THE MARGINS

OF DISCOURSE
The Relation of

;
Literature to Language

BARBARA
ERRNSTEIN S,
 ON THE MARGINS OF DISCOURSE
The Relation of Literature to Language
Barbara Herrnstein Smith

"Barbara Herrnstein Smith is one of few out-

a very
standing theorists of literature writing today, and
her new book. On the Margins of Discourse: The
Relation of Literature to Language, is a major contri-
bution to the contemporary critical debate. She
writes with a rare combination of elegance, verve,
and logical precision. The issues she deals with are
among the most central in the theory of literature,
and many currently fashionable notions do not sur-
vive her incisive analysis. This book is a joy to read."
— John M. Ellis, University of California, Santa Cruz
In this centrally focused collection of articles and
lectures, Barbara Herrnstein Smith examines a funda-
mental problem of literary theory: the location of its
own subject, "literature." Through an analysis of the
dynamics of verbal behavior, she argues that while
terms such as "literature," "fiction," and "poetry"
resist clear-cut and stable definition we nevertheless

learn to make functional distinctions among various

verbal acts and events. Smith asserts that an appre-
ciation of the nature and significance of those dis-
tinctions is crucial to our understanding of literature
and to the methods and goals of literary study.
In Part One, Smith introduces the distinction be-
tween natural discourse and fictive discourse—verbal
structures that function as representations of natural
utterances. She also deals with the relation of utter-
ances to inscriptions and of literary mimesis to repre-
sentation in other art forms In the title essay. Smith
explores a number of borderline cases, including
"found poetry" and proverbs, and considers the
distinctive ways in which we experience and inter-
pret fictive verbal structures.

(Continued on back flap)

 ON
THE MARGINS
OF DISCOURSE
i
 ON
THE MARGINS
OF DISCOURSE
The Relation of
Literature to Language

BARBARA
HERRNSTEIN SMITH

The
University of Chicago Press \
Chicago & London
The University ofChicago Press, Chicago60637
T HE U NI VERSITY OF C HICAGO P RESS, LTD., LONDON
© 1978 by The University of Chicago
All rights reserved. Published 1978
Printed in the United States of America
82 81 80 79 78 543 2 1

Barbara Herrnstein Smith, professor of

English and Communications at the University of
Pennsylvania, is the author of Poetic Closure, also
published by the University of Chicago Press,
which won the Phi Beta Kappa Christian Gauss
Award and the Explicator Award for 1968.

Library of Congress Cataloging in Publication Data

Smith, Barbara Herrnstein,

On the margins of discourse.

Includes bibliographical references and index.

1, — Philosophy.
Literature 2, Style, Literary.

3. Language and languages — Style. 1. Title.

PN54.S6 809 78-18274
ISBN 0-226-76452-4
For Julia and Deirdre,
exemplary daughters
4
 CONTENTS
1

Preface
ix

Part I: Toward a Theory of Fictive Discourse

Literature as Performance, Fiction, and Art 3

Texts and Scores 3
Representations of Discourse 8
Art and Cognitive A ctivity 1

Poetry as Fiction 14
Natural Discourse 15
Fictive Discourse 24

On the Margins of Discourse 41

Definitions and Classifications 41
Fictive Illustrations and Examples 50
Exploiting the Margins 55
Prefabricated Discourse 57
Quotations 64
Saying and Sayings 69
Part II:Exchanging Words: On the Economics
4
AND Ethics of Verbal Transactions
The Witter Bynner Lectures in Poetry for 1977

In the Linguistic Marketplace 79

The Speaker: Utterances as Acts 85
The Listener: Utterances as Events 92

5 7Licensing the Unspeakable 107

and Conventions 107
Constraints
Verbal Art and Cognitive Play 116
Children at the Gates of the Marketplace 124

6 The Ethics of Interpretation 133

Determinate and Indeterminate Meanings 137
Authorial Intentions and Literary Ethics 146

Part III: Linguistics and Literary Theory

Surfacing from the Deep 157
Syntactic Strategies and Analytic Formulas 158
Sound and nonSense 162
The Limits of Linguistic Determinacy 167
Structuralism and the Linguistic Model 1 74
Generating Theories of Narrative 185
Conclusion 197

Notes 203

Index 217
 PREFACE

“And,” quipped a friend, “will you tell us next about poetic

aperture?'' Although I replied at the time (it was some years ago)
that I had no such plans, “aperture” could indeed be seen as the
subject of the present volume: not in the sense of openings but
rather of unendings, interminables, indeterminacies. The turn from
closure* to openness did not, however, require a turnabout, for, as
this study will suggest, it is by virtue of its enclosure that the poem
achieves its amplitude and infinitude. From within the circle that

marks its margins, the poem extends a parabolic curve; disengaged

from historical enactment, it persists in history and acquires
historicity through the continued possibility of its reenactment. Like
Yeats’s golden “form,” it may, “once out of nature ..., sing of

what is passed, or passing, or to come.”

This study was initiated in 1969; a version of it, then entitled Fictive
Discourse, was completed in 1972. The greater part of the present
volume consists of portions of that original manuscript, successively
extracted, expanded, and revised for presentation as public lectures
or publication as individual articles. The germ of the study and an
anticipation of its major topics and arguments appear in the first

essay, “Literature as Performance, Fiction, and Art,” which,

*Poetic Closure: A Study of How Poems End (Chicago, 1968).

IX
 —
PREFACE

though composed for a different occasion, can now be read as an

introduction to the entire volume. Originally the text of a paper
delivered to the American Philosophical Association, it considers a
number of questions raised by Nelson Goodman’s Languages of
Art — and, more generally, by traditional philosophic aesthetics
concerning the relation of literature both to language and to other
artforms. In the course of the paper, 1 propose an alternative
conception of that relation which involves, among other things, the
drawing of a set of distinctions between the texts of verbal artworks
and other texts. That particular problem —the possibility and
propriety of drawing such distinctions, the clarification and refine-
ment of their formulation, and the pursuit of their implications for
literary, linguistic, and aesthetic theory —continued to absorb my
attention and eventually became central to the study that here
completes itself.

Because I was extracting portions of the manuscript for publica-

tion at the same time as I was expanding and revising it (and, over
the course of six years, my views on a number of points themselves
underwent revision), the individual pieces now reassembled are not
in altogether genial alignment with each other. I have not, in

preparing this volume, undertaken the sort of wholesale revision

that would have been required to so align them, but I have filed

some of the rougher edges and, in bracketed footnotes, indicated

where points are developed further or significantly modified in later

portions.

Some difficulties posed by the two key terms of the study, natural
discourse and fictive discourse, may be eased if I briefly discuss
their origin. As noted above, the distinction was initially proposed in

a discussion of the relation of literature to other artforms. Ac-

cordingly, the term natural was chosen as an allusion to the
traditional opposition of “art” and “nature,” in which each member
(as in many traditional oppositions) evokes, and to some extent
defines, the identity of the other. The term fictive had already been
used in Poetic Closure in the corresponding sense of “feigned” or
“represented,” that is, “artificial,” but without the suggestion of an
inferior substitute or deceptive imitation. Although the study focuses

X
 PREFACE

on the relation between fictive and natural discourse, that relation is

to be seen not as unique to language but rather as one instance of

the more general relation between “fictive” and “natural” objects
and events. The former include not only most of what we think of as
art and artifice —
for example, mask, costume, effigy, and illustra-
tion, as well as painting, sculpture, music, and dance —but also
much else that we think of as play, sport, or simply manufacture.
Like the fashioning of tools, the cultivation of plants, or the
domestication of animals, the creation of fictive objects and events
reflects our impatience with nature, with that which merely happens
to exist, happens to happen, or happens to “appear.” They are the
products of our impulse and ability to fabricate, by acts of hand,
eye, or mind, representations {s\m\x\2icvdi, images, quasi-instances) of
natural“phenomena”: that is, to construct or to make appear either
counter-forms of phenomena or counter-feit instantiations of types
of phenomena which, “in nature,” exist or occur independent of our
designs or desires.
Among the phenomena that we may, in this sense, “represent” (to

ourselves and to others) are objects and events that are themselves
“artificial,” both in being the products of human agency and also in
having been fashioned in accord with human designs and desires.

Thus, artworks, toys, and illustrations may represent not only

sunsets and oranges, but also hats, houses, and, indeed, artworks,
toys, and illustrations; and we may represent, in play or drama, not
only people sleeping and stumbling, but also people dancing,
speaking, and, indeed, playing games and performing in dramas.
In so representing them, however (even if the representation is

more strenuous than placing or

created by an act of artistry no
imagining a “frame” around them or placing or imagining them on
a “stage”), we make each of those objects and events, including
those already “fashioned,” serve novel or distinctive functions, and
thus we make them the instruments and products of our own
emergent designs and desires. While discourse itself is certainly the
supreme instrument and product of human designs and desires
and may therefore seem to be naturally artificial, nevertheless ut-
terances, like other objects and events, can be regarded as “natural”
relative to s\xc\\ representations of them (that is, “fictive utterances”)

XI
 —
PREFACE

as we may produce for special ends, aesthetic and other, and which
thus serve novel or distinctive functions.
“Natural discourse,” then, is not equivalent to “spoken language,”
nor is it a substitute for such expressions as “normal,” “everyday,”
or “ordinary” discourse. Similarly, the iQvmfictive, as used here, is

not equivalent to “false,” “imaginary,” or, for that matter, to

“fictional.” Both are terms that 1 have chosen to designate what I

later refer to as “covert categories,” that is, categories which,

though not consistently marked or labeled by the members of a
community, are nevertheless manifested in their behavior as func-
tional discriminations and defined and sustained by relatively

distinctive sets of conventions. (The term covert category is bor-

rowed from Benjamin Lee WhorT s Language, Thought and Reality
[Cambridge, Mass., 1956], pp. 87-101 passim, where it is used in a

similar sense.)
The second chapter in part 1, “Poetry as Fiction,” develops the
suggestion that verbal artworks — poems, plays, stories, novels
may be seen as representations of discourse and, in that sense, as
“mimetic” utterances. Although the term mimetic was later aban-
doned in favor of its alternative, //ct/ve, part of my concern in the

original essay was to suggest how the concept of mimesis, which

figures so largely in Western aesthetic theory and has so often been
invoked crudely or inappropriately with respect to literature, could
be used with some pertinence both to characterize verbal artworks
and to illuminate their relation to other artforms.

Although the concept of fictive discourse was initially developed

to that end, it was clear that the class of verbal structures that could
be described as “representations of utterances” was by no means
confined to “works of literature,” and that the general nature of
fictive discourse was itself a matter worth exploring further. Ac-
cordingly, in the title chapter, “On the Margins of Discourse,”
extensive consideration is given to various verbal structures (for ex-
ample, logician’s examples, advertisements, and quotations) which,
as fictive utterances, commonly serve functions other than “aes-
thetic” ones, and to various “borderline cases” (for example,
greeting-card messages and proverbs) that illustrate the functional
and conventional nature of the distinction between fictive and

Xll
 PREFACE

natural discourse and thus the potentially multiple functions and

contextually variable identity of all verbal structures.
A number of topics and issues touched on only briefly or inciden-
tally in the selections that comprise part 1 had been considered more
systematically in other portions of the original (now almost com-
pletely cannibalized) manuscript. Those portions were revised and ex-
panded for presentation at the University of California, Berkeley, as

The Witter Bynner Lectures in Poetry (“Exchanging Words: On the

Economics and Ethics of Verbal Transactions”), the text of which
forms part 2 of this volume. Among those topics and issues are (1)
the transactional nature of natural discourse and the distinctive
functions of language for speakers and listeners, (2) the joint
origins, in the dynamics or “economics” of verbal transactions, of
the conventions governing natural and fictive discourse, and (3) the
muddle of meaning (and thereby intention), both generally and with
respect to literary interpretation.
While the observations on language offered in this study do not
constitute an original or comprehensive linguistic theory, they do
reflect conceptions of verbal and symbolic behavior that are at
considerable variance with the theory of language represented by
transformational-generative grammar, and at relative distances
from theories or analyses of communication, speech acts, and
meaning offered by, among others, ordinary-language philosophers.
Some key points of those conceptions are made explicit and
elaborated in the first lecture, “In the Linguistic Marketplace.”
The second lecture, “Licensing the Unspeakable,” returns to
fictive discourse, specifically to the possibility (and, indeed, neces-
sity) of verbal “licentiousness,” as created by the economics of
verbal transactions and as learned and revealed in our earliest
experiences with the multiple functions of language. (The discussion
here of the relation between art and what I refer to as “cognitive

play” will be developed further in a study of aesthetic value which I

am now pursuing.)
The term interpretation is used throughout part 1 in a quite broad
sense, to refer to the process or activity of drawing inferences from,
or ascribing meanings to, verbal structures: a process or activity that
directs — and may, in the case of literary works, largely constitute

Xlll
 PREFACE

our response to those structures. As I point out there, the explica-

tions, scholia, “readings,” and so on offered by professional and/or
academic exegetes must be seen as very special cases of “interpreta-
tion,” for we all (including children, illiterates, and members of
cultures where no such professional class exists) must interpret

verbal structures if we are to respond to them at all, and even when

we do articulate our interpretations publicly, it need not be in the

service of instruction. A number of questions concerning meanings

(“determinate” and “indeterminate”) and literary interpretation (in

both the broad and narrow senses), particularly as raised by the

work of certain contemporary American critics and theorists (rep-
resented, synecdochically, by E. D. Hirsch), are the subject of the
third lecture, “The Ethics of Interpretation” —the title of which, as
will be seen, is somewhat ironic.

As a diversion from revising part 2 of this study, I undertook to

write a brief review of a collection of papers edited by Roger Fowler
under the title and Structure in Literature: Essays in the New
Style
Stylistics. The diversion became its own form of obsession and, by

the time it was completed, the “brief review” “Surfacing from the
—
—
Deep” was eighty pages long and something other than a review. It
appears as part 3 of this volume, the original text somewhat
abbreviated, and subsected and subtitled to suggest the various
topics considered.
A number of points made in parts and 2 are here extended in
1

directions not accommodated by the scope and momentum of the

earlier pieces, and are sharpened with respect to their implications
for the study of literature. The most direct extensions appear in the

discussion of various attempts by linguists and others to encircle and

net the parabolas of poetic meaning, and in a brief outline of how a
theory of literary narrative might be pursued in connection with a
more general study of the social and potentially aesthetic functions
of all verbal narrative. The whole of part 3, however, may be taken
as an extension of the conception of literary theory reflected in
Poetic Closure as well as in the earlier sections of this volume, that
is, not as metacriticism, philosophic meditations on literary texts, or
a corner of some other intellectual domain or discipline, and not as
justified by the practices or confined by the purposes of pedagogy;

XIV
 PREFACE

but rather as a continuously evolving inquiry into all the phenomena

and activities of which literature consists and to which it is related,
an inquiry that is responsive and connectible to the general study of
social and symbolic behavior, and that dissolves its own margins in

the very process of discovering them.

This volume is sparing in citational footnotes. I do, however, wish to

acknowledge here some of the more significant sources and con-
spicuous analogues of the study.
As the numerous domestic examples will suggest, many of the
observations made here, especially those on language acquisition
and verbal drawn from homegrown sources, among them
play, are
the two young women to whom this volume is dedicated. (I have,
however, been mindful of the tendency of children to acquire
language in obedient accord with whatever theory of language
acquisition their parents subscribe to.) Other sources, remote in
(personal) time and all but obliterated by contributory streams and
ultimate depositions, include my happy encounter, at an impres-
sionable age, with selected writings by Charles S. Pierce; Kenneth
Burke’s Counter- Statement (Chicago, 1931); and B. F. Skinner’s
Verbal Behavior (New York, 1957: a little-known work, not to be
confused with the well-known review of it by Noam Chomsky
[Language 35 (1959): 26-58]).
Analogues of the study, that is, recently published — or, in many
cases, recently translated —works that, to various extents, parallel or

intersect with it, include the following:

Studies originally published in the 1920s by Russian literary
theorists, especially the essays by Boris Tomasevskij (“Literature
and Biography’’), Jurij Tynjanov (“On Literary Evolution’’), and
V. N. Volosinov (“Reported Speech’’), translated and reprinted in
Readings in Russian Poetics, ed. L. Matejka and K. Pomorska
(Cambridge, Mass., 1971); Volosinov’s “Discourse in Life and
Discourse in Art,’’ reprinted as Appendix 1 in Freudianism: A
Marxist Critique, tr. 1. R. Titunik and ed. with N. H. Bruss (New
York, 1976); and Mikhail Bakhtin’s Problems of Dostoevsky's
Poetics, tr. R. W. Rostel (Ann Arbor, Mich., 1973);
Works by other literary theorists, particularly Roman Ingarden,

XV
 PREFACE

The Literary Work of Art (orig. pub. 1931), tr. George G. Grabo-
wicz (Evanston, 111., 1973); Robert Champigny, Ontology of the
Narrative (The Hague and Paris, 1972); Wolfgang Iser, The Implied
Reader (Baltimore and London, 1974) and “The Reality of Fiction:
A Functionalist Approach to Literature,” Nev^^ Literary History 7
(1975): 7-38; Roland Barthes, S/Z: An Essay (orig. pub. 1970), tr.

Richard Miller (New York, 1974); John The Theory of Literary

Ellis,

Criticism (Berkeley, Calif., 1974); and, encountered as this volume

was going to press, Mary Louise Pratt, Toward a Speech Act Theory
of Literary Discourse {^Xoormn^on, Ind., 1977);
Various studies of art and symbolic behavior, particularly David K.
Lewis, Convention: A Philosophic Analysis (Cambridge, Mass.,
1969); Gregory Bateson, Steps to an Ecology of Mind (New York,
1972); Erving Goffman, Relations in Public (New York, 1971) and
Frame Analysis (New York, 1974); Arthur Danto, “The Trans-
figuration of the Commonplace,” Journal of Aesthetics and Art
Criticism 33 (1974): 138-48; Umberto Eco, A Theory of Semiotics
(Bloomington, Ind., 1976); and, read in manuscript after this

volume was completed, Morse Peckham, Explanation and Power:

An Inquiry into the Control of Human Behavior (New York,
Continuum Books, in press).
A number of friends have given to portions of this study their own
often contentious but always valued attention: at Bennington Col-
lege, Alvin Feinman, Joanna Kirkpatrick, and Richard Tristman; at
the University of Pennsylvania, Larry Gross, Leonard B. Meyer,
Saul Morson, and the much-missed late Sol Worth. I am grateful,
also, to Tom Wick and Eric Reeves for their heterogeneous helpful-
ness as my research assistants, Thomas D. Cohen for his
to
relentless provocations, and to Morse Peckham, who put at my
service his uncanny knack for almost always, ultimately, being
right.

Various phases of the study were aided by leaves and grants from
Bennington College, the University of Pennsylvania, the National
Endowment for the Humanities, and the John Simon Guggenheim
Memorial Foundation. I should also like to thank the Witter Bynner
Foundation and the Department of English at the University of

XVI
 PREFACE

California, Berkeley, for their invitation to deliver the Poetry

Lectures for 1977 and for their hospitality during my visit.

Permission to publish edited versions of the articles and lectures

assembled here is gratefully acknowledged. The original texts
appeared as follows:
“Literature as Performance, Fiction, and Art,’’ The Journal of
Philosophy 47 (1970): 553-63. © 1970 by The Journal of Philos-
ophy, Inc.
“Poetry as Fiction,’’ New Literary History 2 (1971): 259-82.
Reprinted in A^ch’ Directions in Literary History, ed. Ralph Cohen
(Baltimore and London, 1974). © 1971, 1974 by New Literary
History.
“On the Margins of Discourse,’’ Critical Inquiry 1 (1975):
769-98. © 1975 by The University of Chicago.
“Exchanging Words: On the Economics and Ethics of Verbal
Transactions,’’ The Witter Bynner Lectures in Poetry for 1977,
delivered at the University of California, Berkeley, October-Novem-
ber 1977.
“Surfacing from the Deep,’’ PTL: A Journal for Descriptive
Poetics and Theory of Literature! {\911): 151-82. © North-Holland
Publishing Co., 1977.

xvii
i
TOWARD A THEORY
OF FICTIVE DISCOURSE
 ;
t.K
'
i/iAvvo
I »

•-ivj j
r \
 1

LITERATURE AS PEREORMANCE,
FICTION, AND ART

What Nelson Goodman on occasion calls his “heresies” struck me as

not only sound and welcome gospel but of considerable potential
value for literary theory. The otherwise inviting prospect of ex-
tending Languages of Art^ to the arts of language was initially

dimmed, however, by those extensions which Goodman himself

provides; for they are in certain respects quite problematic. My hope
in bringing one particularly disturbing set of problems to attention
here is him to relinquish some lingering orthodoxies and
to persuade
consider some of the more radical and unexpected implications of
Languages of Art for the theory of literature.

I. Texts and Scores

There are two senses in which we commonly speak of literature:

either as the class of linguistic texts or as the class of linguistic

artworks. Thus literature may include, on the one hand, everything
written down in orthographic inscription or, on the other, what is

sometimes awkwardly referred to as “poetry in the broad sense,”

meaning that the distinctive feature is not meter. Although the two
classes overlap, one is not a subclass of the other. Poems are not
certain kinds of linguistic texts; they are certain kinds of linguistic
structures, and inscriptions of them may or may not exist. Since the
distinction I am speaking of is very gross, and since Goodman is a

3
 TOWARD A THEORY OF FICTIVE DISCOURSE

very subtle man, his failure to heed it in Languages of Art must be

taken to reflect not a lack of discernment but rather a deliberate
decision. I shall attempt to suggest here why, given his objectives
and concerns, this decision was mistaken.
In the fifth chapter of the book, “Score, Sketch, and Script”
(pp. 177-221), Goodman examines the relation between a number
of standard artforms and their respective notational systems, actual
or conceivable. He considers in turn music, painting, dance,
architecture, and, along the way, something which he refers to

variously as “literary arts,” “literary works,” and “works of litera-

ture,” and which he exemplifies variously as poems, novels, biog-
raphies, and histories of the Civil War. In view of the examples, it

would seem that he wants us to understand “literary” here in the

relatively undiscriminating sense of written down —which is sur-

prising, of course, in the context of an examination of the various

arts.

It follows from Goodman’s use of the term literary that his

analysis of “the literary arts” applies to all printed texts from the
First Folio Shakespeare to How to Travel in Europe on $5 a Day,
and also to the note I left this morning for the milkman, but does
not apply to a preliterate epic saga or the love poem of an Adaman
Islander or the animal fable spun by an unlettered Appalachian
farmer. Goodman declares that he is “not concerned with what
distinguishes some scripts as ‘truly literary’ works,” but the dis-

claimer does not answer the objection that he should have been
concerned with a distinction very much like that. For my point here
is not that he failed to assume or ensure an honorific definition for
literary art, but rather that he failed to recognize the integrity and
significance of what he would call its relevant “antecedent classifi-

cation,” and that he consequently failed to account for literature

properly in the terms of his own theory.
Goodman might offer to allow for the illiterate’s work of litera-
ture by noting that, for it to qualify, a text of the work need not
someone could produce an inscription for it
actually exist so long as
which, when read aloud, would be accepted as an instance of that
work — presumably in that culture or community. This would not
help much, however; for, given any orthographic system, anything

4
 LITERATURE AS PERFORMANCE, FICTION, AND ART

speakable is also inscribable. Consequently, some other lacking

principle of distinction, every utterance in any linguistic community
that had at least one literate member would be a work of literature.
Goodman has either excluded from his class of “literary works” all
the poems and tales that have never been inscribed, or he has
included as literature virtually everything that has ever been uttered.
Goodman’s initial identification of literary art with written texts is

either responsible for or a reflection of his equally disturbing

identification of a poem with the class of its inscriptions and/or
utterances. I say “and/or” because it is not clear whether he means,
as he says in one place, that “a ... literary work is not the
compliance-class of a text but the text or script itself” or, as he says
in another place, that “all and only inscriptions and utterances of
the text are instances of the work.” In either case, the identification
seems improper, because, aside from concrete poetry (that is,

“picture poems” or verbal constructions dependent on graphic

presentation for their formal properties) and one other important
class of exceptions to be mentioned later, the poem or literary
artwork cannot be identified with its own inscription. Moreover, it

may be identified with the utterance of its inscription only if one

understands “utterance” in the sense of “performance” to be
described hereafter. Given Goodman’s own definitions and anal-
yses, the relation between a poem and its text is readily and, I think,
happily conceived as analogous to the relation between a musical
work and its score or between a dramatic work and its script. Where
a text exists, a “genuine instance” of the poem is a performance
defined by and complying with it. When there is no performance
same ontological status as a score. Thus I
occurring, the text has the
regard as baffling Goodman’s assertion that “an unrecited poem is
not so forlorn as an unsung song.”
Imagine, if you will, that a court musician has performed an air,
both the lyrics and melody of which he has composed himself.
“Lovely,” says the Queen, “we must have a copy.” What does the
musician do? He prepares a score and a text; or, rather, he prepares
a score of the entire song, in which certain features are inscribed in

one symbolic scheme, musical notation, and other features are

inscribed in another symbolic scheme, linguistic orthography. It is

5
 TOWARD A THEORY OF FICTIVE DISCOURSE

not clear why, at this point, or in this state, one inscription should be
thought more forlorn than the other. To produce a genuine instance
of the someone must produce a physical realization of both
air,

words and melody; merely playing the lute and displaying the text
will not do. Moreover, the relation of the text to the work would not

be different if composed only the lyrics that is, a

the composer had —
—
poem and had never intended to set them to music. Or, more to
the point, we should then say that all the music he had designed for
that poem was to be found in its phonetic and prosodic features
insofar as they were specified by the conventions of interpreting
poetic texts. That music could not be found in the text itself because
texts are altogether mute objects. To translate the inscription of that
lyric into an instance of the work, something has to serve as the
instrument of its performance.
It is evident from Goodman’s discussion that he had considered
the possibility of regarding poetry as a mediated or “two-stage” art,
with texts functioning as true scores. He rejected the possibility,
however, because although he could grant that vocal readings might
be construed as performances, the phenomenon of silent reading
apparently remained an insoluble problem. That the problem was
partially created by his own misconception of the matter is suggested
by his argument that, if silent readings of a poem could be
considered instances of a work, then so could “the lookings at a
picture and the listenings to a [musical] performance.” A silent

reading of a poem, however, is or may be a much more specific and

precisely determined activity than looking at a picture or listening to
music. The reader is required to produce, from his correct “spell-
ing” of a spatial array of marks on a page, a temporally organized
and otherwise defined structure of sounds — or, if you like, pseudo-
sounds. The physical or neurophysiological source of the structure
generated by the silent reader is of little significance here: it may
originate somewhere in his musculature or peripheral or central
nervous system, or the source may vary from reader to reader. What
is significant is that the structure itself will not vary.
Although, in a silent reading, the performer and audience are
necessarily the same person, this should not obscure the fact that the
reading consists of two theoretically distinct activities, only one of

6
 LITERATURE AS PERFORMANCE, FICTION, AND ART

which is comparable to listening to music or looking at a picture. To

be sure, the configuration of physical events that constitutes a
genuine instance of an artwork is usually external to the audience’s
own body or sensory system: the music originates outside our ears,
the painting is The surface of the skin,
located outside our eyes.
however, is an arbitrary boundary line here, and if we can conceive
the solitary singer enjoying his own performance, we should not
really have any trouble extending the conception to the solitary silent

reader.
Goodman’s observations on the limits of scores are of consider-
able interest when extended to poetic texts. “The function of a
score,’’ he writes, “is to specify the essential properties a per-
formance must have to belong to the work; the stipulations are only
of certain aspects and only within certain degrees.’’ This is a point
that could be heeded with profit by contemporary linguists and
others who lament what they see as the phonetic inadequacy of texts,
especially in indicating the “correct’’ degree or position of metrical
stresses. But phonetic equivalence is no more required of correct
performances of poems than of musical compositions. What is

required, rather, is compliance with the text according to the

conventions of interpretation established by the tradition within
which the poem and text are composed and interpreted.
It should nevertheless be noted that the conventions for the
interpretation of poetic inscriptions are not the same as those for the
ordinary reading of discursive texts. Not every text is a score,
because not every linguistic inscription is of a literary artwork. One
cannot skim a poem or read it distractedly while listening to a
conversation: not, that is, if one is to produce and experience it as an
artwork. For the structure of sounds and temporal pacing in a poem
are part of what constitutes the work, and the poem has not
occurred unless that pacing and structure have been made manifest
to its audience.
It is curious, in this connection, that Goodman never mentions
one rather striking feature of certain texts, namely their lineation. I

trust he is not under the impression that the lines in which poems are
printed are merely a typographic ritual, or that they identify poems
as a genre but do not otherwise control or define their compliance-

7
 TOWARD A THEORY OF FICTIVE DISCOURSE

classes. For lineation is an integral part of the notational system of

poetry; and although it does not control the rhythmic features of a
performance as specifically as do the indications of timing in

musical notation, it nevertheless does direct and limit them. Con-

sequently, given certain traditionally authorized rules of interpreta-
tion, any performance that does not comply with a poem’s lineation
is not an instance of that poem.

II. Representations of Discourse

I would like to return now to the class of exceptions mentioned
earlier. An made for regarding certain
interesting case can be
literary works as constituted by their own inscriptions, and, al-

though it is not the case that Goodman makes, I think he might find
it congenial. Briefly, the argument would go as follows. As a general
class, literary artworks may be conceived of as depictions or
representations, rather than instances, of natural discourse. (An
analogy may be drawn to the familiar Aristotelian conception of
drama as the representation of human action. Hamlet is a represen-
tation, not an instance, of a man avenging his father’s murder.) The
various genres of literary art — for example, dramatic poems, tales,

odes, lyrics —can to some extent be distinguished according to what

types of discourse — for example, dialogues, anecdotes of past
events, public speeches, and private declarations —they character-
istically represent. Thus, poems typically represent personal
lyric

utterances, or, to use Goodman’s picturesquely unidiomatic terms,

such poems are pictures of utterances. What is significant for the
present point is that certain types of discourse are themselves
characteristically textual inscriptions — for example, chronicles,
journals, letters, biographies, and memoirs — and certain genres of
literature, roughly what we refer to as “prose fiction,’’ character-
istically represent such varieties of inscribed discourse. A literary
work in such a genre would, then, indeed be constituted by the
instances of its own text. (We might say that, if poems are pictures
of utterances, novels are pictures of inscriptions or books; and we
note that novels, a distinctly post-Gutenberg genre, have typically
been representations of chronicles, journals, letters, biographies,
and memoirs.)

8
 LITERATURE AS PERFORMANCE, FICTION, AND ART

One can readily understand, then, what would lead Goodman to

identify at least some literary works with their own inscriptions.
His error, however, was to generalize the identification indiscrimi-
nately, neglecting to distinguish properly among the following three
classes: (1) those instances of “literature,” including all ordinary

texts or inscriptions, which are not literary artworks at all,^ (2) those
literary artworks which, as representations of inscribed discourse,
are indeed constituted by their own texts, and finally (3) those
literary artworks which, as representations of various kinds of
spoken discourse, bear the special relation to their own texts^ that I

have just described here, that is, the relation between a performable
work and its score.

Goodman obser\'es that literary texts may be conceived of in two

ways: first, as phonetic characters in a notational system having
utterances as compliants; second, as characters in a discursive lan-
guage having objects or events as compliants. As to the first part
of this formulation, I have been suggesting that the texts of literary
artworks have certain features that distinguish them —and their
“utterances” —from the inscriptions of ordinary discourse. It is the
second part that concerns me now, for it slights the more funda-
mental distinction between the two senses of literature that I alluded
to at the beginning. The second part of the formulation is inaccurate
because, to the extent that a literary artwork can be conceived as a
character in a discursive language, it has no objects or events as
compliants.
I shall elaborate this in a moment, but first I should say
parenthetically that I think the concept of compliance is severely
strained when Goodman uses it to describe the relation between
natural verbal discourse and the world of objects and events. The
relation between a given phonetic character and the sound that
complies with it is not comparable to the relation between a given
utterance and anything at The world of objects and events does
all.

not comply with our utterances; it contains them and it causes them.**
Granting Goodman his terms, however, and also granting that the
text of a literary artwork becomes, through a performance, a
character in a natural language, it remains true that that character

9
 TOWARD A THEORY OF FICTIVE DISCOURSE

has no compliance-class because the literary artwork is fictive.

Fictiveness, moreover, is precisely what distinguishes the literary

artwork from the more general class of verbal utterances and

inscriptions. Poems and novels, as opposed to biographies and
histories of the Civil War, are linguistic structures whose relation to

the world of objects and events is short-circuited. The short circuit

operates through a convention according to which certain identi-
fiable utterances are understood to be performances of a verbal
action, the occurrence of which as an “action” is entirely confined to
such performances. The child who asks, “Did it really happen or is

it a story?” has already learned the convention, and soon will not
have to ask because he will also have learned the cues that identify
fictiveness in a narration.

By this convention, Keats’s ode “To Autumn,” in which the

speaker is left unspecified, is precisely as fictive as Tennyson’s
poem, “Ulysses,” in which the speaker is represented as the wily
Greek himself. Both poems are understood not as the inscribed
records of utterances actually uttered bymen who spoke poetically,
but rather as linguistic structures composed by men whom we call
poets because they compose such structures. No matter how closely
the statements in such a composition resemble statements the poet
himself as a historical creature might have truly and truthfully
uttered, they remain fictive statements in the poem. As an utterance,
the poem is unmoored from any specific context or occasion in the
world of objects and events, and thus, inGoodman’s terms, it refers
to and denotes nothing, fo object that the poem seems to imply and
evoke such a world is, of course, to speak of the characteristic
marvel and object of poetic composition, but it is not an argument.
To evoke or suggest is not to denote; it is to seem to denote, and that
is what is meant by fictiveness and what has traditionally been
meant by poetry. Ben Jonson’s words: “Hence hee is call’d a Poet,'
In
not hee which writeth in measure only; but that fayneth and formeth
a fable, and writes things like the Truth. For, the Fable and Fiction
is (as it were) the forme and Soule of any Poeticall worke, or Poeme”
{Timber: or. Discoveries, 1641).
Novels and tales are obviously also fictive, but more radically so
than is sometimes supposed. For not only are the characters and the

10
 LITERATURE AS PERFORMANCE, FICTION, AND ART

events narrated in a novel fictional, but so also is the narrator whose

voice relates the events, and, most significantly, so also are the

statements through which the narration is presented. A history of the

Civil War may be conceived, in some very peculiar sense, as having
the Civil War as its compliance-class, but in that sense no war at all

is the compliance-class of the Iliad or, indeed, of War and Peace. A

reader who fails to comprehend the nature of fictions may be as

likely to look for Prince Andrey’s grave as for Napoleon’s Tomb, but
the fact that we can locate the latter and not the former does not
make one part of Tolstoy’s novel any less fictional than another. For
the essential fictiveness of literary artworks is not to be discovered in
the unreality of the characters, objects, and events alluded to, but in
the unreality of the alludings themselves.

III. Art and Cognitive Activity

In the final sections of Languages of Art (pp. 241-65), Goodman
enlists his theory of symbols in the service of a noble cause, namely
the destruction of what he refers to as “the deeply entrenched
dichotomy of the cognitive and the emotive.’’ I am in sympathy with
the cause, and it seems to me that Goodman’s contributions to it

here are of great value. Nevertheless, in demolishing the falsely

conceived dichotomy, he has unwittingly — and, for his purposes,

unnecessarily —obscured a quite validly conceived distinction,

namely that between nature and art.

Those who seek to demonstrate the affinity of science and art

typically focus on the similarities between the scientist and the artist,

observing, for example, that both are engaged in creating or

revealing order in nature or experience. Goodman takes a more
unusual and interesting line, comparing the activity of the scientist

to that of the audience or spectator of a work of art, and empha-

sizing the degree to which the characteristic experience of each is

similar.Thus he speaks of the searching, testing, making of delicate

discriminations, and discerning of subtle relationships that are
involved in one’s experience of art; and certainly the activity of the
scientist is readily described in such terms. What concerns me,
however, is that, whereas the scientist’s activities are directed toward
the exploration of the natural universe, the “reader’’ of the painting

11
 TOWARD A THEORY OF FICTIVE DISCOURSE

or poem or musical composition is engaged in the exploration of a

structure created by a fellow creature and designed by that artist

precisely to encourage and reward such exploration. There is, I

believe, no reason to suppose that the structure of nature was

designed to engage our interest, to entice us into investigations, and
to reward our activities with discovery — although we, of course, have
probably been designed by natural selection to find that sort of
activity gratifying in any case.
My point here is that we— that is, human beings, but perhaps not
uniquely among organisms —do indeed take pleasure in what Good-
man calls cognitive activity, and that this pleasure or satisfaction
may be diminished or enhanced by various circumstances, and that
the artist is one who is skilled at fashioning such enhancing
circumstances, and that we call circumstances so fashioned works of
The
art. artist (a gentleman who has perhaps received an inordinate
amount of attention in Romantic aesthetics but whom Goodman
nevertheless snubs almost, one would think, pointedly)^ —the artist

creates structures that are relatively free of those irrelevancies,

irregularities, and monotonies that often frustrate and attenuate our
satisfaction in cognitive activity, but structures that are, at the same
time, sufficiently complex, rich, or subtle to engage and exercise our
cognitive faculties.
The work of art, then, provides the stimulus to and occasion for a
highly distilled and rewarding cognitive experience, but also a highly
artificial one; for in experiencing a work of art, we may have all the
satisfaction that attends the pursuit and acquisition of knowledge
without necessarily having acquired any knowledge at all. Or,
rather, what we acquire knowledge q/’in a work of art is primarily
the work itself, a fabricated microcosm designed to be knowable in

just that way. What is most important here is that although this
knowledge obviously has its sources in nature and reflects it in

various ways, it cannot be generalized to nature as we generalize

knowledge in science — or indeed in our ordinary experiences —from
one part of nature to another.
We surely must grant, as Goodman insists, that aesthetic experi-

ence is not altogether distinctive. It shares motives and qualities

with the experience of scientists doing science and, of course, of

12
 LITERATURE AS PERFORMANCE, FICTION, AND ART

philosophers doing philosophy. We should acknowledge that it also

shares motives and qualities — perhaps some of the same ones—with
scientists, philosophers, and other people making love and playing
games. We remain, in all these activities, what we always are:
creatures moving about in a world we must seek to know in order to
survive and for whom, as Aristotle noted, learning is the liveliest
pleasure. It is also true, however, that the sources or objects of our
aesthetic experiences may be artificial worlds, fictive “natures,” and
that the consequences of knowing them are confused at one’s peril
with the consequences of knowing nature proper. Of course, we are
usually saved from such peril by the conventions (such as stages,
frames, pedestals, and versification) that operate to inform us that
we are confronted by art, not nature, so that we are never really

deluded by the fictive construction and thus receive no false

knowledge from it.

The connection between these last remarks and the earlier parts of
my paper are, I trust, clear: the distinction that Goodman blurs in
identifying literary art with written inscriptions is, in effect, the
distinction between fiction and history, and between art and nature.
The verbal actions —speech or writing—of other men are a part of
nature, as much a part of the natural environment as the behavior of
sun, stars, rocks, and trees; and those verbal actions are as much a
part of history as all the other actions of our fellow creatures.
Furthermore, our lives do depend on our searching, testing, making
delicate discriminations, and discerning subtle relationships in the

utterances of these fellow creatures. The cognitive explorations thus

directed, however, are not always rewarded with discoveries and
revelations; for these utterances are often obscure and fragmentary,
or predictable and monotonous. If we wish to assure ourselves of a
more satisfying cognitive experience with language, we will turn to
—
poems and novels knowing, of course, that the speech of men in

nature and history is distinct from the language of art.

13
 POETRY AS FICTION

Paradoxes make intriguing titles, but I am not otherwise fond of

them and intend, by the end of this article, to dissolve the one that
entitles it. I mean to do this by elaborating the proposition that
fictiveness is the characteristic quality of what we call “poetry”
when we use the term in the broad sense bequeathed by Aristotle,
that is, to refer to the general class of verbal artworks. My primary
concern will be to develop a conception of poetry that allows us to
distinguish it from and relate it to both nonpoetic discourse and
other artforms. The view presented here was initially, but rather
incidentally, proposed elsewhere.* I have found the elaboration of it

of continuing interest, however, especially since the grounds for

those distinctions and the nature of those relationships remain, to
my mind, extremely problematic in contemporary linguistic and
aesthetic theory.
Before saying anything at all about poetry, I shall, in what
follows, have a few things to say about language generally. Any
theory of poetry inevitably, though not always explicitly, presup-
poses a theory of language. Thus, those who have at various times
regarded poetry as inspired speech, or embellished prose, or the
language of passion, or “emotive” statements, have obviously had
somewhat different notions of what language is when it is not
poetry — for example, uninspired speech, plain prose, the language

14
 POETRY AS FICTION

of reason, or “verifiable statements.” Since, moreover, linguistic

theory is now in a very volatile state, no general propositions
concerning language can be offered casually or taken for granted.
In any case, although I am by no means offering here anything that
could be called a theory of language, the first section of this article
will develop some general observations on nonpoetic or what I call

“natural” discourse, particularly in those respects that are most

significant in distinguishing it from poetry. The second section of
the article will develop some implications of the conception of poetry
as mimetic, or what I shall be calling //cr/ve, discourse.
Although the making of distinctions, definitions, and classifica-

tions will occupy a good deal of the discussion throughout, it should

become clear that my ultimate interest is not in taxonomy but in
poetry as an artform. I am concerned with how, on what basis, we
actually do identify a verbal structure as poetry, and how that
identification directs and modifies our experience and interpreta-
tion of that structure both as distinct from a natural utterance and
as related to other artforms.^

I. Natural Discourse
By “natural discourse” I mean here all utterances — trivial or sub-
lime, ill-wrought or eloquent, true or false, scientific or passion-
ate —that can be taken as someone’s saying something, somewhere,
sometime, that is, as the verbal acts of real persons on particular
occasions in response to particular sets of circumstances. In stress-
ing all these particularities, I wish to emphasize that a natural
utterance is a historical event: like any other event, it occupies a
specific and unique point in time and space. A natural utterance is

thus an event in the same sense as the coronation of Elizabeth I on

15 January 1559, or the departure this morning from Albany of
Allegheny Airlines flight 617, or the falling of a certain leaf from a
certain elm tree. Other events more or less resembling these in

various respects may occur at other times or in other places, but the
event itself —that coronation, that flight, that utterance —cannot
recur, for it is historically unique.

The point requires emphasis because it reflects a fundamental

distinction that may be drawn between natural utterances and cer-

15
 TOWARD A THEORY OF FICTIVE DISCOURSE

tain other linguistic structures which are not historical events and
which can be both defined and described independently of any
particular instance of occurrence. Dictionary entries, for example,
or what we refer to abstractly as “the word /ire” or “the phrase law
and order'" are not themselves particular events; they are, rather,
linguistic /’orms, or the names of certain types or classes of events.
And, as such, certain observations may be made about them: for
example, the morphemic or phonetic features that define all

members of the class, or the syntactic rules governing their accepted

use in English sentences, or, of course, the characteristic features of
the circumstances in which they do occur as part of utterances — in

other words, their “dictionary meanings.” But these linguistic

forms —words, phrases, and so on — are not themselves historical

events unless or until they occur as the verbal responses of particular
persons on particular occasions. Obviously “the word fire" as a
general class is a very different sort of thing from a specific
utterance, “Fire!”, which may warn a man that his life is in danger
or send a bullet speeding toward him, very much depending on the
particular circumstances in which the utterance occurs and to which
it is a response.
A natural utterance not only occurs in a particular set of
circumstances —what is often referred to as its
4
context — but is also
understood as being a response to those circumstances. In other
words, the historical “context” of an utterance does not merely
surround it but occasions it, brings it into existence. The context of
an utterance, then, is best thought of not simply as its gross external
or physical setting, but rather as the total set of conditions that has
in fact determined its occurrence and form.^ That total set of con-
ditions, what makes us say something at a particular time and also
shapes the linguistic structure of our utterance —the specific words
we choose, our syntax, our intonation, and so on —
is likely to be

manifold and complex no matter how simple the utterance. More-

over, the total set of conditions that determines what we say and how
we speak is by no means confined to the objects and events “spoken
about,” or what linguistic theorists of various persuasions refer to as
“referents,” “designations,” “denotations,” or “significations.”
It is worth noting that the existence of an object or event or even.

16
 POETRY AS FICTION

as we say, an “idea,” is never a sufficient reason for responding to it

verbally. In other words, the fact that something is true is never a

sufficient reason for saying it. If I should be heard to say, “It’s five
o’clock,” the reasons for my saying so would clearly include more
than what time of day it was just then, for at any moment it is a
certain time, but I do not announce the time continuously through
the day. Perhaps, on this occasion, I wished to remind someone of
an appointment, or perhaps someone had just asked me‘ for the
correct time. Certainly these circumstances were as significant in
occasioning my utterance as that specific one to which my words,
“It’s five o’clock,” might seem exclusively to “refer,” namely the
time of day.
Given any utterable fact or state of affairs, gross or subtle,
physical or psychological —the state of the weather, the color of
swans, or my opinion of the government —whether or not I will

actually utter it, and how I will utter it, will always depend on other
variables, that is, attendant circumstances other than that fact or
state of affairs. These variables will include, among other things,
the presence of a potential listener, my relationship to him, the
nature of the social occasion, the immediate verbal context (what
either he or I have been saying), and, perhaps most significantly, the
conventions of the linguistic community to which we both belong.
There is no reason to maintain a sharp distinction between the
sort of physical and social variables just mentioned and what might
otherwise be thought of as the internal, personal, mental, or
psychological springs of speech. It is obvious that among the
circumstances that provoke, occasion, and shape an utterance are
conditions peculiar to the speaker’s current state: his emotions, his
feelings, his memories, expectations, beliefs, and desires. I may say
“It’s five o’clock” partly because I am hungry or anxious or bored,
and such conditions must also be recognized as part of the context
of the utterance. We should note, moreover, that the speaker’s
“current state” is inevitably the product of his past as well as his
current experiences, including, most significantly, his past verbal
experiences, and that part of his psychological or mental condi-
tion —and therefore part of the context of his utterance — is how he
has learned to use language.

17
 TOWARD A THEORY OF FICTIVE DISCOURSE

Although we may, for certain purposes, describe an utterance

exclusively in terms of its linguistic form (for example, as a certain
concatenation of lexemes and/or phonemes), a natural utterance
can never be adequately specified or described as an event except in

relation to the context in which it occurred. In other words, a verbal

event, like any other event, is individuated as much by its context as
by its form. Thus, although we could say that two men each pulling
the trigger of a gun are engaged same /orm, it is clear
in acts of the

that Mr. X shooting Mr. Y is not the same event as Mr. A shooting
Mr. B, or as Mr. X shooting Mr. Y again fifteen minutes later.

Similarly, when I say, making introductions at a party, “This is my

daughter,” it may not be a unique event with respect to its linguistic

form, but it is certainly not the some other woman

same event as
saying it of her daughter or, indeed, as my own saying it on some
other occasion, either fifteen minutes later to some other guest or
even absentmindedly to the same one as before.
Moreover, it is unlikely that any two natural utterances would be
even formally identical if one extends attention to the more subtle
aspects of their linguistic form. For although each utterance could
be transcribed with the same symbols, such a transcription preserves
only a fraction of the total physical reality constituting each
utterance, a reality that would include not only a certain sequence of
phonemes, but also intonational features such as pitch contours,
stress, pacing, and usually facial expressions and other gestures as
well. While some linguists may regard the latter aspects of the
utterance with suspicion and dispute their status as linguistic
features, it is nevertheless becoming increasingly evident that there
is no absolute discontinuity between the part of an act or event that
is called “verbal” and the totality of that act or event. In other
words, a natural utterance is always continuous with the speaker’s
total ongoing behavior and also continuous with the total world of
natural events. The professional linguist’s or our own ordinary
description of the utterance reflects an arbitrary demarcation and
abstraction from the fullness, the density, and the spatial, temporal,
and causal continuity of all human action and all events in nature.
Most of us would agree that it is impossible to provide a complete
and exhaustive description of a nonverbal historical event such as

18
 POETRY AS FICTION

the coronation of Elizabeth or the departure of flight 617. What the

historian offers will usually be a selection or abstraction of certain
features of these events at a level thought adequate for the purpose
at hand. It is clear, moreover, that neither an eyewitness report nor,
if we had it, even a videotape, would constitute a total record of the
event; and neither one, of course, would constitute the event itself.

The same limits and distinctions apply to the descriptions and

records of verbal events: Elizabeth’s first speech to Parliament on 4
February 1559, or my farewells this morning to my family. No
description or record would be complete, neither a vocal quotation
nor a tape recording, in either of which many features of the original
event would be lost. The fact, however, that verbal events can be
transcribed in a standard notational system often seems to obscure
for us their similarity to other events. It is true that orthography and
phonetic notation allows us to record or describe natural utterances
with considerable subtlety and specificity of detail through conven-
tionalized symbols. Moreover, a transcription of this kind —that is,

a “text” of the utterance —may be an adequate description or record

of it for most purposes. Nevertheless, we should not confuse a copy
of that text with the verbal event itself, the historical act of a
particular speaker on a particular occasion.
The relation of utterances to texts is of special interest to us here
since, at least in our own culture, we typically encounter poetry as
texts. The relation is extremely complex, however, with respect to
both natural and poetic discourse, and, indeed, it is not always the
same relation. I have just been speaking of texts that serve as
records or descriptions of natural utterances, that is, inscriptions of
verbal events that occurred at some specific time, such as Eliza-
beth’s first address to Parliament. Not all texts bear this relation to

some natural utterance. Many texts personal letters, for ex- —

—
ample are not records or descriptions of utterances, but constitute
utterances themselves, only in written rather than vocal form. It is

true, of course, that there are other very significant aspects to the

relation between writing and vocal speech, and they are not mutually
independent or simply parallel possibilities. Nevertheless, to the
extent that the writer’s act of composing and inscribing is a
historically specific and unique verbal event, it is analogous to the

19
 TOWARD A THEORY OF FICTIVE DISCOURSE

speaker’s act of emitting the sounds that comprise spoken discourse.

And thus we may regard the product of either act as a natural
utterance.
In view of the Gutenberg revolution, the question may arise as to

whether printed (or otherwise duplicated) texts can also be regarded

as natural utterances, and the answer here is sometimes yes and
sometimes no. A printed text may be simply one of many copies of
an inscribed record of a vocal utterance that, like Elizabeth’s
Address, did occur at some specific time and place. In this case, the
text is not a natural utterance, but the transcription of one. But a
printed work may also be a natural utterance itself in written form,
exactly like a personal letter —though the letter, of course, usually
exists as only a single text. It may be initially difficult to conceive of a
printed work as a natural utterance and thus, by our definition here,
a historically unique event. We should recognize, however, that no
matter how many duplications of a text are subsequently produced,
the writer’s actual composition of the linguistic structure that
constitutes that text was and remains a historically unique event.
(“Unique” here does not mean unitary, and it is understood that the
composition of the text will often consist of numerous “acts”
dispersed in time, from the initial jottings to the ultimate revisions.)
To summarize these points, then: whether or not a composition
was written to be printed, and no matter how long it is, or how long
it took to write, and no matter how remote in time or space the
writer from his ultimate audience, or how eloquent its style, or how
culturally significant and otherwise estimable it is, the composition
must still be regarded as a natural utterance so long as it may be
taken as the verbal responses of a historically real person, occa-
sioned and determined by a historically real universe. And this
means that most of what we call “literature” in the general sense of
inscribed compositions does in fact consist of natural utterances.
This would include works ranging from Aristotle’s Metaphysics and
Macaulay’s History of England to an article in a scientific journal or

an editorial in this morning’s New York Times. These are all as

much natural utterances as the remarks exchanged between me and
a colleague a few moments ago.
There remains, however, one other class of texts that are neither

20
 POETRY AS FICTION

natural utterances in written form nor the transcription of natural

utterances that originally occurred in vocal form, and this class
consists of the texts of fictive utterances, including most promi-
nently those compositions that we otherwise refer to as works of
imaginative literature —poems, tales, dramas, and novels. I shall

reserve comment on these texts until later, in connection with the

general discussion of fictive discourse; for, as we shall see, fictive
utterances bear an altogether distinctive relation to their own texts
when indeed (as is not always the case) such texts exist.
But we may return now from the texts to the contexts of natural
utterances, and thereby to the crucial question of meaning and
interpretation. A natural utterance cannot be exclusively identified
or described independent of its meaning be
context, nor can its

understood independent of that context. Indeed, what we often

mean by the “meaning” of an utterance is its context, that is, the set
of conditions that occasioned its occurrence and determined its form.
The view of meaning proposed here is not offered as an analysis of
all the numerous senses in which the term has been or could be used,
and certainly not as a solution to the ever-proliferating number of
problems associated with it in contemporary linguistics and philos-
ophy. Nevertheless, a causal conception of meaning —which this

is —has much to recommend it, particularly here, since it permits us

to appreciate better the distinctive nature of poetic discourse and of
its “interpretation.” Moreover, it is not so idiosyncratic as may first

appear, for “meaning” in the sense of causes or determinants will

often be found to accommodate or correspond to familiar usage of
the term.
I must emphasize that I am speaking here of the meaning not
of words but of utterances, a distinction not always grasped even by
those most concerned with these problems. One may ascertain the
meanings of those abstract classes called words by determining the
conventions governing their usage in the relevant linguistic com-
munity, usually by consulting one’s experience of the language or,

when difficulties arise, either a dictionary —or an analytic philos-

opher. Dictionaries and philosophers are of only limited help,
however, in ascertaining the meaning of particular verbal events.
When we speak ordinarily of the meaning of a particular utterance

21
 TOWARD A THEORY OF FICTIVE DISCOURSE

that is, what someone has said —we are usually concerned not with
the definitions of the words that compose it or even, in a restricted
sense, with what it “refers” to, but rather with why it occurred: the
situation and motives that produced it, the set of conditions,
“external” and “internal,” physical and psychological, that caused
the speaker to utter that statement at that time in that form — in

other words, what we are calling here its context.

For example, definitions and referents are not what interest John
when he asks, “What do you mean?”, in response to his friend’s
remark, “You know, I think Bill is a fool.” Pointing to Bill and
offering an analysis of the “concept” of folly will probably not
answer his question. Knowing more likely to
this, his friend is

describe certain circumstances, observations he has made, impres-

sions he has had (and perhaps also his motives for articulating them
at that moment), and so forth, until John says, “Oh, well, now I

understand what you mean,” meaning that he has located to hisown

satisfaction the reasons for or causes of his friend’s remark. The
qualification here, “to his own satisfaction,” is an important one,
for it is most unlikely that John would in fact have identified all the
determinants involved.
We rarely “understand completely” one another’s utterances, nor
do we need or seek to do so. Criteria for the adequate understanding
of an utterance vary widely, depending on the nature of the
utterance and the primary purposes and interests of the speaker and
listener. —
And although sometimes for example, in a psycho-
analyst’s office —
one may probe for increasingly subtle and obscure
determinants, both speaker and listener are usually satisfied with
considerably less than a total identification of all of them. It is

usually not necessary, and of course it is usually not possible, for the
make up the context of an
listener to ascertain all the conditions that

utterance. It is not necessary because many of them will be trivial

and irrelevant to his concerns. And it is not possible either because
the speaker’s original context is remote in time or space, or because
many of the springs of speech are not apparent from the immediate
context or, as we say, are private or internal to the speaker. The
listener or audience, therefore, is always obliged to “interpret” what
is said or written. That is, to the extent that the listener has an

22
 POETRY AS FICTION

interest in those unavailable determinants, he must hypothesize,

imagine, or infer them.
When we read the inscribed utterance of a friend, such as a letter
from him, we may be more aware of interpreting as such than when
we listen to him speak, but we do so in both instances and by the
same process: partly through inferences based on what we know of
him personally, but mostly through inferences based on all our own
prior experiences, especially our prior experiences with language.
And, when he alludes either in speech or writing to matters of which
—
we have no specific knowledge for example, a third person whom
we have never met, a place we have never visited we supply our —
ignorance by an imaginative projection of what we do know
generally. It is important to emphasize, however, that these projec-
tions are attempts to infer or approximate actual circumstances,
and thus are subject to correction should our knowledge become
more specific. (“Oh, you're Charlie’s brother. From what he said, I

pictured you as much older.’’)

What makes a letter particularly interesting as an utterance is

the fact that, since it lacks the supplementary information usually

conveyed to the listener by intonation and gestures as well as by
shared physical contexts (we cannot point to things in letters), this

sort of information will commonly be supplied by the writer in other

ways: by explicit allusions (“As I write this, I am sitting by my study
window —you know, the one that looks out over the back garden,’’
and so on), by graphic substitutes for intonation (for example,
underlining, punctuation, spacing), and by more subtle modifica-
tions of the language itself (for example, in diction, syntax, turns of
phrasing, and metaphor). Our syntax in letters, because it carries a

greater burden of information than in conversational speech, not

only can be but must be more controlled. To be sure, since we are
often more or less conscious of the generic relation of our letters to
“literature,’’ we will employ forms such as archaisms and meta-
phoric imagery that would seem pretentious or otherwise inappro-
priate in conversational speech. This, however, does not altogether
account for the fact that some of us become, in our letters, rather
uncharacteristically eloquent and “literary’’; for, as we shall see,

there are other reasons why the linguistic features of letters often

23
 TOWARD A THEORY OF FICTIVE DISCOURSE

bear an interesting resemblance to those commonly associated with

poetic discourse.

II. Fictive Discourse

Poems are not natural utterances, not historically unique verbal acts
or events; indeed a poem is not an event at all, and cannot be said
ever to have “occurred” in the usual sense. When we read the text of
a poem or hear it read aloud, our response to it as a linguistic
structure is governed by quite special conventions, and it is the
understanding that these conventions are operating that distin-
guishes the poem as a verbal artwork from natural discourse. The
operation of these conventions is most readily apparent in dramatic
poetry, that is, plays, where it is understood that the acts and events
performed on the stage are not happening but are being represented
as happening. When we see a production of Hamlet, we do not
watch a queen drinking poison, but the enactment of such an event,
which may be said to “occur” only in being thus enacted. But among
the acts and events represented on the stage are also verbal ones. As
the actor who portrays Claudius leans forward and extends his arm
in a gesture of horror and abortive warning, thus representing a man
leaning forward and extending his arm, and so on, that actor also
utters the words, “Gertrude, do not drink,” thus representing a man
uttering those words. We are not aware here of any radical
discontinuity between the enactment of a physical action and the
enactment of an utterance — and of course an utterance is a physical
action, though it has other characteristics that sometimes obscure
that fact.
Most of us would be quite willing to grant the existence of what
could be called mimetic or f ictive discourse — that is, the representa-
tion of speech — at least in dramatic poetry. What I would like to

suggest, however, is that all poetry may be so regarded, that we

could conceive of as fictive discourse not only the representation of
speech in drama, but also lyrics, epics, tales, and novels. The
conception of poetry as mimetic is, of course, quite ancient, and
modern theorists do continue to assert that literature is a representa-
tional art. It is by no means clear, however, what or what kind of
thing it is that the poem “imitates” or represents. One common

24
 POETRY AS FICTION

notion seems to be that poetry, apparently on the analogy of

painting, somehow represents “images in words.’’ Or, in view of the
existence of numerous imageless poems and passages in novels, that

it represents ideas or feelings, either the author’s or those of his

characters. Or, in view of how restrictive even this formulation is, it

is sometimes suggested that literary works, especially narrative

fictions, represent imagined events or even worlds —m, it will

solemnly be added, the medium of language. I will not attempt here

to indicate all the problems entailed by such suggestions,'’ for I wish
only to point out that they all ignore what might be thought most
apparent, namely that what poems do represent “in the medium of
language’’ is language, or more accurately, speech, human utter-
ance, discourse. The conception of poetry as fictive discourse
proposed here attempts to close in on poetry from two directions:
one, as may be distinguished from other mimetic artforms, and
it

two, as it may be distinguished from other verbal compositions. As a

mimetic artform, what a poem distinctively and characteristically
represents is not images, ideas, feelings, characters, scenes, or
worlds, but discourse. Poetry does, like drama, represent actions
and events, but exclusively verbal ones. And, as a verbal compo-
sition, a poem is characteristically taken to be not a natural
utterance, but the representation of one.
A poem represents discourse in the same sense as a play, in its

totality, represents human actions and events, or a painting repre-

sents visual objects. When we speak of the objects represented in or

by a painting, it is understood that they need not correspond to any

particular objects, but rather to an identifiable class of them. A
painting can depict a landscape that exists as a visual object only in
the depiction itself. Thus, when we speak of mimesis or representa-
tion in an artwork, we recognize that it does not constitute the
imitation or reproduction of existing objects or events, but rather
the fabrication of fictive objects and events of which there are
existing or possible instances or types —whether they be rural
landscapes, star-crossed lovers, or laments for dead friends. In
other words, to say that an artist has represented a certain object or
event is to say that he has constructed a fictive member of an
identifiable class of natural (“real”) objects or events.

25
 TOWARD A THEORY OF FICTIVE DISCOURSE

Part of what has obscured the relation of poetic mimesis to

pictorial and other kinds of artistic representation are traditional
notions that identify the various artforms in terms of their charac-
teristic media. Thus, sound is said to be the medium of music,
pigment the medium of painting, and of course words or language
the medium of poetry. The corollary formula —X (artwork) repre-
sents Y (object of imitation) in Z (medium) — has created more
problems than it has illuminated, most conspicuously, perhaps, in
regard to music, where art theorists, under the presumed obligation
to locate the object that music imitates, have come up with an
amazing assortment of chimeras, from shapes of feeling to states of
being. It is another problem, however, that concerns us here. The
plastic materials that are presumably the media of the visual
arts — pigment, stone, metal, and so forth—do not have an expres-
sive function independent of the artworks into which they are
fashioned. These materials, moreover, do not in themselves re-
semble the objects and scenes that they represent. A block of marble
is a very different thing from a human figure. The corresponding
medium of poetry, however, language, is not a “raw” material, but
itself a symbolic system with expressive functions independent of its

use in artworks. For this reason, it has been difficult to conceive of

language as both the medium of an artwork and also what is

represented by it.

The difficulty here, however, is really the traditional concept of

the art medium itself, particularly its implicit dualism of form and
matter. This dualism —the notion of the art medium as formless
matter — not only creates problems with regard to poetry (for
language is obviously not formless matter), but it also obscures the
nature of other artforms. We could just as readily and, I think, more
fruitfully, think of the medium of the visual arts not as pigment and
stone but as the visually perceived properties of matter or, indeed, as
the elements and dynamics of visual perception itself. And,we if

must have a corresponding “medium” for poetry, we would do

better to locate it not simply in words or language conceived
abstractly, but in the whole dynamic complex of verbal behavior and
verbal experience.
But if we are content to do without the traditional notion of the

26
 POETRY AS FICTION

art medium altogether, we may be better able to appreciate the

essential nature of poetic representation and its relation to artistic
mimesis generally. As I suggested above, we may conceive of an
artwork not as the imitation, in some different “matter,” of the
“form” of particular objects or events already existing in nature, but
as the creation of a fictive member of a certain class of natural
objects or events. Thus, paintings are fictive instances of what, in
nature, are visually perceived objects. Musical compositions are
fictive instances of acoustically perceived events, in other words
designed sounds as distinguished from sounds simply occurring in
nature. And poems are fictive utterances. The kinds of natural
events represented in poetry are, of course, quite special: utterances
are themselves human constructions, and in that sense “artificial.”
This should not, however, obscure the sense in which utterances are
nevertheless natural events, like the flight of birds, the falling of
leaves, and all the particular actions of individual men moving
about in, and being moved about by, the natural universe.
We can, I think, readily conceive of a-man-walking as a natural
event and should be able to conceive of a-man-talking as such; for,
as I have already suggested, there is no real discontinuity between
verbal and nonverbal events. A painting can represent, through a
visual configuration of line and color, a man walking or a child
sleeping, because such events are ordinarily perceived primarily as
visual events. And although a visual artist can also represent a man
talking (one may think, for example, of some of Daumier’s prints of
lawyers in animated conversation), he cannot represent pictorially
the utterance itself, for speech is not perceived as a visual event
except of course, when it is in written form, a matter to which I will
return later. But for now let us pursue the example of Daumier a bit
further. As a visual artist, he was of course extraordinarily sensitive

to the expressive and otherwise interesting qualities of the appear-

ances of his fellow creatures: the way they stood and grouped
themselves together, the “expressions” on their faces, the gestures
of their hands, and so forth. Had he also been, as some people are,
extraordinarily sensitive to the expressive and otherwise interesting
qualities of the speech of his fellow creatures, he might have sought
to represent that too. But how could he do so? The answer I am

27
 TOWARD A THEORY OF FICTIVE DISCOURSE

suggesting here is that he could fashion a fictive representation of

speech, that is, a —
poem something, perhaps, like Browning’s “The
Bishop Orders His Tomb,” which I think we might recognize as a
verbal counterpart of a satiric Daumier print: ut pictura poesis.
The relation of “dramatic monologues” to dramatic poetry proper
is, of course, readily appreciated, and we can see how either could

be regarded as mimetic discourse. My claim here, however, is more

general, for what is central to the concept of the poem as a fictive
utterance is not that the “character” or “persona” is distinct from
the poet, or that the audience purportedly addressed, the emo-
tions expressed, and the events alluded to are fictional, but that
the speaking, addressing, expressing, and alluding are themselves
fictive verbal acts. To be sure, a fictive utterance will often re-
semble a possible natural utterance very closely, for the distinc-

tion is not primarily one of linguistic form. Moreover, although

certain formal features — verse, most notably —often do mark and
indeed identify for the reader the fictiveness of an utterance, the
presence of such features are not themselves the crux of the dis-
tinction. The distinction lies, rather, in a set of conventions shared
by poet and reader, according to which certain identifiable lin-

guistic structures are taken to be not the verbal acts they resemble,
but representations of such acts. By this convention, Keafs ode
“To Autumn” and Shakespeare’s sonnets are precisely as fictive
as “The Bishop Orders His Tomb” or Tennyson’s “Ulysses.” The
statements in a poem may, of course, resemble quite closely state-
ments that the poet might have truly and truthfully uttered as a
historical creature in the historical world. Nevertheless, insofar as

they are offered and recognized as statements in a poem, they are

fictive. To the objection, “But know Wordsworth meant what he
I

says in that poem,” we must reply, “You mean he would have meant
them if he had said them, but he is not saying them.” As I shall
explain later, we may choose to regard the composition not as a
poem but as a historical utterance, but then the conventions by
virtue of which its fictiveness is understood and has its appropriate
effects are no longer in operation.

Another matter should, however, be clarified at this point. I have

28
 POETRY AS FICTION

said that novels and tales, as well as lyrics, epics, and dramatic
poems are also fictive representations of discourse. The fictiveness
of prose fiction is, of course, commonly acknowledged, but it is

more radical than is sometimes supposed. For not only are the
characters and events narrated in a novel fictional, and not only
is the narrator whose voice relates the events fictional, but most
significantly, so also is the entire structure of discourse through
which the narration is presented. Indeed, as we all know, many
novels such as War and Peace allude to quite real persons and
events, a consideration that has created theoretical problems for
many literary theorists. The essential fictiveness of novels, however,
is not to be discovered in the unreality of the characters, objects, and
events alluded to, but in the unreality of the alludings themselves. In
other words, in a novel or tale, it is the act of reporting events, the
act of describing persons and referring to places, that is fictive. The
novel represents the verbal action of a man reporting, describing,
and referring.
Consider the following two passages:

He was a gentleman of good family in Buckinghamshire,

and born to a fair fortune, and of a most civil and affable de-
portment. In his entrance into the world, he indulged
himself all the license in sports and exercises and company
which was used by men of the most jolly conversation;
afterwards he retired to a more reserved and melancholy
society.

He had been a member of the Court of Justice, and died at

the age of forty-five. His father had been an official who,
after serving in various ministries and departments in
Petersburg, had made the sort of career which brings men
to positions from which by reason of long service they
cannot be dismissed.

The from the description of John Hampden in Clarendon’s

first is

History of the Rebellion', the second is from Tolstoy’s Death of Ivan

Ilyich. (In both, we might note, allusions are made to real places,

Buckinghamshire and Petersburg.) I am suggesting here that the

relation between the two passages is that the second is a representa-

29
 TOWARD A THEORY OF FICTIVE DISCOURSE

tion of the kind of thing the first really is, namely a biography. The
Death of Ivan Ilyich is not the biography of a fictional character,
but rather a fictive biography. The fiction attaches no more to the
narrated facts of Ilyitch’s life than to the fact of someone’s
narrating them. Tolstoy is, if you like, pretending to be writing a
biography while actually one.
we consider literature from the point of view I am developing
If

here, it becomes evident that the various genres of literary art for —
example, tales, classical odes, and lyrics can often be distin- —
guished from each other according to what types of natural dis-
course they represent: here, respectively, anecdotal reports of past
events, public speeches, and more or less private or personal
utterance.^ Poetry itself, as distinct now from novels and stories,
traditionally represents various kinds of spoken discourse. Certain
types of discourse, however, are themselves typically textual inscrip-
tions: that is, they exist characteristically in written and often in
printed form — for example, chronicles, journals, letters, memoirs,
and biographies. And certain genres of literary art, roughly what we
refer to as “prose fiction,’’ characteristically represent such varieties
of inscribed discourse. Novels, for example, a distinctively post-
Gutenberg genre, have typically been representations of chronicles,
journals, letters, memoirs, and biographies. This aspect of prose
fiction has some interesting implications for the nature of novels as
texts, but they will be better appreciated after we have given some
attention to literary texts generally.
A fictive utterance consists entirely of a linguistic structure,
unlike a natural utterance, which consists of a linguistic event
occurring in a historical context. In a nonliterate culture, for
example, among Northwest American Indian tribes, the linguistic
structure that would be identified as that song or story is preserved
and duplicated, if at all, only in being remembered and recited. But
in a literate culture, the identity of the poem may be preserved and
reproduced through a standard notational system, that is, in a
written text. The poem, however, bears a quite special
text of a

relation to the utterance of which it is presumably an inscribed

counterpart. For it is neither a transcription of an utterance that
actually occurred at some specific prior time, like Elizabeth’s first

30
 POETRY AS FICTION

Speech to Parliament, nor is it a natural utterance in written form,

like a personal letter. It is, rather, like the see re of a musical
composition or the script of a play, that is, formal specifications for
the physical production of certain events. The text of the poem tells

us, in other words, hew to produce the verbal act it represents. This
is evident enough for a playscript, which directs the performer’s
verbal actions along with other more obviously physical actions: for
example, “enter,” “exit,” “is stabbed,” “falls,” says “I am dead,
Horatio; wretched Queen, adieu.” But this is true of any poetic text,
that is, the text of any verbal artwork that represents spoken rather
than written discourse. The text of a novel must be regarded
somewhat differently, as I will explain below. But, allowing for this
exception, the text of any poem is to be interpreted, in the first
instance, as, in effect, a score or stage directions for the perfor-
mance of a purely verbal act that exists only in being thus
performed. A poem is never spoken, not even by the poet himself. It

is always re-cited; for whatever its relation to words the poet could
have spoken, it has, as a poem, no initial historical occurrence.

What the poet composes as a text is not a verbal act but rather a
linguistic structure that becomes, through being read or recited, the
representation of a verbal act.
As I pointed out above, works of prose fiction are characteristi-
cally representations not of spoken but of inscribed utterances, and
for this reason the texts of novels are, interestingly enough, closer

to pictures than to musical scores. What the text of Richardson’s

Clarissa represents is not the speech of certain characters but a
collection of their letters; what David Copperfield represents is not
the spoken reminiscenses of a man, but his autobiography. Each
novel itself, marks printed on its pages and, if you like,
that is, the
the pages themselves, plus covers and binding, is a depiction of —
fictive instance of —a kind of book. Indeed, in view of its three-

dimensionality, the copy of the novel we hold in our hands could be

conceived of as a sculpture, where the sculptor has not satisified
himself in representing the gross physical and visual qualities of a
book, but has sought to represent the very text of one. But, rather
than complicate matters, we may at least agree that what the text of

a novel represents is, precisely, a text.

31
 TOWARD A THEORY OF FICTIVE DISCOURSE

In what follows, I shall be speaking again of poetry in the

narrower sense, that is, as representations of spoken discourse,
usually in verse. Some of the points I shall be making would require
a somewhat different or additional formulation with respect to
novels or representations of written discourse, but I shall not
develop them here.
Although a poem, unlike a natural utterance, consists entirely of
a linguistic struture, we obviously do not respond poems as pure to
forms or merely as organizations of sound, any more than we
respond to plays as purely formal structures of movement or to
traditional paintings as pure configurations of line and color. For
each of these is understood to be a representational artform, and the
spectator readily infers a meaning or context though a fictional —
—
one for the objects, actions, and events represented. The curtain
rises on Hamlet, and we see a human figure blowing his fingers and

stamping his feet on a dimly lit stage. Before a word is uttered, we

have already inferred at least a cold night as the context for his
speech. We read or hear recited a sonnet by Shakespeare: “To me,
fair friend, you never can be old and no matter how little we
know about William Shakespeare of Stratford and the various earls
with whom he may have been intimate, we immediately begin to
create for those words a plausible and appropriate context: at the

minimum, a speaker addressing some other person whom he regards

as fair and, in some sense, as his friend. All our experiences with
language and the contexts in which men speak not only enable us to
make this inference but really oblige us to make it.
Throughout our lives as verbal creatures, we have learned to
respond to linguistic structures in a certain way: namely to interpret
their meanings, to infer their contexts from their forms. Many of
the characteristic and valued effects of poetry as a representational
artform depend on the strength of our habitual tendency to infer
contexts from verbal structures. We should note that Milton, in

Paradise Lost, does not create Eve or Eden; what he creates, rather,
are statements about “Eve“ and “Eden” that lead the reader to
create a woman and a place — in order, as it were, to provide
referents or “meanings” for those statements. Other representa-
tional artforms depend for their effects on comparable tendencies

32
 POETRY AS FICTION

in the spectator: illusionist painting, for example, depends on

fundamental habits of visual perception to transform a configura-
tion of lines and colors on a flat surface into the appearance of a
three-dimensional scene or object. It is only because of perceptual
conditioning produced by our experiences in the natural visual world
that we can see, as a cow grazing in the distance, what is actually
only a few brushstrokes of color on the upper part of a canvas. This
process of interpretive filling-in or perceptual inference is very
similar to the process by which we infer, from a few lines in a poem,
a rich context of motives, feelings, and situations. “To me, fair

friend, you never can be old. ...” Nine small words that summon
up for us a man, his consciousness of the pathos of mutability, and
his impulse to deny its hold on his friend.
Thus, although a poem is a fictive utterance without a particular
historical context, its characteristic effect is to create its own context
or, more accurately, to invite and enable the reader to create a
plausible context for it. And what we mean when we speak of
interpreting a poem is, in large measure, precisely this process of

inference, conjecture, and indeed creation of contexts.* But these

contexts
— —
“meanings” that we half create and half perceive can be
no more than “plausible,” for the poem is a, fictive utterance and its

contexts can be neither discovered nor verified in nature or history.

As we saw earlier, when we interpret a natural utterance, we usually
seek to ascertain its historical determinants, the context that did in

fact occasion its occurrence and form. However complex and elusive
that context, it is nevertheless understood to be historically determi-
nate and particular. The context of a fictive utterance, however, is

understood to be historically indeterminate. This is not to say that

we regard the poem as an anonymous gift dropped from the
empyrean or ignore the fact that it was composed by a real man at a
particular time and place. It is to say, rather, that we may
distinguish between the poet’s act of composing the poem and the
verbal act that the poem represents, just as we would distinguish
William Shakespeare’s act in composing Hamlet and the acts of the
Prince of Denmark represented in the play. Shakespeare composed
the play, let us say, in 1603, but in what year did Hamlet kill
Claudius? In one sense, he kills Claudius every time the play is

33
 TOWARD A THEORY OF FICTIVE DISCOURSE

performed, whether in 1603 or 1970; but in another sense the slaying

of Claudius is an act that never did, never will, and never can occur
in the historical world. It can only be represented as occurring. The
composition of the play, then, was a historically determinate event,
but the events represented in the play are historically indeterminate.
This means, among other things, that when we ask why Hamlet
abuses Ophelia in the nunnery scene, we do not expect to find the
answer in any historical particulars of the life of William Shake-
speare or of the circumstances that occasioned his composition of
the play. Knowledge of these particulars and circumstances may, of
course, help us account for why Shakespeare wrote a play in which a
character named Hamlet abuses a character named Ophelia, but
that is an altogether different question. To understand why Hamlet
abuses Ophelia, the reader must infer from, on the one hand, the
linguistic structure of the play and, on the other hand, everything he
knows about the world of men and the relation of their acts to their
situations and motives, a plausible set of motives and situations for
that act.

Similarly for a sonnet by Shakespeare, say 87, which begins:

“Farewell, thou art too dear for my possessing, /And like enough
thou knowst thy estimate. ...” To interpret it as a poem, to un-

derstand why the speaker is saying “farewell” in such apparently

bitter tones to someone on whom he thought to have some claims of
love, the reader will not require any particulars concerning Shake-
speare’s private life: the identity and moral character of whatever
young men he knew at the time, the specific incidents of personal
betrayal from which he may have suffered, or his opinion of himself

as a lover. What the reader does require is the capacity to conceive

of the kind of situation that might \ea.d a man to feel thus and speak
thus, and the reader can develop that capacity only out of his own
experiences with men, their situations, their feelings, and especially
their language.

The interpretation of a poem as a historical utterance may serve

the special purposes of the literary historian or biographer, but it is

likely to appear shallow, reductive, or “literal-minded” precisely to

the degree that it restricts the context of the poem to historical
particulars and suggests that the meanings of the poem are to be

34
 POETRY AS FICTION

located exclusively in a historically determinate context. For ex-

ample, a recent editor of Shakespeare’s Sonnets prefixes the following
note to Sonnet 107 as part of his running commentary on what was
happening in Shakespeare’s personal life at the very moment he was
writing the poems: “Shakespeare had just escaped from the danger
of his Company’s involvement with the Essex rebellion and the . . .

Queen, furious with Pembroke for fathering Mary Fitton’s child

and refusing to marry her, had sent Pembroke to jail. .’’^
Then . .

comes the sonnet:

Not mine own fears, nor the prophetic soul
Of the wide world dreaming on things to come.
Can yet the lease of my true love control.
Supposed as forfeit to a confined doom ....
Forfeit, indeed, to a confined doom, if interpreted as this editor
suggests. But for the reader who values the sonnet as a verbal
artwork, this sort of interpretation is absurd not only because its

foundations in history are, in fact, quite dubious, but because the

invocation of particulars of this kind —even if they were accurate
have no greater claim to constituting the “meaning’’ of the poem
than an interpretation that infers from it and provides for it an
appropriate and coherent context of circumstances and motives,
quite independent of Pembroke, Mary Fitton, particular jails, and
particular rebellions.
I pointed out earlier that personal letters often exhibit character-
istics that we associate with poetic discourse, a “literariness’’ that is

produced, for example, by unusually well-controlled syntax, pre-

cision of diction, elaborateness or specificity of descriptions, im-
agery, allusion, and metaphor. Since a letter will be read in a
context both temporally and spatially remote from that in which it

was composed, the writer must provide exclusively through its lin-

guistic structure the sort of supplementary information that is other-

wise, in spoken utterance, provided by the physical context shared
by speaker and listener and also by the speaker’s intonation and
gestures. The letter writer, in other words, must exploit all the
expressive possibilities of language itself to enable his reader to infer
and reconstruct properly the meanings and context of his original

utterance.

35
 —
TOWARD A THEORY OF FICTIVE DISCOURSE

The poet is obviously operating under the same limitations, but

even more so. He must convey to his readers not only a context
remote from them in space and time, but one that may never have
existed in history or nature, that may consist entirely of what the
reader will be enabled to construct (rather than reconstruct) from
the verbal form of the poem. Moreover, the poet must suggest,
exclusively through a linguistic structure, the experiences, attitudes,
and feelings — and, indeed, the identity—of a speaker of whom the
reader has no other knowledge. Finally, especially since the text of
the poem will function as the script for its future performance (and
by reciters other than the poet), it must specify or direct its own
vocal realization, including its pacing and other intonational
features.
The poet will, therefore, in the verbal structure he composes, be
straining to the limit all the expressive resources of language. And,
beyond that limit, he will sometimes devise new ones. But what are
sometimes spoken of as “poetic devices” (and we may include here
rhythm or meter) are really the potentially expressive features of

natural discourse. Tropes and figures, distortions of idiomatic

syntax, departures from idiomatic diction, imagery and allusion
all these are certainly not restricted to poetic discourse; nor can they
be taken as the distinctive characteristics of poetic language. They
are not what defines poetry but are, rather, entailed by what does
define it, namely its fictiveness.

Because a poem does not reflect but create the context in which its

meanings are located, its linguistic structure must carry an extraordi-

nary burden. Poetic language seems — and indeed — is richer, more
“suggestive” and “evocative” than the language of natural discourse
precisely because and to the extent that it requires the reader to
participate in the creation of its meanings. In our efforts to interpret
the poem, to construct the context of human situations and motives
it demands in order that its meanings be realized, we will draw on all

our experiences of the world and words of men. Indeed, the activity
of interpreting poetry often becomes the occasion for our recognition
and acknowledgment of otherwise inaccessible feelings and, in a

sense, ourown otherwise unknowable knowledge. The richer and

more extensive our experiences and feelings or as we say, “the —
36
 POETRY AS FICTION

more we bring to the poem” —the more significance it can have for
us, which is why, of course, subsequent readings of a poem “reveal”
more meanings. The language of a poem seems characteristically
“concentrated” because it allows for such an extraordinary and
continuous expansiveness of meaning, not confined to finite and
particular determinants, but drawing on all we know that we can
relate to it. The language of the poem continues to mean as long as
we have meanings to provide for it. Its meanings are exhausted only
at the limits of the reader’s own experience and imagination.
In speaking of the contexts created or projected by the reader, I

have repeatedly used the term plausible; and although I have

seemed meaning fits, wear it,”
to be saying, “If the I have also
implied that the meaning must fit. This fitness and plausibility
relate to significant constraints on interpretation that are themselves
among the conventions of fictive discourse. Though these con-
straints differ in many respects from those involved in our inter-
pretations of natural discourse, they are nevertheless substantial;
and although there are inevitably grounds for argument in deter-

mining them for individual poems, they are nevertheless relatively

determinate.
The poet, in composing the poem, will have made certain assump-
tions regarding his audience, specifically that they are members of a
shared linguistic and cultural community, and thus able and willing
to abide by relevant linguistic, cultural, and indeed literary conven-
tions. Although a poem is a representation of discourse, we can
understand it, infer meanings for it, only through our prior experi-
ences with the sort of thing it does represent, namely natural
utterances in historical contexts. The poet assumes, therefore, that
his reader has a knowledge of the language represented by the poem
and the linguistic conventions that govern the relation of an
utterance to its meanings in that language. However, as we all know,
linguistic convention can hardly be separated from cultural con-
vention. An appropriately informed reader who encounters the word
God in a poem by a seventeenth-century Englishman is not likely to
interpret it as the deity of the Muslims or Hopi Indians, any more
than he is likely to interpret a painting of the coronation of Elizabeth
as the crowning of the Queen of Siam. Furthermore, the poet will

37
 TOWARD A THEORY OF FICTIVE DISCOURSE

assume that his readers are capable of identifying his composition as

one of a kind — a genre—of artwork, and therefore of interpreting it

in relation to those generic and artistic conventions that operated for

him in composing it. Thus the reader who is quite unfamiliar with
the forms and traditional functions of the masque, and mistakes
Comus for the script of an ordinary theatrical comedy, will obviously
be interpreting it inappropriately.® We should note here that the
poet’s assumptions are not to be confused with his intentions.
Whereas the latter — his intentions — are specific, personal, and can
usually only be surmised, the former— his assumptions —are gen-
communal, and are therefore more likely to be
eral, recoverable.
The linguistic, cultural, and generic constraints on interpretation
alluded to here are, of course, what much professional criticism (or
“philology” in the broad sense) is directed toward establishing. And
to the extent that it is engaged in determining the existence and
nature of such assumptions and conventions on the basis of histori-
cal and publicly accessible data, criticism is a cognitively respectable
enterprise issuing in at least theoretically “verifiable” and indeed
cumulative knowledge (granting the probabilistic nature of veri-

fiability in regard to historical facts and the inevitable grounds for

uncertainty and controversy regarding their relevance to individual
poems). It would be well, however, to recognize the distinction
between this enterprise, which is more or less continuous with that of
the cultural historian, and the aspect of professional criticism
mentioned in footnote 6, that is, the public articulation and
elaboration of the critic’s experience as the audience of an artwork.
Both are commonly spoken of as “interpretation” and, of course,
both frequently appear in conjunction, but claims that may be made
for the one cannot be made for the other, and their functions and
value are distinctively different. The meanings of a work that a
philological “interpretation” seeks to establish are those that the
poem bears in relation to the historical universe in which it was
composed, and are themselves historical and determinate; but the
meanings that the poem has by virtue of its characteristics as a
fictive utterance are historically indeterminate and thus cannot be
the object of objective or cumulative knowledge, though we may for
various reasons find their “interpretation” by individual readers

38
 POETRY AS FICTION

interesting and valuable. It might be added that each of these types

of interpretation may, in turn, be distinguished from those inter-
pretations briefly alluded to below as the reader’s (and, when
publicly elaborated, the critic’s) hypotheses and inferences concern-
ing the poet’s governing artistic design. The meanings of “inter-
pretation” are no less multiple than the meanings of “meaning.”
A final observation made here regarding the view of
should be
interpretation I have been developing. To recognize a poem as fictive
rather than natural discourse, as a verbal artwork rather than an
event in nature, is to acknowledge it as the product of a human de-
sign in accord with certain valued effects. I have not discussed here
the very crucial question of the distinction between the effects or
functions of fictive and natural discourse because it is a question
that involves substantial problems in linguistic as well as poetic
theory, and could not be dealt with briefly.’ We should, however, at
least acknowledge the fact that part of the effect of a poem, as
distinct from a natural utterance, derives from the reader’s aware-
ness of the poet standing, as it were, behind the poem as its creator
and artificer. This awareness is commonly reflected in our in-
also
terpretations, for among the meanings we seek for and infer from a
poem are those that, in Aristotelian terms, might be called its final
causes: that is, the motives or intentions, the governing design, of
the poet as an artist, distinct from either a natural speaker or the
fictive speaker of a poem. Thus, we can interpret Hamlet’s abuse of
Ophelia both in terms of a plausible set of human motives projected
for Hamlet and in terms of a plausible set of artistic motives projected
for Shakespeare; and the same sort of double interpretation could

be offered for any poem.

This double aspect of interpretation reflects a more fundamental
doubleness in the nature of poetry, indeed the duplicity of art itself.

As we view the canvas, the myriad spots of paint assume the guise of
natural objects in the visual world, but we are nevertheless always
half-conscious of them as spots of paint. As we watch the play, the
stage recedes and the personal identities of the actors yield to those
of the fictions whom they portray, but when, at the final curtain, we
clap our hands, it is not Hamlet whom we are applauding, but the
performers and the playwright himself. The illusions of art are never

39
 TOWARD A THEORY OF FICTIVE DISCOURSE

c/elusions. The artwork interests, impresses, and moves us both as

the thing represented and as the representing itself: as the actions
and passions of Prince Hamlet and as the achievement of William
Shakespeare, as the speech of men —and as the poet’s fictions.

40
 3

ON THE MARGINS OF DISCOURSE

I. Definitions and Classifications

What a Disease is, almost every Physician defines.
The Anatomy of Melancholy, I.i.2

Asked (or challenged) to define poetry, one is likely to reply with a

sigh, a shrug, a look of exasperation or even one of contempt,

‘indicating not only that the question is oppressive but that anyone
who asks it must be something of a fool, a pest, or a vulgarian.
Though these uncongenial reactions may be interpreted as the signs
of intellectual embarrassment, they are, I think, quite justified. For
the nature of definition and the particular historical fortunes of the
term poetry conjoin to this effect: that a definition of the term will

either be a total chronicle of those fortunes or will constitute merely

one more episode in them. In other words, a definition of poetry is

bound to be either inadequate to the job or, if adequate, then both

unmanageable and uninteresting for any other purpose.
Problems of definition have attended literary theory almost from
the beginning, which is to say, Aristotle. An obsessive and scrupu-
lous classifier, he notes in the Poetics that while the term poet
(woLT^Tris) is applied indiscriminately to all those who compose in

meter, even if the composition is a treatise on medicine or physics,

there is no term to distinguish and label what he evidently believes is

the more significant class, namely, mimetic verbal compositions.

41
 TOWARD A THEORY OF FICTIVE DISCOURSE

whether in meter or not. The same problem bemuses Sidney and

Ben Jonson (the observation of it becomes, in fact, a commonplace
of Renaissance criticism), and it is hardly solved by Coleridge who,
in the space of three pages in the Biographia Literaria, offers a series
of five definitions of poetry: poetry in the lowest sense, a legitimate
poem, poetry of the highest kind, and poetry in the strictest use of

the word, all of which are set aside in favor of a definition of the
poet, which turns out to be a definition of the Imagination.
The difficulties are understandable and are still with us. A
descriptive definition of a term such as poetry is primarily respon-
sible to a history of usage: it attempts to specify the conditions
governing the conventionally sanctioned use of the term in a given
linguistic community. But the conditions governing the use of the
term poetry have expanded, contracted, and shifted many times over
in the history of its occurrence in actual utterances; and the pre-
sumable extension or denotation of the term —that is, poems, or
the actual practices of “poets” — is itself a historical set of phenom-
ena, with characteristics that develop and often change radically
without particular regard for, or sense of obligation to, traditional

definitions.
The difficulties are increased, moreover, by the very fact that the
term poetry' has been subjected to such a long history of explicit and
self-conscious defining, so that any proposed definition must pay
due respects not only to ordinary usage but also to what is, in effect,

a series of independent and often conflicting traditions of usage and

normative definition in the hands of critics, manifesto writers, and
other theorists. One — implied, example, when we
other tradition for
speak of “poetry in the broad sense” — has arisen partly from that
same dissatisfaction that Aristotle felt: the suspicion or conviction
that there exists an interesting and significant class of verbal
compositions, the distinctive nature of which is obscured by the
conventional association of the term poetry with verse. I will return
to that point later, but for the moment we might note that it would
certainly be easier to define poetry if it had not been defined so much
already.
Complicating the enterprise even further is the fact that the
English word poetry has put in long service as a translation of

42
 ON THE MARGINS OF DISCOURSE

presumably equivalent wordforms in other languages, and has

thereby acquired intensions and extensions carried over from signi-
ficantly different cultures and linguistic communities, with histories
of usage that may conflict in fundamental ways with the history of
the term’s usage in English. Finally, the definition becomes even
more problematic as a result of our increasing interest in and
knowledge of the “poetry” of remote and primitive cultures. For
although it is understandable that we should appropriate that term
to refer to compositions that share so patently some of the charac-
teristic features of our own verbal artworks, we inevitably discover in
them other features, functional as well as formal, of a quite different
sort; and these discrepancies in turn threaten to obscure or muddle
whatever coherence and continuity of reference may be claimed for
the term in our own culture and linguistic community. Thus we ask
whether Hopi creation myths are really poems, and whether poetry
really includes or should include anonymous communal chants.
The difficulties described here are reflected in those classic prob-

lems of classification that, even in the most sophisticated circles, are

forever raised, repeated, and reinvented: problems arising either

from intractably divergent views of how poetry should be defined or
from the pressure of borderline cases on inherently feeble borders.
Thus we open the gates between verse and prose in order to allow in
such desirable immigrants as novels or “imaginative” essays, only to
discover that we then have no way to keep out their undesirable
relatives, such as vivid journalism or elegantly written philosophical
papers. Or we establish more stringent criteria for admission in
order to keep out the riffraff, only to find that we have thereby also
excluded some natural-born citizens. Pope’s Essay on Man and
“Thirty Days Hath September” may, we say, be verse (or poetry in
the formal sense, or in the loose sense, or in your sense), but they
certainly are not poetry (or not poetry in the real sense, or in the
most emphatic sense, or in my sense); and Plato’s Dialogues, the
sermons of Jeremy Taylor, and Moby Dick may not be verse (or
poems in the formal sense, and so on) but they certainly are poetry
(or poetry of the highest kind, or in the most universal sense, and so
on). Disputes over such claims cannot be resolved nor conclusive
answers given to comparable questions of classification created by

43
 TOWARD A THEORY OF FICTIVE DISCOURSE

the innovative productions of poets, the discoveries of anthropolo-

gists, or the formulations of new generations of literary critics. For,
as these disputes and questions themselves bear witness, the actual
practices of the English linguistic community with regard to the
wordform poetry have been and remain so divergent and inconsis-
tent that any resolution or answer would constitute merely one more
ad hoc and essentially arbitrary definition of the term, with no
greater claim than any other to descriptive accuracy. It would not,

then, be at all unreasonable or frivolous to conclude that poetry

simply cannot be usefully defined.'
To observe that the term poetry cannot be usefully defined, or that
there is a constantly shifting and dissolving borderline between what
we usually call poetry and all the other things from which we might
like to distinguish it, does not oblige us to deny the possibility of any
relevant distinction between classes of verbal composition. It merely
suggests that the distinctiveness of those classes is not reflected in
the consistency of our labels for them. Their distinctiveness may,
however, be reflected in something else at least as significant,
namely, the distinctiveness of our actual behavior and experiences
with respect to them. It seems clear, for example, that no matter
how vague or naive our literary theories, or how problematic our
explicit definitions, we do xmkQ functional discriminations between,
say, biographies and novels, and between the transcriptions of
actual utterances and the scripts of plays, through the very manner
in which we experience and interpret them, and the sort of value and
implications they have for us. In other words, we take them as
different kinds of things and, accordingly, take them differently.
Most children learn at a relatively early age that some of the
things we tell them are “really true” and others are “just stories’’ or,
more generally, that sometimes we are saying things to them and at
other times using language in a rather different spirit and with a
different force. They learn to make this distinction quite in ignor-

ance of, and independent of, categories such as fact and fiction or
chronicle and tale. Nor do they make the distinction on the basis of
the inherent credibility or “imaginativeness’’ of a narration: for
many contemporary storybooks narrate banal events about banal
characters hardly distinguishable from events and persons in their

44
 ON THE MARGINS OF DISCOURSE

own lives, while many things we tell children truly must seem inher-
ently incredible in terms of a child’s own experiences. (How believ-

able, for example, can a child of four find our statement that men
have traveled to and walked on the moon? Yet the child will appre-
ciate the difference between our him that and our telling him
telling

a story about a boy with a red balloon.) The distinction between, on

the one hand, things that are said and, on the other hand, things
such as stories, nursery rhymes, songs, and verbal games is learned,
rather, on the basis of the child’s own differential experiences with
respect to each: the different contexts in which they occur, the
different vocal tones in which they are delivered, the different stylis-

tic features they may exhibit, but most significantly, the differ-

ent force — implications and consequences —they have as verbal

structures.^
The point I wish to emphasize here is not merely that children
learn to make that sort of discrimination before they have terms for
it, but that even when, as adults, the more sophisticated distinctions

and labels are known, we continue to make functional discrimina-

tions of a comparable sort and on comparable grounds, producing,
experiencing, and responding to various verbal structures differen-
tially, in terms of what might be called “covert” categories, that is,

categories implicitly acknowledged and respected in the culture, and

learned by its members, but cutting through and across the explicit

distinctions presumably reflected in traditional terms such as poetry,

prose, literature, fiction, and nonfiction.
If the reality and significance of such covert functional categories
can be recognized, then one may
abandon the thankless
cheerfully
enterprise of stalking definitions for such terms and direct one’s
attention instead to how we do, in fact, differentially produce and
experience various kinds of verbal events and compositions.
The distinction I draw between “natural” and “fictive” discourse
is offered as an approach to such an alternative enterprise.^ It is not
intended to serve as a solution to the problems of definition men-
tioned earlier but, rather, as a diversion from them. “Fictive dis-
course” is not another name for “poetry” or “imaginative litera-

ture”; nor is “natural discourse” an equivalent of “prose,” “propo-

sitions,” or “ordinary language.” Neither, in fact, is quite an

45
 TOWARD A THEORY OF FICTIVE DISCOURSE

equivalent of anything else for which we have a name. “Natural”

and “fictive” are labels I have chosen to designate what I believe are
functional categories in the sense described above. I also believe that
exploring the nature of these categories, and the rather special
nature of their distinction, may yield an interesting perspective on
much that we do speak of as poetry or literature and on its relation

to all our experiences of language.

It would not be unreasonable, of course, to ask whether the
distinction between fictive and natural discourse developed here
does not itself give rise to problems of classification; ambiguous and
borderline cases comparable to those associated with traditional
distinctions between poetry and whatever it may be distinguished
from. Can every existing or conceivable verbal composition be
assigned to one or the other of these categories without difficulty,
ambiguity, or occasion for controversy? Do these classes, in effect,
represent an absolute distinction? The suggestion that they do is

likely to arouse skepticism in tempermental monists or those who

have learned from intellectual history to be suspicious of all

dualities, polarities, and starkly drawn boundary lines. In recent

years, moreover, a number of influential literary theorists, having
observed the repeated failure and apparent futility of every tradi-
tional attempt to establish a watertight definition of poetry or
literature, have maintained that there cannot be any distinction
drawn between poetic and nonpoetic discourse. “ In view of such
skepticism, the alternative concept of a “continuum” may seem
more attractive and inherently plausible here. Thus, one might be
inclined to argue that the differences between natural and fictive

utterances are fundamentally relative and quantitative, and that

what are described here as discrete kinds are really only distant

points on a spectrum of some sort, a spectrum of degree of

stylization, for example, or of personal expressiveness. I would not
deny that verbal compositions could be arranged on continua of
various kinds (though not without difficulty, as has been demon-
strated many times over by attempts to do just that) but, as it

happens, the principles according to which the items could be thus

aligned have nothing to do with the basis of the distinction between
fictive and natural discourse. To put this another way, there is no

46
 ON THE MARGINS OF DISCOURSE

principle of relative differentiation that could allow us to speak of

any given composition as “more” or “less” fictive or natural and
thereby to assign it its proper place on the continuum.
To give due weight to skepticism, however, the distinction we are
concerned with here may be thought of as relatively absolute. The
apparent paradox arises from the fact that, although, for reasons to
be discussed below, the status of a given composition as either fictive

or natural may be problematic and, indeed, may switch under

various circumstances, nevertheless on any given occasion that
conposition must be taken as one or the other. Natural discourse was
defined earlier as all utterances, spoken or inscribed, that can be
taken as someone’s saying something, sometime, somewhere: all

utterances, that is, that are understood to be the verbal acts of

particular persons on, and in response to, particular occasions.
What is to be emphasized here is the significance of those phrases
“taken as” and “understood to be.” The distinction between natural
and fictive discourse is absolute, as absolute as the distinction
between being married or single, or between being Canadian or
American; but, like these, it is equally a matter of convention. The
Roman Catholic Church may refuse to recognize the civil divorce of
communicants, and the chief of an Amazonian tribe may refuse to
grant one the privileges afforded by one’s American citizenship; for
both these classifications depend on assumptions and conventions
that have meaning and consequence only insofar as they are
recognized and honored within some community. Similarly, the
classification of any particular verbal composition as natural or
fictive has meaning and consequence only insofar as those concerned
with it share certain assumptions regarding how it is to be identified
and interpreted, how it is to be taken.
Since the conventions in question here must be learned, it may
happen that they are imperfectly mastered or, on some occasion,
improperly used or inadequately signaled. Consequently, the appro-
priate classification of a given composition may be mistaken or
doubtful. In quoting a line of verse allusively, we may be mistaken

by our listener to have spoken in our own right. When the comedian
in a nightclub act suddenly exclaims, with good reason, “Sorry,
folks, we’re being raided by the police,” the audience may laugh.

47
 TOWARD A THEORY OF FICTIVE DISCOURSE

having mistaken the remark as part of the act, and thus fictive.

Moreover, since these conventions are conventions, they may, like

the rules of a game, be switched by the agreement of all the

—
players case usually a
in this tacit agreement signaled in various
ways — and what properly taken
is to be a natural utterance on one
occasion may, with equal propriety, be taken as fictive on another,
or vice versa. It may happen, of course, that controversy will arise in
connection with individual works, particularly those exhibiting
internally contradictory or ambiguous features. For example, we
may disagree as to whether Wordsworth’s Prelude is to be taken as
versified autobiography, and thus natural discourse, or the repre-
sentation of a possible autobiography, and thus fictive; and we may
also disagree about how to determine the answer. Similarly, reasons
might be urged for taking Boswell’s Life of Johnson as fictive and
Emily Dickinson’s verse as natural. In each case, however, the classi-

fication one chooses will differentially direct, or be directed by, one’s

experience of the work and the manner in which one interprets it.

Propriety of classification here is not merely an academic or

philosophic concern. As numerous and now familiar psychological
experiments have demonstrated, one’s perception of and/or
response to an event not only determine but are determined by how
one classifies what we “see,” and how we subsequently behave
it:

toward it, will depend on what we see something as. Moreover, since
how we classify an event usually depends on our prior experiences
with events having similar properties, and an event is likely to share
properties with events in a number of other classes, quite practical
problems of classification may arise accordingly: our prior experi-
ences may be relatively limited, and conflicting possibilities of
appropriate classification may present themselves. The guest who
must ask his hostess if he may deposit cigarette ashes in a certain
ceramic dish may have been understandably unsure of whether to
“take” the dish as an ashtray, a saucer, or an objet d'art because it

seemed to share distinctive properties with all three.

The fundamental perceptual or cognitive processes that are

involved in what we call “generalization” (or, by some, “concept
formation”), and by virtue of which prior experiences enable us to
classify and thus deal more effectively with relatively novel situa-

48
 ON THE MARGINS OF DISCOURSE

tions, may also have personally or socially “dysfunctional” conse-

quences, ranging from simple errors and perplexities to prejudice,
superstition, and what Gilbert Ryle calls “category-mistakes.” ^ The
latter, broadly interpreted, would include all those instances in
which the classification of an event or object (or person) made on the
basis of certain of its properties yielded inappropriate assumptions,
expectations, and responses. With respect to discourse, this suggests
that, for us to behave appropriately in response to a given utterance,
we must classify it as, among other things, either fictive or natural.

We must know if a certain narrated anecdote is to be, as we say,

“taken literally” or taken as a joke, and if a certain rather
sensational piece of literature is to be understood as the reflection of
the author’s personal history or as the representation of a “confes-
sion.” Since it is usually to the advantage of the speaker as well as
his listener, or of the poet as well as his reader, that the proper
classification be made, identifying cues or signals, either formal or
contextual, are usually presented: the “joke” is related in a dis-
tinctive style and/or with distinctive vocal intonations; the repre-

sentation of a confession is offered among a group of compositions

labeled or otherwise identifiable as “poems.” In each case, the
classification of the utterance as fictive or natural will entail certain

distinctive assumptions, expectations, and responses. How we take

an utterance, or what we understand it to be, will determine the
dynamics of our experience of it, the nature and process of our
interpretation of its meanings, and the kind of consequences and
value it has for us.
It should be noted, however, that we always have the option of
behaving inappropriately in response to a given verbal structure, not
merely mistaking a natural for a fictive utterance or vice versa, but
knowingly taking it as something other than what it was “given” as.

Thus, the guest, in our analogy above, lacking a suitable receptacle

for his cigarette, may appropriate a “saucer” for the purpose or,
finding that the “ashtray” lends itself to aesthetic observation, may
choose to regard it as an objet d'art. Similarly, correspondence of
intention and reception is not required for a given verbal structure to
function as either natural or fictive discourse on any particular
occasion. We may say that, in taking a verbal structure as something

49
 TOWARD A THEORY OF FICTIVE DISCOURSE

other than what it was given some way other than its original
as, in

constructor intended or expected it to be taken, we have in effect

regiven it to ourselves.**

II. Fictive Illustrations AND Examples

Not all fictive utterances are artworks. We fabricate verbal struc-
tures for a variety of purposes, including many quite remote from
aesthetic ends or effects, and there are numerous kinds of verbal
compositions which no one would speak of as poetry in any sense
but which are, nevertheless, not natural utterances. There are, for
example, those beguiling propositions constructed by logicians to
illustrate a point or a problem: All swans are white. Some swans are
white. No swans are white. This series could, no doubt, be taken as a
minimalist poem, but it is certainly not offered as such by the
logician, and it is not, in any case, natural discourse. For although
the logician has constructed and presented these sentences, he is

not, and will not be taken to be, saying them — nor is anyone saying
them, for they are not being said at all. There are other kinds of
fictive utterances that verge, in interesting ways, on poetry, and yet
other compositions that verge on, or rather totter between, both
fictive and natural discourse. These linguistic curiosities include
such heterogeneous verbal structures as greeting-card messages,
commercial advertisements, quotations, and proverbs. What relates

them to each other is the fact that they all either stand beyond or
straddle the border that separates natural from fictive discourse. It

will be useful to consider these marginal cases here for a number of

reasons: first, in order to clarify the nature and suggest the
significance and interest of that borderline itself; second, in order to
illustrate the range and variety of Aio/zaesthetic fictive utterances;

and finally, in order to indicate both the relation and the distinction
between poetry and its compatriots beyond the border.’
We may begin by considering the kind of verbal structure that
consists of illustrative or exemplary utterances such as the logician’s
swan song or the propositions and sentences constructed by philos-
ophers, linguists, and the authors of foreign-language textbooks to
illustrate a lesson or exemplify a point. A vocabulary list in a Latin
textbook, for example, me, nihil, non, amare, terrere, is followed by

50
 ON THE MARGINS OF DISCOURSE

a series of sentences exhibiting those words: Me non amat. Nihil me

terret, and so on. These sentences are not natural utterances; they
are not being said by anyone, including the author of the textbook.
He is not thought to be making revelations about his social life or
asserting anything about his own character. Nor are they the
transcriptions of utterances actually said by any particular person on
a particular occasion. Of course, utterances having that form could
be and no doubt have been said, but so could and have the
individual words on the vocabulary list: Me? Non! Like the indivi-
dual words, these sentences are verbal forms but not verbal acts or
events. Both can be, or become, natural utterances, but only when
they occur as the verbal responses of specific people to historically
specific sets of circumstances. As textbook examples, they are not
natural utterances and do not have meanings in the sense that
natural utterances do. For the meanings of the latter are located in
the historical universe, confined by and to the particular circum-
stances that actually occasioned them. But the meanings of both the
individual words on the vocabulary list and the sentences that
exemplify them consist of a range of possibilities, the possible
circumstances in which those forms would be appropriate, the
contexts in which they could, given the conventions of the linguistic
community, plausibly and properly occur.
The logician’s exemplary sentences can be regarded in similar
terms. When we encounter, in a philosophic article, such structures
as/1// swans are white or The cat is on the mat, we understand that,

although they have the form of propositions, no one is proposing

them; that, although they are in some sense statements, no one is

stating them. They are not natural utterances, but fictive ones:

verbal structures fabricated, constructed, created by someone, but

not asserted by him.
The same is understood of those often grotesque, if grammatically
“correct,” sentences devised by linguists: John and Harvey like the

play and are disappointed by it respectively: John is easy to please:

John is eager to please: The carpenter struck the nail with the
hammer. Again, one observes that utterances having these forms
—
could occur some of them, anyway. Philosophers and linguists are

marvelously adept at devising sentences that evoke a surrealistic

51
 —
TOWARD A THEORY OF FICTIVE DISCOURSE

society where conversation consists of the exchange of limp, gratu-

itous, and otherwise improbable declarations. I will return to that
later, but for the moment we are noting that although these
sentences con/Jbe said, they are not, in these instances, being said.
It would be absurd to ask Professor Chomsky for an introduction to

this fellow John who is so easy and eager to please.

It would be no less absurd, of course, to inquire after the histori-
cal identity of the young woman to whom Christopher Marlowe’s
passionate shepherd addresses his invitation to pastoral pleasures
Come live with me and be my love — or, indeed, to inquire after the

historical identity of the shepherd himself. The poem, too, was never
really said by anyone to anyone —though, as a Active utterance, it is

in a sense being said eternally. There are, as we shall see in a

moment, significant respects in which the poet’s lyric and the lin-
guist’s example differ, but both are Active discourse: ahistorical,

noncontextual verbal structures, possible utterances, but not actual

ones.®
The nature of the differences between poems and the other Active
utterances we are considering here may be clarified by an analogy
drawn from visual art. The sample sentences and propositions of the
philosopher and grammarian are verbal illustrations', they stand to
poems as certain pictorial illustrations stand to visual artworks. The
drawing of a laurel oak in a botany text, for example, or of a
balalaika in a treatise describing East European musical instru-
ments, is a Active visual structure. Such a depiction may obviously
be distinguished from a particular oak tree or a particular balalaika,
each of which is an object in nature and history, with a specific
identity in time and space. The illustration of the laurel oak, appear-
ing starkly on the white page of a textbook, surrounded by nothing
but printed words, is, in effect, a noncontextual, ahistorical visual
structure. As such, however, the illustration can also be distin-
guished from the representation of an oak tree in a painting. A
painting of an oak tree is also a Active visual structure, but the
artist’s depiction, unlike the illustrator’s, is designed to evoke in the
spectator a more or less vivid impression of the tree’s substantiality,
particularity, and historical identity. In other words, the painting
suggests or implies a specific context in nature and history: the

52
 ON THE MARGINS OF DISCOURSE

artist’s oak tree, composed though it be of spots of pigment on a flat

canvas, nevertheless appears to grow from tufts of grass, appears to

cast a shadow, appears to be “seen” against the line of horizon, and
so forth. Both the textbook illustration and the painting may be
referred to as depictions, but only the painting is a fully mimetic
representation. In the illustration, the expanse of white paper on
which the highly generalized tree is depicted is seen as just that:
white paper. In the painting, however — let us say it is a watercolor
that blank expanse is assimilated into the representation: it becomes
the sky above the horizon or, indeed, the three-dimensional space
which the tree appears to occupy. By implying a context, the painter
creates the illusion of one, eliciting the spectator’s imaginative pro-
jections and perceptual inferences.’
The verbal “illustrations” discussed above may be distinguished
from verbal artworks —poems— in comparable terms. Both are ahis-

torical, noncontextual linguistic structures, but the philosopher’s

The cat is on the mat and the Latin textbook’s Me non amat are not
fully mimetic: they depict utterances, but not utterances-in-contexts.
The natural context and physical substantiality that the painter
implies by colors and lines suggesting shadows and horizons are
implied by the poet only through the expressive resources of language;
but these resources are formidable. The sentence Odi et amo could
appear in a Latin text as an illustration of the first person singular in
first- and fourth-conjugation verbs. As such, it would be serviceable
but hardly more evocative than / eat and I sleep. As the opening
phrase of a lyric by Catullus, however, it effectively summons into

existence an individual speaker and a substantial context of feelings,

motives, and circumstances. In interpreting the poem, we do not, as

with the verbal illustration, merely acknowledge that there is a range

of contexts in which such an utterance could occur; we infer or

project or imagine a particular context for it, a world in which the

speaker is heard to bite out the alternately bitter and melancholy
words that reflect the ambivalent emotions aroused in him by his

faithless lover. “Ot// et amo. ...”

The expressive resources of the poet are formidable because they
draw on all our prior experiences with natural utterances, particu-
larly how we learn to interpret them, and even more particularly the

53
 TOWARD A THEORY OF FICTIVE DISCOURSE

very fact that we are inclined to interpret them. As we noted earlier

in passing, the illustrative sentences constructed by philosophers
and linguists evoke a rather peculiar world, the sort of world where
people make observations such as “The cat is on the mat,” or “The
carpenter hit the nail with the hammer.” They evoke this world,
however, only when the reader yields to the impulse, carried over
from his experiences with natural discourse, to infer historical con-

texts from what appear to be verbal events. The impulse is obviously

misdirected in these instances. No world of cats, mats, carpenters,
nails —or speakers with any reason for talking about them —stands
behind these Active utterances. They are designed to be regarded as
noncontextual. A poem, however, is designed to evoke a context: it

takes advantage of the very impulse that is inappropriate with

respect to the philosopher’s or linguist’s example, that is, the impulse
to interpret, to infer historically particular meanings from verbal
structures.
To conclude this section, we may return to that series of proposi-
tions concerning the whiteness of swans. I suggested earlier that one
might take the series as a minimalist poem; we may note now what
happens if one does so. Titled “Disillusionment,” for example, or
“Disorder and Early Sorrow,” it might suggest a speaker’s mournful
expression of his graduation from naive optimism, to rueful skepti-
cism, to abject nihilism:

Allswans are white.

Some swans are white.
No swans are white.

Or, presented with other titles or without any title at all, it might
suggest other sorts of interpretations: a comment on the decline of
aristocratic values, for example, or a terse resume of an anticoloni-
alist coup in an African nation. What is sometimes called “found
poetry” operates, of course, on the same principle. Any verbal
structure — a set of newspaper headlines, an obituary notice, the list

of ingredients on a box of packaged cereal —can be isolated from its

original context and presented in such a way (lineated, for example,

or read aloud in a studied manner) as to suggest poetry and to invite
our response to it as such, that is, as a verbal artwork, the represen-

54
 ON THE MARGINS OF DISCOURSE

tation of a natural utterance in an implicit dramatic context, de-

signed to invite and gratify the drawing of interpretive inferences.
Poems elicit from us the projection of particular speakers re-

sponding to more or less particular situations. Because we under-

stand them to be fictive utterances, however, we recognize that the
contexts we infer from them are also fictive: unfixable, unlocatable,
in the natural universe — as unreal as the sky we seem to see above
the painter’s oak tree. No matter how vivid the speaker and his
world may be, the poem remains but a possible utterance, and the
meanings that we, as readers, infer from it remain essentially our
own creation, unverifiable in fact and principle. Far from being a
defect, this aspect of poetry is the source of many of its valued
effects —a point to which we will return after we have considered a
number of other nonpoetic but fictive compositions.

III. Exploiting THE Margins

No one takes the philosopher or linguist to be speaking in his own
right in the examples he presents, and the distinction between the
utterances he constructs and those he utters is clear. Deliberately

more ambiguous is the status of those statements in which commer-

cial products are commended in the popular media, particularly
radio and television. “Everybody knows,’’ of course, that the vacu-
ous housewives, earnest doctors, and wide-eyed toddlers who cele-

brate the virtues of various soaps, aspirins, and puddings are actually
actors, their “sincerity’’ part of the performance, their words no
necessary or even likely reflection of their personal feelings and
motives. No one would call such actors liars, for there is a convention
or understanding between advertisers and their audiences regarding
what commercials are and how they are to be taken. Advertisers are
disposed, however, to strain the convention and, in any case, always
count on a certain degree of breakdown of that understanding: a
blurring between act and enactment or at least a carry-over of our
responses to natural discourse.
For our present purposes, several points are worth noting here.
The first is the simple fact that many advertisements are, like poems
and the illustrative sentences of linguists and philosophers discussed
above, fictive verbal structures. They stand in relation to poems as

55
 TOWARD A THEORY OF FICTIVE DISCOURSE

their pictorial counterparts stand in relation to pictorial artworks. A

landscape with milkmaid by Constable or a landscape with beer
*

drinkers by Mr. X: both are Active visual structures, though there

are again significant differences between them in other respects. As
with a poem or a philosopher’s example, what is crucial to the
fictiveness of an advertisement is not that the statements of which it

is composed are false (a certain product may reduce cavities; another

one may bring headache relief faster) but rather that they are not
being by those who “say” them. Such an advertisement is not a
sa/c/

natural utterance, moreover, because it never occurs as a unique

historical event on a particular occasion in, and in response to, a

particular historical context: the same ad looms from the same

electronic void hourly or on twenty consecutive days. Its repeated
presentations are experienced by the viewer not as later re-assertions
of statements asserted previously, but as the replay of the same
ahistorical statements, just as the various performances of a poem
are experienced as the same verbal structure re-cited, and the vari-
ous performances of a play are experienced as the same set of
ahistorical actions re-enacted.
A second point is commer-
that most audiences recognize that, in
cials, the conventions of natural discourse have been suspended and

replaced by other conventions that govern the relation between what

is “said” and what may be inferred therefrom. An ad is certainly not

interpreted the same way a poem is; but neither is it interpreted as a

natural utterance. The audience understands the pretense'® and
responds accordingly: we learn to ignore the apparent “imperative”
force of all those enjoinders to buy, try, use, taste, and wear the
various products mentioned; we learn to shut off what would other-
wise be our reactions to a fellow creature’s enthusiastic recom-
mendations and tones of urgency.
As I suggested above, however, the effectiveness of modern adver-
tising depends on a certain amount of breakdown of the conventions
or leakage between the barriers: our unwitting carry-over of our
assumptions regarding natural discourse, particularly our assump-
tion that people usually mean what they say. The reverse can also
occur: as a consequence of learning to steel ourselves against the
daily blandishments of commercial advertising, we may become

56
 ON THE MARGINS OF DISCOURSE

extremely sensitive to the possibility of both fictiveness and deception

until eventually the assumption that people usually mean what they
say is replaced by the assumption that they usually don’t. That a
skeptical or even cynical attitude toward discourse may carry over to
natural utterances is only one of the ways in which language becomes
a casualty of the advertising industry.
We might remark, in this connection, that the “understanding”
presumably shared by advertisers and their audiences must be rein-
forced by law, sometimes quite explicitly, as when we are informed
that “The following is a paid political announcement” lest we mis-
take the nature of a citizen’s hearty and apparently ingenuous
endorsement of a candidate. Indeed, FCC “rulings” in such instances
can be regarded as the actions of a referee in the interest of what is

quite literally “fair play."' For ad viewing is a sort of contest: one in

which the advertiser tries to compel belief by any means short of
outright deception, and the viewer tries to maintain disbelief as far
as possible short of losing whatever information is there to be gotten.
(It is useful to know that there is a product that reduces cavities — if
it reduces cavities.)
Finally, we can recognize that highly sophisticated audiences are
often capable of something close to an “aesthetic” appreciation of
commercials. Secure in their understanding of the conventions that
operate, they respond not as listeners in a natural verbal transaction,
where the interest of what is said lies in what it reveals of the speaker
and the world to which he is responding, but as the spectators of a
more or less entertaining performance, the skill, novelty, and wit of
which can be appraised with respect to other works in the “genre”
and, indeed, even enjoyed as such
— “disinterestedly.”

IV. Prefabricated Discourse

The commercial possibilities of fictive discourse are not confined to
“commercials” proper but extend to the literal selling of fictive
utterances. Poems and novels are marketable, of course, but what I
have in mind here are certain verbal structures on sale at one’s local
drugstore as well as bookstore, of which the following is a fair
example:

57
 —
TOWARD A THEORY OF FICTIVE DISCOURSE

To Mother, with Love from Both of Us

This world may change in many ways
But you just grow more dear
And closer, Mother, to our hearts
With every passing year.
So this message from the two of us
Is very fondly said
Just keep on being your own sweet self
Through all the years ahead!

This is, to be sure, dreadful stuff. The crudeness of its rhythm,

the triteness and bathos of its sentiments, the vacuousness of its

diction; all this, plus another quality we might be inclined to call

“insincerity,” makes us wince or groan. Would it be proper to sum
up our characterizations and reactions by saying that this is abysmal
poetry? I think not; for it need not be regarded as poetry at all.

Greeting-card verse is like lyric poetry, of course, not only in its

formal pretensions and personal mode but also because it is Active

discourse. The text printed above is obviously not a natural utterance
nor is it the inscription of one (though it is significant that the
typography in which the original was set was evidently designed to

resemble cursive handwriting). Though the statement is personal, it

is not the statement of the person who composed it: it does not
spring from or reflect his emotions and sentiments; it is not his
mother who grows more dear; and he could not very well be “the two
of us” who are purportedly saying the fond message. In fact, no one
is saying it, and it is not a message. Like a poem and like a linguist’s

verbal illustration, it is Active discourse; unlike these, however, it is

designed to he a message under appropriate circumstances.

Actually, there need not be any insincerity involved in the creation
or use of greeting-card verse, for certain conventions obtain here and
are presumably respected. No one will take the author of this verse
to mean what it says, and those who purchase and subsequently send
the verse as a personal message may very well mean what it says. As
the author’s composition, the message is Active; but once it is signed
“John and Mary” and sent to Mrs. Jones in Cincinnati, it becomes
natural discourse and, the conventions having switched, will be
taken, perhaps quite properly, to affirm their sentiments.

58
 ON THE MARGINS OF DISCOURSE

As a fictive utterance, the meaning of the message is indeed

vacuous: like the meaning of the linguist’s or philosopher’s examples,
it consists only of a range of possible conditions (or contexts) in
which it could be an appropriate response. As a natural utterance,
however, it acquires particular meaning with respect to a specific
context: it means what John and Mary feel and want As they
to say.
might have put it themselves: “We chose this card because it was so
fitting.’’

Greeting-card messages are not poems. What they are is something

we may call prefabricated utterances: verbal structures preassembled
for later use as natural utterances. Indeed, it is precisely because the
greeting-card message must meet the expressive needs of so many
people that its language is so vague and general: the more precise
and specific the diction and allusions, the narrower the range of
possible contexts and thus the fewer potential sayers and, of course,
buyers. Had the message included some reference to Mother’s blue
eyes or apple pies, it would have eliminated as potential customers
all those whose mothers had brown eyes and were bad cooks. Greet-
ing-card verse is like American canned (that is, prefabricated) soup,
with its characteristically undistinctive seasoning designed to accom-
modate as wide as possible a range of tastes. In a sense, there is no
poetry more “universal’’ than greeting-card verse.
This last point, concerning generality and universality, has some
bearing on the distinction between greeting-card messages and poetry
proper, that is, verbal artworks. The precision of diction and specifi-
city of allusion that are so conspicuously lacking in these messages
are, in poetry, highly valued qualities. “ In poems also, the more pre-
cise the language and specific the allusions, the more restricted and
particular becomes the context in which the utterance would be
appropriate. But in poetry, that restriction and specificity serve the
characteristic effects noted before, that is, the vividness and force
with which poems do evoke particular contexts and enable us to
project individualized speakers with distinctive and subtly defined
emotions. The world of imaginable circumstances and feelings that
a poem implies, the response to such a world that the poem repre-
sents, are what interest and move us in it. A greeting-card message,
though fictive, is not as mimetically “replete’’ as a poem. Its function

59
 TOWARD A THEORY OF FICTIVE DISCOURSE

is not to represent a natural utterance but rather to become one. It

need not, must not, evoke a particular fictive context because it will

acquire a particular real one.

As we shall see later on, poems can and often do serve a function
quite comparable to greeting-card messages, that is, they are appro-
priated for use as natural utterances. Before turning to that possi-
bility, however, I should like to pursue a bit further the notion of
prefabricated discourse in general. The fact that greeting-card
messages are prefabricated does not make them any the less sincere
(or, of course, “natural”) when they serve as the messages of real

people. Natural discourse does not mean unique or original discourse

and, indeed, the concepts of uniqueness and originality must always
be rather strained in connection with language. A very large propor-
tion of natural discourse consists of verbal formulas—conventional
phrases, idiomatic “expressions,” even whole sentences — we that
have heard and used many times before. In spite of the much-vaunted
creativity of speech and theoretical infinity of possible sentences, it is

clear that most of our informal utterances are composed of a dis-

tinctly finite number of verbal structures: phrases and formulas that

we learn along with individual words and rules of syntax.
As Albert Lord has demonstrated in his study of contemporary
oral epics, the bard or “singer of tales” can recite lengthy heroic

poems on demand, not because he has memorized those poems, but

because he can simultaneously compose and recite them, just as we
all simultaneously compose and speak in natural discourse.'^ The

oral poet learns a large but finite number of traditional verbal

formulas which, in conjunction with a store of suitable variants,
form a substantial repertoire of what are, in effect, prefabricated
lines of verse; and this repertoire can be drawn on instantaneously to
serve a practically infinite number of compositional occasions, that
is, to suit any narrative, thematic, syntactic, or metrical context. For
example, in traditional Yugoslav poetry, most of the characters
apparently have horses and are frequently described as mounting
them. Fhe phrase translated “mounted their horses” is a metrical
unit (specifically, a half line), and it may appear numerous times at
various points in the composition of a singer, each time with suitable
variants:

60
 ON THE MARGINS OF DISCOURSE

All the horsemen mounted their horses . . .

And mounted their horses

the wedding-guests . . .

Then they leaped up and mounted their horses . . .

Moreover, each of the variants here is itself a formula that may

appear in other lines combined with yet other formulaic units:

All the horsemen rode into Baghdad . . .

And the wedding- guests toasted the bridegroom . . .

Then they leaped up and called for their saddles . . .

Each line of the poem may be unique, but each one is composed of
preassembled sections, like an automobile or prefabricated building.
As Lord points out, these bards are none the less original poets, and
each poem reflects continuous creative activity. But originality and
creativity here are evidently not to be equated with absolutely unique
combinations of words.
The generation of speech in natural discourse occurs in a directly
comparable way: in speaking and writing, we draw on a repertoire of
verbal patterns or formulas we have learned and used before, substi-
tuting, as the occasion requires, the variants that are appropriate to
the particular needs of themoment. In the sentences written above,
for example, there occur a number of phrases quite common to
expository discourse: As I suggested earlier, what I have in mind,
something we may call, as Lord points out. These expressions are
clearly prefabricated, and asX points out is clearly a verbal formula
where A" is replaced by a suitable name depending on the immediate
context. In composing these phrases I did not, as it were, start from
scratch. As it were, start from scratch — it is really very hard to be
verbally creative.
Whenever economies of time, space, or energy become significant,
and subtlety or the marks of originality matter little, we are likely to
employ those conspicuously prefabricated verbal structures we call
cliches. In composing telegrams, for example, or interoffice memos,

or impersonal business letters, most of us readily make use of

formulas of expression which are nevertheless meaningful and sin-

cere enough for the purpose at hand: Congratulations on the new'

arrival and best wishes to both of you; Thank you for your letter of
X. . . . I look forw'ard to having your reply, and so forth. We avoid

61
 TOWARD A THEORY OF FICTIVE DISCOURSE

cliches, of course, when we do not wish to appear perfunctory,

impersonal, or uncreative; and young writers and others who take
too seriously the notion that original minds generate continuously
original utterances often become so self-conscious in their search for
unique combinations of words that they can hardly write at all. Even
when we are not using cliches, however, we must still employ a great
number of conventional or formulaic expressions simply in order to
speak with the rapidity and fluency required for natural verbal
transactions; and most of us do take advantage of those preassembled
structures that the linguistic community has evolved, structures that
exhibit a clarity, economy, and euphony of expression that recom-
mend themselves to most native speakers.
We should notice, moreover, that if one breaks it down into small
enough units, any utterance consists wholly of what are, in truth,
prefabricated verbal structures — not merely idomatic phrases and
conventional formulas but individual words themselves, which we
need not buy or borrow because they are freely donated to us for our
immediate use by the linguistic community — all nicely preassembled
from constituent speech-sounds. But it may be asked here: if phrases
and individual words are to be regarded as prefabricated verbal
structures, does this mean that they are also instances of fictive
discourse? And if they are, then what can’t be?
The answer to the first question is, of course, “it depends.’’ Word-
forms (lexemes) or phrases conceived of abstractly (for example,
“the word /ire,’’ “the phrase law and order'') are not natural utter-
ances. Nor are they fictive utterances. They are simply something
else altogether: verbal /’orm5, types or classes of events, but neither
events themselves nor the representation of events.
An individual wordform may be not merely conceived oJ\ however,
but produced vocally or in writing: typed, for example, by a lexi-

cographer on a file card for the dictionary he is compiling, or

dictated by an instructor at a spelling examination, or jotted down
by a poet as a euphonious or otherwise interesting and potentially
useful linguistic item. In such cases (and many others can be thought
of), although the wordform is being typed, voiced, or written by a
particular person on a particular occasion, it is nevertheless not
being said (or inscribed) as natural discourse and would not properly

62
 ON THE MARGINS OF DISCOURSE

be taken or interpreted as such by a listener or reader. If our

lexicographer, spelling instructor, or poet had produced the word
conflagration, for example, it would have been as a response, not to
a fire experienced, recalled, or imagined at that moment, but simply
to the verbal form conflagration itself. (The significance of the dis-

tinction becomes obvious if one thinks of the difference between the

spelling instructor dictating conflagration and his saying “Fire!”)

Here, too, however, it would be improper to speak of the verbal

structures thus produced as fictive utterances. For the Xtvm fictive,
as used in this study, suggests representation of some kind, even if

not always fully mimetic representation: the philosopher’s and lin-

guist’s illustrations, the commercial ads, and the greeting-card mes-

sages discussed earlier, and the quotations and proverbs to be
discussed below — all are representations, either of possible utter-
ances, or kinds of utterances, or past utterances. In other words,
fictive discourse always bears some relation to natural discourse
other than mere distinction from it. Moreover, fictive discourse

suggests the product of fabrication or construction, that is, a verbal

composition presumably consisting of more than a single linguistic
unit and exhibiting at least a rudimentary sort of structure. Thus it

would seem to make sense to reserve the terms fictive utterance and
fictive discourse for representational verbal structures with at least
the minimal degree of length and syntactic organization that we
usually associate with the terms utterance and discourse.
It must be emphasized, however (in answer to the second question
raised above), that there is in principle no minimum length or
degree of syntactic organization required for a verbal structure to
function as an utterance and thus, in effect, nothing (that is, no
verbal form) that cannot be, on some occasion, used and taken as
fictive discourse. Many verbal exchanges consist of brief, syntacti-
cally detached phrases, and many consist of single words:

“Where are the keys?” “/n the car.''

“How’s the family?” "Fine."
“Can I speak to Mrs. Smith?” “Wait, I’ll call her. Mother!"

In the situations suggested by these examples (if one takes them for
the moment as being real), the italicized words and phrases are

63
 TOWARD A THEORY OF FICTIVE DISCOURSE

complete linguistic events. But if an isolated phrase or single word

can be a^natural utterance, then it can also be a fictive one. For, to

turn the tables now, the very examples I just used to illustrate the
possible brevity of natural utterances are, of course, not natural
utterances at all, but verbal illustrations like those of the linguist
and philosopher discussed earlier; and I have therefore just con-
structed fictive utterances consisting of one word. I might add that
there can be and are poems, or at least verbal artworks, consisting of
one word: for example, Pedro Xisto’s ''logograma" entitled “Epitha-
lamium 11” and consisting of the word She (printed, to be sure, in
an interesting way).*'’

V. Quotations
Although poems are not, like greeting-card messages, designed to
become natural utterances, they are often appropriated for that
purpose. We may recall here all those occasions, sometimes trivial,

sometimes not, when phrases, lines, or even longer passages from

literary works come to our lips and pens not in the course of our
specifically citing or referring to them but, rather, as the borrowed
containers for our own meanings. Those with bookish lives are ever

ready to appropriate the words of others, and literature professors

and graduate students are notably inclined to sprinkle even their
casual utterances with literary locutions such as ‘‘Thereby hangs a
tale,” or to leave parties saying things like ‘‘Good night, sweet
ladies, good night,” or ‘‘When shall we three meet again?” The
containers are often borrowed at least partly for the sake of the
residual taste of the original contents. When T. S. Eliot dedicated
The Waste Land to ‘‘Ezra Pound, il miglior fabbro," the borrowed
phrase from Dante’s Purgatorio now carried Eliot’s compliment to
Pound and no longer, as in the original, the poet Guinizelli’s praise
oi' his mentor, Arnaut Daniel; but it may be supposed that Eliot still

wished to retain some of the flavor and resonance of the gracious

allusion as it occurs in the original context. Nevertheless, Eliot was
not simply quoting the phrase from Dante in the sense of citing it;

he was saying it, making it his own by meaning it, though in an

altogether different context and with a totally different referent.

64
 ON THE MARGINS OF DISCOURSE

There is a crucial distinction between speaking q/'and speaking the

words of others.
We often use the term quotation in yet another sense, with refer-
ence not to a citation or a borrowing but rather to a reporting of
someone else’s words. In the course of telling an anecdote, for
example, may quote another person directly: “Well, John was
I

pretty angry at that. 7 Vc had enough,' he said, and slammed his

fist on the table.” Here a verbal action, what John said, is being
reported or described along with other nonverbal actions, what he
did. I can describe John’s words more precisely than his other ac-

tions because Iam describing them in words and thus, we may say,
iconically. I am not, however, reproducing his verbal act exactly,
since the original utterance must have included, among other things,
vocal idiosyncrasies and intonational features that are inevitably lost
or distorted when thus duplicated. Nevertheless, I am not simply
referring to John’s words as I refer to his slamming his fist on the
table; I am, rather, depicting them, adequately enough for the
purposes of the anecdote.
If we recall the discussion of verbal depictions in section III above
and the analogy drawn there to visual depictions, it becomes evident
that a quotation of the sort just described is also a verbal picture.
Here, however, the representation is not of a possible utterance or
an imaginary one, but of a real and particular one —drawn, we
might say, from life. Indeed, we could regard a quotation as a sort
of verbal photograph (or pretechnological p/zowograph) of a natural
utterance, neithermore nor less iconic than a snapshot or home
movie of John’s slamming his fist on the table. The analogies here
are, in every sense, diverting; the significant point is that, in quoting

John’s words, I am understood not actually to be saying them but

rather to be presenting a facsimile of them, and thus a fictive
utterance.
The distinctions among referring to, depicting, and saying the
words of others obtain, we should note, whether the original source
is itself fictive or natural. We may refer to or cite passages of poetry

without affirming them, and we may appropriate the words of real

people as readily as those of characters in plays or novels. We may
remark, “As lago says. What you know, you know," or “As Dr.

65
 —
TOWARD A THEORY OF FICTIVE DISCOURSE

Johnson said, To a poet nothing can be useless." In both cases the

implication is: as he said, so I am saying. We will say what others
have said, be they characters or historical persons, if we can make it

mean or take it to mean what we mean — and are inclined to say.‘^

When we appropriate someone else’s words, we confer on them

our own context of meanings, often quite remote from that of the
original. In such cases, we are using their words in a sense meta-
phorically, taking what was originally a reference to one state of
affairs and applying it to a different but comparable or analogous
state of affairs. It is not the individual words, however, but the
utterance as a whole that becomes metaphoric. An illustration may
clarify the point. A chapter of I. A. Richards’ Principles of Literary
Criticism, entitled “The Chaos of Critical Theories,” bears as an
ironic epigraph the following passage from Henry IV, part I: O
monstrous! But one half- pennyworth of bread to this intolerable

deal ofsack.'*" In the play, this occurs as Prince Hal’s exclamation

on seeing Falstaff’s bill at the Boarshead tavern, the point being
that the proportion of wholesome food to wine in Falstaff’s diet is

appallingly meager. As the epigraph to Richards’ chapter, it implies

that critical theories exhibit a comparable disproportion between
nourishing and merely heady stuff. What originally referred “liter-
ally” to Falstaff’s habits of consumption now refers figuratively to

the quality of critical theories.

We should note, however, that the line from Henry IV never really
referred literally to anything; it was a wholly fictive utterance, not

said on a particular occasion by a historical Prince Hal but con-

structed by William Shakespeare to be recited by any number of
future actors in any number of future performances of the play.
Metaphoric quotations need not, of course, originate in literary
works: numerous epigraphs that function in a way similar to
Richards’ consist of passages quoted from natural utterances
letters, for example, or public speeches. Nevertheless, passages
from poems, novels, plays, and songs do lend themselves more
readily to metaphoric extension of this sort because they are fictive
and we understand that, being so, they are in themselves essentially

free of historically specific “literal” meaning.

The reverse can also occur, that is, we can take what were origi-

66
 ON THE MARGINS OF DISCOURSE

nally natural utterances and make them fictive, thus setting them
loose from their historical moorings and specific meanings and open-

ing them to metaphoric interpretation. In Donald Finkel’s poem,

“Angel Hair,’’ for example, there occurs the following passage:

Most Angels are uneducated. Only one Angel in

10 has steady work.‘^

The passage originally appeared in a magazine article about motor-

cycle gangs (such as “Hell’s Angels’’) and was written without
apparent irony. The fact and effects of its metaphoric extension in

Finkel’s poem are obvious. The poem itself, it might be added, is

but a step away from “found poetry,’’ the marginal aspects of which
were considered above. Here we can note that when we take a
natural utterance —an obituary notice, for example, from the
columns of a newspaper —and present it in or as a poem, we are
inviting the reader to interpret it metaphorically or parabolically,
not with respect to its original context, but as implying some other,
perhaps more general, significance:'*

Albert Molesworth
Eighty-seven years old.
Owner of the nation’s largest
And most prosperous potato farm.
Died yesterday
At his home in Idaho
He left

no
survivors.

The implication becomes a sort of bathetic sic transit gloria mundi.

One may be reminded of Shelley’s “Ozymandias.’’
Some imagination and wit are involved in the finding of found
poems, but certain natural utterances obviously lend themselves
more readily than others to fictive transformation. Conversely, cer-
tain fictive utterances seem especially suitable for appropriation as
natural discourse —eminently quotable, in other words, or likely to
appear in the pages of an anthology such as Bartlett’s. Among the
qualities that make them so is something we commonly refer to as
“universality” but which often amounts simply to generality. The

67
 .

TOWARD A THEORY OF FICTIVE DISCOURSE

listings in one anthology of quotations contain, for example,

numerous passages like the following, culled from various poems
and plays: He that sleeps feels not the toothache; Whoever loved
that loved not at first sight; Everyone can master a grief but he that
has it. He that . . ., whoever . . ., everyone. . . These lines have
that sweep of reference that we associate with proverbs: One man's
meat is another's poison; He that fears you present will hate you
absent, and so on. The pronouns here, he and you, are essentially
vacuous or indeterminate in meaning: like one {on in French, man
in German), they refer to no one in particular and everyone po-
tentially. What is interesting in such passages, as well as in the
proverbs cited, is the fact that, when thus extracted, they do not
evoke particular contexts: they do not imply individualized speakers
responding to historically specific circumstances. On the contrary,
they are very much like greeting-card messages in that they are so
readily utterable in such a wide variety of contexts by speakers quite
different from the original sayers and in circumstances quite remote
from theirs. They are so amenable to repeated utterance because
they are so repeatedly meanable. Recalling our earlier distinction,
then, it would seem that the most universally quotable passages of
poetry are the least poetic ones: the most general and therefore the
least mimetic and evocative. It would be more to our purpose,
however, to observe that in isolating passages of poetry from their
original fictive contexts and pressing them into service as the “con-
tainers” of our own historically determinate meanings, we are re-

authoring them as natural utterances, inviting and expecting from

our listeners a response to them different from that which was
presumably invited and expected by their original authors.
Consider, for example, Shakespeare’s Sonnet 94, the one which
begins “They that have power to hurt and will do none ...” and
concludes with the strikingly proverbial line, “Lilies that fester smell
far worse than weeds.” Excerpted as a proverb, this line will

suggest, as it does in the poem, something like: noble creatures,

when corrupted, are more offensive than merely base creatures. We
may have occasion to “quote” the line, that is, to appropriate it and
assert it “meaningfully” in a particular set of historical circum-
stances. Experienced in the context of the poem, however, the line

68
 ON THE MARGINS OF DISCOURSE

will be interpreted as the fictive speaker’s assertion (or his own

“pointed” application of a proverb); and what interests and moves
us in the line is not its dubious wisdom or its excerptibility for later
quotation but, rather, its expressive quality with respect to that
speaker’s feelings and motives, and the personal circumstances they
appear to reflect.

My point here has not been to malign quoters or to suggest that

the quotation of poetry is morally or aesthetically baleful; as we all

know, it can be done subtly, imaginatively, and quite creatively (for

example, by Keats in his letters). I have been concerned, rather, to use

the various practices we call “quotation” to clarify the relation be-
tween fictive and natural discourse, especially the conditions under
which a given verbal structure can shift from one to the other, and the
consequences of such a shift for our experience of the structure. We
will return to the specific relation of proverbs to poems, touched on
just above, after we have looked more closely at the nature of proverbs
themselves.

VI. Saying and Sayings

Proverbs are often referred to as “sayings.” A saying, we note, is

neither a said nor a says: the gerundive form identifies neither

person nor tense. It suggests speech without a speaker, a self-

sufficient verbal object rather than a verbal act, an utterance that
asserts itself independently of any utterer —continuously, as it were,
or indeed eternally. The proverbs cited earlier, and others such as
Misery loves company and All’s fair in love and war, are anony-
mous: not only are the original sayers of these sayings unknown, but
no trace remains of the occasions on which they were originally said.

They seem, in fact, never to have been initially said at all but rather
to have arisen in toto from the race, culture, or verbal community as

a sort of eternal verbal response to an eternal state of affairs. All

proverbs do, of course, have particular origins even if they are no
more readily traceable than the origins of individual words or in-

ventions such as the wheel. But, like the latter, they have become
cultural artifacts; they are no man’s or, rather, all men’s property,
as much a part of the lexicon of a linguistic community as individual

words or those idiomatic expressions they obviously resemble.

69
 TOWARD A THEORY OF FICTIVE DISCOURSE

Taken as verbal structures, proverbs are, of course, fictive ut-

terances; but, like greeting-card messages, they become natural
utterances under certain circumstances. It is one thing to speak of a.
proverb:

My mother’s favorite proverb is “Live and let live.” My

father’s is “Business is business.’’

and another thing to say it:

She’s always nagging him that he drinks too much or

smokes too much. I [or, “I say’’] live and let live.

Yes, our families are old friends, but you know [or, “like

they say’’] business is business.

When it is not actually being said in (or, as we sometimes express it,

applied to) a particular situation, a proverb is simply a verbal form,

like an individual word. As such, it is noncontextual and its meaning
(like that of a wordform, or a linguist’s example, or a greeting-card
message) is historically indeterminate; that is, it consists only of a
range of possible contexts in which the proverb could be said,

possible conditions under which it could be a natural utterance. But

when we say a saying (as when we sign and dispatch a greeting-card
message), we confer a particular context on it; we transform it from
a verbal form to a verbal act.

It is often said that proverbs embody timeless wisdom or universal

truth. It would be more accurate to say that the range of contexts in

which they could be appropriately affirmed is unusually broad and

recurrent. One reason for that is a characteristic which many
proverbs share with greeting-card messages, namely, generality and
indeed vacuity of language and sentiment. Although proverbs are
not, like greeting-card messages, designed to serve as prefabricated
utterances, they achieve that status by virtue of comparable quali-
ties. By a sort of natural selection, those proverbs that survive are
literally the fittest; that is, they fit the widest variety of circum-
stances or adapt most readily to emergent environments. And that is

because their meanings are indeterminate enough to cover almost all

human, natural, and historical exigencies. If the conditions in which

a proverb could be affirmed disappear, it loses authority and the
proverb itself disappears.

70
 ON THE MARGINS OF DISCOURSE

Proverbs also achieve “universality” by virtue of another quality

that we have encountered before, namely, the readiness with which
they lend themselves to metaphoric or parabolic application. When
we apply a metaphoric proverb (such as One swallow doesn't make a
summer, which is rarely said in connection with swallows or sum-
mers), we are in a sense interpreting it, but no single application
exhausts its meanings:

John: Are you still fixing that salad dressing?

Bill: Rome wasn’t built in a day.
John: I think I’m getting somewhere with Alice. She smiled at me
this morning.
Bill: One swallow doesn’t make a summer.
John: Rome wasn’t built in a day.

There are, of course, some limits to the range of meanings of such

proverbs, usually maintained through a continuous history of inter-
pretation and application. If, however, we are not familiar with the
tradition of application, the metaphoric value (and limits) of a
proverb may be impossible to determine. A certain Ashanti proverb,
for example, goes The okra does not show its seeds through its skin.
It is by no means evident that this botanical observation, which does
not show its seeds through its skin, means “You can’t tell what a
man is thinking from the expression on his face.” Archer Taylor, in
his interesting study of the genre, mentions a number of proverbs
that have come down to us altogether without a tradition of applica-
tion and are thus virtually limitless in meaning. My favorite, I think,
is this one from the Old High German: When it blows, the trees
shake. Very true, of course, and very proverbial sounding, but what
does it mean? Well, we can take it to mean almost anything, from
“Persons in subordinate positions become anxious when their su-
periors are in a bad humor,” to “Certain events are inevitably

produced by other events.” It is, in effect, an all-purpose utterance,

an unimpeachable truth for all seasons.
It was observed earlier that a proverb will lose authority and
eventually disappear if the conditions in which it could be affirmed
no longer hold. Presumably “extinction” of that sort has been the
fate of numerous proverbs that appear in collections but are no

71
 TOWARD A THEORY OF FICTIVE DISCOURSE

longer actually heard, for example, A woman hath no weapon but

her tongue. It should also be recognized, however, that since a
proverb seems to derive its authority from eternal transcendent
truth, it will often be affirmed in the face of immediately contra-
dictory circumstances —sometimes with the implication that, if we
sit the moment out, the reality that the proverb asserts will return
and restore the force of its truth. Because the “saying” has no
known original sayer, it appears uncontaminated by ordinary
human error or bias, and thus oracular. This accounts, we may
suppose, for the exceptional durability of some manifestly foolish
bits of “wisdom.” It also suggests why it can be so maddening to
argue with someone who utters proverbs in reply to one’s carefully
thought-out proposals or empirically verified observations:

“Women make up 40% of the labor force. One third of

all women with children are employed. Government-spon-
sored child-care is available in every enlightened nation in
Europe. ...”
”
“A woman’s place is in the home.

It is like waging battle with an adversary who is both immovable and

immortal.^®
The very anonymity or oracular nature of proverbs often recom-
mends their citation to speakers who would not, for various reasons,
wish or be able quite to say them. As Taylor observes, for example
(speaking of proverbs as distinguished from formal religious doc-
trine), “A sound skepticism pervades proverbial wisdom and ven-
tures to show itself in assertions which no member of the folk would
dare utter on his own.”^'
Along with somewhat similar lines, Ruth Finnegan, discussing the
Central African sanza (which means both proverb and circumlocu-
tionary expression generally), writes:

This hidden and oblique form of speech, then, with its over-
tones of playing safe and avoiding direct commitment, is
one developed to a high degree among the Azande. How-
ever, it seems to be an element in all use of proverbs, one
which comes out particularly in situations of conflict or

72
 ON THE MARGINS OF DISCOURSE

uneasy social relationships and where depths of hidden

meaning are implied.

More generally, we may say that the citer of proverbs manages to get
something said without having to take the responsibility for having
said it. (Often, the speaker cites the proverb as if he must concede
its sorrywisdom although he would himself wish matters otherwise.)
In saying only what “they” say, the speaker disclaims responsibility
for the utterance but does not wholly dissociate himself from either
its general “truth” or its applicability to the particular situation at
hand. He attempts, in other words, to have it both ways. We touch
here on a matter which, though it cannot be pursued on this oc-

casion, is of considerable interest: namely, what might be called the

ethics of verbal transactions. “ For our present purposes, it is suf-
ficient to observe that the assumptions governing the mutual claims
and responsibilities of speakers and listeners in natural discourse
are, in fictive discourse, evidently suspended or substantially quali-
fied. And we might notice, in this connection, that poetry as well as
proverbs may allow a speaker to utter what he could not or would
not say. The poet himself may offer in a poem sentiments which he
would not otherwise wish attributed to him; and, in quoting a poem
written by someone else, we may present sentiments that we are not
quite willing to assert. (A variation of this possibility is familiar to
those who live with —or remember being—adolescents, for whom
the singing of pop lyrics often seems to serve as something between
the occasion for and the evasion of self-expression.)
Proverbs are obviously related to poems in a number of other
ways. Many proverbs have originated or been preserved in works of
literature and, as has often been observed, the formal character-
istics of each are in many respects similar: cohesiveness, euphony, a
high incidence of phonemic repetition, alliteration, parallelism,
figurative language, and so forth. Both proverbs and poems are
verbal structures identifiable exclusively in terms of their linguistic
forms. Both are, or can be taken as, fictive discourse, possible
utterances with possible contexts. And, most significantly for our
concerns here, both are for that reason especially amenable to being
interpreted (or “applied”) metaphorically or parabolically. We do
not, however, characteristically regard proverbs in the same way as

73
 a

TOWARD A THEORY OF FICTIVE DISCOURSE

we do poems, that is, as mimetically replete representations of dis-

course; nor do we characteristically experience or “interpret” them
as such.
There is, as it happens, an eighteenth-century Japanese poem —
haiku — that resembles quite closely the Old High German proverb
mentioned above {When it blows, the trees shake). The poem,
literally translated, goes: “When the west wind blows, fallen leaves
gather in the east.”^'‘ The similarities are evident, but there is a
crucial difference in our experience of and response to each:
whereas the proverb, cited as such, independent of any particular
occasion of application, stands as a vacuous truism, the eternally
sayable saying said by no one, the haiku is read and in effect
“heard” as a human utterance, occasioned by and referring to the
experiences and feelings of a single speaker. Explicit interpretations
of the poem will vary among readers depending on the richness,
subtlety, and specificity of the inferences they are inclined to draw
and articulate. One could, for example, read it as the speaker’s
bemused observation of the simultaneous ferocity and frailty of
nature, or as his elegiac reflection on the inevitability of seasonal
change and of mortality in nature and man. The point, however, is

drawn from the poem but not

that inferences of that sort will be —
from the proverb. The poem invites and allows the construing of
such motives and sentiments for two reasons: partly because of the
relatively greater specificity and vividness of its diction and al-

lusions, but primarily because we take it to be a poem and respond

to it accordingly, that is, by interpreting it.

Interpretation — the construing of a particular set of conditions, a

context, that could plausibly occasion an utterance of that form — is

not merely a possible response to poetry but a characteristic one. In

effect, to take a verbal structure as poetry is implicitly to acknowl-
edge it as a Active utterance designed (or, in the case of “found
poetry,” presented) to invite and reward precisely that kind of
activity. “ It should be emphasized that “interpretation” in this

sense is an activity that differs from both the interpretation of

natural discourse and the interpretation of other forms of Active
discourse. When we speak of interpreting a natural utterance, we
refer not to the construction of a possible context but to the recon-

74
 ON THE MARGINS OF DISCOURSE

struction of an actual one: the attempt to identify or infer the

particular set of conditions that did in fact occasion the occurrence
of that verbal event. The other forms of fictive discourse examined
here (textbook illustrations, logicians’ paradigms, greeting-card
messages, proverbs) are, like poems, understood to be not verbal
events but verbal structures. Such structures, to serve their char-
acteristic functions or purposes (exemplification, illustration, sub-
sequent employment as natural utterances, and so on), must be, in a

sense, “meaningful,” but only in the sense that there must be some
range of contexts in which utterances of that form could appro-
priately occur. The activity of “interpreting” a verbal structure such
as the textbook’s Me non amat or the philosopher’s The cat is on the
mat is complete when that range of possible contexts has been
identified or even just acknowledged to With respect to these
exist.

nonpoetic forms of fictive discourse, richer or more particularized

interpretation is, as we have seen, uninvited and quite gratuitous:
occasionally distracting, sometimes amusing, but, strictly speaking,
not to the purpose. With respect to a structure we take as a poem,
however, interpretation is not thus limited and certainly not gratu-
itous. Here, the activity of construing —imagining, projecting, elab-
orating — a particular plausible context is invited; and we accept the
invitation on the assumption and with the expectation that the
structure has been designed to make that activity an engaging and
gratifying one. In short, with respect to what we take to be poetry,
interpretation is understood to be the purpose.

75
1 t

*
V 'it:
’
-JkUr ^
• ^1 ,
1

*
.
’’•.*•:• i.S • -. -

,1 •IV'i* * »-* • -I

'
* •
J ^ • • • -
'

,I» ' .«•_• i ,

•
•

M . ,
• ,1 •'

:*
 II

EXCHANGING WORDS:
ON THE ECONOMICS
AND ETHICS
OF VERBAL TRANSACTIONS
The Witter Bynner Lectures in Poetry
1977
I
 4
IN THE LINGUISTIC
MARKETPLACE

My pleasure in being invited to offer The Witter Bynner Lectures in

Poetry was pure, but attended by some bemusement. For none of
these lectures, I realized, would be in poetry, and you would have
reason to wonder, especially from the first of them, if they were ever
going to be even on poetry. I can assure you that they shall be,

though the route may appear, at times, a bit oblique. The obliquity
arises from the fact that the question that concerns me here is what
poetry is or, more precisely, what relation it bears to language and
what that relation implies about the experience and value of poetry;
and, for reasons I shall explain shortly, answering that question
requires paying a good bit of attention to what poetry isn't.

The most fundamental problem of literary theory appears to be

the location of its own subject, that is, “literature” or, as we
sometimes call it (already locked into battle with the problem),
“poetry in the broad sense.” The attempt to solve that problem is

not new; and it is not now, and never has been, merely an
“academic” question. But I shall not rehearse here the history of its

proposed solutions and what they have been taken to imply about
how poetry is —or should be —experienced and valued. I shall

instead, by way of reminding you of the nature of the problem and

also introducing my own views, speak of some more or less
contemporary solutions.

79
 EXCHANGING WORDS

You may recall one solution as the distinction between “denota-

tive” and “connotative” meanings, and the related one that dis-

tinguished between “referential” and “emotive” uses of language.

Those distinctions have, for good reason I think, been abandoned,
and nothing appears more dated than the once-confident contrast
between “referential statements” and “pseudo-statements.” One
might, however, sense the ghost or reincarnation of that contrast in
a number of more sophisticated formulations offered in recent years
by literary theorists and linguistic philosophers of various persua-
sions: the distinction drawn, for example, between “judgments” and
“quasi-judgments,” or the one between “reality language” and
“mimetic language”; or the contrasts made between texts with

“engaged-designative meanings” and those with “disengaged-ges-

tural meanings”; we have heard, and still hear, of the difference
between “illocutionary utterances” and “imitation illocutionary ut-

terances” and, most recently, have been presented with the difference
between “serious speech acts” and “nondeceptive pretended speech
acts.”'
Now, it must be granted that there is something vainglorious-
sounding in such a list when the entries are thus lined up. For one
notes a certain similarity of structure in both the categories them-
selves and the basis of the contrasts: on the one side, the referential,

the real, the engaged, the serious; on the other side, the pseudo,
the quasi, the imitated, the pretended. Moreover, the very repetitive-
ness of that structure may suggest obsession and defeat, as if one
hopeful champion after another had set out to capture a certain
beast, and had returned to say he had done so, only to find that the

beast had meanwhile escaped and was still at large. And, of course,
it might also be suspected that the reason he so repeatedly eludes
capture is not that he is so wily or protean but that he does not exist
at all, that the champions have not simply been inadequate to the
challenge but that their conviction that there was a beast to be
apprehended was a contagious delusion.
And so it has been suspected, and so it has been maintained. For
in addition to those who have, by one set of terms or another, sought
to formulate a distinction between literary and nonliterary language,
there are also those who insist that there is no distinction to be

80
 IN THE LINGUISTIC MARKETPLACE

drawn: that all statements are real and all speech-acts serious — or,

in the most recent and beguiling twist of skeptical monism, that no

statements are real because nothing is real.
Now, I should say immediately that my own views on this question
neither mediate nor synthesize the array of positions just outlined.
Moreover, I do not propose to reestablish some traditional or
commonsense view. There is no single traditional view: not only is

the question itself ancient, but so also is the array of conflicting and
unstable answers to it;and what we may think of as the common-
sense view is likely to be only some version of one or another of those
answers, recommended by early exposure and sustained by personal
or institutional inertia. As for the most recent attempts to draw or
erase distinctions, while I appreciate the force of some arguments
that might press one toward a position of monism (either sober
monism or nihilistic monism) I am nevertheless persuaded that, with
respect to the question at hand, distinctions can be made and will
continue to be made.
In that connection, we may return to the list that I presented a
moment ago, those sets of “real vs. quasi” or “serious vs. pre-
tended,” that seemed so familiar and perhaps so dubious. It should
be noted that, insofar as those contrastive categories have been
developed to distinguish literary from nonliterary discourse, they all

reflect a conviction, which I share, that the basis of the distinction

cannot be grounded on any characteristic difference of linguistic
form: in other words, that literary works do not always or necessarily
look or sound any different from nonliterary works, that the identity
of a text as “a work of literature” —or even as “a poem” —cannot
always or necessarily be established by any surface inspection or
even deep analysis of the text itself, and, in fact, that the same text

may change its identity not only in time but at any time under
different circumstances. The basis of the distinction is not linguistic
form, then, but something closer to linguistic /hwct/o/z, which (in the
formulations I have mentioned) is variously identified as purpose or
intention (what the author was up to) or as value, attitude, or
response (what the reader some combination of the
is up to), or as
two: how the author intended the reader to respond, or how the
reader assumes the author intended him to respond, or what each

81
 EXCHANGING WORDS

assumes and expects the other to do or to have done. All of which

brings us very rapidly to a central concept in contemporary literary
theory, namely convention.
These two concepts, function and convention, are, I believe,

indispensable for any discussion of the distinctive character of

literary discourse, but the mere invocation of them does not answer
all questions: it tells us only what sorts of questions we might now try

to answer. It will not do, moreover, to replace the question of the

distinctive forms of literary discourse with the question of its

distinctive functions without recognizing that the latter question

entails answering —or presupposing the answer to —another ques-

tion, namely the functions of nonliterary discourse. Indeed, it seems
clear that the long history of attempts to formulate a functional
definition of literature or poetry (that history, of course, extends
over more than two thousand years) reflects the long history of our
attempts to understand the nature and functions of language itself.

Moreover, the fact that every attempt to formulate such a definition

has proved imperfect or unstable does not reflect successive defeat or
essential futility but is, rather, the necessary and salutary con-
sequence of the fact that our understanding of language (among
other things) continues to enlarge and refine itself. An altogether
perfect, in the sense of eternally stable, theory of the relation of

literature to language is, then, not to be looked for. What we can do,
however, is recognize the problems in successive theories and seek to
adjust those theories accordingly —expecting, of course, to have our
own adjustments adjusted in turn.
To turn, then, to the adjustments to be proposed in the present
series of lectures, I would begin by pointing out that a persistent
problem among earlier theories of the distinctive functions of
literary discourse was a tendency to regard certain types of ut-
terances
— “objective” assertions of fact, for example, or scientific

statements — as the paradigm forms of nonliterary discourse. To

take scientific statements, however (or any specialized type of
utterance, especially one designed to reach its audience as inscrip-
tion or through publication), as exemplifying nonliterary discourse
at its purest, is surely to load, if not to beg, the question of what the
functions of language are. Moreover, even among those who do

82
 IN THE LINGUISTIC MARKETPLACE

recognize that nonliterary discourse embraces a wide variety of

utterances —vocal as well as inscribed, ephemeral as well as pre-
served, commands, promises, excuses and, of course, “emotive”
speech —there remains a tendency to describe language as es-

sentially a medium or channel through which a speaker encodes and

transmits information to a receiving and decoding listener. This
telegraphic model of communication is, however, unduly impov-
erished: not because it is so mechanical but because, unlike that
which it describes, it is a machine without a casement and without a
motor. For discourse is not the transmission of information between
two minds located at opposite ends of a channel but, rather, a series

of complex transactions between two persons who are located in a

rich world of objects, events, experiences, and motives, including
reasons for speaking and listening to each other.
It is important that we recognize not only that discourse is a
transaction, but also that it is an economic transaction, and one in

which the functions or value of an utterance differ significantly for

the speaker and the listener. It is especially important for literary

theory, because the dynamics of that transaction, and the value of
an utterance for both “speaker” and “listener” can, under certain
conditions, change radically; and we can begin to identify the
characteristic functions of language in poetry or literature only when
the nature of those conditions and the consequences of those
changes are appreciated.
In the first of these lectures, I shall be concerned with the
economics of that transaction, that is, the conventions —and, with
them, the claims, responsibilities, and liabilities —that govern the
behavior of speakers and listeners in the linguistic marketplace. In
the second lecture, I shall consider how language functions outside
that marketplace, why we sometimes license the unspeakable, how
we learn to do so, and what difference it makes. In the third lec-

ture, which is the most polemical of the three, I shall argue that
the difference is a crucial one for the experience and value of lit-

erary works, and crucial for what might be called the ethics of
interpretation.
Before I turn to the substance of this evening’s lecture, however, it

will be necessary to say a word about two terms that will figure

83
 EXCHANGING WORDS

throughout this series and are central to my observations here,

namely natural discourse and fictive discourse. Since I have de-
veloped the distinction between them at some length elsewhere, and
wish now to pursue some of its implications, I shall not take the time
here to repeat its precise derivation. I shall, however, present an
initial outline of it and trust that its more specific lineaments will
emerge in the course of the lectures themselves.
Briefly, then, by “natural discourse” I mean all utterances that
are performed as historical acts and taken as historical events. If one
asks what other kind of discourse there is, the answer is simple:
there is no other kind; natural discourse is discourse. There are,
however, verbal structures which constitute, in themselves, neither
historical acts nor historical events, but rather representations of

them and, as such, are understood not to be governed by the same

conventions that obtain for natural utterances: and these verbal
structures I refer to as fictive utterances. Thus, when the French
instructor illustrates a point of grammar by uttering the words Je
m’appelle Jacques, it is understood that these words represent, but
do not constitute, the verbal act of giving one’s name in French, and
do not have the force — make the claims, require the response, or
entail the responsibilities —that they would have a natural verbal in

transaction.
As you will have suspected, fictive discourse includes such verbal
structures as poems, novels, and the scripts of plays — in other
words, much we speak of as “literature” or “poetry in the broad
that
sense” — but clearly not only and not always. There are many
varieties of fictive discourse, serving many functions, from the verbal
playacting of children (“I’m a wicked witch, and I’m going to put
you in a dungeon”) to the exemplary sentences of linguists and
logicians {Je m’appelle Jacques, “The cat is on the mat”). Fictive
discourse can be handy or diverting, and it can also be constructed
for specifically aesthetic ends, that is, as verbal artworks. But just as
not all fictive discourse is literary, not everything that we might call
“literature” is constructed or regarded as The essays of
fictive.

Bacon and the letters of Keats would, at this moment, for certain
members of this culture, be regarded as “literature”; and although

84
 IN THE LINGUISTIC MARKETPLACE

both of them can be read as fictive discourse, neither was designed

to be read as such and neither is only read as such.
In short, “fictive discourse” is not a substitute label for “litera-
ture” or “fiction” or “poetry,” and I am not engaged in defining
these or any other traditional terms or concepts. Natural and fictive
discourse refer to functional categories which, I believe, are fun-
damental to our use and experience of language. I also believe
that the nature of the distinction between them illuminates a
number of problems associated with those traditional terms or
concepts, but precisely because it cuts across the distinctions on
which they have traditionally been based.
I shall have more to say about the nature of fictive discourse in the

second lecture, at which time many of the questions inevitably

provoked by this brief outline will, I trust, be answered. This
evening, however, I shall be concerned almost exclusively with
natural discourse, which —as I emphasized earlier — is a transaction
in which the functions of language differ significantly for speakers

and listeners. I shall consider each party to the transaction in turn,

and then put the parties together.

I. The Speaker: Utterances as Acts

We begin, then, with the speaker, with respect to whom utterances
are, to be sure, acts but which, like most other acts —such as
pushing doors open, lifting books off shelves, or hammering nails

into walls —we perform because they enable us to obtain and achieve
things that are of value to us. We perform verbal acts as well as other
acts, that is, in order to extend our control over a world that is not
naturally disposed to serve our interests. If utterances are recognized
as sharing this basic motive with all other acts, we must regard with
some suspicion the common view that the basic function or purpose
of speech for the speaker is to communicate information to the
listener; for, in the absence of further qualification, that view
implicitly attributes to our verbal actions a uniquely altruistic set of
motives.
This is not to deny that information is communicated by speech;
indeed, it is communicated by all our overt actions, at least to the

85
 EXCHANGING WORDS

extent that any action is potentially expressive, leaving marks and

presenting signs and evidence of the performer’s state and circum-
stances which, in turn, could be correctly inferred by a sufficiently
alert and interested observer. In other words, any overt act or action,
whether voluntary or involuntary, from sleeping to sneezing, from
shifting in one’s chair to rubbing a mosquito bite, will “say” things
about us to other people if they have reason to take note of them.
Thus, observing that my friend holds an unlit cigarette in his hand
and that, after reaching into all his pockets, he is scanning surfaces
and lifting objects, I may readily infer his state from his behavior

and, being amiably disposed toward him, may offer him a pack of
my own matches —without a word on the subject having been
exchanged between us.

Though all actions are potentially expressive, it is also true that

our verbal actions are highly specialized to be expressive, that is, to
make manifest to those who observe them the most subtle features of
our states and circumstances. That is not, however, because we have
a wholly disinterested desire to make such information about
ourselves or our world available to others. What is distinctive about
verbal actions is not the selflessness of our motives in performing
them but, on the contrary, their radical dependence on social
dynamics and, indeed, social economics, for their effect. For it is in

the very nature of verbal actions that they can have appropriate
consequences for the performer — that is, can serve his interests and
enable him to obtain and achieve what he desires —only to the extent
that they affect and control the behavior of other people.
Nonverbal actions —such as pushing doors open —character-
istically have consequences for the performer through their direct
physical impact on his environment. As we move through the world,
we move things (including our own bodies) around in it, performing
an almost continuous succession of actions — placing, pushing,
pulling, and so on — that have the effect of arranging our environ-
ments more to our liking. A verbal act, however, has only trivial
physical consequences. The marks it leaves on air or paper can raise
armies, topple empires, and secure our hearts’ desires only through
the effect those marks have on other people. The enormous dispro-
portion between the direct physical impact of a verbal act and its

86
 IN THE LINGUISTIC MARKETPLACE

ultimate force in modifying the speaker’s environment is related to

the point mentioned above: that whereas all overt acts are poten-
tially expressive, verbal acts are specialized to be so. In that respect,
verbal acts must be considered in relation not to acts in general but
to that particular class of acts which we may term symbolic.
Since this class has been given other names (for example, “com-
municative” or “semiotic”) and since all these terms —including ex-
pressive and verbal —are used in many and overlapping senses, a
full-dress definition will be useful at this point. In the sense of the

term used here, then, an act is symbolic when and to the extent that:

1. the relation between {a) the performance of an act of that

form and {b) the set of conditions in response to which it is
performed is governed by a set of regularities or system of
conventions, and
2. the act can have appropriate consequences for the per-
former only through on other members of a
its effects
community, specifically those who have learned those reg-
ularities or share those conventions and who are thus able to
infer {b) from (a), that is, who are able to infer the
conditions in response to which it was performed from the
fact and form of its occurrence.

It is clearly not the physical form of an act that gives it symbolic

force, for the“same” act can be performed either symbolically or
nonsymbolically, and a single act can have both symbolic and
nonsymbolic force at the same when a door is sticking
time. Thus,
and will not close properly otherwise, my slamming it shut will be a
nonsymbolic act; the slamming will be symbolic, however, when I
want a family member to appreciate the energy of my indignation at
his recent words. Or, when I press the button of a doorbell, my act is

nonsymbolic to the extent that its appropriate effect is to operate an

electrical mechanism, and symbolic to the extent that its appropriate
effect is to summon the inhabitants to the door. In each case, the act
is symbolic when it can have appropriate consequences for me only
through its effect on someone someone who interprets
else, it

correctly, inferring from it the motives and circumstances to which it

was a response.
Ringing doorbells and slamming doors shut are not verbal actions

87
 EXCHANGING WORDS

but, like utterances, as symbolic acts they can be interpreted

correctly only by those who have learned the regularities that govern
the relation of their forms to their motives or who share the
conventions that can give them symbolic force. The same is true of
numerous other forms of behavior that are not, strictly speaking,
verbal or linguistic, but which are nevertheless symbolic in this
sense: gestures such as winking or shrugging; the wearing of certain
items of apparel, such as crowns or tweed jackets; the display of
certain objects, such as badges or fine china — in short, the whole
array of actions that are increasingly recognized as having at least
some specialized expressive or semiotic functions. (I might note that
some of these actions and displays obviously have nonsymbolic as
well as symbolic functions. Moreover, the forms of some of them
obviously have their origins in biological, historical, or other non-
arhitrary connections to what they conventionally imply. That does
not, however, make them any less conventional when they occur as
symbolic acts, for to say that an act is governed by convention is not
to say that it has no basis in nature or history, but that its

effectiveness does not depend on that basis.)

Although the definition of symbolic action proposed here is not

altogether novel, I should like to acknowledge and emphasize some
aspects of it that are a bit unusual and will be of importance to our
general concerns here. First, due acknowledgment must be given to
the fact that two terms, meaning and intention, do not appear in it.

That is because one of them, intention, is superfluous here, its usual

function in distinguishing symbolic from nonsymbolic acts having
been covered here by specifying the distinctive way in which
symbolic acts have consequences for the performer. The other term,
meaning, does not appear because invoking it at this point would
obscure rather than clarify the distinctions I am attempting to draw
here.
I shall not, for the moment, speak further of intention, though I

shall return to it in the final lecture, in connection with authorial

intentions. I do want to say more about meaning, however. First, I

would point out that what is usually covered by the term meaning in

discussions of verbal or symbolic acts appears in this definition as

\{h)\ that is, “the conditions in response to which an act is

88
 IN THE LINGUISTIC MARKETPLACE

performed,” and which can be inferred from the act only by those
who have learned the set of regularities or system of conventions
governing the relation between acts of that form and some range of
conditions. Thus, in the example used earlier, to the extent that my
ringing the doorbell was a symbolic act, its “meaning” can be
crudely described as the set of conditions consisting of my wanting,
for some reason, to summon the inhabitants of the house to the
door. In short, themeaning of my act is equivalent to its causes or
why I performed it. The relevant set of regularities or conventions
here can be crudely described as follows: it is regularly the case and
conventionally understood that people ring doorbells when, for some
reason, they want to summon the inhabitants to the door. In our
culture, children learn that regularity or convention very early, but
they are not born knowing it and those who have not yet learned it do
not ring doorbells when they want to summon the inhabitants and
do not go to doors when doorbells are rung.
For our immediate purposes here, in considering the relation of

verbal acts to other acts, the translation of meaning just offered will
suffice, that is, the conditions in response to which an act is

performed. It is obviously not an adequate translation for all

purposes, however, and when we turn, later, to the listener, for

whom utterances are not acts but events, a richer and more
comprehensive translation will be both possible and necessary. In
the third lecture, I shall expand the translation even further in

connection with the meanings of literary works. We may now,

however, return to the speaker and to the characterization of
utterances as highly specialized acts which have consequences for
the performer in a quite distinctive manner.
The child discovers fairly early in infancy how extraordinary may
be the consequences of certain sounds he makes. Sometimes, when
he has done no more than emitted a syllable, powerful creatures will

enter his presence and do things that minister to his needs and
desires, things that he could not do himself. The child also learns

that by shaping and combining those sounds in certain ways, he can

have an increasingly strong and specific effect on those around him
and thus on his environment. Of course, the child would not learn to
speak as rapidly as he does if he were merely exposed to the verbal

89
 EXCHANGING WORDS

behavior of others or merely rewarded with praise for producing or

imitating linguistic sounds. It is, rather, the very specific services
and attentions he can elicit from other people by producing those
sounds that recommend their rapid acquisition and increasingly
“correct” formation and combination.^
The instrumental functions of language for the speaker are ob-
vious enough when he is issuing commands, making requests, and
asking questions (“Leave the keys on the table, please.” “What time
is it?”). It may be thought, however, that when he is making

statements about the external world or about himself (“The movie

starts at 7:30.” “My feet hurt.”), he is using language essentially to
inform and express, not to affect other people or to direct their
actions. We may still ask, however, why we should tell people things
they do not know, either about the world around us or about our
own ideas and feelings. Since we obviously know more than we ever
affirm and have more feelings and ideas than we ever express, the
question is why we affirm or express anything to particular listeners
or on particular occasions. When we consider the actual circum-
stances under which we speak, the evident answer is, I think, that
what we communicate, when we do so and to whom, are always
determined at least in part by the likelihood of advantage we thereby
derive for ourselves. Information is a commodity of some value,
and we do not normally squander it or donate it gratis to whoever
happens to be around. Rather, we exchange it for specific services,
or invest it for long-term yield. In other words, we are disposed to
provide our listeners with information when our doing so will dispose
them to act in ways that serve our own interests; and our enlighten-
ment of our fellow creatures is always tainted by ulterior motives.
It should be emphasized here, in view of the cynical sound of that
formulation, that the personal interests served by language need not
be grossly selfish and may range as widely in, say, mora/ terms as the
interests served by any of our acts. Thus, we often give information
to our friends and family members out of an independently moti-
vated concern for their welfare. Our “interests,” moreover, are not
always a matter of the immediate satisfaction of specific needs and
desires; and it is usually to our general advantage that our fellow

90
 IN THE LINGUISTIC MARKETPLACE

creatures, whose actions may ultimately affect us directly or affect

the world we inhabit, know what we would have them know about
ourselves and that world.
There are two further aspects to the value of language for the
speaker that require particular attention here, for they are closely
related to the functions and effects of fictive discourse. First, we may
consider that significant but rather obscure set of motives that
impels us, as we say, to “share our experiences” with other people.
Of course, as I suggested a moment ago, we may be inclined to
inform other people of our personal experiences, including our
sensations and feelings, when our doing so will secure their sym-
pathetic attitudes and perhaps practical assistance. Moreover, when
we exclaim to a companion, “Look at this interesting photo,” or
urge someone to taste what we are tasting or listen to what we are
hearing, or when we narrate personal anecdotes (“Say, do you want
to hear something funny ., awful ., wonderful
. .
.,” and so
. . . .

on) we are, no doubt, partly engaged in verbal gift giving, and thus in
,

seeking to please those whose pleasure and goodwill are of general

value to us. When all this is granted, however, there does seem to
remain another aspect to our motives when we attempt, sometimes
straining language to its furthest reaches, to make known to our
companions the most subtle features of our experiences. It might
seem, in fact, that there is an irreducible satisfaction to be gained
from the sense that there are people around us who not only occupy
the same gross physical environment as some we do, but who also to
extent share the same experiential environment, seeing what we see,
tasting what we taste, feeling what we feel.
The notion of a fundamental human desire for experiential com-
panionship is appealing, and it might even be thought that such a
need is innate in our psychological construction, evolving perhaps
through natural selection as a condition for our development as
social creatures and thus our enhanced ability to survive. On the
other hand, it may not be so irreducible a satisfaction after all; for,

by making our experiential environment available to others, we may

not only secure their sympathy, assistance, and gratitude, but also
their very important services as sources of cognitive feedback, that

91
 EXCHANGING WORDS

is, the responses of other people to our experiences can serve us as

an external testing, and either confirmation or qualification, of
our own most elusive perceptions.
The other point to be mentioned here is the fact that we are
almost continuously “saying” things to that most intimate, congenial
and attentive listener whom we carry within our own skins. Interior

speech or verbal thinking is evidently a derived capacity that could

not arise prior to, or independent of, one’s participation, as both
speaker and listener, in a linguistic community. But the functions
and value of language are significantly extended by our ability to
speak (and listen) to ourselves. For one thing, since one’s “self’ is

always part of one’s most immediate physical environment (as phi-

losophers with toothaches have had occasion to observe) and also, in
a sense, part of one’s social environment as well (we are always our
own companions) we have considerable interest in controlling and
modifying its behavior; and so we take great care to feed ourselves
information. Second, when we command, question, warn, and in-

struct the listener within, we can be, well, fairly confident of his
inclination to act appropriately in serving our interests. Also, per-
haps most importantly, by assuming the dual roles of speaker and
listener, we are able to internalize the feedback functions of lan-
guage mentioned just above, that is, to appraise our own perceptions
and responses, to expose, test, compare, and confirm or correct
them. It seems clear that our capacity for interior articulation also
increases our ability to discriminate and classify our experiences,
and therefore enables us to make more effective use of prior exper-
iences in solving problems, making plans, and generally dealing with
the world as we encounter it. Interior speech certainly comprises a
major part of the “activity” of all human beings, and it is likely that
our sense of personal identity and, to a large extent, “consciousness”
itself are products of our internalization of the linguistic trans-
action. But we must now consider the other side of that transaction,
or the functions of language for the listener.

II. The Listener: Utterances as Events

The who discovers that the sounds he makes can marvelously
child
extend his own powers also discovers that the sounds made by other

92
 IN THE LINGUISTIC MARKETPLACE

people may be of great interest and consequence to him. As he

learns more extensively and precisely those conventions that govern
the relation between the sounds people make and the circumstances
in which they make them, he is better able to predict and deal
effectively with his world, including the behavior of the people who
make those sounds.
Whereas an utterance can be properly regarded as an act from the
speaker’s point of view, it is, from the listener’s point of view, an
event, that is, an occurrence in the natural world that affects or
interestshim insofar as it provokes or demands an immediate re-
sponse from him or provides him with potentially valuable informa-
tion. I shall consider the peculiarly “provocative” aspects of verbal
events below. For the moment, concentrating on their function as
sources of information, we might observe (bending J. L. Austin’s
terms a bit) that while we are always using our own words “to do
things,” and that all verbal acts are therefore “performative” /or
the speaker, we are also always using the words of other people to
discover things, and that all verbal events are therefore to some
extent informative for the listener.
Any natural utterance does bear and can communicate informa-
tion, no matter what modal form, whether assertion,
its syntactic or
question, command, or counterfactual conditional, and whether
emotive, metaphysical, or verifiable in fact or in principle. Certainly
when someone asks a question or makes a vow, extends a greeting or
asserts the existence of a supreme being, the listener thereby learns

something about the world, even if it is only something about the

state of the speaker’s needs, attitudes, or beliefs; but of course these
may be things very useful for the listener to know.
We speak so that our listener will infer the conditions that caused
us to speak, specifically so that he will infer that which it is in our
interest that he know or believe about those conditions. And we heed
commands, promises, and
the speech of others, take notice of their
threats as well as their announcements and declarations, because we
can always expect to infer some of the conditions that caused those
utterances and therefore something perhaps important to know and
otherwise unknowable. It is true, of course, that the quantity of

information we receive from the speaker may be quite minimal.

93
 EXCHANGING WORDS

either because we already know it from other sources or because it is

of no inherent interest to us. In such cases, we may resist serving as a

listener; we grow bored, stop paying attention, change the subject,

or flee. Unless, however, our interests are themselves unusually
limited, we do tend to listen to what other people say, having learned
that the circumstances and motives that occasion their utterances
are part of a world that affects us, and that what they are saying
just may be “news.”^ Moreover, although we do not depend exclu-
sively on what people say to learn how they feel and what they
think —for all of their actions are sources of information concerning
their internal states, and we usually learn to check what they “do”
verbally against what they do otherwise —we nevertheless recognize
that verbal responses are sometimes the only perceptible responses a
speaker can make to certain of his personal states or experiences,
and that his speech is therefore the only available source of informa-
tion we have as to them.
It is important, at this point, that we return to the matter of
meaning, specifically to the relation and the distinction between the
meanings of symbolic events and the meanings of events generally. I

would suggest, first, that every event can be conceived of as the

center of a causal nexus, that is, a set of causes and consequences,
corollaries and entailments, both gross and subtle, that obtain at
every level of potential organization; and that total set can be
conceived of as the total meaning of the event. Obviously no mortal
creature can ever determine the total meaning of an event; but then,
no mortal creature is ever likely to be interested in doing so. What we
usually mean by the meaning of an event is not that totality but some
subset of it, that is, some selection from among all the conditions

to which the event ever could be thought to have causal connections;

and which subset we refer to, on a given occasion, as the meaning of
an event will depend on which of its connections are, on that occasion,
of interest to us and also the extent to which those causes, con-
sequences, corollaries, and so on then appear to be inferrable. In
other words, when we speak of, or inquire after, the meanings of an
event,we usually do so with reference to some subset of its total
meaning that is (a) of at least conceivable interest to someone, (b)

94
 IN THE LINGUISTIC MARKETPLACE

not self-evident and therefore problematic or worth remarking, and

(c) at least theoretically some basis.
inferrable on
Let me make these abstractions more concrete. If we encounter,
in a history textbook designed for high school students, a chapter
entitled “The Meaning of the French Revolution,’’ we will expect
something, but of course not everything, to be said about the origins
and implications of the French Revolution, that is, about its histor-
ical causes and historical consequences. Or, learning that the train
is we may remark, “That probably means there was a tie-up at
late,

30th Street,’’ thereby speculating on some of its causes, or “That

means we’ll have to catch a cab to the airport,’’ thereby observing
some of its consequences of interest to us. Jane says: “Look, John, a
falling star! What does it mean?’’ John replies: “Who knows?’’,
meaning that no one —or at least not John —can ascertain its causal
connections; or “Who cares?’’, meaning that no one — at least not
John — is interested in those connections; or “It doesn’t mean any-
thing,’’ meaning that no one — at least not John —can ascertain or is

interested in those connections.

So far, I have been speaking of the meanings of events generally
(the French Revolution, late trains, fallings stars). I should now like
to turn to the meanings of symbolic events, which, of course, in-

cludes natural utterances {utterances, must be emphasized, not

it

words', verbal events, not verbal forms), and the first point to be
made about verbal or symbolic events is that everything I have said
up to this point applies equally well to them. That is, we can
conceive of the total meaning of a symbolic event as all the condi-
tions that ever were or ever will be seen as having some causal
connection to it, some given time, anyone has
whether or not, at

determined or could determine them. And we can go on to observe

that, as with all other events, what we usually mean by the meaning

of a symbolic or verbal event is some subset of that totality, specifi-

cally a subset of conceivable interest to someone and theoretically

ascertainable on some basis. Certain additional characteristics, how-

ever, distinguish what we usually mean by the “meaning’’ of sym-
“
bolic or verbal events.
The first is fairly trivial, namely that, in view of the radically

95
 EXCHANGING WORDS

ambiguous nature of the term meaning, we usually take special

care, when we speak of the meaning of an utterance (or other
symbolic event), to distinguish antecedent from subsequent condi-
tions, often reserving the term meaning for the first and using some
other term, such as significance or import for the second (and also
for various noncausal relations). Thus, we may speak of the “mean-
ing” of someone’s remark as what caused him to say it and its
“significance” as what effects it will have. The distinction of usage
is by no means consistent, but because it is common in English and

also because it has some contemporary currency with respect to

works of literature, I will concentrate my observations in what
follows on meanings-as-causes, though the expansion to meanings-
as-consequences should be understood as possible in all cases. A
further expansion to meanings in the sense of certain noncausal
relations will be of particular interest to us in the third lecture.
The second more important. One of the ways in
characteristic is

which symbolic events differ from nonsymbolic events is that they

are always understood to have occurred by virtue of some personal
—
agency usually, though not necessarily, the agency of some human
being. Nonsymbolic events may be —
and many of them are the —
products of personal action, but symbolic events, by definition,
must be regarded as such. Because a symbolic event is always seen as
the product of someone’s act, we can conceive of its total meaning as
all the conditions that caused that act to be performed, whether or
not the performer or anyone else, at some given time, knows or can
know them. With respect to a natural utterance, its total meaning is

therefore everything that made the speaker speak it.

The third point to be made about the meanings of a symbolic
event is that, although many of those meanings can be inferred on
other bases, to the extent that it is symbolic, some of its meanings
can be inferred only by reference to a set of learned regularities or a
system of arbitrary conventions. And, of course, with respect to
verbal events, that set or system is everything that we mean —or
should mean — by a language, that is, all the regularities and arbi-
trary conventions that govern the relation between the occurrence of
verbal acts or events of some form and the range of conditions in

response to which they are performed.

%
 IN THE LINGUISTIC MARKETPLACE

The fourth and final characteristic of the meanings of symbolic

events follows from the fact that, by definition, at least two persons
are interested in it: the person by whose agency it occurred and the
person for whom it is an event:* in the case of natural utterances, the
speaker and listener respectively. And what follows is that which
particular subset of its meaning we refer to, on any particular
total
occasion, as its “meaning” will depend in part on whose interests are
of interest to us.
An example will be useful here. One evening, in a bar, after a
couple of drinks, John leans over and says to Jane, “You know, I

think you’re really a wonderful person.” We can say that the mean-
ing of John’s utterance, in the broadest sense, consists of the total
set of conditions, circumstantial and psychological, to which that
utterance was a response: in short, everything that made John say it,

whether or not he or anyone else then knows it. We can also specify
various subsets of that totality, each of which might be referred to as
'*the meaning” of his utterance.
Thus, we can say that its meaning consists of all the conditions
that are potentially inferrable by any interested listener —including
some eavesdropper —on the basis of that listener’s knowledge of
various matters. First, there is the listener’s knowledge of linguistic
convention, that is, the general regularities or conventions governing
the relation between the occurrence of utterances of that form and
the conditions under which they characteristically occur. Second,
there is the listener’s knowledge of the particular context of that
utterance, including the physical and social circumstances in which
it occurred, the character of John and his relationship to Jane, what
the conversation between them was like up to that point, and so
forth. Third, there is everything else the listener knows that might
enrich or constrain the inferences he draws on the basis of linguistic
convention and the context. All of the listener’s inferences made on
all these bases, however, will be less than the total set of conditions
that did obtain, and different from the inferences made by some
other listener whose knowledge was more or less extensive or subtle,
and whose interests in John’s utterance were different.
We can also speak of the meanings of John’s utterance as those
conditions that it was in John's interest that Jane infer from it.

97
 EXCHANGING WORDS

which will again be less than the total set though probably more, and
possibly less, than what would be implied by linguistic convention

(We may imagine John

alone. leaning over even closer and saying,
“You know what I mean?’’)
Finally, we can speak of the meaning of John’s utterance as all
those conditions that it was in Jane's interest that she infer, or all

those that she did in fact infer —again, in each case, less than total
set and probably different from the set that it was in John’s interest

that she infer. (We may imagine Jane as replying, “Yes, I do know
what you mean,’’ or “Yeah, I know what you mean.’’)

The implications of this analysis of meaning for our general con-

cerns here will become clear if we return to the discussion of the
value of natural discourse to the listener and particularly to its value
as a source of information. If, as I have been suggesting, we regard
the meanings of any historical event as the total set of conditions

that determined its occurrence, it follows that no natural utterance

is “meaningless,’’ for the very fact that it occurred is evidence that
something determined it and therefore that it has some potentially
inferrable meaning. Even the gibberish uttered by an idiot or a

psychotic has meaning in the sense that it has causes. Of course, to

the extent that linguistic convention does not operate for such a
speaker and that the determinants of his speech cannot thereby be
reliably inferred from its form, his utterance becomes, from the
point of view of the listener, in effect a nonverbal event. As such,
however, it still has meanings, and a listener might arrive at some of
them on the same basis as he would the meanings of any event:
through inferences based on his general knowledge and on his prior
experiences with more or less similar events in more or less similar
circumstances.
Indeed, it is important to recognize that the bases on which we
meanings of a verbal event are fundamentally the
“interpret’’ the
same as those on which we interpret or infer the meanings of any
event. In an upstairs room, I hear a variety of noises — shufflings,
slidings, bangings — and infer, on the basis of all my prior exper-
iences with (among other things) sounds of these sorts, what their
source is and what has caused them to occur: namely, that my

98
 IN THE LINGUISTIC MARKETPLACE

daughter, while preparing to go to bed, is moving around in her

room, opening and closing closet doors and bureau drawers. Later,
as another recognizable sort of clattering sound grows louder, I infer
that she is coming downstairs. A minute later, when she pops her
head intomy study and says, “Hey, my alarm’s broke —don’t forget
to wake me up at seven tomorrow,’’ once again I make inferences as
to the source and cause of those very subtly and precisely formed
sounds. It would be accurate to say that I have, in each case,
interpreted the meaning of a set of sounds on the basis of my prior
experiences with sounds of those sorts occurring in similar circum-
stances, and that, in each case, what I interpret them to mean will be
sharpened, enriched, and perhaps modified by other things I know
or believe, including things about those particular circumstances.
Having said what same about verbal and nonverbal events,
is the
we may now say what is different. First, we note that whereas the
actions my daughter performs in opening and closing closet doors
and walking down a staircase have appropriate consequences for her
that are quite independent of my inferences concerning the sounds
they incidentally produce, that is not the case for her verbal actions,
which will have an effect only through my proper interpretation of
them. Second, the relation between the form of the verbal sounds
she produced and the circumstances that produced them (her dis-
covery of a broken alarm, her concern lest she be late for school, her
need to secure my services, and so on) is not a matter of direct
physical consequences, as when a slammed door produces a slam-
ming sound, but is governed by a highly complex, largely arbitrary,
and extremely subtle set of social conventions.

As I have sought to emphasize, the conventionality of verbal

events does not distinguish them from other events with respect to
the process by which, or basis on which, one draws inferences from
them. One learns the system of rules that governs the relation
between verbal events and their causes in the same way as one learns

the nonarbitrary regularities that obtain in the world of nonverbal

events, that is, by generalizing from prior experiences with events of
that kind. The fact that the rules or regularities governing verbal
events are largely arbitrary and conventional does, however, have a
very important corollary for both the speaker and the listener: for

99
 EXCHANGING WORDS

an arbitrary convention is something that, by its nature, must be

shared if it is to be effective. Consequently, it is not enough to say
that the listener can draw inferences from a verbal event only with
reference to a system of conventions; we must add that the inferences
he draws will be correct only if the form of the speaker’s act has
been governed by that system. In other words, a listener can take the
form of a verbal event as a reliable index of its determinants only on
the assumption that those determinants are implied by it and are
expected to be inferred from it in accord with the relevant rules and
conventions. That assumption —that a speaker means, and expects
to be taken to mean, what he says — the fundamental convention
is

on which all natural verbal transactions are based.

This basic assumption of the linguistic marketplace operates very
much like the one shared by both buyers and sellers that paper or
metal currency has a specific economic value exchangeable, accord-
ing to a relatively stable system of equivalences, for valued goods
and services. When this assumption is violated by the speaker in
natural discourse —when he palms off counterfeit linguistic cur-
rency —we say that he is lying; and lying is a social transgression of
the first order precisely because it undermines the community’s
confidence in that verbal medium of exchange so basic to social
transactions. We do not have a simple term like lying for the
converse violation of that assumption by the listener, though it is

important to recognize its existence, and the situation is familiar

enough in everyday transactions. One may think, for example, of
listeners who refuse to credit what we say in the spirit in which we
have said it, or of those who persist in interpreting latently what we
offer patently, extracting the revelation of unflattering unconscious
motives from our innocent casual remarks, or, generally, of those
who take our utterances as implying something other than what
their forms would, according to linguistic convention, reflect.

Of course, this lying-in-reverse (or what I shall refer to as “false

listening’’) is only an extreme form of what all listeners do when they

interpret our verbal acts as nonsymbolic events, inferring the condi-
tions that caused us to perform them on some basis other than
linguistic convention. We cannot restrict the inferences people draw
from our acts only to those inferences which it is in our interest that

100
 IN THE LINGUISTIC MARKETPLACE

they draw or only to those implied by linguistic convention —and,

when we are listeners rather than speakers, we will do the same.
Nevertheless, if we were never taken to mean what we say, if all our
verbal actions were interpreted contrary to our interests and alto-
gether without reference to linguistic convention, the linguistic mar-
ketplace would rapidly disappear to the loss of all listeners as well as
all speakers.
we do, of course, distinguish lies from
In natural discourse,
erroneous statements and from linguistically improper or obscure
utterances. Similarly, we recognize that listeners can “mis-take” or
draw erroneous inferences from what we say, and that they may, for
various reasons, find our utterances incomprehensible. In all such
cases, however, the assumptions of the verbal transaction break
down because the speaker’s or listener’s knowledge of the world or
of linguistic convention itself is deficient, and such breakdowns are
accepted by the community as understandable accidents that do not
seriously damage the health of the verbal economy. As we shall see,
however, the assumptions of natural discourse may, also without
damage, be deliberately suspended or replaced by other conventions
operating for both “speaker” and “listener.” In such circumstances,
“counterfeit” —that is, fictive —utterances will, like stage money,
have a positive value, as will also the listener’s not crediting the
speaker’s words or not taking what he “says” as meaning what made
him “say” it.

I spoke earlier of the distinctively provocative aspects of language

for the listener. Like other events in our experience, verbal events
may not only provide information about the environment, but also
require responses from us. Moreover, because every linguistic trans-
action is also a social transaction, it inevitably reflects and partakes
of all the responsibilities and moral and psychological pressures that
can characterize our relations with other people. A natural utterance
constitutes, for the listener, not only an invitation and provocation,
but ultimately an obligation, to respond to the speaker. When we
“listen” to someone, as distinguished from merely noticing or over-
hearing what he says — in other words, when we identify ourselves as
his audience —we implicitly agree to make ourselves available to that
speaker as the instrument of his interests. We agree not only to hear

101
 EXCHANGING WORDS

but to heed his promises, excuses, questions, and commands — and

also, of course, his assertions. Most simply, but most significantly,

we agree to understand what he means, that is, to infer the motives

and circumstances that occasioned his utterance, and to act in
accord with the information thereby obtained, or at least, to take it

into account along with whatever else we know and however else we
are inclined to act.
We arecommonly conscious of the social risks that the speaker
incurs when he speaks (or writes, or publishes): his increased vulner-
ability to other people; the exposure of his feelings, attitudes, and
perhaps ignorance; the fact that he has made himself more know-
able to others and that their knowledge may not be used to his
advantage. Hamlet provides the classic image of a speaker attempt-
ing to avoid those dangers, particularly in the scene between the
Prince and his prying schoolmates: “. . . Why do you go about to

recover the wind of me, as if you would drive me into a toil? You . . .

would play upon me, you would seek to know my stops, you would
pluck out the heart of my mystery....” But there is also risk
involved for anyone who identifies himself as a listener, a fact
apparently well-appreciated by children, servants. Army recruits,

and others who, having been exploited as “captive audiences” in the

past, either learn to avoid verbal encounters or develop functional

deafness. Gertrude, in the closet scene, is often portrayed as putting

her hands over her ears: “O Hamlet, speak no more . . . Oh, speak
to me no more. These words like daggers enter in my ears.” Soph-
ocles’ Jocasta, sensing the imminence of “unspeakable” revelations,

urges Oedipus to ignore the messenger from Corinth, and flees

rather than listen. Oedipus himself, however, never flinches: When
the shepherd hesitates, “Ah, I am on the brink of dreadful speech,”
he replies, “And I of dreadful hearing. Yet I must hear.” We might
say that Oedipus was a heroic listener.
In spite of all these risks and liabilities, however, we are usually
willing to accept the implicit responsibilities of listening, and we
accept them primarily for two reasons (that is, aside from the
fundamental reason, namely the chance that we will thereby obtain
useful information): first, because the speaker himself — friend,
guest, child, employer — may be someone whose interests we are

102
 IN THE LINGUISTIC MARKETPLACE

independently disposed or constrained to serve; and second, because

it is to our general advantage, as members of a shared linguistic
community, to preserve the effectiveness of language /or the speaker.
For, as present and future listeners, we have a vested interest in our
fellow creatures’ continued disposition to speak and thus to provide
us with information; moreover, since we will ourselves be speakers
on other occasions and require the mediating services of those who
are our listeners, we have a vested interest in maintaining the
assumption of the listener’s availability and responsiveness.
A final point should be made here concerning the conventions
that govern natural verbal transactions, namely that those conven-
tions are not confined to the contents of dictionaries and grammars,
but include all those rules that govern the relation between the
things people say and the conditions under which they say them,
rules that we learn from our lifelong experiences with language in
use. “Linguistic convention’’ thus includes what we learn to be the
complementary obligations and responsibilities of speakers and lis-

teners. For example, if a speaker says, “Dinner is ready’’ or “Sorry,

I didn’t catch your name,’’ the listener will assume not only that
certain facts obtain about the state of dinner or a failure of appre-
hension, but also that the speaker’s affirming those facts at that
moment reflects other implicit features of his circumstances and
motives: presumably that he expects the listener to rise and enter the
dining room, or that he desires a name to be repeated.
Ordinarily, the listener will take the occurrence of an utterance as
a fairly reliable index of the speaker’s situation, beliefs, and motives
(with, of course, due or paranoid caution as to the possibility of lies

and errors, depending on his prior experiences with that speaker

and on what his sense of the immediate context dictates), and he will

act accordingly; that is, the listener will act on the assumption that
the speaker has taken responsibility not only for the “truth’’ and
“sincerity’’ of his utterance — its correspondence to what he believes
to be the case and to what he believes his motives are for saying

it —but also for what might be called the “spirit’’ of what he says.
Thus, having heard someone say, “I never want to see you again,’’ or
“Jane is wild about you,’’ and having acted accordingly (that is,

avoided the speaker, made a bold pass at Jane), the listener may

103
 EXCHANGING WORDS

justifiably feel that a trust had been betrayed if the speaker later
tells him, “Oh, I was just saying that. You shouldn’t have taken it

seriously.’’ Accused of making irresponsible or misleading state-

ments, however, the speaker may defend himself, pointing out that
he had obviously been “kidding’’ or “being ironic’’ or “speaking
figuratively,’’ and, perhaps with justice, accuse the listener of an
insensitivity to linguistic convention or contextual cues (“Anyone
else would have known what I meant,’’ or “Yes, I said that, but
that’s not the way I said it’’), or of being crudely “literal-minded.”

The listener’s problems and the speaker’s excuses here reveal the
significance and subtlety of certain aspects of the dynamics of
language that have not, until quite recently, been recognized in

formal linguistic theory^ —though we are, of course, familiar enough

with them from the hazards and disasters of our ordinary verbal
dealings. Of particular significance to our concerns here is the fact
that speakers and listeners depend on one another’s knowledge of,

and sensitivity to, various formal and contextual features that signal
how a given utterance is to be taken: with what force, in what
spirit — and, most important, whether or not it is to be taken as a
natural utterance at all. Among the signals a speaker may use
to indicate that what he says is not to be taken “seriously,” or
“straight,” or “literally, ” are intonation (such as exaggerated stress),
attendant gestures (such as a wink or a sidelong smile), and unidiom-
atic or personally uncharacteristic diction or phrasing. The speaker
may also simply rely on the listener’s appreciation of the total context
of the utterance, or his recognition of the fact that what the speaker
says is at odds with what the listener would otherwise expect, given
his knowledge of the speaker’s general situation, beliefs, and at-

titudes. In any case, such formal anomalies or contextual discrep-

ancies would, when properly interpreted, have the effect of qualifying
the conventions and assumptions that would otherwise be under-
stood as governing the relation between what was said and what was
so (or “meant”). If the signals are missed, so that the listener
mistakes the spirit and force of the utterance, his interpretation of

and response to it will be inappropriate; and although the con-

sequences of his error may be nothing more than social embarrass-
ment, the consequences are nevertheless quite real.^

104
 IN THE LINGUISTIC MARKETPLACE

At the beginning of this lecture, I suggested that the distinc-

tions commonly drawn between literary and nonliterary language on
the basis of their differing functions often suffer from an inade-
quate appreciation of the dynamics of discourse itself. The observa-
tions offered here hardly constitute, of course, a full description of
those dynamics. They have been designed, however, to emphasize
those aspects of natural discourse that are particularly significant
for understanding the characteristic functions and value of fictive
discourse, including poetry. It is important to recognize that natural
discourse is a transaction, that it operates through a formidable and
often fragile network of assumptions that extends beyond what we
sometimes think of as linguistic convention —that is, the rules gov-
erning semantic and syntactic propriety —and that this transaction

involves not only complementary advantages but also complementary

responsibilities and risks, for both speaker and listener.

To understand the ethics of verbal transactions, we must appre-

ciate its economics. That means we must appreciate that language is

action, both speaking it and also listening to it, and that it always
operates through the use and control of other people. Like all other
economic markets, the linguistic market is never an altogether free
one: it can be rigged, and it can be floated with counterfeit currency;
the exchanges are not always conducted between those on an other-
wise equal footing and, when attended by the machinery of political
power, the control exerted and services exacted through language
can be literally killing. We sometimes speak of language as a game—
in the sense that, like games, it is a form of behavior governed by
rules. But, in that sense, so is trade, so is politics, and so is war. If
they are games, they are all games that are played for real.

Since I have not, in this first lecture “in poetry” so much as

mentioned a poem. I’ll conclude this evening by reciting a brief one
that touches on this last point and suggests quite incisively the limits
of metaphors of play. It was written by Sir Walter Raleigh, a man
who played every game and fought every war that the period offered:

What is our life? a play of passion;

Our mirth, the music of division;
Our mothers’ wombs the tiring houses be
Where we are dressed for this short comedy;

105
 EXCHANGING WORDS

Heav’n the judicious sharp spectator is,

That sits and marks still who doth act amiss;
"
Our graves, that hide us from the searching sun.
Are drawn curtains when the play is done.
like
Thus march we, playing, to our latest rest:
—
Only we die in earnest that’s no jest.

106
 LICENSING THE UNSPEAKABLE

In the last lecture, natural discourse was described as an economic

transaction between speaker and listener, typically an exchange of
information for services. This evening, I shall be concerned with
how we move out of the linguistic marketplace, why we do so, and
what difference it makes. First, I shall consider some of the
conditions that take us out, conditions that arise from the very
nature of verbal transactions. For as we have seen, the conventions
that make those transactions possible also control and constrain us
as both speakers and listeners; and we would, and do, sometimes
escape them. In the second part of the lecture, I shall be concerned
with the potential aesthetic and ludic functions of language that are
liberated by the removal of discourse from the marketplace, and
shall consider briefly the relation of verbal art and verbal play. In

the third section, I shall focus on the verbal licentiousness of

children and consider how, even as they are learning to enter the
marketplace, they also learn to find its exit gates. At the conclusion
of the lecture, I shall have a few words to say about epistemological
skepticism and what I have referred to as nihilistic monism.

I. Constraints and Conventions

You will recall the observation made last time that being a listener
always involves a liability of some sort. We do not always want to

107
 EXCHANGING WORDS

know the truth; we certainly don’t always want to hear it. Though
usually under circumstances less extreme than those suggested
earlier —the verbal daggers against which Gertrude shields her ears,
the “unspeakable” revelations that Jocasta truly would rather die
than hear —the risks and burdens of serving as someone’s audience
are nevertheless familiar. One may think, for example, of the times
one has found oneself the audience of a friend’s irrepressible and
unabridgeable report of a recent dream, or a child’s rambling
narrative of the plot of a television drama, or a lecture on the
chemical composition of packaged foods delivered by a recently
enlightened colleague.The excruciating boredom we may suffer
under such circumstances arises from the fact that while general
social propriety or our particular relationship to the speaker prevents
us from simply taking flight or tuning out, the amount of interesting
information we receive from such utterances is quite minimal —and
thus out of proportion to, or insufficient payment for, the time and
attention we have been constrained to donate.
We may also consider these poor market conditions from the
point of view of the other party to the transaction, that is, the
speaker, who may have something to sell for which there is no
immediate buyer. Or, to drop the analogy for a moment, it appears
that our impulse to speak is often strong even when there is no
willing or otherwise appropriate listener.
No one has ever said everything he knew or felt. That is not,

however, because we are secretive or, in the ordinary sense, in-

articulate, but rather because the opportunities for speech are quite
limited compared to the motives for it. Because of the various
crucial and unique ways in which language serves us as speakers, we
probably have a generalized tendency to respond verbally to almost
all our experiences. Yet it is obvious that we do not talk continuously,
at least not publicly. Of course, we can always indulge that tendency
by talking to ourselves (either covertly, as in verbal thinking, or
overtly, as in muttering or writing diaries); but it is also obvious that
there are severe limits to the satisfactions to be gained thereby, and
also limits to the services that can be provided by the listener- within.

Of all the not strictly physical afflictions to which we may fall

victim through such misfortunes as exile, imprisonment, illness, and

108
 LICENSING THE UNSPEAKABLE

old age, perhaps the one most acutely felt is loneliness —which
means, among other things, the sheer unavailability of listeners.

When there is “no one to talk to” (note that we far more rarely
complain of there being “no one to listen to”), we are manifoldly
deprived; for, lacking listeners, we lack not only the opportunity to
affect others instrumentally, to secure their services in ministering
to our physical needs and desires, but we also lack their services in
providing what I referred to last time as cognitive feedback, that is,

the occasion they offer for us to verbalize and thus to integrate,

discriminate, appreciate, and indeed experience our own otherwise
elusive perceptions. For we often come into possession of an idea or
feeling precisely through the expression of it, the process required to
make it intelligible to others. Prisoners, exiles, and suburban
mothers often suffer, for what appears to be good reason, from the
fear that they are becoming unresponsive, undiscriminating, torpid,
or indeed not quite sane.
But we may hunger for an audience even in the midst of a feast of
companions. For what we have to say may be too extensive, subtle,
or specialized to be of interest to, or even comprehensible by, those
around us. Or we may, in their company, be unable to select and
organize our words in such a way as to control our listeners’
attention or to insure that they do listen to what we are saying. To a
great extent, the institution of “literature” supplies these needs by
extending the domain of the linguistic marketplace. That is, we may
inscribewhat we cannot speak and put it up for general sale as, for
example, an essay, article, or memoir. “Literature” in this sense,
however, that is, as inscribed natural discourse, is still subject to the
economics of the marketplace. Although no specific service or
immediate response may be asked of the reader (who may be remote
in time as well as otherwise from the author, and therefore under no
pressure of personal or social obligation to him), it is still understood
that the reader will take the inscribed composition as a direct
reflection of the writer’s sentiments, beliefs, and circumstances, and
will find interest and value in it primarily as such.
and publication allow one to “speak” utterances that,
Inscription
because of essentially practical difficulties, would be otherwise
unspeakable. Some utterances, however, are even more radically

109
 EXCHANGING WORDS

unspeakable. Most obviously, we can neither speak nor write what

—
we would say when the “saying” of it in either form would make —
us vulnerable, when the revelation of our sentiments, situation, or
even knowledge would expose us to contempt, hostility, or danger.
Also, we may be moved to address a listener who, for some reason,
cannot or will not heed us. Thus, we cannot really “speak” to dead
friends, estranged lovers, sleeping babies, nightingales, or Grecian
urns, although there may, in a sense, be things we wanted to say to
them. And, of course, even when we do not require a particular
listener, there are few occasions on which we can give unlimited
expression to all of our attitudes, feelings, and recognitions. Not
even lovers or psychoanalysts can be enlisted as continuous
audiences. Finally, even granted endlessly attentive lovers and
analysts, our public voices are always dependent on our public roles,

qualified as well as constrained by our relationships to our listen-

ers. Consequently, there are always sentiments we could express,
knowledge we could display, ways we could speak possible —
utterances, in other words —for which there are no occasions, no
audiences, or no available style. The conventions and assumptions
of fictive discourse, however, may allow us to define the occasion,
enlist the audience, and create or discover the necessary style.

Fictive discourse allows us to speak the unspeakable —but only if

we agree not to say it. This catch could be put in the terms of our
economic analogy as the need for the seller either to lower his price
or to throw in additional goods or otherwise valuable premiums.
Indeed, the characteristic qualities of verbal art are often accounted
for in such terms: meter or rhyme being seen as “additional goods”
(or, as Wordsworth implies in the Preface to Lyrical Ballads,
“superadded pleasure”), and enchanting stories and stylistic colora-

tion being seen as superior (if slightly deceptive) packaging (as in the
classical “sugar-pill” theory of didactic poetry and in some modern
psychoanalytic theories). But these accounts miss the essential
nature of the accommodation made in fictive discourse. For the
“speaker” does not really cheapen, dress up, or disguise his goods
when he offers them as poetry: he transforms them into something
entirely different and, in effect, removes them from the common
verbal marketplace to another kind of arena. What makes it worth

no
 LICENSING THE UNSPEAKABLE

the reader’s while to heed the historically unspeakable utterance

depends on the understanding that he shares with the “speaker”
that it is not, in fact, historically spoken: that, insofar as it is being
offered as fictive discourse, the reader and author have entered a
special relationship, one that is governed by assumptions, claims,
and responsibilities quite different from those that obtain between
the speaker and listener of a natural utterance.
Produced in the theaters of language or displayed in its galleries,
fictive discourse is not subject to the economics of the linguistic
marketplace. Thus, the poet can use language and his audience can
respond to it without the constraints that would otherwise shape and
confine the behavior of each of them. Though this may appear to be
a tendentiousway of describing “poetic license,” its implications are
considerably more far-reaching than what the clich6 suggests. For
the licensing that I am speaking of here extends to the audience as
well as to the poet, and it involves not merely formal or even
thematic features of the utterance, but quite fundamental aspects of
the linguistic transaction itself.

As we have seen, the basic assumption of natural verbal transac-

tions, shared by both parties, is that the speaker means what he says
and that the listener will take him to mean what he says. It is
precisely the suspension of that assumption that defines fictive
discourse. It is not, of course, that the poet is understood to be lying,

but rather that he is understood not to be saying at all. The poet is

not a speaker addressing a listener, but one who composes a verbal

structure that represents a natural utterance. The poem may
represent the poet himself addressing a dead friend or estranged
lover, but the poet, as a historical creature, is not engaged in the
historical act of addressing them — ^just as the language instructor,
whose name really may be Jacques, is not engaged in the historical
act of giving his name when, in illustrating the use of a verb, he
represents that act by presenting the words Je m 'appelle Jacques.
Nor is the poet really “addressing” the audience who may at any
time read his poem, except in the metaphoric (and for that reason
often confusing) sense that any artist
— — painter, sculptor, or com-
poser “addresses” those whose interest and pleasure his work is

designed to secure. Whatever communication may be taken to mean

111
 EXCHANGING WORDS

in regard to the dynamics of art, it is not the same as what it means

in regard to the dynamics of natural discourse.
A representation, we note, does not have the same force as an
event in nature, does not demand and also does not preclude the
same response. We do not have the same attitude or reaction to the
sculpted figure of a nude as we would to a naked person standing or
reclining before us. To mention only a trivial —or perhaps not so
trivial — difference, we cannot stare at persons (clothed or not) as we
can at sculptures or portraits without feeling, as we might say, a bit

voyeuristic. The actual presence of someone both requires certain

social acknowledgments and also constrains our actions in various

ways.
The exceptions here are particularly instructive. We can stare at

naked persons —such as live models in a drawing class or actors on a

stage —when they are in a sense “in our presence” but yet under-
stood as not occupying the same ontological space as we do. In these
circumstances, the conventions that require or preclude certain
responses to other people are understood to be suspended. The
model or actor is, under these conditions, no longer another person,
but the representation of one: in short, a fictive person.
With respect to language, the point may be made concrete if we
consider a situation in which the special assumptions governing
fictive discourse are strained or come into conflict with those
governing natural discourse. A close friend tells us that he has just
composed a poem, and asks us to read it; in the “poem,” we dis-
cover what appear to be allusions to his poor health or domestic
troubles. Under such circumstances, something of a moral dilemma
is likely to arise. On the one hand, in view of our concern for the
writer and the apparent revelations of his griefs or problems, we
may feel that it would be insensitive, an evasion of moral re-

sponsibility or of simple civility, to direct our comments to

“aesthetic” matters: to observe that the metaphor in the third line is

muddy, or that the rhythm toward the end is flat, or even that the
poem is a splendid achievement. On the other hand, we may hesitate
to take explicit notice of the personal allusions, hesitate to offer
sympathy or press for further details, as we would had our friend

112
 LICENSING THE UNSPEAKABLE

Spoken to us (or, of course, written us) about his misfortunes; for to

do so would seem somehow indelicate, presumptuous, a violation of
his privacy or, more significantly, a violation of an understanding
that such a response is not expected and not appropriate.
We may give the ambiguities of the situation a further turn by
supposing that we have good reason to suspect, on independent
grounds, that our friend is in fact using the poem primarily as a way
of informing us of his situation. In that case, although our
sympathies may be genuinely aroused by the substance of the
revelations, and we may appreciate the motives that led to this
indirect means of revealing them, we are still likely to feel to a
certain extent exploited. For, by presenting us with what he calls a
“poem” and yet intimating that he wants us to take it as a reflection
of his immediate circumstances, he has tried to have things both
ways — or, more specifically, tried to say something while evading
the responsibilities of having done so. Our sense of exploitation here
is composed of a number of elements which we would find difficult
to express to him, not only because they are subtle but because they
involve those economic aspects of social transactions that we
commonly hesitate to invoke or acknowledge explicitly.

For one thing, we are always inclined to resent, or experience as in

some sense unjust, situations in which someone “lets us know”

things without actually saying them: the barely articulated or
subsequently denied mutter of annoyance; the pointed inflection
that gives to an otherwise innocuous statement an edge of sarcasm;
faint sighs, “loaded” silences, and a variety of “expressive” though
not conventionalized acts and gestures, such as slamming a door or
fluttering one’s eyelashes. These paralinguistic and nonlinguistic
forms of communication may, of course, be patently symbolic acts
or be used patently to supplement or qualify natural utterances, and
thus be incorporated into what is “said.” The situation becomes
uncomfortable, however, when acknowledgment of what they com-
municate and of the very fact that theycommunicate things, is
evaded: “But I didn’t say I minded your coming home late.” “But I
didn’t ask for help with the dishes.” What is evaded in such cases
is the acknowledgment of an intention to affect or control the other

113
 EXCHANGING WORDS

person, and also acknowledgment of the usual understanding that

the “listener” will and should act in accord with the information
thus delivered. When a speaker does tell us his state or situation
verbally, he implicitly acknowledges his interest in or need for our
consequent response —and thus also acknowledges that he is to that
extent in our debt and that we, in turn, have a counterclaim on
him for a return of services. But the “speaker” who “lets us know”
things without saying them may also be attempting to evade that
acknowledgment, to obtain our services gratis under cover of not
really asking for them.
The various moral derelictions described here obtain for all
speakers who want to have it both ways. There is a special kind of
culpability in the double-dealing of our friend, however, for to
the extent that he indicates that he want the “poem” to affect us
primarily as a revelation of his immediate circumstances, he has
abused the conventions of fictive discourse and thereby contributed
to their undermining. If we see his duplicity as a form of lying,

we must also recognize the possibility of its converse on the part

of the audience, just as we recognized the possibility of reverse
lying or “false listening” in natural discourse. There is, here, an
interesting double reversal that can be spelled out as follows: The
basic assumption of natural verbal transactions is that the speaker
means what he says and will be taken to mean what he says in accord
with linguistic convention. Thus, in natural discourse, the lying
speaker is one who says what he does not mean, and the false

listener is one who interprets an utterance as meaning what the

speaker does not say. But in fictive discourse, where that basic
assumption is understood to be suspended, the “liar” and his coun-
terpart “false listener” are those who do not acknowledge the sus-
pension or who do not permit it to be in force. It would be as if

the French instructor, whose real name

Seymour and was ad-
is

dressed that way by a student, were, for his own amusement, to

comment curtly, “But told you my name was Jacques”; or as if
I

the student, out of malice or waggishness, were to address him as

Jacques, saying, “But that’s what you said your name was.”
The implications of this double reversal will be explored in the

114
 LICENSING THE UNSPEAKABLE

third lecture, when we are concerned with the ethics of interpreta-

tion. For the moment, however, we may consider more generally
how the dilemmas and difficulties discussed a moment ago illustrate

the disparity between the poet’s relation to his reader and what
would otherwise be a speaker’s relation to his listener.
Any utterance occurs in and presumes upon a social context,
and, in natural discourse, that context is constituted in large part
by a relationship between speaker and listener that has been es-

tablished on prior or what might be called external grounds. In

other words, we speak to each other as friend to friend, parent to
child, teacher to student, master to servant, or, it may be, one
twentieth-century intellectual to another. The nature of the ut-
terance and also of the listener’s response to it are always to some
degree determined by the nature of that relationship: the claims and
responsibilities that exist by virtue of it. The relationship between
poet and reader, however, is established on internal grounds, that
is, it is defined with respect to the particular poem, making neither
concessions to nor acknowledgments of whatever external relation
the poet may have to the reader. (This reader or audience, of
course, is to be distinguished from whatever listener — mistress or
nightingale — may be “addressed” in the fictive utterance itself and
also from any audience addressed by the poem offered as a natural
utterance.) The poet and reader do have a “relationship” and cer-
tain social responsibilities toward one another, but these derive from
cultural assumptions regarding not personal relationships but the
nature and value of artworks. In “personal” terms, there need be
no relationship between the artist and his audience. Although this

description may suggest a morally grim and socially chilly state of

affairs, the “impersonality” that characterizes the relation between
poet and reader reflects neither evasion nor abdication of social
responsibility but, on the contrary, a most punctilious respect for

it. Like the impersonality that obtains between painter and model,
or between actor and audience, it is a highly moral state of affairs
that serves each of them and preserves the conventions by virtue of
which art can exist and have value. We shall return to that point also
in the third lecture, where we shall be concerned with claims made

115
 EXCHANGING WORDS

tor the ethical basis of certain constraints on literary interpretation,

claims that depend, in part, on the obliteration of the sorts of
distinctions just made.

II.Verbal Art and Cognitive Play

I should like now to pursue further my earlier observation that
“poetic license” extends to the listener or reader as well as to the
poet. Put another way, we might say that, for both parties, the
removal of the poem from the linguistic marketplace — or, more
generally, the disengagement of any verbal structure from its func-
tions in a marketplace transaction — liberates its potentiality to serve
other functions which, in turn, replace its market value. Verbal
artworks are structures presumably designed to be thus disengaged
and designed accordingly to fulfill other functions. Any verbal struc-
ture, however, has the potentiality to fulfill multiple functions:
just as the metal disc that once served Roman citizens as a coin
of the realm may now be exhibited in a museum case as an ar-

cheological “find” or, duly set by a jeweler, be displayed as a body

adornment, so a text that once was (and still could be) linguistic
currency may serve other functions and thereby acquire a differ-

ent sort of value. The discovery and demonstration of the archeo-

logical value of verbal structures is a characteristic occupation of
the biographer or historian, who, in making “documents” of texts

that originally functioned as natural utterances (or, for that matter,

were originally designed as poems), thus actualizes another of their
potential functions —though not, it might be noted, an originally
intended one.' Similarly, compositions not originally designed to
serve as verbal artworks may be “found” or “set” to do so, that

is, discovered and exhibited to be the occasions for certain percep-

tual and cognitive activities that appear to be associated with what
we sometimes speak of as “aesthetic experience.”^ It will be per-
tinent to our concerns here to consider briefly some aspects of those
activities and their relation to the potential functions of Active
discourse.
As creatures who must know our world in order to survive in it,

we are engaged in almost continuous cognitive activity, scanning

and exploring the environment, searching out principles of regu-

116
 LICENSING THE UNSPEAKABLE

and patterns of conformity, classifying and reclassifying phe-

larity

nomena, testing present against past experiences, examining and

adjusting the categories of our knowledge. Indeed, we are prob-
ably creatures who have survived because of our epistemic hunger
and irritability, our itch to know and our capacity to learn. The
possession of information is obviously a condition for survival, but
so also is the continuous pursuit of it, and it would seem that we are
conditioned by our individual histories —and, in a sense, “condi-

tioned” by the history of our kind — not only to seek and value
knowledge but to find fundamentally gratifying the very process of
coming to know, that is, the activity of learning itself or the sheer
exercise of our cognitive faculties.^ As Aristotle observed, we are
creatures for whom learning is the liveliest pleasure.
We are not commonly conscious, however, of the independent
charms of epistemic activity. For one thing, most of our “learning
experiences” occur in a context of practical concerns: our explora-
tions of the environment are usually directed and constrained by
more or less immediate needs and interests, and the value of our
learning is usually recognized and measured by its instrumental
consequences. Getting off the train in a strange we search for city,

an exit, a restaurant, a taxi, an “information” booth. Picking up a

newspaper, we scan the headlines for “news” that might interest us,
focus on an article, flip the pages quickly to find the continua-
tion of its text. When we are clearly intent on cognitive ends rather
than cognitive processes, the processes themselves are likely to be
attended by anxiety or irritation rather than that lively pleasure of
which Aristotle speaks. Indeed, it would seem that the pleasures of
maximal when the pressure to know is minimal and the
learning are
consequences of knowing are not of immediate concern.
The potential gratifications of learning are, moreover, enhanced
or diminished by various conditions. Certain objects and events
yield discoveries more readily than others to our search for pat-

terns and principles. In a situation where structural relations are elu-

sive or fragmentary, where we encounter mostly clutter and noise or

discontinuity, unpredictability and irregularity, where the objects of
perception resist classification and offer no readily graspable prin-
ciples of order, either internal or in relation to prior experiences.

117
 EXCHANGING WORDS

learning will be continually frustrated. Under such circumstances,

we will feel radical discomfort, disorientation, alienation, and per-
haps panic as the categories by which we organize experience prove
inadequate and our sense of control over the environment slips

from us. It is not the unknown but the unknowable that we dread.
On the other hand, if the patterns or principles are totally man-

ifest and predictable or conform exactly to prior experiences, the

conditions for learning are not really favorable, for there is nothing
to learn. In the language of information theory, one approaches
total redundancy and zero information; in more colloquial terms,
there is no news, and the situation is experienced as boring, bland,
unchallenging, uninteresting.
It would seem, then, that the optimal conditions for engaging in

gratifying cognitive activity are characterized by, among other

things, a combination — either a balance or a particular ratio —of
novelty and familiarity, repetition and variation, conformity and
disparity, redundancy and information. Learning is most graciously
invited by a situation that appears tosome extent unknown but that
promises knowability. Moreover, the process of “coming to know”
seems to be most rewarding when there is what might be called a
gradient of reinforcement, that is, when each step into the blur or
darkness yields increasing clarity and structure, where intuitions of
order may be tested and either confirmed or disproved, and where
local patterns and simple conformant elements are seen as co-
herent in relation both to one another and to more comprehensive
or hierarchic principles of structure. The various characteristics
described here, it should be emphasized, are not independently mea-
surable “properties” or objectively describable “features” of a situa-
tion, and can be specified only relative to some perceiver. What is

“familiar” and “redundant” to me may be novel and rich in infor-

mation to someone else, depending on our individual personal his-

tories, including our cultural experiences. Moreover, because there

are individual, and to some extent innate, differences in our capa-
cities to process information, a degree of novelty and unpredictability
that is agreeably “interesting” to one perceiver may be experienced
by another as intractably chaotic. Nevertheless, there will always be
conditions that are more likely than others to invite and reward the

118
 LICENSING THE UNSPEAKABLE

very process of learning, that is, learning to no immediate end, or

what we might call “cognitive play.”
While the conditions described here are most conspicuously ex-
emplified by situations or activities that we do, in fact, refer to as

“games” or “play” (for example, doing jigsaw puzzles, solving

cryptograms, playing solitaire), they are not confined to them; and,
though the world is not naturally a cognitive playground, we do find,

frame, or fashion gymnasia for the mind in the midst of its market-
places and battlegrounds. Those conditions are perhaps just as
conspicuously exemplified by what we refer to as “works of art” and
by other objects and events that become the occasion for “aesthetic
experiences,” and also (though perhaps less conspicuously) by as-
pects of certain situations and activities which, while generally con-
sidered serious or elevated business, may nevertheless be recognized
as potentially “playful” or “aesthetic”, for example, religious cere-
monies, sexual engagements, doing philosophy.
These observations can obviously be extended to our experiences
with language. To survive in the verbal universe and to profit from
the economics of the linguistic marketplace, we must scan and
explore utterances as well as other visual and auditory structures
and we must attend to the forms of verbal events at least as closely as
to those of any other events. We characteristically obtain informa-
tion from linguistic sights and sounds, however, in relation to the

symbol systems by virtue of which they do bear most of their infor-

mation, and “cognitive activity” with respect to verbal structures

consists primarily of inferring their meanings from their forms in
accord with the conventions of those systems; it consists, in other

words, of interpreting them.

Two points must be added if we are to appreciate the potential
“aesthetic” functions of fictive discourse. The first is that, with

respect to verbal structures, the process of interpretation is typically

energized and rewarded by the economics of the linguistic market-
place and both directed and constrained by contextual considera-
tions: the possible instrumental value of the information thereby to

be obtained and the possible responses thereby required of or pro-

voked from us. In interpreting natural discourse as such, then, our
cognitive activities are always in the service of more or less immediate

119
 EXCHANGING WORDS

cognitive ends. We do not, therefore, usually give ourselves over to

the independent pleasures and interests of merely interpreting a
natural utterance, nor do we usually infer or ascribe meanings that
are irrelevant to its presumed context of origin or to the context of
the transaction itself.

The second point is that although interpretation is the cognitive

activity characteristically evoked by verbal structures, it is not the
only one that is possible. For those structures are not only potential
sources of conventionally coded information; they are also physical
phenomena —shapes and sounds— and, as such, have formal prop-
erties that may become the subject of more or less independent
perceptual and cognitive interest. That potential interest, however,
is usually not actualized and is, commonly surpressed (“ed-
in fact,

ited out” by speakers, “not noticed” by listeners) when we are

engaged in marketplace transactions.
As we know, it sometimes happens that for some reason — for
example, boredom, social alienation, high-spiritedness or inebria-
tion —we will be “distanced” from the marketplace and find our-
selves listening to (or reading) natural discourse with the sort of

desultory concentration, “disinterested interest,” that we associate

with “aesthetic experience,” attending to features of utterances that
would otherwise distract us from the business of verbal transactions:
for example, the intonational contours and rhythms of a dialogue;
chance rhymes, parallelisms, and other formal patterns; evocative
allusions, submerged metaphors, potential puns. Thus decontex-
tualized, the ongoing natural discourse in which we have no “in-
terest” will become quite interesting in another way — or, of course,

amusing or absurd. Or, if we happen to find ourselves in the

congenial company of others who are similarly distanced and dis-
posed, we may engage in forms of verbal behavior and exchanges of
words that have only limited market value: elaborate metaphors,
chains of puns, jokes, verbal impersonations or self-parodies, teasing
or mock arguments, in all of which we play not only with verbal
forms and meanings but also with the very dynamics of natural
verbal transactions and the linguistic and social conventions that
govern them.
But of course we need not wait for such distancing conditions

120
 LICENSING THE UNSPEAKABLE

simply to “happen” or to arise casually; we can create them our-

selves or seek out those already created by others. Poems, novels,
plays, and other “works of literature” are, among other things,
verbal structures that have been designed or discovered to constitute
precisely such conditions, to invite and reward the sorts of activities I

have been describing as “cognitive play”: here, the exploration of

the formal and symbolic properties of language, the contextually
unrestricted interpretation of verbal structures —that is, the playing
out of their potential “meanings” —and the playing with or playing
at the conventions of linguistic transactions.

Clearly works of literature may and do serve other functions as

well, and their potential function as occasions for cognitive play is

not always actualized, not always dominant, and therefore not the
only source of their interest and value. It is, moreover, proper and
possible to distinguish between, on the one hand, verbal artworks
and, on the other, those structures and forms of behavior that are
more readily seen as verbal toys and games. While the distinction is

not, I believe, an absolute one and can never be drawn with preci-
sion, certain relative differences can be suggested.
First, to the extent that a verbal artwork exhibits greater struc-
tural complexity or “relational richness” than a verbal toy, the range
and variety of cognitive activities it evokes will be more extensive.'*

Compared to the solving of a cryptogram, the reading of a poem is

an exercise that both requires and gives scope to more multileveled

information processing and multiple patterning. Also, to the extent
that its formal and thematic elements can be seen not only to enter
into conformant and hierarchical relations with one another but also
to enter into more relations with other — prior and subsequent
experiences, the poem will be more amenable to repatterning. As
new principles of organization are recognized, different internal
relationships and relations to different external experiences are per-

ceived and new “meanings” are manifested. The rereading of a

poem is thus more likely to be rewarding, and rewarding in new
ways, than the resolving of a crossword puzzle, and the jokes and
riddles exchanged community are likely to become stale for its
in a

members more quickly than its songs and stories. Conversely, to the
extent that puzzles, jokes, and riddles do exhibit a relatively high

121
 EXCHANGING WORDS

degree of structural complexity and relational activities, we are more

likely to think of them as works of art or literature (cf. Finnegan's
Wake, which may be, among other things, the most complicated
verbal toy ever devised).
Second, our consciousness of the artificer is likely to be of parti-
cular significance in our experience of a verbal artwork —as distin-

guished here not only from such verbal games as jokes and riddles
(whose original “authors” are usually unknown and, in any case, of
less interest in this respect than their presenters in the context of

verbal play), but also from such “casual” occasions for the aesthetic
experience of language as were described above. As I suggested
earlier, among the conditions that invite our cognitive explorations
of a structure is the assumption of an ultimate reward to those
activities, that is, the promise of its ultimate knowability; and no
matter how’ well that promise is actually made good, it will obviously
be strengthened if the structure has presumably been designed to be
thus explored and “known.” Consequently, although we can engage
as deeply with anonymous poems as with those whose authors are
vivid figures for us, the presumption and/or projection of some

author, and thus the presumption of the “designed” design of a

work, is usually crucial to the nature of our cognitive engagement
with it; and, among the cognitive processes that constitute our
experience of a poem is the process of inferring or hypothesizing
various aspects of that design. It should be noted here that in an
orally transmitted folkloric work, such as a ballad or a tale, the
successive acts of artifice of its various transmitters constitute a
process of gradual “design” which is, in this respect, equivalent to
the various acts of selection and revision that enter into the
composition of a text by an individual author. What is crucial is that
we may assume, in our experience of each, that the structure of the
work has been shaped by either its anticipated or actual effective-
ness in serving an aesthetic function for some audience.^
This brings us to the final and perhaps most significant dis-
tinction, namely our assumption that a verbal artwork, unlike a
verbal toy or game, has been designed to be experienced as fictive
discourse’, for the presumption of its amenability to “interpreta-

122
 LICENSING THE UNSPEAKABLE

tion” as such, that is, as the representation of an utterance, will be a

major source of its interest and value. Whereas the riddle, crypto-
gram, or crossword puzzle characteristically promises knowability in
the form of a “solution” to be found, the joke in the form of a
“point” to be gotten, the poem promises knowability primarily in
the form of “meanings” to be inferred, and the process of inferring
those meanings is central to our cognitive engagement with it. Thus,
that engagement is brought to a conclusion (comparable to the
solving of a puzzle or getting the point of a joke) not only when we
perceive the principles that allow thework to be experienced as a
coherent formal structure, but also when we have the sense that we
have “understood what it means”: when, for example, we have been
able to construe the identity of a “speaker” implied by the repre-
sented utterance and a plausible context of motives and circum-
stances that would account for his speaking thus, and/or when we
have constructed various stable hypotheses of thematic design. The
latter might range from our presumptive identification of various
local and specific “meanings” (for example, an allusion to Queen
Elizabeth, a pun on the word //e, an allegory of Christian re-
demption) to our inference of more general “meanings” attitudes, —
—
sentiments, values, and so on that the work appears intended to
imply or exemplify (for example, that war is hell, that the love of
money is the root of all evil, that death should be faced with
courage).
As the examples here suggest, the material with which we play in

our engagement with a poem is often sober stuff, but the concept of
cognitive play entails neither triviality nor shallowness. When we
watch the Prince of Denmark play with a jester’s skull or hear the
dying Mercutio play with “grave words,” we are entertained, though
we know that death is no jest. Indeed, part of what entertains us in
these plays is the occasion, scope, and license they can provide for
our entertaining thoughts of death: not only for imagining Hamlet’s
thoughts or for inferring Shakespeare’s thoughts, but also for
playing out our own thoughts of death —a playing out for which
there would otherwise be no particular occasion or which, when
there is one (for example, the death of a friend), would be colored

123
 EXCHANGING WORDS

and controlled by the very particularity of the occasion and yet not
"controlled”
*
— fed, focused, and shaped —by the playwright’s
counter-play.
I shall develop this last point further in the next lecture, where I

shall suggest that the "transaction” between the poet and his reader

is indeed more a form of gamesmanship than one of commerce, and

shall have a few more words to say about the nature of the cognitive
activities described here. I should like to conclude the present
discussion of those activities by pointing out that the various
inferences, identifications, and hypotheses alluded to above are not
to be confused with the overt and more or less formal "analyses,”
"readings,” or "interpretations” of poems that are produced as
academic exercises from either side of the professorial desk. Such
exercises reflect an essentially arbitrary arresting, segmentation,
and articulation of a process that is itself always fluid and dynamic
and that may be quite "inarticulate.” Moreover, that process is in

certain respects endless. Although it is typically "concluded” when

we have the sense that we understand what the poem means, we
never "finally” understand a poem. To be sure, we never finally
understand any structure or event, that is, never can (and rarely seek
to) ascertain what was described in the first lecture as its "total

meaning.” Our efforts to understand a natural utterance are usually

satisfied and terminated when we have inferred enough of its

meanings to serve our own interests and those of the speaker in the

context of the verbal transaction. To the extent, however, that a

poem is removed from the marketplace and is not currency in such a
transaction, there are no such interests to be served and there is thus
no corresponding (or "natural”) termination to the activity of
inferring meanings from it. We can be repeatedly motivated,
therefore, to "come to know” a poem, and repeatedly gratified by
the sense that we have understood what it means: not understood
"at last” what it "really” means, but understood again what it might
mean.

Children at the Gates of the Marketplace

III.

We may now turn to the third aspect of verbal licentiousness to be

considered here, namely the linguistic behavior of children and.

124
 LICENSING THE UNSPEAKABLE

thereby, not only the beginnings of verbal play and art but also the
functional origins of the distinction between natural and fictive
discourse.
In the first lecture, I alluded briefly to how children, from the
time of infancy, are introduced to verbal economics, how they learn
to extend their control over their environment though their own
verbal actions, and also how they learn the significance — in all

senses —of the verbal actions of others. It must be added, however,

that from our earliest years, indeed from the very beginnings of our
lives as verbal creatures, we find sources of fascination and pleasure
in language that are largely (though probably not altogether)
independent of its instrumental functions. Infants apparently find
inherently gratifying their own imitation of the linguistic sounds
they hear, and young children seem universally to take delight not
only in storytelling, songs, and verbal impersonation, but also in
producing and hearing rhythmic or repeated linguistic sounds and,
later, in puns, riddles, paradoxes, and other kinds of formal and
thematic wordplay. One may speculate on the developmental origins
and functions of such wordplay, noting, for example, that it allows
children to rehearse and exercise those skills required for the use of
language in the marketplace: perceptual and cognitive, as well as
motor, skills. It seems clear, however, that our interest and pleasure
in exploring the properties of language, and also in exploring our
own responses to it, do not disappear when we come of age, and that
we continue throughout our lives to make excursions from the
marketplaces to the playgrounds of language.
I shall not take the time here to develop the observation,
increasingly supported in recent years, that play is a very serious
matter for children. It has not, however, always been thought so,

and its relation to other serious matters is still often misunderstood

and underestimated. We may recall here the words of Thomas
Gradgrind, the arch-utilitarian of Dickens’ novel Hard Times, on
the occasion of his visit to the Coketown elementary school:

Now, what I want is Facts.Teach these boys and girls

nothing but Facts. . . . You can only form the minds of
reasoning animals upon Facts; nothing else will ever be of
service to them. This is the principle on which I bring up

125
 EXCHANGING WORDS

my own children, and that is the principle on which I bring

up these children. Stick to the Facts, sir!

There remain parents and educators who would banish fairy-

tales from the nursery and schoolroom as pernicious sources of

misinformation or as encouragements to what is taken to be the
child’s innate tendency to confuse fact and fiction. It is likely,

however, that these ideological descendants of Gradgrind not only

underestimate the child’s ability to make the relevant distinctions,
but also fail to recognize the nature and strength of the pressures
that oblige him to do so. Children learn to discriminate at a quite
early age — almost, indeed, from the time they begin to engage in
anything we would call verbal behavior —between, on the one hand,
things that are “said” by and to them and, on the other hand, such
things as nursery rhymes, songs, and verbal play. They do so,
moreover, not on the basis of explicit instructions or distinctive
labels but, rather, on the basis of their own differential experiences
with each, that is, with the relatively distinctive forms and the
absolutely distinctive consequences of natural and fictive discourse.
Fictive verbal structures are commonly marked off from verbal
acts and events by various features, both contextual and internal,

that operate for the child as identifying cues or signals. We need not
specifically instruct a child to regard as fictive any utterance that
begins “Once upon a time ...” and, indeed, traditional formulas of
that sort are probably less crucial than is sometimes supposed, and
not necessarily more effective than purely contextual signals of
fictiveness.

We do not commonly recite a rhyme or begin telling a tale in the

middle of a conversation. Fictive structures are not distributed

haphazardly among natural verbal events but tend to occur in a

marked, framed or self-enclosed context, a “storytime” or

specially
“playtime” that may be named as such, or identified in other ways.
Moreover, even when they are imbedded in natural discourse, the

imbedding itself is signaled. Intonational markers are common, as

are distinctive body motions, and both are readily picked up by
children. We begin a story, “There was once a wise king of England,
and he was named King Arthur ...” with a style of delivery that is

appreciably different from the “matter-of-fact” manner in which

126
 LICENSING THE UNSPEAKABLE

we would say, perhaps in reply to a child’s question, “Elizabeth the

First? Oh, she was queen of England a long time ago, before the
Pilgrims came to America. You remember who the Pilgrims were,
don’t you?’’ We commonly employ a distinctive vocal register when
we narrate fictions or read them aloud; also the rise and fall of pitch
and loudness are often exaggerated, the tempo is often slower or at
least more studied, and even if the material is not itself metrical, we

may be inclined to phrase more rhythmically.

Fictionality, in the sense of historical falsehood, is not, for the
child, the defining quality of a “story.’’ If he says, “Tell me a
story,’’ he may continue with “Tell me the one about Red Riding
Hood’’ or “Tell me the one about how you got stuck in the sticky tar

and Grandma had to wash you with gasoline.’’ What was initially a

family anecdote may be spoken of as a “story’’ even though the child

knows and believes that the events really happened. Should the
parent reply, “But you know what happened with the sticky tar; I

told you about it just yesterday,’’ the child may insist, “No, no, you
have to tell it.’’ The parent who himself appreciates that implicit
distinction between dicere and narrare will construct the story
accordingly, and will deliver it in proper storytelling fashion, “not
speaking’’ (as Sir Philip Sidney put it) “table-talk fashion, words as
they chanceably from the mouth,’’ but employing the intona-
fall

tional markers and distinctive body motions mentioned a moment

ago. Leaning back or forward, lifting or lowering his head, closing
his eyes for a moment or looking into the distance, the teller will
create the sense of a special space, his own entrance into a world

apart from the immediate one: a bardic stage or magic circle, in

short, a playground set off from the verbal marketplace. These
distinctive tones and gestures may be themselves conventional,
learned (or unconsciously imitated) from storytellers we have heard;
they may more or less naturally from the
also be universal, arising
contexts and functions of storytelling. The origin of those features,
however, is not at issue. The point is that, however they arise and
whatever other functions they serve for the storyteller and his
audience, such features also operate as signals of fictiveness that
inform the listener how that discourse is (and is not) to be taken.
In the example just given, the family anecdote becomes a tale, a

127
 EXCHANGING WORD5

fictive “telling,” relished now as much for its structure, rhythms,

predictable details, and bits of quaint dialogue as it was initially,

perhaps", for the amusing image of a once-young parent in a sorry

predicament. An episode of domestic history has acquired the status

of family lore somewhat as in a nonliterate community, a fragment
of presumably true national history becomes part of the culture’s
lore as told by one or more local bards. In both, the thematic core of
the narration may have been first heard or presented as part of a
natural utterance in a context of reportage', but, in becoming a
“tale,” its identity and interest become independent of that context.
The f actuality of the subject does not compromise the fictiveness of
the tale, for it is not the events told that are fictive but the telling of
them. That telling is set apart from reports of past events and from
such allusions to them as may occur in natural discourse. As we have
seen, the listener’s interest in a natural utterance lies primarily in
the information it bears, what it reveals of the world or of the
speaker. For this reason, the repetition of such reports is felt to be

redundant. someone who has just heard a radio broadcast bursts

If

into the room and begins, “Say, did you know there was a terrible
earthquake in Los Angeles?” we are not likely to reply, “Yes, I read
it in the paper this morning, but tell it to me again,” though we

might say, “Yes, I know. Is there any more news about it?” “News,”
or information, is what the listener values in natural discourse. It is
obviously not what he seeks or expects from the twice-told tale.
Children are no less sensitive to redundancy than are adults, and
they are often more openly and even brutally disdainful of it. (“Oh,
I know that already. You don’t have to tell me again.”) But they are
also more openly eager for the special pleasures of “redundancy” in
fiction, and the refrain “Tell it again, tell it again,” appears to be
universal among them.
I mentioned earlier the contextual and paralinguistic features of a
recitation that set it apart from natural discourse for the child.
There are, of course, also internal or stylistic features that he will

learn to associate with fictive discourse: most obviously, rhyme,

meter, and other systematic repetitions, but also archaic diction and
otherwise unidiomatic expressions, and various opening, closing,
and transitional formulas. Like the intonational markers mentioned

128
 LICENSING THE UNSPEAKABLE

earlier, these stylistic features need not have been specifically or

exclusively designed to function as signals of fictiveness, and most
of them have other functions that are related to the valued effects of
verbal artworks. Moreover, stylistic features of this sort do not
define fictive discourse and are not altogether reliable signals of
fictiveness: they occur in household maxims and cautionary homilies
that are said to the child while they may not occur in naturalistic and
crudely didactic storybooks that are nevertheless sto/^books (for
example, John and Mary Visit the Airport). However, when they do
occur, they inform the child that he is probably being presented with
something other than natural discourse.
It is, in fact, this negative definition of fictive discourse that is

most significant for the child, that is, the fact that certain verbal
structures are recognized as not being natural utterances. As I
observed earlier, the child learns to distinguish between natural and
fictive discourse on the basis of his differential experiences with
each. These differential experiences, however, relate not merely to
the more or less distinctive features of fictive utterances, features
that signal them as different from natural utterances, but much
more importantly, to the fact that that difference itself has conse-
quences. If there were no differential consequences, there would be
no reason for the child to learn the distinctive features, and he
would not learn them! In effect, he learns the difference between
natural and fictive discourse because it makes a difference, and his

learning the difference consists basically of his learning the dif-

ference it makes.’
The child’s earliest experiences with the adult use of fictive
discourse usually involve verbal games, songs and lullabies, and
nursery rhymes; later his experiences with fictive discourse involve
not only the stories and tales told to him by adults, but also his own
verbal make-believe, in which he may induce adults and other
children to join. At the very simplest level, these events and activities
introduce the child to a special class of verbal structures that do not
function for him, either as speaker or listener, in the same way as do
the natural utterances he otherwise hears and produces. “Patty-
cake” does not occur in the presence of cakes and ovens; “This little

piggy” is not a piggy, and does not go to market or eat roast beef.

129
 EXCHANGING WORDS

Adults will allow the child to berate them as witches and revile them
as monsters in terms and tones for which he would otherwise be
scolded/ Parents will accept, without murmur, his offers of invisible
tea and his invitations to invisible castles, and will make allusions,

without explanation, to such exotic objects as porridge and tuffets,

and such anomalous events as talking animals and metamorphosed
pumpkins.
Children learn, then, (1) that certain linguistic structures can and
do occur outside the normal context of verbal transactions, and (2)

that the linguistic conventions and social assumptions of the verbal

community do not hold for those structures: specifically, those
conventions that govern the relation of a verbal form to the
speaker’s motives and circumstances, and those assumptions that
govern how the listener is expected to interpret and otherwise
respond to an utterance.
When we consider how the child learns to recognize and respond
to fictive discourse, we must remember that he is still in the process
of learning the forms and effects of natural discourse. Even those
language theorists who are persuaded of the extensiveness of innate
linguistic capacities do not maintain that a child enters the world
with a complete functional mastery of language. From his own
incohate vocalizing and the initially undiscriminated sounds he
hears, the child must learn to produce effective utterances and to
interpret the verbal behavior of others in accord with the conven-
tions of his linguistic community. For the adult, who has mastered
the language, and whose verbal acts and experiences consist largely
of natural utterances, fictive discourse is discriminated as a “special
class’’ of linguistic structures. For the child, however, natural
discourse /ise// gradually emerges as a “special class’’ of linguistic
structures. For even after he has acquired a substantial repertory of
well-formed words, phrases, and syntactic structures, a good deal of
the child’s verbal behavior still consists of what might be called “ex-
tranatural’’ discourse, that is, the production of verbal structures
that are not governed by the conventions of either natural or fictive
discourse: for example, his repetition, to no one in particular,
of “funny’’ words, phrases, or phonemes; the semi-melodic and
minimally verbal “songs’’ that he makes up himself; and long streams

130
 LICENSING THE UNSPEAKABLE

of essentially private —that is, nonsocial though overt —speech and

verbal impersonation. A child may engage for hours in solitary
activity, or even group play with other children, accompanied by
almost continuous extranatural “speech” of this kind —which is,

in effect, interrupted by occasional natural utterances: either his

own (“Now you be the princess”) or those of other people (“A little

less noise, please”).

My point here is that the child actually learns two sets of linguistic
convention, that which governs natural and that which governs
fictive discourse: the distinctive features, distinctive occasions, and
distinctive consequences of each.
Since the child’s earliest verbal behavior is “extranatural,” and
since a large portion of it continues to be such for a few years, it may
be tempting to think of this as evidence for the notion that fictive

discourse (or “poetry”) is more natural than natural discourse, and

to extrapolate the notion to a theory of the poetic origins of speech in
primeval man. It must be emphasized, however, that the earliest

extranatural speech of children is nonnatural or “poetic” only in the

the sense that its forms and occurences imperfectly reflect the

conventions of the linguistic community and are not yet governed

appropriately by the assumptions of natural verbal transactions. It

is nonconventional by default, not by design. Thus, we should not

confuse the child’s verbal play for the poet’s verbal art any more
than we should confuse his delightfully improper generalizations
(for example, his pointing to the sun and saying “Lamp?”) with the
poet’s artful metaphors (for example, alluding to the sun as “the
bright eye of day”). As for extrapolating the child’s verbal play to

a theory of the origins of language, it must be emphasized that the

verbal behavior of every child we observe arises within a functioning

linguistic community and is shaped by its practices from the very

beginning, whereas the “original” speech of primeval man pre-
supposes, by definition, the absence of any such community.
Although extranatural discourse is the first to occur in the child’s
own behavior, it nevertheless follows from the prior existence of
natural discourse in the social world around him.
Granting these distinctions, however, we may still suppose that
our tendency, as children, to explore the properties of verbal forms

131
 EXCHANGING WORDS

and verbal behavior is related to the pleasure we take, as adults, in

creating and experiencing verbal art. Wherever we find a linguistic
marketplace, we also find, just outside its gates, and adjacent to
each other, a playground and a theater; wherever we exchange
words, we also play with them and play —that is, represent —the
exchange of them. Fictive discourse appears to be as universal as
natural discourse and we may suppose it to be as ancient as language
itself.

In concluding this evening, I should like to return very briefly to

Thomas Gradgrind. Gradgrind, you recall, maintained that the
Coketown children should be taught “nothing but facts.” For
philosophers and others who observe, on epistemological grounds,
that there are no facts, that the universe is a structure of fictions,
Gradgrind’s absurdity lies in the fact that his facts are themselves
fictions. Since nothing is really real, they argue, there can be
nothing that is unreal — or nothing that is not unreal. I would reply
that it doesn’t quite follow. That is, while it may be that there are no
ultimate realities, it does not follow that there are no distinctions
among unrealities. The distinction is often in the kinds of conse-
quences they have. We are at some peril if we fail to distinguish the
unreality of dreams from the unreality of waking experiences or
unreal representations of tigers from unreal tigers —a peril that I

shall not call “real” but, as Sir Walter Raleigh put it, at least “in
earnest.” It may be argued that all natural utterances are themselves
fictive. But the distinction between natural and fictive discourse

does not thereby collapse: it simply, so to speak, moves over one

ontological notch.

132
 6
THE ETHICS OF
INTERPRETATION

The observations presented in the two previous lectures can be

brought to bear on certain questions of literary theory. My topic this
evening is their bearing on questions of literary interpretation. I

shall, in a moment, be reading to you a passage from a recently

published essay by Professor E. D. Hirsch. The volume in which it

appears is The Aims of Interpretation; the essay is entitled

entitled
“Three Dimensions of Hermeneutics.” Although I shall be taking
issue with a number of points in it, my argument is not with
Professor Hirsch specifically, but rather with the general and not
uncommon practice —which the passage handily exemplifies —of
enjoining certain forms of literary interpretation and interdicting
others through an appeal to the ethical basis of verbal transactions.
The passage reads as follows:

Therefore, let me state what I consider to be a funda-

mental ethical maxim for interpretation,
a maxim that
claims no privileged sanction from metaphysics or analysis,
but only from general ethical tenets, generally shared.
Unless there is a powerful overriding value in disregarding
an author's intentions {i.e., original meaning), we who
interpret as a vocation should not disregard it. [Hirsch’s
italics] . . . When we simply use an author’s words for our
own purposes without respecting his intention, we trans-

133
 EXCHANGING WORDS

gress what Charles Stevenson in another context called “the

ethics of language,” just aswe transgress ethical norms
when we use another person merely for our own ends. Kant
held it moral action that men should
to be a foundation of
be conceived as ends in themselves, and not as instruments
of other men. This imperative is transferable to the words of
men because speech an extension and expression of men
is

in the social domain, and also because when we fail to

conjoin a man’s intentions to his words we lose the soul of

speech, which is to convey meaning and to understand what
is intended to be conveyed.
I am not impressed with the view that this ethical impera-
tive of speech, to which we all submit in ordinary discourse,
is not applicable to written speech or, in particular, to
literary texts. No literary theorist from Coleridge to the
present has succeeded in formulating a viable distinction
between the nature of ordinary written speech and the
nature of literary written speech. For reasons I shall not
pause to detail in this place, 1 believe the distinction can
never be successfully formulated, and the futility of attempt-
ing the distinction will come to be generally recognized.
Moreover, if it is seen that there is no viable distinction
between “literature” and other classifications of written
speech, it will also come to be recognized that the ethics of
language hold good in all uses of language, oral and written,
in poetry as well as in philosophy. All are ethically governed
by the intentions of the author.*

One may readily sympathize with the sort of impulse that pro-
duces so strong a statement as this, especially if one recalls various
ignorant, inept, perverse, or preposterous interpretations of literary
works; and one may appreciate the effort to devise an argument
exposing the impropriety of such interpretations in principle. The
grounds of this particular argument are nevertheless weak, for they
consist primarily of begging the fundamental questions: Are there
no relevant distinctions to be made between a man’s “speech” and
any literary work? Do the “ethical imperatives” that we acknowl-
edge in ordinary discourse hold good in “all uses of language”?
fhose who, like Hirsch, argue for the reader’s —or professional
interpreter’s — moral obligation to discover the “original meaning”

134
 THE ETHICS OF INTERPRETATION

of a work assume, assert, and insist that the answer to these questions
is The burden of the previous two
yes. lectures has obviously been to
answer them otherwise.
Before turning to the substance of the disagreement, I should like
to consider briefly the precise nature of what Hirsch alludes to as
“the soul of speech,’’ that is, “to convey meaning and to understand
what is intended to be conveyed.” So defined, this is clearly quite
close to what I have been describing as the fundamental assumption
governing natural verbal transactions, namely that the speaker
means, and will be taken to mean, what he says in accord with
linguistic convention. What follows, however, is that all uses of
language are ethically governed not by the author’s intentions but by
the conventions of the linguistic community. In his formulation and
discussion of this assumption, Hirsch introduces a gratuitous and
ultimately obscurantist note of piety into the dynamics of an
essentially pragmatic situation. It is true enough that when we fail to
respect that assumption, “we lose the soul of speech”; like all

categorical imperatives, however, those that govern verbal trans-

actions derive their force from the fact that their transgression would
make the universe we share with our fellow creatures a poorer place
for all of us, in this case poorer for the loss of the benefits of
language. We “submit” to these assumptions not because we
conceive men as “ends in themselves” rather than as “instruments
of other men,” but because to violate them would be to reduce the
effectiveness of language for all speakers and listeners, among whom
we ourselves number. Indeed, it seems clear that to use language at

all is inevitably to use other people as instruments, though there is

no reason to regard speaking and listening as therefore any more

unethical than, say, grabbing hold of someone’s arm on an icy road
without first asking his permission or consulting his interests. We
graciously serve others as instruments because, in doing so, we
contribute to, and preserve, the network of social assumptions that
permits us to use them as instruments; and it is just as well for all of

us that the vitality and effectiveness of language do not depend on

anything so vagrant as the moral rectitude of any of us.
We may now turn to the more fundamental issues, which are (1)

whether the assumptions (or, if one prefers, imperatives) governing

135
 EXCHANGING WORDS

natural verbal transactions do “hold good in all uses of languages,

oral as well as written, in poetry as well as in philosophy,’’ (2) whether
there ate no relevant and viable distinctions to be made in this

regard, and (3) if they do not uniformly hold good, and there are
distinctions to be made, what the ethics of interpretation are.
One would agree, of course, that the assumptions governing
“ordinary speech’’ extend to ''written speech’’ and to “literary texts"
when, by these latter terms, we refer to inscribed natural utterances,
which include not only notes to the milkman and love letters but also
essays, memoirs, and works of philosophy — and thus many compo-
sitions that we do consider “literature.” What I have been suggest-
ing, however, is that there is a class of verbal structures to which the
assumptions governing ordinary speech cannot be extended because
the suspension of those assumptions is precisely what defines that
class. For Hirsch to maintain his resounding generalization, “the
ethics of language hold good in all uses of language,” he would have
to deny the existence of any verbal composition of the sort referred to
here as fictive discourse. This would require him not merely to deny
that such “literary texts” as poems, novels, and playscripts are
distinguishable from “written speech,” but also to deny that such
uses of language as the presentation of sample sentences by linguists
and logicians, the verbal play of children, the citation of proverbs,
and the recitation of stories are, any of them, in any relevant way
distinguishable from “ordinary speech.”
That there is a relevant distinction to be drawn is obscured, in
Hirsch’s argument, by the shifting and otherwise problematic senses
in which he uses certain expressions: among them, “ordinary
discourse,” which seems in some instances to mean specifically vocal
or conversational speech but, in other instances, to mean any form
of nonliterary discourse; and also “original meaning,” “authorial
meaning,” and “the intentions of the author,” all of which seem to
be equivalent at some points, but not all points.
In the article under discussion, Hirsch is arguing for the reader’s
(or interpreter’s)^ obligation to identify and articulate “original
meaning,” primarily as opposed to what he calls “anachronistic” or
“nonauthorial” meanings, that is, meanings which there is reason to
believe (presumably on linguistic or historical grounds) that an

136
 THE ETHICS OF INTERPRETATION

author did not intend, or could not have intended, his work to
convey. The opposition and preference may appear reasonable
enough, but when they are examined more closely in connection with
the nature of literary meaning and literary interpretation, the
opposition becomes tenuous, the preference dubious, and the ethics
of interpretation are seen to lie elsewhere.

I. Determinate and Indeterminate Meanings

I would begin by suggesting that Hirsch’s tendency to equate
meanings with authorial intentions, conjoined with his failure to
distinguish between natural and fictive discourse, have made it

impossible for him to recognize that although the intentions of all

authors are historically determinate, the meanings of all utterances

are not. I shall elaborate both halves of that observation, beginning
with the meanings of utterances and proceeding to the nature of
authorial intentions. It will be useful at this point, in order to clarify
the distinction between determinate and indeterminate meanings in
natural and fictive utterances, to draw together a number of
observations made earlier in these lectures:

1. Because every natural utterance, from a conversational

remark to a philosophical treatise, is a historical act and a historical
event, some of its meanings are historically determinate. That is, the
conditions that occasioned the utterance and shaped its form, the
conditions to which it was a response, are understood to have
occurred in the historical universe. Therefore, like the meanings of
any historical event, from the explosion of a supernova to the falling

of a leaf, some of the meanings of a natural utterance can be more or

less validly inferred by an interested observer, and the validity of his
inferences can be challenged or supported by, among other things,
appeals to historical evidence.
2. Because a natural utterance is also a verbal act and event,
some of its historically determinate meanings presumably can be
and are expected to be inferred by the listener on the basis of
linguistic convention, and the propriety of those inferences can be
challenged or supported by appeals to those conventions.
3. Because the composition (N.B., composition) of a fictive

utterance is a historical act and event, some of the meanings of that

137
 EXCHANGING WORDS

act or event are historically determinate; that is, it is understood that

the conditions that occasioned the composition of the utterance and
shaped its form as a historical act occurred in the historical universe
and that some of those conditions are at least theoretically ascer-

tainable on some basis. If the fictive utterance is a literary work,

such as a poem, play, or novel, the meanings of the work, in this

sense, would be no different from the meanings of a nonliterary

composition. They would include everything from the author’s most
intimate motives in composing it to all the social and intellectual
circumstances that could be conceived of as having occasioned its

composition or shaped its form; and inferences concerning such

meanings could be supported or challenged by appeals to anything
from personal diaries and theater records to the traces of an entire
set we have the efforts of (and
of social and cultural conditions. Thus
the disputes among) literary scholars, biographers, and historians.
4. Because a fictive utterance is a verbal structure, some mean-

ings can be assigned to it — and were presumably expected and

intended by the author to be assigned to it —on the basis of linguistic

convention, and assignments of meaning made on that basis can be

supported or challenged by appeals to those conventions. Thus, if

the utterance is a literary work, we have the efforts of (and disputes

among) literary explicators. Of course, especially if the work was

composed in a remote time, the establishment of those conventions
themselves may require historical validation and therefore be sup-
ported and challenged by appeals to historical evidence. Thus we
have efforts of (and disputes among) literary philologists.
5. However, because a fictive utterance is not itself a historical
act or event, because it is understood that that verbal structure was
not “performed” and did not “occur” in the historical universe,
some of its meanings are historically indeterminate and therefore
not even theoretically ascertainable on the basis of historical evi-
dence. In other words, to speak of the meanings of a fictive

utterance as historically indeterminate is not to override — ignore,

mistake, or betray —something that is there, but to acknowledge the
fact that something is not there.
4'hese observations will be made more concrete if we consider the
following two situations:

138
 THE ETHICS OF INTERPRETATION

A. We discover, in an old attic trunk, the fragment of what

appears to be a letter written by some unknown correspondent to
some equally unknown recipient.
B. We encounter, in the current issue of a magazine, a verbal
structure identical to that of the letter fragment, but labeled
“poem.” The author, let us say, is someone about whom we know
nothing but his name.
In both the letter and the poem, there are an unidentified “I” and
“you,” and various objects, places, and events are alluded to in
general terms: a “tree in the garden,” the time “I saw you walking
by the seashore.”
As we read the letter, we will “interpret” it as an eavesdropper
might interpret an overheard conversational remark; that is, we will
infer meanings from it on the basis of linguistic convention and

whatever general knowledge seems relevant. The writer appears to

have been a woman; she was, it seems, rather imperiously demand-
ing someone’s return. We gather that she and the person addressed
had lived together in a house by some seashore and that there had
been a garden attached to it.

Of course, because the letter is an inscribed utterance, certain

textually uncodable features of vocal speech —such as intonation
will be lacking, and we may find ourselves unsure of its precise tone.
Perhaps it is not so much imperious as cold; and perhaps the
coldness is a mask for self-doubt or pride. Thus, in interpreting the

letter, we will inevitably be drawn into speculation about various

conditions not inferrable from it on the basis of linguistic con-
vention: all those conditions that would enrich or constrain the
inferences drawn on that basis — for example, the character of the
writer, her relationship to the person addressed, and the exact
motives and circumstances that occasioned the letter. And we may
also be drawn into speculation about the particular objects and
places alluded to.
We would recognize, however, that our speculations were just

that: hypotheses entertained in ignorance but at least theoretically

falsifiable by further or more specific information (“Oh, it was
written by Aunt Louise; and it was her daughter, not her lover, she

was asking to return —she was very ill at the time, you see —and the
139
 EXCHANGING WORDS

house had been in California, not Maine,” and so on). We would

recognize, in short, that if the letter was a natural utterance,
composed as a historical act and regarded as a historical event, some
of its meanings were historically determinate, at least theoretically
locatable in the historical universe.
As we read the poem, indistinguishable from the letter as a verbal

structure, we will again ascribe meanings to it on the basis of

linguistic convention and general knowledge. As with the letter, we
may again be drawn into speculation about the tone of the utterance
and about the circumstances and motives that occasioned it. And we
may again supply particulars for the general allusions: the tree will

be pictured as an oak, the seashore will be imagined as that of

Maine, and the “I” and “you” will acquire personal features and
personal histories as evoked by that verbal structure but also as
supplied by our own projections, drawn from our own experiences.
As with the letter, we will recognize (1) that the particulars we have
supplied are not “in” the poem (that is, not inferrable from it on the
basis of linguistic convention: “tree” does not imply oak any more
than it implies birch, “seashore” does not imply Maine to the
exclusion of California), (2) that these particulars may not cor-
respond to those the author had in mind when composing it, and (3)

that another reader would be likely to supply somewhat different

particulars and also to imagine somewhat different circumstances
and motives as occasioning such an utterance.
What distinguishes our speculations here is that none of them are
falsifiable by subsequent information. An adequate or valid inter-
pretation of the poem — an interpretation that served our as interests
readers and the poet’s interests as an — need not wait upon and
artist

does not depend upon our identifying the particular tree the poet
had in mind. For the species of that tree and the location of that
seashore will be understood by both the poet and the reader as being
historically indeterminate — as will, also, the unspecified features
and character of that “I” and “you,” and the unspecified motives
and occasion of that entire utterance. All of these meanings, to the
extent that the poem is offered and taken as a fictive utterance, will
be understood to be unfixable, unbeatable in the historical
universe.

140
 THE ETHICS OF INTERPRETATION

I have emphasized the condition “offered and taken as a fictive

utterance’’ because that does, in fact, make all the difference. If the
structure labeled “poem’’ was designed by its author to be taken as a
natural utterance, the label was unfortunately chosen, for it would
misdirect the assumptions of at least a large number of its readers.
The same would be true had the structure been presented with a title

or lineation or any other features suggestive of such structures as are

commonly labeled “poem’’ and had not been otherwise dis-
tinguished from them. Conversely, the fragment that we took to be a
letter may have been designed to be taken as a poem or fictive
utterance but, lacking signals to that effect, have been misidenti-
fied. Errors of identification produce erroneous assumptions and
bring into play inappropriate conventions. Conventions are con-
ventions, however, and they may change over time and, under
varying conditions, be altered. Since I have discussed both these
matters elsewhere at some length,^ I shall not elaborate them further
here. We may now return to the interpretation of our poem, taken as
a fictive utterance, and our letter, taken as a natural utterance.
Now, in addition to all the meanings we have considered up to this
point, there is another set of meanings that may be attached to any
utterance, natural or fictive, and indeed to any object or event.
What I refer to here are all those relations, connections, or
implications which, as I mentioned in the first lecture, are often

distinguished from antecedent or causal meanings by some other

term, such as significance or import. These other meanings include
various metaphoric or analogic implications, as when a tree in a
garden is said to “signify’’ nature, endurance, traditional family
values, or the temptation of Adam and Eve. Objects and events
“mean’’ or way when they
“signify’’ other objects or events in this

are seen by someone as belonging to the same class by virtue of what

are perceived as shared properties. These other noncausal meanings
also include those more or less general propositions that an event
may be seen to instantiate or exemplify, for example that Russian
landowners were oblivious to the condition of the peasants or that
provincial life produces intellectual stultification or that suffering is

inescapable. Events “mean’’ or “imply’’ such propositions when they

are seen by someone as being instances or exemplifications of them.

141
 —
EXCHANGING WORDS

It will be worth our while to focus here on the second type of

noncausal meaning, that is, meaning as exemplification. If we
return to our hypothetical poem and letter, what appears to be
different about the poem in this respect is that it was meant to have
meanings of that kind. The implication — or, as it is sometimes said,

the “communication” —of such meanings, that is, more or less

general propositions about the universe, would seem to be one of the
primary motives for writing poems, and also plays and novels. The
authors of such works, it seems, would have us know what they have
been blessed or cursed to know: that love is ruthless or redeeming,
that men are craven or heroic, that life is, after all, meaningful — or,

after all, meaningless. To this extent, we may say not only that the
line between didactic poetry and pure poetry is hazy, but that all

poetry is didactic. We usually refer to a work as “didactic” when

such propositions are explicitly formulated within them. But all

works of literature may be seen to imply propositions, most of them

not stated explicitly and many of them unstatable unspeakable —
in terms of the formulated wisdom of the culture. Indeed, every
literary work may be regarded as Xht fictive exemplification of some
set of propositions and, as Sir Philip Sidney observed, “a feigned
example hath as much force to teach as [doth] a true.” It is as if each
novel opens with the invisible words, “For example”: “[For ex-
ample,] about thirty years ago. Miss Maria Ward, of Huntington,
with only seven thousand pounds, had the good luck to captivate Sir
Thomas Bertram, of Mansfield Park and as if every play
opens with an invisible prologuist who says “For example, events
such as these could occur,” and every lyric with the invisible words,
“For example, I (or someone) could say....”** Furthermore, it

might be maintained that the interpretation of a literary work

should be directed largely or most centrally to inferring its meanings
in this sense, that is, to identifying and making explicit the
propositions that it exemplifies, these propositions being seen as
precisely the point, the significance, the import — in short, the
meaning —of the work.
The difficulty with all this, however, is that every literary work,
like every object or event in the universe, can imply —can be seen to
exemplify — an infinite set of propositions and can correspond ana-

142
 THE ETHICS OF INTERPRETATION

logically to an infinite number of other objects and events. To be

sure, when we are speaking of an object or event that exists by virtue
of personal agency, such as an utterance or a poem, among that
infinite set of propositions and correspondences may be some partic-
ular sets that it implied for the author and that the author intended
it to imply for his audiences. The question, then, for the ethics of
interpretation, what extent, and for what reason, the reader or
is to
interpreter of a literary work is obliged to identify and articulate
those particular sets and obliged to infer no other sets. And an
adequate answer would, I think, acknowledge the following: (a) that
we cannot ever identify those sets, but only hypothesize them, {b) that

while the provisional identification of those sets may be of interest for

some purposes, it may also be the very process of hypothesizing them
that, in our experience of the poem, is “interesting,” (c) that the
reader inevitably will infer sets of propositions and correspondences
that are different from those the work implied for its author, for no
two persons could infer exactly the same sets or, of course, articulate

them exactly the same way, {cD that it is most unlikely that the work
implied only one particular set for the author himself, even while he
was composing it, and, finally, (c) that we are under no ethical
obligation to do what cannot be done, or to refrain from doing not
only what we inevitably will do but the doing of which may be a
significant aspect of our engagement with a literary work and a
central source of its interest and value for us.

Some of these points will be considered more fully when we turn to

the matter of authorial intentions. One of them, however, can be

considered immediately, and that is how the distinctive nature of

literary interpretation contributes to the characteristic value of

literary works. We may, one final time, compare our hypothetical
poem and letter.

I referred earlier to the speculations we may entertain when we

read each of them, speculations about such matters as the character
of the “I” and “you,” the nature of the relationship between them,
and various particulars of place, circumstance, and motive. The
difference that I should like to emphasize now is that, with respect to

the poem, our speculations constitute our “interpretation.” In other

words, since these matters cannot be even theoretically determined.

143
 EXCHANGING WORDS

there is nothing we can do but endlessly hypothesize them. I would

suggest, however, that in that very endlessness is a major source of
the poem’s vitality and a continuing source of its value for us. While
we may transform the letter, originally designed to function in a

natural verbal transaction, into an occasion for such speculative

inferences (that is, into an occasion for “cognitive play”), we may
have reason to suppose that the verbal structure labeled “poem” was
designed (meant) to be interpreted in that way — and designed
(constructed) accordingly. What I would add now, in view of the
discussion of meaning as “import” or “significance” (that is, the
infinite set of metaphoric correspondences and general propositions
implied by the poem), is that here, also, it is precisely their endless-
ness that gives the poem its life.

A poem may be described as, among other things, a structure of

parabolic meanings, “parabolic” in both senses, that is, displaying
the infinitely open curve of a parabola, and forming parables for an
infinite number of propositions. It is partly by virtue of the parabolic
nature —or indeterminacy—of “meaning” in poetry, and the con-
sequent difference in the process by which we arrive at that
meaning, that poems acquire value for us. For they thereby be-
come the occasion for the exercise of the reader’s own imaginative
powers, specifically as the occasion for unusually creative cognitive
activity. Engaging in this activity is not the mark of the reader’s
solipsism, self-indulgence, or amorality; nor is it an incidental or
gratuitous by-product of his experience of the poem; on the con-
trary, as I suggested in the last lecture, it can be quite central to that
experience. I should now like to make a few observations on another
aspect of that activity.
At every moment, throughout our lives, we are the subjects of
potential “experiences,” but we are not always aware of them as
such. Given the practical demands of ordinary existence, we cannot
give equal attention to all the events that impinge on us, either from
the external world or from the world of our own feelings; nor do we
recall them all with equal precision. Our perceptions are not only
directed but selected by the demands of the immediate occasion,
and our experiences are usually preserved in memory only to the
extent that they continue to serve our cognitive needs. Thus, much

144
 THE ETHICS OF INTERPRETATION

that is potentially knowable to us, because it is part of what has, in

some happened
sense, to us, slips by apparently unknown or at least
unacknowledged
For the same reason, however, some particular subsequent occa-
sion can evoke those forgotten or unacknowledged experiences,
causing us to remember what we did not even know we knew: a
perception never before quite articulated, an emotion we had
sustained on the periphery of consciousness, a sense —barely
grasped before —of the “import” of some incident. Our memories
are thus triggered; our knowledge is magnetized to a center, ordered
by and organized around it; and those prior “experiences” become
available to us because the present occasion presents us with some
reason for acknowledging them. Thus we are put in full possession
of what was always our property but kept in reserve, as it were, until
we came of age and found some way to use it.

Poetry may be eminently an occasion of this kind for the reader.

To the extent that the meanings of a poem are understood to be
indeterminate, we must supply meanings for it; and by obliging us to
do so, the poem creates a need and therefore a use for knowledge
that might otherwise remain unavailable to us. This is not to say,
however, that a poem is a springboard for daydreaming. On the
contrary, its value lies not simply in provoking that activity but also
in shaping and, indeed, resisting it. Even as certain possibilities of

interpretation are opened, they are also directed, lured, and re-

directed by the poet through the verbal structure he has designed.

I we can see, in this description, the outlines of an
think
engagement between the poet and reader that is very close to that of
a game. Moreover, as in any game, there may be masters and
amateurs on both sides of the board: readers who, less attentive than
others to the poet’s own moves, play a wilder but perhaps less
gratifying game — and, for an observer, a less elegant one; “unimag-
inative” or “literal-minded” readers who, not taking advantage of
the openings, play a tighter, more timid, and, for the observer, less
interesting game.
Having mentioned possible observers, we might pursue the
analogy a bit further, noting that literary interpretation can also be a
spectator sport and, in fact, has been one for some time. The

145
 EXCHANGING WORDS

“interpretation” of a literary work may be seen as the descriptive

report or reenactment of a game that has already been played. And
since there can be masters on both sides of the board —great readers
as well as great poets, matches for each other in the boldness or
subtlety of their moves —there can also be master games, as
engaging for the spectator as those he plays himself, or even more
engaging. Thus, to fill out an incomplete list presented earlier, we
have the achievements not of literary historians or explicators as such
but, precisely, of literary interpreters, those who offer not to give us
the poem —which often means taking it away —but rather to take it

on for us, as one plays a match with an opponent.

Authorial Intentions and Literary Ethics

II.

But where, in all this, are the intentions of the author? Do we not, in
making sport of him, in turning his words into cognitive toys, use
him merely for our own ends and thus transgress ethical norms? I
do not think so, and we have, in fact, been speaking of the intentions
of the author all along.
1 observed earlier that although the intentions of all authors are
historically determinate, the meanings of all utterances are not.
Having considered the second part of that statement, we may now
consider the first and, finally, the implications of both for the ethics
of interpretation.
The intentions of an author with respect to his composition,
whether originally designed as fictive or natural discourse, can
indeed be seen as historically determinate — and therefore at least
theoretically ascertainable — but only in the sense in which the
intentions of anyone, with respect to any act he performs, are
historically determinate. Thus, if we see a boy throw a basketball
iuvv'»rd a hoop, we may infer that he “intends” to get it through: in
other wo.ds, iiiat gening it through is, for him, the appropriate
consequence of that act. Our inference may, of course, be incorrect:
perhaps he was, in fact, mischievously intent on breaking the garage
window just above the hoop. Whether or not we inferred them
correctly, however, his intentions were historically determinate and
someone who knew more about the thrower and the context of his
throwing might have inferred them more accurately.

146
 THE ETHICS OF INTERPRETATION

What is distinctive about a symbolic act, we recall, is that the

performer’s intentions can be realized, that is, his act can have
appropriate consequences for him, only to the extent that the
conditions in response to which he performed it — its meanings in

that sense —are correctly inferred by someone. If the boy’s act was to

—
some extent symbolic if he wanted it to *'say something” to
someone (about his hostility perhaps, toward the owner of the ga-
rage window) —
then his intentions could be realized only if they
were correctly inferred by an appropriate witness.
Consider, now, the situation in which a young girl announces
meanacingly to her father, ‘‘I’m a wicked witch and I’m going to put
you in a dungeon.” Although she is, we assume, presenting a fictive
utterance, and thus does not intend her father to believe that she is a
witch or will put him in a dungeon, nevertheless, she does intend to
have some effect on him — perhaps to amuse him, divert his atten-
tion or engage him in play— and that historically determinate
intention may be correctly or incorrectly inferred. If, in the midst of
play, she had directed her announcement to a much younger child
and thereby ‘‘unintentionally” frightened him, we would say that the
younger child had misinterpreted her intentions —but precisely, we
should note, because he had taken her to mean what she said, had
interpreted her words in accord with the conventions of natural
discourse — or, in Hirsch’s terms, in accord with the ‘‘imperatives”
of ‘‘ordinary speech.” Clearly, then, it may be among the historically

determinate intentions of some ‘‘authors” that we not take them to

mean what they ‘‘say.”
In natural verbal transactions, the speaker’s act has appropriate
consequences, and his intentions are realized, only insofar as the
listener correctly infers the motives, sentiments, and circumstances
that occasioned his utterance and are implied by it in accord with
linguistic convention. When my daughter remarks growlingly at

breakfast, ‘‘Damn it, I didn’t get a minute’s sleep last night,” I

assume that she expects me to infer, from the fact and form of her
utterance, that she spent a more or less sleepless night, that she is

now in an irritable mood, and, furthermore, that, having inferred

these things, I will behave accordingly, that is, offer commiseration
and make few demands on her energy. Had I not made those

147
 —
EXCHANGING WORDS

inferences, her act in speaking would have failed of its intentions,

which were primarily to secure my services in her behalf.
The “appropriate consequences” of an author’s act in composing
and offering a poem or play or novel are not, however, usually a
matter of securing our services or affecting our behavior in that

sense. Indeed, the composition of such a work is not, in that sense, a

symbolic act at all. The author may have found satisfaction in
anticipating (and thus would have “intended”) his readers’ recogni-
tion of certain allusions, their appreciation of certain experiences

and illuminations, and their inference of certain “propositions”

(perhaps otherwise unspeakable and inaudible) that the work could
be seen to exemplify. It also would have been his intention, we
assume, to produce a work which, in engaging our interest and
providing for us the pleasures of verbal art, would secure our
admiration for his artistry. None of these intentions, however,
depend for their realization on the reader’s identifying the partic-
ular historical determinants of his composition, the personal motives
and the specific sentiments, states, and circumstances that occa-
sioned its creation; nor are those meanings implied by or “expressed”
in the poem.
may be presumed, for example, that when Keats wrote his
It

sonnet “To Sleep,” not only did he not expect Sleep or anyone else to
act in response to the imperatives of which it consists

O soothest Sleep! if it so please thee, close.

In midst of this thine hymn, my willing eyes . . .

— but he also did not expect anyone to infer anything from the poem
concerning his wakefulness on any particular night or, indeed,
anything in particular concerning him as a historical person. His act
in composing the sonnet did, of course, have particular historical
determinants (which may have included the insomnia produced by
his nascent tuberculosis, or his current fascination with Shake-
speare’s sonnets), and we may, for various purposes, seek to identify
some of them on the basis of biographical and other historical
evidence. It is most doubtful, however, that such determinants were
what the poem was designed to “express,” or that the inference of
them is required for the poet’s act in composing it to have

148
 THE ETHICS OF INTERPRETATION

“appropriate consequences.’’ On the contrary, it is more likely that

in designing his poem as a verbal artwork, Keats would have
expected and intended his readers not to infer those determinants,
not to direct their interpretations and certainly not to confine them
to ascertaining the particular conditions that occasioned its compo-
sition. Indeed, the crudest violation of a poet’s artistic intentions,
depriving his work of its potential effectiveness in providing the
particular pleasures and interests he did design it to afford, would
be to insist on interpreting it as a natural utterance. It would, we
might note, be the converse violation of linguistic convention alluded
to in the last lecture as reverse lying or “false listening.’’

Let us turn now to the general question of authorial intentions as

raised by Professor Hirsch’s argument. As we have seen, “authorial
intention’’ can be taken to mean many things, among them the sort
of ejfect that any author or speaker designs his words to have on his
audience. These effects may range from such specific and local ones

as “eliciting information on the time of day’’ or “evoking the

passage in Canto V of the Inferno concerning Paolo and Francesca’’
to such general and comprehensive effects as “being instructive’’ or
“being amusing.’’ If that is, at least in part, what Hirsch means by
it, we could agree that a listener or reader who missed or miscon-
strued such authorial intentions would, to that extent, have mis-
interpreted the utterance. We could also agree that the misinterpre-
tation would be, if not quite unethical, then at least unfortunate, the
degree of the misfortune depending on the nature of the author’s
intention and the consequences for both him and his listener
entailed by its correct or incorrect interpretation. If a stranger asks
us for directions toTimes Square and we reply by telling him how to
get to Grand Central Station we have, in so misconstruing his
intentions, behaved rather stupidly, to his distress. Or if someone
shouts a well-intentioned warning to us in traffic, and we mistake it

for an obscene we may put ourselves in otherwise avoidable

call,

physical jeopardy. Normally, however, we take care as speakers and

listeners to make intentions readily inferrable and to identify them

as best we can, since it is usually to the advantage of both parties in

natural verbal transactions that a speaker’s intentions be correctly
inferred.

149
 EXCHANGING WORDS

To a certain extent we may, as Hirsch suggests, extend these

observations to poems, plays, and other literary works, noting, for
example, that it is always unfortunate for a reader when, through
ignorance or incompetence, he misses out on whatever delight or
interest the author designed his work to provide, and that, if the

ignorance or incompetence is grave enough (for example, leading

him to mistake a parody for the real thing), it may effectively deprive
him of the work altogether. Moreover, although in the case of dead
poets it is hard to say what misfortune they suffer from a misidentifi-
cation of their intentions, it could certainly be maintained that the
effectiveness and thus continued production of all art depends on
the existence of competent and informed audiences who can experi-
ence the effects that individual artists designed their works to
produce.
This much said, however, there are a number of further points
that may be made if we consider our actual practices and what might
seem, in the terms of Hirsch’s formulation, to be our morally
unregenerate behavior with respect to literary works. For one thing,
we may always choose, for aesthetic or other ends, for example in
book titles (For Whom the Bell Tolls), epigraph quotations (//
rniglior fabhro), and “found poetry,’’ to decontextualize a verbal

composition, detaching it from its original moorings and exhibiting

it in a new context that evokes interpretations that may be quite

remote from those presumably, or even certifiably, intended by its

author. 1 do not, however, believe that we need think of ourselves as

violating authorial intentions when we do so; for, in knowingly and
candidly appropriating for one purpose (and to evoke one set of
meanings) a verbal structure originally designed for another purpose
(and, as such, designed to evoke other meanings), we have in effect
regiven it to ourselves and thus reauthored it. The composer of a
verbal structure does not, and presumably does not expect to, retain
eternal proprietary rights to the manner of its employment.
Second, it seems obvious that much of the canon of “literature’’
consists of works that have been pressed into services they were not
designed to perform, works that attain the status of literature and
have value as such largely to the extent that we do “disregard” or
“fail to respect” their author’s original intentions. Thus we read for

150
 THE ETHICS OF INTERPRETATION

“entertainment” what was intended for instruction or inspiration,

and now find aesthetically engaging what was originally meant to
rouse us to political action or convert us to pious practices, or to
warn, accuse, blandish, or seduce the readers to whom they were
originally addressed.
Finally, if we consult the literary history of any culture, we find
that the works that survive —the plays that continue to be produced,
the poems and novels that continue to be read, the proverbs that
continue to be cited, the tales that continue to be recited —are those
and exemplify emergent meanings. It is true that literary
that evoke
works may engage our interest and touch our spirits as records and
images of an otherwise irrecoverable past; but to endure as some-
thing other than vivid historical artifacts, they must also be able to
serve as metaphors and parables of an unpredictable future. They
must, in short, continue to have meanings independent of the
particular context that occasioned their composition, which will
inevitably include meanings that the author did not intend and
could not have intended to convey.

I may seem to be maintaining that there are no ethical considera-

tions in literary interpretation. That is not the case. There is an
ethics of interpretation; what it governs, however, is not the behavior
of interpreters toward authors but rather of interpreters toward their
own audiences. For the interpreter, if he speaks, presumably speaks
natural discourse. His utterances are therefore governed by the
ethics of the linguistic marketplace, where the fundamental impera-
tive for all speakers is that they mean what they say and take
responsibility for having said it. For professional exegetes, this

means, among other things, that they should acknowledge the

nature, limits, motives, and consequences of their activities; speci-
fically, they should recognize and acknowledge that the publicly
articulated “interpretation” of literary works (for example, in

classrooms and in the pages of professional journals) is a highly

specialized activity, the forms of which reflect the particular histori-
cal and cultural conditions in which it arose and in which it is now
pursued, and the pursuit of which serves particular social and
institutional functions.

151
 EXCHANGING WORDS

To the extent that interpreters claim to have identified the

historically determinate meanings of a work, their claims may, as I

pointed out earlier, be supported or challenged by appeals to

historical evidence and the proprieties of evidentiary logic. An
interpretation can be demonstrably “wrong,” however, only to the
extent that it claims to be demonstrably “right,” and it is clear that
there are forms of “interpretation” that need not and should not
make such claims: those, for example, that I alluded to earlier as
reports or reenactments of a reader’s individual engagement with a
poem. Such interpretations may themselves be seen as specialized
(that is, extended, and more or less formalized) versions of such
common social activities as informal chatter about songs and stories,

casual allusions to passages of poetry, exchanged recollections of

characters and incidents in plays and novels, and the invocation of
elements of particular literary works as illustrative analogies to
currently interesting situations or as exemplifications of currently
interesting propositions. In their most highly elaborated versions,
the potential interest and value of such interpretations would consist
not in the interpreter’s putative identification of the historically
determinate meanings of a work but, rather, in the intellectual
subtlety and imaginative fertility that he displays in playing out
various of its historically indeterminate meanings, and also in the

general or immediate interest (for example, philosophic or political)

of such propositions as the work may then be perceived to exemplify.
Conversely, to the extent that such an interpretation exhibits the
intellectual crudeness and imaginative barrenness with which the
interpreter has experienced the work, and also to the extent that the
propositions or attitudes he uses the work to exemplify are them-
selves seen as shallow, vulgar, or otherwise noxious, the interpreta-
tion will be of little interest or value and disdained accordingly.
The two forms of interpretation described here are not, of course,
clearly separable. Indeed, it is doubtful if either could be delivered
in its pure form, that is, as an “explication” uncontaminated by the
individual interests and pattern processing of the interpreter, or
as a “reading” that reflects no assumptions about the historical
determinants or intended meanings of a work. On the contrary,
as I suggested in the last lecture, a significant source of the pleasure

152
 THE ETHICS OF INTERPRETATION

and interest we take in our cognitive engagement with a poem is our

presumption and projection (or hypothesizing) of the artist’s

“design,” and a corresponding source of interest and pleasure for

the spectator of such a reported engagement —or reenacted
“game” — is precisely its consequent quality of inter-play. This is a
quality, it might be added, that extends to the transaction between
the interpreter and his own audience, since the latter will inevitably
“match” his own experience of the work against that represented by
the offered “reading,” and, to the extent that the interpreter’s game
thus itself becomes an occasion for cognitive play, it serves an
“aesthetic” function. Since a given interpretation may serve a
number of functions simultaneously, there is no particular reason
why, in any case, such separation or purity should be sought.
Controversy inevitably arises, however, when the different claims
that can be made by each are confused (either by the interpreter or
his audience) or when the relatively distinctive sources of the
interest and value of each are seen as opposed.

Which brings me to my concluding remarks. The two forms of

publicly articulated “interpretation” that have been roughly delin-
eated here (and other forms that could be seen as variants or variant
combinations of them) are often set in polarized opposition to each
other. As I have noted, each form serves more or less distinctive

functions and each may be valued accordingly. While it may be

observed that the functions served by each will themselves be valued
more or less highly insofar as they are seen as instrumental to other
valued ends (for example, political or intellectual, communal or
individual), I do not myself believe that either has any absolute
priority of interest or unique claim to value, nor do I believe that the
functions served by each of them are necessarily competitive.
The effort to identify the historical determinants of literary works
(for example, the cultural circumstances and personal motives that
occasioned their composition and shaped their form; the linguistic
and literary conventions in accord with which they were originally
intended to be interpreted and experienced; the attitudes, values, or
sentiments they may have been designed to exemplify and perhaps to
promote) is a pursuit that clearly serves certain functions; and those

153
 EXCHANGING WORDS

functions are, to my mind, of considerable social value and intel-

lectual interest. My argument with Professor Hirsch has not been an
argument against the practice of “historicist criticism” but rather,
as I stated at the beginning of this lecture, with the attempt to give
certain forms of interpretation privileged status on what appear to
me to be the extremely tenuous grounds of their unique claim to
ethical propriety. As part of the counterargument developed here, I

have sought to emphasize that whatever historical investigations can

determine, they cannot determine the historically indeterminate
meanings of a poem. Those meanings remain indeterminate by
nature, them reserved for the
the inference or “supplying” of
individual reader as part of his engagement with and experience of
the poem. I have also meant to suggest that it would be grossly
impoverishing to restrict the activity of “interpreting” literary works
to the establishment of their historically determinate meanings, no
matter how much comfort we might take in the conviction that we
had thereby banished, or branded in advance, all invalid or
unvalidatible interpretations.
I should like to conclude by observing that the impoverishment
alluded to here would be a quite radical one and, by thus making the
universe a poorer place for ourselves and our fellow creatures,
subject to a categorical imperative. For if, as I have suggested, it is

at least in part as the occasion for individual cognitive activity that

literary works acquire value for us, it would follow that any
hermeneutic principle or pedagogic practice that aborted that
activity would, to that extent, diminish the value of those works.
Indeed, were it possible for the strenuous, concerted, and cumula-
tive efforts of professional interpreters ever to persuade us that all

the meanings that could be validly inferred from a poem were now
“known,” that poem would effectively cease to function as one. In
short, if the ascetic view of interpretation against which I have
argued here were somehow to prevail, not only would readers have
little reason to read poems, but poets would have little reason to
write them: they would be better advised (as, of course, they
sometimes have been) to come right out and say what they mean.

154
LINGUISTICS AND
LITERARY THEORY
•
I j
 SURFACING FROM THE DEEP

It does not, I think, require an eye made especially sober by the light
of too many word new a certain pathos.
setting suns to find in the
New clothes, new toys, bear beneath stiff folds and bright surfaces
images of the tattering and chipping to come. The newborn and
newlywed both figure forth in their very names the shadow of a
temporality made more poignant by their own ignorance of it. Or,
where a measure of self-consciousness must be presumed, as in the

numerous recent announcements, from the expeditionary forces of

literary study, of new approaches and new directions toward new

horizons, there is pathos in such outbraving it in the face of the sorry

destinies of so many of those who went before. Thus, confronted by

this volume,' one might, even before opening its covers, find some-
thing unhappy in its subtitle —both for the reasons mentioned and
also because, in implying obsequies for the old stylistics and thus
acknowledgment of the brief life span of its most immediate fore-
bear, it announces a birth manifestly tainted by genetic weakness.
There is, of course, also belligerence and bravado in that subtitle:
for the New Stylistics, cradled in the grave of the old, must, in laying
claim to its inheritance, allay the natural suspicion that it will be,

like its ancestor, subject to similar rapid obsolescence. Such a

suspicion is, however, not altogether quelled by closer examination
of the papers here assembled and thus christened.

157
 LINGUISTICS AND LITERARY THEORY

The editor of the volume, Roger Fowler, has written an intro-

duction to it and also contributed one of its seven papers. Since the
burden of his introduction and the argument of his paper are closely
related, I shall consider them together, below. It will be instructive,
however, to look first at the two papers that precede Fowler’s own;
for one of the features that the new stylistics shares with the old is a
frequent disparity between its general claims and its particular
achievements, reflected in the present volume in the discrepancy
between the actual character of a number of the articles in it and the
editor’s introductory comments on them.

I. Syntactic Strategies and Analytic Formulas

Donald C. Freeman’s examination of “syntactic strategies’’ in three

poems by Dylan Thomas (“The Strategy of Fusion: Dylan Thomas’s

Syntax’’) is offered both as evidence for a hypothesis concerning
language in poetry and as a demonstration of a recommended
method of critical analysis. The hypothesis, in its fullest formula-
tion, is as follows: “One way in which poetic language differs from
ordinary language ... is that a poet’s deployment of his language’s

transformational apparatus, its syntactic patterns, not only reflects

cognitive preferences, a way of seeing the world; perhaps more
importantly, it reflects the fundamental principles of artistic design
by which the poet orders the world that is the poem’’ (19-20).^ Aside
from the noxious jargon, the most problematic terms here are
“reflects” and “artistic design.” Reflects could mean anything from
“is determined by” to “manifests” or “instantiates” and seems, in

the discussion that follows, to mean them variously and

all of
inconsistently. Artistic design is not only vague but radically am-
biguous, its meaning hovering between something like the poet’s

aesthetic purpose (or intentions or ideas) and something else like the
poem’s aesthetic structure, and thus anything from the overall
thematic principles by which the poem is organized to the specific
meanings (or effects) the poet wishes to suggest (or create) or his

vision of the universe. Thus although it is clear that Freeman wishes

to hypothesize some relation between a poem’s (or poet’s) syntax and
something else about it (or him), it is not clear to what the syntax is

related and it is also not clear if the relation is one of causality (and

158
 SURFACING FROM THE DEEP

if SO, in which direction), correlation, covariance, or strictly any-

thing at all.

Not only is the hypothesis itself so vague as to be vacuous but,

since the evidence offered to support it consists of a series of
analyses that assume it, the question is begged all through. More-
over, the analyses themselves, which Freeman offers as a model
method by which we can “lay bare the deep form of particular
poems’’ (39) and “gain a deeper insight into the poem’s inner form
and aesthetic center’’ (20), consist of the familiar tortuous parsings
which, here, yield simplistic, dubious, or incomprehensible revela-
tions which in turn, by yet another familiar circular route, are
clearly restatements of the interpretations of the poems with which
the analyst began. Freeman’s summary gives the flavor of much of
what precedes it: show how a number of syntactic
“I have tried to
strategies — first the yoking of immediate syntactic constituents
which are contradictory; second, intensive relative clause formation
and related transformations; and finally, various kinds of preposing
transformations —have functioned with increasing force in three
poems to help achieve the unity of man and natural process which is

one of the central principles of Dylan Thomas’s poetics’’ (39).

In the introduction, Fowler alludes to Freeman’s paper as “typical

of recent stylistics in adopting a powerful and complex central
hypothesis as the point of departure.’’ He continues:

Freeman’s chosen text —three poems by Dylan Thomas — is

not offered as virgin territory to be explored by a perfectly

neutral analytic machine [as presumably might have been
the case in the old stylistics] . . . Freeman approaches
Thomas with a specific line of enquiry or field of interest:
the poet’s habitual synthesizing, identifying, generalizing
preoccupations.... This preoccupation assumedly en- is

coded in the artistic design of Thomas’s poems; and the

artistic design both resides in the language and is discovered

by the reader in his language-induced experience of the

poems. TG [that is, transformational-generative gram-
. . .

mar] provides a delicate analytic apparatus for just the

kinds of [syntactic] structures Freeman needs to de-
scribe. Freeman shows how such structures, in Dylan
. . .

159
 LINGUISTICS AND LITERARY THEORY

Thomas, arrest the reader and, as he unravels their con-

torted syntax, guide him to the heart of the author’s artistic
design. In the relation he sets up between grammatical
model and critical problem. Freeman illustrates not a
mechanical transposition of text into linguistic jargon, but
a dialectical process in which the linguistics and the litera-
ture are mutually responsive (5-7).

It is true that in the new stylistics, unlike the old, “the reader’’ is

conspicuously hooked into the machinery, so that whereas earlier we

might have been offered simply a description of features and devia-
tions, we
now to understand that the features and deviations are
are
described by way of accounting for the reader’s “language-induced
experience of the poem.’’ As in Freeman’s paper, however, where
the peculiarities of Thomas’s syntax are alleged to be of interest
because they “foreground” for the reader the most crucial features
of the poet’s “artistic design” (24), the hookup often simply attaches
a new piece of apparatus onto the machine (here an ill-fitting cog
imported from Russian and Czech Formalism) without appreciably
changing either its product or efficiency. When, as before, the key
concepts are vague, the assumptions dubious, and the aims con-
fused or quixotic, the results of the analysis, despite the array of
technical terms, tree diagrams, charts and formulas, will demon-
strate nothing other than the operations of the machine itself. When
logic is slack, it matters how rigorous method is; and, in the
little

new stylistics as well as in much of the old, the rigor continues to be

misplaced, all of it being invested in the cranking of the machine
and none in its casements and connections.
Before turning to the other essays in the volume, but with
reference to them, I should like to make some further observations
on Freeman’s hypothesis and the analytic method that follows from
it, the general form of which could be stated as follows: something
in a literary work that is more or less manifest or “surface,” S, bears
some relation, R, to something else that is more or less obscure,
inner, central or “deep,” X\ therefore, by analyzing S, one may
discover X. Or, represented as a formula:

S(K)

160
 SURFACING FROM THE DEEP

This, with some variations, constitutes the central argument of

almost every paper in the volume. In Freeman’s paper, S, the
surface phenomenon, is “syntactic strategy’’; in other papers, it is

“syntactic patterning’’ or “phonological schemata’’ or “the lan-

guage’’ of the poem or its “form’’ or “the observable patterns in a
literary work’’ or simply “the text.’’ X, that which is deep and
which, in Freeman’s paper, was the poem’s “inner form,’’ is else-

where “the kernel theme’’ of a novel or story, or the work’s “literary

significance’’ or its “content’’ or “meaning.’’ The relation, R, can
be, as in Freeman, that S “reflects’’ X or, as in other papers, that it

“mediates,’’ “mimes,’’ “encodes,’’ is a “realization’’ or “concretiza-

tion’’ or “actualization’’ of it, or is “generated’’ by it, or “expresses’’
it.

There is, of course, nothing surprising in the formula itself: it is a

paradigm of many forms of analysis, from chemical to Freudian.
Moreover, it is familiar from more traditional forms of literary
analysis, particularly those performed over the past few decades by
the legatees of the New Criticism, who commonly operated on the
assumption that the “themes,’’ “meanings,’’ and “literary signif-

icances’’ of a work were “expressed’’ or “reflected’’ by its “form’’ or

“language.’’ In fact, the only thing that is surprising about the for-
mula how familiar it is, though the familiarity has been obscured
is

by the new values given to its variables or, in some cases, by the new
terms given to the old values. Thus, where S might, some years ago,
have been image clusters or symbolism, it is now almost exclusively
syntax. Or where it might, then, have been “sound patterns,’’ it is
now “phonological schemata,’’ representing here, I think, not so
much the greater precision of the analyst as his bewitchment by the
sound pattern of those polysyllables. Of course, “encodes’’ is not
quite the same thing as “expresses’’; but perhaps it is not always
altogether different either.
If, as I have been suggesting, there is discernible in the sounds of
the new stylistics the melody of an old song, two conclusions might
be drawn, depending on the degree of one’s attachment to that song.
One is that, since there is evidently nothing new under the sun, we
might as well go back to bed. The other is that what is affectionately
referred to as the form-content dichotomy continues, like the fre-

161
 LINGUISTICS AND LITERARY THEORY

quently reglued crack in an old teacup, to appear under pressure,

and that its prominence among the essays in this volume is at least

one indication of the weakness of the approaches they represent.

The issue itself is by no means avoided. Fowler remarks in his
introductory comments on the second paper: “Epstein boldly con-
fronts the age-old problem of the variability of literary form in

relation to content. He assumes, as I believe one must, that it is

possible to ‘say the same thing’ ‘in different words’ ’’

(8). Must one
assume it? I think not, though I also think the alternative should
consist of more than the pious incantation of its denial or the
repetition of credos concerning organic form or the indissolubility of
form and content. What is needed, rather, is a richer conception of
language than the one on which most criticism currently draws: a
conception that does justice to the actual dynamics of verbal be-
havior generally, that permits a precise formulation of the various
relations which, under various circumstances, obtain between the
formal features of an utterance or text and whatever else (its causes
or consequences, “meanings’’ or “effects’’) may be of interest to us,
and that indicates where useful distinctions, in those regards, may
be drawn among classes of texts or utterances, for example, “literary’’

and “nonliterary.’’ Such a conception might, among other things,

lead us to clarify means to “say the same thing’’ and explain
what it

under what circumstances we can or cannot say it “in different

words.’’ As long as we remain vague about the nature and identity of
both that “thing” and its relation to the linguistic form of a text or
utterance, no matter how precise or refined our methods of analyz-
ing the syntactic, prosodic, or lexical features of a work, the
analyses will remain pointless. Moreover, naming the thing “inner
form” or “deep structure” and replacing “expresses” with “en-
codes” or “actualizes” simply obscures the vacuity of the exercise:
the crack is still in the cup, the tea is still leaking into our laps.

II. Sound AND nonSense

The sort of mischief it creates can be seen in the second paper in the
volume, E. L. Epstein’s “The Self-Retlexive Artefact: The Function
of Mimesis in an Approach to a Theory of Value for Literature,”
which Fowler had praised for boldly confronting that “age-old

162
 SURFACING FROM THE DEEP

problem.” Toward the end of his paper, Epstein claims to have

presented a method whereby judgments of poetic value, ‘‘now made
on an intuitive basis,” can henceforward be arrived at ‘‘by a com-
paratively objective procedure” (74). Encountered at its beginning,
the heroism of such a claim would have induced immediate feelings
of pity and terror: nothing but folly could be expected to support it

and swift defeat is written in its face. Little but folly does support it,

and if the defeat is not as swift or total as one might have expected,
it is because the claim itself is so volatile that when, by the end of the
article, it is actually made, it has already dispersed into thin air.

The method is derived from an assumed ‘‘model of language-

production” that is, to my knowledge, unique:

... a speaker or writer first constructs a lexical constella-

tion which mimes a state of affairs; this constellation is then
realized in linear and segmental form syntactically, and
then either phonologically or graphemically. This realiza-
tion may be produced automatically, that is, with no prin-
ciple of selection operating among its linear elements other
than the style of the speaker or writer (and that style
operating outside of awareness), in which case the final
speech-act is casual prose. On the other hand, there may be
conscious or quasi-conscious selection and arrangement of
syntactic and phonological linear elements of form, in

which case a ‘‘poetic function” is operating (40-41, italics

in text).

Epstein’s conception of the sequence of stages we follow in speaking

or writing certainly gives one pause, as do the grounds of the
distinction he draws between ‘‘casual prose” and ‘‘poetic function,”

and his peculiar use of the term speech-act. More significant, how-
ever, is his invocation of mimesis, a term which (with mime) is

transformed here and throughout the article into an instrument of

tendentious obfuscation. Epstein can make the curious statement
that ‘‘lexical constellates” mime ‘‘states of affairs” not, as one
might otherwise have suspected, because he holds to a primitive
conception of words as images of things, but because mime is for

him an all-purpose term that takes on whatever value is required to

relate anything to anything else. Thus, when, in a particular analy-

163
 LINGUISTICS AND LITERARY THEORY

sis, A is said to mime B, it is glossed or paraphrased variously as A

“echoing,” “corresponding to,” being “analogous to,” “reinforc-
ing,” “conveying,” “expressing,” and, to be sure, “imitating” B.
As we have seen, Epstein is not unique among stylisticians in

playing fast and loose with key terms. Rather than linger over the
point, however, we may proceed directly to a consideration of the
method itself, which works as follows: On the assumption that the
value of a poem is (or seems generally thought to be, or is felt to be

“in this stage of history”) directly proportional to the extent to

which “the elements of its form” (specifically its “phonological and
syntactic schemata”) mime (or are “reflexive of” or “express” or

“correspond to”) its “content” (or “lexical constellate” or “the state

of affairs” mimed by that constellate), and that “the highest
grades” are assigned to works in which the most intensely “sub-
jective” (“personal” and “emotional”) content is most closely and
thoroughly mimed by all its schemata, we may, by examining the
degree of subjectivity in a poem’s content and the degree of corre-
spondence between that content and the poem’s form, arrive at its

poetic value. ^ In short, the sound (and syntax) must be an echo to

the (preferably subjective) sense.
A full-dress demonstration of the method is offered for Blake’s
“ I'he 7'yger,” “a work universally acknowledged as great” (60).

Observing that all the “phonological structuring” in the poem is

“non-mimetic” (for example, there is, in its “commonplace metrical

and rhyme-schemes,” “nothing that corresponds to claws, stripes,

fire, roars, teeth, ferocity”), Epstein states: “What Blake achieves

in lyf^er [s/c] is a subjective mimesis, based on syntax, for a very
high degree of value. It seems obvious from the poem that what
Blake is conveying is his own awe at a complete mystical perception
of the energy that drives the universe and the poet, a force here
symbolized in a tiger, a power beyond good and evil” (61-62). This

much being “obvious from the poem,” it remains for Epstein to

provide a “close analysis of its syntax,” which “reveals complex
mimetic schemata which reinforce and convey this subjective state to

the reader” (60).

Epstein’s analysis of the poem, which Fowler characterizes in the
introduction as “a sharply focused description” of the “significant

164
 SURFACING FROM THE DEEP

features” of its ‘‘verbal particulars,” selected ‘‘according to clearly

articulated criteria of rhetorical significance” and ‘‘rendered in the

reliable vocabulary of linguistic science” (8), occupies fifteen pages

and is a tissue of absurdities. To be sure, it is rendered in ‘‘the

reliable vocabulary of linguistic science,” which, for Epstein, means

a fatal attraction to the jargon of every literary and linguistic
theorist from Aristotle to J. L. Austin. Nevertheless, the analysis is

shot through with circularity and arbitrariness, and what would,

under less scientific auspices, have been called the ‘‘reading” of the
poem is marked by the belaboring of the obvious, heavy-handed-
ness, literal-mindedness, and sheer silliness — as illustrated in the
passages already quoted and in Epstein’s summary:

In the description of some structures in Tyger we have seen

how two subtle, complex, and apparently contradictory
emotional states, complete knowledge and questioning awe,
are mimed, reinforced, and finally reconciled by the satis-
faction and frustration of syntactic expectations in the
reader. Not only does this entitle the poem to a very high
grade of value; it actually makes the poem possible. It is

difficult to see how just this highly subjective proportion of

differing elements, awe and illumination, could otherwise
have been conveyed (74).

Thus form and content, pried apart for analysis, are once more
announced as ultimately inseparable, and Fowler’s bold confronter
finds it hard to see how, after all, the same thing could have been
said in different words.
We recall, however, that the analysis is not itself the object of the
article but only a demonstration of a method, though what it is a
method for remains an open question. Specifically, it becomes
increasingly unclear whether Epstein’s comparatively objective pro-
cedure is designed to determine value or to explain value judgments.
Thus, when he states, ‘‘in this paper, the second type of speech-act
[that is, the one in which ‘‘a ‘poetic function’ is operating”] will be
closely examined and analyzed, to [among other things] . . . suggest
a reason for formal mimesis of content as part of a criterion of value
for literature” (41), the angularity of the last clause leaves one in
doubt whether he plans to suggest a reason for making mimesis a

165
 —
LINGUISTICS AND LITERARY THEORY

criterion for value or the fact that (and/or a reason why?) mimesis is

a criterion of value. Similarly, the conclusion characterizes the

approach as “having the virtue of accounting for certain judgments
of value” and, in the next sentence, as being “an [incomplete]
indicator of value” and, finally, as being a “standard of value”
which, however, “applies mainly to varieties of Renaissance and
post-Renaissance poetry from technologically advanced countries of
Western Europe and America” (74-75). It appears, therefore, either
that Epstein fails to appreciate the difference between a procedure
(and/or set of criteria) for determining value and a theory that
accounts for value judgments, or that he wants to have it both ways,
or that he does not quite want to have it either way.
There is a difference, of course, and it is a significant one for
literary theory. An objective procedure (even a “comparatively” ob-
jective one) for determining the value of individual poems is, I

think, about as wild a goose as an objective procedure for determin-

ing the meaning of individual poems more below). A
(of which,
theory of value judgments, however, though hardly a tame goose, is
both conceivable and perhaps worth chasing. Epstein does seem to
recognize fitfully that we cannot account for value judgments solely
by examining the objects that are valued. Hence the qualifications
“in this stage of history” or “at least in a post-Romantic age,” and
so on. While these allusions are, in themselves, preposterously
vague, they do reflect his acknowledgment of the existence of
cultural variables among the determinants of value judgments (what
Fowler refers to as his “generous cultural-historical premise” [9]). A
theory that did attempt to account for judgments of literary value,
however, would do more than make polite gestures toward cultural
and historical variability. It would begin with the recognition that
the sources of value in a literary work are numerous and variable
and that individual value judgments are a function of multiple and
subtle conditions. Neither “the culture” nor “the age” ever operates
as a monolithic determinant of value or value judgments: there are
many cultures within any “culture,” and we do not all occupy the
same “age” at any given historical moment. Moreover, individual
judgments are always to some extent responsive to social and situa-
tional contexts and — in this respect like individual interpretations

166
 SURFACING FROM THE DEEP

determined in part by the reader’s personal history and experience.

A theory of literary evaluation might be expected to attempt to
identify all such relevant conditions and to determine the sources of
variability and constancy among them —including, of course, what
may be the independently describable properties of literary works.
In doing so, it would presumably draw on the anthropology and
sociology of art and on cognitive, perceptual, and developmental
psychology — as well, no doubt, as on linguistics. Finally, it might be
expected to be connectable to a general theory of aesthetic evalua-
tion and, ultimately, to a general theory of evaluative behavior. It
would be an enterprise of some magnitude, then, and one not likely

to issue in the assignment of “grades” to individual works or in

cavalier allusions to other folk elsewhere. It would, in short, be an

enterprise that looked quite different from Epstein’s.

III. The Limits of Linguistic Determinacy

Fowler’s charitable characterizations of Freeman’s and Epstein’s
papers are edged with a strong tone of defensiveness, perceptible
elsewhere in the introduction and in Fowler’s own paper. It seems
clear that, after twenty years of large claims and minimal achieve-
ments, stylistics has sustained some telling blows —delivered by,
now, two generations of critics and theorists, not all of whom could
be dismissed as merely expressing reactionary sentiments or making
“the standard objections to all professedly empiricist versions of
induction in science” (81). The earlier tones of arrogance are now
muffled, bold territorial claims are replaced by modest offers of
assistance and collaboration, former errors are acknowledged and
abjured, former standard-bearers are ruthlessly purged, and the
truce is cemented by a marriage —or perhaps it is that overtures of
alliance are made toward a bordering tribe that seems to be enjoying
greater prosperity. In any case, the marriage or alliance is repre-
sented here in the volume’s and Structure. ) and in the
title {Style . .

yoking, within its pages, of three more or less conventionally

“stylistic” papers (all on poetry) and four more or less “struc-

turalist” papers (all on narrative). That the yoking is an uneasy and

probably unstable one is suggested by Fowler’s own characterization
of the relation between the two groups of papers, a characterization

167
 LINGUISTICS AND LITERARY THEORY

that distorts both and, I think, to some extent misunderstands the

nature of the second, a point to which I shall return later.
In his own contribution to the volume (“Language and the
Reader: Shakespeare’s Sonnet 73’’), Fowler joins party with those
who have attacked the old stylistics for its “positivism,’’ for its
“presentation of poems as spatial (rather than temporal) and static
(rather than engagingly kinetic) constructs’’ (89), and for its con-
centration on tallying up the “linguistic features’’ of a text without
reference to any reader’s potential or even possible experience of
them (83). At the same time, he wishes to maintain some integrity of

identity for stylistics and some rationale for its procedures, which he
had described in the introduction as “assailing [the problems of
literary criticism] with the equipment of an established battery of
language sciences’’ (5). Thus, although he sees Stanley Fish’s
saturation bombing of stylistics^ as “chiming” with some of his own
“complaints” (89), he finds Fish’s own literary analyses (in, for
example. Self- Consuming Artifacts [Berkeley, 1972], deficient be-

cause, among other reasons, they do not, “where they involve

syntactic considerations, make any specific use of the delicate
metalanguage of syntactic description which has been developed by
Chomsky and by his recent successors” (90). But, more generally,
with respect to Fish and others who show proper regard for the
reader’s experience and for its “linearity” and “temporality,”
Fowler’s central objection is that they “provide no guarantee that in
following their technique we do not absolutely throw caution to the
winds” (92-93). His own techniques, as demonstrated in his analysis

of Sonnet 73, are designed “[to avoid] the limitations of all the
above approaches,” (93), that is, those of the old stylistics and of its

recently developed alternatives.

Independent of the argument in which it is imbedded, Fowler’s
commentary on Sonnet 73 (“That time of year thou may’st in me
behold . . .”) is in many ways admirable: the reading he offers is

rich and subtle, attentive to the poem’s formal qualities and often
illuminating in its description of them. These are substantial virtues,
but they are also the virtues of literary criticism as practiced by
sophisticated and sensitive readers. For the commentary to qualify
as stylistics, it must offer something more or other than such virtues.

168
 SURFACING FROM THE DEEP

and its problems lie precisely in what else it offers: namely, the sort
of guarantee earlier found lacking in traditional literary criticism
and even in “affective stylistics,” a guarantee not only that caution
has not been thrown to the winds, not only that the interpretation is

more than “just any old response” (102), not even only that it is

"'controlled by the verbal structure of the poem'" (102, italics in

text), but that it is uniquely so controlled and controlled by nothing
else. In spite of many disclaimers of the sort of spurious “objectiv-
ity” and “mechanical discovery procedures” associated with the old
stylistics and some gestures of acknowledgment toward “the license
of interpretive insight allowed to criticism” (120) and "sets of
permissible readings” (102, my italics), Fowler makes it clear that
he believes that the interpretation of Sonnet 73 that he offers
necessarily follows from his analysis of its linguistic features. But of
course it does no such thing, nor could it have; for no analytic
method can produce an interpretation and none can validate one.
Fowler’s reading of the sonnet is extensive (almost thirty pages long)
and thorough; it is not, however, “exhaustive.” It is also plausible,
but no more plausible than a number of others that he had earlier
dismissed as not sufficiently controlled by the syntactic structure of
the sonnet. Most significantly, it is no less vulnerable to charges of
interpretive circularity than other readings he had derided for
“assumfing] that the poem’s meaning is clear, known in advance”
(102).
Briefly: after presenting a summary of how “commentators
usually regard this sonnet” (that is, “as a dramatization of the
horror of mortality; the poet reflects on his own imminent death,
prettily figures it, and entreats his lover to love him the more for the

inevitability; thereby the general reader is drawn into a depressing

recognition of the fact of human transience” [102-3]), and remark-

ing the tendency of those commentators to be distracted by the
poem’s formal symmetries and “spectacular” metaphoric structure
and their “surprising” failure to consider the syntax more closely,

Fowler develops his own reading of the poem, a reading which, as he

had put it in the introduction, “shows how the linear organization of
the syntax directs the reading process in the retrieval of meaning”
(10). The upshot of the difference is that whereas other critics.

169
 LINGUISTICS AND LITERARY THEORY

preoccupied with what the poem ostensibly “states” and “figures,”

read it as death-anticipating and, in the couplet (“This thou
perceiv’st, which makes thy love more strong”), love- entreating,

Fowler, by paying close attention to “the lexical constituents which

perform the figuring” (116), demonstrates that it is lingering-
vitality-affirming and love- ascribing. The fact is, however, that there
is considerably more variability (not to mention subtlety) among the
interpretations offered by those other commentators than Fowler’s
oversimplified, distorted, and debased summary versions of their
observations suggest and that his own interpretation constitutes
neither a correction of their misemphases nor a supplying of their
failures of attention, but simply another reading — and one that, for

the points at issue, does not especially commend itself.

I shall not attempt here to pursue the details of Fowler’s lengthy

commentary on the The nature of his difficulty in making
sonnet.
good his claims, however, may be indicated by reference to what
becomes a rather striking feature of his own style, namely the high
frequency of self-qualifying locutions and, in conjunction with them,
the recurrence of the formula S {R) X, where S is, as usual, a
“surface” linguistic feature and X is the ascribed meaning, but here
R, the relational term, is a word or phrase of exceptional vagueness.
For example, S “might suggest” X (111); S “may well carry a
secondary meaning of” X (111); “it is fair to suppose that” S
“embraces” X (111); S “manages to connote” X (112); S “un-
expectedly delivers connotations of” X (1 14); S “bears witness to” X
(115); “there is a delicate suggestion of” A" in 5" (116).
Now, there is, to my mind, nothing objectionable in Fowler’s or

any other critic’s using phrases such as “might suggest” or “seems

to connote” in referring to the relation between a linguistic feature
and an ascribed meaning. On the contrary, such phrases are, I

think, precisely descriptive. What is notable about their recurrence

in Fowler’s paper is that, if such expressions are not reflections of his
timidity or evasiveness, they can only reflect his implicit apprecia-
tion of what he explicitly or otherwise refuses to appreciate: namely,

that there is a significant aspect of meaning in the poem that is

necessarily variable, irreducibly indeterminate, and, therefore, that

although “the verbal structure” of the poem may direct one’s

170
 SURFACING FROM THE DEEP

experience and interpretation of it, that structure cannot “control,”

in the sense of unequivocally determine, either of them.
One may grant that, by virtue of the fact that a poem is a verbal
structure, its meanings are presumably inferrable (and intended to
be inferred) in accord with linguistic conventions, that we cannot
“understand” the poem or read it “properly” unless we know and
respect those conventions, that questions of meaning are properly
settled by appeal to them and, where the relevant conventions are
themselves in question, that philologists, cultural historians, and
linguists may be called on to identify, recover, or specify them. Not
all meanings are equally susceptible to coding, however, and some
remain elusive and hazardous even in ordinary nonliterary verbal

transactions: for example, those which arise from subtle features of

the particular circumstantial and psychological context of an
utterance, which might include the speaker’s frame of mind or exact
motives in speaking and his relationship to and attitudes toward his
listener. (“He was awfully abrupt, but I’m not sure if he was angry
or just distracted.” “Don’t you think that comment sounded
sarcastic? . . . disrespectful? . . . flirtatious?”) These aspects of
meaning (which to some extent overlap with what speech-act
theorists refer to as the “illocutionary force” of an utterance) are
what, with respect to a poem, we sometimes speak of as its “tone”;
and after the philologists, linguists, and cultural historians have
provided their expert glosses on every word and grammatical
construction in it, there is nothing that remains more resistant to
agreement among readers. For, of course, it is that particular
context that is, in a poem, least explicit and most open to surmise
or, indeed (as I have argued elsewhere), understood to be his-

torically indeterminate and thus, by convention, radically indeter-

minable, eternally open to constructive rather than reconstructive
interpretation.
It is clear, I think, that the aspect of meaning I have been
discussing here is precisely what Fowler refers to as “the super-

structure of connotations” (117) in Sonnet 73 and from his “analy-

sis” of which he derives his own particular angle on the tone of the
sonnet. As I have suggested, however, the analysis is no analysis, but

a set of construals, and the superstructure is indeed a superstruc-

171
 ”

LINGUISTICS AND LITERARY THEORY

ture, but one constructed by the reader (that is, each reader) upon
the poern. The particular superstructure that any reader (including
Fowler) constructs will reflect the particular choices he has made
from among the numerous possible “connotations” of the poem’s
“linguistic features” and also the particular way he organizes and
integrates them; and his choices will reflect who he is and everything
he knows, not only about words but also about the worlds in which
and out of which they are spoken. He may, of course, be absolutely
convinced that his reading has been “controlled by the verbal
structure” of the poem, for although he has made choices, he is
usually not aware of having chosen them from among alternatives.
There is perhaps no illusion so compelling as the sense one
sometimes has of not only the propriety but the necessity of one’s
own construction;

“But look! It’s right there on the page

—
“Whereon do you look?”
“Do you see nothing there?”
“Nothing at all; yet all that is I see.”

Some note should be taken here of the set of notions associated

with “the ideal reader,” a creature who figures largely in Fowler’s
article and to some extent in a number of others in the volume. As
we have seen, Fowler, though associating himself with Stanley Fish’s
demolition of the old stylistics, cannot accept Fish’s own alternative,
“affective stylistics.” One reason is that, although Fish’s “appeal to
the analogy of generative linguistics” seems to be “a useful point of
departure,” Fish does not take the analogy seriously enough:

... we are invited to witness the reader responding se-

quentially to language in literature, but the reader cannot
be kept in decent communal order because he is not
controlled by the constraints of precise communal syntax.
The reader is not the “ideal reader” we might expect in a
generative stylistics based on generative grammar, but a less
organized, regrettably substantial, actual reader (90-91).

Similarly, incommenting on the work of Stephen Booth, who

attempts to show how the reader constructs interpretive frameworks
for responding to Shakespeare’s sonnets, Fowler finds “indecisive

172
 SURFACING FROM THE DEEP

and evasive” Booth’s imperturbability in the face of “incompatible,”

“variable,” and “switching” interpretations, and he is forced to
question the qualifications of Booth’s blithely unqualified ''[the]

reader”: “How can we be sure that the reader Booth employs to lead
us through the sonnets is the right kind of reader” (92)? By contrast,
the “interpretively significant formal structures” that Fowler
analyzes in his own reading of Sonnet 73 are “validated in the ‘ideal
reader’s’ experience because they reflect culturally coded knowledge
activated in the process of reading” (93).
Of all the courts of appeal ever proposed for the ultimate
validation of an interpretation, none could be more sublimely
imposing than that presided over by this ideal reader. Who can
gainsay his decisions? Who could accuse him of fallibility, sub-
jectivity, or partiality? Above all mortal battles and free of all mortal
limits, all natural or historical contingencies, he simply reads and
knows. Have mortal readers been at odds for years on the reading of
a line? a passage? a poem? Bring them to court: the ideal reader will
settle the case, cracking the message by consulting the cultural code.
But, it may be asked, how will the ideal reader make his knowledge
known to us? Since his sole mode of being consists of reading and
knowing, how can he express himself? The answer is evident:
through his mediator, the real but totally self-effacing critic. Only
the ideal reader knows for sure, but only the critic knows for sure

what the ideal reader knows.

Put briefly, my point here is that, as invoked by Fowler, the ideal
reader cannot rescue us from the hermeneutic circle because he
constitutes merely another loop in it. I shall return to the subject

below, in connection with another article in the volume by Jonathan

Culler. While we still have Fowler’s article before us, however, I

should like to comment on his enthusiastic association of himself

with critical positions and analytic techniques that emphasize the

“linearity” and “temporality” of “the reading experience.” It is

certainly important to recognize that reading is a process and that it

occurs in time. Fowler, however (and some other critics and theorists
whom he cites, and some whom he doesn’t), goes far toward
transforming that recognition into a new critical dogma, attended by
a new set of critical pieties. We do, of course, usually read the words

173
 LINGUISTICS AND LITERARY THEORY

of a text sequentially, but (1) we normally scan as we read, taking in

(and anticipating) more than one word at a time; (2) our expecta-
tions and recollections are organized not only linearly but also
hierarchically, comprehending and subsuming simultaneously a
number of structural patterns of various unit sizes, for example,
from phoneme to canto, from morpheme to total semantic field; (3)
—
we continuously modify our impressions recollections as well as
expectations —
as we read; (4) we often reread a text, sometimes
immediately, especially such brief texts as sonnets; and (5) we
always perceive a text spatially as well as temporally and, in verse,
the specific array of words on the page usually contains information
that affects our expectations and recollections. In other words, we
do not perceive or experience texts as “strings” of words or the way
one pulls in wash off the laundry line; “temporality” and “linearity”
do not control the reader’s experience with the rigidity that Fowler et

al. sometimes suggest, nor is the reader locked once and for all into
a given linear sequence.

IV. Structuralism and the Linguistic Model

In the introduction to the volume, Fowler takes explicit note of the
two approaches to literary study reflected in its title and represented
in its pages by, on the one hand, the three papers already discussed
here and, on the other, the four that follow. His characterization
and contrast of the stylistic and structuralist papers is even-handed
and he evidently regards the two approaches as “complementary”
(10): both “can be attached to a basically Saussurean model of
language” and both “assail [problems in literary criticism] with the
equipment of an established battery of language sciences”: while
one “entails close attention to the surface structure of literary texts,”
the other investigates “deeper, more abstract patterns”; while one
“emphasizes particularity, individuality, concreteness,” the other
“ismore given to generalization and abstraction”; and while one
makes the individual work “more recognizable,” the other makes it
“less recognizable” (10-11). As that last apportionment suggests,
the evenhandedness is sometimes a bit strained, but Fowler is

certainly generous and sympathetic in his remarks on the struc-

turalist papers — to the point, indeed, where he suppresses what

174
 SURFACING FROM THE DEEP

some structuralists themselves are often at pains to make clear,

namely that the two approaches are not complementary but, at least
in their own view, something closer to antagonistic.
The nature of the antagonism and the degree of its sharpness can
be indicated by quoting some observations made elsewhere by one of
the volume’s contributors, Jonathan Culler:

The type of literary study which structuralism helps one to

envisage would not be primarily interpretation .... Rather
than a criticism which discovers or assigns meanings, it

would be a poetics which strives to define the conditions of

meaning. The study of literature, as opposed to the
. . .

perusal or discussion of individual works, would become an

attempt to understand the conventions which make litera-
ture possible (viii) .... Granting precedence to the task of
formulating a theory of literary competence and relegating
critical interpretation to a secondary role, [structuralism]
leads one to reformulate as conventions of literature and
operations of reading what others may think of as facts
about literary texts.’

As I shall explain below. Culler’s views of how literary study should

be pursued are themselves not altogether unproblematic. For the
moment, however, we may note (1) that he sees structuralism not
merely as another “approach to literature’’ but as a new definition
of the total enterprise of literary study; (2) that the enterprise is con-
ceived of as comprehensive rather than atomistic: it is literature as

a whole (or, as the structuralists often put it, “as a system’’) rather
than individual works, even the total corpus of literary works, that is

the subject of study; and (3) that he views the objective of literary
study not as the elucidation, evaluation, or even description of
individual texts, but as the explanation of literary phenomena
generally. Structuralism, in other words, is not a new(er) criticism; it

does not offer to do better (more “objectively,’’ “rigorously,’’ or

“solidly’’) what literary critics attempt to do; it offers to do
something else. Moreover, unlike literary criticism and the new
stylistics, it is not essentially pedagogic in aim or justification: the
“understanding” it seeks would be valued, not as ancillary, but as
such.

175
 LINGUISTICS AND LITERARY THEORY

So characterized, there much is in the structuralist program that

commends itself. Moreover, much of the work done in its name is, I

think, engaging and impressive. I believe, however, that the pro-

*

gram generally and the individual studies pursued in furtherance of

it are, at present, in a state of captivity — specifically, captivity to its

own intellectual origins — and that it cannot truly prosper until it

manages to shake itself loose from the egg that hatched it.

The egg, of course, is linguistics,® and the nature and consequences

of the captivity happen to be illustrated fairly dramatically in Culler’s
own contribution to the volume, an article that manages, with great
intelligence and elegance, to draw a set of interesting and valuable
conclusions from a set of highly questionable premises and distinctly
strained arguments.
The immediate aim of the article is indicated by its title,

“Defining Narrative Units”; its broader aim is to contribute to the

development of a general theory of narrative structure. In view of the
number of such theories recently proposed. Culler observes that it

is necessary “[to think] seriously about the criteria to be used in

evaluating competing approaches and hence about the goals of an
analysis of plot structure” (125). After considering and rejecting the
grounds on which other theories have been promoted —for example,
that they provide a “metalanguage” which “enables the analyst to
describe a range of stories” or that “there is evidence about the
general validity of the presuppositions on which [the] theory might
be thought to be based” (126) — he proposes his own set of criteria:

. . . the only was to demonstrate the superiority of a theory

of plot structure is to show that the descriptions of parti-
cular stories which it permits correspond with our intuitive
sense of and that it is sufficiently precise to prohibit
its plot
descriptions which are manifestly wrong. ... In short,
competing theories of plot structure can only be evaluated
by their success in serving as models of a particular aspect
of literary competence: readers’ abilities to recognize and
summarize plots, to group together similar plots, etc. This
intuitive knowledge constitutes the facts to be explained,
and without this knowledge, which we display every time we

176
 SURFACING FROM THE DEEP

recount or discuss a plot, there is simply no such subject

as the analysis of plot structure because there is nothing
for the analyst to be right or wrong about (127).

To illustrate these points, Culler presents three plot summaries of

Joyce’s story “Eveline,” ranging from a single sentence to a brief
paragraph; there would, he maintains, “be substantial agreement
among readers about what ought to be included at a particular level
of generality” (128). There is, he concludes, an “implicit knowledge
which enables the reader to process the text,” that is, “rules and
procedures he is unconsciously following when he picks out the
crucial items in the plot”; and these “intuitions of the reader” are
what “a theory of plot structure ought to explicate” (130).
Culler then discusses and evaluates a number of competing
theories of narrative structure proposed by, among others, L^vi-
Strauss, Greimas, Todorov, and Barthes. He concludes the article
by formulating the outline of a theory of plot structure which, while
drawing on and combining key concepts that (by one set of terms or
another) had appeared in each of the theories discussed, neverthe-
less satisfies better than any of them those criteria he had initially

established.
I shall return below to Culler’s final formulation. First, however, I

should like to examine more closely the process by which he arrived

at it, particularly the origins of that set of criteria. As the quoted
passages suggest. Culler uses a number of terms
— “competence,”
“intuitiveknowledge,” and so on — that are familiar from trans-
formational-generative linguistics. These reflect more than termin-
ological contagion, however: they are the consequence of Culler’s
quite explicit conviction that if literary theory is to become a fruitful
discipline, it must model itself after “the theory of language”; and
the borrowing of terms is only a small indication of the thoroughness
and consistency with which Culler does in fact hew his work in the

image of Chomsky’s. The sources and implications of this conviction

are detailed in his Structuralist Poetics, the following passage from
which is a central statement of the reasoning behind it:

. . . since literature itself is a system of signs and in this

respect like a language, one postulates a poetics which

177
 —
LINGUISTICS AND LITERARY THEORY

would Study literature as linguistics studies language, taking

its cue from linguistics wherever that seemed possible.’

To one who is not a true believer, this statement is striking for the
number of questions that it begs and the flagrancy of the non
sequitur at its center. Is literature a system of signs? Is a language a
system of signs? If so, and they are in that respect alike, does the
inference still follow? Specifically, are the differences between
literature-as-a-whole and a language of no significance for the
nature of the discipline that studies each? Is linguistics a monolithic
discipline? Specifically, is it equivalent to transformational-genera-
tive grammar? If not, is Chomsky’s theory of language and the
transformational-generative grammarians’ pursuit of its study the
only and/or the most suitable model for the theory of literature and
the pursuit of its study? And, in any case, are the assumptions,
procedures, concepts, and conclusions of linguistics themselves so
well established, so free from internal problems or external criticism
that literary theorists are well advised to adopt and apply them
unreflectingly?'®
Since a good case could be made for answering every one of these
questions no, the premises of the argument are at least dubious.
Moreover, even if one would answer some of them yes, the argument
itself remains illogical. For, in accord with such reasoning, one
could just as easily begin with a different characterization of
literature, but one of presumptively equal validity — say, as a form of
art or form of human behavior —and go on to observe that it is “in
this respect like’’ any number of other things — say, painting or
sex — and conclude that the study of literature should “take its cue’’
from some method by which one of those other things was studied
say, iconography or the investigation of animal sexuality. Or, if one
does begin by characterizing literature as a system of signs, one
could just as easily, and rather more logically, conclude (as, in fact,

some other theorists have concluded) that the study of literature is

properly the domain of the general study of sign systems —that is,

semiotics or communications theory —where it would take its place

beside the study of its sister sign systems, including natural
languages.
One might expect some fairly peculiar things to follow from such

178
 SURFACING FROM THE DEEP

premises and conclusions, as indeed they do in the hands of other

theorists similarly captivated. As it turns out, however, they serve
Culler surprisingly well for, like the basically honest but happily
inept character in a comic “caper” movie, he almost always
manages to pull the job off anyway or at least to end up with
something of value to show for his efforts. He does this, I think,

partly by either strategically or unwittingly misunderstanding

Chomsky, even when he literally parrots his words, and partly by
tailing to produce exact counterparts in the analogies he draws from
linguistics to literary theory. The cumulative effect of the individual
discrepancies is quite substantial and have permitted Culler to make
some strikingly original and fruitful contributions to literary theory
in spite of his unpromising convictions. As is evident in the present
article, however, the cost of those convictions is the expenditure of a
certain amount of wayward energy.
For example, alluding to the alleged facts that, when asked, we all

produce comparable plot summaries and that we agree about what

should be included in a summary at any level of generality. Culler
writes:

These examples illustrate the kinds of facts that a theory of

plot structure ought to explicate. A theory is “descriptively
adequate,” to use Chomsky’s phrase, “to the extent that it
correctly describes the intrinsic competence of the idealized
native speaker.” The structural descriptions which it assigns
to a text must correspond to the intuitions of the reader, or
to put it the other way around, the theory must account for
the intuitions of the reader by providing a model of the
competence which enables him to perceive structure (130,
italics mine).

The angularity of the italicized passages is explained by their source,

a passage in Chomsky’s Aspects of the Theory of Syntax, which
it will be useful to have before us:

A grammar can be regarded as a theory of language: it is

descriptively adequate to the extent that it correctly de-
scribes the intrinsic competence of the idealized native
speaker. The structural descriptions assigned to sentences
by the grammar, the distinctions it makes between well-

179
 LINGUISTICS AND LITERARY THEORY

formed and deviant, and so on, must, for descriptive

adequacy, correspond to the linguistic intuitions of the
native speaker. “ . . .

The nearly wholesale appropriation is obvious; what is less obvious

but more significant are the discrepancies, which is to say the
off-the-mark analogies.
1. The difficulty begins with the term theory itself: although
taken in the broad sense of an explanatory model or set of
explanatory statements about something, Chomsky’s own theory of
language is quite far-reaching, embracing everything from chil-

dren’s acquisition of language to linguistic universals and what they

allegedly reveal about the human mind. Chomsky also uses the
phrase “theory of language’’ in a much more restricted sense, as
here, to refer specifically to a generative grammar. It is clear that
Culler has taken Chomsky’s perhaps reasonable criteria for an
adequate grammar of a language as his model in formulating
criteria for an adequate “theory of plot structure’’ (and, in Struc-

turalist Poetics, for a general theory of literature). A language,

however, is a very different sort and order of thing from plot
structures, and there is no good reason to assume a priori that the

specific criteria for a grammar of one would be at all relevant to a

theory of the other. It is not surprising, then, that Culler, having
adopted Chomsky’s criteria for evaluating grammars, is obliged to
take a very circuitous route toward his own very different destina-
tion. Thus, whereas it makes sense for Chomsky to invoke the
intuitions of the native speaker regarding well-formed and deviant
sentences as both the facts to be accounted for by a generative
grammar and as validating the rules it specifies (for it is the
specification of those rules that, by definition, constitutes the
grammar of a language), it makes little sense to extend that aspect
of the theory of syntax to a theory of narrative structure, and
Culler’s efforts to do so produce some of the most tortured
argumentation in the article.

2. What is referred to in Chomsky’s own work as the native

speaker’s “intuitions,’’ “tacit knowledge,’’ or “competence’’ is

understood to reflect a substantial component of literally innate

knowledge, more or less activated by a certain amount of pure

180
 SURFACING FROM THE DEEP

“exposure” to a particular language. Indeed, Chomsky is often at

some pains to emphasize {contra learning theorists) that language is

not, in his view, acquired by “learning” (that is, conditioned

generalizations and discriminations, and so on) and certainly not by
explicit instruction. For Culler, however, “intuitive” seems to bear
no implication of innateness, and literary competence is seen as the
product of a variety of cultural experiences, not excluding formal
education. In Structuralist Poetics, for example, he observes:

That achievement [that is, of literary competence] requires

acquaintance with a range of literature and in many cases
some form of guidance. The time and effort devoted to
literary education by generations of students and teachers
creates a strong presumption that there is something to be
learned .... [In studying literature] one gains not only
points of comparison but a sense of how to read.‘^

One need have no quarrel with Culler’s observations here, as such.

On the contrary, one may observe that whereas Chomsky’s own
appeal to innate mechanisms often has the consequence of closing
the doors to inquiry prematurely. Culler’s emphasis on the cultural
transmission of literary conventions permits an empirical definition
of literary competence and encourages an empirical investigation of
its nature, origins and the processes by which it is acquired.
Nevertheless, it is from Chomsky’s view of the relation between
linguistic competence and a speaker’s intuitions that Culler derives
his criteria for a theory of plot structure, particularly the very
troublesome suggestion that what such a theory must do is specify
the rules and procedures that readers intuitively follow in describing
and summarizing plots. We should note that if this principle were
extended, the theory of literature would end up accounting not for
the characteristic nature or functions of literary works or the
characteristic ways in which we experience them, but rather the
characteristic ways in which we have learned to talk about them.
That a theory of literature modeled after a theory of how we talk

should end up telling us only how we talk about literature is perhaps

not so surprising. Because Culler, however, really is interested in the
nature of literary texts and literary behavior, he eventually trans-

181
 LINGUISTICS AND LITERARY THEORY

forms the essentially misbegotten concept of literary “competence”

into an exceptionally useful instrument for investigating both.
3. Although Culler does not, as do Fowler and others, invoke “the
ideal reader” to validate interpretations of literary works, he does

conceive of him along distinctly Chomskian lines as analogous to

“the idealized native speaker.” One notes, however, that whereas
the native speaker is also presumably a native listener, the ideal

reader cannot be presumed to be also an ideal writer. Indeed, if the

analogy from linguistics to literature were strictly pursued here, it is

the writer (who would be not only also a reader but presumably a
highly competent one) whose intuitions about plots would have to be
accounted for. Moreover, since those intuitions would seem to be
most obviously “displayed” in the plots of actual literary narratives

actually constructed by actual writers (assuming that those con-

structed by less-than-ideal writers would not have survived as
“literature”), the theorist of plot structures would seem to be
justified in concerning himself directly with the structure of actual
plots rather than having to approach his subject by way of putative
readers’ putative intuitions invisibly manifested in putative text
processings or occasionally displayed in elicited plot summaries.
To be sure, there are good reasons why a theory of literary
narrative or any theory of literature should be particularly con-
cerned with the knowledge, expectations, and responses of readers,
and good reasons to maintain that any theory of narrative structure
that did not account for, or take into account, the ways that readers
experience narratives would be to that extent incomplete. None of
those reasons, however, are supplied by a conception of literary
theory derived from linguistics. All of them, in fact, derive directly
or indirectly from the diff erences between literature and language. I

would suggest, then, that Culler’s disanalogy here is perfectly

justified and that, in developing his theory of narrative structure, he
might have avoided some strained arguments and angular formula-
tions had he recognized that it was a disanalogy.
As it is. Culler’s insistence on locating both his “facts” and the
source of his theory’s validation in the reader’s intuitions saddles
that theory with an altogether gratuitous double burden that notably
impedes its forward motion. For it not only creates the problem of

182
 SURFACING FROM THE DEEP

how one is to identify those intuitions or locate their manifestations

(and there certainly seems to be something more vagrant and
variable about readers’ constructions and assessments of plot sum-
maries than there is about the speech of native speakers or their
assessments of the well-formedness of sentences), but also raises or
begs the questions of whose intuitions are being accounted for
and/or who has the competence to validate them. Culler’s example
of “Eveline” does little to persuade one of the generalization it is

designed to illustrate, namely the “substantial agreement” alleged

to obtain among readers regarding the propriety of plot summaries
(and, obviously, adding the qualification, *Hdeal readers,” would do
nothing but beg the question directly). The extent to which we
produce comparable plot summaries may have a good deal to do
with who “we” are and how we learned both what “plots” are and
also what it means to “summarize” them. It is doubtful, for
example, that a South American tribesman would produce the same
plot summary of one of his myths as would Claude Levi-Strauss. Or
to put it another way, wouldn’t our explaining to the tribesman (or,

of course, to a child) what we wanted when we asked for a “plot

summary” be the same as telling him how to “process” the story the

way “we” have learned to? In any case, the derivation of narrative
units and universals from the hypothesized plot summaries of
hypothetical ideal readers is justified only by a series of analogies in

which narratives are identified with sentences, plot summaries with

the intuitions that comprise literary competence, literary compe-
tence with linguistic competence, and readers with native speakers,
each of which is strained and the sum of which is highly dubious.
That Culler ultimately arrives at his destination, that is, a plausible
and fruitful theory of narrative, owes to the fact that his instincts

are better than his maps.

We may now consider Culler’s concluding observations. Following
Barthes,'^ he writes: “It is when the reader begins to place actions in

sequences, when he perceives teleologically organized structures,

that he begins to grasp the plot” (137). He goes on to postulate that
in “processing the text,” that is, in reading the story or summarizing
its plot, the reader’s perception of narrative units “is governed by his
desire to reach an ultimate summary in which the plot as a whole is

183
 LINGUISTICS AND LITERARY THEORY

grasped in a satisfying form”; and, following the suggestions of

Levi-Strauss, Greimas, Kristeva, and Todorov, he suggests that, at
its most fundamental level of generality, that form is always some-
thing like ‘‘the modification of a situation” or ‘‘a passage from one
situation to another.” Consequently, ‘‘the incidents of the plot must
be organized into two groups and these groups must be named in

such a way that they represent either an opposition ... or a logical

development. ...” Finally, ‘‘each of the two groups can in turn be
organized either as a series of actions with a common unifying factor
which serve as a name for the series or as a dialectical movement in

which incidents are related as contraries and named either by a

temporary synthesis or by a transcendent term which covers both
members of a contrast” (139, italics mine). Culler adds that a model
of this sort might prove more fruitful than the usual taxonomical
approaches to narrative structure because it ‘‘would lead us to think
seriously about the expectations of readers and their role in the
perception of plot structure” (140).
Aside from the curious emphasis here on naming (which evidently
reflects Culler’s conviction that the reader’s ‘‘processing” of a
narrative always issues in some more or less explicit verbal formula-
tion and/or that a theory of plot structures can only account for how
readers talk about plots), these hypotheses are plausible, their
formulations are precise enough to permit testing and modification,
and they are likely to be fruitful for literary theory for pretty much
the reasons Culler states: that is, they suggest the dynamic interplay
between the fundamental features or structure of literary narratives

and the fundamental features or structure of human perception and

cognition. (The interplay, I would add, operates in both directions:
we respond to narrative structures in a certain way because of
certain psychological predispositions, and so on, but also literary

narratives are constructed to conform to and gratify those pre-

dispositions.) In short, the theory of narrative with which Culler
emerges in this article has many recommend it. Validation
things to
from linguistic theory is, however, not among them. Indeed, one
cannot avoid remarking that his concluding observations and pro-
posals may be understood and defended quite independently of their

184
 SURFACING FROM THE DEEP

origins in the Chomskian model and, inspirational effects aside,

could have been formulated without that model.
I should like to make one further point about Culler’s conclu-
sions, namely that their generality may be both more limited and
broader than he indicates here: more limited in view of the fact that

there are obviously genres of fictional narrative (for example,

“postmodern” works by Beckett, Robbe-Grillet, Sarraute, and so
on) the structures of which cannot be so described and are not so
perceived; and broader in that many, perhaps most, nonfictional or
nonliterary narratives can also be so described and are also so
perceived. In fact, it may be that the innermost kernel of narrative
structure is simply “our” (Western? modern? . . . universally
human?) shared sense of what it means for something to have
happened, and therefore equivalent or related to the basic motive
and occasion for all [storyjtelling (= narrare). I shall return to and
expand this suggestion below.

V. Generating Theories OF Narrative

As we have seen, although Culler appropriates for a theory of
narrative Chomsky’s criteria for a theory of language, he does not,
as Chomsky sometimes does, equate “theory” with “generative
grammar” and thus does not call for or propose a grammar of
narratives. A number of other theorists, however, do just that,
among them the authors of the next two papers in the volume,
L. M. O’Toole and John Rutherford. O’Toole, in his “Analytic and

Synthetic Approaches to Narrative Structure: Sherlock Holmes and

‘The Sussex Vampire,’ ” observes:

Just as linguistics cannot limit itself long to the study of

syntax alone, but must constantly refer back to semantics,
so poetics . . . must analyze not merely the formal patterns
but what is being expressed by those patterns. . .
.
[0]ne
definition of the literary work of
might be that art it is the
syntactic expression of a deep semantic opposition (145).

Accordingly, and also in accord with the work of two contemporary

V V
Soviet structuralists, A. Zolkovskij and Y. Sceglov (cited, 148), one

185
 LINGUISTICS AND LITERARY THEORY

seeks to provide a “generative model” of narrative structure,

“mapping the ‘transformational history’ of narrative texts and
episodes much as the generative linguist maps the transformational
history of a sentence” (148).
He continues: “The starting point for such a process must be
some kind of ‘kernel’ in the deep structure of the work”; this kernel

may be called its “theme,” “a scientific abstraction, a formulation

in a more or less abstract ‘metalanguage’ of the irreducible meaning
of the text” (148) or, in the words of Zolkovskij and Sceglov, “that
invariant of which everything else in a work is a variation ...”
(quoted, 148-49). By “illustrating how elements of plot, character
and the object world are generated through various operations on
the theme,” we may “account adequately for many features of the
style of a work” (148-49).

The argument of O’Toole’s article can obviously be represented as

a variation of our formula

s (R) X,
where S = the surface features of the text (here plot, characteriza-
tion, and “object world” as well as more specifically verbal fea-
tures), X= the kernel theme, and R = “is generated by.” As the
title of the paper suggests, however, the analysis is here supple-
mented by a synthesis:

S (R) X from X ^
with X, the theme, arrrived at provisionally by “a combination of
analysis and intuition” and seen as “a hypothesis which can then be
tested through the process of synthesizing the text” (149).
O’Toole demonstrates the operation of the method by analyzing
and synthesizing Conan Doyle’s story, “The Sussex Vampire.”
Initial analysis suggesting that the kernel theme is “the triumph of
reason over the irrational,” this provisional statement of it is tested
systematically “against various structures in the work represented by
the labels: Fable, Plot, Narrative Structure, Point of View,
Character, Setting'' and found to hold good (150-51, italics in text).

That is, it is possible to see in each of these aspects of the

story —even, O’Toole notes, in the “oxymoron” of its title —that
186
 SURFACING FROM THE DEEP

“basic semantic opposition.” The synthesis of the story follows, in

pursuit of which the theme is further analyzed into a set of
antithetically related key factors (for example, adventure-security,
apparent-real) which are shown to be “realized” through various
“operations” (for example, combination, repetition, juxtaposition),
in numerous elements of the story, ranging from the array of
characters and their respective traits to the antithetical structure of
specific sentences.
The synthesis is, however, little more than an extension of the
analysis to other features of the story: the only difference is that the
second part of the analysis is explicitly directed toward demonstrat-
ing the tyranny of the kernel theme whereas, in the first part, it was
directed toward demonstrating the propriety of the identification of
the theme. And, of course, for all the fantasies of genetic or
biochemical engineering suggested by such terms, O’Toole does not
“generate” the text of “The Sussex Vampire” either. Obviously the
same kernel (that is, “the triumph of reason over the irrational”)
could have (and has) sprouted numerous other stories, and the
“mapping of the transformational history” of the story is what
would elsewhere be referred to as showing —methodically, to be
sure, and in great detail —how that theme is “reflected” (here,
“realized”; in Rutherford, “actualized” or “concretized”) in

various aspects of the story.

Nevertheless, the demonstration is in many ways impressive: the
analysis (whether in forward or reverse gear) is systematic and
thorough; and although the apparatus involves the familiar alge-
braic abstraction and notation, it is handled with some geniality and
tact. O’Toole’s own style is lucid and energetic, and he is certainly
an attentive reader. The analysis is, in fact, persuasive in its details

and quite intriguing to follow somewhat the same way as the text
(in

which is its subject and, perhaps, for some of the same reasons), and
one finishes it with the sense that something quite neat has been
accomplished.
What has been accomplished is the question. Without doubt
O’Toole has discovered what is, in effect, the formula for “The
Sussex Vampire.” One may suspect, however, that what he has
thereby demonstrated is the extent to which that story (which, as he

187
 LINGUISTICS AND LITERARY THEORY

had himself pointed out earlier, “is one of a vast series of stories of a

very highly stylized and conventionally structured type” [151]), was

written to formula. In other words, the neatness of the demonstra-
tion seems to be a function of the neatness of the material, and the
generality of both the method that it illustrates and the model of
narrative fiction that it assumes remain dubious.

John Rutherford’s article (“Story, Character, Setting and Narrative

Mode in El Amigo Manso") seems to be in some measure
designed to answer this objection — or, as Fowler remarks of the
article in his introduction, “the technique of structural analysis is

here demonstrated in relation to a full-length novel in the modern

European realistic tradition: an implicit rebuttal of the Anglo-Saxon
criticism that structuralist analysis ‘works’ only for exceptionally
schematic narratives (detective stories, fairy-tales) or exceptionally

short ones” (17).

Although the article is cluttered and its central argument some-
what fractured and diffuse, Rutherford’s major concern is to
describe a set of methods for analyzing narrative fictions, illustrat-
ing them through an examination of El Amigo Manso. In the course
of the examination, he draws explicitly on O’Toole’s study of “The
Sussex Vampire,” observing:

O’Toole has necessarily taken a simple example to illustrate

the theory he expounds; yet the same theory is applicable to
complete texts, in which there are more basic (or “deep”)
themes and hence many more possible permutations
(182-83).

And, a bit later, he remarks that “the same method of analysis that
was used for the simple characterization of ‘The Sussex Vampire’
can, without strain, be applied to rich narrative texts” (184). For
reasons I shall outline below, however, 1 believe that the application
not only strains the method but cracks it at its weakest point.
Like almost every other contribution to this volume, including the
editor’s introduction, Rutherford’s article is marked by an excep-
tional degree of methodological self-consciousness. Thus, although

188
 SURFACING FROM THE DEEP

the major portion of the article is an analysis of a particular novel,

and the immediate aim of that analysis is to illustrate a set of
analytic methods, Rutherford’s broader purpose is “to tackle the
fundamental and elementary . .
.
problem of the global segmenta-
tion of narrative texts’’ (178). The problem arises, he writes, because
“Even though the aim of structural or semiological poetics is the
comprehension of literature as a system the poetician can . .
. ,

hardly dispense with the analysis of [individual] texts as a basis for

his investigations’’; and although the problem of segmentation
(“What are the component parts of literary works?’’) is not great “at
the purely verbal or stylistic level,’’ nevertheless “narrative literature
has another dimension which some extent independent of
is to
language’’ (177). Aside, however, from the conviction that there is
nothing that is not the better for being examined by way of analogies
from linguistics, there is no reason to suppose that this other
dimension of narrative texts (
= “all that which can be para-
phrased’’ [177]) has “segments’’ corresponding to those which
linguists identify in their verbal level. The units of linguistic
segmentation —phoneme,morpheme, word, sentence, and so on
are more or less discrete entities that can be more or less precisely
identified and defined and which, moreover, are linearly and
hierarchically ordered with respect to one another. (Other signifi-
cant aspects of verbal structures, for example, prosodic and intona-
tional features, are, for good reason, referred to as “supra-
segmental.’’) What Rutherford proposes as the “segments’’ of the
nonverbal dimension of literary narratives — namely, story, charac-

ter, and setting, with narrative mode (

= “all that which concerns
the presentation of the world of the novel, rather than that world
itself as the reader is invited to imagine it’’ [178-79]) as a “bridge’’
between the two dimensions (179) — are not segments in that or, in

fact, in any other sense, and further analogies drawn from linguis-

tics on the basis of an implied correspondence between the two are

therefore unjustified.
Segmentation, however, is only one aspect of Rutherford’s
broader purpose. A second one is suggested by the following
passage:

189
 LINGUISTICS AND LITERARY THEORY

Poetics [as opposed to “traditional novel criticism’’), con-

cerned to formulate the norms governing literature as a
system of communication, must direct its attention towards
that which is and so the poetician’s
specific to literature;
task is not to calculate the degree to which a novel ap-
parently copies but rather to establish the principles
life,

according to which the novelist, consciously or uncon-

sciously, makes a selection for the purpose, possibly, of
giving the impression of a copy of life. The poetician will,

then, look for non-psychological and non-sociological ways

of discussing character, non-historical ways of discussing
story, and non-geographical ways of discussing setting
(178).

It is in accord with this curiously negative formulation of the

poetician’s objectives that Rutherford proposes a set of four analytic
methods (one per segment) “that seem to offer interesting possi-
bilities’’ (179) (presumably for doing things differently from literary
critics), illustrating each method with reference to each segment of
El Amigo Manso.
Rutherford prefaces the analysis of the novel with a two-page
summary of it, noting: “The only important parts of the text I have
omitted in this summary are its anti-realistic opening and closing
chapters. ... It is not, however, my purpose here to discuss this
.’’
fascinating aspect of the narrative mode of El Amigo Manso . .

(181).(One might point out, in passing, that the acknowledgment of

such an omission, when added to the extraordinary flatness and
awkwardness of the summary itself, suggests the interesting possi-
bility of unreliable poeticians and, of course, renews one’s skepti-
cism concerning the degree of “general agreement’’ Culler had
ascribed to the construction and assessment of plot summaries.) He
then analyzes in turn the characterization, setting, story ( = plot),

and narrative mode of the novel. It is in the analysis of characteriza-

tion that Rutherford draws on O’Toole’s method, observing that
“what it points towards is ... a system which has certain affinities
with the analysis of distinctive features in phonetics’’ (184). Accord-
ingly, “the basic characterization’’ in the novel (of which “the actual
characterization in the text is an expansion or elaboration’’) is

190
 SURFACING FROM THE DEEP

represented in tabular form, each character being assigned a plus or

minus with regard to his (or her) possession, in reality and in

appearance, of four traits: rationality, superiority, nobility, and

worldliness. Noting that two question marks appear in the table,

reflecting the fact that, in the novel, “the question of the moral
status [that is, “nobility”] of Irene’s and Manso’s behavior is left

unanswered,” Rutherford goes on to state: “This type of ambiguity

is, of course, characteristic of literature, and any method of analysis
should take it into account” (185). And nothing further is said of
the matter. This is, however, surely facile and evasive; for since
poetics was earlier defined as “directing its attention towards that
which is specific to literature such a characteristic, rather than
something to be “taken into account” is presumably what is to be
accounted for. Moreover, precisely because such ambiguity or
indeterminacy of classification is so characteristic of the traits of
literary characters, the more specifically “literary” the characters,

the less useful the distinctive-feature model, for all the entries on the
table would then tend to be question marks!
As I suggested above, the fact that such a method worked well for
O’Toole’s analysis of “The Sussex Vampire” may tell us more about
the formulaic quality of that story than about the general validity of
the method. And, to the extent that the method can be applied by
Rutherford to El Amigo Manso, or by any other analyst to any other
novel, it can tell us only something about the stereotyping of that
novel’s characters or about the analyst’s insensitivity to their literary
qualities. The “Anglo-Saxon criticism” is, then, by no means
rebutted.
Rutherford’s observations on the setting of the novel are per-
functory or, at best, not notably more revealing or interesting than
comparable observations in more traditional forms of literary

criticism. The analysis of “story” brings us to familiar analogic

grounds: “Just as the linguistician who is concerned with syntax has
the task of discovering the principles governing the strings of words
that we recognize as sentences, so the poetician who is concerned
with story analysis should try to discover those strings of actions and
situations which we recognize as stories” (186). The analytic method
that follows from this argument is adopted from Todorov’s

191
 .

LINGUISTICS AND LITERARY THEORY

Grammaire du Decameron, in which stories or story sequences are

taken as analogous to propositions, characters to nouns, their
attributes to' adjectives and actions to verbs. Analysis produces an
algebraic formula that represents an abstract model of the novel’s
story, such that “the incidental details can be seen as expansions or
elaborations of the underlying structure it reveals” (192). The
justification of the production of such abstract models of individual
novels and stories is that “they give the poetician, in principle, the
chance of constructing a finite number of models which, together
with the application of a set of transformational rules, will generate
all the stories possible in narrative literature” (192). Well . .
.
just a
chance.
The analysis of the novel’s story is followed by an analysis of the
fourth of its segments, narrative mode, defined by Rutherford as
“the ways in which the novel’s objective world of character, story
and setting may be actualized in the text itself” (195). He observes:
“The interest ofmany narrative texts, especially modem ones, is

centered on narrative mode rather than story, character or setting”

(196, my italics). As I shall suggest below, in connection with other

more general questions raised by the article, it might have seemed

more to the poetician’s point to observe that the “especially” in this
respect is probably not so much modern texts as, in fact, literary
ones.
The article concludes with the hope that “the reader will now . .

be in a position to judge whether the division of texts into these four

segments, and the methods of analysis I have suggested for each of
them, show possibilities of contributing to a general theory of
narrative” (211), and with the observation that, although the
methods do not meet the criteria set forth by Culler —insofar as
“they do not provide any direct explanations about how a reader
assigns meanings to a text as he reads it” but only “after he has read
it” — nevertheless, “the two approaches complement each other”
(211-12).

The three structuralist papers in the volume discussed above (the

fourth, by Seymour Chatman, is anomalous among them and will be
considered separately, below), both individually and taken together.

192
 SURFACING FROM THE DEEP

have, I think, a better claim to the attention of literary theorists than

do the papers offered by the stylisticians: they reflect a more
spacious and sophisticated sense of the possibilities of literary
theory; the general objectives for which they argue are explicitly and
interestingly related to each other; and, although I believe each of
them is radically flawed or limited, their authors do not seem, as
do the stylisticians, to be running in place at a dead end. As I have
been suggesting, however, it may be that in their attachment to
linguistic models and attraction to linguistic analogies, the structur-

alists represented in this volume have created unnecessary obstacles

for themselves and taken strangely circuitous routes toward their
destinations.
In that connection, I should like to consider briefly two of
Rutherford’s central assumptions, beginning with the one that is

apparently fundamental to the structuralist use of the linguistic

model, namely that literature, like language, is in some sense a
system —a “system of signs’’ in Culler’s characterization, a “system
of communication’’ in Rutherford’s. One may agree that our
expectations of and responses to various verbal structures are
shaped by distinctive sets of conventions and, moreover, that it is

possible to regard and study “literature’’ as the system constituted

by one set of such conventions. It does not follow, however, that that
system is directly analogous to the one that constitutes a natural
language or, more importantly, that literature is, like language, a
system of communication. On the contrary, one may observe that
there is a fundamental way, in this respect, in which they differ, and
that its implications for the pursuit of literary theory are substantial:
for the distinctive characteristics of the system that defines and
constitutes “literature’’ —the conventions in accord with which a
verbal structure does become or is regarded as literature — originate
in and continue to be shaped not by a communicative function but
by an aesthetic one. Thus, whereas the conventions that constitute a
language (or any essentially communicative system) exhibit charac-
teristics —such as relative stability, economy and perspicuity —that
reflect its fundamentally instrumental functions in communication,
the conventions that constitute literature exhibit rather different
characteristics —such as relatively continuous innovation, “un-

193
 LINGUISTICS AND LITERARY THEORY

economical” elaboration, and semantic ambiguity or indetermi-

nancy — that
*
reflect its aesthetic functions.*^

The significance of the distinction becomes apparent when we

consider another of Rutherford’s assumptions, also familiar in
structuralist writings, namely, that poetics ‘‘must direct its attention
to that which is specific to literature.” We recall that, for Ruther-
ford, this implied that the poetician must attempt to establish ‘‘the
principles according to which the novelist makes a selection
. . .

from life and orders that selection for the purpose, possibly, of
giving the impression of a copy of life” (178). It is difficult to see,
however, how that implication follows. Is it the selection and
ordering from life that is ‘‘specific to literature " Are not all

narratives, literary or otherwise, selections and orderings from life

in accord with some principle? And is it the purpose (‘‘possibly”) of

that selection and ordering to give the impression of a copy of
life? —that is, is verisimilitude ever the basic principle or purpose of
the novelist’s art? By way of ansewring the questions raised here,
and also to draw together some points made earlier in connection
with the articles by Culler and O’Toole, I should like to offer some
general observations on narratives and some suggestions for an
alternative conception of narrative theory.

It is, I think, quite possible that every ‘‘story,” fictional or non-

fictional, has what could be called a kernel theme or indeed a
‘‘generating” principle, in accord with which its elements have been
selected from all the possible things the narrator could at that
moment have said or written: something quite close to what, with
regard to a conversational anecdote or news-story, we speak of as its

‘‘point,” and which could also be seen as its motive or, most simply,
the reason why it was told. If, as I suggested earlier, the basic
or minimal plot of every story (change, reversal of fortune, or
peripeteia, as structuralists from Aristotle to Barthes have main-
tained) can be reduced to the assertion ‘‘something happened,” then
its basic theme or generating principle would seem to constitute the
reply-in-advance to the listener’s always potential question, ‘‘So
what?”; and both basic plot structure and basic theme could be seen
as indeed universals of narrative, grounded in the recurrent occa-

194
 SURFACING FROM THE DEEP

sions, contexts, dynamics, functions, and motives of one of the most

universal of all forms of verbal behavior (or “speech-acts”), namely
telling.

In ordinary verbal transactions, we do not tell stories (for

example, relate personal anecdotes, repeat news events) merely to
inform our listeners that something happened: things are always
happening, but there is not always any reason to tell them to
someone. One reason we do sometimes tell them is that the fortunes
of those involved in the story (for example, the speaker himself) are
presumably of some independent interest to the listener. If that is

not the case (and of course it is usually not the case in fictional
narratives), the story is usually told because it exemplifies or
indicates (or apparently contradicts) some general proposition
again, one that is presumably of some interest to the listener —which
may or may not be explicitly stated.
Like nonfictional narratives, fictional ones also relate something
that happened; and what happens in them also commonly illustrates

or indicates (or subverts) general propositions, for example, that

loveless marriages encourage adultery, that sublunary events are
providentially designed, that friends in need are friends indeed. Not
all fictional narratives are (or are designed to be) “literary,” of

course: whatever aesthetic interest they may have, parables, myths,

and fables obviously also serve didactic or socializing functions and,
as I think we now recognize, the “literariness” of a verbal structure
is always both a relative and variable matter. Nevertheless, if we are
seeking for that in narratives “which is specific to literature,” we
might reasonably begin by looking for those characteristic modifica-
tions of narrative universals that seem most clearly to be shaped
by —or to serve —an aesthetic function. Among the more prominent
of such modifications are the following:
1. In the elaboration of “what happened,” the structure and
formal linguistic features of the narrative themselves secure the
listener’s interest and become sources of gratification. For example,
the “suspense” that is potential in any narrative is, in literary

narratives, characteristically enhanced by that structure of “enig-

ma” described by Barthes in 5/Z; and what would otherwise be the
listener’s natural impulse to prod the speaker to “get to the point”

195
 LINGUISTICS AND LITERARY THEORY

is, with respect to a literary narrative, characteristically countered

by the pleasures of getting there.
2. The general propositions that are exemplified or indicated by
the events of the narrative are characteristically left unstated and
therefore indeterminate, the construal or inference of them by the
individual reader being among the characteristic sources of interest
in literary narratives.

3. In accord with one of the central conventions of “the system

of literature,” the act of narration is itself understood to be fictive

and the entire verbal structure of the work is understood to be not a

tale told but the representation of a telling —a modification that
simultaneously establishes and protects the aesthetic function of the
literary narrative as well as serving it.

This list could obviously be extended and each item in it elaborated

and refined. I offer these observations not as a theory of narrative
but to suggest that a theory along such lines could be pursued and,
furthermore, pursued without direction or analogies from the
criteria, objectives, or analytic or synthetic methods of transforma-
tional-generative grammar or, for that matter, from concepts or
methods in any other discipline, from physics to philosophy,
occasioned by and designed to solve quite different problems. I

would suggest, then, that a general theory of narrative might offer

to discover the social and psychological functions of all tellings, to
relate the characteristic features of narratives to those functions,

and and account for variations among those

to locate universal
functions and features; and that, in association with it, a theory of
literary narrative might offer to investigate and describe character-
istic modifications of those features that appear to serve and be
shaped by specifically aesthetic functions. While those who pursue
such theories of narrative might be expected to draw on linguistics
for whatever understanding of language it may provide that is

actually relevant to their enterprise, and would also be expected to

draw on the data of, and integrate their own findings and formula-
tions with, studies in anthropology, folkloristics, sociology, psychol-

ogy, and communications, they nevertheless would not define their

1 %
 SURFACING FROM THE DEEP

domain or objectives or devise their methods of analysis in the image

of those of any other field of study.

VI. Conclusion
Seymour Chatman’s article, “The Structure of Narrative Transmis-
sion,’’ is the last in the volume and, to my mind, the most sub-
stantial: its subject is sharply focused, its aims plausible and clearly
defined, and, although Chatman, like his co-contributors, is con-
cerned with methodological questions, he is not consumed by them.
Thus the major part of the article is neither a methodological
warm-up exercise nor a dress rehearsal, but a full-fledged per-
formance or (the article being one section of a larger study) at least
the complete scene of one.
The questions that Chatman’s study is, at least in part, designed

to answer are as follows: for any narrative work, if a narrator is

present, how is his presence recognized and how strongly is it felt by

the audience; and, more generally, since “in all but the scenic
arts —
like drama and the ballet —
pure mimesis ... is an illusion,’’
how is this illusion achieved, “by what convention does a reader, for
example, accept the idea that it is ‘as if he were personally on the
scene, though the fact is that he comes to it by turning pages and
reading words’’ (215). Two assumptions, not completely self-

evident, are implied by this last question: that the idea is accepted
(or the illusion of pure mimesis experienced) by the reader, and that
the illusion is achieved by some “convention.’’ With regard to the
first, one might quibble over the terms illusion and mimesis and
suspect that the purity of either is rare, but I would be inclined to
grant that we do, when we read literary narratives, experience
something like an imaginative substantiality (or “real presence’’)
and that our doing so is of significance in the value those narratives
acquire for us and thus for literary theory. With regard to the
second, the assumption is justified in the subsequent discussion, in

which the term convention is used in a nonmetaphoric sense with

reference to the basis on which quite specific linguistic features of a
text direct and control the reader’s quite specific inferences and
experiences.

197
 LINGUISTICS AND LITERARY THEORY

Chatman observes: “Preliminary to any discussion of the struc-

ture of discourse in literary narratives is an understanding of the
linguistic and linguistic-philosophic basis for reports of speech,

thought, physical action, and so on” (in effect, those conventions,

strictly linguistic and other, that govern the inferences we draw from
narrative statements generally), “since it is at least partly on these
grounds that the reader makes his decision about who is speaking,
thinking or whatever, and in particular whether there is an express
narrator or not” (218). This issomewhat awkwardly stated, but the
point is, I think, clear enough and important. In elaborating it,
Chatman draws on the theory of speech-acts developed by J. L.

Austin and John Searle and on Richard Ohmann’s distinction,

based on that theory, between fictive and nonfictive narrations.*^

The borrowings here, though digressive and perhaps not strictly

necessary, are nevertheless tactful and discriminating, and permit

Chatman to describe, with impressive precision, the conventions
that direct our identification and interpretation of narrative state-

ments in literary discourse.

The methodological concerns of the article are reflected in

Chatman’s trenchant critiques of studies by Ohmann, David Lodge,

and F. Stanzel,*® and in the contrast he notes between his own
descriptive analysis, based on the assumption that there is a variety
of “discourse features” which combine in various ways in individual
works, and taxonomies that propose to establish a finite number of
“homogeneous and fixed genres” (233). He observes:

Variety among narratives is thus [that is, in his own

analysis] accounted for in terms of various mixtures of
independent features, not by an endless proliferation of
categories or a Procrustean reduction of instances into
normative types. Literary theory in general and narrative
theories in particular have suffered from too powerful a
reduction into a small number of genres, with the conse-
quence that the full discursive complexities of individual
cases are missed because they don’t “fit” or get inter-
preted somehow as exceptional, or even worse, aberrant
(233-34).

Since Chatman is not himself given to self-promotion, it should be

198
 SURFACING FROM THE DEEP

added that among the salient virtues of his own taxonomy are the
following: that it is sparing in its assumptions and that those
assumptions it makes are drawn from what is relatively well
established in our present understanding of language and the
dynamics of communication; and that, because they are based on
clear and consistent principles systematically pursued, its categories
are genuinely parallel, comprehensive, conceivably exhaustive, and
lend themselves readily to refinement and elaboration by other
theorists.

In the article, Chatman develops only one section of what is

ultimately to be “a description of the variety of narrative trans-

mission,” specifically that section which deals with ‘‘structures with
the least presumption of a narrator’s presence” (233), that is,

narrative works and passages close to ‘‘unmediated story or pure

transcript or record” (237). Accordingly, he describes how sources
of narrative transmission are implied and inferred (1) in those forms
of literary narrative, such as epistolary novels, that imitate written
records and ‘‘reduce the implied author to a mere collector of
documents” (240); (2) in those, such as Faulkner’s As I Lay Dying,
which imitate a transcription of speech and in which ‘‘the implied
author is presumed to be nothing more than a stenographer” (242);
and finally (3) in those, such as passages in Joyce’s Ulysses and
Woolf’s Mrs. Dalloway, in which a character’s thoughts and
feelings, or ‘‘consciousness,” are represented without the overt
mediation of a narrative voice.
The distinctions here are sharp, but admit of subtlety, and the
generalizations are uncluttered and persuasive. Without making
much ado of the fact, Chatman consistently relates distinctive
linguistic features to their characteristic implications or expressive

effects, supporting his observations in these regards neither by

invocations of the competence of ideal readers nor by allusions to
cultural or linguistic codes but by demonstrating, with concrete
examples from a number of novels and stories, how variations
among those features produce corresponding differences of impli-
cation or effect. Also, it might be noted that although the general
framework of the analysis as well as its details reflect Chatman’s
command of the contemporary study of language, its domain.

199
 —
LINGUISTICS AND LITERARY THEORY

objectives, and methods are neither derived from nor validated

through analogies to linguistics.
*

In fact, Chatman makes only one direct allusion to transforma-

tional-generative grammar, but it is, for our present purposes, an
instructive one. In the course of examining the narrative functions
and effects of certain linguistic features in the opening sentences of
The Brothers Karamazov, among which are a number of passive
constructions, Chatman observes that “in the early versions of
transformational grammar, passive constructions were derived from
their active counterparts.” He continues:

But fundamental narrative units the story statements —

are not to be equated simply with the underlying deep struc-
tures of sentences. On the contrary, the transformations
whose effects appear on the surface manifestation of
literary narratives may be clearer indications of narrative
structure (223).

Although the specific point is given substance in the discussion that

follows, it is made essentially in passing; that is, it is not further
explored as a methodological issue. A more general point is

suggested, however, and it is one of considerable significance, not

only for Chatman’s study and the broader investigation of narrative
but also for the general pursuit of literary theory: namely, that the
methods used by linguists to analyze, explain, and represent the
grammatical features of verbal structures may be quite irrelevant to
the analysis and explanation of their literary functions and effects,

that the employment of those methods may in fact obscure rather

than illuminate the literary theorist’s subject and, finally, that it

may be in the investigation of the ^'surface linguistic manifesta-

tions” of literary texts, rather than in the effort to locate any
putative ''deep structures” in them, that the labors of the literary
theorist will be most rewarded.
It is pertinent to observe that Chatman’s article resists reduction
to a variation of our formula

S(R) xl££!EiL2LlA-.
Or, put another way, in his study of the conventions of narrative
transmission, X drops out of the formula and R is the relation not of

200
 SURFACING FROM THE DEEP

surface phenomena to deep matters but of various linguistic fea-

tures to their convention-governed implications and, thereby, to
their literary-narrative functions and effects. The relation here, we
might note, is not a vague or volatile one, such as “reflects” or
“actualizes,” but the relatively unambiguous, stable, and reversible
one of implication, here of the type specified-implication-in-accord-
with-specified-convention. Moreover, since both the features and
their effects are manifest, open to direct observation and even to
manipulation (as when, by altering a feature slightly or comparing it

to a slightly different one, Chatman demonstrates the alteration or

difference of its effects), he can concentrate his efforts on the
precise and systematic description of those features and can provide
an explanation of their functions and effects in various (and
potentially alt) narrative texts. Deep-sea diving is not required at any
point.
I do not wish to overburden Chatman’s article with praise. Its
aims and achievements are relatively limited, its claims modest.
Nothing in its revelations is startling or altogether novel, which may
be part of the reason why it reads so easily, invites so few objections,
and occasions so little exasperation. Nor do I wish to suggest that all

that is admirable and exemplary in the article owes to the fact that
Chatman, uniquely among the contributors to the volume, has not
felt compelled to dog Chomsky’s tracks. Nevertheless, reaching this

article at the end of a volume in which so much that is new is old,

and so much of what is both new and old is dismaying, one might
very well come to the conclusion that only by surfacing from the
deep can we discover the salutary pleasures of air and light, acquire
a less subterranean and more sunlit view of the continents there are

to explore, and have the hope of dry land at the end of our journeys.

201
 I • ' , . ' 1 } ' .J

» \

* •
.-I. iuB nsi — ' •*" •*

*J ,¥ V ,
i'" tnt”. * *
(

’
jJ.^i i:.-i fff ' * 1 . I'l- ..

' 1 •

.... •
|U‘ V’-V <

r.^« ' 7
’
' .
•
, u. i|ki4i/.

• •, -. .A ^ ‘
-
>
t *
^ •
•
tsf-iv * vj-'tr •tAiv

« •'•» . rt«4u *. • 1 1

‘
«
w .1 fikipik, n” r ni'tiip’U # #
*
- • . • i ,

* .

’VI .1 •. i*r ‘.If Ri.ccCi‘*l

^

t 1 -
. b nry> flt-j •
t\ « • •
•
) 1

••
• .- M 1
'•' i» - « . 1 If. %
4. ; .
.

'}*'!* *• . >•
'
» -*

•
' 'i:*? .
>1
; *
-j it V
> »
!?«*'

^ ^ f

t '•41* 4
'
• .1 M '
^'.r 1 V ft 1
-4 .11
W 1 . f.l . •
'
’
. .t .
;• , ,
•».,• fi.y •
: >

ft*r >\ •
. . .
t'U’' ^ .
r -; - .Vi.ftt-t'
vn
'
.
•• ’ 1 1
' *
i
'* V .
* .•
.
> .• »ttii
•
.
•
*Tr.'*

- » . iv’ ’! •
•
( ,^TT3i:* i/i * -* >' .^1 ihfil

• I,
,
1
,
t*. *. Tn u- i*
np
1

^
'
I (
.^* .
. r **
. . . r> • *
1. V jCul
1 .

^ . « !l ' . .
!' ' ’!•• - n, ;
» . lO* • ij,

If • ’ ' ' ' .. i| rSy^fr'd

*'T*- -•^'t ;
.
.
,
. ,
*4.1 .ft
' ”.,T ii
1 » v!M’ '"
,':) . .'i‘!. ffi *tifi

» *r'/' M
G"*
•
>
1 r
, f . .
.
t
‘i ,

. . n f* 1
'
' i
'' t 1
'1 , .1 i tli"'/ t

i&'Hi
1

t* V \r* V.

« • ^ tv » h. t’rtX-"-*’'

’
•
T Vl: i». («<sj -'lU'**'" fit

m
It u. t \
 NOTES

Chapter One
Nelson Goodman, Languages of Art: An Approach to a Theory
1.

of Symbols (Indianapolis and New York, 1968).

2. Since the classifications offered here are not based on simple
considerations of aesthetic “merit” or cultural significance, this
would include the works of Herodotus as well as the note
first class

for the milkman. What distinguishes the literary or linguistic art-

work is what I describe below as “fictiveness” and have elaborated
elsewhere as the distinction between ahistorical (or mimetic) and
historical discourse {Poetic Closure [Chicago, 1968], esp. pp. 14-
23). There are certainly borderline and otherwise difficult cases, for
the distinctions we are dealing with here are not strictly logical ones,
but involve those “antecedent classifications” that Goodman some-
times acknowledges, that is, cultural traditions and conventions that
govern how any utterance, on any particular occasion, is conceived
of by its speaker (or author) and “taken” by its audience.
3. That is, if such texts exist. This third class would also include
all instances of uninscribed oral “literature” and, of course, all

nonce works of verbal art such as the rhymes a child might compose
while playing. By making no provision for these events or neglecting
to acknowledge their logical and historical relation to inscribed
works of verbal art, Goodman secures a spurious simplicity for his
own class of “literary works” in Languages of Art.
4. Goodman conceives of the relation between natural discourse

203
 —
NOTES TO PAGES 9-13

and “the world of objects and events’’ as identical with the relation
between scores and their compliant performances or inscribed
characters and their phonetic compliants. Since this conception
seems to partake of all the defects of any referential theory of
language, I find it patently unserviceable. [The nature of the causal
relation between utterances and “the world of objects and events’’
iselaborated below (pp. 16, 21-22) and developed more fully in part
2 (pp. 86-100) in connection with a nonreferential conception of
“meaning.’’]
5. Goodman does not, to my mind, give sufficient attention to the
extent to which our response to a work of
opposed to any
art, as

configuration or occurrence of natural events, is determined by our

recognition of it as an artifact. When we “interpret’’ a work of art
that is, hypothesize its generating or structural principles, or ascribe
causal relations to it —we are engaged in a cognitive activity with
possibilities and limits significantly different from those which ob-
tain for the scientist interpreting a set of natural events. For
example, we inevitably experience and interpret works of art in a
characteristic cultural contextand at least partly in terms of pre-
vious experiences with the genre and style of the work, considera-
tions that could be extended to the interpretation of natural events
only by gross metaphor.
6. Goodman suggests that “scientific systems’’ are as “artificial’’
as works of art (cf. his reply, “Some Notes on Languages of Art,''
The Journal of Philosophy 47 (1970): 563-73). The sort of scientific
system he has in mind, however, is presumably not a disembodied
concept but a theory or hypothesis made available to us in the form
of a discursive statement or set of statements. Such statements are
artificial only in the restricted sense that they are the products of
human activity. They are surely not artificial, however, in the sense
of constituting a fictive nature; for the scientist does intend, and his
audience understand him to have intended, that his alludings be
will

taken as real alludings. That is, the conventions determining or

indicating fictiveness do not operate for the statement of scientific
systems any more than they operate in my note for the milkman.
Would Goodman maintain that there was no relevant
really wish to
difference between the account of the motion of the planets in
Copernicus’s De Revolutionibus (“hypothetical” though it was) and
the account of the motions of The Flying Island of Laputa in
Gulliver’s Travels!

204
 NOTES TO PAGES 14-16

The phrase “nature proper’’ is perhaps ambiguous. There is an

obvious sense of nafure which any aspiring monist, myself included,
would have to grant embraced cultural artifacts, including artworks
themselves, along with all other objects and events. I am not using
nature in that sense here, but in the other familiar sense that
contrasts “nature’’ with artifacts, the natural object or event with
the artificial one that may be taken to represent it.

Chapter Two
1. Poetic Closure: A Study of How Poems End (Chicago, 1%8),
esp. pp. 14-25, and “Literature as Performance, Fiction, and Art’’
[above, pp. 3-16].
2. [Since the population denoted by that “we’’ is not delimited in
any way here, the consequent implication could be that “we’’ are
everyone and/or that “actually c/o’’ means “actually should.'' It was
not my intention, however, here or below, either to invoke what
would certainly be a very questionable universal or to present a
normative statement in the guise of a descriptive one. As I point out
later in the study, especially in parts 2 and
what “we’’ actually do
3,
depends a good deal on who we actually are and on the circum-
stances in which we do it; and what I allude to below as the
conventions and assumptions that direct “our’’ experience of verbal
structures must be understood as themselves historically, culturally,
contextually, and otherwise variable.]
3. Since the term context has been acquiring increased currency

in comtemporary aesthetics and linguistics, I should point out that it

is not my intention here to quarrel with or qualify the sense it bears
for other theorists. It might have been better to discover or devise
another term altogether for what I am here defining and later
elaborating, but the alternatives that presented themselves seemed
just as likely to create comparable confusions, and I confess to a
temperamental loathing of neologisms. It should also be noted that,
in proposing that we view the context of an utterance not merely as
its physical setting but as the totality of its determinants, I am not so
much broadening the ordinary reference of the term as affirming
the existence and significance of a particular relation, namely
causality, between a verbal event and the universe in which it occurs.
Defined in terms of that relationship, the “context’’ of an utterance
inevitably refers to something more extensive than what the common
use of the term suggests, but also something more particular.

205
 NOTES TO PAGES 25-46

have considered the matter elsewhere: see “The

4. I New Imag-
ism,” Midway 9 (1969): 29-44.
5. “Private or personal utterance” may be extended to include not
only overt but interior speech. The representation of the latter,
particularly in romantic and modern lyrics, is discussed in Poetic
Closure, pp. 139-50.
6. I should emphasize that I am not specifically referring here to
those formally and publicly articulated “interpretations” of poetry
that we associate with academic or professional criticism, but rather
to the informaland often enough private activities of the reader as
such, or what we might otherwise speak of as his response to or
experience of the poem. Of course much formal criticism is an
extension of these informal activities, but the very fact that pro-
fessional critics are offering public statements entails other concerns
and responsibilities, and I am not presuming here either to limit or
to account for them. See, however, pp. 38-39, below.
7. The Sonnets, Songs, and Poems of Shakespeare, ed. Oscar

James Campbell (New York and Toronto, 1964), p. 136.

8. [That is, with respect to those conventions. As
emphasize I

later, the reader is not constrained in any absolute sense by the

poet’s assumptions, and interpretations of a work that are “in-

appropriate” in that respect may yet be “appropriate” in relation to
other emergent functions, conventions, and conditions of encounter.]
9. [It is dealt with at length in part 2.]

Chapter Three
1.Comparable observations may, of course, be made with respect
to the term literature. For a discussion along somewhat the same
lines as presented here, see Christopher Butler, “What Is a Literary
Work?'' New Literary History S (1973): 15-29. [See also John Ellis’s

recent and illuminating discussion of the definition of literature in

The Theory of Literary Criticism: A Logical Analysis (Berkeley,
1974), pp. 24-53.]
2. [An analysis of the child’s functional discrimination of fictive
discourse is presented in greater detail in partpp. 124-32.] 2,
3. The distinction initially outlined in “Poetry as Fiction” is

redrawn here with some modifications.

4. have in mind here statements such as the following:
I

No literary theorist from Coleridge to the present has suc-

ceeded in formulating a viable distinction between the
nature of ordinary written speech and the nature of literary

206
 NOTES TO PAGES 49-53

written speech. For reasons I shall not pause to detail in this

place, I believe the distinction can never be successfully
formulated, and the futility of attempting the distinction
will come be generally recognized (E. D. Hirsch, Jr.,
to
“Three Dimensions of Hermeneutics,’’ New Literary His-
tory 3 [1972]: 260).

Since nobody has ever managed to devise any workable

criterion for distinguishing “poetic language’’ from “ordi-
nary language’’, seems foolish to retain a spurious termin-
it

ological distinction which effectively denies common sense

(Roger Fowler, “The Structure of Criticism and the Lan-
guage of Poetry,’’ Contemporary Criticism, Stratford-upon-
Avon Studies [London, 1970], 12:183-84).

[I return in part 2 to “skeptical monism,’’ and consider more

closely Hirsch’s statement and the argument that attends it (pp.
133-54.]
5. The Concept of Mind (London, 1949), pp. 16 ff.
6. Signals and cues for the classification of utterances, and the
possibility and consequences of misclassification, have received
extensive treatment in the works of Gregory Bateson and Erving
Goffman. See Bateson’s notion of “frames’’ and “metacommunica-
tion’’ in “A Theory of Play and Fantasy,’’ Steps to an Ecology of

Mind (New York, 1972), pp. 177-93, and Goffman’s most recent
study. Frame Analysis: An Essay on the Organization of Experience
(New York, 1974), esp. pp. 40-47.
7. Terms are troublesome. In what follows, I shall use the terms

poem and poetry to refer to (1) those compositions commonly and

readily so designated, and (2) a particular subset of fictive utter-
ances characterized later in the discussion, the members of which
usually, but not necessarily, correspond to (1).
8. [The terms ahistorical and noncontextual are subject to some
misunderstanding. As observed earlier (pp. 33-34), I am not
I

suggesting that a fictive verbal structure comes into being inde-

pendent of any human agent or historical context. I do distinguish,
however, (a) and contextual act that constitutes a
the historical
natural utterance, {b) the historical and contextual act that con-
stitutes the composition of a fictive utterance, and (c) the ahistori-
cal, noncontextual structure that constitutes a fictive utterance.]

9. The contrasts drawn here between paintings and pictorial il-

lustrations will not, of course, hold good for all paintings and all

207
 NOTES TO PAGES 56-66

illustrations, even we confine ourselves

if just to the “representa-
tional” ones. A number of painters have cultivated a deliberately
flat, anti-illusionist quality (Matisse, for example, or Milton Avery),
and others (for example, Roy Lichtenstein) have been interested in
suggesting the pictorial quality of — or, in fact, in representing —
precisely the sort of illustration described above. The analogy does
not require a hard-and-fast distinction between paintings and il-

lustrations, however, because the distinction between poems and

verbal illustrations is not itself a hard-and-fast one. All fictive
structures are, by definition, to some extent mimetic, but not all to
the same extent. What may be “mimetic repleteness'' that
called —
is, the degree to which a fictive structure not merely represents an

identifiable member of a certain general class of things but also

evokes the illusion or impression of its historical particularity — is a
and
relative matter, will vary from poem to poem, from painting to
painting, and from illustration (visual or verbal) to illustration.
10. [Not all audiences, of course, especially not those, such as
children, who have not yet learned the conventions. (See pp. 212-13,
footnote 7.)]
1 1 . Compare the greeting-card message printed above with these
lines from Theodore Roethke’s “My Papa’s Waltz”:
The hand that held my wrist
Was battered on one knuckle;
At every step you missed
My right ear scraped a buckle.
{The Collected Poems of Theodore Roethke [New York, 1%1], p. 45)
12. The Singer of Tales (Cambridge, Mass., 1960), esp. pp.
30-67.
13. The examples given here have been adapted and simplified
from Lord’s, pp. 50-53, and some consist of wholly invented
analogues.
14. Reprinted in Concrete Poetry: A World View, ed. Mary Ellen
Solt (Bloomington, Ind., 1968), p. 114.
15. Of all the relations a speaker may have to someone else’s
words, perhaps the most interesting performing of them, as
is his
when an actor recites the lines of a play or when we read a poem,
either aloud or to ourselves. “Performing” is quite distinct from
either quoting, depicting, or referring to an utterance or, of —
course, saying it. The relation is, however, a complex matter in its
own right. A brief discussion of literary performance within the

208
 NOTES TO PAGES 66-90

theoreticalframework of the present study may be found in Poetic

Closure (Chicago, 1968), pp. 9-10 and above, pp. 5-8.
16. London, 1924, p. 5.
17. Answer Back (New York, 1968), p. 33.
18. [“Parabolic” meaning is explored further in part 2, pp.
141-44.]
19. The Proverb (Cambridge, Mass., 1931).
20. It may be worth pointing out, in connection with our earlier
discussion of advertisements, that advertising slogans often achieve
the semblance and acquire the force of household “sayings” at least
partly because, like those sayings, no one knows who first said them.
Thus, electronically transmitted gnomes infiltrate and swell the
store of culturally transmitted proverbs and maxims until we can
virtually regulate our entire lives in accord with what some totally
and eternally unidentifiable “they” say.
21. The Proverb, p. 169.
22. Oral Literature in Africa (Oxford, 1970), p. 424.
23. [The topic is treated at length in part 2.]
24. Harold G. Henderson, An Introduction to Haiku (New York,
1958), p. 112. The translation given aboveadapted from the two
is

translations of the poem provided by Henderson.

25. [The characteristic activity of interpreting a poem is not con-
fined to the construing of contexts. Other distinctive aspects of
literary meaning and interpretation are considered in part 2, esp.

pp. 121-24 and 137-54.

Chapter Four
1. I allude here to the following studies: Roman Ingarden, The
Literary Work of Art (orig. George G. Grabowicz
pub. 1931), tr.

(Evanston, 111., 1973); Kate Hamburger, The Logic of Literature

(orig. pub. 1957), tr. Marilynn J. Rose (Bloomington, Ind., 1973);
Robert Champigny, Ontology of the Narrative (The Hague, 1972);
Richard Ohmann, “Speech, Literature, and the Space Between,”
New Literary History ^ (1972): 47-64; John R. Searle, “The Logical
Status of Fictional Discourse,” New Literary History 6 (1975):
319-32.
2. Studies by Roger Brown and Camille Hanson support the
suggestion that parental approval and disapproval do not, in them-
selves, operate as “selection pressure” for the increased “correct-
ness” of a child’s speech. See “Derivational Complexity and the

209
 NOTES TO PAGES 94-104

Order of Acquisition in Child Speech,” Cognition and the Develop-

ment of Language, ed. John R. Hayes (New York, 1970), pp. 155-
207, and the discussion by Brown in A First Language (Cambridge,
Mass., 1973), pp. 410-12.
3. The quantity of information conveyed by an event is always
relative to a particular “receiver” at a particular time. Thus, while
the assertion that the GeorgeWashington Bridge is intact may
contain no measurable “news” for most listeners, it may be highly
informative to someone who had recently planned or chronically
fears its collapse.
So distinguished, the meaning of symbolic events is close to
4.

what H. P. Grice terms “non-natural meaning” (“Meaning,” The

Philosophical Review 64 [1957]: 377-88, and “Utterer’s Meaning
and Intentions,” The Philosophical Review 1% [1969]: 147-77). The
formulation of the distinction presented here, however, differs in
several respects from Grice’s. For a view of convention-governed
behavior more in accord with the one developed in the present study,
see David K. Lewis, Convention: A Philosophical Analysis (Cam-
bridge, Mass., 1969).
The “two persons” may, of course, be the same person, that
5. is,

we may respond to our own symbolic acts as symbolic events, as

when we leave notes for ourselves.
6.During the past decade, considerable effort has been directed
to remedying this deficiency. The effort is reflected in the work of
speech-act theorists (J. L. Austin, How to Do Things with Words
[New York, 1965], and John R. Searle, Speech Acts [Cambridge,
England, 1969]) and in studies such as Erving Goffman, Relations
in Public (New York, 1971); Gregory Bateson, Steps to an Ecology

of Mind (New York, 1972); Robin Lakoff, “Language in Context,”

Language 48 (1972): 907-27; Dell Hymes, Foundations in Socio-
linguistics (Philadelphia, 1974); Roy Turner, ed., Ethnomethodology
(London, 1974), esp. Turner’s article, “Words, Utterances and
Activities,” pp. 197-215.
Gregory Bateson’s discussions of “metalanguage,” “metacom-
7.

munication,” and, generally, of “the messages that identify what

sort of message a message is” (see especially “A Theory of Play and
Fantasy” and “Epidemiology of a Schizophrenia,” Steps to an Ecol-
ogy of Mind, pp. 177-200) correspond very closely to the observa-
tions outlined in this paragraph. Schizophrenia is, of course, a more
serious misfortune than social embarrassment, but readers familiar

210
 NOTES TO PAGES 1 16-21

with Bateson’s theory of “the double bind’’ will appreciate the

relationbetween mistaking the spirit and force of an utterance and
being given contradictory messages about them.

Chapter Five
1. What is described here as the potential multiplicity of function
of any verbal structure is evidently similar and perhaps equivalent to
what Jacques Derrida refers to as the “iterability’’ of all “written
signs’’ or all “marks.’’ See “Signature Event Context,’’ Glyph 1

(1977): 179-80.
2. As the heavy qualifications indicate, I am not offering here to
provide a definition of “artwork’’ or “aesthetic,’’ nor, as will be
seen, am I suggesting or assuming that there is some kind of activity
or experience uniquely evoked by artworks or some single function
that such works serve. Moreover, I should emphasize that the “cog-
nitive activities’’ described below are not to be thought of as speci-
fically “intellectual’’ or “mental.’’ Although there is still much to
discover about the mechanisms by which we acquire (or “process’’)
information, it is clear that they involve the entire organism and
that, with respect to our engagement with artworks, no distinctions
can be drawn between purely intellectual responses and either sen-
sory, emotional or physiological responses.
3. Cf. Eleanor J. Gibson, Principles of Perceptual Learning and

Development (New York, 1969): “. getting information from the

.

environment is intrinsically rewarding, not secondarily so’’ (p. 127).

Gibson does not, however, break down that “getting’’ into a process
and end (that is, learning as distinguished from knowing), which
appears to be why she finds puzzling or “paradoxical’’ the fact that
subjects in studies of perception seem both to enjoy uncertainty but
also to be rewarded by a reduction of uncertainty (p. 128). She solves
the problem, improperly I think, but suggestively, by concluding
that those experiments in which subjects appear to enjoy uncertainty
are “artificial.’’
4. borrow the term relational richness from Leonard B. Meyer,
I

to whose studies of musical structure I am indebted here at many

points. See especially his discussion of “conformant elements’’ and
“hierarchic structures’’ in Explaining Music: Essays and Explora-
tions (Berkeley, 1973) and “Grammatical Simplicity and Relational
Richness: The Trio of Mozart’s G Minor Symphony,’’ Critical
Inquiry 2 (1976): 693-779. Recent studies that illustrate the pos-

211
 NOTES TO PAGES 122-29

sibilities of multiple pattern-processing in literary works include

Stephen Booth, An Essay on Shakespeare’s Sonnets (New Haven,
Conn., 1969); Roland Barthes, S/Z: An Essay (orig. pub. 1970), tr.
Richard Miller (New York, 1974); and Benjamin Hrushovski, “Seg-
mentation and Motivation in the Text Continuum of Literary Prose:
The First Episode of War and Peace,'" Papers on Poetics and
Semiotics, no. 5 (The Israeli Institute for Poetics and Semiotics,
Tel-Aviv University, 1976).
5. In the very process of composing a poem, the individual author
continuously tests its an artwork on himself (by
effectiveness as
assuming the role of a stand-in for some presumptive or potential
audience), and he shapes or revises his text accordingly. He may
also, of course, explicitly “try it out” on various associates, and
further revise it in the light of their criticisms and reactions. Sim-
ilarly, the folkloric work is continuously being tested by and tried
out on the various successive audiences to whom it is presented in
the context of performance, and the responses of those audiences
enter into the successive shappings and revisings of the work.
6. See Susanna Millar, The Psychology of Play, (London, 1968)

for a useful historical introduction and a review of recent research. A

number of later studies of specifically verbal play and a compre-
hensive bibliography of the subject may be found in Speech Play:
Research and Resources for Studying Linguistic Creativity, ed.
Barbara Kirshenblatt-Gimblett (Philadelphia, 1976).
7. It should be emphasized that I am concerned here specifically
with how children learn to discriminate natural and fictive dis-
course, that is, verbal and, for the most part, vocal acts and events.
Although there is reason to believe that children learn to discrimi-
nate all fictive and events in comparable ways, a thorough study
acts
of the question would require special attention to technologically,
especially electronically, mediated events. For example, it seems
clear that the identifying signals of fictiveness in popular magazines,
motion pictures, and radio and television broadcasts are not as
reliable as those in discourse, not only because they are sometimes
deliberately more ambiguous (cf. the discussion above, pp. 55-57)
but also because they are relatively less well-established and less
culturally redundant. Moreover, while the child learns language
through social transactions and under a wide variety of conditions,
his encounters with magazines and television screens are more likely

212
 NOTES TO PAGES 134-164

to be solitary and to occur under relatively invariant conditions. The

opportunities for corrective cognitive feedback are therefore mini-
mal, and the differential consequences of fictiveness in “the media”
are certainly far less “im-mediate.”

Chapter Six
1. The Aims of Interpretation (Chicago, 1976), pp. 90-91.
2. Hirsch tends to speak not of readers but of interpreters, for
example, “those who practise interpretation,” and “we who interpret
as a vocation,” and seems to imply that there should be not merely
readers who engage professionally in the public articulation of their
interpretations (as teachers, “critics,” essayists, and so on), but a
corps of highly disciplined exegetes who serve as self-abnegating
mediators between authors and the laity. As I suggest below, how-
ever, it is not clear why such a ministry should be necessary or what
useful functions its devotees would perform that are not already
performed by literary philologists and cultural historians. On the
question, see also pp. 38-39, above.
3. See esp. pp. 47-50, above.
4. The propositions exemplified by a work may, of course, be
propositions about anything, and at any level of generality. Thus a
novel may exemplify propositions about the nature of narrative as
well as the nature of society; and, as we know, much “postmodern-
ist” poetry can be seen as exemplifying propositions about language.
(Propositional exemplification is also considered in part 3, pp. 195-
%, in connection with theories of literary narrative.)

Chapter Seven
1. Roger Fowler, ed.. Style and Structure in Literature: Essays in

the New Stylistics (Ithaca, N.Y., 1975). Numbers in parentheses

below refer to its pages.
2. The first clause of Freeman’s sentence is indebted to an ob-
servation made by Richard Ohmann in Shaw: The Style and the
Man (Middletown, Conn., 1%2), p. 22. Freeman dubs it “the
Ohmann hypothesis” and states that it is among the aims of his
article “to expand this principle with reference to aspects of the
grammar of poetry, a topic so far little studied from a transforma-
tional generative point of view” (19).
3. Epstein himself does not explicitly describe the method or

213
 NOTES TO PAGES 167-84

indicate its relation to its assumptions. The statement provided here

is a compendium of points scattered between pages 41 and 60.
4. Among recently published studies that exhibit something of the
scope one would expect from the kind of theory described here are
Jan Mukarovsky, Aesthetic Function, Norm and Value as Social
Facts (orig. pub. Prague, 1936), Mark E. Suino (Ann Arbor,
tr.

Mich., 1970); Morse Peckham, Man’s Rage for Chaos (Philadelphia

and New York, 1967); and Hans Kreitler and Shulamith Kreitler,
Psychology of the Arts (Durham, N.C., 1972).
5. “What Is Stylistics and Why Are They Saying Such Terrible

Things about It?” Approaches to Poetics, ed. Seymour Chatman

(New York and London, 1973).
6. An Essay on Shakespeare's Sonnets (New Haven, Conn., 1969).
7. Structuralist Poetics (Ithaca, N.Y., 1975), p. 128.
would be Fowler’s contention, one presumes, that stylistics
8. It

was hatched from the same egg or that the two approaches are at
least fraternal twins. Historically, that is more or less true. Neverthe-
less, insofar as stylistics defines itself as the application of linguistic
methods of analysis problems of literary criticism, it is more
to
essentially tied to linguistics than is the structuralist program for the
study of literature, at least as projected by Culler in the passage
quoted above. Of course, characterizations of stylistics and literary
structuralism and accounts of their historical and intellectual fore-
bears and siblings vary considerably among those who have under-
taken to describe either or both of them and to trace their joint or
respective lineages (cf. Karl D. Uitti, Linguistics and Literary Theory
[Englewood Cliffs, N.J., 1969J; Fredric Jameson, The Prison-House
of Language [Princeton, N.J., and London, 1972]; Robert Scholes,
Structuralism in Literature [New Haven, Conn., and London, 1974];
Morton W. Bloomfield, “Stylistics and the Theory of Literature,’’
New Literary History 1 [1976]: 271-311).
9. Structuralist Poetics, p. 96.

10. Comparable questions were raised by Gerald Prince in his

excellent review oi' Structuralist Poetics in PTL: A Journal for De-
scriptive Poetics and Theory of Language 1 (1976): 197-202.
11. Noam Chomsky (Cambridge, Mass., 1965), p. 24.
12. Structuralist Poetics, p. 121.
Roland Barthes, 'Tntroduction a I'analyse structural des
13.

recits,” Communications 8 (1966): 1-27.

14. Claude Levi-Strauss, Anthropologic structurale {Pdir'xs, 1958)

214
 NOTES TO PAGES 194-98

and ''L'analyse morphologique des contes russes/* International

Journal of Slavic Linguistics and Poetics 3 (1960): 122-49; A. J.
Greimas, Semantique structurale (Paris, 1966); Julia Kristeva, Le
Texte du roman (The Hague, 1970); Tzvetan Todorov, Grammaire
du Decameron (The Hague, 1969).
15. The precise nature of those “aesthetic functions’’ are not,
I realize, self-evident. For the present discussion, one need only grant
that a set of distinctive functions exists. An
and valuable
interesting
analysis of the aesthetic use of language (drawing a good deal from
East European literary theory) appears in Umberto Eco, A Theory
of Semiotics (Bloomington, Ind., and London, 1976), pp. 261-76.
[See also the discussion above, pp. 116-24.]
16. Rather than “in literary narratives,’’ here and below, it would
be more accurate, if more cumbersome, to say: “to the extent that a
’’
narrative is, or can be, regarded as ‘literary.’
17. Austin, How Do
Things with Words (New York, 1962);
to
Searle, Speech Acts (London, 1969); Ohmann, “Speech Acts and
the Definition of Literature,’’ Philosophy and Rhetoric 4 (1971):
1-19.
18. Ohmann, “Speech, Action, and Style,’’ Literary Style: A
Symposium, ed. Seymour Chatman (London and New York, 1971);
Lodge, Language of Fiction (London, 1966); Stanzel, Narrative
Situations in the Novel (Bloomington, Ind., 1969).

215
 . '

! r- » M

^»4 «r«nit«L *».!

111^4 . >'t

* 1 *. V 1 -J*! .*»!<’* •

't 5
•.t .rt,* •

‘ ,*•' :* I.'
'
- • '
« 1 lv»»

'
» jR i
*
' ^if »t» 1, .

.... •
•'
* -,
.' '

- ; I

Vi T '

• '
1 \ « '%V<^ 4 ti . .
.'

r<s-
». ,4

.
«'’.^.>* ' -
•’
• *•’-’• ***<' f< - '
. #v j‘

»
r «
,*
.
r ii •
.< > 4

SC-

k 4

* .• i 4

•
t
f
«

t I • »

I
 INDEX

Acts. See Symbolic acts; Utterances, 148-50, 153. See also Intentions,
as acts; Verbal acts authorial
Advertisements, 55-57, 209 Audience. See Listener; Reader
Aesthetic experience, 11-13, 57, Austin, J. 210
L., 93, 165, 198,
116-24, 211. See also Language, Autobiography and memoirs, 8, 31,
aesthetic functions of 48, 49, 136
Ahistorical structures, 15-16, 18, Avery, Milton, 208
21,51,56, 70, 95, 138, 207
Aristotle, 8, 13, 14, 20, 39, 41, 42, Bacon, Francis, essays of, 84-85
117, 165, 194 Bakhtin, Mikhail, xvii
Art:and cognitive activity, 11-13, Barthes, Roland, xviii, 177, 183,
116-24; and competent audi- 194, 195,212
ences, 150; and illustration, 52- Bartlett (anthology of quotations),
55, 56, 207-8; impersonality of 67-68
115; literature as a form of, 3-9, Bateson, Gregory, xviii, 207, 210-11
13, 24-28, 32, 119-24; medium, Beckett, Samuel, 185
concept of, 25-27; vs. nature or Biography, 4, 8, 10, 29-30, 44
natural events, xii-xiv, 11-13, Blake, William, “TheTyger,”
204-5; and perception, 26-27, 164-65
32-33, 48-49, 52-55, 116-24 pas- Bloomfield, Morton W., 214
sim, 211; and play, 116-24, 131- Body motion. See Gestures
32; visual, and poetry, 25-28, 32- Booth, Stephen, 172-73, 212
33, 52-55, 208. See also Mimesis; Boswell, James, Life ofJohnson, 48
Representation Brown, Roger, 209-10
Artistic design, 12, 39, 122-23, 145, Browning, Robert, “The Bishop

217
 1

INDEX

Orders His Tomb," 28 Context: and interpretation, 23,

Burke, Kenneth, xvii 32-37,52-55, 97, 104, 119-20,
Butler, Christopher, 206 121, 123-24, 146, 150-51, 171,
204; as meaning, 16-17, 21-22,
Campbell, Oscar James, 206 205; as signal of fictiveness, 126,
Categorical imperatives, 134-35, 129-30; of verbal transaction, 16-
154 17, 115
Catullus, “Oi/f et amo,” 53 Conventions: and advertisements,
Causal nexus and meaning, 94. See 55-57; as “antecedent classifica-
also Meaning, concept of tions,” 203; arbitrariness of, 88,
Champigny, Robert, xviii, 209 96, 99-100; concept of, in literary
Chatman, Seymour, 192, 197-201 theory, 82;and greeting-card
Children: and language, 10, 44-45, messages, 58-60; and identifica-
89-90, 92-93, 124-32, 147, 209- tion of fictive discourse, 10, 24,
10, 212; and mass media, 208, 28,47-48,84, 101, 104, 112-16,
212-13 204; in literature vs. in language,
Chomsky, Noam, 52, 177-82, 185, 193-94; and symbolic acts, 87-89;
201 variability of, 205; of verbal trans-
Citation:and proverbs, 70-73, 74; actions, 107-16, \20-2\. See also
as quotation, 64-69 Linguistic conventions
Clarendon, Edward Hyde, earl of. Copernicus, De Revolutionibus, 204
History of the Rebellion, 29-30 Covert categories, xiv, 45-46, 85
Cognitive activity: and art, 11-13, Criticism, literary. See Interpreta-
1 16-24, 204; and literary interpre- tion, literary; Narrative, theories
tation, 144-46; use of term, 21 of; New Criticism; Structuralism;
Cognitive feedback, 91-92, 109, 213 Stylistics

Cognitive play: and aesthetic experi- Culler, Jonathan, 175-85, 192

ence, 1 19-24, 153; and poetry,
144-46
Coleridge, Samuel Taylor, 42 Danto, Arthur, xviii

Communication: in art vs. dis- Daumier, Honore, 27-28

course, 111-12; duplicitous, 1 13- Definition, as problem in literary
14; and functions of language, theory, 43-50, 79-82, 206
85-86, 90, 193-96; telegraphic Derrida, Jacques, 211
model of, 83; theory of, and lit- Dickens, Charles: David Copper-
erary study, 178 field, 31; Hard Times, 125-26,
Competence, literary, 37-38, 150, 132
180-82, 199. See also Reader; Dickinson, Emily, 48
Reading Distancing and aesthetic expe-
Composition, literary, 20, 33-35, rience, 120-21
60-62, 148-49, 212 Drama, 8, 21 , 24, 25, 39-40, 44, 66,
Concrete poetry, 5, 64 84, 136, 142
Constable, John, 56 Dramatic monologues, 28

218
 INDEX

Eco, Umberto, xviii, 215 as, 69-75; quotations as, 65-69;

Economics. See Verbal transactions, signals of, 49, 104, 126-30, 207,
economics of 212-13; storytelling as, 127-28;
Eliot, T. S., The Waste Land, 64 term discussed, xii-xiv, 45-46,
Ellis, John D., xviii, 206 62-64, 83-85; texts of, 9-11, 20-
Epigraphs, 64, 66, 150 21, 30-31. See also Literature;
Epstein, E. L., 162-67 Poetry; Representation, of dis-
Ethics: of literary interpretation, course; Verbal art
114-16, 142-43, 146-54; of verbal Fictive persons, 112
transactions, 73, 90-91, 100-105, Finkel, Donald, “Angel Hair,’’ 67
112-16, 133-37 Finnegan, Ruth, 72-73
Events: historical, vs. ahistorical Fish, Stanley, 168-69, 172
structures, 15-16, 18, 21, 51, 56, Folklore. See Oral literature
70, 95, 138, 207; meanings of, Form-content relation, 161-62,
94-98, 137-38; natural, vs. art, 164-65
xii-xiv, 204, 205; representations Formula for linguistic analysis, 160-
of, 34, 112; symbolic, 94-98; 61, 170, 200-201
utterances as, 15-20, 21-23, 27, Formulas, verbal, 60-62, 126, 128
92-103, 137. a/so Verbal Framing, xiii, 119, 207. See also
events Mistakes and misinterpretations;
Extranatural discourse, 130-31 Reauthoring texts; Switching of
categories
Found poetry, 54-55, 66-67, 74,
False listening, 100-101, 114, 149 150
FCC (Federal Communications Fowler, Roger, 158-60, 162, 164-
Commission), 57 74, 188, 207, 213, 214
Fiction. See Narrative; Novels; Freeman, Donald C., 158-61, 167
Stories and storytelling Functional distinctions, xiv, 44-46,
Fictive discourse (Active utterances): 81-83,85, 105
aesthetic vs. nonaesthetic, 50-55,
74-75, 84-85, 119-20; greeting-
card messages as, 57-60; literary Games: verbal, vs. verbal art, 121-
works as, 10-11, 24-40, 110-16, 23; verbal transactions as, 105,
122-23, 127-28, 137-41; logi- 124, 145-46, 152-53. See also
cians’and linguists’ examples as, Play
50-55; meanings and interpreta- Genres, literary: and interpretation,
tion of, 32-40, 74-75, 122-24, 38, 204; as representations of
136, 137-54; vs. natural dis- types of discourse, 8, 30-32; tax-
course, xiii-xiv, 24-32, 46-50, onomies of, 198-99
74-75, 84-85, 110-16, 119-20, Gestures, 18-19,86,88, 113, 126-27
124-32, 136, 137-44, 148-49, 203 Gibson, Eleanor J., 211
{see also Reauthoring texts; Goffman, Erving, xviii, 207, 210
Switching of categories); proverbs Goodman, Nelson, xii, 3-13, 203-4

219
 1

INDEX

Greeting-card messages, 57-60, 63, terance, 49-50, 81-82, 116, 150-

64, 70, 75 51 {see also Reauthoring texts); as
Greimas", A.-J., 177, 184 designed effects, 148-50; in greet-
Grice, H. P., 210 ing-card verse, 58; as historically
determinate, 146-48; and literary
Haiku, 74 interpretation, 135-38, 140-43,
Hamburger, Kate, 209 146-54. See also Artistic design
Hanson, Camille, 209-10 Interior speech, 92, 108, 206
Henderson, Harold G., 209 Interpretation, general: as activity,
Hermeneutic: circle, 173; princi- 22-23, 32-37, 53-55, 66-67, 74-
ples, 154. See also Interpretation, 75, 119-24, 139-40, 143-46, 204;
general of letters, 23-24, 35, 139-44; of
Herodotus, 203 natural events vs. artworks, 204;
Hirsch, E. D., Jr., 133-37, 147, of proverbs, 68-69, 70-74; of
149-50, 154, 207, 213 symbolic acts, 87-89; of symbolic
Historical determinacy: of authorial events, %-104; of texts vs.

intentions, 146-48; of meaning, speech, 23; use of term, xv-xvi;

vs. historical indeterminacy, 33- of verbal vs. nonverbal events,
39,51,59, 68, 70-71, 137-46, 98-100. See also Interpretation,
152-54, 170-71 literary; Listener; Meaning; Mis-
Historical events. See Events takes and misinterpretations;
Hrushovski, Benjamin, 212 Understanding
Hymes, Dell, 210 Interpretation, literary: as cognitive
activity, 36-37, 119-24, 154; con-
and ethics of
straints on, 37-38;
“Ideal reader,” 172-73, 182-83, language, 133-54; and evalua-
199. See also Reader; Reading tion, 166-67; vs. nonliterary
Iliad, 1 interpretation, 32-40, 53-55, 73-
Illustrations, verbal and visual, 50- 75, 119-24, 137-44; and stylistics,

55, 64, 65, 75, 207-8 159, 168-73; types of, distin-
Indeterminacy. See Historical guished, 38-39, 124, 146, 151-54,
determinacy 206, 213. See also Interpretation,
Information: and aesthetic experi- general; Meaning; Reader;
ence, 1 16-23, 211; and cognitive Reading
activity, 1 17-18; as commodity, Intonation and tone, 18, 49, 104,
90, 107, 108; and language, 83, 120, 126-27, 128-29, 139, 169-70,
85-86, 90, 93-94, 98-100, 102-3, 171
128; relative to listener, 210 Iser, Wolfgang, xviii

Ingarden, Roman, xvii, 209

Inscription. See Texts Jameson, Fredric, 214
Intention, concept of, 88, 146-49. Jocasta (Oedipus Tyrannus), 102
See also Intentions, authorial Jokes, 120-23
Intentions, authorial: vs. assump- Jonson, Ben, 10, 42
tions, 38; and classification of ut- Joyce, James, Finnegan's Wake, 122

220
 INDEX

Keats, John: letters, 69, 84-85; “To tive theory, 197-201. See also
Autumn,” 10, 28; “To Sleep,” Conventions
148-49 Linguistics, 7, 16, 104; examples in,
“Kernel,” of narratives, 161, 186- as fictive discourse, 50-54, 58-59,
87, 194 136; and literary study, 14-15,
Kirshenblatt-Gimblett, Barbara, 157-201 passim, 214; transforma-
212 tional-generative, 168, 177-85,
Knowledge. See Cognitive activity 185-86, 200
Kreitler, Hans, 214 Listener: effect on speaker, 17, 108-
Kreitler, Shulamith, 214 10, 115; functions of language
Kristeva, Julia, 184 for, 92-94, 98-101; risks and
liabilities of language for, 101-5,
107-8. See also False listening
Lakoff, Robin, 210 Literature: as fictive discourse, 8-
Language: acquisition, see Children 11, 24-40, 137-38; as inscribed
and language; aesthetic functions natural discourse, 20, 109, 136;
of, 119-24, 193-%, 215; concept vs. language, as system, 178, 193-
of, in literary criticism, 162; func- 94, 196; problem of term and
tions of, for listener, 92-94, 98- concept, xii, 3-5, 8-11, 13,45-46,
104; functions of, for speaker, 79-85, 121, 150-51, 194, 195,
85-92, 93, 108-9; as game, 105; 203, 206. a/50 Composition,
literary vs. nonliterary, 79-85, literary; Interpretation, literary;
105; origins of, 131-32; poetic, Oral literature; Poetry; Texts;
vs. natural discourse, 36-37; Verbal art
referential theory of, 204; as sys- Logic, examples in, as fictive dis-
tem of conventions, %, 193-94. course, 50-54, 58-59, 63, 75, 84,
See also Fictive discourse; Natural 136
discourse; Utterances; Verbal Lord, Albert, 60-61
acts; Verbal events; Verbal play; Lying, 100-101, 103, 111, WA.See
Verbal transactions also False listening
Learning. See Cognitive activity
Letters: interpretation of, 23-24,
35, 139-44; as literature, 23-24, Macaulay, Thomas, History of
84-85; and novels, 8, 30, 31; and England, 20
quotation, 66, 69; texts of, 19-20, Marlowe, Christopher, “The Pas-
23-24, 35; and verbal formulas, sionate Shepherd,” 52
61 Matisse, Henri, 208
Levi-Strauss, Claude, 177, 183, 184 Meaning: concept of, 21-22, 36-39,
Lewis, David K., xviii, 210 51, 59, 70, 88-89, 94-98, 137-46,
Lichtenstein, Roy, 208 147; historically determinate vs.
Linguistic conventions: concept of, indeterminate, 33-39, 51, 59, 68,
96, 103-4, 193-94; and interpre- 70-71, 137-46, 152-54, 170-71;
tation, 37-38, %-101, 135, 137- metaphoric and parabolic, 66-68,
40, 147, 153, 171-72; and narra- 71, 141-44; new or emergent.

221
 INDEX

121, 124, 150-51; “original” vs. ances):and Active discourse,

“anachronistic,” 134-37; as pro- compared, xiii-xiv, 24-32, 46-50,
ositiohal exemplification, 123, 74-75, 84-85, 110-16, 119-20,
124, 141-44, 148, 152, 153; of 124-32, 136-44, 148-49, 203;
symbolic acts, 88-89, 147; of term discussed, xii-xiv, 15, 20,
symbolic vs. nonsymbolic events, 45-46, 84-85. See also Language;
94-98; total, 22, 94-98, 124; of Utterances; Verbal acts; Verbal
utterances vs. word-meaning, 21- events; Verbal transactions
22, 95. See also Interpretation, Nature vs. art or artifacts, xii-xiv,

general; Interpretation, literary 11-13, 205

Media, mass, 55-57, 212-13 New Criticism, 161
Medium, concept of, in aesthetics, Novels, 8, 10-1 1, 28-31, 43, 66,
26-27 136, 142, 151. a/50 Narrative
Herman, Moby Dick, 43
Melville,
Metacommunication, 207, 210-11 Oedipus (Oedipus Tyrannus), 102
Meter, 7-8,41-42, 110, 127, 128 Ohmann, Richard, 198, 209, 213
Meyer, Leonard B., 211 Oral literature: children’s, 126-29;
Millar, Susanna, 212 composition of, 60-61, 122, 212;
Milton, John: Comas, 38; Paradise and concept of literature or po-
Lost, 32 etry, 4-5, 43, 203. See also Prov-
Mimesis, xiv, 24-28, 162-66, 197. erbs; Riddles; Songs; Stories and
See also Mimetic repleteness; storytelling; Verbal play
Representation O’Toole, L. M., 185-88, 190-91
Mimetic repleteness, 52-55, 59-60,
74, 208 Painting. See Art, visual
Mistakes and misinterpretations, Parabolic meaning and interpreta-
47-49, 101, 103-4, 146-50, 207, 66-68, 71, 73, 144
tion,
210-11. See also Reauthoring Peckham, Morse, xviii, 214
texts; Switching of categories Perception and art, 26-27, 32-33,
Monism: and art vs. nature, 205; 48-49, 52-55, 116-24 passim,
and facts vs. fictions, 132; and 211. See also Aesthetic experience
literary vs. nonliterary discourse, Performance and literature, 4-10,
80-81, 206-7; and natural vs. Ac- 31, 56, 208-9
tive discourse, 46-47, 132 Persona, 28
Mukafovsky, Jan, 214 Pierce, Charles S., xvii
Music, 4-8, 27, 31 Plato, dialogues of, 43
Play, xiii, 105, 119-24, 147, 212.
Narrative: literary vs. nonliterary, See also Cognitive play; Verbal
185, 194-96; structural analyses play
of, 186-92; theories of, 176-201; Plays and playscripts. See Drama
universal, 183, 194-96. a /50 “Poetic license,” 111, 116
Stories and storytelling Poetry: attempts to define, 14-15,
Natural discourse (natural utter- 41-44, 45, 46, 79, 84-85; com-

222
 INDEX

position of, 33-34, 137-38, 148- Representation: in art, xii-xiii, 24-

49, 207, 212; as fictive discourse, 28, 32-33, 52-55, 65, 112, 207-8;
8, 24-40, 52-55, 68-69, 73-75, of discourse, xii-xiv, 8-11, 24-40,
84-85, 110-12, 115, 136-45; vs. 52-55, 59-60, 63, 65, 74, 84, 123,
greeting-card messages, 57-60, 208
64; interpretation of, 32-39, 74- Richards, I. A., 66
75, 138-46, 171-72 a /50 In- Richardson, Samuel, Clarissa, 31
terpretation, literary); and letters, Riddles, 121-23, 125
compared, 23-24, 35-36, 139-44; Robbe-Grillet, Alain, 185
as natural discourse, 28, 64-69, Roethke, Theodore, “My Papa’s
115, 149; vs. proverbs, 73-75; Waltz,’’ 208
and visual art, 25-28, 32-33, 52- Rutherford, John, 185, 188-94
55, 208. See also Literature; Ryle, Gilbert, 49
Verbal art
Pope, Alexander, Essay on Man, 43 Sanza (Central African proverb),
Pratt, Mary Louise, xviii 72-73
Prefabricated discourse, 57-64, 70 Sarraute, Nathalie, 185
Prince, Gerald, 214 Sayings. See Proverbs
Propositions, exemplification of, Sceglov, Y., 185-86
in literature, 141-43, 152-53, Scholes, Robert, 214
195-96, 213 Searle,John R., 198, 209, 210
Proverbs, 63, 68-69, 69-75, 136, Segmentation, 189
151,209 Semiotic. See Symbolic
Punctuation, 23 Semiotics, 178
Puns, 120, 123, 125 Shakespeare, William: Hamlet, 8,
Puzzles, 119, 121-23 24, 32-34, 39-40, 102, 123;
Henry IV, part 1, 66; Romeo and
Quotation, 19, 63, 64-69, 150, 208 Juliet, 123; Sonnets, 28, 32, 34,
35, 68-69, 168-73
Raleigh, Sir Walter, 105-6, 132 Shelley, Percy Bysshe, “Ozyman-
Reader: in literary theory, 81, 160, dias,’’ 67
168-74, 182-84, 197, 199; and Sidney, Sir Philip, 42, 127, 142
meaning, 32-40, 55, 150-51, Sincerity and insincerity, 58, 60,
170-72 (see also Interpretation, 61-62, 103
literary); relation to author, 109- Skinner, B. F., xvii

16, 124,150-51. 566 a /50 Reading Smith, Barbara Herrnstein, 203,

Reading, as activity or process, 6- 205, 206
8, 11-12, 121-24, 144-46, 173- Songs, 5-6, 45, 66, 73, 126, 129
74, 183-84. See also Interpreta- Speaker: and economics of verbal
tion, literary; Reader transactions, 85-92, 93, 102,
Reauthoring texts, 49-50, 58, 64- 108-10, 147-49; and ethics of
69, 70, 150 verbal transactions, 73, 99-105,
Relational richness, 121-22, 211 112-15, 135, 151

223
 INDEX

Speech. See Language; Natural dis- ances, 19-21, 23, 30-31, 36, 109,
course; Speaker; Utterances; Ver- 116, 136, 139
bal acts; Verbal events; Verbal “Thirty Days Hath September,” 43
transactions Todorov, Tzvetan, 177, 184, 191-92
Speech-act theory, 80, 81, 171, 198, Tolstoy, Leo: Death of Ivan Ilyich,
210 29-30; War and Peace, 11, 29
Stories and storytelling, 60-61, 125- Tomasevskij, Boris, xvii

30, 136, 185, 194-%. See also “To Mother” (greeting-card verse),
Narrative; Novels 58-59
Structuralism: and linguistics, 174- Truth, 28, 55-57, 70-73, 103-4,
85; and narrative theory, 174- 107-8
201; and stylistics, 167-68, 174- Turner, Roy, 210
75, 192-93, 214 Tynjanov, Jurij, xvii

Stylistics: as method of literary

study, 157-74 passim; and struc- Uitti, Karl D., 214

turalism, 167-68, 174-75, 192- Understanding, 22-23, 101-2, 124.
93, 214 See also Interpretation
Swift, Jonathan, Gulliver's Travels, Universality, concept of, in poems
204 and proverbs, 59-60, 67-68,
Switching of categories, 47-49, 54- 70-71
55,64-70,81, 112-15, \ \b. See Utterances: as acts, 18, 24, 85-92,
also Reauthoring texts 105, 137; as events, 15-23, 27,
Symbolic and
acts: defined, 87; 85-
92-103, 137; unspeakable, 108-
literary composition, 148; mean-
86- vs. words, 15-16,62-63,69-
11;
ings and intentions of, 147-49; vs. 70, 95. See also Language; Nat-
nonsymbolic acts, 87-88. See also ural discourse; Verbal acts;
Verbal acts Verbal events
Symbolic events, 94-98. See also
Verbal events Value, literary, theory of, 162-67
Verbal acts, 13, 85-92; and nonver-
Talking to oneself. See Interior bal acts, 18, 24, 31, 65, 69-70,
speech 87, 137; and symbolic acts,
Taylor, Archer, 71, 72 89; vs. verbal events, 93; vs.
Taylor, Jeremy, sermons of, 43 verbal forms, 69-70. See also
Telegrams, 61 Language; Natural discourse;
Tennyson, Alfred Lord, “Ulysses,” Utterances
10, 28 Verbal and authorial intentions,
art:

Texts: and concept of literature, 3- 148-50; and cognitive activity,

11, 13, 30-31, 81, 136; and mu- 1 16-24; and verbal play, 121-23,

sical scores, compared, 4-8, 9, 131-32, 203. See also Art; Lan-
31 ;
reading of, 6-8, 36, 173-74; guage, aesthetic functions of;

(see also Interpretation, literary; Literature; Oral literature; Poetry

Reading); relation of, to utter- Verbal events: vs. nonverbal events.

224
 INDEX

18-19, 94-100; vs. verbal forms, 111-16, 124, 143, 145-49. See
15-16, 18, 21, 51, 70, 95. See also also Language; Listeners; Natural
Events; Symbolic events discourse; Speakers
Verbal forms, 15-16, 18, 21, 51, Verse, 32, 42-43, 58-59. See also
69-70, 95 Meter
Verbal formulas, 60-62 Volosinov, V. N., xvii
Verbal play, 45, 84, 124-32, 136.
See also Cognitive play; Play
Whorf, Benjamin Lee, xiv
Verbal transactions, 85-116,
Words. See Verbal forms
133-54: economics of, 57, 62, 83,
Wordsworth, William, 28, 110; The
85-86, 90-91, 93, 100-101, 105,
Prelude, 48
107-12, 116, 119-20, 124, 135,
149; ethics of, 73, 84, 90-91, 100-
105, 112-16, 133-37, 146-54;
Xisto, Pedro, “Epithalamium II,”

between interpreter and his au-

64

dience, 151-53; narratives as,

195; between poet and reader. Zolkovskij, A., 185-86

225
 %

“« •*
/»rnT^} (•;
'
• •'
»

,U .' **;,*’ ••
' n it<i>»‘ v #f*i‘

» • ?
•
•
> fc
.*

• •
• •* (r.i*
f »(ltM I i 1

I t ^ I .
'

’ •^’
•
‘nyJ*

» *

.*•*
I <
V

. M
•*
\*

II >•
»

'f /I ( I
»

I * (^ 1 *
’

t -
 Part Two, originally delivered as the Witter Byn-
ner Lectures Poetry for 1977, examines the charac-
in

teristic functions of literature in relation to the social

economics of verbal transactions and describes the
conditions under which verbal structures acquire
value by serving aesthetic functions. Questioning the
extent to which literary hermeneutics can be based
upon appeals to a general "ethics of language," she
those aspects of literary meaning that
tries to clarify

are not merely elusive but essentially indeterminate.

Finally, in Part Three, Smith presents a vigorous
critique of the claims, methods, and theoretical as-
sumptions of contemporary stylistics and literary
structuralism. On the Margins of Discourse demands
that readers reexamine their basic notions of the
relation of literature to language and of literary
interpretation itself.

BARBARA HERRNSTEIN SMITH, professor of English

and communications at the University of Pennsyl-
vania, is the author of Poetic Closure, also published
by the University of Chicago which won the
Press,
Phi Beta Kappa Christian Gauss Award and the Ex-
plicator Award for 1968.

ISBN: 0-226-76452-4
Printed in U.S.A.

For further information on books of related interest,

or for a catalog of new publications, write:
Marketing Department
The University of Chicago Press
5801 S. Ellis Avenue
Chicago, IL 60637
 A Study of How Poems End
Barbara Herrnstein Smith H
Winner of the 1968 Phi Beta Kappa Christian Gauss Award in literary M
criticism and the Explicator Award for 1968
"Ranging from the Elizabethan lyric through free and syllabic verse
and concrete poetry, Poetic Closure is a learned, witty and richly
illustrated study of the behavior of poems. ... It can be read,
enjoyed, -'.i

studied by people who like reading poetry, including I would ;

— poets."—
""

suspect New York Times Book Review

"Barbara Herrnstein Smith emerges in her book on 'how poems end'
as one of the handful of pioneers in poetry criticism
today."— Poetry
1968 xvi,289 pages p

Cloth ISBN: 0-226-76342-0

Paper P381 ISBN 0-226-76343-9
:

LANGUAGE AND THE POET

Verbal Artistry in Frost, Stevens, and Moore
Marie Borroff
Language and the Poet undoubtedly one of the most significant criti-
is

cal studies of Robert Frost, Wallace Stevens, and

Marianne Moore yet
to appear. Marie Borroff, William Lampson Professor
of English at
of
Yale University, demonstrates an original approach to the study
literary style in which she examines language not
only in terms of
its intrinsic characteristics, but as part of the cultural inheritance in

which all speakers participate.

1979
Cloth ISBN: 0-226-06651-7

WRITING AND DIFFERENCE

Jacques Derrida
Translated, with an Introduction
and Additional Notes, by Alan Bass
"Derrida can safely be called the leading philosopher in France today
and, together with Jacques Lacan and Michel Foucault, the most
—
important intellectual presence." Geoffrey Hartman, Yale University
First published in France in 1967, Writing and Difference,
a collec-

tion of Jacques Derrida's essays written between 1959 and 1966,

has

become landmark of contemporary Fre.nch thought. Scholars and

a
students from all disciplines will find Writing and Difference an in-'
spiring and challenging work.

1978
Cloth ISBN: 0-226-14328-7

The University of Chicago Press will publish several other works by^
Jacques Derrida, including La dissemination translated by Barbara^
Johnson.

THE UNIVERSITY OF CHICAGO PRESS