and completion.
Our normal expectation is of progressive change and growth. A figure

which is repeated over and over again arouses a strong expectation of
change both because continuation is inhibited and because the figure is not
allowed to reach completion (Meyer, 1956: 135).
These cravings, however, are transformed in this poem by the interaction of

its various layers. The resulting affect may be described, perhaps, by a more
moderate phrase: "powerful cravings for improvement, the clarification of
grouping and a sense of certainty and stability".
We may, then, point out two opposite aspects in the prosodic organisation
of Ratosh's poem. On the metric and line levels it induces a marked feeling of
security, whereas on the level of line-grouping it induces an unusually strong,
almost chaotic sense of insecurity. We claim that the different ratings of low-
and high-absorption subjects result from their different degree of willingness to
allow the chaotic quality of the grouping of verse lines. Low-absorption
subjects would attempt to keep grouping as far as possible under control,
whereas high-absorption subjects would not hesitate—would even enjoy—
letting it escape from under control. Thus, in the experience of high-absorption
subjects, regular metre will give "false security to the Platonic censor in us",
allowing them to contemplate this overall affect of insecurity; whereas in the
experience of low-absorption subjects regular metre will combine with ele
ments that are perceived as essentially non-chaotic, increasing their sense of
"genuine" security.
In the first two studies reported above we asked subjects to respond to a
four-line stanza with its rhyme pattern manipulated in a variety of ways,
within experimental procedures that were similar to the present ones. With
respect to the aaaa version we found that low-absorption subjects viewed it as
being relatively closed on the OPEN-CLOSED scale (M=5.9), whereas high-
absorption subjects viewed it as relatively open (M=3.0). Since subjects
received no instruction as to what aspects of the text they were requested to
respond to, each of them seems to have sought out and emphasised the aspect
that most suited his or her absorption style. High-absorption subjects seem to
have sought out the more fluid, less stable aspects, low-absorption subjects
seem to have sought out some relatively firm, stable aspect.
This difference seems to be quite consistent, and is emphasised in their
responses to two of the three versions. The aaaa version is ambiguous in the
sense of being "indistinct". This rhyme-scheme does not give rise to an impres
sion of controlled organisation (except in so far as the mind imposes its own
arbitrary differentiation upon the stimuli). In other words, it exerts little
resistance to the mind's attempts to impose a closural quality upon the lowly
differentiated sequence of a-rhymes. More specifically, low-absorption sub
jects will tend to impose some degree of closure upon the undifferentiated
sequence of the aaaa rhyme-scheme, offers little resistance. High-absorption
subjects, by contrast, will be rather tolerant to being exposed to the lowly-
differentiated aaaa rhyme-scheme, and may even enjoy this Gestalt-quality for
its own sake.
In Ratosh's poem, low-absorption subjects may take a clue from the
graphic arrangement of lines on the page into groups of four, imposing arbitra
ry closures at the end of each such group. Such good organisation, though
arbitrary, "creates a psychological atmosphere of certainty, security, and
patent purpose"; it exhibits definite directions, as if the preceding lines led up
to the closural line. In such a mental organisation, no chaotic quality of
ungrouped lines may arise, and the security given to the "Platonic censor" in
low-absorption subjects is not "false" but "genuine". The psychological
atmosphere of patent purpose thus imposed upon the undifferentiated rhyme-
pattern will then be in harmony with the basic attitudes characteristic of the
instrumental set.
High-absorption and low-absorption subjects thus respond to objects
organised in different ways from the same verbal material. Low-absorption
subjects respond to a flat sequence of monotonously rhyming verse lines.
High-absorption subjects, by contrast, respond to a hierarchic system in which
regular metre and predictable line-endings (that arouse a sense of security and
control) are played up against an over-all chaotic quality of insufficiently-
grouped lines, any grouping of which would be an arbitrary imposition. This
may explain why high-absorption subjects rated this poem as tense and
complex, whereas low-absorption subjects rated it as relatively relaxed and
simple. Likewise, the sequence of monotonously rhyming verse lines is felt to
be relatively boring—and, consequently, unpleasant—by low-absorption sub
jects, whereas by high-absorption subjects it is felt to contribute to an over-all
hypnotic quality, which is judged to be interesting and pleasant. This over-all
hypnotic quality is of considerable intensity, and is best actualised when
experienced for its own sake, by dwelling on it, rather than going beyond it.
That is why it is so pleasurable to high-absorption subjects adopting an expe
riential set.
Conclusion and Wider Perspectives
In a personal communication Norman Holland characterised the difference

between my approach to reader-response and his own approach as a text-
active vs. a reader-active approach. We claim that our approach is both text-
active and reader-active. We admit, though, that we tried to create an experi
mental setup in which personal, idiosyncratic responses are eliminated, and
only such responses are elicited that can systematically be related to the text or
to the respondent's decision style. This is most apparent in the results
reported in Figure 3. In our view, a text-active ingredient in one's approach is
indispensable, if one wishes to distinguish between the reader's structural
knowledge and his performance in reading. One must know what aspects of a
text give rise to what potential effects, in order to know how to interpret the
reader's response to it. We can, perhaps, rephrase our position in purely
reader-active terms. Rather than attributing an action to the text, we may point
out in it certain constraints3 on the reader's activity. We may, then, distinguish
between readers who observe more, and those who observe less such con
straints. Eventually we may find significant correlations between decision style
and professional training on the one hand, and readers' relative compliance
with the constraints on the other.
Many experimental studies of poetic structures are based on artificial
examples, thought up for purposes of the experiment. Instances of genuine
literature are usually too complex for experimental purposes, and it is quite fre
quently impossible to know to what aspects of the structure the subject has
responded. If the experimenter makes it clear to what aspect of the text he
wishes the subject to respond, he may pre-determine the results of his exper
iment; if he does not, he may not know to what aspect his subject has re
sponded. In our study we used one genuine instance of literature and two
manipulated versions, in order to foreground certain structural differences,
while attempting to keep other factors constant.
One of the aims of Cognitive Poetics is to relate perceived effects to poetic
structures in a consistent and principled manner. The first part of the present
chapter attempted to do just that, within a Gestalt-theoretical framework. The
purpose of the second part of the chapter was to obtain experimental evidence
for the suggested Gestalt qualities associated with the structure of the various
versions of the Rubáiyát under discussion. The primary Gestalt quality in
question was what Herrnstein-Smith termed poetic closure. In our experiments
we employed highly-qualified subjects (whether formally or not formally
qualified in literature), not because we hoped through their responses to obtain
direct insight into the workings of poetic competence, but because we hoped in
this way to minimise the interference of irrelevant factors. As it turned out we
detected in the responses of some of the formally-qualified subjects indications
of interfering factors that were not conspicuous in the responses of the other
subjects, which indicated the formers' reluctance to deploy all their profession
al skills "with respect to such 'simple' (or 'bad') examples of poetry".
Since subjects received no instruction as to what aspects of the text they
were requested to respond, each seems to have sought out and emphasised the
aspect that most suited his or her absorption style. High-absorption subjects
seem to have sought out the more fluid, less stable aspects, and low-absorption
It is difficult to talk about these matters without attributing some action to the
text. Even the word constraint (defined as "something that limits freedom of
action or choice") conceals rather than solves the problem, by relegating the
verb to the relative clause of the definition.
subjects seem to have sought out a relatively firm, stable aspect. This we have
found with reference to both structural closure and closural allusion.
Take the example of closural allusion as reflected in such a classical example
as "and went his way" in Omar Khayyam's verse (aaba) or, more emphatical
ly, "And went his way then, never to return", in the aabb version. Both
endings have two aspects: one aspect suggests closure—the termination of
everything that previously went on in the poem; the second aspect suggests a
certain indefiniteness, by being tacit on what happened after the departure.
This interacts with structural closure. Our results concerning the aabb version
("And went his way then, never to return") show that both high- and low-
absorption subjects judged the ending of this version to be considerably closed.
As Figure 3 shows, the difference between their judgments was insignificant.
This similarity may be accounted for in terms of the text-active approach: The
rhyme pattern of this version displays a symmetrical, "good" gestalt,
effectively constraining the reader's response. Still, there was a slight
difference between their judgments—in the direction that could be expected
from an analysis of their absorption ratings.
This difference seems to be quite consistent, and is emphasised in the
responses to the aaaa version and the aaba version. The reason seems to be
that the ending of both versions is ambiguous—though in different senses. The
aaba version imposes distinct, though conflicting constraints on the reader's
response. The aaaa version, on the contrary, is ambiguous in the sense of
being "indistinct". Whereas the aaba rhyme-scheme gives rise to a larger
rhythmic pattern by virtue of the differentiation of line-endings (that is, by
virtue of the "b"-rhyme at the end of the third line), the aaaa rhyme-scheme
will not give rise to an impression of controlled organisation (except in so far as
the mind imposes its own arbitrary differentiation upon the stimuli). In other
words, whereas the conflicting aspects of the aaba version constrain the
reader's response in opposing directions, the aaaa version exerts little resis
tance to the mind's attempts to impose a closural quality upon the lowly-
differentiated sequence of "a"-rhymes. More specifically, low-absorption
subjects will tend to impose some degree of closure upon the undifferentiated
sequence of the aaaa rhyme-scheme, while the rhyme-pattern will exert little
resistance to this. High-absorption subjects, by contrast, will be rather tolerant
of being exposed to the lowly-differentiated aaaa rhyme-scheme, whose
indistinct character may be reinforced by certain aspects of the semantic and
thematic structure of the last line, and may even enjoy this Gestalt-quality for
its own sake, "without any irritable reaching after fact and reason"—in Keats's
famous formulation.
It would seem that high-absorption subjects attribute as great a weight as
possible to the open-aspect of the closural allusion, whereas low-absorption
subjects attribute as small a weight as possible—within, of course, the
constraints of the text. As we have seen, on the prosodic level the rhyme-
scheme of the aaba version allows both a "sharpening" and a "nonsharpening"

performance. The more you exaggerate the deviance of the third line, the more
the stanza is felt to close firmly, with a "click"; this effect is reinforced by the
closing-aspect of the "closural allusion" on the thematic level. It is these ele
ments that induced the low-absorption subjects to judge this version as
considerably closed. Now, the Hebrew translation of Omar Khayyam's poem
(the aaba version) says "and descended into shadow" for "and went his way".
These Hebrew words perform the function of "closural allusion"; however, at
the same time they amplify the vague, elusive, "open" element in it: shadow
denotes, in Ehrenzweig's (1965) terms, a thing-free and Gestalt-free quality.
This may enhance the high-absorption subjects' rating this version as very
considerably open-ended.
It is through a comparative analysis of the responses to the three versions,
within each absorption group, that one can uncover the dynamics of reader-
response. According to our analysis there is a feeling of saturation arising from
the four lines of the aaaa rhyme-scheme. One would thus expect this version
to be perceived as more boring and less pleasant than the others. This indeed is
what one finds for low-absorption subjects. By contrast, high-absorption sub
jects tend to find this version—and indeed all versions—more pleasant and less
boring than do low-absorption subjects, which of course is what one would
expect. In fact, when the same subjects evaluated a much longer 24-line poetic
text by Ratosh, of the same monotonous aaaa rhyme-scheme, this difference
was accentuated (the mean rating on the BORING ~ INTERESTING scale for low-
absorption subjects was 1.6, and for high-absorption subjects 4.6). Regarding
the Gestalt qualities, our results seem to indicate that low-absorption subjects
tend to respond primarily to "strong" (closed) Gestalts, whereas high-absorp
tion subjects tend to respond primarily to "weak" (open) Gestalts.
We would assume that a competent reader would be aware of these oppos
ing qualities of closure and indefiniteness. When we reported our experiment at
the Cognitive Poetics Workshop at the Katz Research Institute one of the
participants exclaimed: "But I can point out at the end of this poem (the aaba
version) elements that 'count toward' both a closed and an open ending!" The
two aspects of the ending—suggesting closure or indefiniteness—are in essence
in competition in determining the final rating on the scale. A competent reader
who is unaffected by emotional involvement or any form of bias could use the
format of the open question to discuss these opposing qualities. Nonetheless,
our subjects realised these aspects only partially—according to the require
ments of their absorption-styles.
Our second study brought to the open an additional issue of great import
ance. George Klein (1951) established leveling-and-sharpening as a personality
variable. Some people seem to have a permanent preference for achieving cog
nitive stability in their environment by leveling differences, others by sharpen
ing them. Students of psychology leveled the differences between the versions,
students of literature sharpened them. These results, however, do not indicate

that students of psychology tend to be "levelers" or that students of literature
tend to be "sharpeners" in their personality styles, but rather that students of
literature are equipped by their training to handle the evasive differences
between the various versions, whereas students of psychology are not. As a
result, the former tend to sharpen, the latter to suppress such differeces. Thus,
one must distinguish between general attitudes and specific mental strategies.
Individuals may have recourse to the latter either in the service of a cognitive
style, or some other general attitude determined by their professional training,
or by being a member of an aesthetic or critical school. The study results
reported in this chapter have shown an important aspect emphasised time and
again in our discussion of the implied critic's decision style. In critical and
theoretical texts one can at best trace mental strategies, not the cognitive style
or the professional group of the writer: what they indicate depends on inter
pretation. When one resorts consistently to a certain kind of strategy in a
variety of situations, it may indicate a certain decision style.
Performing "Kubla Khan"
An Instrumental Study of Four Readings
Preliminary
An anonymous reviewer of this book suggested: "Because so much of the main

argument of the book centres on an analysis of prosodic structure (i.e. the
claim that Kubla Khan is a hypnotic poem and the proposed justification of
that claim, as well as the experimentally studied effects of rhyme patterns on
real readers) the discussion could have benefited from more recent work on
related issues, such as, to mention just one example, Chafe's (1994) analysis of
the intonation unit and how it can be manipulated to influence the flow of
information". Since this suggestion hits upon my present research interest, I
am delighted to take up the challenge and add this third part to my book. I will
argue, however, that the study of intonation may throw little light on the text
itself, but may yield significant insights into the vocal performance of poetry.
So I have submitted to instrumental investigation four commercially-available
readings of "Kubla Khan", in which I will explore selected aspects of the
opening and the closing passage.1
In a paper on "Kubla Khan", William Benzon (2003, online) has already
related Wallace Chafe's work on intonation to this poem. Since a major feature
of this book is a discourse on the method of studying "Kubla Khan", I will
devote slightly more attention to his handling of Chafe than I would otherwise
have this (it is by no means a criticism of Chafe). This also gives me an
opportunity to introduce certain methodological considerations that will be
useful in the ensuing discussion.
Chafe identifies three different kinds of intonation units. Substantive units

tend to be roughly five words long on average and, as the term suggests,
present the substance of one's thought. Regulatory units are generally a
word or so long (e.g. and then, maybe, mhm, oh, and so forth), and serve
to regulate the flow of ideas, rather than to present their substance. Given
The sound files for this chapter are available online:

http: //www.tau.ac.il/~tsurxx/KublaEmpirical_folder/KublaEmpirical.html
these durations, a single line of poetry can readily encompass a substantive

unit or both a substantive and a regulatory unit.
The third kind of unit, fragmentary, results when one of the other
types is aborted in mid-execution. That is to say, one is always listening to
one's own speech and is never quite sure, at the outset of a phrase,
whether or not one's toss of the syntactic line will reel-in the right fish. If
things do not go as intended, the phrase may be aborted. Fragments do not
concern us, as we are dealing with a text that has been thought-out and,
presumably, edited, rather than with free speech, which is what Chafe
studied.
I criticised this argument in a paper (Tsur, forthcoming) in which I discuss,

among other things, two readings by Sir John Gielgud of line 14 in
Shakespeare's Sonnet 129.2 One striking thing about the two readings of this
line is that Gielgud divides the verse line into at least three sentence fragments
(sentence fragment is a word, phrase, or clause that in speech usually has the
intonation of a sentence but lacks the grammatical structure usually found in
the sentences of formal and especially written composition). This is a salient
feature of poetry reading in general: stretches of speech are broken up into
smaller than usual intonation units, so as to exert rigorous control over every
syllable and metrical position. This may have emotive effects too. Apparently,
therefore, "fragments do not concern us, as we are dealing with a text that has
been thought-out"; but, in fact, the longish pauses in Gielgud 1, especially the
one after "shun", give an impression that fluent intonation "is aborted in mid-
execution". One "is never quite sure, at the outset of a phrase, whether or not
one's toss of the syntactic line will reel-in the right fish". Consequently, the
pauses in Gielgud 1 may be perceived as a source of uncertainty, of lack of
control. The reading in Gielgud 2 is divided into no fewer chunks, but by other
means than pauses. Consequently, it is not perceived as hesitant, but on the
contrary, as a stretch of speech every bit of which is under strict control.
This issue points up a crucial theoretical problem. Chafe gives a phonolog
ical and phonetic description of syntactic structures in prose writing and in
free speech. If we transfer his categories and principles to the description of
poetry, we get a description in phonological language of what can be described
in parallel syntactic language.3 This usually sounds very scientific, but yields
very little insight into poetic structure. Indeed, after an elaborate exposition of
Chafe's distinction Benzon applies it only once, making only one brief distinc
tion in the poem "Kubla Khan": "Clearly the exclamation points early in lines
12 and 14 mark interjections and not substantive intonation units (in Chafe's
terms)". This information is not very illuminating. The problem arises because
2
The sound files for that article are available online:
http ://www.tau.ac.il/~tsurxx/Gielgud_Experiment/page_for_sound_files.html
3
I have elsewhere discussed, at considerable length, the problem of reductionism
and of re-stating the same facts in parallel languages (e.g., Tsur, 1997).
Performing "Kubla Khan' 145
Benzon applies what he knows about intonation to the printed text, and not to
a specific oral rendering of it—actual or imaginary. You may assign to the
printed text only intonation contours that are typically related to its meaning(s).
And such typical contours give you only information you already know about
the text. You know that exclamation points mark interjections; and thus, the
intonation units must be "regulatory units" and not "substantive units".
Discussions of intonation may become illuminating when text and intonation
are independent variables; when, for instance, a performer uses intonation to
solve a problem: to settle, for example, conflicting syntactic and versification
structures. These conflicting structures are part of the text; and performers
may have recourse to a stock of intonational devices prevalent in their speech
community, to offer a wide range of possible solutions; or they may choose
not to solve the problem at all.
This problem, namely that intonation is no part of the poem but of
accidental performances, causes confusion also when approaching the issue
from the performance end. Much empirical and theoretical research is
bedevilled by having mistaken an accidental performance for the poem itself.
One cannot measure pitch, duration, or amplitude relations in a poem, only in a
particular performance which, in turn, is a perceptual solution to a problem
posed by conflicting linguistic and versification patterns. That is, a
performance is, at best, constrained by the conflicting patterns of the text. At
worst, it is quite arbitrary.
I will explore two aspects of vocal performances: poetic rhythm and the
tone or mood of the speaking voice in the poem. The former is a solution for a
perceptual problem. The latter is related to interpretation. It should be noted
that both aspects are simultaneously conveyed by the same single human
voice. I am working within a theoretical framework which I developed for
handling conflicting patterns of language and versification by accommodating
them in a third pattern, that of performance. My views on the tone and mood
of the speaking voice have been crucially influenced by Iván Fónagy's "vocal
gestures" approach. In the investigation of poetry readings by professional
actors we must remember Fónagy's (2001: 88) caveat: "The actor does not
express his true emotions: he displays emotive attitudes. The psycho-physio
logical process of simulated and real emotions probably differ". Both poetic
rhythm and emotive attitudes are perceptual qualities. Thus I am exploring
vocal correlates which, in turn, may illuminate perceptual qualities. Fónagy
discusses at great length (ibid., 87-173) the articulatory and acoustic correlates
of the vocal expression of emotions. I submit that in many instances actors
have recourse to a subset of these vocal and articulatory devices, sometimes
just enough to have the audience perceive the special mood or emotion
suggested by the speaking voice. Such a conception would be supported by
Fónagy's experiments with synthesised speech, in which certain vocal
variables believed to be of emotive importance were systematically varied.
Listeners "recognise" certain configurations of as few as two acoustic cues as

expressing some tone or emotion; and when a small change is made in one of
the cues, they "recognise" another tone or emotion. Thus, for instance, to
indicate threat, "tonal rise had to be supported by a gradual and considerable
rise in intensity; without such a parallel rise of intensity level, the utterance
was usually interpreted as a simple warning" (ibid., 104). Or, "a sudden final
rise of the F0 curve [...], accompanied by a radical fall in intensity, suggested
coquetry in somewhat more than 50% of the cases. The same F0 function
failed to elicit any coquetry votes if the intensity diminished less abruptly.
Such a variant suggested only 'interest'" (ibid., 105). This conception must be
complemented by Christine Bartels' (1999) work on intonational meaning
where, too, subtle and minimal oppositions such as that between a falling
boundary tone and a final fall-rise may affect the emotive attitudes of state
ments, though rarely affecting their truth value.
Finally, I wish to make a comment on the computer graphs displayed
throughout this study of recorded readings. Whatever the information they
convey to the trained phonetician, the uninitiated reader may treat them as
illustrations, as visual aids that may help him follow the discussion in a field
where no appropriate metalanguage is available to describe the elusive per
ceived effects.
The First Five Lines

Continuity and Discontinuity
In this section I will present some of my analytic tools, by way of scrutinising

Sheen's performance of the first five lines of the poem. In Chapter 2 I dis
cussed at considerable length the run-on sentence in the following lines.

I pointed out an element of restlessness in the peculiar correspondence (or

rather lack of correspondence) of the syntactic structure with the prosodic
unit. I also pointed out the principle that the nearer the syntactic break is to
the end of a verse line, the greater our relief when the missing part is supplied.
On the other hand, the nearer the beginning of a run-on syntactic unit is to the
end of a verse-line, the greater the tension thus generated. Thus, the re-
quiredness of ran is very high in this instance, and so is the "momentum"
generated by the run-on sentence, underlining the speed-aspect of ran.
Consider the stretch of words "river, ran": ran begins a run-on sentence;
Performing "Kubla Khan" 147
punctuation enhances the break before the end of the line, and by the same
token heightens the impetus of the run-on sentence.
Such statements as the foregoing are usually made while tacitly assuming an
unspecified appropriate performance. Comments on intonation contours may
be made only if one explicitly envisages a particular performance.
This sense of requiredness is further reinforced by a more elusive form,
which I pointed out elsewhere in relation to Lewis Carroll's "The Walrus and
the Carpenter" and Shakespeare's Sonnets (Tsur, 2002). The requiredness of a
stressed syllable in the last strong position of a line increases, and yields a
stronger closure, when it is a monosyllable preceded by a polysyllabic lexical
word with its stress on the penultimate syllable (as in "'The tíme has come,'
the Wálrus sáid" as opposed to "The Wálrus sáid: 'The tíme has cóme'"). This
is the case with the phrases Kubla Khan, river ran, and sunless sea.
It is exceptionally difficult to perform such a sequence of verse lines
properly. Punctuation requires a halt after "river", but the verse line demands
continuation. Then it is versification that requires a halt after "ran", but syntax
presses on to run from line to line. A peculiar feature of this passage is that
every line can be perceived as end-stopped; only when one reads on is it
perceived as a run-on. When one reaches river ran, the sentence and the verse
line apparently have a coinciding boundary, leading to a sense of rest and
stability. The ensuing preposition through sabotages this sense, and urges to
"run on". After man, again, the sentence and the line apparently have a
coinciding boundary which, again, is sabotaged by the adverb down. Thus, an
impetuous, "boundless" stream of speech is directly perceived, not inferred.
This increases the difficulty of performing the sequence. Performance should
be such that both versification and syntactic units preserve their "warring iden
tity". When we listen to Sheen's performance of these lines, that is precisely
what we hear. On the one hand, there is remarkable continuity at these crucial
points, after "river", "ran", and "man"; on the other, one perceives marked
discontinuity at those same points. We hear no pauses there nor, indeed, can
any be measured. Intuitively, this effect is due to certain acoustic manipula
tions of the words "ran, man, down". Quite conspicuously, they are exception
ally overstressed. Looking at the lower windows of Figures 1 and 2, at the
"waveplot", one may notice a thick lump right over these words, considerably
larger than the adjacent parts of the plot. This indicates great "amplitude"
(loudness).4 The same information is more conspicuous in the higher window
of Figure 3 regarding ran. These three key words are quite loud, have excep
tional duration, and are assigned a rising-falling intonation contour—ran and
down have "flat hat" patterns, man a "pointed hat" pattern (cf. Ladd, 1996:
15-17). Two of these words begin, and three end, with a liquid consonant
4
"Amplitude" is the absolute value of the maximum displacement from a zero
value during one period of an oscillation. It appears to consciousness as
"loudness", that is, the magnitude of the auditory sensation produced.
(m.n.r). A considerable part of the rising contours, and all of the falling
contours, occur on the liquids. Consequently, the vowels end with a sustained
(or even rising) high pitch, indicating expectation of continuity—and something
else besides (cf. Christine Bartels, 1999: 34). In this case, this something else is
the clear articulation of the word boundary and, by the same token, of the line
boundary.
One key term here would be "overdetermination". Loudness is an acoustic
cue for stress. Duration (lengthening) may be an acoustic cue for stress, as well
as indicating an ensuing boundary, that is, discontinuation. The rising-falling
pitch contour may be an acoustic cue for stress, as well as may serve to
articulate syllable or word boundaries. A steeply-rising pitch contour may
generate a forward-pressing "perceptual force", especially when its peak "hits"
the speech sound or the syllable after the middle (as in down). Such "late
peaking" has a forward-pressing perceptual force; and it may increase the
prominence of a syllable too.
Figure 1 Waveplot and pitch extract of "Where Alph, the sacred river, ran /
Through caverns" in Sheen's reading.5
Consider, again, the words river, ran. The duration of river is 476 msec
(milliseconds); of ran 471 msec. Within river, in turn, the duration of ri- is 194
msec; of-ver—282 msec. Now consider the following. The monosyllable ran
is of roughly equal duration with the preceding bisyllabic river. The second,
unstressed syllable of river is considerably longer than its stressed syllable. In
the case of ran duration contributes to both overstressing and indicating
discontinuation; in the case of -ver it contributes to a perception of discontinu-
5
The lower window presents the waveplot display which shows wave amplitude
(in volts) on the vertical axis, as a function of time (in milliseconds) on the
horizontal axis. The upper window presents a fundamental frequency plot,
which displays time on the horizontal axis and the estimated glottal frequency
(FO) in Hz (=pitch) on the vertical axis.
ity enhancing, in the present instance, the requiredness of ran. The words man
(583 msec) and down (611 msec) are even longer than ran. The "hat" patterns
of intonation on these salient words contribute to their perceptual isolation
from both what precedes and what follows them.
Compare now the pitch contour of sacred to that of river in Figure 1. In
both, the second (unstressed) syllable is lower than the first. However, the
higher syllable of river is lower than the lower syllable of sacred. In a
genuinely hypnotic performance, we would expect the pitch contours of the
first syllables of the two words to be of roughly equal height, so as to
emphasise the alliteration of [r] (in river, ran) and the obtrusive hypnotic
rhythm. Sheen seems to suppress this deliberately. On the other hand, the fall
from the first to the second syllable is considerably longer in river than in
sacred. What is more, at the bottom, the contour changes direction and initiates
a steep rising movement. The relatively deep pitch fall and the change of
direction enhance at this point the sense of an intruding break and, by the same
token, a compelling sense of requiredness, even though there is no measurable
pause there. Consequently, the overstressed and overarticulated ran arouses a
strong sense of relief and satisfaction and, by the same token, a strong sense of
closure at the line boundary (reinforced by the exceptionally long duration, and
the falling pitch curve on the [n]). At the same time, the sustained high pitch
on the vowel and the lack of measurable pause after the word suggest
continuation.
Figure 2 Waveplot and pitch extract of "Where Alph, the sacred river, ran /
Through caverns measureless to man / Down to a sunless sea" in Sheen's
reading.
An overarticulated and overstressed syllable in the last strong position of a

verse line generates a powerful sense of closure, of completeness, of stability
(as in ran and man in the present instance, though this sense is then sabotaged
here by the enjambment). The same kind of overstressed and overarticulated
syllable in a weak position (as Down in the present instance) generates a sense
of instability, forward drive, urge for completion and return to regularity. The
following perceptual dynamics may result from the state of affairs described
above. The stream of speech is strongly segmented into perceptual units (verse
lines) with well-articulated boundaries, reinforced by feigned phrase-endings. A
powerful propelling perceptual force is thrusting across these boundaries,
sweeping until sea, where the conflicting syntactic and versification units have
a coinciding boundary. Here the sweeping power comes to rest. If one
perceives this syntactic-prosodic propelling force as related to its object of de
scription—a forward-thrusting river that comes to rest when it reaches the
bottom level, the sunless sea—it can lend enormous perceptual reinforcement
to the description.
Figure 3 Wavepiot and amplitude envelope of "Where Alph, the sacred river, ran
/ Through caverns" in Sheen's reading.
We are dealing here with the principle that a perceptual unit tends to
preserve its integrity by resisting interruptions. This principle is predicted by
Gestalt theory, and supported by Garrett, Bever and Fodor's (1966) "click"
experiments concerning the psychological reality of linguistic segments.
Intruding events may "hit" some perceptual unit at the middle and generate
balance; or at the boundary, reinforcing stability; or between the middle and the
boundary. In the latter case, the "defensive" perceptual unit tries to push the
intruding event out, toward the boundary, to cast it out, generating what gestalt
theorists call a "perceptual force". This principle is at work on various levels,
for some of which the poet is responsible, for some—the reciter. There may be
a syntactic break within the line, sometimes marked on the printed page by
punctuation (as in river, ran). The nearer the break to the line boundary, the
stronger the "perceptual force" generated. A break (with or without a pause)
may be inserted by the performer even where syntax does not require it. A
heavily stressed syllable may reinforce stability when it occurs in a strong
position, especially in the last strong position of the line. But it may also
enhance fluidity, if it occurs in a weak position (as Down). Finally, this

principle is at work on the sub-phonemic level too. An intonation peak, just
like other intruding events, may reinforce balance at the middle of a phoneme;
and may reinforce fluidity when it hits the perceptual unit anywhere between
the middle and the boundary. Late peaking has a foreward-thrusting, early
peaking a backward-grouping effect. Consider the intonation peak in Figure 1
on the syllable ri-, and in Figure 2 on the diphthong [ow] in Down. (For an
excellent overview of research on early and late peaking see House and Wich
mann, 1996).
An additional aspect must be pointed out with respect to the exceptionally
salient words ran and man. Look at Figure 4 and listen to the phrases river,
ran and measureless to man in Sheen's reading. Ran and man have here a
peculiar character. They sound as if the narrator wanted to "hammer home"
something to a rather naive audience. As I said above, these monosyllables are
performed as excessively long and excessively loud, emphatically punctuating
the speech sequence. On top of this, the intonation curves of the two words
look and sound as being of a similar shape and pitch. What is more, ran jumps
up to this pitch from a much lower note; man jumps down from a much higher
one. It is as if no matter where you are, you return always to the same per
ceived pitch, well above the bottom line, where no rest can be achieved. Even
the rhyme sound is perceived in this context as reinforcing the "repeated
blows" of the "obtrusive" pitch.
Figure 4 Wavepiot and pitch extract of the phrases "river, ran" and "measureless
to man" in Sheen's reading.
Such an obstinate, hammering beat could make a crucial contribution to the

hypnotic quality of the poem. The reciter, however, prefers to strike a
different note in the first line: the tone of a fairytale-teller, lending an affected
quality to the narrator's voice. Thus, the hammering beat on ran and man is
reinterpreted as part of this affected tone. Moreover, it would appear that the
reciter deliberately avoids hypnotic rhythms. As we have seen, he suppresses

possible hypnotic accents in his reading.
I have claimed that in the phrase river, ran Sheen deliberately suppresses
possible hypnotic rhythms. He does the same with the phrase "Kubla Khan".
Consider these words, as excised from the title and the first line. The difference
may teach us a lot about the reciter's artistic conception. In the title phrase,
the speaker gives the more or less normal stress pattern of spoken English. The
intonation of Khan begins near the sustained pitch of Kubla; a hypnotic
rhythm would require only a slightly stronger stress (or higher pitch) on Khan
for obtrusive reiteration. In the first line, by contrast, the [u] of Kubla leaps up
to 183.750 Hz, well into the typical female pitch range, and then smoothly
falls down to 97.566 Hz (my speech analysis application, SoundScope,
specifies the typical male range as 80-150 Hz, the typical female range as 120-
280 Hz). This unusually high reset and smooth fall efficiently obliterates the
impression of reiterative rhythm. Furthermore, infused with a certain voice
quality indicating a friendly patronising attitude, such a "drawn out" intonation
contour may suggest the narrator's voice "spinning out" his story (to a naive
audience, perhaps).
Figure 5 The words "Kubla Khan", excised from the title and the first line in
Sheen's reading. Superscript [h] indicates aspiration. 6
An aspirated speech sound is pronounced with or accompanied by aspiration.

This can be illustrated by putting the back of one's hand before the mouth
when uttering the pairs of words "which — witch", or "pit — spit". In the
first member of each pair a stronger puff of air is felt. The phonetic feature
[±ASPIRATED] is phonemic in many languages; but when aspiration occurs
where not required by phonemic contrast, it may suggest a wide range of
emotions. Extreme emotional states, in general, are physiologically associated
with the emission of abrupt blasts of air.
Tone, Intonation, and Articulation: A Comparison
I have displayed online five readings of the phrase "Kubla Khan" (in two
sound files)—four excised from the first lines of four readings of this poem,
and one from the Merriam-Webster Collegiate Dictionary —Audio Edition. In
three of them (Richardson, Pack, and the dictionary) intonation rises or resets
at the onset of Khan before falling; in the other two (Sheen and Jennings) it
smoothly continues, in a straight line, the falling contour of Kubla.7 In spite of
this similarity, the latter two readings sound very different. Sheen's pronuncia
tion of Kubla Khan is crucial in establishing the "tale-teller's" tone; Jennings'
sounds more like ordinary English speech.
Figure 6 The words "Kubla Khan", excised from the first line in Sheen's and
Jennings' readings. Superscript [h]s indicate aspiration.
I wish to point out three conspicuous differences between them. First, as we

have seen, Sheen's intonation contour juts well into the typical female range
whereas Jennings' remains well within the typical male range. Second, there is
a difference of duration: 1181 msec (Sheen), versus 977 msec (Jennings). Such
a long-falling intonation contour spread over a prolonged phrase in Sheen's
reading to suggests a tone of spinning out, of stretching the story. Third, in
Sheen's reading the [k]s are strongly aspirated and overarticulated.8 Here
For some reason, I can't help hearing a small rise of intonation at the onset of
Jennings' Khan. The pitch extract, however, clearly shows a 1.5 Hz descent
from the bottom of Kubla to the top of Khan. When excising the first [a] and
the beginning of the second [a], one unambiguously hears a downward step.
Listen to them online.
The huge triangular lump in the waveplot before ub la of the first line in
Sheen's reading (see Figure 6) indicates overarticulated and aspirated [k], and
has no counterpart in the other readings.
aspiration and overarticulation seem to indicate strong, unspecified emotions.

They enable the reciter to speak slowly and tone down the emotional dyna
mics of the loudness dimension, and still display an emotional response.
As to pace of speech, one may fruitfully compare the above durations to the
duration of this phrase in the other readings: Richardson—1350 msec; Pack—
963 msec; dictionary—851 msec. Jennings, Pack, and the dictionary attempt to
sound as near as possible to ordinary English speech; Sheen and Richardson
adjust their intonation and protracted duration to the peculiar out-of-the-way
speaking tones they have adopted (tale-telling and incantation, respectively).
Figure 7 The words "Kubla Khan", excised from the first line in Richardson's and
Pack's readings, and as spoken by a female speaker in the Merriam-
Webster Collegiate Dictionary —Audio Edition.
A similar but more elusive difference can be noticed when one compares the
word caverns in the two readings reflected in Figures 1 and 14. The difference
between the strongly aspirated and unaspirated [k]s is, again, very conspic
uous. But there is also a great similarity and difference in the emotive effect of
intonation. They sound as essentially "the same" performance, but with an
elusive emotive difference. Jennings' intonation sounds somehow "matter-of-
Performing "Kubla Khan 155
fact", reserved, "kept back"; Sheen's intonation, by contrast, sounds more

"rounded out", "outgoing", as if making a gesture of "driving home" an emo
tion. The difference is minute, but clearly discernible. This perceived difference
has to do with the relative jutting of intonation, the relative length of fall, and
the shapes of the contours. In Figure 1, the intonation contour on caverns juts
out to 120.163 Hz, and then falls on the second syllable to 78.610 (a fall of
over 42 Hz). In Figure 14, it is roughly level with the preceding part of the
contour; and it falls from 105.000 Hz to 92.453 Hz (a fall of about 12 Hz). But
there is also a difference in the shapes of the contours. In Figure 1, the curve on
the first syllable of caverns begins with a convex "knee"; in Figure 14, with a
concave depression. This difference is more easily discernible in the larger,
online version of these figures. The convex "knee" has a decisive contribution
to the "rounded-out", "forward-moving" emotional quality of Sheen's reading;
the concave depression in the intonation curve in Jennings' reading, by con
trast, is associated with a more "austere", "kept back", relatively dispassionate
tone. We will find a similar difference of tone between the two tokens of
"Beware!" in the last stanza, in Richardson's reading, conspicuously asso
ciated, again, with the respective convex and concave curves.
Figure 8 Wavepiot and pitch extract of "In Zanadu did Kubla Khan / A stately
pleasure-dome decree" read by Richardson.
According to the received view, as formulated by Chatman, it is impossible

to convey conflicting intonation patterns in oral performance. The performer
must, he says, choose one or the other intonation pattern, with the related
meanings. Since the early nineteen seventies I have been arguing that one can
convey conflicting intonation contours by having recourse to conflicting
acoustic cues. By now I have collected (and published) an ever-increasing
corpus of such conflicting intonation contours as did Tom Barney (1990). The
present instance is perhaps unique in my corpus. In his recordings, Michael
Sheen is very consistent in suppressing versification in favour of syntax when

conflicting intonation contours are demanded. This seems to be the only place
where he so meticulously observes the enjambment. And he does it excellently
well. I strongly suspect that his motive was not so much to preserve the run-
on sentence, but to generate the "hammering home" quality of the fairytale-
teller's voice.
Now let us listen to the first stanza in Sir Ralph Richardson's reading. The
tone of this reading is very different from Sheen's. If Sheen's tone can be
located somewhere between the tale-teller and the ordinary speaker, Richard
son's strikes at once an unnatural, "bardic", incantatory tone. In certain
expressions Sheen has the tone of bringing news from Wonderland to ordinary
people like you and me. Sometimes he expresses his wonder on behalf of his
audience. Sheen expresses his wonder; Richardson generates wonder. Richard
son's tone is elevated, intended to remove his story as much as possible from
ordinary life to, perhaps, magic ceremonies. This is not a ham's overacting,
even though it may sound offensive to some present-day listeners. Contemp
orary taste favours realistic speech in the theatre, and is suspicious of such an
elevated, out-of-the-way delivery style. It is, however, very carefully con
trolled, and has a conspicuous artistic purpose: to confer upon the text what
might be described as ritualistic significance. "There is a chaunt in the recitation
both of Coleridge and Wordsworth," William Hazlitt recorded in 1823, "which
acts as a spell upon the hearer, and disarms the judgment" (Quartermain, 1998:
219). It is difficult to know today what kind of "chaunt Hazzlit had in mind.
But Richardson's incantatory tone may have precedent in Coleridge's own
reading.
Figure 9 Wavepiot and pitch extract of "Where Alf, the sacred river, ran /
through caverns" read by Richardson. The marker marks the vowel
boundary in ran, showing a peak delayed to [n].
Figures 8 and 9 show a very unfamiliar picture (corresponding to the

unfamiliar pitch movement of the reading). Not only does it climb up again and
again to roughly the same exceptionally high pitch, but the "hat" shapes ("flat"
or "pointed") are very much in evidence. What is more, these pitch movements
occur across a span of 77-170 Hz, exceeding both upwards and downwards
the "typical male range". Though the "hats" persistently jut well into the
"typical female range", Richardson's voice is perceived as a resonant deep male
voice—"resonant" both in the sense of "intensified and enriched by or as if by
resonance" and "marked by grandiloquence".
Based on Eugene Morton's interspecies findings and on his own well-
controlled experiments in human phonology, John Ohala (1994) argues that
deeper voices sound "more dominant" and higher voices more submissive. This
position has been adopted in several recent works on emotion and meaning in
intonation. I have elsewhere argued that things are more complicated, and that,
depending on other elements in the context, rising and falling intonation
contours may assume very different emotional qualities. Here, for instance, the
exceptionally high tones are not perceived as "submissive", owing to their
combination with a rich resonant voice quality even in the higher registers. The
high tones do not suggest submissiveness here, but rather an elevated tone. In
the present case, the combination of a deep resonant voice with persistent high
pitches produces what might be described as ritualistic efficacy combined with
an authoritative tone. It is as if the reciter attempted to generate an atmosphere
characterised by at least two elements which Rudolf Otto (1959) discerned in
the "numinous": "the wholly other", and "Mysterium tremendum". The mar
gins of the "hat" shapes serve to articulate the word or syllable boundaries, cue
considerable stress on the syllable, and may generate, as we shall see, certain
perceptual dynamics.
The incompleteness of the line is intensely felt, and demands the missing part.
Ran in the last position delivers the goods, generating a strong sense of relief
and closure. The exceptionally long duration of this word (684 msec), like a
pre-boundary fermata in music, arouses a sense of stability and lack of forward
motion; no new information is expected. Thus, it emphatically confirms the
line ending. The stream of speech, however, displays conflicting cues at this
point—both of continuity and discontinuity, generating a sense of firm closure
at the line ending, and an impetuous forward drive beginning a run-on sentence
at the same time. The most conspicuous cue for continuation is the contiguity
of the word-final [n] and the onset of the fricative [9] (separated by the line
boundary). In the waveplot, at the end of the [n] there is a minute protrusion
to the right, which is the onset of the [9] at the beginning of the next line. The
marker in Figure 9 marks the vowel boundary in ran, and is followed by a
minute "late peak" on the [n]. Late peaking, as we have seen, generates a per
ceptual drive forward. In addition, on the [r] of ran there is an exceptionally
long rising intonation curve (from 77.098 to 162.132 Hz), constituting one side
of the "hat". A rising sequence of sounds in music may have, as suggested by
Cooper and Meyer (1960: 15), a forward-grouping effect. It "leads" toward the
end. The other side of the "hat" falls only half way, just enough to indicate a
discontinuity at the word (and line) boundary; but since it falls only half way,
it anticipates continuation as well.
Both these initial and terminal pitch movements occur not on the vowel, but
on the liquids [r] and [n]. This leaves the vowel at an exceptionally high pitch
which, again, predicts continuation. Extreme pitch movement on liquid
consonants is a favourite resource of artistic recital. One can easily manipulate
them for artistic purposes (such as affect or poetic rhythm), interfering as little
as possible with the phonological and paralinguistic significance of the
message. We have encountered this in Sheen's reading as well.
Figure 11 Waveplot and spectrogram of "In Zanadu did Kubla Khan" read by
Richardson.
A similar story, with the necessary changes, can be told of the transition
from the first to the second line (Figure 8). As I claimed above, a disyllabic like
"Kubla", with the main stress on the penultimate syllable, may generate a quite
intense sense of requiredness for the last (stressed) syllable in the last (strong)
position. The monosyllable Khan is exceedingly long (669 msec), suggesting a
rest, bestowing a stable closure on the first line. Yet, continuity too is secured
between the two lines: the [n] of Khan is inseparably run into the next two
speech sounds at the onset of the next line: Khanas[tately], as it were. It
should be noted that there is a 156-msec silence between the [s] and [t] of
stately. This, however, is not perceived as a pause, but as an articulatory
gesture overarticulating the [t]. Again, the emphatic closure of the first line is
secured, but, at the same time, the sentence is run on, ignoring the line
boundary. The last marker marks the end of [n] in Khan, foregrounding the
continuity of Khanas, as well as a late peak on [n].
In Figure 10, too, man runs inseparably into Down; but a perceptual bound
ary between them is suggested by three means: a longish "terminal" intonation
contour on man, reset of intonation from the end of man to the onset of down,
and the exceptional duration of man (753 msec). The relative length of this
monosyllable will be apparent in comparison to the duration of the preceding
trisyllable, measureless (685 ms). There is an additional terminal contour on
sea, where verse lines and run-on sentence have a coinciding boundary.
As D.B Fry's (1958) experiments have shown, pitch change, duration, and
overall loudness, in this descending order of effectiveness, are the acoustic cues
for stress. In her dissertation, Agaath Sluijter (1995) recently demonstrated
that change of "spectral balance" (when the loudness of higher formants9 is
relatively emphasised) is almost as effective a cue for stress as duration. In this
reading of the sequence In Xanadu did Kubla Khan a stately there are no
significant differences in the pitch peaks (see Figure 8). As a result, the
alternation of stressed and unstressed syllables is considerably backgrounded.
Figure 12 Amplitude envelope of "In Zanadu did Kubla Khan / A stately pleasure-
dome decree" read by Richardson.
9
A formant is one of several regions of concentration of energy, prominent on
a sound spectrogram, that collectively constitute the frequency spectrum of a
speech sound. In Figure 11 four or more formants can be seen.
The first syllable of Zanadu sounds more strongly stressed than the second.
As Figures 8 and 12 show, this is due to the combination of an insignificantly
higher pitch and an insignificantly greater amplitude envelope with an
insignificantly longer duration of the first syllable {Za = 223 msec; na = 200
msec). At the same time, as Figure 11 shows, the higher formants in Za- are
louder, relative to the lower formants, than in na-. In Kyubla, too, the first
syllable is perceived as more stressed than the second. This is due to a
different combination of acoustic cues. Here the higher formants in la (the
unstressed syllable) are louder, relative to the lower formants, than in Kyub-
(the stressed syllable). The intonation peaks of the two words are practically
equal. The amplitude envelope is insignificantly greater in Kyub-. The signi
ficant difference is in relative duration {Kyub = 344 msec; la = 175 msec). Now
the relative prominence of du and Khan is cued very differently. Their most
salient feature is their exceptionally long duration {du =619 msec; Khan = 669
msec). As to spectral balance, by contrast, the higher formants are remarkably
faint in these two exceedingly long syllables. As Figure 12 shows, the ampli
tude envelope of Khan, but not of du, somewhat juts out relative to the
adjacent syllables. In my experience, this difference of cues may be quite
functional, suggesting rhythmic prominence and boundaries without grossly
interfering with linguistic prominence and boundaries. Here, quite obviously,
they serve to generate an obtrusive, hypnotic rhythm which, nevertheless,
does not sound mechanical or childish. At the same time, they cue a caesura
and a line ending intruding upon the continuous stream of syntax.
I have pointed out that Sheen deliberately suppresses possible hypnotic
rhythms in the phrase Kubla Khan in the first line, and in sacred river, ran in
the third line. In these instances, Khan has a considerably lower pitch than
Kubla; and river has considerably lower pitch than either sacred or ran. We
may add that in the word Xanadu too the syllables display a marked gradual
fall of pitch, though the reciter does considerably prolong the unstressed
syllable du. On the whole, the rhythm is preserved but toned down, made
relatively natural, with its possible obtrusive quality suppressed. Richardson,
by contrast, preserves a persistently high pitch in his reading, rendering his
intonation deliberately unnatural from the very beginning. The sustained
exceptionally high pitch peaks generate a "bardic", incantatory tone; by the
same token, they equalise, so to speak, the stresses, rendering the rhythm
obtrusive, hypnotic. As shown in Figure 8, the peak of Khan is almost as high
as that of Kubla; and its excessive duration amply compensates for this
"almost". As shown in Figure 9, the peak of river, not as in Sheen's reading, is
even insignificantly higher than those of the adjacent words. The other acoustic
cues are used to differentiate between the stresses. This is most unusual be
cause, as I said, pitch change is the most effective cue for stress.
Both Richardson's and Sheen's tones have an element of stylisation: Sheen

suggests a tale-teller's, Richardson an incantatory tone. Richardson's tone is
more thorough, Sheen's more sporadic. Both reciters handle the strained en-
jambment of the first stanza masterfully. They prove (if it still needs proof)
that the received view, that in oral performance it is impossible to convey the
conflicting intonation contours of enjambment, is inadequate. However, they
handle the pattern of recurring stresses in different ways. Sheen attempts to
preserve rhythmic cadences that are as near as possible to ordinary English,
suppressing any sign of obtrusive rhythms. Richardson, by contrast, combines
an incantatory tone with obtrusive, hypnotic rhythms.
Alex Jennings aims at a delicate balance between colloquial tone and poetic
rhythm, without the stylisation of either Sheen or Richardson. In this reading
too, as in Sheen's, the syllables of the word Xanadu display a marked gradual
fall of pitch; and here, too, the reciter considerably prolongs the unstressed
syllable du (Za= 197 msec, na= 119 msec, du= 384 msec). And here too, as in
Sheen's reading, the intonation of the word Khan smoothly continues the long-
falling movement started on -la. This falling contour and the prolongation,
together, have two rhythmic (not necessarily phonological) functions: to
indicate a strong metrical position occupied by an unstressed syllable, and a
break without pause at the caesura. The duration of Khan is, however, much
shorter in this reading (513 msec) than in either Sheen's or Richardson's (in the
Merriam-Webster-Dictionary phrase "Kubla Khan", the duration of Khan is
only 440 msec). On the other hand, not as in the other readings, it is followed
by a 350-msec pause. Thus, disregarding the enjambment from the first to the
second line, Jennings imposes a closure on the line by a straightforward,
longish pause, which lengthens Khan only insignificantly beyond the
dictionary.
Figure 13 Wavepiot and pitch extract of "In Zanadu did Kubla Khan / A stately
pleasure-dome decree" read by Jennings.
In this reading, too, ran is run into through, so that there is no measurable
pause at the line boundary. The break is indicated only by the lengthening of
the word ran (448 msec). Nearly half of it is [n] (201 msec), further empha
sised by a late peak and an additional curl on it. A forward drive is generated
across this boundary by the late peak. Thus, again, the longish ran generates
both stability and a forward drive at the same time. Being much shorter than
ran in Sheen's and Richardson's readings, it is just enough to hint at this
double function, but renders the flow of speech less unnatural. On the other
hand, the intonation peaks on the stressed syllables of sacred river ran reach
up roughly to the same pitch, and thus do not deliberately suppress the
obtrusive potential of the rhythm; they even corroborate it.
Figure 14 Waveplot and pitch extract of "Where Alf, the sacred river, ran /
through caverns" read by Jennings. The marker marks the vowel boundary
in ran, showing a late peak on [n].
As Figure 15 shows, here too, as in the other performances, man and down
are longer (man—569 msec; down—434 msec) and louder than the adjacent
syllables; and the two words are contiguous. To be precise, there is a 31 msec
pause between the two words. But such brief periods of silence are not
perceived as pauses, but as an articulatory gesture, when the vocal track is
closed for articulating a consonant. In the present instance it is perceived as a
careful distinction between the consecutive alveolar consonants [n] and [d]. In
other words, contiguity is preserved, but the articulatory gesture is perceived
as a cue for discontinuity between cognate consonants. In ordinary speech
such similar consonants are less clearly distinguished; their careful distinction
here is what Gerry Knowles (1991) calls "segmental discontinuation", suggest
ing a break without a pause. This discontinuation is reinforced by the falling
"terminal" contour on man. It is noteworthy that the syntax does not require
such close contiguity; nevertheless, all three performers insist here on contin
uity reflecting the flow of speech and, at the same time, discontinuity reflecting
versification. It is also noteworthy that down occurs in a weak position, and as
such, its stress is deviant. Notwithstanding this, all three performers overstress
(rather than understress) it. The reason seems to be that, as I said above, a
stressed syllable in the first weak position of a line initiates a forward
pressure, pushing toward the end; in the present instance, to a point at which
the verse lines and the long run-on sentence have, at long last, a coinciding
boundary. This forward push is perceptually reinforced in two of the readings
(Sheen and Jennings) by a late peak on down. Even the minute late peak in
Jennings' reading is discernible (even in the continuous flow of speech) as a
slight rise of intonation. Straightforward sound imitation would demand a
steeply-falling intonation curve on this word; two of three reciters preferred to
have recourse to a late peak, reinforcing the forward drive.
Table 1 Duration of ran and man in msecs in their

two occurrences in the three readings
ran1 ran2 manl man2
Sheen 471 444 583 531
Richardson 684 523 753 627
Jennings 448 392 569 488
In Chapter 2 I pointed out that the phrases sacred river ran and measure
less to man occur, letter by letter, in two different passages of "Kubla Khan". I
compared these two passages in great detail, and found that the same words
contributed to considerably different perceived qualities in the two passages.
For our present purpose, the important thing to notice is that in the second
passage there are three clauses, the boundaries of which converge with the line
boundaries, generating a relatively relaxed tone. The first passage consists of a
single clause which runs over three lines, yielding a strained enjambment. The
theory propounded here predicts that the words ran and man would be shorter
in their second occurrence than in their first one. This is a challenging way to
test the validity of the present theory. Table 1 shows that, indeed, this is the
case with reference to both pairs of words, in all three performances.
We already had another finding that is consistent with this. In two of the
readings the sentence in the first line is run on to the second line. In these read
ings the word Khan is considerably lengthened, in harmony with the predic
tions of the present theory. Jennings marks the end of the first line by a
straightforward pause. Accordingly, in his reading the duration of Khan ex
ceeds only insignificantly its duration in the dictionary.
Listening to Roger Lloyd Pack's reading of this poem one notices a very odd
thing. The reciter segments the verse lines into two, three, four segments,
separated by straightforward pauses, making no effort to realise any of the
effects we observed in the other three readings. It sounds as if the reciter forgot
his text and needed the pauses to recall the words. Or, to put it more mildly, it
may be an idiosyncratic mannerism for which I could discover no tradition or
artistic purpose. So I am not going to submit the first stanza of this reading to
the same kind of analysis as the others. In the last stanza, however, there are
good reasons, as we shall see, to give this reading close attention.
As I have already said, the received view on enjambment is that in vocal
performance the performer can choose only one of the conflicting intonation
contours, and must suppress the other . The present conception is that both
continuation and discontinuation are cued by redundant cues in ordinary
speech. In delivering an instance of enjambment, performers may have recourse
to conflicting cues. The three reciters scrutinised above emphatically confirm
this conception, with reference to an exceptionally strained enjambment. What
is more, they have recourse to remarkably similar cues, while adopting very
different tones.
The Last Stanza
The remainder of this chapter will be devoted to selected features of the four
performances of the last stanza.
Rounded and Unrounded Contours
At first I propose to dwell on the feature which I call the "rounding effect".
Let us listen to the last line of the poem as performed by Sheen and Jennings,
paying special attention to the words "And drunk the milk". There is a
conspicuous difference between the two delivery styles. In both performances
this phrase is uttered in a higher than usual pitch range. In Jennings' reading,
Performing "Kubla Khan'' 165
however, the two lexical words "drunk" and "milk" bear fairly equal, excep
tionally strong stress. One hears two steeply-rising, longish intonation curves,
which run parallel, and reach up roughly to the same pitch. The rising contours
in Figure 17 confirm this impression. These contours end, each, in a small,
rounded humpback. Intuitively, the two contours suggest some kind of
semantic or rhythmic equivalence. In Figure 16, too, the intonation on "milk"
consists of a (less steeply) rising longish contour ending in a small humpback.
The intonation contour on the vowel of "drunk" consists of a roughly flat sur
face (the rising part of the contour occurs on the [r]; the contour on the vowel
is almost flat, with a trough at its beginning). The movement from "drunk" to
"milk" is clearly perceived as a "flat" upward step.
Figure 16 Wavepiot and pitch extract of "And drunk the milk" read by Sheen.
Figure 17 Waveplot and pitch extract of "And drunk the milk" read by Jennings.
To use Gerry Knowles' distinction, the two reciters exploited for rhythmic
purposes an opposition between two types of intonation contour, available for
bringing out a semantic contrast (personal communication; cf. Tsur, 1998:
253). When we enumerate several items of equal weight, we use just minute
humpback intonation contours. When the speaker wishes to indicate that a

later phrase is not just one more item in the list but provides additional
information about the preceding item, he will have recourse to a flat intonation
contour or with a depression in it, similar to Sheen's. Semantically, then,
Sheen's solution reflects ordinary prose usage much better. The following joke
may illuminate the issue. A panda orders a steak sandwich. After eating it he
shoots the bar-tender and walks out without paying. The manager stops him at
the entrance: "What do you think you are doing?" The panda answers: "As
you may have noticed, I am a panda; look me up in the dictionary; it says
"bearlike animal; eats, shoots, and leaves". In "eats, shoots, and leaves" (three
verbs) the intonation contour will be "rounded out" on "eats", whereas in "eats
shoots and leaves" (verb + two direct objects) it will be kept lower—flat, or
depressed. The stress of the verb is subordinated to that of the ensuing nouns,
which give some additional information about it.
Returning now to our recitation, there are two word-final [k] sounds. Intu
itively, they too contribute to the equal weight of the two stressed syllables in
Jennings' performance, and to the rhythmic subordination of "drunk" to
"milk" in Sheen's performance. One may account for this intuitive difference
with the following observations. In Sheen's reading the [n] of "drunk" is
inseparably run into the [k]; in "milk" there is a huge pause (153 msec) in
midword (!), before the stop release10 of [k]. In Jennings' reading, there are
two even longer pauses before both [k]s (169 msec and 237 msec, respective
ly). A pause in midword, before a stop, is not necessarily perceived as a period
of silence. Even such long pauses as these may be perceived as indicating an
articulatory gesture: as the period during which the vocal track is closed as part
of the articulation of a stop consonant. In this case, the longer the pause, the
more overarticulated is the consonant—as long as an articulatory gesture and
not a pause is perceived. In three instances, then, word-final [k]s are overartic
ulated; and this, in turn, increases the prominence of the syllables in which
they occur. In one instance, by contrast, [k] is not overarticulated, and
corroborates the relatively low prominence of the syllable in which it occurs.
Thus, in the present case, the relative articulation of the [k] sounds corrobor
ates the perceived subordination of "drunk" to "milk" in Sheen's reading, and
their equal rhythmic weight in Jennings' reading.
Intonation contours have no meanings of their own; contrasting intonation
contours may suggest a choice between contrasting phrase markers, if we are
aware of them (cf. Lieberman, 1967: 110-122). But if we are aware of con
trasting metric organisations, they may suggest a choice between them. The
same is true of contrasting emotional qualities. We are dealing here with intrigu
ing minimal pairs. The panda joke offers contrasting semantic and syntactic
structures. The phrase "and drunk the milk" contrasts delivery styles regarding
10
Stop release is the movement of one or more vocal organs in quitting the
position for a speech sound, experienced as a click.
poetic rhythm; but may have far-reaching emotional implications too. In

Sheen's reading the verb "drunk", bearing a flat intonation contour, is sub
ordinated to the rightmost stress of the phrase. This conforms with common
prosaic usage, and is in harmony with Sheen's general tendency to tone down
or suppress versification in favour of prose rhythm. In Jennings' reading, the
humpbacks and parallel rising curves suggest equally heavy rhythmic events,
contributing to a possible hypnotic rhythm (in Richardson's and Pack's
readings, too, the two words bear roughly equal stress). A pun on the word
"flat" may bring out the emotional implications of these contours. To the
intonation structure, this adjective applies in the sense "horizontally level", or
"characterised by no significant rise or decline"; when applied to human
qualities, it may suggest "inactive", or "lacking in animation, zest, or vigour".
In our explorations we will encounter not only rounded "humpbacks" at the
crest of intonation contours, but also rounded "knees" after the crest,
"thrusting forward", so to speak. Such rounded contours serve as a "curved"
vocal gesture, and tend to suggest an "outgoing", flexible, emotionally active
attitude, that has a "tender" ingredient. The rounded curve reflects "smooth
articulation, slow gradual transitions, in the case of tenderness" (Fónagy, 2001:
107). In music, says Fónagy (ibid., 125), "tenderness induces a gently
undulating melodic line in a relatively narrow pitch range". Even or concave
falling contours, by contrast, tend to suggest an emotionally restrained, or even
stiff—indeed, "flat" attitude. Such rounded "knees" have no meanings of their
own, but may display certain general emotional tendencies, pointing the
direction to something: it may indicate some sort of attachment or affect
toward an object, or an attempt to affect the attitudes of an addressee.11
Next, let us consider the following two lines:
2. And all should cry, Beware! Beware!

His flashing eyes, his floating hair!
The first clause in these lines is a statement. The rest are exclamations: noun
phrases or imperative verbs. Both exclamations and imperative verbs indicate
forceful utterances. Imperatives can be used toward addressees who are
present, or imagined to be present. Exclamations tend to express strong
feelings; imperative verbs—the will to influence the behaviour of another. Both
typically suggest immediacy and an unreserved, conclusive attitude. In oral
performance such an attitude is frequently suggested by an increase of loud
ness and by a falling intonation curve indicating resolvedness, finality, and self-
containedness.
Let us listen online to these two lines in Richardson's reading, and look at
Figure 18. All these exclamations jut well into the typical female range of pitch.
11
According to Magda B. Arnold, emotions are a felt tendency toward an object
judged suitable, or awayfroman object judged unsuitable
The vocal folds fully resonate in the noun phrases of the second line, lending
considerable force to the utterance. In the imperative verbs of the first line,
Richardson displays his masterful control over his voice. Anguished or fright
ened cry typically results in unusually loud voice which, in turn, tends to be of
a higher than usual pitch which, in real life or with bad actors, may result in
uncontrolled shrill shrieking. Richardson suggests great loudness in his voice,
not by increasing its amplitude, but by raising the "top" of his voice to about
220 Hz, and providing some additional vocal features of excited speech. To
avoid the uncontrolled shrieking vibration of his vocal folds, he restrains their
resonance. As a result, he merely indicates excited cries. This enables him to
express great astonishment, without bursting the constraints of his art. My
speech analysis software cannot satisfactorily account for this effect. But
Fónagy's (2001: 93) discussion of the muscular strategies applied in anger and
hatred may be illuminating of what we hear here (even though I could not
assess, like Fónagy, the correlation between my informant's muscular strate
gies and the loudness level of his vocal output):
Figure 18 Wavepiot and pitch extract of "And all should cry, Beware! Beware! /
His flashing eyes, his floating hair!" read by Richardson.
In spite of the considerable effort of the expiratory muscles and the

resulting high subglottal pressure, the intensity of the sound produced
might be lower than that produced by much less effort in tender or neutral
speech [...]. The intense expiratory effort is counteracted by an equally
intense effort of the antagonistic sphincter group at the glottal level.
To be sure, Richardson expresses no anger or hatred, but excitement. He

indicates great subglottal pressure, but applies muscular strategies to it that
prevent it from producing a shriek. So it is a mere "stylised" expression of
great excitement. In the next two phrases the high, attenuated, cry fades into a
lower, fully resonant voice expressing excitement at the sight of the frenzied
youth with "His flashing eyes, his floating hair!". The intonation peak of hair
is reduced to about 170 Hz (which is still within the typical female range), and
the vocal folds are freely resonating again. The reading compensates with a
fuller, resonant voice for the loss of pitch. The falling intonation curves are
very much in evidence in these two lines, emphatically closing the brief
phrases. Thus, the intonation curves of the exclamations suggest an unreserved,
"total" attitude on the onlookers' part. The only exception is "his floating
hair!", where the falling intonation curve does not reach the established bottom
line, and suggests lesser resoluteness and finality.
Figure 19 Waveplot and pitch extract of "Beware! Beware!" read by Richardson.

The first two graphs represent the intonation contours of the full words.
The next two graphs represent the intonation chunks immediately
preceding and following the peak of the second vowel. The last two graphs
represent the contours articulating the vowel boundaries.
If we listen now to the two tokens of Beware! in isolation and in context, we

may discern an intriguing difference between them. The tone of the first token
has an ingredient that may be described as urging, imploring, pressing (in the
sense of trying hard to persuade). But the tone of the second token is per
ceived, rather, as "curt". The intonation of the latter suggests a more rigid, that
of the former a more pliant attitude. Both have a falling terminal contour. Nor
may their relative duration account for this difference. Though the first token is
somewhat longer (758 msec) than the second (666 msec), both are unusually
long, and the shorter is far from "curt". After some experimentation I came to
the conclusion that the difference is in the rising and the falling parts of the
intonation contours. I isolated very small chunks at the beginning of the
contours, and listened to them once, twice, three times, to discriminate their
pitch movement. Then I added a small chunk, and then an additional chunk.
This procedure enabled me to discern certain crucial parts of the pitch move
ment. Then I listened to the falling part of the curves. You too may listen
online to these experiments.
In Figure 19 and the accompanying sound file I present the two tokens of
"Beware!", then a segment that contains the intonation curves on both sides of
the peak; finally, the falling part of the intonation contour. When you listen to
the intonation curve on either side of the peak excised from the first token, you
may hear, though fleetingly, the change of direction. In the one excised from the
second token I could discern only a falling direction. The other difference
concerns the falling part of the contour. In the first token one may see and hear
a convex "knee". At the same point in the second token one may see and hear a
"recession" in the contour. It is these perceptible inflections in the intonation
of the first token that render it soft rather than "curt", urging rather than
peremptory. Owing to a relatively gradual, smooth bend in the intonation
curve before and after the peak, there is a tender persuasive ingredient in the
first imperative. We may recall that we saw and heard similar differences
between Sheen's and Jennings' readings of the word caverns.
The Voices of the Voice
In contrast to the beginning of his reading, Pack achieves some fluency in the
last stanza. Here, most pauses occur at places where syntax and punctuation
do require some break (there is, though, a minute—but greater than neces
sary—pause after "all"). Still, if we compare Pack's reading to the other ones,
his breaks are more frequently indicated by a pause; also, these pauses are
considerably longer, and unfilled (in Sheen's reading, for instance, the longish
pause between the two tokens of beware is filled by loud breathing). As we
shall see, in these two lines Pack compensates for the pauses by a sophisti
cated use of intonation. At any rate, these two short lines (eight-syllable-long
each) are divided by syntax into three and two intonation units, respectively.
Into such a wealth of breaks it is difficult to insert additional unmotivated
pauses. Having thus overcome this problem, Pack introduces an intriguing
twist into his text, displaying the voices of two speakers one atop the other.
There is the excited voice of "all" who would "hear me"; and the voice of "I"
who reports it. Sheen, for one, conflates the two voices into one stream of
excited exclamations. The onlookers cry excitedly; and the speaker in the actual
situation imitates, as it were, their excited cry. Pack conveys the two as
different, even though he himself, as a reciter, has only one voice. He performs
this by having recourse to the boundary intonation known as "fall-rise".
Imperative utterances frequently end with a falling contour. Christine
Bartels (1999: 269) offers two alternative explanations for this. The falling
phrasal intonation of imperatives may be due "to their being complete propo
sitions accompanied by an ASSERT morpheme in their semantico-pragmatic
representation". Alternatively, one might say "in an implicatural account" that
the basic meaning of a pre-boundary "Low" tone "across sentence modes is
'impositive' or 'dominant'". This would account for the falling contours ending
the imperatives in both Richardson's and Pack's readings, and might be
applied, with the necessary changes, to exclamatory noun phrases, as in the

second line, too. Adapting again to our purpose one of Christine Bartels'
observations: "both the [Low] nuclear accent and the [High] boundary tone are
marked options that appear to contribute to the utterance's special connot
ation" (part of which, in our case, may be an inconclusive, pensive mood), but
the basic illocutionary force of this utterance is still that of an assertion, or im
position, or exclamation (1999: 35).
The falling intonation curves suggest here a powerful unqualified attitude—
warning or horror, or both. It is the attitude of the onlookers in the
hypothetical situation reported. The rising boundary tone, by contrast, con
veys the unassertive, inconclusive mood of the speaker in the actual situation.
The rising boundary tone by itself cannot give more specific information than,
e.g., "an unassertive, inconclusive mood"; all the rest must be gathered from the
context. Thus, for instance, the speaker in the actual situation may be less than
certain that he would ever recover the lost melody, and so have doubts whether
those unqualified cries would ever be heard. In the present case one may notice
an evasive change of voice quality between the first and second tokens of
"Beware!". This change takes place precisely when the speaker shifts from the
falling boundary tones at the end of "and all should cry", and the first token of
"Beware" to the fall-rise at the end of the second "Beware", and of each of the
two noun phrases "his flashing eye" and "his floating hair".
Figure 20 Waveplot and pitch extract of "And all should cry, Beware! Beware! /
His flashing eyes, hisfloatinghair!" read by Pack.
As I said, my software cannot demonstrate this change of voice quality so as

to account for it in a principled manner. I can only characterise it impres-
sionistically as coloured with a pensive or wistful tint. Thus, the emerging
voice quality tinges the emerging unassertive, inconclusive mood with a pen
sive or wistful tint. One-and-the-same intonation contour conveys, then, two
very different attitudes, belonging to two different voices; the voice of the
wonder-stricken or horrified hypothetical onlookers crying out at the sight of

the ecstatic youth; and the wistful voice of the speaker, reporting those
exclamations in the actual situation.
To quote Christine again, a broad notion that appears safely applicable to
the distinction in boundary tones following a pre-boundary Low tone "is that
statements bearing 'declarative fall' [...] convey a sense of self-containedness,
closedness, 'finality', whereas statements bearing a fall-rise [...] convey
reservations, openendedness, or 'continuity'" (59). In these two lines, in
Pack's reading, the "falling" curves suggest the unconstrainedness, self-
containedness, 'finality' of the imperatives and exclamations expressing intense
exaltation of mind and feelings—the frenzy of the youth and the horrified
response of the onlookers in the hypothetical circumstances. The rising bound
ary tone conveys the sense of reservation, irresoluteness and openendedness
associated with pensiveness. At the same time, as I hinted above, the same
falling-rising curve contributes to an (over)articulation of the line-ending but
also takes care of a sense of continuation.
Moving now to Sheen's reading of these two lines, I wish to point out three
conspicuous aspects. First, as I have said, unlike Pack, he conveys the
horrified calls of the crowd by an excited pitch movement and voice quality.
Second, he, too, like Richardson, cries out in a controlled, soft voice. Third, he
disambiguates the syntactic ambiguity which I pointed out in these lines in
Chapter 2, in the service of an aesthetic conception at which I hinted at the
beginning of the present chapter. My theoretical framework is insufficient for
handling the excited voice quality in a reasoned manner comparable to certain
other ingredients of the readings. Here, and elsewhere, I can only draw atten
tion to significant aspects of voice quality, occasionally adding sporadic
comments. I can, however, point out here the loud breathings in the pauses
between the two tokens of "Beware!", and after "hair", as if the excited
speaker were gasping for air. These gasps bestow a dramatic functionality on
the pauses.
At the beginning of the present chapter I pointed out that Sheen's handling
there of the enjambment is quite exceptional in his recordings. In his album of
Great Romantic Poets, he is very consistent in suppressing versification in
favour of syntax when conflicting intonation contours are demanded. Here
seems to be the only place where he so meticulously observes the conflicting
cues of the enjambment. And he does it excellently well. I strongly suspect,
however, that his motive was not so much to preserve the run-on sentence but
to generate the "hammering home" quality of the fairytale-teller's voice. In the
present instance he goes even a longer way than usual to generate syntactic
continuity suppressing the line boundary. He invents syntactic continuity
even where punctuation demands discontinuity. In Chapter 2 I commented on
these two lines: The noun phrases in line 50,I said, could be the direct objects
of the repeated imperative verb in line 49 (as, for instance, in "Beware the ides
of March"). However, the exclamation marks separate them from the verbs,
turning them into ostensive exclamations, as if the onlookers pointed at the
speaker and exclaimed with horror: "His flashing eyes, his floating hair!". The
emotional effect of this device is obvious. But Sheen suppresses this ambiguity
and abandons the emotional impact of the ostensive gesture. He subordinates
the two noun phrases to the verb "Beware" as its direct objects. As the
waveplot of Figure 21 indicates, there is no measurable pause between
"Beware!" and "his"—separated on the paper by a line boundary. What is
more, as the listening ear may discern, the intonation pitch of "his" directly
continues the falling pitch contour of "Beware!", indicating that verb and noun
phrase are to be read as one unit.
His flashing eyes, his floating hair!" read by Sheen.
Range of Emotional Hues
If we listen to the four readings of the last stanza, one of the most impressive
things is that all four reciters cover a wide range of emotional hues, which fade
gradually into one another even over a very short span of text, within one verse
line, or even between two words, or within one single word. We have already
encountered such a shift of tone between the two tokens of "Beware!" both in
Richardson's and Pack's readings. In the former the shift proceeded from the
less to the more conclusive tone; in the latter the other way around.
I cannot do justice to all this at a reasonable length. So I will pay close
attention to only a small selection of small-scale paradigmatic instances. I will
briefly outline such a close succession of attitudes in Jennings' performance.
Even this account will be necessarily sketchy. Let us listen to the first two
lines of the last stanza in Jennings' reading.
3. A damsel with a dulcimer

In a vision once I saw:
The second line sounds more or less like ordinary speech, relatively
unemotional or mildly emotional—at least, as compared to the other lines in
the passage. The first line deviates from this tone in an interesting way. There
are small segments in it which differ slightly from ordinary pronunciation.
Alliteration foregrounds the [d] in damsel and dulcimer. Not as in the Mer-
riam-Webster Dictionary pronunciation of damsel, Jennings assigns a falling
intonation curve in damsel and the two tokens of dulcimer on that part of the
sound in which [d] and the vowels are coarticulated. This renders the [d]s
exceptionally emphatic. The gradual transition of the "humpback" or "knee"
begins only after that point.
Figure 22 Wavepiot and pitch extract of "damsel" excised from the Merriam-
Webster Collegiate Dictionary—Audio Edition and Jennings' reading. The
markers indicate that portion of the vowel where no trace of the [d] or
[m] can be heard.
Jennings attaches a certain emotive attitude to the two nouns in this line,
which has no trace in the other three readings. It would appear that the
"ordinary" pronunciation of the first vowel of damsel is somehow disturbed:
the centre of phonetic gravity is somehow displaced to the right. Let us
compare Jennings' utterance of this word to the way it is uttered in the Mer-
riam-Webster Collegiate Dictionary—Audio Edition. In the former, this shift of
the centre of gravity is quite conspicuous. Figure 22 shows a moderate late
peak in both vowels, a salient humpback in the dictionary, and a much smaller
one in the poem. But the ensuing falling curve is more outward bending in
Jennings' than in the Dictionary's reading, and is spread over a longer time
span. Moreover, in Jennings' reading, the voice quality of the vowel [æ] differs
considerably from the Dictionary's ordinary voice quality. It can be described
as "incomplete phonation": "half whisper" or "breathy voice". 12 In the

present context the incomplete phonation may suggest unspecified emotional
arousal, or perhaps the incomplete existence of the unattainable object of
yearning. This voice quality may account for the emotional character of the
vowel, but not for the displacement of the centre of gravity.
Measurements of the duration of the sequence [dæ] cannot account either
for the perceived difference between the two tokens of damsel. There is only
an insignificant difference of duration, but in the expected direction: in
Jennings' reading it is 197-msec-long, in the dictionary 191-msec-long. The key
to the centre displacement is in some other duration measurements. There is a
significant difference in the duration of the whole word: Jennings—714 msec,
Dictionary only 538 msec. This great difference of duration lies not in the
sequence [dæ], but in the continuant consonants following it: In Jennings' read
ing [m] lasts 181 msec, [z]—106 msec; in the dictionary, [m] lasts only 64
msec, [z] only 59 msec. Jennings' reading contains a slightly rounded falling
intonation contour and the longish continuants stretch it over a longer period of
time than in the dictionary.
Figure 23 Wavepiot and pitch extract of "A damsel with a dulcimer" read by
Jennings.
This lengthening causes the speaker to linger on the word longer than
expected without significantly affecting the stressed vowel. The emotional tint
of the vowel [æ] lends to this lingering a peculiar hue, as if the speaker were
clinging to an object, or clinging onto a pleasant memory. I would say that the
lingering on these continuants becomes an iconic representation of the clinging
to an object or a memory. One of the meanings of "nostalgia" can be suggestive
here: "a wistful or excessively sentimental yearning for return to or of some
12
"The typical half whisper can be seen as reflection of concealed passion [...].
It may also refer to sexual excitement, which is often reflected in imperfect
phonation" (Fónagy, 2001: 118).
past period or irrecoverable condition". If I am right, such a delivery style

would be in harmony with my interpretation (in Chapter 2) of the relationship
between the last stanza and the description of Kubla's earthly Paradise which,
from the perspective of the last stanza, turns into an irrecoverable Paradise
Lost, the object of the speaker's yearning.
When listening in context to this line, I have an intuition that the speaker is,
as it were, clinging to the dulcimer very much like to the damsel, but in a very
toned-down version. Here, however, the only culprit that generates that
intuition concerning the speaker's attachment to the object is a perfectly
rounded intonation "knee" (see also Figures 23-24). Or, listen to the two
tokens of dulcimer both in and out of context in the lines "A damsel with a
dulcimer" and "And on her dulcimer she played". In the first token, a clear
"rounded" effect can be heard on dul-; in the second token one hears a straight
downward line. Looking at the graphs, in both tokens the initial downward curl
marks the voice onset of the voiced plosive [d]. In the first token there is,
again, a rounded "knee"; in the second one—an acute angle initiating an abrupt
fall..
Figure 24 Wavepiot and pitch extract of the two tokens of "dulcimer" read by
Jennings.
Thus, the last stanza begins, in Jennings' reading, on a mildly emotional

tone. The emotional quality is generated by rather slight deviations from
matter-of-fact spoken English: lengthening of the continuants [m, z], the in
complete phonation in damsel, and the "rounded knee" in dulcimer, indicating
a tender tone, e.g., reminiscence, clinging to the memory of a damsel and a
dulcimer. Sweet thoughts recollected in tranquillity. This tone fades, over a
very short time span, into a highly passionate tone. This passionate tone,
however, is very different from the one suggesting the speaker's frenzy and the
onlookers' horror in Richardson's or Sheen's readings. Rather, it may be
characterised by such features of "admiration" as "regarding with wonder,
pleasure and approval". Neither the tone adopted by Jennings nor those
adopted by Richardson and Sheen are arbitrary. They rely on different

elements of the text. The preceding stanza ends with the lines "It was a miracle
of rare device, / A sunny pleasure-dome with caves of ice!" If the speaker
could revive the music within himself and rebuild the dome and the caves "with
music loud and long", it would be a miracle upon miracle. Jennings' emotional
tone expresses the wonder at this great miracle. For Richardson and Sheen, the
miraculous revival of the music and reconstruction of the caves and the dome
are part of, or induce, an ecstatic state of mind in the speaker.
In Chapter 2 I discussed this part of the poem at great length. There I argued
that there was no need to rely on allusion to Plato or Shakespeare: the reader
would readily recognise here the peak of an emotional experience. What
appears to be of great importance here, I said, is, first, that the speaker (who
ever he may be) arouses, when in the mental state described, "holy dread" in
his audience, not unlike the numinous: "For man shall not see me and live"
{Exodus, 33: 20). Second, this mental state is somehow related to his ability to
revive within himself the symphony and song of an Abyssinian maid, and to
rebuild with music loud and long the impressive sight described in detail in all
the preceding stanzas. For Jennings clinging to, and yearning for, a remote past,
followed by excited wonder, are separate stages leading to the overwhelming
peak experience.
Figure 25 Wavepiot and pitch extract of "To such a deep delight 'twould win me"
read by Jennings.
In Jennings' reading the line "To such a deep delight 'twould win me" is a
significant stage in the transition from the less to the more passionate parts of
the stanza. The line actually gains emotional significance merely owing to its
foregrounding by the overarticulation of its consonants. Consider [p] in
"deep". It is separated from dee- by a longish pause (158 msec). As I have
already said, such pauses in midword are not usually perceived as silent
periods of time, but as an articulatory gesture: as if the lips were strongly
closed for quite long, overarticulating the [p]. This involves an interesting
perceptual process. Suppose you play back the word without the [p]; you will
hear what you see on the screen: the speech sounds [di:] and a period of
silence. But if you play back the whole word, the period of silence is not
perceived as such, but as an articulatory gesture, after the event. I call this
"back-structuring". The longer the pause before the stop release, the stronger
the perceived articulation of the plosive. A similar story can be told about the
[t] at the end of delight Now consider the word-initial [d]s in "deep delight". If
you compare the two [d]s in Figures 26-27, the first [d] is marked by a greater
lump in the waveplot than the second. The horizontal difference indicates the
difference in duration; the vertical difference indicates the difference of
loudness.
Figure 26 Waveplot and pitch extract of "deep delight" read by Jennings.
The difference in loudness can be better observed in the upper window of

Figure 27. In both words, the amplitude of [d] is marked by a smaller acute
protrusion, the ensuing vowel with a higher and wider protrusion. The curve
on the first [d] is nearly as high as that on the vowel; the curve on the second
[d] is barely visible at the bottom of the graph. The first [d] is more carefully,
perhaps over-, articulated. Now notice that "deep" is preceded by a huge, 304-
msec pause. It occurs between two words and is long enough so that it should
be perceived as a pause proper. I have, however, the impression that when
listening to the whole line, back-structuring occurs here too, owing to the
already over-articulated [d] following the pause, so that this plosive is per
ceived as exceptionally well articulated. This prominence affects the alliteration
too. Now notice the two back-to-back [t]s in "delight 'twould". In Figure 25
they are marked by two exceptionally large lumps with an 86 msec pause
between them. Here, too, the pause is perceived as a refractory period between
two similar articulatory gestures rather than a straightforward pause. This is
most unusual. In ordinary speech one pronounces in such instances only one
[t]. That is how Sheen performs it. Richardson pronounces a single, but very
emphatic, [t]. Pack, too, separates the articulation of the two [t]s, but
inconspicuously, in the background. In Jennings' reading the articulation of the
two [t]s is most salient. The only reason for such an exaggerated articulation of
two separated [t]s may be the performer's wish to overarticulate the
consonants for the emotional effect mentioned above. Likewise, the [s] of such
is overarticulated: it is by far the longest (217 msec) and loudest speech sound
in this line.
Figure 27 Wavepiot and amplitude envelope of "deep delight", read by Jennings.
Figure 28 Waveplot and pitch extract of "That with music loud and long" read by
Jennings. The marker marks the onset of the vowel.
Let us turn now to Jennings' reading of the lines "That with music loud and
long, /I would build that dome in air". They serve as a transition from a mildly
emotional to an exceptionally passionate tone. I wish to point out a few salient
features here. In the first line the sound seems, somehow, "an echo of the
sense". In a sense this is odd, because the word "long" is the shortest of the
three lexical words in the line. Furthermore, the stressed vowels of these three
words are relatively short too. To be sure, alliteration draws attention to the
meanings of the words "loud" and "long". But most of the burden of expres
siveness is carried by the pre-vowel consonants. These consonants ([m], [j],
and [1]) are continuous, periodic, and voiced. The semantic feature [+LONG] is
suggested by the long rising intonation curve on these consonants, especially in
the first two words. Now consider this. Musical tones differ from mere noises
in that the former are periodic, the latter aperiodic. Speech consists of both
periodic and aperiodic sounds. That is why certain consonants (e.g. [1, m, n, j])
are regularly said to be "musical". Periodicity is related to voicing. All vowels
are periodic. Acoustically, periodicity means that a wave of the same shape
recurs indefinitely. We take for granted that vowels are periodic. Here the long
rising intonation contours draw attention not only to the [+LONG] feature in
the semantic dimension, but also to the [+PERIODIC] feature in the phonetic
dimension.
Figure 29 Wavepiot and pitch extract of "I would build that dome in air" read by
Jennings. The markers mark the [i].
What is more, the intonation contours of the three words have exactly the
same shape; that is, the macro-shapes too are "periodic". The difference in size
of these three contours has to do with a general property of intonation. For
good physiological reasons, falling intonation is the unmarked form. We
breathe in air at the beginning of utterances and air pressure under the larynx
gradually decreases with every syllable we utter. This causes the intonational
phenomenon usually dubbed as "declination". "Declination refers specifically
to the trend of the top and the bottom lines that define the limits of the local
pitch movements [...]. This means that even when nothing is 'happening'
phonologically in the contour, F0 continues to go down slightly; it also means
that a pitch movement at the beginning of the phrase will be higher than the
same pitch movement later in the phrase" (Ladd, 1996:18). Thus, the three
contours are perceived as equivalent. Briefly, intonation foregrounds periodici

ty in this line in the phonetic dimension on two levels and length in the
semantic dimension. Thus the sound is an echo of the sense expressed by the
words "music [...] long".
Figure 29a Waveplot and pitch extract of "I would build that dome in air" read by
Richardson. The markers mark the [i].
Listening to these two lines brings out two further conspicuous aspects.
First, the words build and dome with their initial voiced plosives appear to
bear exceptionally strong stress, preparing, as it were, for the highly passionate
tone of the ensuing lines. Secondly, in the intonation of the stressed syllables
of music, build, and dome one may discern that "foreward-pressing", "rounded
-out" quality we already discerned elsewhere. In music, it is the long rising
contour that foregrounds the periodic consonants that appears to generate this
impression. In build the rising curve is shorter, but ends in a late peak, changes
direction and continues in a long, falling contour. The major part of this
contour occurs on the relatively long periodic consonant [1]. The exceptionally
strong stress of build must be attributed to the intonation peak that massively
juts out of the other peaks in these lines. Sudden rise of pitch may indicate
excitement, whether joy or anger. Here, obviously, positive excitement is
suggested. The special affectionate or pleasurable tone of "build" in this
performance will be more readily appreciated if we compare it to "build" in
Richardson's performance. He too assigns exceptionally strong stress to this
word; but the rounded affectionate or pleasurable tone is missing. The vocal
gestures that make the difference can be clearly seen in Figures 29 and 29a. The
intonation contours jut out in both readings far above the other intonation
crests, but in Jennings' reaches higher up (174.308 Hz) than in Richardson's
(152.428 Hz). The vowel [i] is considerably longer in Jennings' (188 msec)
than in Richardson's (86 msec) reading. As the figures show, in Richardson's
reading the peak is aligned with the middle of the vowel; in Jennings' there is a
salient late peak. Owing to these three factors, when one isolates the [i], one
can hear in Richardson's reading a flat stable pitch; in Jennings' one can hear a
quite sophisticated pitch movement. As a result, in Jennings' reading one can
hear an "outgoing", affectionate pitch movement, and a stable, flat pitch in
Richardson's, conveying a "determined" attitude.
Curiously enough, something like the perceptual qualities we have pointed
out in build can be discerned in dome as well, even though its intonation peak
is among the lowest, and occurs in the middle of the vowel. The affectionate,
"foreward-pressing", "rounded-out" effect can be attributed to the convex
"knee" or "humpback" at the beginning of its intonation contour, and the
relatively long falling contour, most of which occurs on a periodic consonant
([m]). The strong stress can be attributed to the relatively long duration of this
word: 606 msec. The durations of the other stressed syllables are: mju-: All
msec; loud: 562 msec; long: 510 msec; and build: 562 msec. The voiced
plosives [b] and [d], in themselves, are not very loud. Still, they are perceived
as overstressed, because the whole word is exceptionally foregrounded. These
heavily stressed voiced plosives are perceived after the event as anticipating
the highly passionate tone in the ensuing lines.
Caves of Ice
As I said earlier, I am giving here only a sketchy outline of the successive

emotive tones in Jennings' reading of the last stanza. The reader may follow
for himself the rounding effects on the word air and the first syllable of sunny.
At present, I will dwell for some time on the following two extracts from the
poem:
4. It was a miracle of rare device,

A sunny pleasure-dome and caves of ice!
5. I would build that dome in air,
That sunny dome! those caves of ice!
In order to discern the special character of excerpt 5, let us compare it to ex

cerpt 4. Both contain exclamations that express wonder and admiration toward
the same objects. But the second one has an additional ingredient. According to
my interpretation in Chapter 2, that additional ingredient would be something
like strong yearning. Such a conception requires a difference in the wonder and
admiration expressed in the second exclamation. I excised the phrase "caves of
ice!" from the two passages, as read by Jennings. Listen to them.
Both are heavily loaded with emotion. But "caves of ice!" (2) sounds some
how more heavy, more forceful, more significant. Thus, for instance, in "caves
of ice!" (1) the phrase-final pitch contour is relatively sustained, suggesting an
inconclusive openness to wonder; in "caves of ice!" (2) it falls relatively deep,

suggesting a more resolute attitude. A comparison of additional acoustic
correlates may prove very illuminating, especially since in both phrases the
performer has recourse mostly to the same vocal cues, except that in "caves of
ice!" (2) these cues appear to be more forceful than in "caves of ice!" (1). First,
the whole utterance is slightly longer in "caves of ice!" (2) (2035 msec) than in
"caves of ice!" (1) (1994 msec). The word "caves" is exceptionally long in both
tokens; but in "caves of ice!" (2) it is longer (900: 846 msec), even relative to
the duration of the whole phrase. In "caves of ice!" (2) one hears, again, that
strong attachment to the object of wonder. Relative duration may, of course,
contribute to this sense of "attachment". But there are more decisive elements
too, which rely on more evasive cues. Impressionistically, the onsets of syn
tactic units are more emphatic, more vigorous in "caves of ice!" (2) than in
"caves of ice!" (1).
c
Figure 30 Wavepiot and pitch extract of "caves of ice!" (1), read by Jennings.
signifies glottal stop.
In both tokens there are glottal stop before of and ice. A glottal stop is the
speech sound we insert before aim when we say "I said an aim, not a name".
According to Fónagy (2001: 21), the glottal stop may be, among other things, a
sign of emphasis. In citation form no word begins with a vowel; it is preceded
by a glottal stop. In connected speech, however, words are usually run one
into the other, and glottal stops are omitted. Consequently, when a glottal stop
is inserted in spite of all, listeners perceive what Gerry Knowles (1991) calls
"segmental discontinuation", even if no pause is involved. In both utterances,
of and ice are preceded by a pause and then begin with a glottal stop. This
double device of discontinuity overarticulates the onsets of these words, and
renders them "harder", more emphatic. In "caves of ice!" (2) the glottal stops
are longer and louder; hence the impression that the word onsets are more
emphatic, more vigorous.
Figure 31 Waveplot and pitch extract of "caves of ice!" (2), read by Jennings. c
Figure 32 Wavepiot and amplitude envelope of "caves of ice!" (1) read by

Jennings. The markers mark the word-initial [k] and aspiration. c signifies
glottal stop.
In both tokens there are three kinds of deviation from ordinary phonation,
the purpose of which is to increase the emotive effect of the utterance:
aspiration, creaky voice and whispering. Both utterances begin with a strongly
articulated [k] followed by strong aspiration (aspiration is indicated by a
superscript h). The former overarticulates the word onset; the latter is a
phonetic expression of strong emotion. In Figures 32 and 33 I isolated the [kh]
stretch by two markers. Notice that the [kh] stretch is insignificantly longer in
"caves of ice!" (1) (132: 127 msec), but as Figures 32-33 indicate, it is very
significantly louder in "caves of ice!" (2), both in absolute terms and relative to
the ensuing vowel. Indeed, when one listens to the two tokens one after the
other, one is struck by the relative obtrusiveness of the aspiration in the

second token.
Creaky voice is produced when the vocal folds are held more tightly together
than in regular voicing. It is "the sound that often occurs at the beginning of a
sentence beginning with a vowel, or at the end of a sentence when a speaker
goes down into a very low pitch. Creaky voice is not used in English, except as
part of the intonation" (Ladefoged, 2001: 127). This mode of production lends,
in the present case, additional force and emphasis to these words, in mid-
sentence. The creaky portions of these two words are considerably longer in
"caves of ice!" (2). Phonation becomes more and more imperfect. Eventually it
turns into loud whispering. In Figures 30-31 there is no pitch contour on ice.
This reflects the fact that in the utterance of this word there is no voicing. In
the first token of the phrase there is a rising-falling intonation contour on
caves, with additional smooth curlings (see Figure 30), suggesting a relatively
mild disposition. There is no pitch contour on ice; but one hears a rising-falling
"humpback", reinforcing this disposition.
figure 33 Waveplot and amplitude envelope of "caves of ice!" (2) read by

Jennings. The markers mark the word-initial [k] and aspiration. c signifies
glottal stop.
In Figure 31, too, the intonation contour begins with a smooth "humpback"
or "knee" suggesting a tender disposition (it sounds more affectionate than the
one represented in Figure 30) which gradually but quite quickly stiffens into a
more resolute one, owing to the straightish falling contour. The rest of the
phrase is uttered at the bottom-line pitch. In Figure 31 too, as in Figure 30,
there is no pitch contour on ice; but one can hear the sustained low pitch, with
a final falling boundary tone. As Figures 30-35 show, the phrase-final [s] in ice
is considerably longer in the second utterance (432 msec) than in the first (360
msec). In Figure 32, the beginning of the [s] has a certain intensity, but then it
fast fades away. In Figure 33, it begins at a comparable intensity level, but then
it fails to fade away for some time. This is most conspicuous when one listens
to the recordings. I have quoted Bartels, Christine saying "statements bearing
'declarative fall [...] convey a sense of self-containedness, closedness, 'final
ity'". This sense coupled with loud whispering that fails to fade away
generates a sense of continuous tension between a resolved and an unresolved
emotional state.
Agaath Sluijter (1995) made a fruitful distinction between overall loudness
and spectral loudness of a word. Relative formant intensity affects the quality
of speech. The darker the patches in a spectrogram, the greater the intensity.
On comparing Figures 34 and 35 one finds that the higher light (or colour)
patches are considerably darker in the latter than in the former, adding to the
perceived forceful quality of the second utterance.
The utterance "caves of ice!" (2) begins, then, with a tender tone clinging to
the caves of ice, but then stiffens into a more forceful, tense emotional attitude.
The various kinds of incomplete phonation, aspiration, creaky voice and loud
whispering lend a strong emotional tone to both utterances, but a stronger one
to the second utterance. The wonder and admiration in excerpt 4 are direct, ex
pressed in an exclamation. In excerpt 5 they consist of a more restricted
impulse, of less direct gratification. This is reflected also in more restricted and
constricted articulatory gestures, such as more intense glottal stops, and longer
creaky voice.
Figure 34 Waveplot and spectrogram of "caves of ice!" (1) read by Jennings.

The foregoing kind of analysis cannot detect the specific contents of the
emotions conveyed by the acoustic correlates of the utterance. But it can point
out general tendencies that are in harmony with one or another emotional
tendency. In this way one can meaningfully compare excerpts 4 and 5. The
former expresses wonder and admiration. The latter too expresses wonder and
admiration, but has an additional emotional ingredient, of the longing or yearn-
ing kind. The Random House College Dictionary characterises these two nouns
among the synonyms of "desire" as follows: LONGING is an intense wish [...]
for something that is at the moment beyond reach but may be attainable at
somefixturetime. YEARNING suggests persistent, uneasy, and sometimes wist
ful or tender longing. Our analysis has revealed both the tender and the uneasy,
as well as the persistent ingredients in the vocal correlates.
Figure 35 Waveplot and spectrogram of "caves of ice! (2) read by Jennings.

Jennings ' "Stylised Loudness "
At this point, Jennings reaches the emotional peak of the experience reported:
"And all should cry, Beware! Beware! / His flashing eyes, his floating hair!". In
certain respects, his performance is similar to Richardson's and Sheen's. But it
is much more restrained, the overwhelming emotion is much more under artistic
control. All three shout in a very moderate, soft voice. All three resort to what
we have described, following Fónagy, as "incomplete phonation", "half whis
per", "breathy voice" rather than loud cry. In Fónagy's terms, "the intensity
of the sound produced [is] lower than that produced by much less effort in
tender or neutral speech", in spite of "the considerable effort of the expiratory
muscles and the resulting high subglottal pressure". It is a kind of "stylised
loudness". But, whereas Richardson's pitch juts up beyond 220 Hz, and
Sheen's almost to 210 Hz, Jennings' juts "only" to about 154 Hz (on all)); the
peak of the rest is even lower: about 136 Hz on "Beware! Beware!", and about
133 Hz on the ensuing noun phrases (which still is within the typical female
range). And, again, while the intonation contours of Sheen's exclamations fall
to around 150 Hz (except the last one), Jennings' fall to around 105, or 80, or
74 Hz (which is well within the typical male range).
As Figures 36 and 37 (as well as listening to the recording) suggest,

Jennings' exclamation is clearly articulated into four chunks by minute pauses,
the intonation contours of which reach up and then down to roughly the same
levels, over a significant pitch span. This generates a sense of obtrusive
rhythm, contributing, perhaps, to a hypnotic quality. The "incomplete phon
ation" is relatively brief in this reading; it begins toward the end of cry, and
reaches its peak on Beware! Beware! Thence, phonation becomes increasingly
complete. Thus, while Sheen keeps his exclamations on a consistently high
pitch level, and on a consistently high level of incomplete phonation, Jennings
is relatively moderate in these two respects, and returns to regular levels rela
tively soon. We should bear in mind that the 70 Hz frequency range 140-210
Hz covers only half the perceived pitch span than the frequency range 70-140
Hz. Sheen's voice is not only higher and of less complete phonation, but also
its falling contours lend it less stability.
His flashing eyes, his floating hair!" read by Jennings.
Rather than staying on a sustained high pitch, Jennings finds a different

vocal gesture to indicate a terror-stricken emotional quality. For this end he
exploits the alliteration in "flashing" and "floating". He overarticulates these
fricatives by three different means. They begin with a forceful, closer than
usual closure of the lips, almost [p]; then they continue with a fricative stream
that is longer and much louder than in Sheen's reading, for instance. In the
waveplots of Figures 37 and 38 one may clearly see that the graph of [f] is
much thicker (that is, louder) in the former than in the latter. Such a vocal ges
ture may suggest an attitude like trembling with fear. At the same time, the
stressed vowel of "flashing" is exceptionally brusque.
The best way to grasp Jennings' peculiar pronunciation of [f] in "floating" is
to compare it to his own pronunciation of "floated" in "floated midway". The
light touch of the lower lip in articulating [f] may serve as a gesture that
imitates "floating", when meaning supports this. Indeed, Jennings articulates it

smoothly in "floated". In "floating" (and, to some extent in "flashing"), by
contrast, Jennings performs the consonant [f] more roughly and more force
fully. This is perhaps supported by the meaning of "flashing" too. As shown
in Figure 37a, in "floating" the [f] is much louder (both in absolute terms and
relative to the vowel) and longer (203 vs 188 msec) than in "floated"; what is
more, its amplitude envelope is much more perturbed. Listen online to the two
recorded words. In "floated" the [f] is floating, so to speak; in "floating" it is
pronounced with lips more tightly closed, almost a [p] or [pf]. When I tried to
cut the [f] at the onset like a salami, I always heard a [pf] rather than [f].
Figure 37 Wavepiot and pitch extract of "His flashing eyes, his floating hair!"
read by Jennings.
Figure 37a Waveplot and amplitude envelope of "floated" excised from "floated
midway" and "floating" excised from "floating hair!" read by Jennings.
In Pack's reading of these same lines I pointed out that he conveyed two
different speaking voices by resorting to the falling-rising boundary tone.
Earlier in this chapter I pointed out a vocal gesture in Jennings' reading, that
consists of lingering on continuous speech sounds and suggests, at the same
time, an emotional quality by some other vocal device. This was perceived as
clinging to an object or memory, yearning, for instance, for an inaccessible
reality. Something like a combination of these two may be observed in the
reading reflected in Figure 37. The vowels of eyes and hair are lengthened to a
considerable extent. The falling intonation curves too are drawn out, with a
rounded "knee" at the beginning. These are, perhaps, the most outstanding
instances of the "rounding effect" I discussed above. It typically indicates a
tender attitude. The combination of vocal gestures in this line indicates two
different speaking voices: the terror-stricken "all" who would cry in the imag
ined situation, and the person clinging on the pleasurable fantasy in the actual
situation.
Figure 38 Wavepiot and pitch extract of "His flashing eyes, his floating hair!"
read by Sheen.
The Problem of Closure
How should such a poem end? In her admirable classic of cognitive poetics,
Poetic Closure, Barbara Herrnstein Smith (1968: 34) characterises closure as
follows: "Closure allows the reader to be satisfied by the failure of continu
ation or, put another way, it creates in the reader the expectation of nothing.
That expectation of nothing, the sense of ultimate composure [...] is variously
referred to as stability, resolution, or equilibrium". Closure "gives ultimate
unity and coherence to the reader's experience of the poem by providing a
point from which all the preceding elements may be viewed comprehensively
and their relations grasped as part of a significant design" (ibid., 36). As to the
possible contribution of intonation to closure, I quoted above Christine Bartels
to the effect "that statements bearing 'declarative fall' [...] convey a sense of
self-containedness, closedness, 'finality', whereas statements bearing a fall-rise
[...] convey reservations, openendedness, or 'continuity'" (59). In this poem,

closure is supposed to occur at the peak of an emotive crescendo.
When we compare the four readings of the last line of the poem, with special
attention to the word Paradise, we find something quite surprising. Jennings
and Sheen end their readings with a falling contour, while Richardson and Pack
end their performances on a sustained high pitch, or even with a rising pitch.
This suggests unresolvedness, non-finality, open-endedness. There may be
many aesthetic reasons for such a sabotage of closure. Here I wish to point out
one. These two endings may reflect two different conceptions of the poem's
nature. My analysis of the poem in Chapter 2 implies that it has the significant
structure of a hypnotic-ecstatic poem, the end of which constitutes the peak of
an ecstatic experience. From this peak, "all the preceding elements may be
viewed comprehensively and their relations grasped as part of' an emotive
crescendo. Coleridge himself, however, designates this poem as "A fragment",
and in his introductory remarks tells us the story of how it became a fragment.
Jennings seems to conceive of this poem as having a significant structure;
Richardson—as of a fragment. Consequently, Jennings attempts to give
"ultimate unity and coherence to the reader's experience"; Richardson does
not. The vocal gestures involved in these processes also have far-reaching
consequences regarding the attitudes of the speaking voice. Perhaps the interac
tion of the lack of closure and the resulting unassertive attitude of the speaking
voice is significant too.
Figure 39 Wavepiot and pitch extract of "And drunk the milk of Paradise" read
by Richardson (the straight vertical lines are artefacts of the machine).
The section from "And all should cry" to "holy dread" conveys an agitated
state of mind. In the text, "honey-dew" and "the milk of Paradise" are related
to this agitated state of mind via the causal conjunction "for", implying that
such food and drink marked the speaker as "wholly other" who may arouse
numinous awe. However, if we ask people to make a forced choice to relate
"honey-dew" and "the milk of Paradise" to numinous awe or a serene state of
mind, the odds are that most people would relate them to the latter. I have
quoted Iván Fónagy saying that the typical half-whisper and other types of
incomplete phonation can be seen as reflection of concealed (or perhaps
restrained) passion. Such low-intensity half-whispers may be produced, then,
both in tender and in agitated states of mind: "In spite of the considerable
effort of the expiratory muscles and the resulting high subglottal pressure, the
intensity of the sound produced might be lower than that produced by much
less effort in tender or neutral speech". All the reciters discussed resort, to
some extent, to such effortful half-whispering at the peak of the ecstatic
experience. In a surprise move, Jennings returns to half-whispering when he
utters the words "honey dew"; but this time, it is a mild, effortless pronuncia
tion. In this way he preserves the contrast between the state of mind suggested
by "beware!" or "holy dread" and that suggested by "honey dew"; notwith
standing, he can present them as parts of the same emotive crescendo.
Figure 40 Wavepiot and pitch extract of "And drunk the milk of Paradise ' read
by Pack.
Figure 41 Waveplot and pitch extract of "And drunk the milk of Paradise" read by Jennings
(the straight vertical lines are artefacts of the machine).
The issue at stake remains whether he can relate the last line to this
mounting sequence of excitement, and still achieve the "ultimate repose" of
closure. In my judgment, he does it in a most impressive way. The forcible
articulation of [p] lends a resolute quality to the onset of Paradise. As we have
seen, intonation on the words "drunk" and "milk" rises in this reading from a
low to a high level. In the last line, there is an abrupt fall of intonation, not a
gradually-descending continuous melodic line usual in terminal intonation con
tours, but a discontinuous pitch movement, jumping down from 128 Hz on
milk, through 111 Hz on of, to 85 Hz on Pa- and 71 Hz on -ra-, No pitch
contour is available on the whispered -dise; but a slightly-falling boundary tone
can be heard. The intonation on Paradise is perceived as a prolonged low pitch
after the great fall, bestowing exceptional stability on the utterance. This move
ment ensures an effective closure, but also deviation from what is normal.
Figure 42 Wavepiot and pitch extract of "And drunk the milk of Paradise" read by Sheen.
At the same time, Jennings secures a highly emotional quality by other

means than high tones. In the first place, the very deviation from normal
gradual, continuous intonation has a dramatic emotive effect. Secondly, he
overarticulates and aspirates the word-initial [p]. Both the overarticulation of
the bilabial [p] and aspiration in general have a strong emotive effect. Look at
the waveplot in Figure 41. The vertical dimension of the waveplot indicates
loudness; the horizontal—duration. The [p] is brief, but of exceptional ampli
tude; the aspiration ([h])—of exceptional amplitude and duration. This is amp
ly confirmed by what we hear: due to the overarticulated [p], word onset is
abrupt and emphatic; the aspiration is more than obtrusive. Finally, the last
syllable of Paradise fades into a whisper. That is why it bears no pitch
contour. It is illuminating to compare these features to Sheen's rendition of
Paradise in Figure 42, displaying similar tendencies, but much more moderate
ly. Listen to Sheen's line. It too ends with a falling intonation. But its contour
is a continuous, smoothly rounded rise-fall on Pa-, further falling on -radise.
Sheen's [p] is considerably softer, and the aspiration, though relatively long, is
of a lower amplitude.
Richardson displays a very different conception in this last line. He presents

the last two lines not as the culmination of a rising emotional pattern, but in
opposition to it. In contrast to Jennings' forceful ending, he offers a very mild
ending. As I have already suggested, "honey dew" and "the milk of Paradise"
have mild connotations, as opposed to the violent agitation of mind culminat
ing in "holy dread". These two lines thus fade into what Barbara Herrnstein-
Smith calls "anti-closure", suggesting something like affirmation of uncertainty
or inconclusiveness, foregrounding this mild quality in these lines. Richard
son's resonant voice becomes markedly faint in the last line, and assumes a
tender emotional tone. We noted at the beginning of this discussion that
Richardson's voice covered an exceptionally wide pitch range. In this last line,
by contrast, the peaks move quite monotonously around 100 Hz. The sentence
(and the whole poem) end on this relatively high note. Far from ending with a
falling "terminal" contour, the last curve rises from roughly 60 Hz to roughly
102 Hz, at the main level of this line, suggesting open-endedness, lack of
closure, an unresolved, nonconclusive mood, such as sadness, pensiveness,
wistfulness, longing. Much of the perception and interpretation of these lines
depends crucially on how rigorously we take the causal conjunction for. I
would say that Richardson treats this conjunction not in its strict causal sense,
but in one which "is logically nearly equivalent to since, or because, but
connects less closely, and is sometimes used as a very general introduction to
something suggested by what has gone before" (Webster's 1913 Dictionary).
For him, the last two lines are somewhat loosely related to what was going on
in the preceding few lines, reinforcing the emerging psychological atmosphere
of indecisiveness.13
Such highly active or assertive emotions as happiness/joy, excitement, confi
dence, anger or fear are typically associated with high pitch level, and some
times with wide pitch range too and large pitch variability; such low-activity
emotions as grief/sadness are typically associated with their opposites (cf.
Fónagy, 2001: 91). Richardson's tone on the last word cannot be characterised
as grief, but perhaps as "elegiac"; it certainly suggests some "indrawn",
"minor-scale" attitude.
"Unwanted finality effects in the concluding lines can be weakened or
obscured by metrical deviations" (Herrnstein-Smith, 244). Coleridge himself
provides Richardson with such a rather trivial metrical deviation, that certainly
13
"The philosophy of the new aesthetics [of anti-closure in modern art—RT] is
summarised by Meyer as follows: 'The denial of the reality of relationships and
the relevance of purpose, the belief that only individual sensations and not the
connection between them are real, and the assertion that predictions and goals
depend not upon an order existing in nature, but upon the accumulated habits
and preconceptions of men—all these rest upon a less explicit but even more
fundamental denial: a denial of the reality of cause and effect'" (Herrnstein-
Smith, 1968: 239n).
weakens the finality effect. The last strong position of the verse line (and of
the whole poem) is occupied by the last (unstressed) syllable of Paradise,
which fails to close the line properly. I would say that Richardson deliberately
refrains from counteracting this lack of proper closure, even shifts to a
conspicuously unassertive tone in his voice. Jennings, too, deliberately leaves
the last syllable unstressed. But, by contrast, he emphatically closes the verse
line (and the poem) by the abrupt jump of pitch from a higher to a lower level,
as discussed above.14
Herrnstein-Smith carefully makes the point that anti-closure "is rarely
realised as the total absence of closural effects. [...] The poet will avoid the
expressive qualities of closure while securing, in various ways, the reader's
sense of the poem's integrity" (ibid., 244). Here we are concerned with expres
sive qualities generated both by the poet and the reciter. In the last two lines of
the poem there is a rather inconspicuous built-in closural feature: parallelism.
The two lines are parallel owing to the distribution of the hackneyed biblical
phrase "milk and honey" between them; and the parallel verbs "fed" and
"drunk". The structure is chiastic too, that is, there is a reversal of the order of
words in one of two otherwise parallel phrases (here, OBJECT + VERB followed
by VERB + OBJECT). What is more, the direction of change is from the less
natural to the more natural word order. A sense of "resolution" is achieved
through progression from a marked to an unmarked word order.
But in Richardson's oral performance, the "sense of combined continuity
and stability is achieved" (ibid., 245) by the opposite potentials of one intona-
tional feature. The sustained high intonation contour, that is, the lack of a fall
at the end of a syntactic unit, suggests, as we have seen, open-endedness,
expectation for continuation. In music, too, says Leonard B. Meyer, the
sustained pitch normally should arouse "saturation", an expectation of change.
Here, however, "because this is the end of the piece, lack of forward motion
[...] is expected and desirable" (1956: 136). Thus, Richardson may indicate
unassertiveness and fragmentariness at the end of the poem, and still preserve a
minimum degree of integrity and separation from the world, required by a work
of art.
The lack of forward motion has a completely different effect in the middle of
a perceptual unit. In Richardson's performance of the last line one perceives an
extremely strong forward-pressing perceptual force in the sequence "milk of'.
In light of the foregoing discussions it is not difficult to discover the reason for
this. There is a huge 235-msec pause in the middle of the word milk, before the
[k]. At the other end, the [k] is inseparably run into of. Of course, the pause
14
My statements in this discussion are descriptive, not evaluative. I do not want
to imply that one of the two performances of the last line is better than the
other. I am exploring the aesthetic conceptions underlying Jennings' and
Richardson's performances of these lines, and how they are indicated by the
vocal gestures.
suggests overarticulation of the [k]. But the huge pause in midword on the one
hand, and the attachment of [k] to of on the other is perceived as if the stop
release intruded into the stream of speech at the "wrong" point; and our
cognitive system handles the intruder just like other intruding events we have
considered, by pushing them outward. I repeatedly perceive this pause as the
period of closing the vocal track for (over)articulating the [k]; and, by the same
token, for "gathering force" for a forward thrust. Put differently, this long
period of silence is marked by a lack of forward motion not only in mid-line,
but in mid-word, suffused with energy (invested in overarticulation); conse
quently, it presses to reach a point of proper cessation. I would guess that
some other readers too have similar perceptions. No such perceptual force is
perceived in relation to [k] in "drunk", which is perceived as a point of stabil
ity and articulateness. It is preceded as well as followed by a pause, so that the
stop release hits the period of silence nearer to the middle. Moreover, the
preceding pause is shorter than in "milk"; so the sense of lack of forward
motion is less obtrusive.
Summary and Some Wider Perspectives
For the past ten years or so I have been conducting instrumental investigations
of the rhythmical performance of poetry. I have been exploring how the same
phonetic cues may signal, at the same time, linguistic and versification units, or
even resolve possible conflicts between them. On one occasion at least, I have
gone one small step further, exploring how the same vocal cues may have a
third function too, indicating emotive attitudes as well. I called this "triple-
encodedness" (Tsur, 2000; 2002a).
In the present work I submitted to instrumental investigation commercially-
available recordings of "Kubla Khan" by four experienced British actors. I anal
ysed performances of only the first five lines and of the last stanza. Only
occasionally did I touch upon the other parts of the poem, for the sake of
comparison. I started out, as in my earlier work, with the conflicting patterns
of language and versification in the enjambment in lines 3-5. According to the
received view, in such instances the reciter must make up his mind whether he
wants to preserve the intonation of the sentence or of the verse line (see
"Performance" in The New Princeton Encyclopedia of Poetry and Poetics,
1993). The expanding corpus I explored during the past ten years suggests that
this is not so, that one can have one's cake and eat it, too. Conflicting acoustic
cues may indicate conflicting intonation contours, in the same delivery
instance. My present corpus is no exception. Three out of the four actors
successfully accommodated the conflicting linguistic and versification patterns
in a third pattern of performance. What is more, all three had recourse to
similar vocal devices in solving the problem, in a way that was in perfect har
mony with the predictions of my cognitive theory.
In my earlier work (e.g., Tsur, 1998) I argued that the same vocal resources
as the ones deployed in the rhythmical performance of enjambment are usually
deployed in resolving the conflicting patterns of stress and metre too. As I
pointed out in Chapter 2, "Kubla Khan" is one of the most regular poems in
English from the metric point of view. This was one feature of the text that
prompted my reading of "Kubla Khan" as a hypnotic poem. Here I could not
analyse the performance of consecutive stressed syllables, or of stress maxima
in weak positions, as I had done, extensively, in Shakespeare's, Milton's,
Shelley's, and Keats' poems. I had to content myself with considering solu
tions for minor deviations, or the variety of delivery styles offered by the
various performers. As I proceeded, I drifted further and further into issues re
lated to the vocal expression of emotions, attitudes, and, eventually, to the
actor's basic conception or interpretation of the poem. These latter issues are
quite unlike anything I have done before.
In the present state of our knowledge, an instrumental investigation of the
vocal output of an actor cannot predict the rhythmic or emotional character of
what the listener will perceive. There is no escape from relying on
impressionistic responses of listeners both to the rhythm and the emotion
conveyed. As a further step, one must propose a plausible hypothesis to relate
the measurable vocal (or articulatory) gestures to the perceptions of listeners.
Such a hypothesis must be based on phonetic, psychological and literary
resources. But even such a view is too optimistic regarding experimental rigour.
In another paper (Tsur, forthcoming) I report an experiment, in which I
asked members of the PSYART and the Coglit online forum lists to respond to
the rhythmic character of one verse line in two different readings by the same
actor (displayed online). Only five persons responded to my request, but in
their details they covered a surprisingly wide range of responses. All five
respondents happened to be highly qualified listeners (three literary theorists,
one musicologist, and a psychiatrist who had written extensively on the poem
in question). I concluded that the aesthetic event of poetry recital is very com
plex. The responses I received suggest that in some instances my informants
responded to different aspects of the same event. The event may be consis
tently and quite thoroughly described by the tools offered here. One must,
however, realise that in a poll like the one reported each respondent will re
spond only to a small subset of aspects. Thus, even widely different responses
may be consistent with one another.
Iván Fónagy found that we need not go to such complex events as poetry
recital to encounter such a welter of seemingly incompatible responses, and
offered a very similar explanation to mine. Fónagy had a French actress mimic
certain emotions by uttering stock phrases in a wide range of situations reflect-
ing a variety of emotive attitudes. In response to these utterances, his subjects

had to choose the name of an emotion from a list of fifteen.
In the semantic tests based on laryngographic recordings of emotions

portrayed by Marie-Claude Mestral, fear was frequently interpreted as
reproach and repressed anger; repressed anger as disdain or reproach;
reproach as repressed anger and also as joy; joy could be confused with
coquetry; tenderness with sorrow and joy; coquetry with irony (Fónagy,
2001: 120).
"The perception of an attitude is, in fact, a result of componential analysis

[...]. Errors in decoding may be due to partial neglect of vocal features, or to a
recombination of these features which differs from that intended by the
speaker" (Fónagy, 2001: 124). I suggested a rudimentary instance of compo
nential analysis and combination of features in my discussion of the word
damsel in Jennings' reading. Prolongation of the continuants [m] and [z]
generates a sense of "lingering", "attachment". The "incomplete phonation" of
the vowel [æ] in this reading is perceived as a tender emotional quality. In the
present context, prolongation and incomplete phonation may be integrated as
the perception of an emotional attitude which may be characterised as "clinging
to a memory", or "longing for a lost reality"; or, out of context, "attachment,
affection".15 This is in harmony with my own impression of the sound I hear.
I suppose that some other listeners too would discern the same quality; others,
perhaps, would not. Contrary to Fónagy, who could independently assess the
actress' intentions or vocal and articulatory gestures, I cannot claim that my
perception is more correct, or more accurate, than theirs. But I could, perhaps,
guide some of them to perceive what I perceive by pointing out those features
and their combination. And if we still disagree, we could, at least, have a rea
soned disagreement, each side knowing what the other is talking about. De
gustibus disputandum est.
As I insisted time and again during my discussion of the vocal performances,
vocal gestures cannot convey specific emotions, only general emotional ten
dencies at best. The actual contents of emotions must be looked for in the
semantic dimension of the words. This is true of the instance mentioned in the
preceding paragraph too. I also mentioned that Fónagy empirically investigated
the huge repertoire of articulatory and acoustic correlates of the vocal
expression of emotions. He pointed out that an actor does not express his true
emotions: he displays emotive attitudes. I further argued that actors do not
usually simulate the full vocal repertoire of an emotion, but have recourse to a
subset sufficient to merely indicate the emotion intended. This, I suppose, is
more true of poetry recital than of realistic play-acting in the theatre. On
15
Notice the common meaning components of "lingering", "clinging", and
"attachment".
comparing Jennings' and Richardson's performances of the two lines "And all
should cry, Beware! Beware! / His flashing eyes, his floating hair!" we find
that they have recourse to different vocal gestures for characterising the cry of
the terror-stricken audience, foregrounding different aspects of the emotional
state. In Jennings' and Pack's readings of these lines I pointed out a more
subtle issue concerning the mere indication of the emotion intended. Here the
combination of vocal gestures indicates two different speaking voices in one
actor's voice: the terror-stricken or astonished and admiring "all" who would
cry out in the imagined situation, and the pensive voice or the person clinging
to the pleasurable fantasy, reporting those cries in the actual situation.
Relating vocal gestures to interpretation seems to be a more complex issue.
In many (or perhaps all) instances it would involve a kind of "Chinese-box" ar
rangement. It would be odd to say: "This discourse consists of one diph
thong". But if we utter "I" in response to a question, say "Who did it?", it
would make sense to say that the diphthong [aj] constitutes a word, which
constitutes a sentence, which constitutes a paragraph, which constitutes a
discourse. Likewise, a falling or rising intonation contour cannot determine the
interpretation of a poem. But, as we have seen throughout the present study,
they may convey, respectively, a sense of self-containedness, closedness,
'finality' on the one hand, and reservations, open-endedness, or 'continuity',
on the other. By the same token, they may indicate a wide range of emotional
attitudes. Now if the sentence in question happens to be the last sentence of a
poem, the falling or rising intonation contour may have a crucial contribution to
poetic closure, or the lack of it. This in itself, is not enough to determine the
interpretation or underlying conception of the poem. But, if we have two
alternative conceptions of a poem, say, "Kubla Khan" as a hypnotic-ecstatic
poem culminating in an emotional peak vs. "Kubla Khan" as a magnificent
fragment fading away at the end, the two kinds of ending may, as we have
seen, imply far-reaching consequences for the general nature or the overall
structure of the poem.
I also discussed Jennings' pronunciation of the word "damsel". On the
phonetic level I pointed out the lengthening of the continuants [m] and [z], and
the "incomplete phonation" of the vowel [æ]. I claimed that the former may be
perceived as lingering, the latter as indicating a tender attitude. I further
suggested that the two may be integrated as the perception of an emotional
attitude which may be characterised as "clinging to" some object, or "longing
for" some reality. Now in Chapter 2 I propounded an interpretation of the
great change of tone and theme that occurs in the last stanza of the poem. The
"factual, detailed, matter-of-fact" presentation of Kubla's earthly paradise,
transferring the semblance of truth to it, is experienced in the perspective of
the last stanza as a kind of Paradise Lost. The earthly paradise with its realistic
plenitude and matter-of-fact details becomes a fleeting vision, very much like a
prenatal or other-worldly experience that the speaker is attempting in vain to
recapture. It is turned into an inaccessible reality, referred to, but beyond direct
apprehension. The sense of yearning conveyed by the vocal gestures in pro
nouncing "damsel" and some other words in Jennings' performance (but not
the others') corroborates this interpretation. Again, it would be odd to say that
the lengthening of the continuants and the incomplete phonation of the vowel
corroborate my interpretation. One must go through the Chinese boxes all the
way.
Postscript
The graph in the middle window of Figure 43 reflects the intonation contour of
the last line of "Kubla Khan" as read by Pack. The graph in the bottom
window reflects my attempt to manipulate its final portion. Such a manipula
tion was impossible twenty five years ago. In my 1977 book I speculated on
how stress maxima in weak positions could be performed rhythmically, in
spite of all. One particularly difficult (and rare) construction consisted of verse
lines ending with a stress maximum in the ninth position, resulting from con-
trastive stress, such as
6. Bright Star! would I were steadfast as thou art (Keats)
or
7. Cannot be damned, alas! why should /be? (Donne)
The closure of these lines is blatantly sabotaged by a stressed syllable in the

ninth (weak) position between two unstressed syllables in strong positions. I
had an intuition that such lines could be performed rhythmically with the help
of a certain kind of falling terminal contour. I propounded the hypothesis that
the performer could compensate for this lack of closure by a falling terminal
intonation contour covering a cadential pitch interval (of a musical third or
fifth). Such a chord sequence of sounds, I argued, might generate a feeling of
stability, of "home coming".
When I was first initiated into the mysteries of speech analysis in 1980 at
the Haskins Laboratories, I was told, to my great dismay, that it was impos
sible to measure on the computer the perceived pitch of speech, or perceived
pitch intervals. The middle (genuine) contour in Figure 43, for instance, ends
with a curly curve on -dajs. There is no way to predict, from the machine's
output, which part of it determines the perceived pitch.
My dream was to manipulate the intonation contour so as to generate
cadential and non-cadential falling terminal contours at the end of verse lines,
and to elicit rhythmicality judgments from listeners. I was told that there was
no chance in the foreseeable future to have an application that would manipu

late intonation. It appeared to be theoretically impossible to control perceived
pitch or pitch intervals. Now, however, it seems that my dream (or part of it)
may come true. After having completed the foregoing study I discovered that
the speech analyser application "Praat" has two features which when com
bined might do just what I dreamed of in 1980. It allows one to manipulate the
pitch contours, yielding credible intonation curves; and pitch units can be set in
Hertzs or semitones. The graph in the lowest window of Figure 43 reflects my
attempt to manipulate the final curly intonation contour into a proper falling
terminal contour. The results may be heard online.
Figure 43 Wavepiot with genuine and manipulated pitch extract of "And drunk
the milk of Paradise" read by Pack. The middle window shows the genuine
contour, the low window the manipulated one.
Ironically enough, by the time I got such control over the manipulation of
intonation contours, I had less use for it than I had hoped. Over the years I
have changed my conception in this respect. In 1980 I was looking for
differences between the intonation contours in normal, everyday speech and
the performance of stress maxima in the ninth position in poetry reading. I
expected to find that in verse lines ending with a stress maximum in the ninth
position, a falling intonation contour covering a musical third or fifth would
elicit more "rhythmical" judgments than non-cadential contours (assumed to be
found in ordinary speech). However, interval judgments by musically trained
listeners indicated that this was not necessarily the case. So, by the time I
published my 1998 book I had changed my expectations. I expected that in or
der to solve the problems arising from metrical deviations, performers would
have recourse to intonational resources not contrasted to, but available in, or-
dinary speech. To render a verse line with a stress maximum in the ninth
position rhythmical, a long-falling terminal contour may suffice, irrespective of
musical cadence. "It is not unusual for contrastive elements", says Chafe
(1994: 61) "to show exaggerated pitch deviation as well as exaggerated
volume". The higher the exaggerated pitch deviation required by the contrastive
stress on "thou" or "I" in the preceding excerpts, the longer the falling terminal
contour made possible. The theory predicts that deviant stresses arouse
tension only if the verse line is properly closed. A stress maximum in the ninth
position prevents proper closure of the line in the tenth (strong) position. A
longer than usual terminal contour may be instrumental in solving the rhythmic
problem posed by the stress maximum in the ninth position: it may
perceptually close the verse line not properly closed by a stressed syllable in
the last (strong) position.
As to the reliability of the measurements of pitch intervals in semitones, I
have my doubts. To be sure, very accurate measurements are possible. The
issue at stake is how these measured intervals correspond to the pitch intervals
perceived by flesh-and-blood listeners. This requires extensive experimen
tation. I believe that we are still far away from the possibility of exerting
rigorous control over cadential and non-cadential terminal contours as experi
mental stimuli.
Norman Cook used the semitones feature of "Praat" in an attempt to point
up certain analogies between major and minor scales in music on the one hand,
and certain moods reflected by intonation contours in ordinary speech on the
other. He, too, seems to have hoped for finding a correspondence between the
diatonic nature of music and intonation. But this did not turn out to be the
case. In this way, however, he got around the discrepancy between the mea
sured and the perceived pitch intervals. Cook worked with a highly competent
team who made the instrumental and statistical analyses for his study. He
summed up his pioneering work as follows:
Strong conclusions cannot be drawn from only one experiment using a

new analytic technique. The experimental evidence presented above is
certainly suggestive, but raises almost as many questions as it answers
concerning the relationship between diatonic harmony and non-scalar
pitch combinations. At the very least, however, these results are clear in
dication that the relative interval substructure in speech intonation should
be investigated—not within the rigid framework of the fixed scales of
diatonic music, but taking important hints from music theory and music
psychology concerning the emotive force of pitch combinations (Cook,
2002: 117-118).
In other words, there is no evidence for diatonic musical scales underlying

speech intonation. But "the relative interval substructure" may be significant.
Analogies between pitch movements, or even the measured pitch intervals in
music and intonation may be illuminating.16 However, as I have said, the

correspondence between the measured and the perceived pitch intervals in
speech intonation should be investigated too.17
In Chapter 3 above we systematically manipulated the rhyme patterns of a
quatrain, and elicited responses to the changing structures. In these experi
ments, however, not all other things were equal: by changing the rhyme pat
tern, we necessarily changed words and images too. In 1980 I hoped systemat
ically to manipulate the pitch intervals of the falling terminal contour of a
deviant line in a similar manner and elicit rhythmicality judgments from sub
jects. The manipulation reported in figure 43 is a major step in this direction.
For the time being we must nevertheless content ourselves with less ambitious
empirical feats. Consider my manipulation of Pack's intonation. In the genuine
reading the final intonation contour curls around 87 to 107 Hz. In the doctored
version it falls to roughly 50 Hz, in a slightly rounded curve. From what we
know about falling and not-falling boundary tones, we may predict, quite
safely, that the genuine version will be perceived as relatively open-ended; the
doctored version as generating a considerable sense of closure. The results of an
empirical test with flesh-and-blood listeners are very unlikely to refute such
predictions. Nevertheless, the manipulation of intonation reflected in Figure 43
may prove illuminating: we obtained two alternative versions that can be com
pared. The indecisive, suspensive tone of one version, and the conclusive,
assertive tone of the other may directly be experienced, while all other things
are really equal.
16
Fónagy (2001: 124-137) investigated in considerable detail the similarities and
contrasts between music and intonation. He found that "in everyday speech,
pure tones and perfect intervals are the exception. In contrast to tonal music,
we find no trace of tonal tension or rules of harmony. In a five-minute con
versation put into musical notes, the dissonant intervals (minor second, aug
mented fourth) predominated, representing 76.4%, as against 23.0% in two
Schubert songs" (Fónagy, 2001: 130).
17
I have asked an outstanding authority in instrumental phonetics, as to the
semitone pitch intervals: has anybody tested whether the perceived semitones
are the same as the measured ones? I received the answer "Yes, this has been
measured extensively in the literature. Try scholar.google.com, and JASA".
Searching for "semitones" turned up about 1,470 items, partly in music, partly
in phonetics. Among the first one hundred items or so I couldn't find a single
item that investigated issues relevant to my question. They typically
investigated such phenomena as that an increase in semitones of the FO
differences between two utterances (ranging from isolated vowels to sentences)
facilitated a wide range of perceptual tasks. These experiments did not try to
ascertain the correspondence between measured and perceived intervals, but
correlations between measured intervals and performance in a different kind of
perceptual task.
Iván Fónagy, a dear friend and mentor, died a few months before I sent the
camera-ready copy of this book to the publisher. In his last days he managed to
have a look into Chapter 4, and made some encouraging comments. During his
long career he had innumberable wonderful insights in a wide range of
disciplines: linguistics, music theory, anthropology, psychology, and literary
theory. Whenever I thought I discovered some new territory, I had to realise that
Iván, like the Vikings, had already preceded me there.
Afterword
Integration and Wider Perspectives
Before I added the third part to this book it was suggested to me that it basical
ly consists of two very loosely connected parts, and that it would be good to
have a final chapter which binds together those two parts, in terms of how the
findings of the two affect or support each other. Later a third part was added
to the book. The purpose of this chapter is to explore how the three parts
affect or support each other.
My claim is that all three parts are governed by similar aesthetic concep
tions and hypotheses, but follow different methodologies. The first part is en
tirely speculative, in the manner prevalent in literary criticism. The other two
parts are empirical, based on tests conducted within different disciplines, on
different aspects of the aesthetic event. The second part explores gestalt
qualities of the text and "reader-response", with methodologies drawn from
Gestalt Theory and experimental Cognitive Psychology. The third part em
ploys Instrumental Phonetics, submitting to instrumental investigation four
commercially-available vocal performances of "Kubla Khan" by leading British
actors. In all three parts I discuss the interaction of the structure of poems
with a performing consciousness. We are faced, however, with two different
kinds of output.
Wellek and Warren argue that a poem is a stratified system of norms, full of
potentialities. Every performance (mental or vocal) realises a subset of these
norms, rendering the abstract system more concrete. Eddie Zemach says that a
work of art is ontologically incomplete. According to Margolis, an interpreta
tion offers a mental schema for imposing unity and coherence on a work of art,
where it is "defective" or incomplete in these respects. All of these characteri
sations refer to critical activities in which the interpretive output is in a
conceptual metalanguage. Critics are not usually aware that a vocal perfor
mance of a poem too fulfils a similar function, namely the concretion of an
ontologically incomplete entity, except that a critic expresses himself in a
metalanguage within a conceptual system whereas an actor actualises the poem
in a stream of speech sounds. Rather than a conceptual metalanguage, actors
use acoustic and phonetic cues to indicate concrete emotions, attitudes, or
rhythmic complexities. This part of their work is subject to immediate per
ception. The critic uses words to draw attention to emotional and rhythmic
qualities generated by the words of the poem. He makes such statements as

"This passage displays a sad emotional quality", or "This verse line contains
three consecutive stressed syllables". The actor, by contrast, utters the actual
words of the poem, using his voice to suggest emotions and rhythmic qualities.
The poem as an abstract system of norms can be accessed only through some
kind of vocal or mental performance. The actor provides an actual vocal
performance of the poem; the critic gives the reader instructions how to per
form it for himself.
In the emerging structure of this book, the theoretical discussion of the poem
and of the implied critic's decision style of the first part become a coatrack
upon which the second and the third parts are hung in different ways. I will,
therefore, have to discuss separately the interaction between the first and third
part of this book, and that between the first and the second part. I will treat
the former only briefly; the latter I will treat at considerable length. In fact, the
latter will serve as the hub for a much wider discussion.
Part I & Part III: Critical and Vocal Performance
As to the vocal actualisation of "Kubla Khan" as a stratified system of norms,

I discussed at considerable length lines 3-5, for instance, where the syntactic
unit conflicts with the versification units: a single clause runs through these
three lines ignoring, as it were, the line endings. According to the received view
such conflict cannot be rendered in vocal performance; the performer must
make up his mind whether he preserves the syntactic or the versification
structure. I have claimed for the past thirty years that both structures can be
conveyed simultaneously by having recourse to conflicting acoustic and
phonetic cues. Three of the four readings analysed here do, indeed, realise the
conflicting patterns by having recourse to conflicting cues, as predicted by the
theory.
"Kubla Khan" is one of the metrically most regular poems in the English lan
guage. The stress and versification patterns tend to converge in it. In generative
terminology, a greater than usual proportion of strong positions is realised in
stressed syllables, and a greater than usual proportion of weak positions in un
stressed syllables. Some of the alliteration patterns occur in stressed syllables
in strong positions, with one or more intervening positions, reinforcing this
convergence. Other alliteration patterns yield an interlacing network which
generates an elusive texture, reinforcing the element of uncertainty. The two
types of alliteration combine to inspire the reader with "false security". These
convergent and divergent sound patterns combine with several other elements
to suggest a conception of "Kubla Khan" as a hypnotic poem. When, in the
beginning, I propounded this conception theoretically, I was thinking in terms
of the gestalts of these configurations, independently of any possible
Afterword 207
performance. In harmony with certain prevalent phonological theories, I

believed that "a stress is a stress", no matter what phonetic cues actualise it.
The foregoing instrumental study revealed that some reciters enhance the
regular hypnotic rhythm by assigning equal stresses to alliterating stressed
syllables in strong positions; others suppress regularity by differentiating
between those stresses. In these instances, the instrumental study of the third
part reveals how performance may support or suppress the hypnotic structure
analysed in the first part.
The vocal performances analysed in part three sometimes support the pro
jected world of the poem too, as I discussed it in part one. Thus, for instance,
in Chapter 2 I interpreted the break between the first thirty six lines of the
poem and the last stanza as a sudden shift from a direct perception of Kubla's
building enterprise to yearning for a lost, inaccessible reality. The Abyssinian
maid in the last stanza, too, is part of an irrecoverable vision. With reference to
the words damsel, dulcimer, and the second occurrence of caves I pointed out a
vocal gesture in Jennings' reading that consists of lingering on continuous
speech sounds and suggests, at the same time, an emotional quality by some
other vocal device. This was perceived as clinging to some object or memory,
yearning, for instance, for an inaccessible reality.
In some instances, however, I pointed out that performances not only sup
ported, but also affected the analysis of the poem in interesting ways.
Consider, for instances, the lines "And all should cry, Beware! Beware! / His
flashing eyes, his floating hair!". Here we find a statement and then a series of
reported exclamations. This observation is trivial enough and, to the best of my
knowledge, no critic has ever bothered to point it out. Nor did I, in part one.
Indeed, two of the performers conflate the two voices into one exclamation.
Two of them, however, generate something of considerable aesthetic interest.
They use their single voice to convey the reporting voice of the speaker and
the reported horrified cry at the same time. In his reading Jennings exploits the
alliteration in "flashing" and "floating" to indicate a terror-stricken emotional
quality, overarticulating these fricatives by three different means, as if the lips
were quivering—in the present context this may be perceived as quivering with
fear. At the same time, in pronouncing the stressed vowels of these words he
deploys vocal gestures similar to those which I mentioned in the preceding
paragraph. As a result, the combination of vocal gestures in this line indicates
two different speaking voices: the terror-stricken "all" who would cry out in
the imagined situation, and the person clinging to the pleasurable fantasy in the
actual situation.
Pack too indicates here two different speaking voices, but by different vocal
gestures and, indeed, he attributes to the reporting voice a different character.
As I pointed out in part three, he uses a falling-rising intonation contour at the
end of the exclamations. The "falling" curves suggest the unreservedness,
decisiveness, 'finality' of the imperatives and exclamations—the unqualified
response of the horrified onlookers in the hypothetical circumstances. The

rising boundary tone conveys a sense of reservation, irresoluteness and open-
endedness associated with a reflective mood. One-and-the-same intonation
contour thus conveys two very different attitudes, of two different voices: the
voice of the wonder-stricken or horrified hypothetical onlookers crying out at
the sight of the ecstatic youth; and the wistful voice of the speaker, reporting
those exclamations in the actual situation.
Our awareness of these two voices in the last stanza affects the discussion
in Chapter 2 in an interesting way. To be sure, I could eventually have arrived
at this conclusion without relying on the performances; but the fact is that I
did not. I had an uneasy feeling when, in Chapter 2,I discussed the issues of
depersonalisation and emotive crescendo. The last stanza, I said, leads to the
ecstatic peak of the emotive crescendo. On the other hand, we noticed a sudden
shift from a depersonalised description in the first thirty-six lines, to an
abundance of personal pronouns in the last stanza. Depersonalisation, in turn,
is said to be characteristic of certain altered states of consciousness resulting in
a loss of selfhood, whereas the abundant use of personal pronouns indicates
sustained self-presence. This observation would contradict the observation
that the last stanza constitutes the ecstatic peak of the poem. I was haunted by
an uneasy feeling that I may have been imposing my models on the text, dis
regarding evidence to the contrary. When the performances attracted my atten
tion to the two voices precisely at the ecstatic (or frenzied) peak, everything
fell into place. The reporting voice suggests the ordinary consciousness of the
perceiving self; but the situation reported is one of frenzy or ecstasy. In
relation to the neurobiological study of consciousness, Antonio Damasio
speaks of
discovering the biological underpinnings for the curious ability we humans

have of constructing, not just the mental patterns of an object—the
images of persons, places, melodies, and of their relationships, in short,
the temporally and spatially integrated mental images of something-to-be-
known—but also the mental patterns which convey, automatically and
naturally, the sense of a self in the act of knowing. [...] The neurobiology
of consciousness faces two problems: the problem of how the movie-in-
the-brain is generated, and the problem of how the brain also generates the
sense that there is an owner and observer for that movie (Damasio, 1999:
11).
We are, of course, interested here in the linguistic, not the neurobiological

correlates of the knowing self. "There is a presence of you in a particular
relationship with some object. If there were no such presence, how would your
thoughts belong to you? Who could tell that they did?" (ibid., 10). The
"movie-in-the-brain" proceeds along the entire first thirty-six lines. But it is
the massive appearance of personal pronouns in the last stanza that "generates
Afterword 209
the sense that there is an owner and observer for that movie", and draws
attention to the absence thereof in the first thirty-six lines. The ongoing
"movie-in-the-brain" is still present in the last stanza; it even constitutes the
culmination of the emotive crescendo. At the same time, it is separated from
the "owner and observer for that movie". The simultaneous presence of the
two voices (and the two kinds of consciousness) in the last stanza may induce
a vivid perception of the disparity between the actual and the yearned for but
unattainable reality.
Part I & Part II: Hypnotic Quality and Cognitive Style

The Reader and Real Readers
I wrote the rest of this chapter, on the mutual support of the first and second
parts of this book, when I did not know yet that there would be a third part
too. In what follows I propose to return to "Kubla Khan" in the light of
experimental evidence—not only the evidence presented in Chapter 3, but also
experimental results of other researchers concerning speech perception and
visual and auditory perception, not necessarily applied to aesthetic issues. As
I suggested above, the first two parts of this book are less loosely connected
than would appear at first sight. In fact, the research of the second part was
deliberately guided by the theories underlying the first part. Chapter 1 explores
the effect of cognitive and personality style on reader response and critical
activities in "Kubla Khan" criticism. Chapter 2 treats "Kubla Khan" as a
hypnotic-ecstatic poem. The research reported in Chapter 3 assesses the effect
of cognitive and personality style on reader response to a text variable (rhyme
pattern) which, according to the theory put forward in the preceding chapters,
is crucial for an appreciation of hypnotic poetry. The personality trait as
sessed is the one used to predict susceptibility to hypnosis. In the third
experiment, the effect of this personality trait on the appreciation of hypnotic
poetry is directly assessed: high-absorption readers are inclined to report more
pleasing, low-absorption readers more displeasing features in a hypnotic poem.
The present chapter attempts to propound an integrated, fine-grained model
of response to hypnotic poetry. The experiments reported in Chapter 3 spot a
crucial stage of reader response, midway in a top-down and bottom-up pro
cess. Persons of various cognitive styles rate certain rhyme patterns differently
on an OPEN ~ CLOSED scale. This may account for their diverse responses to
the overall prosodic structure of hypnotic poetry. By the same token, such
responses may affect their subliminal perception of the rich precategorial audi
tory information associated with speech sounds, and its interaction within or
across gestalt boundaries. This, in turn, may account for their respective
propensity to perceive what some nineteenth-century readers considered the
"verbalised music" of "Kubla Khan".
In my work on critics' decision style I carefully distinguish between the

"implied critic" and the "flesh-and-blood" critic. The former is inferred from
the critical decisions reflected in the latter's writings. I assume that the two are
loosely related: neither totally separate, nor totally identical. In a Hebrew pa
per I applied the theoretical battery of Chapter 1 above to a corpus of criticism
on a corpus of fiction designated as "Literature of Extreme Situations". I drew
attention to a Hebrew critic who quite consistently displayed critical decisions
characteristical of "the quest for certitude". A psychologist wrote to me with
reference to this paper: "What would you do if you had a chance to test that
critic's empirical personality, and found it flexible, non-authoritarian, tolerant
of ambiguity, and so forth?" Well, I would be in considerable trouble. It is,
however, impracticable to ask this critic to take a full battery of clinical tests. I
could hide behind that impracticability—just like Don Quixote, who did not
put to test the helmet he had prepared, because he trusted it would endure.
Our experiments, reported in Chapter 3, clearly established, with reference to
one personality trait at least, systematic relationships between personality
style and literary response, in the expected direction.
My first exposition of the two critical attitudes (negative capability and
quest for certitude) was published in College English (then edited by Richard
Ohmann), in a special issue subtitled "The Reader and Real Readers", a dicho
tomy which I adopted here. In Chapter 3, above, I suggested that in much
criticism "the reader", or "the qualified reader", or "the competent reader"
means "a hypothetical reader whose response conforms with my analysis". In
the final resort, however, there is no escape from trying to find out whether
there is any correlation between personality style and real readers' response to
a piece of literature.
At the outset I propose to make a comment on my preoccupation with per
ceived effects in light of the relationship between aesthetic response and
experimental rigour. The most unique feature of my work is, perhaps, my
attempt to relate consistently perceived effects to poetic structures. I also
claim that such effects cannot be inferred, only directly perceived, as one
directly perceives that "the book is red" or "the tea is sweet" (cf. Sibley,
1962). This, however, requires the researcher to assess what flesh-and-blood
readers perceive in works of literature. It would be nice to use a stop watch
with reference to perceived effects, too, as in the assessment of metaphor
understanding. Aesthetic objects have extremely complex structures, and the
proper response to them must be a response to a "regional quality" emerging
from the interaction of a wide variety of elements. Different readers may utilise
different subsets of these elements, weight differently the same elements, and
organise them in different structures. I have called such activities "mental per
formance" (a rudimentary example of varying mental performances can be
found in Chapter 1, apropos of my discussion of Elizabeth Schneider's
handling of the sound patterns of "Kubla Khan"). These organisations cannot
Afterword 211
be predicted from the poem's structure alone. Different mental performances

yield different perceived effects. And conversely, we must rely on the
perceived effects reported by readers to infer their respective mental perfor
mances. The isolation of a single element in a controlled experiment may, there
fore, oversimplify the aesthetic issue. The response to a complex aesthetic
object would, therefore, require some less straightforward ways of appraisal.
Theorising and Experimenting
As we suggested in Chapter 3, this poses a communication difficulty between

experimenter and "experimental subjects". If the experimenter formulates the
experimental task in general terms, he cannot be sure to what aspect of the
work his respondent would respond. On the other hand, if he puts the task in
more specific terms, he may predetermine the results of his experiment by
suggesting the expected answer. Consider, for instance, the experiments report
ed in Chapter 3. When I told a friend of mine about them, he expressed his
wish to answer the questionnaire. His OPEN-CLOSED ratings made absolutely
no sense to me. It turned out that he was not judging "poetic closure", but
whether a rhyme word ended with an open or closed syllable. In this particular
case, the solution seemed to be quite simple: we added at the top of the
questionnaire, without any comments, Barbara Herrnstein-Smith's definition
of "poetic closure". This solution was feasible only because we did not try to
assess the man-in-the-street's response, but that of respondents with consider
able academic experience.
A situation thus arises in which the more rigorous the experimental set-up,
the more trivial the experimental task must be. I therefore had to resort to less
formal modes of appraisal. I relied on "the received view", on what critics and
theoreticians wrote in their published works, on spontaneous responses of my
students in the classroom, and on my own introspection. This is not very
rigorous, nor very respectable from a scientific point of view, but, I claim, may
yield significant insights inaccessible to more rigorous methods. Still, there is
some evidence that experimental researchers found my work fairly useful. To
quote just one example, Mario Borillo (Institut de Recherche en Informatique
de Toulouse, CRNS et Université Paul-Sabatier) organised an international, in
terdisciplinary workshop on "Cognition and Creation" at the Centre Culturel
de L'Albigeois, Albi (12-14 January, 1995). To the participants he sent a
"Manifesto", in which he wrote:
Even though few cognitive empirical inquiries were dedicated so far to

artistic creation, philosophic thought, formal analysis and cognitive
psychology have mapped the way to the next stage of systematic
experimental study: I would mention for instance the questions raised by
Nelson Goodman about the nature of perception, symbolic languages and
the work of art, by Roger Shepard about the visual anomalies and their role
in the visual arts, Reuven Tsur's attempt at founding a cognitive version
of poetics, Philip Johnson-Laird's suggested theory of musical creation,
Ray Jackendoff's investigation of relations between musical forms and
affects.
Speaking for myself, my own "way to the next stage of systematic experi
mental study" was guided by a methodology which, in retrospect, seems to me
quite clear. As I claimed in Chapter 1, theoretical models or frameworks cannot
be directly applied to works of literature; they must be applied to texts via
critical terms that are properly articulated and have considerable descriptive
contents. Critical terms, on the other hand, with their descriptive contents,
have little significance unless they are understood in relation to a theoretical
framework or model.1 A critical statement can be true, and still trivial and de
void of interest, unless a theoretical framework or model imputes significance
to it (cf. Tsur, 1992a: 500-535). I suggested that a metalinguistic statement like
This poem contains 41 lines may be perfectly true but trivial from the aesthetic
point of view. On the other hand, inverting the digits, a statement like This
poem contains 14 lines may be regarded as meaningful from the poetic point of
view, when viewed within the theoretical model of the sonnet form. Experi
mental research may be regarded as a particular instance of this principle. The
empirical data yielded by experiments are trivial in themselves. They become
significant only when a hypothesis (based on a theory or theoretical model) is
applied to them. In this respect, it makes little difference whether one inter
prets one's own or other people's experimental results. But conducting one's
own experiments has the advantage of being able to generate data relevant to
one's inquiry, or putting one's hypotheses to further test. Ideally, we ought to
have tested all the speculations propounded below.
Accounting for the effects of hypnotic poems requires, of course, more
complex descriptive terms than counting verse lines, and theoretical models
that are more sophisticated than the sonnet form. Consider Figures 1-2 in
Chapter 1.I adapted them from Elizabeth Schneider, who uses them to demon
strate the richness and intricate interlacing of "Kubla Khan'"s sound patterns.
In terms of the foregoing, we might say that she used critical terms with con
siderable descriptive contents (including two diagrams) to point out those
sound patterns. These descriptive statements become aesthetically significant
in view of a commonsense psychological theory, partly explicit in her work,
partly implied, that can be stated as follows: "Intricately 'overlaid or interwo
ven' perceptual patterns tend to blur each other, yielding a fine backgrounded
texture; in verbal art this is perceived as 'subtle harmonies', commonly associ
ated with Milton and Bridges". This quality, in turn, is justified by an implicit
1
Prof. Marcelo Dascal comments that this is, in fact, the case of any theore
tical and even observational term: all observation is theory laden.
Afterword 213
commonsense theory at the aesthetic level:—"Such subtle harmonies are a good-

making feature of poems". I went two steps further. First, I applied to the
sound patterns of this poem Chatman's distinction between a Pope-type allit
eration (which cooperates with metre and syntax) and a Donne-type allitera
tion (which runs counter to metre and syntax). Secondly, I integrated the ef
fects of these two types of sound patterns within a model of hypnotic poetry.
According to this model, hypnotic effects in poetry crucially depend on a
unique balance of elements that generate a strong effect of certainty (e.g., regu
lar metric events) and elements that generate a strong effect of uncertainty (e.g.,
irregular or uncertain grouping of rhymes, and the blurring of rhymes by run-on
sentences). As I argued in Chapter 2, Pope-type alliterations reinforce the
regular events, contributing to the effect of certainty; Donne-type alliterations
blur the regular events and diffuse the focus of perception, contributing to an
effect of uncertainty.
I further speculated that certain personality types would have difficulties in
responding to the uncertainty effects of such poems. I expounded this concep
tion long before I could hope for empirical testing. But our experiments
supported much of these speculations. On the textual level they tested only
one link in this complex: response to the various groupings of rhymes. We ob
tained conflicting results; but these could be systematically accounted for with
reference to a distinction between high- and low-absorption respondents. The
findings unambiguously supported my speculations that various personality
types respond differently to the uncertainty effects of certain poems. Since we
used a personality trait that was devised to predict susceptibility to hypnosis,
one could conjecture that our findings might be relevant to the response to
hypnotic poems (though it was found relevant to non-hypnotic poems as
well). This conjecture was supported by our finding that low-absorption
respondents (unlike their high-absorption counterparts) tended to judge a puta
tive hypnotic poem boring (rather than interesting). One possible explanation
for such judgments is that low-absorption respondents realise the monotonous
regular metric events better than the irregular or uncertain grouping of rhymes,
upsetting the balance required for experiencing a poem as hypnotic.
The aims of experimental work must be controlled by exceptionally well-
defined, operational terms. These usually involve very simple, statistically
quantifiable steps. The results of the experiments presented in Chapter 3 are
due to the unique encounter of a speculatively elaborated theory that already
involved operationally defined descriptive terms as presented in Chapters 1
and 2, with a more rigorous experimental approach that contributed an addi
tional set of quantifiable terms. I must confess that I was very skeptical when
Joseph Glicksohn proposed his experimental design, and doubted very much
that anything aesthetically meaningful might come out of it. Eventually I had to
admit that the above principle (that descriptive terms receive their significance
from some theoretical framework) did wonders in the present instance too.
This became possible only by joining together two very different approaches
of scholars with vastly different academic skills. On the one hand, I had been
elaborating theoretical discussions in great detail for years with no hope for
ever putting them to empirical test. On the other hand, Joseph Glicksohn
brought to the joint adventure brilliant experimental aptitude and proficiency.
This was manifest not only in the experimental design, but he also proved a
wizard in interpreting the experimental results, even with reference to unfore
seen questions, after the event.
We began our experimental work after I had elaborated a theory of hypnotic
poetry and applied it in great detail to "Kubla Khan" and a range of Hebrew
poems. Among other features, hypnotic poetry is characterised by obtrusive
rhythms which direct attention away from the poem's contents to its sound.
In this respect, one must distinguish between several poetic effects regularly
associated with regular rhythms. Predictable, regular events may induce
security or boredom depending, in some instances at least, on the reader's
temporary or permanent predisposition. Certain textual distinctions, in turn,
may explain that more than usually regular metric structures can have three
different perceived effects: witty, hypnotic, and an effect of simplified mas
tery of reality (as in nursery rhymes); but, as we found in our experiment, they
may also sound boring, monotonous to some readers (note that the first three
terms have considerable descriptive contents; in the latter terms the evaluative
ingredient is stronger than possible descriptive ingredients).
As to the three kinds of effect above, they are based on a sense of security
induced by predictable, regular events. The source of the difference between
them is that the hypnotic quality (unlike the witty and simplifying qualities) is
due to false security. In hypnotic poetry, regular metre may induce security,
so as to allow the reader to contemplate the most irrational contents and
qualities conveyed by poems. On the prosodic level, too, at the rank-above-
the-line, insecurity may be enhanced by rhyme patterns. In this respect, both
irregularity and indefinite grouping of rhyme sequences may increase inse
curity. The present book is devoted to "Kubla Khan", in whose rhyme struc
ture possible good shapes are blurred by a variety of means. In our experiment,
by contrast, we used a Hebrew poem we judged hypnotic, Ratosh's "Bircat
Shoshanim" ("Greeting with Roses"), rhymed throughout on the same rhyme
sounds (monorhyme). "Shape", says Meyer (1956: 161) "may be regarded as a
kind of stylistic 'mean' lying between the extremes of chaotic overdiffer
entiation and primordial homogeneity". The rhyme patterns of the two poems
mentioned deviate in opposite directions from "good shape", but both render
the security induced by the obtrusive rhythms of these poems false. The
events that constitute the verse line are more than usually regular, whereas the
grouping of lines within the units-above is either unpredictable or arbitrary,
respectively. My evidence for all this was based on gestalt theory, metrical
theory, the analysis of poetic texts, and introspection. Our controlled experi-
Afterword 215
ments in Chapter 3 supported my speculations, in a way that is by no means

trivial.
The foregoing characterisation implies that my version of hypnotic poetry
crucially depends on regarding the irregular or blurred rhyme patterns of
"Kubla Khan" and the "primordial homogeneity" of the all too regular, undiffe
rentiated monorhymed sequence of verse lines in Ratosh's "Bircat Shoshanim"
as somehow equivalent. The question was what do "real readers" think. Our
experiments did, indeed, confirm the functional equivalence between mono-
rhymes and irregular rhyme patterns postulated above. It is most remarkable
that this equivalence is valid for both low- and high-absorption respondents—
in opposite ways, though.
Figure 1 Version x Absorption interaction for OPEN-CLOSED ratings
As reported in Chapter 3, we manipulated the rhyme pattern of one of

Omar Khayyam's Rubáiyáts, obtaining three versions, with aaaa, aabb, and
aaba rhyme patterns. The Rubáiyát was chosen not because it is an instance of
hypnotic poetry (it is not), But rather as a relatively simple text of
considerable aesthetic merit, in which controlled systematic changes could be
made. This setup was derived from my speculative book on Hebrew hypnotic
poetry (Tsur, 1988), where it was used in a thought experiment, that explored
certain processes of aesthetic response deemed relevant to the response to
hypnotic poetry. It reflects, on a miniature scale, the issues related to rhyme
patterns mentioned in the preceding two paragraphs. Our real experiment was
designed to find out whether competent real readers do indeed respond as
predicted by the thought experiment. What we found confirmed our expecta-
tions, but revealed a much more complex pattern which proved to be full of
further possibilities.
This "real experiment" must be regarded as a pilot study. There are, in fact,
some methodological problems in our procedure. When one changes the rhyme
pattern of a quatrain, one cannot avoid making other changes as well, on the
syntactic, semantic and thematic levels. So, in theory at least, we could not be
sure whether our subjects responded to changes in rhyme patterns or to some
other concomitant changes. To solve this problem we ought to have repeated
the experiment with a great number of Rubáiyáts, with comparable batteries of
respondents. However, as we commented in Chapter 3, some groups of our
subjects responded in our empirical tests to this English text, some to its
Hebrew translation by Jabotinsky. Though this was a matter of convenience
rather than deliberate planning, it turned out to be convincing evidence that we
were dealing with genuine Gestalt organisation and not with language-specific
accidents. Thus, for instance, the possibility that some of our English respon
dents treated "Caravanserai" and "Day" as rhymes for the ear and some as
rhymes merely for the eye did not affect our experimental results as far as
language differences were concerned. Or, consider the third line in the aaba
(original) version, and the aabb version. I claim that in the former it is
perceived as deviation from, or intrusion upon, a sequence of a-rhymes,
whereas in the latter as the onset of a new perceptual unit that repeats the ini
tial couplet pattern. The generation of a credible aabb version forced us to
make some additional changes in the English text, so that we could not be
certain anymore as to which changes our English-speaking respondents re
sponded to. In the Hebrew texts, however, the third line is letter-by-letter
identical in the two versions. Thus, since there were no significant differences
between our results with the English and the Hebrew texts, we could be pretty
certain that our results referred to gestalt groupings of rhymes.
Consider the issue of poetic closure, and the OPEN~CLOSED ratings of the
three versions. In the aaaa and aaba rhyme structures the "stimulus control
[is] weak enough to leave the observer with a margin of freedom". The aabb
rhyme pattern has much stronger gestalt than the other two; so the stimulus
exerts much more rigorous control over readers' response. We expected low-
absorption respondents to judge all three versions as more closed than high-ab
sorption respondents would; and high-absorption respondents to judge them
as more open. Indeed, this was the case. But, to our great surprise, their
judgments yielded two V-shaped graphs, one of them inverted (see Figure 1).
The aabb version with its symmetrical, strong shape exerts relatively rigorous
control over readers' mental performance; hence the small difference (in the
expected direction) between the responses of low- and high-absorption
subjects. The aaaa and aaba rhyme patterns, by contrast, have weaker, more
ambiguous shapes, for different reasons. As a result, they do not resist mental
manipulations, and are perceived as relatively closed by one group, and
Afterword 217
relatively open by the other. At the same time, these versions are equivalent for
both groups in that stimulus control is kept weak enough to leave the observer
with a margin of freedom to perform them as "open" or as having "strong
closure"—as their respective cognitive styles may demand. Since subjects
received no instruction as to what aspects of the text they were requested to
respond to, each of them seems to have sought out and emphasised the aspect
that most suited his or her absorption style. High-absorption subjects seem to
have sought out the more fluid, less stable aspects, low-absorption subjects
seem to have sought out relatively firm, stable aspects. In Chapter 3 we point
out these specific aspects.
This result has two facets. First, it supports the speculation that both un
predictable grouping (aaba) and indefinite grouping (aaaa) of verse lines are
less determinate than rhyme patterns that yield strong, symmetrical gestalt
(aabb). Second, this indeterminacy may be exploited in the service of the
emotional needs of different personality styles. Thus, for instance, stable
closures imputed by low-absorption respondents on these structures may
inspire them with a relatively strong sense of security, rendering the security
induced by the rhythm of the poem genuine. From this analysis of our results
with rhyme patterns and absorption style we may infer that respondents in
clined toward such closure-judgments will perceive hypnotic poems as regular
at the lower rank, but will be less prone to perceive the unit-above-the-line as
open. Consequently, they will tend to judge certain poems with exceptionally
regular rhythms boring rather than hypnotic. Our experimental results with a
poem which we considered trance-inductive amply confirmed these predic
tions. Low-absorption respondents judged it more "boring", whereas high-
absorption respondents judged it more "interesting". Now note this. The ab
sorption scale was originally devised to predict susceptibility to hypnosis. So,
our commonsense expectation would be that high-absorption respondents
would resort more frequently to a "hypnotic" mental performance of poems
and would pass a more favourable judgment on hypnotic poems than low-
absorption respondents. Our finding that this is indeed the case is not, there
fore, very revolutionary. Still, it is quite reassuring. What is more, when exa
mined in the wider perspective of what we know about speech perception and
about the interaction of gestalts with gestalt-free elements, these findings may
also yield an insight into the phonetic source of the difference between low-
and high-absorption respondents' response to hypnotic poetry.
Hypnotic Poetry and Precategorial Sensory Information
I will explore this with reference to empirical evidence produced by other re
searchers, in widely different contexts. I have suggested that hypnotic poems
tend to shift attention away from their contents to their music. As reported in
218 "Kubla Bulan"—Poetic Structure, Hypnotic Quality & Cognitive Style
the Introduction, some contemporaries of Coleridge regarded "Kubla Khan",

for better or worse, as "verbalised music". The foregoing discussion explored
the gestalt grouping of abstract phonetic categories, but practically ignored the
auditory information that conveys those categories. In what follows, I will ex
plore the cognitive mechanisms underlying the ability to appreciate hypnotic-
ecstatic poems as "verbalised music". I have already alluded to these mechan
isms in Chapter 1, in my discussion of Schneider's treatment of alliteration. As
speech research of the past fifty years or so has demonstrated, the rich pre-
categorial auditory information that communicates the phonetic categories is
restructured in the speech mode of aural perception and becomes inaccessible
to conscious introspection. I have been arguing (e.g., Tsur, 1992b; 2001) that
there is some experimental evidence that in what I have called "the poetic mode
of speech perception", some of that information becomes subliminally avail
able; I claim, in fact, that it is crucial for the perception of musicality in verse.
As we will see, the diffuse precategorial auditory information that carries the
speech categories is also indispensable for the performance of certain non-aes
thetic memory tasks. As gestalt theory and the Rorschach inkblot test suggest,
in some extreme instances it may also be crucial for the perception of an emo
tional quality in verse.
Elsewhere (Tsur, 1992a: 113-114) I pointed out that both in gestalt theory
and the Rorschach inkblot test, good shapes are associated with the rational
faculty, gestalt-free elements with emotional qualities. In the Rorschach inkblot
test, form responses indicate an ability to exercise strong intellectual control,
whereas the coloured component is regarded as an affectively charged stimulus.
A large number of colour responses indicates emotional excitability (Ror
schach, 1951: 99). "Although much effort has been expended in experimental
procedures designed to disprove this theory, it has been supported to a
remarkable degree" (Alcock, 1963: 54). A special case in the Rorschach test is
the integration, not merely the co-presence of strong gestalts with gestalt-free
qualities. This is particularly conspicuous in the Form/ Diffusion (FK) score,
for instance, which refers to a combination of good form perception with
shading, when "the shading is used to present diffusion, darkness, or depth"
(Alcock, 1963: 48). I claim that the concurrent perception of phonetic catego
ries and precategorial auditory information, too, may, in certain circumstances,
be similar to such a Form/Diffusion response. Alcock claims that scoring well
on FK is a difficult achievement. It involves good form realisation combined
with shading organised to present an impression of perspective or other three-
dimensional quality. It implies an ability to exercise intellectual control when
faced with stimuli that are commonly found to be disturbing, such as the light-
dark amorphous shadings of the blots. It is as if the perceiver, in boldly orga
nising this shading as a vista, demonstrated that he can contemplate the powers
of darkness with equanimity. I suggest that, likewise, the realisation of a
Afterword 219
hypnotic poem is a difficult achievement, for similar reasons, and reflects an

ability to "contemplate the powers of darkness with equanimity".
As we sae in Chapter 1, some critics (e.g., Schneider) emphasise gestalt-free,
meaningless qualities generated by interlacing sound patterns; and some (e.g.,
Fleissner) direct attention away from the sound dimension to meaning. In this
respect, there are considerable differences between critics. In the next few
paragraphs I will argue that persons with certain personality styles, those
related to the quest for certitude, are less prone to perceive the rich precatego-
rial auditory information than other persons, due to the combination of two
different reasons: owing to reluctance to notice the unique, unclassifiable sensa
tion; and to the perceptual dynamics of gestalts and gestalt-free information (in
the light of the cognitive processes concerning closure discovered in our
experiments). As to the first reason, in Chapter 1 I quoted Richard Ohmann:
The leveler is more anxious to categorize sensations and less willing to

give up a category once he has established it. Red is red, and there's an end
on't. He levels (suppresses) differences and emphasizes similarities in the
interest of perceptual stability. For him the unique, unclassifiable sensation
is particularly offensive, while the sharpener at least tolerates such anoma
lies, and may actually seek out ambiguity and variability of classification
(Ohmann, 1970: 231).
I claim that this "anxiousness to categorize sensations" and the intolerance

of unique, unclassifiable sensations applies not only to such gross perceptual
categories as "red", but also in the domain of specific phonetic categories and
subliminal precategorial sensory information. As to the second reason, my
earlier work on rhyme (based on experimental work done by others in the
phonetic laboratory) may illuminate the different responses of low- and high-
absorption respondents to hypnotic poetry as discussed above.
When we emit strings of speech sounds we obviously have recourse to pho
netic coding. It is much less obvious that phonetic coding is a major resource in
the performance of a wide range of cognitive operations. Liberman et al.
(1972), for instance, interpret the results of a series of experiments by Crow-
der and Morton (1969) as follows: "In the experiment by Crowder, we may
suppose that these same auditory characteristics [...], held for several seconds
in an echoic sensory register, provide the subjects with rich, precategorical
information that enables him to recall the most recently presented items with
relative ease". Liberman, and Crowder and Morton were concerned with dis
tinguishing between phonetic categories which are more, and those which are
less thoroughly restructured. I am going to distinguish, by contrast, between
those language users who are more, and those who are less inclined to rely on
precategorial sensory information, or to perform thorough restructuring. Else
where I have claimed (e.g., Tsur, 1992b) that more and less thoroughly
restructured speech sounds have different emotional effects.2 Here I am claim

ing that an aesthetic context may direct attention from the desired results of the
cognitive task away to the "rich, precategorial information"—the cognitive
device is turned to aesthetic ends. In terms of Tellegen's distinction discussed
in Chapter 3, this involves a switch from instrumental to experiential set, the
latter being "a state of receptivity or openness to undergo whatever
experiential events, sensory or imaginai, that may occur, with a tendency to
dwell on, rather than go beyond, the experiences themselves and the objects
they represent"; in the former, "input from receptors is not used to enhance
experiencing but to make the discrimination needed for guiding instrumental
acts and evaluating achievements". In Chapter 1 we encountered Fleissner who,
most conspicuously, displays little tendency to dwell on the sound texture of
the poem "to enhance experiencing", but rather to "go beyond the experiences
themselves and the objects they represent". Schneider displays an opposite
tendency.
Researchers at the Haskins Laboratories (e. g., Liberman and Mann, 1981:
128-129; Brady et. al., 1983: 349-355; Mann, 1984: 1-10) in their study on the
possible causes of some children's difficulty to learn to read have revealed a
deficiency in the use of phonetic representation by poor readers; good readers,
by contrast, seem to make an excellent use of it. In one experimental task, poor
readers had greater difficulty than good readers in tapping once or three times
in response to the number of syllables in such spoken words as pig or
elephant, or once, twice or three times in response to the number of speech
sounds in such words as eye, pie or spy. This has been interpreted as a deficien
cy in the use of phonetic representation. In light of the Crowder-Morton expe
riment mentioned above, we may assume that rich, precategorial information,
held for several seconds in an echoic sensory register, renders the phonological
categories available longer to efficient readers for conscious processing than to
poor readers.
In another task, respondents had to memorise groups of words—either
rhymed or unrhymed—as in the following:
(1) chain train brain rain pain

cat fly score meat scale
Sibilants like [s] and [ʃ], for instance, are frequently used to hush a crying baby,
or as an imitation of natural noises. The hushing potential is due to the rich
precategorial sensory information available when uttering, e.g., sh-sh-sh-sh-sh.
The precategorial sensory information consists, in turn, of irregular noises,
generating the noise-imitation potential of these speech sounds. The voiceless
plosives [p, t, k], by contrast, are thoroughly restructured, and are frequently
associated with a hard, or aggressive quality.
Afterword 221
Good readers did consistently better with both kinds of groups than poor
readers. However, with the rhymed groups, their performance seriously deteri
orated. While their reliance on phonetic representation increased their overall
performance, the similar sounds of the rhyming words seem to have caused
confusion in their acoustic memory. Thus, this experiment reveals an intimate
relationship between rhyme and the cognitive mechanisms involved in certain
memory tasks. Good readers made efficient use of phonetic representation. We
are confronted with two mental strategies: one in which the "precategorial"
auditory characteristics of the speech signal can be preserved for a while, and
one in which it cannot. Since the poor readers made inefficient use of the
acoustic information in short-term memory, they were not penalised by the
similar sounds of the rhyming words.
Crowder and Wagner (1992: 228-230) summarise a related experiment by
Byrne and Shea, devised to discover what representation, if not phonetic, poor
readers make use of. Subjects had to take a "reading test", reading out lists of
words and then were given a surprise memory test, in which they were pre
sented with the words read earlier, interspersed with a number of additional
words, to which they had to respond "old" or "new". The "new" words were
either phonetically or semantically related to the "old" words. "Assume the
prior items were home and carpet: house and rug would be the semantically
similar foils and comb and market would be the phonetically similar foils".
Good readers tended to confuse both phonetically and semantically related
words, poor readers only semantically related words. Crowder and Wagner in
sist: "The fact that poor readers seem to be using 'too much' meaningful pro
cessing does not imply that they are better at top-down processing than good
readers. It is just that they may be so deficient in bottom-up processing they
have no other recourse".
Professors of literature and professional critics would certainly count as
very efficient readers by the criteria of the foregoing experiments. Moreover,
they are usually well-trained to handle the phonetic component in poetry. The
more remarkable it is, therefore, that one may discern consistent differences in
their handling of the phonetic component relative to the thematic and semantic
elements (see the section Alliteration, Onomatopoeia and Decision Style in
Chapter 1). The afore-mentioned experiments reveal a cognitive mechanism
that may account for some effects of the sound patterns of poetry on the one
hand, and for individual differences in handling them, on the other. Close
scrutiny of their published work reveals that some of these professors and
critics rely more, others less, on phonetic representation (cf. also Tsur, 2003:
215-222). This difference as we know it partly reflects the critic's decision
style, partly his/her relative reliance on phonetic representation.3 Since I have
3
Characteristically enough, an anonymous reviewer of the two-parts version of
this book commented on it: "The high point of the discussion is the painstak
ingly attentive analysis of prosodic structure in the poem", and suggested that I
not yet worked out a methodology to investigate this, I will bracket the issue,
and go on to explore some of the other aspects of the problem.
My approach to cognitive poetics assumes that in the response to poetry
cognitive or adaptive devices are turned to aesthetic ends. What in the pre
ceding experiments is described as acoustic confusion, in the aesthetic context
of rhyme patterns becomes the musical fusion of rich precategorial sensory
information in the rhyming words. I submit that the OPEN~CLOSED ratings
discussed earlier are relevant here. The strong shapes imputed to the text in the
low-absorption respondents' reading impede this fusion, whereas the weak
shapes tolerated by high-absorption respondents are more favourable to it.
When the reader of a poem displays tolerance for the weakening of interme
diary shapes by run-on lines or an asymmetrical rhyme pattern, that is, when
the control of strong shapes over the cognitive system is loosened, the rich
precategorial sensory information reverberates in echoic memory, and the
whole is perceived as more spacious, more plastic, having a fuller body. From
what we know about closed gestalts, the interaction of gestalt-free sensory
information across clear boundaries is impeded. Let me quote just one piece of
empirical evidence.
When Chevreuil, in the early nineteenth century, made artists familiar

with colour induction, he did not touch directly on the problem of form
and colour. The experiment which demonstrated interaction most clearly
was to place a small grey square on a large ground of colour. On a green
ground the grey square would turn a distinct pink. Obviously the more
saturated the surrounding green, the stronger was the induced pink in the
square. A few years later a most paradoxical phenomenon was observed;
when a sheet of semi-transparent tissue paper was placed over the whole
area the saturation of the green ground was of course severely diminished.
One would have expected that the colour induction in the grey square
would also become much paler. But the opposite happened: the pinkness
of the grey square became more pronounced. Many years passed until no
less a man than the great Helmholtz gave the trite explanation of the
paradox. The tissue paper made the outline of the grey square fuzzier and
this weakening of its form increased colour interaction. A more impressive
documentation of the overriding importance of line and form could
scarcely be imagined. A comparatively crude weakening of the line was
sufficient to compensate—indeed more than compensate—for the
enormous loss in the saturation of colours (Ehrenzweig, 1970: 170-171).
extend it to intonation (from which the third part was born). Another review
er, by contrast, commented: "A shorter book would eliminate one of the faults
of the manuscript at present—its tediousness in attending to details of prosodic
response". Now irrespective of the merits of what I wrote, the two reviewers
obviously have different thresholds of tolerance for attending to the phonetic
component of poetry.
Afterword 223
Ehrenzweig further elaborates on this finding:
We can summarise therefore that colour-interaction between figure and

ground stands in inverse proportion to the good gestalt of the figure [...]
the ambiguity of a weak figure on a strong ground immensely increases
colour interaction. Gestalt psychology, without the help of Helmholtz's
discovery, stated independently that colour interaction increases within
the boundaries of a good gestalt while it is inhibited across its boundaries
(ibid., 172).
Colour interaction, in turn, may strongly affect the perceived quality of

outlines.
As in all relationship between form and colour the reverse effect can also
happen. Strong colour interaction tends to make sharp outlines seem much
softer than they are; it levels down differences in tone (ibid., 171).
Ehrenzweig's foregoing examples concern the visual mode. He also pointed

out similar dynamics in "true polyphony" in music: "to the extent to which a
musical note is fitted into a clean melodic 'line' it is prevented from fusing into
harmonic tone 'colour'; conversely, a strong chord will temporarily fuse the
loose strands of polyphony into solid tone colour so that the separate melodic
lines disappear altogether. I have mentioned that the ear constantly oscillates
between the harmonic fusion and polyphonic separation of the melodic lines;
this conflict between 'form' and 'colour' belongs to the very life of music"
(Ehrenzweig, ibid: 173). What I have elsewhere called "convergent" and
"divergent" style displays, with the necessary changes, exactly the same kind
of dynamics. In the former, stressed syllables typically converge with strong,
unstressed syllables with weak positions; alliterating sound patterns typically
converge with stressed syllables in strong positions; sentence endings typically
converge with line endings, yielding a clear versification line. In the latter there
is a tendency of the same elements to diverge and blur each other, very much
like "the loose strands of polyphony" in a "strong chord". As I mentioned in
Chapter 1, Schneider pointed out similar dynamics of such convergent allitera
tion patterns as "ceaseless turmoil seething," "mazy motion," "river ran" etc.,
which are "partly concealed by the interlacing of other patterns" (she even
uses the same verb as Ehrenzweig, "oscillate").
Gestalt Theory has explored the conditions that maximise our tendency to
perceive a stimulus as a coherent whole (even if it is made up of discontinuous
parts). "Good" or "strong" shapes are those which maximise this tendency.
When the perceived patterns fall short of meeting these conditions, the shapes
become correspondingly "poorer" or "weaker". Some of the key terms are
"symmetry", "closure", "grouping", "good continuation", "good return",
"predictability", "similarity", "proximity", "figure-ground relationship", etc4.

These terms have well-articulated descriptive contents. The terms "good" or
"strong", "poor" or "weak" do not suggest here "aesthetically good" or "aes
thetically poor"; in fact, as in the present instance, "poor" or "weak" shapes
may be vehicles for certain hard-to-achieve aesthetic effects. Good shapes are
generated not only by closed areas indicated by continuous outlines (as, e.g.,
Chevreuil's "small grey square on a large ground of colour"). A series of
discrete stimuli too may be perceived as good shape, provided that they meet
the conditions that maximise our tendency to respond to groups of individual
stimuli as unified percepts. This holds true for music and the visual arts as well
as for the verbal signs of poetry. In this sense, rhyme patterns may display
stronger or weaker shapes.
How can we know whether similar dynamics occur in speech perception as
well? There is good experimental evidence that the tendency to perceive certain
rhythmic phenomena applies within intonation contours, but not across
contour boundaries (cf. Tsur, 1998: 300-301). C. I. Darwin and A. Donovan of
Sussex University have demonstrated this in a series of brilliant experiments
(1979, two papers and demo tape5).
Translating Ehrenzweigs terms into the terms of our foregoing discussion, an
inclination to close the shapes of e.g. verse lines may obstruct the fusion of
rich precategorial sensory information in rhyming words, whereas an inclina
tion to leave them open will tend to promote it. Interaction or musical fusion in
the present context suggest precategorial information in a relatively diffuse and
activated state. One might suggest that such interaction across (weak) gestalt
boundaries generates the musical fusion of diffuse, inarticulate, precategorial
auditory information. Consequently, low- and high-absorption readers may
experience hypnotic poems differently at the subphonemic level as well. The
4
These conditions are surprisingly similar in the visual and the aural mode. One
can take Arnheim's Art and Visual Perception and Meyer's Emotion and
Meaning in Music, and compare them condition by condition. Cooper and
Meyer applied these principles also to The Rhythmic Structure of Music;
Ehrenzweig (1965; 1970) applied them to both music and the visual arts;
Barbara Herrnstein-Smith applied them to poetry in her ground-breaking book
Poetic Closure. And I myself applied them to poetic rhythm (e.g., Tsur 1977;
1998).
5
The gist of the experiment and the files with the sound stimuli are available
online at
http://www.tau.ac.il/~tsurxx/Rhythm_Book_mp/Chapter9b_Sound_Files.html
and
http://www.tau.ac.il/~tsurxx/Rhythm_Book_ai/Chapter9b_Sound_Files.html
In Darwin and Donovan's case, the rhythmic phenomenon concerned is
perceptual isochrony. This may support my conjecture (based mainly on
introspection) that in end-stopped lines or couplets regular rhythmic events
are perceived as more rigid than in run-on lines or couplets.
Afterword 225
former may tend to hear the rhyme words as relatively flat hollow; the latter as
rich and resonant.
The second of Ehrenzweig's paragraphs quoted above requires some eluci
dation. In our context, I propose to substitute "diffuse, precategorial sensory
information" for "colour". Consider the assertions "the interaction of [diffuse,
precategorial sensory information] increases within the boundaries of a good
gestalt while it is inhibited across its boundaries"; and, conversely, "strong
interaction of [diffuse, precategorial sensory information] tends to make sharp
outlines perceived as much softer than they are". The prototypical text
structure most favourable to this principle would be what I have elsewhere
called the "divergent passage" (see, e.g., Tsur 1992a: 146-153; 455-470). This
consists of a more than usually fluid structure followed by an effective closure.
The last eight lines of the second stanza of "Kubla Khan", for instance, contain
a simplified version of it: there are three consecutive couplets each run on by
syntax to the next. Only the fourth (last) couplet entirely converges with syn
tax, yielding an effective structural closure. According to Ehrenzweig's second
paragraph and Darwin and Donovan's experiments, the interaction would in
crease across the boundaries of the enjambed couplets, and would be further
enhanced by the closing couplet of the passage, across which it would be
inhibited.6
As part three of the present book demonstrates, performers display
considerable freedom in manipulating intonation contours for the solution of
rhythmic problems. I used to believe that weak or strong shapes yielded by
complex part-whole relationships were a matter of mental, not vocal per
formance. In my instrumental analyses of vocal performances I found, how
ever, that vocal performances too may make a decisive contribution to such
perceptions. One of the most noteworthy examples in my corpus in this
respect is the first quatrain of Shakespeare's Sonnet 107, which contains a
single run-on sentence, where different actors may preserve to different degrees
both the distinct lines and the perceptual whole of the stanza (Tsur, 1998:
234-243).7 They can have recourse to intonation contours and other phonetic
cues to indicate line endings as well as to integrate the units-above-the-line.
In Chapter 2 above I noted that in Donne's alliterations, the repeated sound

clusters are sometimes separated by syntactic boundaries. Assuming that in ac
tual performance these boundaries are sometimes cued by intonation contours,
they may inhibit the interaction of precategorial auditory information across
them, contributing to Donne's notorious lack of musicality (I have elsewhere
discussed other means by which Donne deliberately sabotages the musicality of
his verse: Tsur, 1998: 49-53).
Three readings of the quatrain by outstanding British actors are available
online (excerpt 5) at
http://www.tau.ac.il/~tsurxx/Rhythm_Book_mp/Chapter7b_Sound_Files.html
The Emerging Model
Let us now step back and review what we have done so far. At the outset I
proposed a methodology to handle aesthetic effects, according to which
aesthetic objects must be described by critical terms that have rich descriptive
contents. Such descriptions may be trivial, unless they are rendered significant
by some more abstract model or theory, that cannot directly be applied to
aesthetic objects. The foregoing discussion offers a fairly complex instance of
this conception.
The inquiry begins with the pedestrian empirical observation that some
readers judge certain texts as "closed", some as "open"; some readers judge
them "boring" and some "interesting". Such findings would be in perfect
harmony with the widespread belief that in literary response "anything goes";
but the adherents of this conception need no controlled empirical evidence to
support it. Psychologists, by contrast, look for significant correlations of these
numbers with other scales of judgment and other variables of the aesthetic
situation; in our case, the variables are text structure, personality style, and
professional qualification. To determine what is "significant" we need a theory.
We started with fairly well-articulated theories concerning personality style,
and concerning a structuralist-gestaltist conception of text structures and their
perceived effects. But concerning professional qualification we had only vague
commonsense expectations that expert readers would be more competent than
others in realising the meanings and structures of a poem. In this respect, our
research came up with some quite unexpected details. It would appear that
professional qualification may hinder aesthetic response for low-absorption
respondents, but not for high-absorption respondents.8
The structuralist description of the respective structures of the three
versions of the Rubáiyát was reinterpreted in terms of gestalt theory, account
ing for certain perceived effects. Conflicting perceived effects do not
necessarily indicate random whims of the observers, but, rather, different
With reference to the three versions of the Rubáiyáth, significant group x

absorption interactions were found for two scales: BORING ~ INTERESTING,
and UNEMOTIONAL ~ EMOTIONAL (see figures 2-3 in Chapter 3). One notes
that the low-absorption "formally-qualified" group found the three versions
(overall) to be relatively more boring and less emotional than did the other
three groups, which rated the texts midway. This may suggest that high-
absorption subjects, both professional and nonprofessional, are equally respon
sive to these texts; and so are low-absorption subjects when not habituated by
professional experience. To put it more bluntly, professional training seems to
impede low-absorption (but not high-absorption) readers in responding to
aesthetically relevant aspects of the poems. This impediment appears quite
consistent: the two graphs show roughly the same pattern.
Afterword 227
mental performances of the text. In other words, the various responses are
governed by the same psychological principles, but are responses to different
texts; that is, various readers may respond to different poetic structures,
achieved through different mental organisations. These mental organisations, in
turn, are constrained by the phonetic, semantic, syntactic, and versification
elements that constitute the verbal structure. These differences of mental
performance become significant in terms of personality theory. On the
OPEN-CLOSED scale, for instance, symmetrical, "good" shapes control quite
efficiently both high- and low-absorption readers' responses. When, however,
the gestalts are weakened in one way or other, high- and low-absorption
respondents apply different mental strategies. The former are quite comfort
able with weak shapes, and may do nothing to improve them, or may even
actually search them out. The latter will tend to improve the shapes—within
the constraints of the verse structure. In the "primordial homogeneity" of the
aaaa rhyme pattern, differentiation—and, by the same token, closure—may be
achieved by slightly increasing the weight of the last line, without changing the
words of the text. In the aaba rhyme pattern law of good continuation is vio
lated. Good continuation cannot be restored without changing the actual words
of the text. But exceptionally "good return" can be achieved by exaggerating
("sharpening") the deviating line, leading to strong closure. The more exagger
ated the deviation, the more gratifying the "return" underlying the closure, the
punch line.
These results can be used as a starting point from which to draw inferences
about trance-inductive poems. The combination of an exceptionally regular
verse structure and an irregular (or, alternatively, homogeneous) rhyme struc
ture of the unit-above-the-line distinguish certain poems to which hypnotic or
ecstatic qualities are frequently attributed. Taken in itself, an elaborate struc
turalist description of this may arouse the notorious "so what" question. A
cognitive theory of hypnotic poetry may render the relationship between these
structural ranks meaningful. A "hypnotic" judgment of such a poem testifies to
a mental performance that appropriately realises both ranks. A "monotonous"
judgment, by contrast, indicates a reading that realises only (or mainly) the ex
ceptionally regular verse structure. A personality theory concerning tolerance
or intolerance of ambiguity, or concerning the relationship between the
"absorption"-scale and susceptibility to hypnosis may, again, render these
judgments significant. Furthermore, similar hierarchies may be constructed
concerning other elements as well: for instance, at the subphonemic level,
concerning phonetic categories, categorial perception and precategorial sensory
information, interpreted in perspective of ever higher cognitive tasks, and the
gestalt constraints on colour induction. Eventually, a wide range of such
hierarchies may be integrated into a quite sophisticated cognitive theory that
may account for rather complex aesthetic qualities.
Finally, one may apply wider aesthetic considerations for preferring one
reading to another. Of the wide variety of possible considerations I will
propose here only one. Concerning Omar Khayyam's original Rubáiyát, its
structure can be left ambiguous; but the sharpening of the deviant line may
yield much more pleasurable results than leaving the structure ambiguous. And
conversely, there are texts whose weak shapes may elicit pleasurable ambigu
ity much more efficiently. Furthermore, the resulting punch line may strongly
corroborate the epigrammatical style of this poem, as well as of the whole
poem sequence. So, the reading spontaneously applied by low-absorption
readers may be more satisfactory, though high-absorption readers, too, may
easily master it. In Ratosh's poem, by contrast, high-absorption readers are
inclined to realise the structural levels more fully, achieving a quite unique
aesthetic quality. In this respect, the aim of critical discourse in both instances
would be to point out certain ignored possibilities (rather than to announce the
correct interpretation).
But even such a mental performance would not be a guarantee for actualising
the hypnotic-ecstatic quality of, e.g., "Kubla Khan". In the Introduction I
quoted John Spencer Hill's survey of evaluations of this poem. For William
Hazlitt (and many other romantics), for instance, "Kubla Khan" only demon
strated that "Mr. Coleridge can write better nonsense verses than any man in
England"—and then he added, proleptically, "It is not a poem, but a musical
composition". This evaluation suggests that the respondent has appropriately
actualised a considerable part of the interacting auditory levels of the poem;
but the highly sophisticated rhythmic and musical elements are insufficient in
themselves, and must interact with contents. Assuming that rhythm functions
in ecstatic poetry similarly to rhythm in ritual, the following comment may be
relevant. "The effects of ritual behavior can affect the most complex functions
of the animal's mind, giving ritual the capability to alter emotion and mood. In
humans, because of the brain's great complexity, ritual behavior almost always
involves the highest level of thought and feeling. Our rituals are about some
thing; they tell stories, and these stories give them meaning and power. The
stories are crucial to the effectiveness of human ritual, chosen and shaped to
meet specific cultural needs" (Newberg et al., 2001: 85). Hazlitt seems to have
realised mainly the complex music of "Kubla Khan", rather than its interaction
with the "story"; alternatively, he may have realised both, but kept them
apart.
So, how do the experimental findings reported in Chapter 3 enhance our un
derstanding of decision style and hypnotic poetry in general, or of "Kubla
Khan" in particular? In Chapters 1 and 2 I started out with two general
assumptions: first, that cognitive and personality styles are general attitudes
that guide readers' responses to pieces of poetry as well as their critical deci
sions; and second, that persons endowed with negative capability are better
equipped to realise trance-inductive poems than those who are not. These
Afterword 229
assumptions relied mainly on a rather vague (though well-established) notion

of "intolerance of ambiguity". Persons who do not tolerate ambiguity will
defend themselves against the uncertainties induced by certain kinds of poetry.
Our empirical research has obtained fairly convincing evidence for these as
sumptions. But it did much more. It broke down the general attitudes into
specific activities that may account for the fine-grained processes that underlie
both assumptions. My general model was something like this: There are poems
characterised as "hypnotic", "trance-inductive", "spell-weaving". Such poems
have certain discernible characteristic features. This is the input fed into a black
box, the head, in which we hypothesise a mediational system that determines
the respective responses of various readers which, in turn, determine their re
spective critical outputs. There are certain more-or-less established dichotomic
spectra of decision styles, with associated cognitive strategies. From the
strategies deployed in the critic's output one may infer the general kind of
mediational system that generated that output, but not its details. We also had
a third kind of expectation, that the gestalt rules of perception would be
relevant to all stages of these activities.
Our experiments gave us a clue to the fine-grained structure of the hitherto
hidden mediational system, at the level of immediate perception. This enabled
me to insert into the model some of the empirical findings on speech percep
tion and colour interaction. In my speculative book on Hebrew hypnotic
poetry I relied on the gestalt-rules of perception in an attempt to predict the
reader's differential responses to the three versions of the Rubáiyát. One
purpose of our experiment was to ascertain that readers did, indeed, respond in
the expected way. However, having distinguished between high- and low-
absorption respondents, we also discovered that in some instances the two
groups of respondents handled rhyme patterns in opposite ways—but in a
manner that was compatible with gestalt theory. Where the control of strong
shape was effective, their responses were fairly uniform; they diverged where
the control of strong shapes was loosened. Respondents resorted to different
mental performances. By a happy accident, the rating scale concerned was the
OPEN ~ CLOSED judgments on rhyme patterns. The same texts were treated as
considerably open by one group, and as considerably closed by the other. Such
differences of perception have far-reaching consequences at two very different
stages of response to hypnotic poems. At one stage, a distinction between
false and genuine security induced by regular rhythm is crucial for distinguish
ing between witty and hypnotic poetry. When rhythm is regular and the
rhymes too are perceived as rigidly patterned, yielding a closed unit, security
tends to be perceived as genuine; when the rigid patterning of rhymes is
loosened, security tends to be perceived as false. High-absorption readers were
inclined to leave rhyme patterns more ambiguous than low-absorption readers.
At another stage, closed and open gestalts were hypothesised to crucially in-
fluence the interaction and fusion of the rich precategorial auditory information
that carries the speech sounds, thus affecting the perceived musicality of verse.
Thus, our experimental results, complemented by those of other researchers,
enable me, at the end of this book, to construct a fairly fine-grained model of
the mental performances that fill the gap between hypnotic poems' complex
structures and the different responses of low- and high-absorption readers to
them. We started out with a bottom-up and top-down approach concurrently.
We experimented with a monorhymed poem which we had reasons to choose
as a paradigmatic instance of a trance-inductive poem. In a top-down view,
high-absorption respondents tended to judge it "interesting", low-absorption
respondents "boring". In a bottom-up perspective, the process began with rich
precategorial auditory information at the sub-phonemic level, proceeding to
regularly alternating events at the line level, then to the grouping of rhymes at
the unit-above-the-line level.
From what we know about phonetic coding, diffuse precategorial auditory
information is recoded into clear-cut phonetic categories (in some phonetic
categories it is more thoroughly recoded than in others). From experiments
with efficient and poor readers we know that some persons are more inclined
than others to rely in the performance of certain cognitive tasks on the
auditory information associated with phonetic coding. From what we know
about personality styles, there are persons of certain styles for whom "the
unique, unclassifiable sensation is particularly offensive"; such persons will be
anxious to completely re-structure the auditory information into clear-cut
phonetic categories as fast as possible. Persons of other personality styles
may attempt to hold it "for several seconds in an echoic sensory register".
From what we have discovered in our experiments about personality style and
poetic closure, some cognitive styles will attempt to close versification units
whenever possible; others will enjoy keeping them open. From what we know
about the figure-ground relationship and colour interaction, we should expect
the interaction of the precategorial auditory information to be strongest when
control of strong shapes over the cognitive system is considerably loosened.
When line-endings and couplet-endings are blurred by run-on sentences (as in
the last eight lines of the second stanza of "Kubla Khan"), or when (as in the
first stanza and the first five lines of the second stanza of "Kubla Khan") a
fifth line intrudes upon what otherwise would be a symmetrical quatrain, or
when (as in Ratosh's "Bircat Shoshanim") monorhyme renders the grouping of
lines uncertain, the interaction of the precategorial auditory information is
greatly increased across the boundaries of weak prosodic shapes. When the
verse lines and couplets are end-stopped, or when a quatrain can be divided
into two symmetrical halves, interaction is inhibited. Likewise, the witty or
hypnotic effect of regular rhythms depends on whether they are contained
within rigid, clear-cut higher prosodic units, or can interact across the bound
aries of weak, flexible higher units. Thus, we may assume that high-absorption
Afterword 231
readers will experience "Kubla Khan" as having a richer, more resonant audit
ory texture; low-absorption readers as having a relatively dull, hollow sound.
From what we know about the relationship between gestalt-free, diffuse sen
sory information and emotional responsiveness according to the Rorschach
inkblot test, we expect that being exposed to the precategorial auditory
information in the poetic mode of speech perception would greatly increase the
emotional quality perceived in a text. Our experiments have established a
crucial stage in this complex process, midway between the "bottom" and the
"top". One's inclination to perceive shapes as open or closed may crucially
affect the dynamics of the precategorial auditory information. I have suggested
that hypnotic poems tend to shift attention away from their contents to their
music. The above considerations would indicate that high-absorption readers
perceive more music to shift attention to than low-absorption readers. This
distinction appears to be valid with reference to poetry in general, but is
absolutely critical with reference to hypnotic poetry.
Thus, we may know quite a lot about the subliminal and semi-conscious
mental processes of high- and low-absorption readers that may account for
their different responses to Coleridge's or Ratosh's hypnotic poems. In a wide
range of my writings I have warned against the reductionist use of cognitive
mechanisms in critical discourse. Here I propose to make a positive comment.
When we use "phonetics"-language or "colour-interaction"-language to describe
our responses to these poems and others, we do not merely describe the cogni
tive mechanisms underlying them. Rather, this is also a way to draw attention
to subliminal perceptions that take place during reading where no better meta
language is available. It should be also mentioned that for a variety of cultural
reasons it is unlikely that a present-day critic would squarely declare a poem
like "Kubla Khan" boring or the like. He will have recourse to some round
about, academically more respectable strategy. One such strategy we encoun
tered in Chapter 1. The critic may direct attention away from the poem to
"more interesting" stuff: to its alleged or real sources, or some sexual symbol
ism. That is very scientific, academically most respectable, and above all, it ex
empts us from the need to face the excruciating ecstatic quality of the poem
that forces us to "contemplate the powers of darkness with equanimity".
The Archimedean Point
An anonymous reviewer argued as follows against my approach in Chapter 1.1

gave several answers to this criticism, of which I will reproduce here only one.
In the opening the author suggests that certain reading practices are as
maladaptive to certain kinds of poetry as certain behaviors are to social
relationships. Sounds fine, until I start asking—wait a minute. Social
relations are one thing (there is living feedback, success, failure, happiness,
rejection and acceptance), but what is the negative result of reading a

hypnotic poem with a higher degree of closure?
The aim of my argument in Chapter 1 is not to diagnose the critic's person

ality style for its own sake. It has, in fact, a double theoretical focus: validity
in interpretation and the implied critic's decision style, resulting in reading
strategies. One may independently follow two parallel but separate courses.
One may establish and apply a methodology for assessing validity in interpret
ation. And one may observe in critical discourse cognitive strategies
characteristic of certain cognitive styles, empirically established by psycholo
gists (in the psychological laboratory and clinic). Then one may check whether
interpretations judged invalid on methodological grounds correspond in any
way to mental strategies judged maladaptive on psychological grounds. I will
briefly demonstrate this through a relatively simple instance; but Chapter 1
abounds in more sophisticated examples as well.
In Chapter 1 I discussed Fleissner's comment on sunless sea. "Again an
onomatopoeic meaning dominates: continuity of sibilants suggests the surge of
the sea". There is no doubt that the sibilants are there; and that repeated sibi
lants can serve to imitate all sorts of noises. However, in Benjamin Harshav's
typology, sunless sea is a "neutral sound pattern", that is, a sound pattern that
draws attention to itself, but does not connect to corresponding salient
meanings in the text. The "surge of the sea" does not occur in the poem. What
is more, the phrase lifeless ocean makes it quite implausible that a resounding
"surge" would be a salient feature of this particular sea. This might suffice to
conclude that Fleissner's interpretation fails to be valid (in Chapter 1 I pro
pound a much more elaborate methodological argument). As to the cognitive
strategies deployed, they are precisely the ones regularly attributed by
psychologists to the quest for certitude. "Neutral sound pattern" is a "unique,
unclassifiable sensation", and as such, "particularly offensive" for the quest for
certitude. Indeed, Fleissner eliminates it, by relating it to a (nonexistent) mean
ing. Such an interpretation requires the critic to achieve rapid closure, so as to
separate the cognitive processing of the phrase sunless sea from that of lifeless
ocean, resulting in compartmentalisation rather than integration of meaning.
Finally, this is an exquisite instance of what Harvey described as "a greater
insensitivity to subtle and minimal cues and hence a greater susceptibility to
false but obtrusive cues". The onomatopoeic potential of the sibilants is here
precisely such a "false but obtrusive cue". It obviates the need to face the
music of a "meaningless" sound texture. As I have shown in Chapter 1, this
strategy is persistent in Fleissner's handling of sound patterns in "Kubla
Khan"; and is consistently opposed to, e.g., Elizabeth Schneider's handling.
Thus, validity in interpretation and adaptation strategies may be indepen
dently established, and maladaptive cognitive strategies in reading may receive
"living feedback" in terms of success and failure. Such failure, in terms of
Afterword 233
validity judgments, may count as a "negative result" by any standard; that is,
unless one concludes that in interpretation anything goes.
Archimedes, inventor of the lever, was reported to have said: give me a point
outside the universe, and I'll dislodge it. When confronted with conflicting
readings or responses to a piece of literature, the readers' decision style and,
eventually, mental performance, may serve as just such an Archimedean point.
I would adapt to this issue an old comment of mine (Tsur, 1992: 303). We
perceive the world (including poems) through our peculiar "lenses". We cannot
escape this limitation. This, however, does not mean that "anything goes".
David Pears criticised the old Kantian microscope analogy in his introduct
ory remarks on Kant's and Wittgenstein's critical philosophy as follows:
If a scientist became convinced that what he saw through his microscope

was an effect of a flaw in the lens, he would start all over again. But the
analogy is imperfect at the essential point. A microscope yields one set of
observations, whereas what comes through the lens of the mind is the to
tality of human experience. So in this case there is no possibility of side
stepping and no independent check, and the very idea, that this lens might
be flawed, seems to be empty (Pears, 1971: 30-31).
At some variance with Pears's implication, my point is that we do not build

worlds of individual fantasies under the pretext of reading a poem. We try to
make sure that the many perceptions of a poem are sufficiently similar, so as
to afford meaningful discussions between a great variety of perceivers. This
can be done only if we find out as much as possible about our shared tools and
strategies of perception, as well as about the peculiar biases of our individual
lenses that seem to be responsible for individual differences in perception. I
submit that my discussions of "Mental Performance" and of "The Implied
Critic's Decision Style" are major steps in this direction.
References
Alcock, Theodora (1963) The Rorschach in Practice. London: Tavistock

Publications.
Arnheim, Rudolf (1967) Art and Visual Perception. London: Faber.
Auerbach, Erich (1962 [19521]) Odysseus' Scar, in George Steiner and Robert
Fagles (eds.), Homer — A Collection of Critical Essays, 19-37.
Englewood Cliffs, N.J.: Prentice-Hall.
Austin, James H. (2000) Consciousness Evolves When Self Dissolves, in
Jensine Andresen and K.C. Forman (eds.) Cognitive Models and
Spiritual Maps, 209-230. Exeter: Imprint Academic.
Barney, Tom (1990) "The Forms of Enjambment". University of Lancaster
unpublished MA dissertation.
Bartels, Christine (1999) The Intonation of English Statements and Questions
—A Compositional Interpretation. New York & London: Garland
Publishing.
Beardsley, Monroe C. (1958) Aesthetics: Problems in the Philosophy of
Criticism. New York & Burlingame: Harcourt, Brace & World.
Beardsley, Monroe C. (1966) Aesthetics: From Classical Greece to the
Present. New York: MacMillan.
Beer, J.B. (1959) Coleridge the Visionary. London: Chatto & Windus.
Benzon, William L. (2003) 'Kubla Khan' and the Embodied Mind. PSYART: A
Hyperlink Journal for the Psychological Study of the Arts. Available:
http://www.clas.ufl.edu/ipsa/journal/2003/benzon02.htm
Berek, Peter (1978) Interpretation, Allegory and Allegoresis. College English,
40:117-132.
Black, Max (1962 [1954-19551]) Metaphor, in Joseph Margolis (ed.),
Philosophy Looks at the Arts, 218-235. New York: Scribner.
Blake, Robert R. and Glenn V. Ramsey (eds.) (1951) Perception—An
Approach to Personality. New York: Ronald Press.
Bloom, Harold (1963) The Visionary Company: A Reading of English
Romantic Poetry. Garden City: Doubleday Anchor.
Bodkin, Maud (1963 [19341]) Archetypal Patterns in Poetry. Oxford: Oxford
UP.
Booth, Wayne C. (1961) The Rhetoric of Fiction. Chicago: Chicago UP.
Borillo, Mario (1995) Cognition & Creation: A Manifesto for the Pursuit of
Convergences between Theories of Artistic Creation and Cognitive
Science. Toulouse: Université Paul-Sabatier.
Brady, Susan, Donald Shankweiler & Virginia Mann (1983) Speech Perception
and Memory Coding in Relation to Reading Ability. Journal of
Experimental Child Psychology 35: 345-367.
Brooke-Rose, Christine (1958) A Grammar of Metaphor. London: Secker.
Brown, Roger (1968) Words and Things. New York: Free Press.
Bruner, Jerome S. and David Kretch (eds.) (19682) Perception and Personality.
New York: Greenwood Press.
Burke, Kenneth (1962) A Grammar of Motives and A Rhetoric of Motives.
Cleveland & New York: Meridian Books.
Chafe, Wallace (1994) Discourse, Consciousness and Time: The Flow and
Displacement of Conscious Experience in Speaking and Writing. Chicago
and London: The University of Chicago Press.
Chatman, Seymour (1960) Comparing Metrical Styles, in Thomas Sebeok
(ed.), Style in Language, 149-172. Cmabridge Mass.: MIT.
Chatman, Seymour (1965) A Theory of Meter. The Hague: Mouton.
Cook, Norman D. (2002) Tone of Voice and Mind. Amsterdam: John
Benjamins Publishing Company.
Cooper, C. W. and L.B. Meyer (1960) The Rhythmic Structure of Music.
Chicago: Chicago UP.
Crowder, Robert G, and Richard K. Wagner (1992) The Psychology of
Reading. New York and Oxford: Oxford UP.
Damasio, Antonio (1994) Descartes ' Error: Emotion, Reason and the Human
Brain. New York: Grosset/Putnam Books.
Damasio, Antonio (1999) The Feeling of What Happens: Body and Emotion in
the Making of Consciousness. New York: Harcourt Brace & Company,
1999.
Darwin, C. I. and A. Donovan "Perceptual Isochrony" (demo tape).
Darwin, C. I. and A. Donovan (1979) Perceptual Studies of Speech: Isochro
ny and Intonation. Paper presented at the Conference on Language Gen
eration and Understanding, Bonas (Gers).
Donovan A. and C. I. Darwin (1979) The Perceived Rhythm of Speech. Paper
presented at the IXth International Congress of Phonetic Sciences,
Copenhagen.
Ehrenzweig, Anton (1965) The Psychoanalysis of Artistic Vision and Hearing.
New York: Braziller.
Ehrenzweig, Anton (1970) The Hidden Order of Art. London: Paladin.
Fellows, B. J. (1986) The Concept of Trance, in P. L. N. Naish (ed.), What is
Hypnosis? Current Theory and Research. Milton Keynes: Open Uni
versity Press. 37-56.
Fleissner, Robert F. (2000) Sources, Meaning, and Influences of Coleridge's
Kubla Khan: Xanadu Re-routed—A Study in the Ways of Romantic
Variety. Lewiston, N.Y.: E. Mellen Press.
References 237
Fónagy, Iván (2001) Languages within Language: An Evolutive Approach.

Amsterdam: John Benjamins (Foundations of Semiotics Series).
Frenkel-Brunswick, Else (1968 [19481]) Intolerance of Ambiguity as an
Emotional and Perceptual Variable in J.S. Bruner and D. Kretch (eds.),
Perception and Personality. New York: Greenwood Press.
Fry, D. B. (1958) Experiments in the Perception of Speech. Language and
Speech 1: 126-151.
Fruman, Norman (1972) Coleridge: The Damaged Archangel. London: Allan
& Unwin.
Garrett, M., T. Bever and J. A. Fodor (1966) The Active Use of Grammar in
Speech Perception. Perception and Psychophysics 1: 30-32.
Glicksohn, J. (1987) Hypnotic Behaviour Revisited: A Trait-Context Interac
tion. Behavioral and Brain Sciences. 10 (4). 774-775.
Glicksohn, J. (1989) Cutting the 'Gordonian Knot' Using Absorption and
Dream Recall. Journal of Mental Imagery, in press.
Glicksohn, J., R. Tsur and Ch. Goodblatt (1991) Absorption and Trance-
Inductive Poetry. Empirical Studies of the Arts 9/2: 115-122.
Goldstein, K. and M. Scheerer (1941) Abstract and Concrete Behavior.
Psychological Monographs 53,2: 1-10.
Gombrich, E.H. (1961 [19511]) Meditations on a Hobby Horse or the Roots of
Artistic Form, in Morris Philipson (ed.), Aesthetics Today, 114-128
Cleveland & New York: Meridian Books.
Gombrich, E.H. (1969 [19651]) The Use of Art for the Study of Symbols, in
James Hogg (ed.), Psychology and the Visual Arts, 149-170. Harmonds-
worth: Penguin.
Halle, Morris and Samuel Jay Keyser (1971) English Stress: Its Growth and
Its Role in Verse. New York: Harper and Row.
Harvey, O.J. (1970 [19671]) Conceptual Systems and Attitude Change, in
Peter B. Warr(ed.), Thought and Personality, 315-333. Harmondsworth:
Penguin.
Hasenfus, N., Martindale, C , & Birnbaum, D. (1983) Psychological Reality
and Cross-Media Artistic Styles. Journal of Experimental Psychology:
Human Perception and Performance, 9 (6), 841-863.
Hepburn, R.W. (1968) Emotion and Emotional Qualities, in Harold Osborne
(ed.), Aesthetics in the Modern World. London: Hudson.
Herrnstein-Smith, Barbara (1968) Poetic Closure: A Study of How Poems End.
Chicago: Chicago UP.
Hilgard E. R. (1981) Imagery and Imagination in American Psychology. Jour
nal of Mental Imagery 5 (1). 5-66.
Hilgard, J. (1979) Imagination and Sensory-Effective Involvements in Every
day Life and in Hypnosis, in E. Fromm & R. E. Shor (eds.), Hypnosis:
Developments in Research and New Perspectives (second ed.). New
York: Aldine. (483-517).
Hill, John Spencer (1983) A Coleridge Companion. London: Macmillan.
House, Humphrey (1973 [19531]) 'Kubla Khan' and 'Christabel', in Alun R.

Jones and William Tydeman (eds.), Coleridge: The Ancient Mariner
and Other Poems -A Casebook, 200-216. London: Macmillan.
House, Jill and Anne Wichmann (1996) Investigating Peak Timing in
Naturally-Occurring Speech: From Segmental Constraints to Discourse
Structure. Speech, Hearing and Language: Volume 9. Available online:
http://www.phon.ucl.ac.uk/home/shl9/jill/house.htm
Hrushovski, Benjamin (1968) Do Sounds Have Meaning? The Problem of
Expressiveness of Sound-Patterns in Poetry. HaSifrut 1: 410-420 (in
Hebrew). English Summary: 444.
Hrushovski, Benjamin (1980) The Meaning of Sound Patterns in Poetry: An
Interaction View. Poetics Today 2: 39-56.
Jakobson, Roman (1956) Two Aspects of Language and Two Types of Apha
sic Disturbances, in Roman Jakobson & Morris Halle, Fundamentals of
Language. The Hague: Mouton. 55-82.
Jakobson, Roman (1980) Brain and Language. Columbus Ohio: Slavica
Publishers.
Kane, Julie (2004) Poetry as Right-Hemispheric Language. Journal of Con
sciousness Studies—Controversies in Science & Humanities: 21-59.
Klein, George S. (1951) The Personal World through Perception, in R. R.
Blake and G.V. Ramsey (eds.), Perception—An Approach to Personality.
New York: Ronald Press. 328-355.
Knight, George Wilson (1960[19411]) Coleridge's Divine Comedy, in M.H.
Abrams (ed.), English Romantic Poets: Modern Essays in Criticism,
158-169. Oxford & London: Oxford UP.
Knowles, Gerry (1991) "prosodic Labelling: The Problem of Tone Group
Boundaries", in Stig Johannson and Anna-Brita Stenström (eds.),
English Computer Corpora. Selected Papers and Research Guide.
(Topics in English Linguistics 3) Berlin: Mouton de Gruyter. 149-163.
Köhler, Wolfgang (1972) The Task of Gestalt Psychology. Princeton: Princeton
UP.
Kreitler H., & Kreitler S. (1972). Psychology of the Arts. Durham NC: Duke
UP.
Kumar, V. K., Pekala, R. J. (1988) Hypnotizability, Absorption, and Individu
al Differences in Phenomenological Experience. International Journal of
Clinical and Experimental Hypnosis, 36. 80-88.
Ladd, Robert D. (1996) Intonational Phonology. Cambridge University Press.
Ladefoged, Peter (2001) Vowels and Consonants: An Introduction to the
Sounds of Languages. Oxford: Blackwell.
Liberman, A.M., I. M. Mattingly, & M. T. Turvey (1972) Language Codes
and Memory Codes, in Melton, A. W. and E. Martin (eds.), Coding
Processes in Human Memory. New York: Winston.
Liberman, Isabelle Y., and Virginia A. Mann (1981) Should Reading Instruction
and Remediation Vary with the Sex of the Child? Status Report on
Speech Research SR-65: Haskins Laboratories.
References 239
Lieberman, Philip (1967) Intonation, Perception and Language, Cambridge

Mass.: MIT.
Lowes, John Livingston (1927) The Road to Xanadu: A Study in the Ways of
Imagination. Boston & New York: Houghton Mifflin Co.
Mann, Virginia A. (1984) Reading Skill and Language Skill. Developmental
Review 4: 1-15.
Margolis, Joseph (1962 [19581]) The Logic of Interpretation, in Joseph
Margolis (ed.), Philosophy Looks at the Arts, 108-118. New York:
Scribner.
Meyer, Leonard B. (1956) Emotion and Meaning in Music. Chicago: Chicago
UP.
Miller, George A. (1979) Images and Models, Similes and Metaphors, in
Andrew Ortony (ed.), Metaphor and Thought, 202-250. Cambridge:
Cambridge UP.
Miller, J.G. (1951) Unconscious Processes and Perception, in Blake &
Ramsey (eds.), Perception—An Approach to Personality. New York:
Ronald Press.
Monteiro, K. P., Macdonald H., & Hilgard E. R. (1980) Imagery, Absorption,
and Hypnosis: A Factorial Study. Journal of Mental Imagery, 4. 63-81.
Newberg, A., D'Aquili E. & Rause V. (2001) Why God Won't Go Away: Brain
Science and the Biology of Belief New York: Ballantine Books.
Ohala, John J. (1994) The Frequency Code Underlies The Sound-Symbolic
Use of Voice Pitch, in Hinton, Leanne, Johanna Nichols, and John J.
Ohala eds. Sound symbolism. Cambridge: Cambridge University Press.
325-347.
Ohmann, Richard (1970 [19621]) Modes of Order in Donald C. Freeman (ed.),
Linguistics and Literary Style, 209-242. New York: Holt, Rinehart &
Winston.
Ornstein, Robert E. (1975) The Nature of Consciousness. Harmondsworth:
Penguin Books.
Otto, Rudolf (1959) The Idea of the Holy, John W. Harvey (trans.). Har-
mondsworth: Penguin.
Pater, Walter (1951[18731]) Conclusion, from The Renaissance, in James
Harry Smith and Edd Winfield Parks (eds.), The Great Critics, 894-897.
New York: Norton.
Pears, David (1971) Wittgenstein. London: Fontana.
Pekala, R. J., Wenger, C. F., & Levine, R. L. (1985) Individual Differences in
Phenomenological Experience: States of Consciousness as a Function of
Absorption. Journal of Personality and Social Psychology, 48 (1), US
UI).
Plutchik, R. (1968 [19621] The Evolutionary Basis of Emotional Behaviour, in
Magda B. Arnold (ed.), The Nature of Emotion, 67-80. Harmondsworth:
Penguin.
Persinger, Michael A. (1987) Neuropsychological Bases of God Beliefs. New
York, Westport, Connecticut, London: Praeger.
Preminger, Alex and T. V. F. Brogan (1993) The New Princeton Encyclopedia

of Poetry and Poetics. Princeton: Princeton UP.
Quartermain, Peter (1998) Sound Reading, in Charles Bernstein (ed.), Close
Listening: Poetry and the Performed Word. New York and Oxford:
Oxford University Press.
Rader, C. M., & Teilegen A. (1987) An Investigation of Synesthesia. Journal
of Personality and Social Psychology, 52 (5). 981-987.
Richards, I. A. (1925) Principles of Literary Criticism. London: Kegan Paul.
Rorschach, Hermann (1951) Psychodiagnostics. Bern: Huber.
Schneider, Elisabeth (1975 [19531]) Coleridge, Opium and Kubla Khan. New
York: Octagon Books.
Sibley, Frank (1962 [19591]) Aesthetic Qualities, in Joseph Margolis (ed.),
Philosophy Looks at the Arts, 63-87. New York: Scribner.
Sluijter, Agaath (1995) Phonetic Correlates of Stress and Accent. The Hague:
Holland Institute of Generative Linguistics.
Snyder, E. D. (1930) Hypnotic Poetry: A Study of Trance-Inducing Techniques
in Certain Poems and its Literary Significance. Philadelphia: University
of Pennsylvania Press.
Snyder, E. D. & Shore, R. E. (1983). Trance-Inductive Poetry: A Brief
Communication. International Journal of Clinical and Experimental
Hypnosis 31, 1-7.
Spanos, N. P. (1987) Hypnotic Behavior: Special Process Accounts are Still
not Required. Behavioral and Brain Sciences. 10 (4). 776-781.
Spanos, N. P., Brett, P. J., Menary E. P., & Cross, W. P. (1987) A Measure of
Attitudes Towards Hypnosis: Relationships with Absorption and
Hypnotic Susceptibility. American Journal of Clinical Hypnosis, 30 (2).
139-150.
Sutherland, Keith (2001) Consciousness and Emotion. Journal of Conscious
ness Studies 8 (12) 79-82.
Tellegen, A. (1981). Practicing the Two Disciplines for Relaxation and Enlight
enment: Comment on 'Role of the Feedback Signal in Electromyograph
Biofeedback: The Relevance of Attention' by Quails and Sheehan.
Journal of Experimental Psychology: General, 110 (2), 217-226.
Tellegen, A. (1982) Multidimensional Personality Questionnaire. University of
Minnesota.
Tellegen, A., & Atkinson, G. (1974) Openness to Absorbing and Self-Altering
Experiences ('Absorption'), a Trait Related to Hypnotic Susceptibility.
Journal of Abnormal Psychology, 83 (3). 268-277.
Tsur, Reuven (1969) 'Dabhar' (Word) by Bialik: A Relationship between the
Teaching of Literature and the Education for Values (in Hebrew).
Methodica, Summer: 70-78.
Tsur, Reuven (1972) Articulateness and Requiredness in Iambic Verse, Style
6: 123-148.
Tsur, Reuven (1975) Two Critical Attitudes: Quest for Certitude and Negative
Capability. College English 36: 776-788.
References 241
Tsur, Reuven (1977) A Perception-Oriented Theory of Metre. Tel Aviv: The

Porter Israeli Institute for Poetics and Semiotics.
Tsur, Reuven (1978) Emotion, Emotional Qualities and Poetry. Psychocultural
Review 2: 165-180.
Tsur, Reuven (1979) Levels of Information Processing in Reading Poetry.
Critical Inquiry 5: 751-759.
Tsur, Reuven (1983a) What is Cognitive Poetics? Tel Aviv: The Katz
Research Institute for Hebrew Literature.
Tsur, Reuven (1983b) Critical Terms and Insight: The Mental Dictionary of
"Critical Competence". Tel Aviv: The Katz Research Institute for
Hebrew Literature.
Tsur, Reuven (1985a) Contrast, Ambiguity, Double-Edgedness. Poetics Today
6, 417-445.
Tsur, Reuven (1985b) Decoding Secret Messages, and the Quest for Certitude:
Hebrew Critics Confronting Literature of Extreme Situations (in
Hebrew). Hassifrut 34: 142-160.
Tsur, Reuven (1988) Hebrew Hypnotic Poetry — Preliminary Observations.
(in Hebrew). Tel Aviv: The Katz Research Institute for Hebrew
Literature.
Tsur, Reuven (1992a) Toward a Theory of Cognitive Poetics. Amsterdam:
Elsevier (North Holland) Science Publishers.
Tsur, Reuven (1992b) What Makes Sound Patterns Expressive: The Poetic
Mode of Speech-Perception. Durham N, C: Duke UP.
Tsur, Reuven (1994) Hipnotikus Költészet. Holmi 6: 243-265. (in Hungarian).
Tsur, Reuven (1996) Text, Reader, World—The Implied Critic's and the Real
Reader's Decision Style (in Hebrew). Tel Aviv: Hakibuts Hameukhad
Publishing House.
Tsur, Reuven (1997) Sound Affects of Poetry—Critical Impressionism,
Reductionism, and Cognitive Poetics Pragmatics and Cognition 5: 283-
304.
Tsur, Reuven (1998) Poetic Rhythm: Structure and Performance—An
Empirical Study in Cognitive Poetics. Bern: Peter Lang.
Tsur, Reuven (2000) Phonetic Cues and Dramatic Function—Artistic
Recitation of Metered Speech. Assaph—Studies in the Theatre: 173-196.
Tsur, Reuven (2001) Onomatopoeia: Cuckoo-Language and Tick-Tocking—
The Constraints of Semiotic Systems. IconicitylnLanguage. Available
Online:
http://www.trismegistos.com/Iconici1yInLanguage/Articles/Tsur/default.
html
Tsur, Reuven (2002) Phonetic Cues and Dramatic Function—Artistic Recita
tion of Metered Speech. (Expanded version). PSYART: A Hyperlink
Journal for the Psychological Study of the Arts.
http://www.clas.ufl.edu/ipsa/journal/2002/tsur05.htm
Tsur, Reuven and Motti Benari (2002) 'Composition of Place', Experiential

Set, and the Meditative Poem (A Cognitive-Pragmatic Approach).
Pragmatics and Cognition. 203 - 237.
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 (2003b) Deixis and Abstractions: Adventures in Space and Time
in Joanna Gavins and Gerard Steen (eds.), Cognitive Poetics in Practice.
London: Routledge. 41-54.
Tsur, Reuven (forthcoming) Delivery Style and Listener Response in the
Rhythmical Performance of Shakespeare's Sonnets, College Literature
(special issue on "Cognitive Shakespeare: Criticism and Theory in the
Age of Neuroscience", guest edited by Patrick Colm Hogan and Lalita
Pandit).
Tsur, Reuven, Glicksohn J., & Goodblatt C. (1988) Gestalt Qualities in Poetry
and the Reader's Absorption Style. Paper given at the international
symposium on Text and Context (Cognitive Symposium III), Tel Aviv
University.
Tsur, R., Glicksohn, J., & Goodblatt, C. (1990) Perceptual Organization,
Absorption and Aesthetic Qualities of Poetry. Halász László (ed.),
Proceedings of the 11th International Congress on Empirical Aesthetics.
Budapest: Institute for Psychology of the Hungarian Academy of
Sciences. 301-304.
Tsur, R., Glicksohn, J., & Goodblatt, C. (1991) Gestalt Qualities in Poetry
and the Reader's Absorption Style. Journal of Pragmatics 16/4: 193-
206.
Valentine, C. W. (1962) The Experimental Psychology of Beauty. London:
Methuen.
Watson, George (1973 [19661]) "Kubla Khan", in Alun R. Jones and William
Tydeman (eds.), Coleridge: The Ancient Mariner and Other Poems - A
Casebook, 221-234. London: Macmillan.
Weisstein, Ulrich (1974) Impressionism, in Alex Preminger (ed.), Princeton
Encyclopedia of Poetry and Poetics. Princeton: Princeton UP.
Weitz, Morris (1972 [19641]) Hamlet and the Philosophy of Literary
Criticism. London: Faber & Faber.
Wellek, René (1963) Concepts of Criticism. New Haven & London: Yale UP.
Wimsatt, W.K. (1954) The Verbal Icon. New York: The Noonday Press.
Wittgenstein, Ludvig (1976 [19531]) Philosophical Investigations, trans.
G.E.M. Anscombe. Oxford: Blackwell.
Yarlott, Geoffrey (1967) Coleridge & the Abyssinian Maid. London: Methuen.
Recorded Readings
Pack, Roger Lloyd reading S.T Coleridge. Rough Winds Productions. (2000)
References 243
Richardson, Sir Ralph reading The Poetry of Coleridge. HarperCollinsAudio-

Books (Caedmon, 19581).
Rowe, John, Alex Jennings & Stephen Critchlow reading The Nations
Favourite Lakeland Poems. BBC Radio Collection. (1999)
Sheen, Michael reading Great Poets of the Romantic Age. Naxos AudioBooks.
(1994).
Index
absence, 27, 82 analytical philosophy of criticism,

powerful absence, 81 70-76
absorption, 1, 2, 3, 4, 6, 7, 8, 115- Apollo, 76
141, 227 Apollonian, 98
high- and low, 6, 15, 209, 213, archetypal
215-217, 228, 229, 230, 231 patterns, 27, 113
abstract vs. concrete personality, 15, approach, 91
31-39 contents vs. archetypal patterns,
abstraction, 41 53, 58, 98
abstractive relevance, 56, 57 interpretation, 55
absurd, 6 archetype , 30
theatre of , 77 Death and Rebirth, 53, 54, 75, 92-
Addison, 79 100
aesthetic of Paradise and Hades (of Heaven
effect, 110, 128 and Hell), 27, 29, 72
evaluation, 132 Archimedes, 65, 233
judgment, 128 Ariosto, 79
quality, 50, 84, 120, 227, 228 Arnheim, 116-118, 128, 134, 224
response, 57, 60, 210, 215 Arnold, Magda B., 167
affect, 158 ars gratia artis, 4
affective fallacy, 69 art history, 9
Alcock, 218 aspiration, 152, 153, 154, 184, 185,
alliteration, 39-49, 103-106, 110, 186, 193
174, 178, 180, 188, 206-207, assonance, 39-49
212-213, 218, 221, 223 Auerbach, 86
altered states of consciousness, 3, Austin, James H., 96
92, 96, 208 Bacchic maidens, 37
ambiguity, 5, 7, 8, 77, 102, 173, Bacchus, 52
223, 228, 135, see also tolerance back-structuring, 178
and intolerance of ambiguity Barney, Tom, 155
ambiguous, 8, 105, 110, 136, 139, Bartels, Christine, 146, 148, 170-
228, 229 171, 172, 186, 190
ambivalence (ambivalent), 16, 24 Bartlett, 89
amygdala, 14 Beardsley, 50, 71, 80, 81, 98
Beckett, 55
Beer, 25-26, 62, 69, 75 158, 159, 161, 190-196, 199,

Benzon, 17, 79, 82, 85, 95, 143445 200, 203, 209, 211, 216-217,
Berek, 12, 15 219, 227, 229, 230
Bialik, 16 structural, 119, 139, 225
Black, Max, 114 anti-closure, 194-195
Blake and Ramsey, 5 cognitive psychology, 1
Blake, William, 101 cognitive strategies, 130, 134, 135
Bloom, Harold, 32-36 cognitive style, 7, 11-77, 209-233
Bodkin, Maud, 27-30, 38, 39, 50-56, see also decision style, personality
57, 58, 59-75, 81, 91, 98-99, 113 style
Booth, Waine, 14 Collins, 79
Borillo, Mario, 211 colour induction, 222-223, 227
bottom-up and top-down approach, colour interaction, 45, 229, 230
230 competence/performance, 16
Brady et. al., 220 concave, 155, 167
breathy voice, 175, 187 see also concrete and abstract
incomplete phonation, voice functioning, 30-39, 48
quality personality, 6, 17, 22
Bridges, 42 concrete universal, 32, 38, 50
Brooke-Rose, 16 contrastive stress, 200, 202
Brooks, Cleanth, 66 convergent and divergent
Brown, Roger, 31 sound patterns, 41-45, 206-207
Bruner, 56 rhythms, 102
Bruner and Kretch, 5 style, 223
Bühler, 56 convex, 155, 170, 182
Burke, Kenneth, 3, 70, 81, 82 Cook, Norman, 202
Carroll, Lewis, 147 Cooper and Meyer, 158
Caspi, Moshe, 6 couplet, 105, 116, 117, 118, 120,
categorial perception, 227 216, 225, 230
caverns, 25, 27-30, 33, 39, 46, 47, creaky voice, 184, 185, 186 see also
49-50, 51, 53, 72, 80, 81 incomplete phonation, voice
Chafe, 143-145, 202 quality
chaotic overdifferentiation, 214 Crowder, 40, 219
Chatman, 102, 105-106, 110, 135, Crowder and Morton, 219
155, 213 crucial recommendation, 50, 73
Chevreuil, 222, 224 Dalbiez, 63-65
chiasmus, 195 Damasio, 5, 6, 14, 208-209
Chinese-box arrangement, 199-200 Dante, 71, 72, 91
Chomsky and Halle, 103 Darwin and Donovan, 224, 225
closural decision style, 5, 6, 7, 9, 11, 14-19,
allusion, 119, 139, 140 77, 137, 141, 206, 210, 221, 232,
quality, 136, 139 233 aee also cognitive style,
closure (cognitive), 56-61, 232 personality style
delayed/rapid, 23, 30, 39, 40, 58, declination, 180
75, 232 depersonalisation, 92, 95, 96-97,
closure (poetic), 4, 108, 110, 118- 208-209
130, 136, 138-140, 147, 149,
Index 247
depth-psychological interpretation, evaluation, 23, 24, 71, 112, 132

49, 56 see also Freudian, Jungian evaluative ingredient, 21, 74, 112
depth-psychological model, 66 experiential vs. instrumental set, 6,
Descartes, 5, 14 8, 15, 112, 121, 220
diatonic musical scales, 200-202 see experimental phonetics, 39-40
also major and minor scales experimental rigour, 197, 210, 211,
Dickens, 57 213
Dionysian, 98 expressive sound patterns, 40-41
Dionysus, 37, 90, 75, 92 F-scale, 18
divergent, 59 see also convergent factual statement, 71, 73
and divergent factualism, 15, 61
metre, 103 Fellows, 131
passage, 225 Fergusson, Francis, 75
Donne, 105, 110, 200, 213 Fitzgerald, 117
double-edged, 40, 101, 102, 129, FK score, 218
134, 135 Fleissner, 46-49, 83, 219, 220, 232
double-edgedness of rhythm, 9 flexible vs. rigid personality, 6, 15,
triple-edged, 68, 102-103 23
ecstasy, 90, 92-100 floating, emotional potentials of,
ecstatic, 135, 208 91-92
effect, 101 focusing sound pattern, 40-41
experience, 92, 103, 131 Fónagy, Iván, 145, 146, 167, 168,
peak, 208 175, 183, 187, 192, 194, 197-
poems, 11, 91, 95, 135 199, 203
poetry, 15, 92, 109 foreward-pressing, 181, 182
quality, 11, 101, 103, 110, 231 Forster, 46
Ehrenzweig, 89, 108, 140, 222, 223, Frenkel-Brunswick, 3, 6, 7, 19, 60,
224-225 73, 102, 135
emotional Freud, 49, 54
attitudes, 199 Freudian, 49, 50, 54, 63, 99
experience, 177 Fruman, 48, 50-56, 57, 99
patterns, 53 Fry, D.B, 159
peak, 187 Garrett, Bever and Fodor, 150
quality, 2, 8, 10, 53, 102, 113, gating, 56
135, 155, 206, 207, 231 genre, 79
response, 102, 134, 135, 154 gestalt, 65, 45, 106, 130, 134, 206,
sound effects, 220 216, 219, 222, 223, 227
symbolism, 27, 53, 98 approach, 81, 128
emotive boundaries, 209, 224, 225, 230
attitude, 145, 146, 174, 196, 198 closed and open, 140, 229
crescendo, 92, 97, 191, 192, 208 constraints, 227
tones, 182 grouping, 216, 218
empirical aesthetics, 128 principles, 116
enjambed couplets, 225 qualities, 3, 4, 8, 115-141
enjambment, 86-87, 104, 146-151, rules of perception, 9, 229
155-156, 157-159, 161, 163, 164,
172, 196, 197, 206, 225, 230
theory, 1, 3, 8, 39, 40, 44, 106, Hilgard, 131

205, 214, 115, 138, 140, 150, Hill, John Spencer, 4-5, 228
218, 223, 226, 229 Holinshed, 20
poor/good, 4, 44, 89, 101, 106, Holland, Norman, 137
128, 129, 134, 218, 223, 225 Homer, 25
gestalt-free, 81, 82, 88, 97, 103, House and Wichmann, 151
140, 218, 219, 222, 231 House, Humphrey, 38, 84
gesture, 155 Hrushovski, 40, 41 see also Harshav
articulatory gesture, 159, 163, humpback, 165, 166, 167, 174, 182,
166, 177, 178, 186, 197, 198 185
vocal gesture, 145, 167, 181, 188, Hunt, Leigh, 4
189, 190, 191, 195, 198, 199, hypnotic, 6, 42, 121
200, 207 effect, 130, 213
Gielgud, 10, 144 poem, 2, 66-70, 143, 197, 206,
Glicksohn, 121, 131, 132, 133, 213, 212
214 poetry, 1, 3, 4, 8, 43, 110-111,
Glicksohn, Tsur and Goodblatt, 115 130-137, 209-233
glottal stop, 183 quality, 2, 66, 209-233
God experience, 91-92 rhythm, 152, 160, 167, 207
Goethe, 83, 86 suggestions, 68
Goldstein, 18-19 susceptibility, 121, 131, 209, 213,
Goldstein and Scheerer, 19 227
Gombrich, 31, 56-58, 60 hypnotic-ecstatic quality, 6, 10, 80,
Goodblatt, 115 111, 112, 151, 188, 191, 199,
Goodman, Nelson, 211 209, 228
Gray, 66, 79 iconic representation of the clinging
grouping, 64, 104, 105, 108, 116- to an object or a memory, 175-
130, 135-137, 151, 158, 213, 176
214, 217, 223, 230 impersonal syntactic structures, 23
half whisper, 175, 187, 192 see also see also depersonalisation
incomplete phonation, voice implausible, 232 see also plausible
quality, whisper implied
Halle and Keyser, 101 author, 14
Halliday, 39 critic, 6, 7, 11, 14-19, 77, 206,
Harshav, 232 210, 232, 233
Hartley, 62 reader, 7
Harvey, 17-19, 22, 31, 35, 48, 60, Impressionism, 17, 61, 70
232 inaccessible reality, 94, 95, 190,
Haskins Laboratories, 40, 200, 220 200, 207, 209
hat patterns, 147, 149, 157 incantatory tone, 110, 154, 156,
Hazenfus, Martindale & Birnbaum, 160, 161
122, 132 incomplete phonation, 175, 176,
Hazlitt, William, 4, 5, 156, 228 185, 187, 188, 192, 198, 199 see
Helmholtz, 222-223 also breathy voice, creaky voice,
Hepburn, 131 half whisper, voice quality and
Herrnstein-Smith, 110, 116-120, whisper
138, 190, 194, 195, 211, 224 insight, 65-66, 92, 95
Index 249
interpretation, methodology, 70-76 leveling-and-sharpening, 6, 8-9, 15,

interpretive 50, 128-130, 140-141, 219, 227,
community, 9 228
strategies, 115 liberal vs. authoritarian personality,
intonation, 2, 143-203, 225 15
manipulating intonation contours, Liberman, (Al) et al., 219
200-203 Liberman, (Isabelle) and Mann, 220
instrumental investigation, 143-203 Liberman, Al, 219
irony, 4, 35-36, 66,198 Lieberman, Philip 166
irrelevant details, 32 literature of extreme situations, 77,
Jabotinsky, 121, 216 , 210
Jackendoff, Ray, 212 Lowes, 25, 61-69, 79, 90
Jakobson, 42, 107-108, 112 major and minor scales in music, 202
Jennings, 153-203, 207 see also diatonic musical scales
Johnson, Dr. Samuel, 79 Mann, Virginia, 220
Johnson-Laird, 212 mannerism, 164
Jung, 30 Margolis, 3, 70, 71-75, 113, 205
Jungian, 30, 49, 50, 54, 55 meditative and mystic poetry, 96
Kafka, 77 merely possible, 19, 48, 73, 82
Kane, Julie, 14 Mestral, Marie-Claude, 198
Kant, 8, 80-81, 84, 233 metalanguage, 146, 205, 231
Kantian microscope analogy, 233 metaphor, 92, 119
Keats, 6, 15, 76, 82, 92, 96, 102, Metman, 55
103, 139, 200 metre, 5
Kékulé, 63, 65 Meyer, Leonard B., 116-120, 136,
Kipling, 82 194, 195, 214, 224
Klanfer, 57 Miller, George, 114
Klein, George, 9, 140 Miller, J.G., 23
Kna'an-Kedar, 9 Milton, 28, 42, 79, 101, 108
knee, 155, 167, 170, 174, 176, 182, modernist, 12
185, 190 monorhyme, 214, 215, 230
Knight, Wilson, 38, 59, 70, 88, 93 Monteiro, Macdonald and Hilgard,
Knowles, Gerry, 163, 165, 183 131
Köhler, 65-66 Morton, Eugene, 157
Kreitler & Kreitler, 120 Multidimensional Personality
Kumar & Pekala, 131 Questionnaire (MPQ), 122, 132
Ladd, 147, 180 mysticism, 95
Ladefoged, 185 naive quality, 101
Lamb, 4 Napoleon, 83
late peaking, 148-151, 158, 159, negative capability, 6, 11, 15, 26-
162, 163, 174, 181, 182 30, 93, 102, 103, 210, 228
lateral inhibition, 40, 44 Neo-Classicists, 117
lateralisation, 13, 107 neurobiological, 6
law of good continuation, 116-117, neuropsychological, 6, 91-92, 96
128, 227 neuroscience and evolutionary
law of return, 116, 117, 118, 227 biology, 5
Lemaître, Jules, 16 neutral sound patterns, 40-42, 232
New Criticism, 7 perceptual organisation, 117, 120

Newberg, D'Aquili and Rause, 100, performance,
228 hypnotic, 149
Nietzsche, 97, 98 mental 5, 9, 39, 43-45, 136-141,
noisy attributes, 58 206, 210, 211, 216, 217, 225,
numinous, 52, 91, 157, 177 227, 228, 229, 230, 233
awe, 191 vocal, 143-203, 205, 206-209,
nursery rhyme, 101, 109, 214 225
obtrusive, 151, 152, 162, 193 rhythmical, 10 see also vocal
rhythm, 100, 103-109, 135, 149, performance
160, 161, 188, 214 periodic sounds, 43
quality, 160 periodicity, 180-181
Occam's Razor, 14 Persinger, 91-92
Oedipus, 59 personality style, 6, 7, 8, 10, 11, 57-
Ohala, John, 157 77, 141, 209, 217, 219-220, 226-
Ohmann, Richard, 16, 42, 210, 219 233 see also cognitive style,
Omar Khayyám, 3, 8, 117-130, 139, decision style
140, 215-217, 228 personality theory, 227
onomatopoeia, 39-49, 221, 232 personality variable, 1, 2, 7, 8, 57,
open vs. closed mind, 6, 15 121, 140
opium, 11 philology, 83
orbitofrontal cortex, 5, 14 Piaget, 18
organised whole, 134 Plato, 28, 36-38, 90, 92, 98, 177
Ornstein, 13-14 Platonic censor, 5, 102, 103, 106,
Orphic cults, 37, 90, 92 108, 109, 111, 112, 135, 136, 137
ostensión, 107-108, 112, 173 plausible, 3, 30, 36, 71-76, 113, 197
Otto, Rudolf, 52, 157 see also implausible, merely
overdetermination, 148 possible
overtone interaction, 45 Plutchik, 100
Pack, 153, 164, 167, 170-203, 207 Poe, 42, 48, 67, 69, 111
panda joke, 166 poetic effects, 214
parallel languages, 144 poetic mode of speech perception,
paradox, 222 218, 231
Pater, Walter, 16 poetic quality, 61
Patterson, 83 poetic rhythm, 2, 158
peak experience, 91, 177 Poincaré, 63
Pears, David, 233 Polo, Marco, 34
Pearson, 133 Pope, 15, 68, 101, 104, 105, 109,
Pekala et al, 121 110, 117, 213
perceived effects, 118, 138, 210, Porphyry, 25
211, 214, 226, 146 positions (metrical),
perceived pitch, 200-203 weak and strong, 40
perceived quality, 3, 8, 9, 11, 66, stress maxima in weak positions,
115, 131, 145 200-202
perception and personality, 5-7 positivism, 15
perceptual constancy, 89 post-hypnotic suggestions, 111
perceptual force, 148, 195, 196
Index 251
precategorial auditory information, nature description, 79, 83-87

see precategorial sensory traditions, 79
information Rorschach inkblot test, 218, 231
precategorial sensory information, rounded and unrounded contours,
4, 45, 89, 209, 217-225, 227, 164-170, 181, 182 see also
230, 231 concave, convex, humback, knee
prosodic organisation, 115-141 rounding effect, 182, 190
prosodic structure, 2, 86-87, 100- Rubáiyát, 3, 8, 9, 117-130, 138,
110, 143 215-217, 228, 229
prosody, 5, 40, 79, 214, 230 saturation, 116, 119-120, 135, 140,
psychoanalytical critic, 49 195
psychological models, 61-70 Schiller, 86
psychological reality, 150 Schneider, 26-27, 34, 37-38, 41-46,
Purchas, 20, 21-22, 24, 28, 33, 34, 48, 49, 57, 59-61, 62, 70-72, 79-
35, 112 80, 82, 90, 93-94, 98, 106, 108,
purposiveness without purpose, 8, 84 210, 212, 218, 219, 220, 223, 232
Quartermain, 156 security, false and genuine, 5, 102,
quatrain, 104, 108, 110, 116, 117, 106, 109, 111, 112, 135, 136, 137,
118, 119, 203, 230 206, 214, 217, 229
quest for certitude, 6, 11, 15-20, 29, Seneca, 28
77, 83, 93, 210, 219, 232 sexual symbols, 51, 52, 231
Racine, 117 shading, 218
Rader and Tellegen, 121 Shakespeare, 10, 20, 36, 61, 144,
Ransom, John Crowe 5, 102, 135 147, 177, 225
Ratosh, 132, 140, 214, 228, 230, Shaw, Bernard, 11, 18
231 Sheen, 146-203
readers, efficient and poor, 220-221, Shepard, Roger, 212
230 short-term memory, 9
realistic plenitude, 80, 94 Sibley, 50, 210
reductionism, 14, 144, 231 sign limit, 56, 57, 60
regional quality, 11, 50, 210 simplicity, 118
Renaissance, 79 simplifying qualities, 214
Repp, 40 simultaneous kinds of consciousness,
requiredness, 86, 146, 147, 149, 208-209
157, 159 simultaneous speaking voices, 170-
retro-relate, 97, 108 172, 188-190, 199, 207-209
rhyme, 39, 40, 41, 43-45, 105, 106- Sluijter, Agaath, 159, 186
109, 110, 115-141, 143, 151, Snyder, 4, 43, 66-70, 110-111, 121,
203, 209, 211, 213-217, 224, 229 130-132
rhyme patterns, 1, 2, 3, 8, 86, 227, Snyder and Shore, 121, 131, 132
229 Socrates, 37
Richards, 47, 130 solution of a problem, 134
Richardson, 153-203 sonnet, 93, 212
Rimbaud, 20 sound patterns, 210
ritual, 228 Spanos, 131
romantic, 12 Spanos, Brett, Menary and Cross,
narrative poem, 93 121, 131
speech mode of aural perception, tolerance of discordant attitudes, 59

218 top-down and bottom-up, 30, 209
speech perception, 45, 229 trance-inductive see hypnotic
speech processing, 4 Tsur, 132, 134, 144, 196, 197, 212,
spell-weaving, see hypnotic 215, 218, 221, 224, 225, 233
Spenser, 79 Tsur and Benari, 14
stratified system of norms, 205, 206 Tsur, Glicksohn, and Goodblatt, 115
structural knowledge, 7, 10, 115 trough, 165
structuralist models, 39, 40 true polyphony, 223
sublime, 80-81, 84, 85, 88, 95 Tsur, 14, 40, 43, 55, 77, 86, 89, 91,
supersensuous presence, 81 96, 101, 104, 109, 121, 129, 131,
Sutherland, Keith 6 typical male pitch range, 187, 152,
Swinburne, 4, 5, 67, 111 153, 157
symbol, 3, 11 typical female pitch range, 152,
and allegory, 11, 15, 17 153, 157, 167, 169, 187
and the ineffable, 11-14 Valentine, 130
mongering, 49 verbalised music, 4, 209, 218
symbolic vestibular sensations, 92
criticism, 59, 60 voice quality, 152, 157, 171, 172,
interpretation, 26-30, 70 174, 175 see also breathy voice,
meaning, 60, 87, 113 creaky voice, half-whisper,
Symbolism, French, 16 incomplete phonation, whisper
symmetrical, 118 Warren, Robert Penn, 70
synaesthesia, 121 Watson, 36-38, 83, 94
synecdoche, 107, 112 Weisstein, 17
tale-teller, 151, 153, 154, 156, 161, Weitz, 3, 70, 71-73
172 Wellek, 66
Tellegen, 112, 115, 121, 122, 132, Wellek and Warren, 205
220 Wertheimer, 116
Tellegen and Atkinson, 121 whisper, 184, 185, 186, 193 see also
temporal lobe transients, 92, 96 half-whisper, incomplete
Tennyson, 47, 67, 104, 111 phonation, voice quality
text-active and reader-active, 137, Wimsatt, 32, 50
139 Wimsatt and Beardsley, 69
theorising and experimenting, 211- Wittgenstein, 73, 233
217 witty poetry, 101, 109, 135, 229,
thing-free, 28, 81, 82, 87-92, 97, 230
103, 112, 140 Wordsworth, 91, 109, 156
think-aloud experiments, 7 Yarlott, 20-25, 27, 32-36, 48, 74,
tolerance and intolerance of 84, 85, 97, 112-113
ambiguity / uncertainty, 6, 7, 9, Zeitgeist, 5
14, 15, 40, 42, 44, 76, 102, 135, Zemach, Eddie, 205
210, 213, 227, 229
tolerance and intolerance of
uncertainty see tolerance and
intolerance of ambiguity
In the series Human Cognitive Processing the following titles have been published thus far or are
scheduled for publication:
17 LANGLOTZ, Andreas: Idiomatic Creativity. A cognitive-linguistic model of idiom-representation and idiom-

variation in English, xii, 315 pp. + index. Expected April 2006
16 TSUR, Reuven: 'Kubla Khan' - Poetic Structure, Hypnotic Quality and Cognitive Style. A study in mental, vocal
and critical performance. 2006. xi, 252 pp.
15 LUCHJENBROERS, June (ed.): Cognitive Linguistics Investigations. Across languages,fieldsand philosophical
boundaries. xiv, 328 pp. + index. Expected April 2006
14 ITKONEN, Esa: Analogy as Structure and Process. Approaches in linguistics, cognitive psychology and
philosophy of science. 2005. xiv, 249 pp.
13 PRANDI, Michele: The Building Blocks of Meaning. Ideas for a philosophical grammar. 2004. xviii, 521 pp.
12 EVANS, Vyvyan: The Structure of Time. Language, meaning and temporal cognition. 2004. x, 286 pp.
11 SHELLEY, Cameron: Multiple Analogies in Science and Philosophy. 2003. xvi, 168 pp.
10 SKOUSEN, Royal, Deryle LONSDALE and Dilworth B. PARKINSON (eds.): Analogical Modeling. An
exemplar-based approach to language. 2002. x, 417 pp.
9 GRAUMANN, Carl Friedrich and Werner KALLMEYER (eds.): Perspective and Perspectivation in Discourse.
2002. vi, 401 pp.
8 SANDERS, Ted J.M., Joost SCHILPEROORD and Wilbert SPOOREN (eds.): Text Representation. Linguistic
and psycholinguistic aspects. 2001. viii, 364 pp.
7 SCHLESINGER, Izchak M., Tamar KEREN-PORTNOY and Tamar PARUSH: The Structure of Arguments.
2001. xx, 264 pp.
6 FORTESCUE, Michael: Pattern and Process. A Whiteheadian perspective on linguistics. 2001. viii, 312 pp.
5 NUYTS, Jan: Epistemic Modality, Language, and Conceptualization. A cognitive-pragmatic perspective. 2001.
xx, 429 pp.
4 PANTHER, Klaus-Uwe and Günter RADDEN (eds.): Metonymy in Language and Thought. 1999. vii, 410 pp.
3 FUCHS, Catherine and Stéphane ROBERT (eds.): Language Diversity and Cognitive Representations. 1999.
x, 229 pp.
2 COOPER, David L.: Linguistic Attractors. The cognitive dynamics of language acquisition and change. 1999.
xv, 375 PP.
1 YU, Ning: The Contemporary Theory of Metaphor. A perspective from Chinese. 1998. x, 278 pp.

Kubla Khan - Poetic Structure, Hypnotic Quality and Cognitive Style - A Study in Mental, Vocal, and Critical Performance

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

Kubla Khan - Poetic Structure, Hypnotic Quality and Cognitive Style - A Study in Mental, Vocal, and Critical Performance

Uploaded by

Copyright:

Available Formats

'Kubla Khan'

Poetic Structure, Hypnotic Quality and Cognitive Style

Editorial Advisory Board

John Benjamins Publishing Company

PR4480.K83 T78 2006

© 2006 - John Benjamins B.V.

Kubla Khan: Or a Vision in a Dream, a Fragment

Part One: Criticism and Meta-Criticism

Part Two: The Reader and Real Readers

Part Three: Vox Humana

Then all the charm

Poor youth! who scarcely dar’st lift up thine eyes—

But oh! that deep romantic chasm which slanted

Five miles meandering with a mazy motion

1 To sing a sweeter song tomorrow.

And sank in tumult to a lifeless ocean:

A damsel with a dulcimer

tuting a wide range of dichotomic scales as expounded by psychologists of

more comprehensive theory. Second, I greatly extended the scope of my

responses depend, largely, on the reader's co-operation. Hence the multiplicity

different verdicts from different readers, guided by a Zeitgeist. My assumption

acted in a decidedly irrational way" (Sutherland, 2001: 81). I began to explore

because the critics seem to be incapable "of being in uncertainties, Mysteries,

sessed on a questionnaire (though I expect to find some affinity between the

structure of the Rubáiyát bars a leveling strategy, whereas sharpening may

It is quite reasonable to suppose that students in different disciplines have

The only man who behaved sensibly was my tailor: he took my

Symbols and the Ineffable

an unsayable reality. Precisely this is said to be the difference between symbol

is characterised by a translucence of the special in the individual, or of the

In a paper abounding in wise formulations, Peter Berek (1978:121) makes

Symbolism is a literary resource based on a metaphysical assumption: the

And later, again,

Symbolism is perhaps a yearning after allegory in the absence of positive

Coming back now to Coleridge's passage, it emphasises the revelation "of

It will be noted that I have substituted revelation for translucence when

The Implied Critic's Decision Style

When in a piece of criticism, or in the output of a critic, certain cognitive

which Shakespeare possessed so enormously—I mean Negative Capability,

In that paper I attempted to show that in the middle of that spectrum of

The leveler is more anxious to categorise sensations and less willing to

Un symbole est, en somme, une comparaison prolongée dont on ne nous

I do not pretend to know anything about Lemaître's cognitive style or, in

* A simpler cognitive structure, comprised of fewer differentiations and

The categorical or conceptual attitude is characterised by ability or readi­

and to isolate them voluntarily, to abstract common properties, to plan

This can be taken for a description of the smooth, undisturbed functioning

If we restrict ourselves to what is 'given', appealing to the poem as a

In Xamdu did Cublai Can build a stately Palace, encompassing sixteene

Yarlott comments on this passage: "Coleridge seems to have deliberately

The substitution of "bright/sinuous" for "pleasant/delightful" produces sin­

I can only commend this critical technique of comparing a poem to its

quite clear. Purchas uses adjectives in which positive evaluation is unambig­

According to the conception suggested here, neither expression should be

one's enjoyment of it. Whether it is morally undesirable, or whether the

This is a way in which some organisms handle the problem of ignorance by

The bald ambivalence of "did Kubla Khan / A stately pleasure-dome

neoplatonic mysticism, they pretend to account for the poem's ecstatic or

Negative Capability and Symbolic Interpretation

One of the most illuminating studies of "Kubla Khan" is Schneider's (1975)

This air of importance without visible foundation contributes to the sug­

This paragraph follows a long discussion of the foregrounding of dome in

7 This phonetic discussion too is indicative of an attitude of negative capability.

The categorical or conceptual attitude is characterised by ability or readi

The substitution of "bright/sinuous" for "pleasant/delightful" produces sin

quite clear. Purchas uses adjectives in which positive evaluation is unambig

This air of importance without visible foundation contributes to the sug

Examining my own response to the cavern image, as it occurs in Co

There are dozens of parallels in Renaissance English to this account of po

exerting lesser or greater pressure on one's tolerance of uncertainty. Ono

ALTERNATIVE MENTAL PERFORMANCES. Schneider is one of the few liter

That is puzzling. I know what kind of sound the "murmuring of innumer

Features of this kind, appearing in reveries or dreams, easily lend them

One important distinction this passage makes is between archetypal con

my 1975 paper I discussed at length Eva Metman's Jungian archetypal inter

One important activity involved in interpretation is the attempt to abstract in