Intention and Interpretation in Literary Criticism

Author(s): Göran Hermerén

Source: New Literary History, Vol. 7, No. 1, Critical Challenges: The Bellagio Symposium
(Autumn, 1975), pp. 57-82
Published by: The Johns Hopkins University Press
Stable URL: http://www.jstor.org/stable/468278 .
Accessed: 26/01/2014 19:41

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to
New Literary History.

http://www.jstor.org

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
 Intention and Interpretationin Literary Criticism*
GdranHermer6n

I. Introduction
HE INTENTIONAL FALLACY is surelyone of the most livelydebated
topics in the theory of literature. Monroe Beardsley and William
Wimsatt's by now classical article on this fallacy has been reprinted
in several anthologies, and the referencesto it are countless.1 If capacity
to stimulate and provoke scholarlydiscussionsis a measure of importance,
then theirpaper is probably one of the most importantpapers in aesthetics
published during this century,at least in the Anglo-Saxon world. How-
ever, in spite of the numerous articles and books that since then have
appeared on the intentionalfallacy, the problems raised by Beardsley and
Wimsatt cannot, in my opinion, be regarded as solved.
This opinion is not shared by every scholar. For a differentview I
would like to quote a statementfrom Richard Palmer's well-known book
on the principles of hermeneutics: "The New Criticism is essentially
right about the autonomy of the literary work; to look in a work for
the subjectivityof the author is rightlyheld to be a fallacy (the intentional
fallacy), and the testimonyof the author as to his own intentionsis cor-
rectlyregarded as inadmissibleevidence." 2 Here Palmer makes a number
of interestingassumptions,which I would like to make explicit, since they
do not at all seem to be uncontroversial.
For example, Palmer's use of the phrase "the autonomy of the literary
work" without explanation suggests (1) that according to him the mean-
ing of this expression is clear, and (2) that a sharp line can be drawn
between on the one hand the interpretation,understanding,and enjoy-
ment of a literarywork and on the other facts about the author's intentions
and the society in which he lived. Moreover, it is not clear to me what
Palmer means by "to look for the subjectivityof the author"; perhaps he
wants to suggest (3) that it is impossible to find in a work any signs or
indications of subjective states of mind of the author. Furthermore,he
takes for granted (4) that it is clear what the intentional fallacy con-
sists of, and (5) that it is indeed a genuine fallacy. Finally, he agrees
with the frequentlymade claim (6) that "the testimonyof an author as
to his own intention is . . . inadmissible evidence."
However, all the statements mentioned above seem to me clearly
dubious, and some even plainly wrong; and in the rest of this paper I
shall tryto show why.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
58 NEW LITERARY HISTORY

II. A Storyand SomeFacts

Some of the difficultiesraised by statementsabout intentionsof authors
can be introduced by the following story,which I heard once in the
lunchroom in Trinity College, Dublin.
As is well known, Samuel Beckett worked for some time as James
Joyce'ssecretary. One day when Joycewas dictating to Beckett,somebody
knocked at the door. "Come in," said Joyce (presumably without chang-
ing his voice very much) and continued to dictate. Beckett took down
"Come in"; and when he read to Joyce what he had taken down, Joyce
said: "Where did that 'Come in' come from?" Beckett explained that
Joyce had said it himself and started to erase it. But Joyce interfered.
He wanted it to be where it was. He was writing about strange coin-
cidences, and here he had a supreme example in his own text.
Looking at this storymore closely, we can see that intentionsof one
kind or another are involved in many ways. First Joyce did not intend
"Come in" to be in the text. Then he changed his mind and wanted to
include it. Beckett made two mistakes as to Joyce's intentions: firsthe
mistakenly thought that Joyce intended to include "Come in" in the
text, and later he mistakenlythought that Joyce wanted him to erase it.
Moreover, in addition to the problem whether Joyce intended "Come
in" to be in the text at a particular point, there is the furtherproblem
of what he wanted to say by including it in his book.
Whether this story is true or not does not matter very much in the
present context. The important thing is that it illustratesa number of
problems connected with the role of statements about intentions in
literarycriticism. It is, of course, easy enough to imagine endless specula-
tions by criticsas to what is meant by this sudden "Come in" in the text,
if we did not know the explanation. The storyalso shows that we must
distinguishbetween intentionsof differentkinds and on differentlevels,
which perhaps can be explicated by something like Austin's theory of
speech acts.3 Clearly, we must not confuse writing with the intention
of conveying a certain message and writing with the intention of mak-
ing money.
Two general problems suggest themselvesat this point. First: to what
extent can statementsabout what the author intended be tested? Cer-
tainly, there does not in this particular case seem to be any obvious
linguistic or literaryconventions we can appeal to in the way we can if
we want to find out what is meant by "plastic arm" in a poem from the
eighteenthcentury (written before "plastic" in the modern sense was in-
vented). Second, if these problems can somehow be solved and we have
reliable informationabout what the author intended, another question
arises: what relevance does this informationhave as evidence for and
against differentcritical hypotheses?
After these preliminaries,it may be well to state some facts. One fact
is that there is a great deal of explicit reference to the intentions of
authors even in the writings of critics who are generally regarded as

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 59

formalists,and who regard themselvesin this way. For example, in his

book on Finnegans Wake William York Tindall describes himself as "a
text man, concerned less with Joyce than with what he wrote."4 Never-
theless,it is not difficultto find statementsabout Joyce's intentionsin his
book. Thus at one place he makes the following point: "With this
effectin mind he inserted passages, beyond common readers, in uncom-
mon tongues.... Deliberately, he inserted little puzzles (Letters, I, 228,
250). Suppose we solve one or two of what he called his 'ninetyninean-
gular' mysteries. There are always more. Best not to worryabout them;
for they are an intended part of our experience and his effect. Joyce
meant the Wake, like the world around it, to be inexhaustible."5 Here
I have italicized a number of referencesto what Joyce intended-accord-
ing to Tindall, Joyce meant the Wake to be inexhaustible, and Tindall
talks about the intended part of our experience and about what Joyce did
deliberately and what he did in order to achieve certain effects.
Moreover, it is clear fromsome of Tindall's comments on the firstpage
of Finnegans Wake that he considers statementsby the author about how
his work is to be read (what he intended) to be admissible and relevant-
perhaps also conclusive--evidence for certain critical analyses, at least
sometimes: "That my painful-yet partial and imperfect-analysis is
more or less on the right track is proved by Joyce's own partial and im-
perfectexplanation of the firstpage in a letterto Harriet Weaver (Letters,
I, 247-48)." 6 The strong words "is proved by" in the quotation above
are interesting.They suggestthat Tindall does not regard Joycemerelyas
an ordinaryreader, as good or bad as any other; otherwisehe would not
attach that much importance to Joyce's own explanation.
Examples like these could easily be multiplied. But there is no need
to do so here. Instead I want to call attention to another fact which
likewise should be kept in mind during the subsequent discussion; and
that is simplythe fact that what criticswrite can sometimesbe interpreted
in many ways. Consider, for example, the followinginterestingpassage, in
which Tindall discusses the bedroom scene in chapter 16 of Finnegans
Wake: "Problem: is this bedroom scene an interruptionof the dream
or part of it? Is Earwicker-if he is the dreamer-dreaming that A. L. P.
gets up to comforta child or, waking from the dream, does he encounter
realityfor a minute or two before going to sleep again and dreaming the
dawn in? My guess is that all in this chapter is part of the dream." 7
How should this be understood?
The key question raised by this quotation seems to be this: what kind
of problem is Tindall here calling attention to? It can very well be
understood as a problem involving Joyce's intentions. Then the question
is: did Joyce intend the text to be read in this or in that way? But the
problem can also be construed in a differentway: can the text, given the
linguisticand literaryconventions of the author's time, be read in this or
in that way? Finally, the problem can furtherbe understood as follows:
should we read the text in this or that way, if we want to make the best
of the text,i.e., make it as interestingand rewarding-aesthetically and in

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
60o NEW LITERARY HISTORY

other ways-as possible, given certain linguistic, historic, and literary

facts?

III. SomeDescriptions
oftheIntentional
Fallacy
In addition to all this, it is not quite clear what the intentionalfallacy
is supposed to consist in. In their firstpaper on this topic, Beardsley and
Wimsatt make the followingpoint: "the design or intentionof the author
is neither available nor desirable as a standard for judging the success of
a literary work of art."8 This is a somewhat unfortunate formulation
which does not convey what I think the authors wanted to say; they
surelywanted to make more than the rather trivialpoint that the author's
intentionsare not relevant fordeterminingthe receptionof a literarywork
of art: every new writerprobably wants his firstnovel to be a master-
piece, but that does not mean that his work will be a success.
The authors have also restated the fallacy in somewhat differentways
later on. Thus in a paper on what they called the "affective fallacy,"
Beardsley and Wimsatt make the following summary of the basic ideas
in their original paper: "The Intentional Fallacy is a confusionbetween
the poem and its origins,a special case of what is known to philosophersas
the Genetic Fallacy. It begins by tryingto derive the standard of criticism
from the psychological causes of the poem and ends in biography and
relativism."9 But here the expression the standard of criticism is am-
biguous. Moreover, the definite article in that expression suggests that
there is only one standard of criticism,which certainly is not true.
However, in a later paper Wimsatt has corrected and restated his and
Beardsley's earlier formulation in the Sewanee Review article of 1946:
it, he says, "should certainlyhave read: 'The design or intention of the
author is neither available nor desirable as a standard for judging either
the meaning or the value of a work of literaryart.' "10 In what follows,
I shall use this last formulationas a point of departure for my discussion
of the intentional fallacy, though also this formulation is controversial
in a numberof ways.
For example, the expressionthe meaning of a literarywork of art raises
the same kind of problem as "the visual appearance of this object." An
object will certainly look quite differentto persons, observing it with
differentbackground knowledge and expectations from differentangles
under differentconditions. But which of these many appearances is the
appearance, and what principle could be used to decide whether one of
these many appearances is the appearance or not of that object? Similarly,
which of the many possible meanings of a literarywork of art is the
meaning? 11
But apart from that it is clear from this last quotation that whatever
the intentional fallacy may be, it is not one fallacy but several which
should be discussed separately; and in the next two sectionsI shall proceed
to discuss what I take to be the main forms or types of the intentional
fallacy.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 61

fromIntention
IV. Inferences toValue
I shall begin by considering some examples of reasoning, which might.
be cited as plausible instances of the intentionalfallacy:

Pattern I
(1) Premise 1: The author intended to achieve (convey, .. .) X in
creatingr
(2) Premise 2: The author did not achieve what he intended

(3) Conclusion: The literarywork r is good (bad, indifferent)

Now if reasoning according to this pattern is held to be a fallacy, then

probably the following kind of argument would also be considered
fallacious:
Pattern II
(1) Premise: The literarywork r is good (bad, indifferent)

(2) Conclusion: In creating r the author did not achieve what he

intended

But the question is: are these two patterns of argument fallacious?
In trying to answer this question, I shall concentrate on pattern I.
The firstthing that should be noted is that I deliberately,and provoca-
tively,have stated the general conclusion as "The literarywork is good"
rather than-as might have been expected-"The literarywork is bad."
The reason for this is simple. If the alleged discrepancy between the in-
tentions and the work is relevant at all, it may count both for and
against the literarywork, depending on whether the intentionsare good
or bad. If they are bad, we may be inclined to say that the author-
fortunatelydid not achieve what he intended.
But let us leave this complication aside and concentrate on the relations
between premisesand conclusions in pattern I, assuming that what holds
for that pattern will also hold, mutatis mutandis, for pattern II. Clearly,
the conclusion does not follow from the premises. But that is not enough
to show that the argument above is a fallacy. Nor does it sufficeto show
that the premises in many cases or even in most cases are not relevant as
evidence for the conclusion. The reasoning above may be poor and
dangerous, considered as an inductive argument,without being fallacious.
The basic question here is this: are the premises in pattern I ever
relevant as evidence for the conclusion? Tentatively, I would be inclined
to answer this question in the affirmative.This answer is based on the
assumption that informationabout the author's intentions (of a certain
kind) is relevant for determininghow a work of art is to be understood,

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
62 NEW LITERARY HISTORY

e.g., if it is to be taken ironically or not; and informationof this kind

appears at least sometimesto be relevant for determiningwhether a work
of art is good or bad. I shall return to this assumption later on and
try to defend it; meanwhile I shall give some examples to explain what
I have in mind.
Suppose that a poem can be read as a pastiche or as a parody, and sup-
pose that these notionsare at least partlyintentional,i.e., to decide whether
a poem is a parody or not, we need some informationabout the author's
intentions. Suppose also that originalityis artisticallyvaluable. Clearly,
if the poem in question is to be taken as a parody, it may be an original
work of art in the sense that nobody ever before has made a parody of
this kind. But if it is not to be taken as a parody or a pastiche, then it is
not a very original poem. In cases like this, it does not seem to be too
far-fetchedto say that informationabout the author's intentionsis indeed
revelant for judgments of value.
Let me add some other and perhaps more obvious examples. Suppose,
for instance, that we find out that a certain poem was intended to be
read at the funeral of an old friend of the poet. Suppose also that this
poem in formnn and tone is very gay. Would we not in such a case-know-
ing that the poet liked his friend and was sad to see him pass away-be
inclined to hold this lack of correspondence between form and intended
function against the poem? Is there not something odd about gay
funeral marches, sad wedding poems, and so forth?
Of course, it is easily possible to construe a fallacy in which statements
about the author's intentionsare involved. For example, if firstinforma-
tion about the author's intentions is used to decide what the work in
question means, and then this informationabout what the work means is
used to decide what the author intended, we clearly have an instance of
a fallacy. The reasoning is circular, and the circle is vicious. But there
is no need to introduce a specificname on this particular kind of circular
argument and call it "the intentional fallacy"; if we did, and wanted to
be consistent,we would have to introduce an indefinitenumber of names
on differentkinds of circular reasoning, depending on what the premises
were about.
By saying this I do not in the least want to reduce the historical im-
portance of what Beardsley, Wimsatt, and others have written on the
intentionalfallacy. Their papers were veryimportantas a reaction against
much loose and unverifiabletalk about the author's intentions,and they
pointed out a number of genuine difficultiesin checking statementsabout
what the author did or did not intend. I am only saying that the con-
siderations above, if sound, show that the so-called intentional fallacy is
not a logical fallacy like "denying the antecedent" or any other of these
classical fallacies which Beardsley himself has described very well in his
excellentbook, Thinking Straight.12

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 63

V. Inferences to Meaning
fromIntention
I shall now consider another line of reasoning, which perhaps may be
regarded as the standard instance of the intentional fallacy:
Pattern III
(1) Premise: The author intended to achieve (convey,...) X in creat-
ing r

(2) Conclusion: The work r means X

For example, it may be suggested that a novel is a satire of the federal
bureaucracy, since this is what the author wanted to make.
As in the previous case it might be argued that if reasoning according
to pattern III is fallacious, then the following kind of argument should
also be regarded as a fallacy:
Pattern IV
(1) Premise: The work r means X

(2) Conclusion: The author intended to achieve (convey, . .. ) X in

creatingr

But are these two patterns fallacious?

The crucial question here is this: is information about the author's
intentions ever relevant for or against statements about what a literary
work of art means? The answer to this question obviously depends on
how the key expressions, notably means and intended, are interpreted.
They could very well be interpretedin such a way that reasoning accord-
ing to the patterns above becomes triviallytrue, as "A, because B," if
A is defined in termsof B. But means and intended could also be defined
in such a way that reasoning of this kind becomes clearly false.
I admit that it is difficultto draw a precise line here between termino-
logical and substantial philosophical issues; theyare intertwinedin a very
complex way. One could argue, and I have myself argued elsewhere,
that we should sharply distinguish between three differentconcepts of
meaning: intended meaning (authorial meaning), meaning according
to conventions and traditions (textual meaning), and meaning defined
in terms of effects (response meaning). If it is furthersupposed that
these concepts may never be combined with each other,and that meaning
in discussionsof literaryworks always means "meaning according to con-
ventions and traditions,"then reasoning according to these patterns turns
out to be fallacious.
Like Monroe Beardsley, George Dickie, and others, I think that the
distinctions between the three kinds of meaning mentioned above are
very important. But it does not follow fromthis that the last two assump-

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
64 NEW LITERARY HISTORY

tions are correct, i.e., that the three concepts may never be combined,
and that meaning in literarydiscussionsalways means "textual meaning."
Thus the thesis that patterns III and IV are fallacious should be rela-
tivized to certain assumptions and definitions: if they hold, the thesis
is correct, otherwise not. This shows the importance of the theory of
meaning taken for granted by literarycritics.
Before going any further into this very difficultarea, I would like
to referto an example where a literarycritic has tried to illuminate a text
by autobiographical material. In a paper on Kafka's Das Urteil, J. P.
Stern begins by explaining why he has chosen to write about this book:
"I have chosen it for interpretationfor three closely connected reasons.
First,as an outstandingand outstandinglyclear example of the way Kafka
draws autobiographical material into his fictions. An interpretationof
the storyshould thereforeilluminate one side of his literaryundertaking,
the other side of which is the fictionalizingof his autobiographical and
private writings."13 This example is especially interesting,since Stern is
well aware of the theoreticaland methodological problems raised by using
autobiographical material.
On the same page Stern continues with these programmatic remarks:

The factthatall Kafka'simportant worksare markedby strongautobiographical

elementsis one among several reasonswhy any interpretation of them should
proceedwithparticularcaution. They remainof coursefictions.There is, how-
ever,an especialtemptation to read intothemattitudes,
ideas,and evaluationsnot
foundtherebut in his diaries,letters,or otherdocuments.The autobiographical
pendantto 'Das Urteil'is Kafka's 'Briefan den Vater,'writtensevenyearslater.
I shallquote fromit herebecauseit providesseveralpointersto an understanding
of a storythatis difficult,
notin orderto imposeon the storyan evaluationand a
meaningderivedfrom'Der Brief'but missingfromthe storyitself(whichis what
manycriticshave done). The storyis as completeas any Kafka everwrote.How
completethatis,we shallsee.14

One of the interesting of Stern'sinterpretation

features is thathe is
usinga textthatwas written sevenyearsafterDas Urteilto shedlight
on it. If Sternby thatstrategy
wantsto tellus whatKafkameant,or
could have meant, by certain passages in Das Urteil,he is clearly assuming
that thereis a constancyin the author'sway of thinkingand feelingof
such a kind that we are entitled to disregard the seven years that have
passed betweenthe two works.
It should be noted that I am not saying that this assumption is either
tenableor untenable.Nor am I sayingor implyingthatwhatSternwants
to do is to tellus whatKafka meant. I am onlysayingthatif thisis what
he wantsto do,thenhe is making
thatparticular He could,
assumption.
of course, also be doing a number of other things; the concept of inter-
pretation is by no means very clear. For example, he could be suggesting
a possibleway of readingthe text,a way thatmakesthe bestof the text
and has the additional virtue of fittingin with what we know about the
author's attitudes and feelings.
I mentionthis in order to call attentionto the complexityof the

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 65

notion of meaning in theses like "The author's intentionhas no relevance

whatsoever in determiningthe meaning of a literarywork of art." The
key concepts in this thesis are intention and meaning, and it seems that
every stand taken as to whether the anti-intentionalistthesis stated above
is correct or not presupposes an explicit or implict theoryof meaning--
and that part of semantics is one of the most difficultand also most con-
fused areas in contemporaryphilosophy.
To sum up: the thesis that the intentional fallacy is indeed a fallacy
should be relativized to a theory of meaning. If something like the
theoriesof meaning proposed by Austin, Grice, Searle, and their followers
is accepted, then it follows trivially that some authorial intentions are
always relevant; at least one kind of intention is built into the very con-
cept of meaning. In that case the intentional fallacy is of course no
fallacy. But I shall argue that intentionsare relevant even if these theories
of meaning are not accepted, namely, in the cases of irony, parody,
pastiche, and satire. If what I will say below about these literarymodes
of expression is true, then general statements about the irrelevance of
biographical informationcannot be correct.

VI. Literaryand Nonliterary

Intentions
But it is not enough to try to clarifythe concept of meaning; we must
also focus on the concept of intention. Suppose we can distinguishat
least in a crude way between literaryand nonliterarymeanings, intentions,
and so forth. Then the simple and unprecise problem-how is meaning
related to intention?-may be replaced by at least four, using the follow-
ing combinations as a point of departure:

(A) literarymeaning-literary intention

(B) literarymeaning-nonliterary intention
(C) nonliterarymeaning--literaryintention
(D) nonliterarymeaning-nonliterary intention

The anti-intentionalistposition may be modified,and saved, by bringing

in some of the distinctionsabove.
For example, it may be argued that if the concept of literature is to
have any use at all, then information about the author's nonliterary
intentionsare irrelevantfor determiningthe literarymeaning of his text.
This modified thesis is much weaker-and more plausible-than the un-
qualified stronger thesis. Whether the weaker thesis is tenable or not
depends of course on whether the distinctionsbetween literaryand non-
literarymeaning (and respectiveintentions) can be made intelligible,and
how theyare drawn. This is a problem I shall now turn to. The scholarly
discussions of, e g., Balzac's writingsshow that it is by no means an easy
task to state these distinctionsin a clear and yet uncontroversialway.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
66 NEW LITERARY HISTORY

Two crucial questions in this context are: What is intended by the

author? How is his intentionexpressed? If the formerquestion is taken
as a point of departure, we can at least distinguishbetween, say, the case
in which an author in or by writinga novel

(1) wants to convey a certain message,

(2) wants to become rich and famous.

By classifyingmessages in differentways, furtherrefinementscan of course

be obtained. It would seem quite natural to draw the distinctionbetween
literaryand nonliteraryintentionsin such a way that only (1)--or only
some subset of (1) -should count as a literaryintention.
If, however, the second question is taken as a point of departure, the
distinction between a literary and a nonliteraryintention concerns the
way in which the intentionis expressed rather than the object of the in-
tention. For example, here we may consider the case when the intentionis

(a) expressedin a text,i.e., by words and sentences,

(b) expressedin otherways,e.g., by gestures.

By saying that an intentionis expressed in a text I mean to say that the

intention is recognizable in the text by intelligentand informedreaders,
i.e., by readers familiar with the linguistic and stylisticconventions used
in formulatingthe text.
These two bases of division can verywell be combined with each other.
Without going into details at this point, I would like to suggest that
such a combination could fruitfullybe used as a point of departure for
the attempt to define literaryintention and literarymeaning. This sug-
gests that biographical informationcan very well be used in the process
of discovery,as a way of directing the critic's attention to possible ways
of understandinga literarytext. But then his hypothesesabout how the
work is to be understood must be checked against the text and what we
know about the linguistic and stylisticconventions of the author's time.
If the alleged intentionscannot be recognized in the text, however hard
critics try to find them there, then that hypothesis is discarded; only
intentionsrecognizable in the text are relevant.
However, what happens if some readers recognize the intentionsin the
text whereas others fail to do so? The majority is not always right,so
we cannot simplycount the number of people on both sides. The standard
procedure in such cases seems to be this: those who recognize the in-
tentionswill have to explain why they think they are there, and this will
involve (a) describing the linguistic and other conventions the author
can have known and used, and (b) showing that given these conventions,
the text could or should be read as expressing the intentionin question.
But let us leave this complication aside for a moment and return to
the modified thesis that informationsabout the author's nonliteraryin-
tentions are not relevant for determining the literary meaning of his

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 67

text. Is this thesis acceptable? It seems to be a compromise between the

extreme intentionalistand anti-intentionalistpositions that may well be
worthexploring.
Interestinglyenough, Wimsatt appears to go a long way in the direction
of thismodifiedthesisin his 1968 paper. Here he seems prepared to accept
arguments from intention "as it is found in, or inferredfrom, the work
itself,"15 though it is not quite clear what he means by "the work itself"
in this context. Moreover, anyone who has read E. D. Hirsch's book on
the validity of interpretationscannot fail to be struck by how the sharp
contrast outlined in the firstchapter (where he attacks Beardsley and
Wimsatt) between the intentionalistand the anti-intentionalistposition
becomes less and less clear.16Hirsch's comments on how statementsabout
intentions are checked suggests that he is probably not talking about
authorial intentions of the kind Beardsley and Wimsatt were attacking
in their firstpaper, judging from the examples used to illustrate their
thesis.
Even so, there are some literaryphenomena and modes of expression
that will constitute serious difficultieseven for the weaker anti-inten-
tionalist thesis; these include irony, satire, parody, pastiche, and (what
is sometimescalled) Freudian symbolism. Here it seems that information
about the author's nonliteraryintentionsmay be relevant for determining
the meaning of the text. Since space is limited, I shall here single out
so-called Freudian symbolismand irony for special consideration, rather
than argue for and against the anti-intentionalistthesis in a general and
abstract way. I shall begin with the former,since that gives me the op-
portunityof commenting on the distinctionbetween conscious and sub-
conscious intentions.

Intentions
VII. Consciousand Subconscious
The distinction between literary and nonliterary intentions is not
enough, in however sophisticated a way it is drawn; we must also take
into consideration that there may be. various degrees of consciousness of
intention. The distinctionbetween conscious and subconscious intention,
gradual and filled with problems as it is, crosses the distinctiondiscussed
in the previous section. It would obviously be a mistake to assume that
an author always is aware of all the motivations behind his creative
work. But to say that is not to say that artistic creation is a random
process. Much interestingresearch has been done in this area over the
last decades, but more remainsto be done.
The term Freudian symbolismis sometimes used in this context. But
this is a rather unfortunateand misleading expression. For one thing, it
suggeststhat iconographical symbols (like a unicorn, standing for purity,
or an hourglass, standing for the brevityof life) and Freudian symbols
(like a snake, standing for a phallus) have something in common, by
virtue of which they both belong to the same class or category, though

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
68 NEW LITERARY HISTORY

they differfrom each other in some respects. But it is not obvious that
Freudian and iconographical symbolsare related to each other in the way
red and white wines are. If X is a Freudian symbol of T, then there
is a causal relation between X and T-if someone dreams (writes, etc.)
about snakes in a certain way, then this is according to Freud a symptom
or an indication of hidden or repressed sexual conflicts,desires, and so
forth. But if X is an iconographical symbol of r, then no causal con-
nection of this kind is involved; instead, there is a tradition or a con-
vention, according to which X stands for 2.17
However, the fact that there is a clear distinction between typical
instancesof each kind of symboldoes not exclude that there are a number
of cases which seem to fall between the typical cases or to be a combina-
tion of them. For example, if a snake according to Freud is a phallic
symbol in the sense explained above, and if Freud's writingsare widely
read, then authors and artistscan include snakes in their works as sexual
symbols,assuming that people are familiar with at least certain aspects
of Freud's works. In this way we can observe a transitionfrom Freudian
symbolismto iconographical; and the distinctionbetween Freudian and
iconographical symbolism must not be drawn in such a way that this
is excluded.
I shall not try to complicate the argument of the present paper by
distinguishingamong kinds of unconscious and subconscious intentions.
A rather more important distinction in the present context (from the
point of view of literary scholarship) is that between differentuses of
Freud's and Jung's theories about subconscious conflicts,archetypal pat-
terns,and so forth. I shall firstdescribe some differentuses in a schematic
manner, and then I shall illustrate them by examples from literary
criticismand finallyexplain why I think the distinctionsare important.
The case of Freudian symbolismis complex. To begin with, we may
focus on the following use-or perhaps, rather, family of uses-of psy-
chological and psychoanalytical theories in literarycriticism:

(1) The critic tries to show that the author (a) has got the idea to,
say, a part of the story from such a theory, or (b) writes in order to
illustratesome aspect of that theory.
For example, Tindall and other writers on Joyce have recognized the
importance of Freud's influence on some aspects of Joyce's work. Joyce
used verbal puns like "freudful" in Finnegans Wake, and according to
Tindall he got some of his jokes fromFreud,18as well as the father-killing
theme. Tindall also connects some stylisticfeatures of the book with
Freud: "Freud's dream-work--censorshipwith all its displacements and
condensations-may account in part for the extreme obscurityof Butt's
Oedipal story... ."19
As far as I can see, such an attempt to connect literaryworks of art
with psychological or psychoanalyticaltheories is no more than a special
case of tryingto show that the author in some sense or other was in-
fluenced by the theorydiscussed. The form and extent of the influence

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 69

may of course vary considerablyfrom case to case, and all the theoretical
and methodological problems that arise in testing hypotheses about
artisticand literaryinfluencewill arise in this context also. But it would
carry us too far away to discuss these problems here.20
In (1), the critic is tryingto say something about the author. Where
did he get some of his ideas? Was he, or was he not, aware of this? But
the author need not be involved in psychoanalytic criticism, and (1)
should be distinguishedfrom:
(2) The critic uses a theoryto describe, explain, organize, and illumi-
nate certain features of a literarywork, e.g., the motives and actions of
the characters in a play, or the stylisticfeatures of the work.

For instance, Erwin Steinberg characterizes Bloom's fetischisticinclina-

tions and his masochism with quotations from Freud's writingson these
topics; after a long quotation from Freud, Steinberg continues: "This
passage was quoted at length, because all three categories fit Bloom so
well. His erotogenic masochism is plainly exhibited all through the Circe
episode, but particularly in his excited grovelling before Bella Cohen
(passim) and in his watching through the keyhole the love-making of
Molly and Boylan." 21 This example is instructive,since it shows that it
may be difficultto maintain a clear distinctionbetween (1) and (2) in
cases when the author (as here) is contemporarywith, or later than, the
psychological or psychoanalyticalwriter. But it is not difficultto find an
instance of (2) which is not an instance of (1) : Ernest Jones's classical
explanation of Hamlet's delay.22 Jones is here applying Freud's theoriesto
the persons in the play, treating them as if they were persons from real
life. In so doing, Jones does not have to presuppose that Shakespeare had
read Freud; thus no anachronism is involved.
As I have already pointed out, nothing need be said about the author
at all in case (2). But a careful reading of what Jones and some of his
followershave writtenon Hamlet shows that occasionally they do make
references to the author; and then we are in a methodologically quite
differentsituation:

(3) The critic uses a literarywork-together with generalizationsfrom

a psychological or psychoanalytical theory-as evidence for or against
hypotheses about the existence of subconscious intentions,conflicts,etc.,
in the author.

For example, Jones argues in his book on Hamlet that Shakespeare leaves
Hamlet's behavior as puzzling as it is and in need of explanation, because
Shakespeare had unconsciouslyprojected his own reactivated and formerly
repressed oedipal complexes.23 And in an earlier work he had argued
that the conflictin Hamlet "is an echo of a similar one in Shakespeare
himself,to a greater or lesser extent with all men." 24
The main reason for the distinction among the three cases above is
the differencein methodological situations. In (1), the critic obviously
needs to demonstrate that the author was familiar with the relevant

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
70 NEW LITERARY HISTORY

theorybut not that this theoryis correct,or at presentis considered to be

correct by experts on psychology and psychoanalysis. But the situation
is exactly reversed in (3). Whether the author is or is not familiar with
the theoryused is of no significanceat all, but it is essential that the theory
be correct. Thus objections against critical interpretationsthat are rele-
vant in case (1) are irrelevant in case (3), and conversely.
Moreover, it can very well be illuminating to characterize the plot in,
say, a Greek classic tragedy in Freudian terms-as is done in case (2) -
even though the tragedy was writtenseveral centuries before Freud was
born. Subconscious intentions,desires, conflicts,and so forth,obviously
existed long before Freud had made an attempt to describe and explain
them; thus the critic cannot in that case be accused of anachronism. But
in case (1) it would of course be a valid objection if it could be shown
that the author of the literarywork in question was not, and could not
possibly have been, familiar with the relevant psychological theory.
Now, to what extent are the author's intentions, conscious or sub-
conscious, involved in the three cases above? As to (1), his intentionsare
clearly involved, and they may be both conscious and subconscious. As
to (2), it does not seem necessary to say or assume anything about the
author's intentions; the actions and motives of the persons in the play
are described in terms of some theory,and the main thing here is that
what happens in the play fitsthe theory. Finally, subconscious intentions,
conflicts,and so forth, are clearly involved in (3); the point of that
approach is preciselyto describe featuresof the author's psychologythat
he himselfmay not be aware of and mighteven have denied emphatically.
What followsfromthisconcerningthe relevance of intentionsin literary
criticism? To begin with, we can here distinguish at least three anti-
intentionalistpositions:

(A) None of the three kinds of investigationsdescribed above can

be carried out, since the necessary evidence simply is not available.
(B) Even if such evidence were available, literaryscholars should not
tryto carryout investigationsof these three types,particularlynot the last
one.
(C) It is both possible and desirable to carryout investigationsof these
three kinds, but this can be done without informationabout the author's
nonliteraryintentions.

It is probably very difficultto defend (A). But even if (A) is false, (C)
may be true; and I shall thereforeconcentrateon the last position.
Since one counterexample is sufficient,let us consider case (3). Sup-
pose the generalizations in the relevant psychological or psychoanalytical
theories always were formulated in such a way that the if-clause only
described the verbal behavior of human beings. Then we would never
have to consider information about the author's nonliteraryintentions
and behavior. But even a most cursoryreading of psychoanalytictheories

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION
71I

shows that this is not so; and thus I fail to see how project (3) could be
carried out if only the author's literaryintentionsare considered. In that
case, the evidence would be far too scanty, and the difficultiesof con-
necting the psychoanalytictheorywith the literarywork overwhelming.
But the tough anti-intentionalistneed not give in yet. Perhaps he will
admit that informationabout an author's nonliteraryintentionsmight be
useful, and perhaps even indispensable, if one wants to carry out certain
research projects, e.g., something like (3). But he can always deny that
projects of that kind are worthwhile. It may be difficultto settle such an
issue, but then it is at least clear that in the last analysis the basic dis-
agreement between the intentionalistand the anti-intentionalistin this
particular case is of a normative, rather than empirical, kind.

VIII. Irony
As we have seen, the radical anti-intentionalistposition can be stated
as follows: facts about the author's intentionsare never relevant as evi-
dence for or against critical hypotheses. Consider now the followingkind
of case, which I am sure any literaryscholar could illustrateby examples
fromhis own experience.25
An author has used a particular sequence of words in a novel or a
poem. According to the linguistic and stylisticconventions (traditions,
etc.) of that time, these words may be read as genuine praise of someone's
fine qualities, but they may also be read ironically. Thus we cannot by
consulting a dictionarysingle out one of these interpretationsas anachro-
nistic. Moreover, both interpretationsfitin with the context. We cannot
thereforesingle out one of the proposed interpretationsby saying that
it makes nonsense of the work. Furthermore,let us also suppose that the
proposed interpretationsare quite different,and that the author intended
the poem to be read ironically. In such cases it would seem fairlyobvious
that informationabout what the author intended is relevant in deciding
how the work is to be understood.
Well, is it? The anti-intentionalistmay not be willing to give in so
easily. He can say that this simply shows that the poem is ambiguous.
Moreover, if irony is defined in a suitable way, his thesis will turn out to
be true. Thus let us not jump to conclusions; before going any further,
we must tryto make explicit the views on irony presupposed by both the
intentionalistand the anti-intentionalist.Perhaps their theses in this case
are also relative to certain definitions and theories (or fragments of
theories).26
It is possible to distinguish among at least the following three kinds
of views on irony:

(a) The internal view. Irony is an internal feature of a literarywork.

Whether a work is ironic or not depends on the presence of certain
specifiable features in the work.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
72 NEW LITERARY HISTORY

(b) The circumstantial view. Irony is not an internal feature of a

literarywork. Whethera workis ironicor not dependson the circum-
stancesin whichit is publishedor read.
(c) The intentionalview. Neitherthe internalnor the circumstantial
view is correct.Whethera literaryworkis ironicor not dependson the
intentionsof the author.

Needlessto say,each of theseviewscan be made moreprecise-and ac-

cordinglybe replacedby a familyof others-byclarifying the keyexpres-
sions depends,internalfeature,circumstance, and intentions.But for
the limitedpurposeof the presentpaper it is not necessaryto go into
furtherdetailsat thispoint.
In a personalcommunication to the author,Monroe Beardsleyseems
to preferthe firstof the threeviewsabove, thoughhe sayshe sometimes
waverson the questionof circumstances:
There has to be a signal,an internalclash of somesortthat givesthe workthe
ironic twist (as I would argue we can find cues in, say, Swift's 'Modest Pro-
posal . . .'). Without that signal, the work is simply not ironic. Now the author
may have wished it to be ironic, may even think he has succeeded in making it
ironic-that is what we find out (I would say) in consultingbiographical sources:
what he intended it to be. But such informationdoes not increase the probability
that it is ironic unless it leads us to redouble our effortsto find the subtle signal,
and thus may indirectlycontribute to our discovery that it has a subtle irony,
afterall. The issue that is settled by biographical informationis not the same issue
that we are in doubt about: namely whetherthere is ironypresent.27

I shall now makea fewbriefcommentson the pointsmade above.

Clearly,Beardsleyis rightin sayingthatbiographicalinformation does
not settlethe questionwhetherthereare literarysignalsof the kind he
suggestspresentin the work. Thus a good case can be made for dis-
tinguishing amongthe threeviewson ironydescribedabove. One would
only wish that literaryscholarsshould devotemore timeand efforts on
makingthorough and well-documented studies of these literarysignals.
But even so, one mustask: what does the alleged importanceof these
distinctions prove?
The distinctions among (a), (b), and (c) suggestthe followingap-
proaches to a more comprehensive and systematictheoryof irony. In
the firstcase, we need an inventory and a classification
of different kinds
of literarysignals,like exaggerations, etc., and a theoryabout how they
work. In the secondcase, we need a similarinventory and classification
of different kindsof circumstances, and a theoryabout how theywork,
i.e., affectthe understanding of literature.In the thirdcase, finally,we
would want a furtheranalysisof the relevantkind of intention.This
analysisshould be at once psychological,conceptual,and phenomeno-
logical.
The keyphrasein Beardsley'sstatement above seemsto be this: "There
has to be a signal,an internalclash of some sortthatgivesthe workthe
ironictwist."Now what is meantby this? If one were to argue against

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 73

the hard-line,anti-intentionalistposition one should stressthe importance

of distinguishingbetween ironyand evidence of irony. The literarysignals,
the circumstances of publication, the information about the intentions
of the author-all this can be used as evidence for or against hypotheses
about irony,but this evidence should not be confused with the concept of
irony.
Thus I suggest that the situation could be described in this way. The
literarysignals-internalclashes,exaggerations, etc.-mentionedby
Beardsley do notconstitute the irony;rather, theyare evidenceof the
presence ofirony.Theyindicate thattheworkis ironical in thesensethat
the authorwas not he did not want his wordsto be taken
literally serious;,
or at theirface value. Couldwe always,bytheway,merely by
contemplating a text,decidewhether theproposals madeby theauthor
are exaggerated, and whether theseexaggerations are suchthatthetext
shouldbe takenironically? In myviewthisquestionmustbe answered
negatively.
Butdoesthismeanthatironyis a totally intentionalistic
concept, and
thatthe possibility of failure-i.e.,thatthe authorintendedto write
ironicallybut failed-is excluded?No. In such a case, the situation
seemsto be this: theauthordid notwantto be takenseriously, and he
useda number ofliterarysignalsto makethisclear,and he believedthat
hehadsucceeded.Buthewaswrong.The signals heusedweretoosubtle,
theaudiencetoounsophisticated, or thecircumstances ofpublication and
readingunhappy.Thustheworkwas intended to be ironical,butit was
notunderstood inthisway.
As to the definitionof the keyconceptof irony,I wouldpersonally
notbe preparedto advocatea definition onlyin termsof intentions, or
onlyin terms ofresponses,butrather a definitionin terms ofa combina-
tionofa number ofdifferentfactors:a literary workis ironical, if and
onlyif theauthorusedcertaindevicesin orderto makeclearthathe
did notintendhisreadersto takehimliterally, and he wantedthemto
understand thatthiswashisintention. To thiskindof definition we can
easilyadd clausesto theeffect thatthe readersdid understand his in-
tention.Thuswe mustconsider thecommunication situationas a whole
whenwe are aboutto decidewhether a literary
workis ironicalor not.
Thus,to sumup, givena certaindefinition of theconceptof irony,
it can be arguedthatthethoroughgoing anti-intentionalist
confuses the
evidenceof thepresenceof ironywiththeconceptof irony.Whether
thischargeis correct or notthenboilsdownto thequestionof howthe
conceptof ironyis to be defined;and in orderto decidethis,we must
startby indicating whatkindof definition we want. Is it a descriptive
definition,a stipulation,
or somekindof explication? Clearly,it is not
a stipulation. Andevenif it is an explication, it shouldnotdeparttoo
muchfromordinary usage,and thisusageis described in dictionarieslike
theOxfordEnglishDictionary.
Turningnow to the definitions
of ironyin thatdictionary,
we findthe
followingexplanationof the meaningof that term: "A figureof speech

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
74 NEW LITERARY HISTORY

in which the intendedmeaning is opposite of that expressedby the

words used; usuallytakingthe formof sarcasmor ridiculein which
laudatoryexpressionsare used to imply condemnationor contempt."
Though thisdefinition does not definitelysettlethe issue,it clearlylends
supportto the attemptto defineironypartlyin intentionalistic terms,and
to thecriticism made of theanti-intentionalisticposition above.
That ironyin literary is used in something
criticism verymuchlikethis
sense is supportednot onlyby what is said in some workson the theory
of literature,28but also by quotationslike the followingone, takenfrom
Lawrence E. Harvey'srecentbook on Samuel Beckett: "We recognize
the bitterironythatmimicscrueltypartlyin rage,partlyin the certainty
of understanding the case, partlyto jolt the readerinto awareness;and
we can guesstheadmixtureof personalpain and pugnaciouscouragein a
pun like 'cough up yourT.B.' "-29 In this commentto Beckett'spoem
Ooftisch,the critic'spoint seems to be that the intendedmeaningis
oppositethe literalone; thuswe have ironyin the sensedefinedabove.
Since exactlythe same kind of argumentcan be made concerning
parody,pastiche,and satireas concerningirony,I shall not repeatwhat
I have writtenabove. It is also clear that parody,pastiche,and satire
are intentionalnotions,both in the sense thattheyrequirean object-a
parodyis alwaysa parodyofsomething-andin thesensethattheauthor's
intentions are involvedin the definition of the concepts. The definitions
of these termsin standarddictionariesare also explicitlyor implicitly
stated in intentionalterms. But this does not, of course,exclude non-
intentionalevidencein the attemptto check the hypothesis that a par-
ticularpoem is, say,a pastiche.

IX. Externaland InternalEvidence

The thesisthat the intentionalfallacyis indeed a fallacyis combined
with,and based on, (1) a distinctionbetweenwhat is called externaland
internalevidence,and (2) a thesisthatonlyinternalevidenceis relevant,
and that externalevidencesuch as biographicalinformation is inadmis-
sible. It has oftenbeen pointedout thatthisdistinction
is not veryclear,
and thereis an examplein the originalarticleby Beardsleyand Wimsatt
that bringsthisout rathernicely: theirdiscussionof T. S. Eliot's The
Waste Land.
Beardsleyand WimsattconsiderEliot's notesboth relevantand illumi-
natingfor thosewho want to understandthe poem, and theysolve this
problembyregardingthenotesas partof thepoem: "We mean to suggest
by theabove analysisthatwhereasnotestendto seemto justifythemselves
as externalindexesto the author'sintention, yettheyoughtto be judged
like any otherpart of a composition.. ..." 30 Of course,if the notesare
part of the composition,theymustbe classifiedas internalevidence;and
we can use themto supportcriticalinterpretations withgood conscience,
accordingto Beardsleyand Wimsatt.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 75

But this raises a number of intriguingproblems. Suppose that T. S.

Eliot (a) had not published these notes to The Waste Land but kept
them to himself instead, or (b) that he made notes-which he did not
publish-to some of his other poems. After his death, these notes are
found among his unpublished papers. If it is permissibleto use the pub-
lished notes as evidence, why can we not use the unpublished notes? Let
us consider both the possibilitythat we can use the unpublished notes as
evidence for or against interpretationsof Eliot's poems, and the possibility
that we cannot do so.
Suppose firstthat we can (i.e., are allowed to) use the notes also
in case (a) and (b). Then we have to face the followingquestion: why
can we not--if they are alive-ask the authors of other poems to make
notes to their poems or-if they are dead-try to reconstructhow they
wanted theirpoems to be understood? It seems to me that once you allow
the author's comments as evidence in one situation, it is difficultto stop.
It then also becomes difficultto maintain a clear and nonarbitrarydistinc-
tion between external and internal evidence.
Suppose, however, that we cannot use the notes in case (a) and (b) as
evidence for or against critical interpretations.Then we must ask: why?
The only differencebetween the notes that we can use with good con-
science as evidence-because they are regarded as part of the poem and
hence as internal evidence-and the notes in case (a) and (b), which
we are not supposed to use as evidence, seems to be this: the latter notes
are not published.
But this differenceis quite irrelevantin the present context. Whether
an author's comments on his poem are published or not, they have to be
examined criticallyand carefully,if they are to be used as evidence. If
an author can make misleading statementsabout his own intentions in
his diary (either without being aware of it, or in order to deceive future
generations of literary scholars), then why can he not put into print
misleading statementsabout his own intentions?
Nevertheless, Beardsley and Wimsatt have made a very interesting
suggestion by arguing that the notes should be regarded as part of The
Waste Land. This is a good point which provides much food for thought.
For one thing, it shows that the boundaries of a poem, as it were, are
not given once and for all. But what does this prove? Cannot this sug-
gestion be used against the distinction between external and internal
evidence, and against the talk about the "autonomy of the literarywork"?

X. Literatureand Communication
If the general argument in this paper is roughly correct, what con-
sequences follow for the studyof literature?How could and should litera-
ture be approached, if some intended meanings are not only relevant
but, at least sometimes,important?
One way of studyingthe intended meanings is to study the linguistic,

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
76 NEW LITERARY HISTORY

stylistic,and symbolicconventions (traditions,etc.) at the time and place

where the work was created. At this point, Wittgenstein's notion of
language games suggestsitself. We can imagine the author and his readers
playing a complex and sophisticated language game, much more compli-
cated than the ones discussed by Wittgenstein. The author cannot follow
a rule alone, and if he uses a "private language" (in a strictsense), com-
munication breaks down. The development of a complex language game
has a history. In the course of this history,traditions have been estab-
lished, and knowledge of these traditions is indispensable to those who
want to know what is going on.
This approach to literature-the study of traditionsor conventions of
various kinds-is based on the assumption that the author wants to com-
municate something to his readers. In order to be able to do this, he
must use a code that is familiar to his intended audience. Even Joyce
does this-also in his neologisms. To understand what, say, finiskey
means, one must know what finish,or possiblythe Latin finis,and whiskey
means, and this of course every literarycritic does.
But this is not all. The author uses a code the readers are familiar
with,or gives them clues in differentways, and he wants them to recognize
that he has used this particular code. Thus even in this case, wants and
intentionsare built into the very process of communication. In this study
of intentionson differentlevels, a wide range of historical and linguistic
methods could be used. The evidence for or against particular critical
hypotheses has, of course, to be sifted carefully; and techniques for this
have been developed in differentlinguisticand historical disciplines.
Finally, the following complication in the interpretativework is im-
portant. Facts about traditions existing at a certain time and place do
not explain everything. Artists and writers do not merely work within
a tradition, they also work against a tradition; they depart from tradi-
tions and create new traditions. Joyce is perhaps an extreme example of
this, but he is far from the only one. In fact, this seems to be part of a
general explanation of how our language develops and is enriched with
new metaphors, how words acquire new or partly new meanings, and
so on.
From what I have said so far, it should be obvious that I would like to
challenge a commonly held dogma in literarycriticism,namely, the idea
that poets and writersin their literaryworks do not communicate any-
thing, and that talk about communication in literature somehow misses
the point of what literature is about or makes the concept of literature
superfluous. This dogma is sometimes supported by the following kind
of argument. Consider, it is said, some lines in a poem, e.g., the following
lines from one of W. B. Yeats's later poems:

That old man climbed; the day grew dim;

Two swanscameflying up to him.
Linkedbya goldchaineach toeach,
Andwithlowmurmuring laughingspeech,
Alightedin thewindygrass.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATIONI AND INTERPRETATION 77

Is what Yeats writes here true? Did two swans come flyingup to the
old man? And were they really linked to each other by a gold chain?
Since these questions are absurd, it is argued that the poet does not com-
municate anything; obviously it is then taken for granted that only true
and false statementscan be communicated.
Clearly the question of truthand falsityin the ordinarysense is irrelevant
as far as poetic and fictional statements go. I would not only accept
but insiston this. But I would also insistthat it does not follow from this
that poets and writersnever communicate anything. Thus, if the propo-
sitional theoryis the view that fictionalstatementsexpress propositionsand
are true and false in the ordinary sense, and the communication theory
is the view that authors communicate somethingto their readers,we must
not conclude that the communication theoryis wrong since the proposi-
tion theoryis wrong.
On the contrary,the assumption mentioned above (that only true and
false statementscan be communicated) can easily be criticized. It seems
to remain from the days when the main functionsof language was con-
sidered to be descriptiveor informative. But this was long ago. The con-
cepts of communication, and the models of communication processes
developed in a number of disciplines, can in my view provide a fresh
point of departure, if we want to get a clear picture of the role of state-
ments about intentions in literaryscholarship. Of course, we must not
assume that there are no differencesbetween what is communicated in
literature and outside literature, and how this is communicated. This
has to be decided by empirical investigations. But the concept of litera-
ture, and the distinction between literature and not-literature,is by no
means very clear. In that situation it should be temptingto test a simple
model of communication on literaturein order to find out if and to what
extentit works.

VI. A SimpleModel and SomeofIts Consequences

The main features of this admittedly simplified model are indicated
by the followingdiagram:
socialand artisticenvironment
and
linguistic
knowledge traditions
literary knowledge
interests author..t text- reader interests
values values
intentions whatthetextis intentions
needs,etc. about (describes, needs,etc.
symbo-
expresses,
lizes)
feedback
socialand artisticenvironment

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
78 NEW LITERARY HISTORY

A diagram of this kind could, among other things,be used to distinguish

two radically differentapproaches to literature. In the firstapproach,
the scholar's interestis focused on clarificationof the text. To elucidate
the literarywork he uses (independent of the text) informationabout
the author's knowledge and interests,about the literary and linguistic
traditions of the time, about the responses of contemporaryreaders and
what the text is about.
In the second approach, the situation is reversed. The text is used
as evidence for or against other hypotheses,e.g., about the author's knowl-
edge and interests,about the literaryand linguistictraditionsof the time,
about the responses of contemporaryreaders, or about what is described,
symbolized,etc., in the text. Obviously Dickens' novels contain a great
deal of informationabout how people in his day behaved, were dressed,
about the roles of men and women, about the relations between parents
and children, rich and poor, and so forth. Similarly,a literarywork can
be studied as a symptom of neurotic tensions in the author or of class
conflictsin his society.
These two approaches can very well be combined with each other,and
with differentkinds of studies of other works of art. The importance of
distinguishingbetween them should be obvious: if firstthe literarytext
is used as evidence of what the author intended, and then this information
about what the author intended is used to decide what the text meant, a
vicious circle is not far away. But fortunately,literaryscholars do not
start fromscratch, and they do not study just one work at a time.
It is sometimes said that there is no intended effectwith a literary
work beyond securing what Austin called uptake,32i.e., the understanding
of what the speaker or writer wanted to say. The purist conception of
literature seems untenable to me, however. Many satires, parodies, and
pastiches have indeed had other intended effects;the authors wanted to
call attention to injustice and inhumanity,to change society in one way
or other. Some of the works of Jonathan Swift, for example, clearly
belong to this category.33 And today we find an increasing number of
mixtures between fictionand documentary social studies. But according
to the purist conception, they are by definitionnot literature.
Be that as it may, if an author wants to achieve or avoid a particular
effect-whether uptake only or other effectsas well-he writes in such
a way that he thinkswill achieve this. The way in which he writes will
then be based on assumptions about the readers, partly of psychological
character, partly concerning their interests,expectations, and knowledge
of literaryand linguistictraditions. If the reader is intelligentand critical,
he will tryto figureout what the author wants to say; and this attempt
will then be based on a number of assumptions concerning the writer's
interests,expectations, and knowledge. Thus the situation is more com-
plex than is indicated by the diagram.
The relations between the text and what the text is about is another
controversial topic, discussed under many differentheadings such as
"literatureand reality,""meaning in literature,"or "truthand literature."

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 79

But what reality? What relations? These problems do not disappear if

we follow the suggestionof the Russian scholar Juri Lotman and say that
literature is a model of reality. We would then have to distinguishbe-
tween differentkinds of literarymodels and between differentrelations
between a model and what that model is a model of. Thus a number
of refinementscould be suggested at this point.
Of course, the model outlined above could also be developed in other
directions,and there are many other models of communication in theo-
retical literature. But these models do not necessarilyexclude each other.
Rather, they focus on differentaspects of the process of communication.
For example, those who are interestedin the reception of literaryworks
and the responses to them, rather than in the interplay between writer
and reader, may want to focus on partly differentaspects of the process
of communication, aspects which are described more fully in the models
suggestedby scholars like Schramm, Westleyand Maclean, and Maletzke.
But on the whole, it seems to me that models of the kind outlined in
this section give a fairlyadequate picture of what is going on in literary
criticism. Literary scholars want to discover meanings that fit the text
and can have been intended. They want to avoid anachronisms, and
for that reason they study the linguistic and literary traditions of the
author's time, on the assumption that they presumably are known to the
author. They study whether a proposed interpretationor explanation fits
with what has been said by the author in his other literaryworks, and
with what is known about his nonliteraryviews and behavior. They try
to find out how the work has been read and understood by different
readers contemporarywith the author, and so on. In other words, what
the critics are doing, it seems to me, is shedding light on some of the
factorsin the communication situation according to the model above.

and Interpretation
XII. Philosophy
In concluding this paper, I would suggest that a diagram of the kind
discussed above also can be used to make clear what assumptions have
been made in interpretinga literary text, and to clarify the so-called
hermeneutic circle and the role of what in the German philosophical
traditionis called the "Vor-Verstlindniss"in interpretation. Then we are
moving on to a number of intriguingand important philosophical prob-
lems.
As is obvious from Palmer's survey,34there are many theories dealing
with problems of this kind. Unfortunatelythese theories are stated in
a difficultand obscure terminology,which is not likelyto appeal to literary
critics unfamiliar with the German metaphysical traditions. Moreover,
these theoriesare veryabstract; theyare far fromwhat is actually done in
literary criticism and from the more detailed methodological rules de-
veloped in literary scholarship. These abstract and general theories
thereforeneed to be tested against empirical material.
Such a test could be carried out in several ways. For example, one

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
80 NEW LITERARY HISTORY

could start along the lines suggested by Morris Weitz in his book on
Hamlet and make comparative studies of what has been writtenover the
years about a particular literarywork, such as Dante's Divina Commedia,
Fielding's Tom Jones, Stendhal's Le Rouge et le noir, or Arthur Miller's
Death of a Salesman. If one chooses a fairlycomplicated and well-known
work as object of analysis, the material is likely to be very rewarding.
Thus the point of departure will be quotations fromwritingsof literary
critics and scholars, i.e., specific and concrete examples of how different
scholars at differenttimes and places have worked with differentper-
spectives (theoretical and normative frames of reference) and suggested
a variety of interpretations. This material can then be analyzed both
historicallyand systematically.In the firstcase, the analysis will pave the
way fora more accurate and precise historyof literarycriticismthan those
presentlyavailable. In the second case, the analysis can clear the ground
for a discussion of a number of intriguingphilosophical problems.
Such analyses will show that these interpretationssometimes supple-
ment each other and sometimes compete, or at least appear to compete.
But it is not always easy to decide whether differentproposed interpreta-
tions exclude each other or not. If, however, theydo exclude each other,
one must of course ask what methods could be used to choose among
them, assuming that literarycriticism is not to be based on guesses or
wishfulthinking.
Material of this kind will provide a powerful challenge to analytical
philosophers. It illustratesdifferentmethodological approaches and what
can be achieved by differentmethods. Clearly, it calls attention to the
role of a number of implicit assumptions in interpretationof various
kinds, empirical assumptions as well as presuppositions of understand-
ing, world views, and views on man. Moreover, such material demon-
strates that literaryworks in a certain sense are "open" or have no fixed
boundaries; critics and scholars have interpretedthem in differentways
in the past and there is no reason to think that they are not going to
continue to do so in the future.
Discussions of material of this kind will shed interestinglight on a
number of concepts used in literarycriticism,for instance, understanding,
interpretation,explanation, intention, and literature. But it will also
highlight a number of methodological assumptions often tacitly taken
for granted rather than discussed openly and systematically. Such dis-
cussions can probably also help to lay the foundation for a synthesisof
studies of the syntacticand semantic structuresof literarytexts and, for
instance,studies of these textsas psychologicaldocuments or as expressions
(in several senses) of social and economic conditions.

XIII. ConcludingRemarks
This paper has been metacritical rather than critical in the sense that
I have not tried to defend or criticize any particular interpretationof
any of the literary works I have referredto. Rather, I have tried to

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
INTENTION, COMMUNICATION, AND INTERPRETATION 8I

study some arguments used by critics to support or criticize certain

hypothesesabout literaryworks of art.
I began by discussingthe intentionalfallacy and have then systematically
tried to widen the perspective. It appears that both the intentionalist
and the nonintentionalistpositions are based on theorieswith an ideologi-
cal function: the theoriesare attemptsto articulate and defend rationally
an approach to literature which is considered fruitfuland worthwhile.
This is not to say that the stand one takes on thisissue is arbitrarybut that
the fundamental issues are nonempirical; they concern normative ques-
tions.
Therefore, it is not likely that the basic problems raised by Beardsley
and Wimsatt will ever be definitelysolved. The discussion of the in-
tentional fallacy does not concern a narrow theoretical issue. Though it
is now more than twenty-five years since Beardsley and Wimsatt's original
article appeared, it is still being discussed in the scholarly journals; and
these discussions are likely to continue as long as we are seriouslyinter-
ested in literatureand in literaryhistory.
Clearly, the issue of intentionalismis important in the sense that the
position one takes on it has a bearing on several other issues in aesthetics
and in the methodology of literature. For example, it has relevance for
the question as to what methods should be used in studies of literature,
i.e., for the future development of literary scholarship. Thus should,
as some have proposed, the traditional biographic literary history be
replaced by synchronicand diachronic studies of literaryforms? Or, as
has been suggested by others, by studies of the impact and reception of
literaryworks? Or, as still others have argued, by the social historyof
writers?
I would also like to add that the problems raised by the intentional
fallacy concern a number of academic subjects. Here we have an area
where interdisciplinarycooperation is both possible and desirable. Such
cooperation would, I am sure, be of mutual benefit to linguists,phi-
losophers, and psychologists,as well as to literaryhistorians.

UNIVERSITY OF LUND

NOTES
* This paperis based on notesmade fora conference on thetheoryand method-
ologyof literaryhistoryin Bellagio,Italy, 1973. Earlierversionsof partsof the
presentpaper have since been presentedin lecturesat Abo Academy,Finland;
Ume&University, Sweden; and TrinityCollege,Dublin.
1 "The IntentionalFallacy,"firstpublishedin theSewaneeReview (1946); rpt.,
e.g.,in W. K. Wimsatt,Jr.,The VerbalIcon (Lexington,1954) and in Philosophy
Looksat theArts,ed. JosephMargolis(New York,1962).
2 Richard E. Palmer,Hermeneutics:Interpretation Theoryin Schleiermacher,
Dilthey,Heidegger,and Gadamer(Evanston,Ill., 1969), p. 246.
3 Cf. J. L. Austin,How to Do ThingswithWords(Oxford,1962); WilliamP.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions
82 NEW LITERARY HISTORY

Alston, Philosophy of Language (Englewood Cliffs,N. J., 1964); and John Searle,
Speech Acts: An Essay in the Philosophy of Language (Cambridge, 1970).
4 William York Tindall, A Reader's Guide to Finnegans Wake (New York, 1969),
p. 25.
5 Ibid., p. 23; my italics except the thirdand the last one.
6 Ibid., p. 32; firstitalics mine.
7 Ibid., pp. 284-85.
8 "The IntentionalFallacy," Sewanee Review, 54 (1946), 468.
9 The Verbal Icon, p. 21.
10 Wimsatt, "Genesis: A Fallacy Revisited," The Discipline of Criticism, ed.
P. Demetz et al. (New Haven and London, 1968), p. 222.
11 Cf. the discussion in my Representation and Meaning in the Visual Arts
(Lund: Miromedelsf6rlagen, 1969), p. 31.
12 Monroe Beardsley,ThinkingStraight (Englewood Cliffs,N. J., 1950).
13 J. P. Stern, "Franz Kafka's 'Das Urteil': An Interpretation,"German Quarter-
ly,45 (1972), 114.
14 Ibid., 114-15.
15 Wimsatt,"Genesis," p. 210.
16 E. D. Hirsch, Jr., Validity in Interpretation(New Haven, 1967). For Beards-
ley's reply,see Monroe Beardsley, The Possibilityof Criticism (Detroit, 1972).
17 For a furtherdiscussion of iconographical symbolism,see Ch. 4 of my Repre-
sentation and Meaning.
18 Tindall, Finnegans Wake, p. 52.
19 Ibid., p. 197.
20 These problemsare more fullydiscussed in my Influence in Art and Literature
(Princeton, 1975).
21 Erwin Steinberg,The Stream of Consciousness and Beyond in Ulysses (Pitts-
burgh, 1973), pp. 214-15.
22 ErnestJones,Hamlet and Oedipus (New York, 1949).
23 Ibid., Ch. 4.
24 Ernest Jones,Essays in Applied Psychoanalysis (London, 1932), p. 59.
25 There are also similar examples in the historyof ideas, e.g., Henry Dodwell's
ChristianityNot Founded on Argument (1742), to which David Berman has
called my attention. It was firstinterpretedliterallyand accordingly attacked by
the left (the freethinkers)and praised by the right (the conservatives). Then it
was interpretedironicallyand attacked by the rightand praised by the left. Though
asked publicly (in print) to clarifyhis intentions,the author refused to comment
on his work. Cf. also Leslie Stephen, English Thought in the Eighteenth Century
(New York, 1962), I, 148.
26 Cf. ?V above, where the thesis that the intentional fallacy is a fallacy is rel-
ativized to certain conceptsand theoriesof meaning.
27 In a letterto the author.
28 Cf. Wolfgang Kayser, Das Sprachliche Kunstwerk (Basel, 1948), p. 111.
29 Lawrence E. Harvey, Samuel Beckett: Poet and Critic (Princeton, 1972),
p. 157.
30 Wimsatt,The Verbal Icon, p. 16.
31 See ?I and n. 2 above.
32 Stein H. Olsen, "Authorial Intention,"BritishJournalof Aesthetics,13 (1973),
228.
33 In his JonathanSwift and Ireland (Urbana, 1962), Oliver W. Ferguson makes
the followingpoint concerningthe parody employedby Swiftin A Modest Proposal:
"The rhetoricalfunctionof his ironic use of the economists'technique is to create a
feelingof horrorthat will be commensurablewith his terribleanger" (p. 175).
34 Seen. 2.

This content downloaded from 139.82.115.34 on Sun, 26 Jan 2014 19:41:49 PM

All use subject to JSTOR Terms and Conditions