M O N R O E C. B E A R D S L E Y

FICTION

AS R E P R E S E N T A T I O N

As humankind descends from its Age of Gold to ages marked by baser metals, fulfilling the gloomy prophecies of far-sighted philosophers, a stiff penalty for such cognitive progress as may be achieved along the way is, we have been warned, the degeneration of poetry into prose. When Sir Philip Sidney argued that "the p o e t . . , nothing affirms, and therefore never lieth," he was gracefully rebutting an ancient indictment against the poets that threatened again in the new shape of Puritanism- the immediate target of his Defense of Poetry being Stephen Gosson's School of Abuse; or, a Sundry Pleasant

Invective against Poets, Pipers, Players, and Other Caterpillars of the Commonwealth. There have been those who thought that Sidney's concession was a high price to pay even for exemption from such serious guilt as that. of being a caterpillar of the commonwealth. But in our time there is very wide acceptance of the view that the price is right-except that in recent years developments in the philosophy of language have shown us a better, if much more prosaic, way to formulate and generalize Sidney's thesis: the poet does not perform illocutionary actions, and therefore does not run the risk of failing to meet the responsibilities incurred in the performance of such actions. I do not say that everyone who has a theory of fictional discourse would agree with this contemporary version of Sidney. Anyway, I agree with it, and I will offer my defense of it. In the course of doing so, I will also give my answer to what is evidently the next question: if the act of writing a poem does not constitute (or, in Alvin Goldman's sense, generate) t h e performance of an illocutionary action, then what kind of action does it constitute (or generate)? There are now in circulation a number of answers to this second question, making essential use of such terms as "mimesis," "make-believe," "pretending," "imitating." And philosophers who have given these answers have contributed greatly to our understanding of the nature of fiction. The particular view I present is not radically different from the others, but I hope it is somewhat more satisfactory in important respects. Synthese ,16 (1981) 291-313. 0039-7857/81]0463-029t $02.30. Copyright © 1981 by D. Reidel Publishing Co., Dordrecht, Holland, and Boston, U.S.A.

292

MONROE

C. B E A R D S L E Y

Before coming to the main argument, I should make two preliminary remarks, designed to discourage expectations that would otherwise be legitimate but disappointed. First, there is an easy solution that I do not propose to adopt, or even to discuss at length. It commends itself to those attached, or at least attracted, to the new schools of critical theory that emphasize the reader's response, and to others that embrace a form of neostructuralism. On this view, fiction is not a kind of text but a way of taking or treating texts. No doubt there is a frame of mind in which we can approach any text, however poignant or eloquent, even the most absurd or malicious, bracketing our concern for its communicative aspect, lifting it out of the stream of life. But I do not believe we can give an adequate account of what it is to take a text as fiction in this way without giving an adequate account of what it is for a text to be fiction. For to take x as fiction must be somehow to treat x as though it is fiction. Moreover, even if we can achieve that triumph of the aesthetic point of view, treating all texts as fictive, we shall not have erased the distinction between those texts that are appropriately so treated and those that are not, between those that invite play and those that call for engagement as real discourse. Second, I have to set aside here one of the most complex and difficult philosophical problems about fiction. This concerns the ontological status of fictional entities (Kilgore Trout, the Mill on the Floss, the ransom of Red C h i e f ) - o r better, the logical analysis of sentences containing names and definite descriptions of nonexistent persons, objects, and events. A great deal of ingenious and sophisticated work has gone into developing theories to deal with this problem, and the main possibilities have been sorted out, developed, and, of course, attacked. ~ One familiar way of posing the central puzzle is to note the paradoxical fact that, although in one quite plain sense, it cannot be true that Anna and Vronsky were lovers, since there were no such persons in reality, yet in another equally plain sense it must be true that they were l o v e r s - a t least it is not false to say so, in the way it would be false to say than Anna and Vronsky were married, or that they never met. Somehow, although Tolstoy was not composing history or biography when he wrote the words of Anna K a r e n i n a - h i s tale is fiction, if anything i s - nevertheless, his writing those words provides a ground, or a base, that makes true our statements reporting the

FICTION

AS REPRESENTATION

293

doings and undergoings, the actions and sufferings, of his characters. But the logical connection here has not yet been worked out to everyone's satisfaction, and until that is accomplished we find it hard to avoid talking, in a philosophically troubling way, about giving true descriptions of unreal worlds, as though the contents of such worlds could be described just like the contents of the actual world in which w e - but not Anna and V r o n s k y - live. So I do not promise any help with this problem here. We may agree from the start that if a text includes sentences containing names or descriptions of the nonexistent, then that text is fiction. But then our question will still be: what are we attributing to the text when we call it fiction, when we so classify it? Moreover, it may be that not all fiction contains such names or descriptions. In that case we must still ask what this larger class consists of, and try to give an answer that will (among other things) explain how texts that refer to the nonexistent are a subclass of that larger class of fictional texts.

The account of fiction that I favor has been worked out in different ways. My version begins with the concept of a text, which (in order to start somewhere) I take as comparatively unproblematic. A text is a syntactically ordered string of words in a language, and it may be written or oral. It is, of course, not an individual inscription or utterance, but a text-type, of which inscriptions and utterances are instances. Producing a text is, strictly speaking, producing an instance of a text. Usually we produce a text not just for the sake of doing so, but in order to perform an action of a quite different sort that can only be performed via text-production: that is, an illocutionary action. Thus if I should say "Pardon me!" under certain circumstances (say, after bumping into you), I would be doing at least two things: (1) uttering these words and (2) begging your pardon. Begging pardon is one kind of illocutionary action. If you were to reply, "That's all right," you would be performing an illocutionary action of another kind, namely pardoning me. Illocutionary actions are of numerous general k i n d s - s u c h as inviting, threatening, deploring, regretting, and affirming (what Sidney said the poet doesn't do) and also of indefinitely many more specific kinds - such as inviting one's friend to

294

MONROE

C. B E A R D S L E Y

a party, threatening to burn down a house, and affirming that the gods are licentious (which is what Plato said poets do but ought not to do). Sidney also said that the poet doesn't lie. Lying, too, is an illocutionary act, but a more complex one, which involves affirming or denying, but also other conditions-for example, that the speaker believe that the proposition he is asserting is not true. If this condition is lacking, it doesn't matter what the speaker says: he simply fails to lie. Here is a clear example of the essentially rule-governed character of all illocutionary actions: they are possible only because there are rules or conventions built into the language that take the following form: Producing text T under conditions C generates illocutionary action/. For example: If the patient is terminally ill, and if the doctor knows that the patient is terminally ill, and if the doctor speaks in a serious straightforward tone of voice, and if the doctor addresses his words to the patient, and if the patient hears and understands what the doctor is saying, then in uttering the words "You will recover," the doctor is lying to the patient. Perhaps I have left out some of the requisite illocutionary-act generating conditions; in any case, I have not put in too many, I'm quite sure, for if any of my conditions is lacking, the illocutionary action will not come off. Each illocutionary action belongs to an act-type, and has a set (or sets) of conditions under which actions of the type are generated. Sometimes some of the conditions are vague, or uncertain, or in dispute; but always some are unquestionably requisite and others unquestionably not. As with every action that has some degree of complexity, so with illocutionary actions: they can not only be performed, when circumstances permit, but also represented. Waiting on table is a complex action that involves various elements, including the actual handling of food. But the gifted mime can represent the action with the most minimal gestures - without food, or a diner to serve, or even a table or tray. He must somehow select a distinctive or selective movement or sequence of m o v e m e n t s - s a y a way of holding his arm that would hardly be used by anyone who was not carrying a tray. But once he

FICTION

AS REPRESENTATION

295

has sketched in this feature, which singles out his movements as imitative of waiting on table, we can read what he represents. My use of the term "represent" here is close to its use of visual art. There is still much controversy about what visual representation involves- in the sense for which I use the term "depicting," as when we say that a painting depicts a barn or a wedding. But I hold that here, too, it is a matter of selective matching-some shapes, colors, line-segments, regional qualities, or gestalts are captured and perceived to be those characteristic of the kind of object or event depicted. So when the television actor says "Pardon me," and we know he knows that he has not really committed a fault, we cannot say he is actually begging pardon, but only that he is representing an illocutionary action of begging pardon; he has dramatically depicted a begging pardon. And when we read the opening words of a sonnet by John Keats, Bright Star! Would I were steadfast as thou a r t . . . we recognize at once a serious deficiency in the conditions under which the text was produced. It is addressed to a celestial body, and reads like an attempt to praise it or express admiration to it. But illocutionary actions require what J. L. Austin called "uptake": what I alluded to above as the addressee's hearing and understanding what has been said. Since the star cannot fulfill this condition, the purported illocutionary act does not get generated by the words; but since there is enough in these words, in their context of the whole sonnet, for us to fill out a kind of illocutionary action that would have been performed were all requisite conditions present, we must read the words as representing an illocutionary action of that kind. Fiction, then, is the representation, not the performance, of illocutionary actions. My account is, of course, extremely sketchy, and perhaps will remain sketchier than we could all wish, even after I have tried to plug some of its holes and explore its implications. One major question is whether we need a distinction, apparently unprovided so far, between representing an illocutionary action and merely failing to perform one. An actor on the stage who tries to shoot another a c t o r - b u t m i s s e s - s e e m s to have gone beyond the mere representation of homicide, which calls for sticking to blank cartridges. Maybe he was carried away by a passion for realism, but Actor's Equity will

296

MONROE

C. B E A R D S L E Y

take a dim view of the proceedings. No doubt a simple way of making this distinction would be to invoke intention: if he intends or tries to kill and fails, that's one thing; if he intends not to kill, but performs some of the actions requisite for killing (e.g., pointing the gun, pulling the trigger), then he is representing a killing. But that's too simple, I think. Certainly if he attempts to kill, but does not kill, then he has failed; but attempting does not entail intending any more than intending entails attempting. Killings manqu~ (so to speak) cannot be divided into failures and representations, for these concepts belong to different contrasting pairs. Representation can be intentional or nonintentional, and a failure may or may not be a representation. There are some nice questions about representation, and there is too much dispute for it to be resolved briefly. But consider visual representation, or depiction: and let us agree with Nelson Goodman that this is a kind of reference (though not that it is denotation). In recognizing the picture of a barn as a barn, we may be, as he says, reading the picture in terms of some set of conventions that he calls a "system of representation.'" Since this concept has not really been explicated, I am free to interpret it in my own way: any depiction works by selective similarity, as I said, but different systems of depiction use different types of features of things to single out things of a certain k i n d - w e can represent a hockey player by a few lines that sketch a characteristic momentary posture of hockey players or by a mass of colored shapes that echo patterns of distribution of bodies in a hockey game. Learning a way to draw is no doubt acquiring proficiency in using such a system. Now I don't suppose that such pictorial conventions could get established if depicting could not be done intentionally; and perhaps to sustain the convention it is even necessary that depicting is typically intentional. But once the system of representation exists, unintentional representation can occur. An unskilled draughtsman who starts out by trying to draw a dog might end up by depicting a cat, and one who is no draftsman at all but simply a compulsive doodler in department meetings may discover that he has unwittingly drawn a monster, with gaping jaws and fearsome teeth, who vaguely resembles the chairman. In short, we can talk not only about the painter representing, but also about the painting representing-perhaps independently of the painter's will. This process I call the detachment of reference-that is, the process by which a mode of referring is devised and used intention-

FICTION

AS REPRESENTATION

297

ally by human referrers, but once the convention becomes understood and accepted in a community, the locus of reference is shifted from the person to the object used to refer, and unintentional reference become possible. So it is with the representation of actions in general; once it is understood that synecdoche can be a form of reference, in which the part is intentionally offered for the whole, actions of certain kinds can be represented by actions of other kinds: the mime's holding his hand up in a certain way represents someone's carrying a tray. Given the rule, it is no longer necessary that action-representation be intentional, even if in fact it often or even generally is. And it seems to me that a person who tries to perform an action of a certain kind, but does not succeed, may come close enough to succeeding so that from what he does one can recognize what he was trying to do. Then the action he performed represents an action of the sort he tried to perform. In other cases, of course, as when the complete novice first tries a complex dive from a springboard or a new disco dance or the tying of a difficult knot, the failure may not capture a sufficiently distinctive feature of the kind of act intended so that it can be said to represent. I apply these general principles also to the representation of illocutionary actions. When a text is produced that could be used to perform an illocutionary action of some recognizable kind, and enough of the requisite conditions are present or presupposed to permit the kind to be recognized- but not enough for such an action actually to o c c u r - t h e n we have representation, a verbal depiction. Normally, of course, that is what the author will have intended. But it's possible-it violates no conceptual t r u t h - t h a t the author was trying to perform an illocutionary action, and only succeeded in representing one. And he may fail to perform the action, and fail so badly he does not even represent o n e - for example, if he can't get the right words out or doesn't provide a sufficient context then there is a kind of unintelligibility. The represented illocutionary action does not, of course, have to be of a sort that could be performed: hence one can represent the action of expressing admiration to a star, though one couldn't actually do so. (Keats's sonnet belongs to a common type of lyric in this respect: I think of Gilbert and Sullivan's Major General Stanley singing, "Pale Moon! To thee I sing. Bright regent of the h e a v e n s " - w h i c h purports

298

MONROE

C. B E A R D S L E Y

to ask a question of an addressee that is manifestly incapable of providing the requested information.) Similarly, one can draw a picture of a Spiderman, or Wonderwoman, though no such thing could exist, or even of M. C. Escher's physically impossible objects. And we can see how the presence of non-referring names and descriptions also fits in,to the scheme. If the speaker in the poem carries on about his recently departed sweetheart, whom he calls "Lenore," and in fact this name (in context) is the name of no real person, then the poet, in producing this text, cannot himself be performing the itlocutionary action of complaining about his loss (since Lenore couldn't really be lost), but is representing an action of complaining. ii I hope it is evident that illocutionary actions can be represented in texts, though there may be legitimate doubts about the adequacy of my account of such representation. The next step is to apply the concept of illocutionary-act representation to those literary texts we acknowledge to b e - o r might be persuaded to accept as-fictional ones. The two kinds I am specially concerned with are fictional prose narratives and lyric poems. A number of linguists, notably Kate Hamburger, 2 have studied linguistic features that seem to be typical of prose fiction, in search of objective marks of fictionality. Linguistic features that signify, because they evidently presuppose, an omniscient narrator are arguably such marks of fiction. For if one of the requisite conditions for performing a particular illocutionary action (say, that of reporting that a certain event occurred in the world of the work) is that the speaker be omniscient, then the speaker cannot be identified with the flesh-and-blood author (who is merely human, and therefore not omniscient). Hence the author must not be performing, but only representing, the action of recording the adventures of Anna Karenina or Holden Caulfield. It has been suggested that the use of indirect discourse about the inner states and processes of characters in the novel implies such a claim to omniscience. Thus here is Count Vronsky sitting in his carriage: Yes, it was the same thing over and over again. Every one, his mother, his brother, every one thought fit to interfere in the affairs of his heart. This interference aroused in him a feeling of angry h a t r e d - a feeling he had rarely known b e f o r e ?

FICTION

AS REPRESENTATION

299

The first sentence has enough irony in it to detach the narrator himself from Vronsky: the narrator does not necessarily sympathize with Vronsky's petulant resentment. The second sentence is an account of an occurrent emotion and of a general disposition or trait of Vronsky. But it doesn't seem that an omniscient narrator is required here; there is only the implicit claim that the narrator could have known about these feelings, was in a position to learn of them. Since feelings can be inferred from behavior and from verbal expressions, we would not necessarily consider it out of place to find quite similar feelings ascribed to historical figures in a nonfictional narrative: Nixon, according to his own published recollections, became very suspicious of his aides after John Dean defected; he preserved his tapes because he feared betrayal and hoped they would afford some protection against it.

A more special and interesting linguistic feature, much discussed in recent years, 4 has been claimed as a distinguishing mark of fictionality in narrative: it is the so-called "free indirect style," or "narrated monologue." Here is Vronsky again: "What a pity!" thought Vronsky, putting up the roof of the carriage. "It was muddy before, now it will be a perfect swamp. ''5

The word "now" here accomplishes a drastic switch from the narrator's point of view to that of Vronsky himself: for the narrator it would be "then," in accordance with the prevailing past tense; "now" could only be thought by Vronsky himself, referring to the time at which he thought it. The combination of a temporal adverb referring to present (or future) time with a verb in the past tense marks the "epic preterite," and is, according to K/ite Hamburger, an infallible indication of fictionality, 6 " N o w " and "will be" take the pastness out of "thought Vronsky" and substitute a "fictional present." Hence, it is said, the free indirect style would not be appropriate for historiography-or, as I would put it, the free indirect style would prevent the text from being historiography and transform it into the depiction of a historiographic iUocutionary action. There are many interesting controversies about the proper linguistic analysis of the free indirect style and about its significance, especially for the question whether all narratives have an implicit n a r r a t o r - a matter of controversy. It does seem to me, however, that we do not have here a certain sign of fictionality. First, the word "now" can be used to mark a shift of point of view back in time, even if it is still the narrator's point of view. That late lamented journal, The Civil Liber-

300

MONROE

C. B E A R D S L E Y

ties Review, published a long account by David Hamlin, Executive Director of the Illinois Branch of the American Civil Liberties Union, of the ACLU's efforts to defend the First Amendment against the extraordinary, unconstitutional restrictions imposed by the town of Skokie, Illinois, to prevent a demonstration by American Nazis. At one point the narrative runs: . . . The court agreed with our position t h a t . . . [etc.] An appeal was taken from the ruling. Now, the battle was being waged on four f r o n t s : . . 7

followed by a summary of the various legal actions. The expression "Now, the battle was being waged" is the epic preterite; here the narrator identifies his interests with that of the Illinois ACLU, and it is as though he were reporting the way things appeared to the ACLU at that stage. Clearly no omniscient narrator is called for, but only one close to what was going on. And similarly, I want to suggest, the report that Vronsky thought the road would soon become a swamp does not assume any conditions that could not, even in principle, be fulfilled by an actual author. Even if it should turn out (and this is still an open question) that there are no particular linguistic features that are decisive marks of fictionality, there may be many that serve as cues for the reader. Moreover, our brief discussion suggests the possibility of looking further afield for other marks that may be similar in having an epistemic aspect. Certain illocutionary actions require a particular epistemic relationship between the speaker and, say, the events he is describing: for example, one can (strictly speaking) report only events one has actually witnessed or heard about from a witness. I can describe the battle of Borodino, but I can't report it. Tolstoy's description of that battle cannot be a report of it, either, though of course he might pass on someone else's report. This line of inquiry needs to be pursued further than it has been, but I'm somewhat doubtful about its promise. The sentences in fictional narrative mainly purport to be descriptions of events (including human actions); I can't see that any very clear of definite epistemic conditions are among those requisite for such descriptions, and so I can't see how such conditions are typically violated in novels. The main feature of novels that makes them fictional must be the presence of nonreferring names and descriptions. When Tolstoy

FICTION

AS REPRESENTATION

301

writes that Vronsky put up the roof of the carriage, h e is producing a text that could be used in describing an event. But since the name "Vronsky" does not refer to any actual person, no illocutionary act of describing occurs; it is only represented-as the fictional act of a fictional narrator. In a broader perspective (but here we get into other problems) we can say that the text implicitly contains nominalizations of all its event-describing sentences: for example, "Vronsky's putting up the carriage-roof." Another character might later on allude to this action of Vronsky's; the action is one of the entities in the fictional world he inhabits. But no such action really occurred, so the participal phrase, "Vronsky's putting up the carriage-roof," is also a nonreferring expression. And the novel is replete- at least implicitly- with them. Indeed, if we include nominalized event-descriptions in the implicit corpus of a novel, all novels contain at least some non-referring names or descriptions, since all describe at least some events that never occurred. III

Even those who will cheerfully assent to my conclusion that novels and short stories a r e - at least generally speaking-representations of (compound) illocutionary actions, may be more uneasy when it comes to lyric poems. But we use the same unsatisfactory rubric to cover b o t h - t h e y are "imaginative literature" - and nonreferring expressions often occur in both. Yet it may seem that somehow there is a difference: the Anna/Vronsky affair was "made up" by Tolstoy (even if he got the idea from a newspaper report or a dinner-table conversation), but lyric poems are more like reports of events that actually occurred. These are seldom physical events, though they might be: Wordsworth did wander around the waterfalls above Tintern Abbey. Mostly they are psychological events of various sorts. When, one day in 1877, for example, Gerard Manley Hopkins began a sonnet by writing, "The World is charged with the grandeur of God," he was displaying a religious emotion he actually felt, wasn't he? And so in writing those words, why was he not performing the illocutionary action of expressing that emotion or attitude? Literary theorists have found it convenient to introduce a construct called the "speaker" or "persona" of the poem, as a way of collecting all that the sentences of the text purport about their ostensible

302

MONROE

C, B E A R D S L E Y

producer: for example, that he (or she?) is a deeply devotional person. It is the illocutionary actions of this fictional speaker that are, on the view I am defending, represented by the words of Hopkins's poem. Even if it is true, and we know it to be true, that the sentiments of this speaker and the sentiments of Hopkins coincide in certain respects - which is what we learn from studying the extant documents of Hopkins's life and comparing them with his p o e m - this will not, by itself, permit us to identify the speaker with the author. They remain two, not one-comparable, and perhaps in some respects contrastable, but distinct. I want to argue that to regard them as the same person would lead to absurd and unacceptable consequences. If Gerard Manley Hopkins and the speaker of his sonnet are one and the same person, then what is true (extensionally) of the former must be true of the latter. The speaker of the poem, it turns out, is 33 years old (since Hopkins was born in 1844), and on the verge of ordination (since Hopkins was ordained later in 1877). Surprisingly enough, the speaker of the poem is someone who has never been to Afghanistan or been allergic to strawberries. But it seems to me a serious mistake to ascribe such properties, and indefinitely many others, to the speaker of the poem; therefore, the speaker is not Hopkins. But then we can draw a further conclusion. If Hopkins were performing an illocutionary action in writing the words of this poem, he would be the speaker of the poem. For he can perform only his own actions. Thus if he is not the speaker of the poem, he must be representing the illocutionary actions performed by the speaker. He can state his religious sentiments in a diary or in a letter to a friend or to the editor, but not in a lyric poem. You may wish to take exception to my list of allegedly illegitimate inferences from author to speaker: can't we distinguish between the permissible and the impermissible ones? Many a literary historian with biographical interests would want to say that it is of no consequence or relevance whether Hopkins himself travelled abroad or walked with a limp or liked Stilton cheese, but it may be of great interest and relevance that he was a Roman Catholic and about to become a priest. My reply is somewhat inflexible, I suppose. First, I don't see any nonquestionbegging way of distinguishing between the relevant and the irrelevant facts about Hopkins himself. For, second, the question is simply whether we identify author with speaker: if we

FICTION

AS REPRESENTATION

303

do, then everything (extensionally) true of either necessarily becomes true of the other; if we do not, then whatever we know of either is not necessarily true of the other. We must be careful here: we can't rule out inductive or probabilistic inferences, in some cases. Thus if we know a lot about Hopkins in 1877 we may infer that if he wrote a poem then the illocutionary actions he would represent would probably be those of a religious person. But the inference is finally tested by appeal to the text, for it is the text that reveals what sorts of iUocutionary actions he actually did represent. The Hopkins poem is a lyric, but it verges on the didactic, at least in beginning with a firm metaphysical declaration. I use it partly to counter a view that could be advanced to narrow the scope of what I have proposed. Perhaps I am right about lyric poems, but wrong about the other major species, the didactic: with the latter, it may seem, genuine illocutionary actions are performed in the writing. Thus, for example, John Reichert has ably criticized a version of my position that I defended some time ago, 8 and his example is that fine little poem by Robert Frost that begins: Nature's first green is gold, Her hardest hue to hold . . . . and ends So leaf goes down to leaf, So Eden came to grief. So dawn goes down to day. Nothing gold can stay. And he argues that the last line "Nothing gold can stay" is a genuine assertion. I do not deny the distinction between didactic and lyric poetry, but I treat didactic poetry as a species of lyric. I mean that what distinguishes it from the rest is not at all that it involves performances rather than mere representations of illocutionary actions, but that it represents certain kinds of illocutionary action: affirmations and denials of general principles, declarations and pronouncements, arguments and refutations (as in Pope's Essay on Man). We are still in the realm of fiction, but it is philosophy fiction (taking this term broadly and somewhat charitably). It is still basically concerned with human attitudes, but articulates them through a play of ideas.

304

MONROE

C. B E A R D S L E Y

"But what reason is there," Reichert asks, 9 "for calling Frost's poem fictional, or for thinking that he did not mean precisely what he wrote?" I give these reasons: One: we can apply to this poem the same sort of argument I applied to the Hopkins poem. It is clear that we should not identify the speaker with the author, because then we would have to say that, for example, the speaker in this poem milked cows and appeared at the Kennedy inaugural (which is absurd); and if the speaker is not the author then his illocutionary actions are not the author's. Two: if the line "Nothing gold can stay" were actually affirmed here (contrary to Sir Philip Sidney), we should have to take it seriously and raise embarrassing questions. "Gold" in this poem, I suppose, must be short for "gold-colored." Those of us who have taken a course in elementary logic or Thinking Straight know well that we must mistrust universal negative statements, and when we are confronted with one we should immediately cast about a possible counterexample to test it. In this case that is easy: one thing that is gold-colored is gold itself (the metal). And in fact gold has been around the earth's surface for billions of years. According to the 1969 edition of the Encyclopedia Britannica, "Gold is extremely inert and is considered to be virtually indestructible." This is silly, you will reply, and quite rightly. By which reply, I argue, you acknowledge that in writing his last line Frost was not performing an illocutionary action. Third: there is the artifice in a r t - i n poems: meter, rhyme, mannered or formal features, which suggest a strong concern for aesthetic effect, independently of illocutionary success. This artifice or artificiality does not absolutely preclude performance, of course, but adds a possibly competing m o t i v e - since there is always the lurking likelihood of conflict between the substance of some chosen illocutionary action and other illocutionary actions added to make a complex saying that is interesting independently of its truth or other appropriate felicity. So the presence of formal features is a criterion of fictionality. Fourth: to my mind, the prime mark of fictionality in poetry is what might be called "address without access." I mean that the ostensible speaker evidently wishes to communicate with someone directly and immediately- there is an ostensible someone whom he is addressing-but the author has cut his text loose from any particular

FICTION

AS REPRESENTATION

305

occasion that might afford access to such an addressee. We hear a personal voice, but we know that it cannot be speaking to us, since we have not received the text from the author, but found it in a library or book-store or magazine. Of course genuine illocutionary actions are published a b r o a d - a s history, news, polemic, e t c . - b u t then in a more formal style that preserves their illocutionary force: they are not withdrawn from occasions and circumstances, but addressed to a particular audience. As soon as a consciousness of the speaker as an individual speaking to some other individual creeps into the text, its character changes. We see this in the newspaper writings of such humorists as Art Buchwald, Russell Baker, Erma Bombeck: even if what they tell happens to be true, it becomes fictional in the telling. There is a claim to intimacy with the reader that would have to be fulfilled for proper uptake to be secured and illocutionary actions to be generated- but the author has not fulfilled it. IV

Although, as I remarked at the start, I am by no means alone in holding this view of fiction, and of imaginative literature as fiction, 1° I can't claim to be among the majority of literary theorists, and you may well feel that I have pushed the theory and its implications beyond most of my colleagues, if not beyond the bounds of reason. So you will understand if, in my exposed position, I cast about for corroborative evidence and argument in the work of other recent writers. By way of conclusion I will adduce a few of these corroborations, and also take advantage of them to note further implications of my v i e w - w h i c h will, I hope, help to point up the significance of the issue over the nature of fictional texts. Most encouragingly, I note the substantial agreement on many points between Barbara Herrnstein Smith and myself, though she has developed her position a good deal more fully in some respects. In her recent book, On the Margins of Discourse, H she argues that all poetry is "representation," in contrast to what she calls "natural discourse" or "natural utterances" or "may be so regarded. ''~2 She too emphasizes the affinity of poetry, as "mimetic," with pictorial and dramatic representation, la and has many highly original and illuminating things to say about the ways in which the fictional character of a text affects our interaction with it. ~4

306

MONROE

C. B E A R D S L E Y

I have been reading a recent essay by Kendall Walton; ~5 among other good things, it contains a discussion of the problem of how it is that literary works can survive repeated readings, that we can even feel suspense when we know how things in a novel will turn out. I am not concerned here with his explanation, but his discussion suggested to me another argument- perhaps not a very strong one - f o r the view that poems as well as novels are not records of illocutions but rather representations. If Frost's poem, or Hopkins's, were not fictions, but the vehicles of assertions, we would at least find it much harder to take repeated interest in t h e m - to go back to them again and again. Or at least our interest would tend to be very different. Once we got the message, that would be it; there would be nothing further to learn, and re-reading would be superfluous. It is the recognition that we have to do with a representation that enables us, or at least helps us, to enjoy its content for its own sake, to enter into its world with our own make-believe, to make-believe over and over again that we are in the presence of someone who believes that the world is charged with the grandeur of God or that precious things cannot be held on to. Recognizing a fiction, we are not encouraged to note a definite bounded message and run, for there may be more in the picture than even the author realized, and there is sure to be more than would be needed to convey a message. Thus the poem takes on a thickness and complexity that makes it serve as an object of aesthetic contemplation. Something like this, I think, has been meant by those who s a y very misleadingly - that poems ought to give us "what oft was thought but ne'er so well expressed," or that it is not what is said, but how it is said, that counts. I strongly repudiate these formulas-even if one carries the authority of Alexander Pope. Expressing in a poem can only be done by representing illocutionary actions, and that's just the thought that gets into the p o e m - s o different expression is always different thought. The how is part of the what: to say something in a different way, or manner, is simply to say that thing and also say something else. ~6 Example: to say you play the piano is one thing; to say it modestly is just to say at the same time (implicitly or by suggestion) that you do not regard this accomplishment with vast pride. So an interest in how things are said in poems is precisely an interest in all that is said (all the illocutionary actions represented), however subtly or softly or delicately.

FICTION

AS R E P R E S E N T A T I O N

307

One of the most important implications of the fictional view of poetry concerns the theory of interpretation it logically leads to. 17 Attempts have been made, notably by E. D. Hirsch, to formulate a very general account of interpretation, one that will encompass the strikingly varied activities that go on under this heading when we are talking about, say, the Bill of Rights, Lycidas, a dream, the invasion of Afghanistan, tea leaves, or a message in code. Surely it is of some interest and importance to discover how far this systematic inquiry can be carried. But even within the limited realm of texts, it is also essential, I think, to distinguish between two different, and in some ways even opposed, activities that are called "interpretation." An ordinary t e x t - and Barbara Smith has some very good passages making this clear-belongs to a set of circumstances, or an ecompassing situational context. If we are in doubt about what it s a y s that is, about exactly what iliocutionary actions were being performed in the act of composing i t - o u r task is to reconstruct the situational context and examine the act-generating conditions that were present. Though we can always in principle distinguish between the question: Which illocutionary actions did the author perform? and the question: Which illocutionary actions did the author intend to perform, we are often mainly interested in the latter. So if the question is raised what the drafters of the Bill of Rights meant by "the right of the people peaceably to assemble and to petition the Government for a redress of grievances," and whether this applies to American Nazis, then we will properly investigate records of the original debates, political circumstances, contemporary writings by political philosophers, etc. And if we take a poem as a text that is to be interpreted in this reconstructionist fashion, we will study the letters of the author and the memoirs of his friends. If, on the other hand, the poem is a representation of an illocutionary action (that is, of a series of them), there is no question of reconstructing an actual action. We have a text containing sentences that could be used, would naturally be used, to perform illocutionary actions of certain kinds, but with semantic and syntactical deviations of various interesting sorts. Our task is rather to construct appropriate actions by filling in the gaps of the text in a reasonable way. To borrow Barbara Smith's example, we read "To me, fair friend, you never can be o l d . . . " : we have to take these words, along with those that follow, giving the closest attention to every word and every

308

M O N R O E C. B E A R D S L E Y

grammatical feature, and imagine in detail a situation in which such words would be used to perform an illocutionary action, and what, in that situation, they would be saying about the speaker and the addressee and their relationship. This is not a job for mere impressionists, of course; it is a genuine task, with its own regulations. Not all such interpretations will be correct; and all are, to some extent, testable by the words of the poem. But biographical facts will not help us in this kind of interpretation, since we are, as I said, not trying to reconstruct an illocutionary action that really took place, but rather to construct an implied or projected one that is represented by its verbal component. Thus, in his masterly commentary of Shakespeare's Sonnets, TM Stephen Booth is concerned to understand how the relevant meanings of Shakespeare's words and phrases and the contexts they bring with them combine, intertwine, fuse, and conflict in the potentially dizzying complexity from which a reader's sense of straightforward simplicity emerges.

As C. L. Barber noted in a long review, Booth excludes consideration of the relationship between the sonnets and Shakespeare as a man writing them out of human situations, a man using poetry to relate himself to a well-born young man and a promiscuous mistress. He is concerned only with Shakespeare as a fabulous artificer. The focus gives his commentary its unprecedented v i r t u e s . . . And it permits him to turn his back on the quagmire of biographical speculations. 19

As Barber says, it is the "action of the poetry" that matters in the reader's experience of it, and that calls for the critic's attentions; yet he feels that something is missing: But the action of the poetry can be properly understood, I think, only by recognizing that it is working to transform or cope with "situations and relationships which," as J. B. Leishman puts it, "cannot have been invented, if for no other reason than that they have been left so tantalizingly obscure. ''~° [Booth's alleged failure to explain adequately the sonnets' coherence] results partly from Booth's refusal to consider the poems' relationship to the situations implicit in them.

But in so far as the situations are actually "implicit," they must be part of the poem's a c t i o n - w h a t the fictional speaker is concerned about. On the other hand, how writing the ~onnets helped Shakespeare to "cope" with his real-life problems is something we really can't discover and have no right to drag into the poem even if we could. It may sound strange and even presumptuous to say that such

FICTION

AS REPRESENTATION

309

inveterate biographical speculators as A. L. Rowse simply fail to understand the basic nature of lyric poetry and the proper way to read it. But this conclusion seems to follow from the position we must take in order to avoid even greater absurdities. Probably it will not go without saying, and therefore I should say explicitly, that of course 'I do not deny that poems can be used to perform illocutionary actions, to make a religious or philosophical or political statement, to protest angrily or to exhort to action. Some time ago I read of Gaston Miron, called the "poetic voice of Quebec nationalism," in the N.Y. Times. z~ Certainly to his fellow Qu~becois his poems are often more than representations. One of them, "October," was written after violent demonstrations, or nonpeaceful assemblies, in 1970, which were (naturally) suppressed by armed force and with mass arrests. The final stanza: We shall make of you, land of Quebec, A bed of resurrections And of a thousand flashing metamorphoses Of our leavenings from which the future will arise And of our uncompromising wills Men will hear your pulse beating in history It is rippling through the October autumn It is the russet noise of deer in the sunlight With the future open, the future committed. Now I still want to maintain that the act of writing these words was an act of representation; this is evident from the fact that the act of making promises to an abstract entity like the "land of Quebec" is one that is in fact impossible to perform, since no uptake can be obtained. But of course if Gaston Miron reads his poem from a platform during a rally in which thousands of people are petitioning the Canadian government for a redress of grievances, and shouting their resolution to secede from the other nine provinces, then the act of producing a token of that text under those circumstances will be political, and will generate genuine illocutionary actions. But that, I would say, is a use of the t e x t - a perfectly legitimate o n e - and does not affect the nature of the text itself, as originally produced: whatever else may be done with it, Miron's lines are (part of) a poem, and thus a fiction. They are serious fiction, true; but we have known since Aristotle's Poetics that poems are serious, even when they present a

310

M O N R O E C. B E A R D S L E Y

mimesis of the impossible - indeed, that they can be more serious than much nonfiction. The case of Gaston Miron reminds me of an even more famous contemporary poet, who is interestingly discussed by Mary Frances HopKins in a recent essay arguing for sincerity as an important critical consideration in dealing with "confessional poetry. ''22 She describes the characteristic qualities of Andrei Voznesensky's public readings of his poems, the ways in which he manifests his "total commitment to the poem." But the most striking evidence of commitment occurred at Town Hall during Vosnesensky's reading of his poem 'Winter at the Track.' Halfway through the poem, Vosnesensky apparently forgot what came next. He stopped completely, dropped his arms, and bowed his head in a gesture of apology and finality. William Jay Smith, who shared the stage with him, rushed over with the text, as did others. Vosnesensky declined. The mood was gone, and he would not force it. That audience never heard the end of the Russian poem. 23

Now I submit that this is a most revealing episode, involving as it does a poet famous for his participation in anti-war, anti-capitalist, anti-imperialist, and other occasions. If he had been interrupted in the performance of an illocutionary action, one would surely expect him to resume and complete it as soon as feasible. Only if we take him, even in his public oral reading, to be representing an illocutionary action for an aesthetic purpose can we understand why he broke it off once "the mood was gone." I don't assume, of course, that aesthetic purposes and revolutionary (or political) purposes cannot ever be reconciled in the writing or the oral reading of a poem. But there is a deep tension between these purposes, I think; and it is to be expected from what we know of each and of the conditions for achieving each. We have seen that recognizing a poem as the representation of an illocutionary action goes along with refusing to identify the poem's speaker with the poem's author. This distinction preserves the "ontological incompleteness" of the speaker-that is, it is not the case that one member of each pair of contradictory statements is true of the speaker. This truth-value gappiness is conspicuously characteristic of fictional entities and of fictional worlds, in contrast to actual worlds. 24It is a necessary consequence of being an imagined world, for no world could be imagined in infinite detail, even by Jose Luis Borges. It is part of what makes fiction appeal to our own imaginations, as readers, since it

F I C T I O N AS R E P R E S E N T A T I O N

311

leaves possibilities o p e n - indeterminacies - f o r us to fill in as we like. Finally, it helps fiction to produce a degree of that emotional detachment that is a mark of the aesthetic response, however moved we may also be. It is no derogation of poems, or of any works of art, to find in them this kind of purposeful withdrawal from full engagement in the affairs of real life; nor does this imply a denial of the fact that it is real life to which we return from art and to which poems and other artworks are bound in the end to make their splendid contribution.

Temple University

NOTES Some excellent recent discussions of these ontological problems are Nicholas Wolterstorff, 'Worlds of works of art,' Journal oy Aesthetics and Art Criticism 35 (1976), 121-32; Peter van Inwagen, 'Creatures of fiction,' American Philosophical Quarterly 14 (1977), 299-308; David Lewis, 'Truth in fiction,' American Philosophical Quarterly 15 (1978), 37--46; Kendall Walton, 'How remote are fictional worlds from the real world?' Journal of Aesthetics and Art Criticism 37 (1978), 11-23; Robert Howell, 'Fictional objects: How they are and how they aren't,' Poetics 8 (1979), 129-77. The Logic of Literature, trans. Marilynn J. Rose, 2d ed. rev., Bloomington: Indiana University Press, 1973, ch. 3. 3 Anna Karenina, Part II, ch. 21; trans. Constance Garnett, N.Y.: Modern Library [n.d.], p. 217. 4 See Hamburger, above; Ann Banfield, 'Narrative style and the grammar of direct and indirect speech,' Foundations oy Language 10 (1973), 1-39; G. L. Dillon and F. Kirchhoff, 'On the form and function of free indirect style,' PTL: Poetics and Theory of Literature 1 (1976), 431-40; S. Y. Kuroda, 'Reflections on the foundations of narrative theory-from a linguistic point of view,' in Teun Van Dijk (ed.), Pragmatics of Language and Literature, Amsterdam: North-Holland, 1976; Banfield, 'The formal coherence of represented speech and thought,' PTL 3 (1978), 289-314; Dorrit Cohn, Transparent Minds: Narrative Modes for Presenting Consciousness in Fiction, Princeton: Princeton University Press, 1978, ch. 3; Seymour Chatman, 'The structure of narrative transmission,' in Roger Fowler (ed.), Style and Structure in Literature: Essays in the New Linguistics, Ithaca: Cornell University Press, 1975; Janet Holmgren McKay, 'Some problems in the analysis of point of view in reported discourse,' Centrum 6 (1978), 5-16; Joel Rudinow, 'Representation, voyeurism, and the vacant point of view,' Philosophy and Literature 3 (1979), 173-86. 5 Anna Karenina, Ioc. cit. 6 Hamburger, pp. 71, 82, 95.

312

M O N R O E C. B E A R D S L E Y

7 David M. Hamlin, 'Swastikas and survivors: Inside the Skokie-Nazi free speech case,' The Civil Liberties Review 4, March-April 1978, p. 26. 8 In The Possibility of Criticism, Detroit: Wayne State University Press, 1970. 9 John Reichert, Making Sense of Literature, Chicago: University of Chicago Press, 1977, p. 129. 10 See references in 'Aesthetic intentions and fictive illocutions,' in Paul Hernadi (ed.), What is Literature? Bloomington: Indiana University Press, 1978. 11 Barbara Herrnstein Smith, On the Margins of Discourse: The Relation of Literature to Language, Chicago: University of Chicago Press, 1978. Marcia Eaton's analysis of pretending comes fairly close to my concept of (verbal) depicting (see 'Liars, ranters, and dramatic speakers,' in Benjamin R. Tilghman (ed.), Language and Aesthetics, University of Kansas, 1973, p. 55). But I think she is wrong in saying that acting on stage mostly consists in actually performing the pretended action, though not being the character pretended. She is acutely right that (in my language) the actor represents not a man walking (which he is) but a French Count walking (which he is not). But it should also be noted that the actor represents a man walking across an eighteenth-century drawing room floor (which he is not doing, since there is only the stage set). 12 Smith, p. 24. 13 Ibid., p. 25. 14 I can perhaps claim some support for my theory of fiction from the considerations advanced by Samuel R. Levin on behalf of his own interesting proposal, in 'What kind of speech act a poem is,' in Teun A. van Dijk (eds.), Pragmatics of Language and Literature, Amsterdam: North-Holland, 1976. According to him every poem contains (in its deep structure) an implicit "higher sentence" that introduces it and "expresses the kind of illocutionary force that the poem is to be taken as having" (pp. 149-50). It runs: "I imagine myself in and invite you to conceive a world in which (I say to you): The unpurged images of day recede . . . " or whatever. To say that the text of a poem is a representation of illocutionary actions (and thus has no iUocutionary force) seems to me (1) a simpler account than Levin's, that (2) explains everything about poems that his account explains and yet (3) avoids the awkward questions that can be raised about the idea that Yeats does not just imagine his world (and create it in purported descriptions) but also imagines himself somehow in that world. 15 'Fearing fictions,' The Journal of Philosophy 75 (January 1978), 5-27. Cf. 'How remote are fictional worlds from the real world?' Journal of Aesthetics and Art Criticism 37 (Fall 1978), 11-23. 16 See 'Verbal style and illocutionary action,' in Berel Lang (ed.), The Concept of Style, Philadelphia: The University of Pennsylvania Press, 1979. 17 Barbara Herrnstein Smith has some valuable discussion of this point; see her chs. 2 and 6. See also my essay 'Intentions and interpretations: A fallacy revived,' forthcoming in The Aesthetic Point of View. 18 New Haven: Yale University Press, 1978. 19 New York Review, April 6, 1978, p. 32. 20 Themes and Variations in Shakespeare's Sonnets, Atlantic Highlands: Humanities Press, 1961, p. 11. 21 April 2, 1978.

FICTION

AS R E P R E S E N T A T I O N

313

22 'Sincerity and the performing artist: An old critical concept reestablished,' in Esther M. Doyle and Virginia Hastings Floyd (eds.), Studies in Interpretation, Vol. II, Amsterdam: Rodopi, 1977. z3 Ibid., p. 218. 24 Cf. Nicholas Wolterstorff, 'Worlds of works of art,' pp. 131-32. See Note 1.