Truth in Fiction: A Reexamination of Audiences STOR Peter J. Rabinowitz Critical Inquiry, Vol. 4, No.1 (Autumn, 1977), 121-141. Stable URL: http://links.jstor.org/sici?sici=0093-1896%28197723%294%3A1%3C121%3ATIFARO%3E2.0.CO%3B2-U Critical Inquiry is currently published by The University of Chicago Press. Your use of the JSTOR archive indicates your acceptance of JSTOR' s Terms and Conditions of Use, available at http://www.jstor.org/aboutiterms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www .j stor .org/journals/ucpress .html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support@jstor.org. http://www .j stor.org/ Tue Dec 20 20:54:49 2005 ® Truth in Fiction: A Reexamination of Audiences Peter J. Rabinowitz 1 Some literary arguments, such as those about the fourth book of Gulliver's Travels, clarify and enrich the works around which they center. Others-the debate between the New Critics and the Chicago School or the battle over la nouvelle critique in France-are valuable because they reveal differences in critical ideologies and suggest alternative ways of looking at texts. But in many ways, the most fascinating are those controversies which force a radical reexamination of our critical vocabulary. These quarrels often deal ostensibly with a specific text, but they unintentionally and unconsciously reveal something more general: a fundamental inadequacy in the way we talk about literature. Not only can't we resolve the differences between the critics, but lacking the terminology to explain the disagreement fully, we are left with the sense that we don't even understand just what it is about or why it has occurred. If pursued, such quarrels can transcend their initial subjects and lead us to substantial critical insights. One such controversy surrounds Vladimir Nabokov'sPale Fire. The novel itself is in the form of a poem by the recently deceased professorpoet John Shade, with an introduction, commentary, and index by Shade's nextdoor neighbor and colleague, Charles Kinbote. Kinbote's elaborate exegesis relates the poem to historical events in a country called Zembla, events which climax in the escape of King Charles at the outbreak of the Zemblan Revolution. The reader soon learns that Kinbote believes himself to be King Charles and that his colleagues consider him to be either a lunatic or a fool. 121 122 Peter J. Rabinowitz. Truth in Fiction But perhaps I shouldn't even describe the novel that way, for as soon as it was published, critics disagreed not only about what it meant and how good it was but even about what was "really happening" in the novel. Mary McCarthy argued that Kinbote was really a member of the Russian department named Botkin;1 Andrew Field claimed that Kinbote had in fact been invented by Shade;2 and Page Stegner suggested that perhaps, to the contrary, Shade had actually been invented by Kinbote. 3 Kevin Pilon wrote a chronology of Pale Fire as if all events-including those in Zembla and those in Shade's poem-had really occurred. 4 John Stark, on the other hand, insisted that actually only "Nabokov and Pale Fire (in a sense) are real; any layer inside them (actually in the novel) is imagined, and none of those inside layers is more real than any other," although, curiously, he also criticized Shade for the realism of his poetry ("it recreates the past instead of creating and exhausting new possibilities") and praised Kinbote for a commentary that is "purely imaginary."5 Who is right? The answer to this question, of course, must come prior to any satisfactory discussion of the overall meaning of the novel: until we know what is going on, we can hardly interpret it, much less evaluate it. But although the flame of controversy has died down as the novel has aged, the question has not been answered. I think that John Stark's contradictory arguments help explain why. Obviously, he is right when he says that only Nabokov and his novel "really" exist-but he is just as obviously wrong. The word "real" has several different meanings here, and we simply do not have, at the present time, a sufficiently precise vocabulary for distinguishing them. It seems, then, that in order to address the question, "What is really happening in Pale Fire?" we must first ask ourselves, "How do we even begin to talk about truth in fiction?" Questions about the status of literary truth are as old as literary 1. Mary McCarthy, "A Bolt from the Blue," The New Republic 146, no. 23 (4 June 1962): 21. 2. Andrew Field, Nabokov: His Life in Art (New York and Toronto, 1967), pp. 316-18. 3. Page Stegner, Escape into Aesthetics: The Art of Vladimir Nabokov (New York, 1966), p. 129. 4. Kevin Pilon, "A Chronology of Pale Fire," Russian Literature Triquarterly, no. 3 (1972): 370-78. 5. John Stark, "Borges' Tlong, Uqbar, Orbis Tertius' and Nabokov'sPale Fire: Literature of Exhaustion," Texas Studies in Literature and Language 14, no. 1 (Spring 1972): 142-43. Peter J. Rabinowitz, assistant professor at Kirkland College, is currently working on articles on Raymond Chandler, Faulkner, and Dostoyevsky and is, as well, a music critic for the Syracuse Guide. He wrote his doctoral dissertation in comparative literature on the philosophical implications of Nabokov's use of humor and terror. "Truth in Fiction" is the first article he has had published in a scholarly journal. Critical Inquiry Autumn 1977 123 criticism, but they have become both more intricate and more compelling as literature has grown progressively more self-conscious and labyrinth ian in its dealings with "reality." One might perhaps read The Iliad or even David Copperfield without raising such issues. But authors like Gide (especially in The Counterfeiters), Nabokov, Borges, and RobbeGrillet seem continually to remind their readers of the complex nature ofliterary truth. How, for instance, are we to deal with a passage like the following from William Demby's novel The Catacombs: When I began this novel, I secretly decided that, though I would exercise a strict selection of the facts to write down, be they "fictional" facts or "true" facts taken from newspapers or directly observed events in my own life, once I had written something down I would neither edit nor censor it (myself).6 What does this sentence mean? When an apparently fictional narrator (who, to make matters more confusing, has the same name as his author and is also writing a novel entitled The Catacombs) distinguishes between "fictional" and "true" facts, what is the status of the word "true"? It clearly does not mean the same as "fictional," for he opposes it to that term. Yet it cannot mean "true" in the sense that historians would use, for he calls what he is writing a novel, and even if he quotes accurately from newspapers, the events of a narrator's life are not "historically" true. This is but a small version of other more famous literary questions. What precisely do we mean when we ask whether the governess in Henry James' The Turn of the Screw is really a trustworthy witness? Or when we ask whether Dostoyevsky's Golyadkin (The Double) really has a double, or what really happened last year at Marienbad? The means of dealing with the problem of truth in the novel have been various, but not, I think, wholly satisfactory. In this essay, therefore, I would like to set forth a model which allows us to talk more clearly about literary truth. The model does not work equally well for all novels, but it works for most, and is particularly useful in clarifying (although not in "solving") the difficulties encountered in texts with involved narrative structures. It gives us a way to talk about The Double and The Turn of the Screw, and, as I will show at the close of my argument, it explains just what is at issue in the Pale Fire dispute. My model is centered less on the novel's text than on the novel's reader. Such reader-oriented criticism has, of course, become increasingly fashionable of late. Critics like Wayne C. Booth, Stanley E. Fish, Norman Holland, Wolfgang Iser, John Preston, and Walter Slatoff, however diverse their views, have all moved away from the 6. William Demby, The Catacombs (New York, 1970), p. 93. 124 Peter J. Rabinowitz Truth in Fiction New-Critical emphasis on the "text itself" and have begun to study in detail the ways in which those texts interact with readers.7 For the purposes of my argument, however, two essays which deal with different levels of audience interaction seem particularly relevant: Walker Gibson's classic, "Authors, Speakers, Readers, and Mock Readers,"8 and Walter J. Ong's more recent development of Gibson's ideas in "The Writer's Audience Is Always a Fiction."9 Both of these critics have recognized that the act of reading demands a certain pretense. Making an analogy to the familiar distinction between real author and "speaker," Gibson notes: I am arguing, then, that there are two readers distinguishable in every literary experience. First, there is the "real" individual upon whose crossed knee rests the open volume, and whose personality is as complex and ultimately inexpressible as any dead poet's. Second, there is the fictitious reader-I shall call him the "mock reader"whose mask and costume the individual takes on in order to experience the language. 1o This is a good first step, but unfortunately neither Gibson nor Father Ong is sufficiently concerned with the distinction between fictional and nonfictional modes of address, or with the related distinction between speaker (or narrator) and implied author. Both critics treat autobiography and novel in much the same way; Gibson, indeed, moves wittily 7. See, in particular, Wayne C. Booth, The Rhetoric of Fiction (Chicago, 1961) and A Rhetoric of Irony (Chicago, 1974); Stanley E. Fish, Self-Consuming Artifacts: The Experience of Seventeenth-Century Literature (Berkeley, Los Angeles, and London, 1972); Norman Holland, The Dynamics of Literary Response (New York, 1968) and 5 Readers Reading (New Haven and London, 1975); Wolfgang !ser, Der implizite Leser (Munich, 1972) appearing in English as The Implied Reader: Patterns of Communication in Prose Fiction from Bunyan to Beckett (Baltimore and London, 1974); John Preston, The Created Self: The Reader's Role in Eighteenth-Century Fiction (London, 1970); and Walter Slatoff, With Respect to Readers (Ithaca and London, 1970). The practitioners of the French nouvelle critique and nouveau roman, although working in a substantially different tradition, have also paid considerable attention to the active role of the reader and especially the critic. See, for instance, Roland Barthes, Critique et verite (Paris, 1966); Serge Doubrovsky, Pourquoi la nouvelle critique: Critique et objectivite (Paris, 1966) appearing in English as The New Criticism in France, trans. Derek Coltman (Chicago, 1973); Alain Robbe-Grillet, Pour un nouveau roman (Paris, 1963) appearing in English as For a New Novel, trans. Richard Howard (New York, 1966), especially the essay, "Temps et description dans Ie recit d'aujord'hui" ("Time and Description in Fiction Today"). The interest in reader-oriented criticism can also be seen in the success of the reader seminars and workshops at the 1975 and 1976 MLA Conventions. 8. Walker Gibson, "Authors, Speakers, Readers, and Mock Readers," College English 11, no. 5 (February 1950): 265-69. 9. Walter J. Ong, S.J., "The Writer's Audience Is Always a Fiction," PMLA 90, no. 1 (January 1975): 9-21. 10. Gibson, pp. 265-66. See also Booth's discussion of this subject in The Rhetoric of Fiction, pp. 137-44 and passim. Critical Inquiry Autumn 1977 125 but carelessly from advertising copy to novel, as if the latter were merely a more elaborate version of the former. But this two-sided model is far too simple to account for the complexity of literary experience; and while they both, toward the ends of their essays, hint of the existence of still further audiences, neither critic develops this idea. In fact, there are at least four audiences implied in any narrative literary text, and Gibson's "mock audience" and Father Ong's "fictionalized audience" lump several of them together. While the dichotomy proposed by Gibson and Father Ong may be helpful in analyzing criticism, advertising copy, or very simple narratives, it fails with more intricate and ironic works. 2. The First Three Audiences Defined My model grows from the simple initial observation that all works of representational art-including novels-are "imitations" in the sense that they appear to be something that they are not. A piece of canvas, for example, appears to be the mayor or the Madonna; a tale about a nonexistent clerk and his overcoat appears to be a "true account." As a result, the aesthetic experience of such works exists on two levels at once. We can treat the work neither as what it is nor as what it appears to be; we must be aware simultaneously of both aspects. A viewer is hardly responding appropriately to Othello if he rushes on the stage to protect Desdemona from the Moor's wrath; 11 nor is the reader of the Sherlock Holmes stories who treats his idol as a historical being, makes pilgrimages to his home on Baker Street, and uses weather reports to determine when certain stories "actually" took place. 12 Neither, however, is it proper to refuse to mourn Desdemona simply because we know that she will soon rise, return to her dressing room, remove her makeup, and go out for a beer with Roderigo. Similarly, anyone who argues that Holmes is simply a fiction, and thus refuses to fear for his safety as he battles Moriarty, is missing the point of the whole experience. In the proper reading of a novel, then, events which are portrayed must be treated as both "true" and "untrue" at the same time. Although there are many ways to understand this duality, I propose to analyze the four audiences which it generates. More complex works-novelswithin-novels, novels with frames, epistolary novels, novels addressed to 11. One may recall a scene in Godard's film us Carabiniers in which, while at the movies, one of the characters knocks down the screen trying to join the bathing girl projected on it. 12. For an encyclopedia of this sort of reaction see William S. Baring-Gould, The Annotated Sherlock Holmes, 2 vols. (New York, 1967). Much of such research is, of course, tongue-in-cheek; but I wonder, reading through it all, just how much. 126 Peter J. Rabinowitz Truth in Fiction internal characters, novels with multiple narrators, certain ironic novels-may appear to have more than four, but these are only variations of these basic forms. (1) First, there is the actual audience. This consists of the ftesh-andblood people who read the book. While this is the audience in which booksellers have the most interest, it is the only audience which is entirely "real," and the only one over which the author has no guaranteed control. 13 (2) Second, the author of a novel designs his work rhetorically for a specific hypothetical audience. Like a philosopher, historian, or journalist, he cannot write without making certain assumptions about his readers' beliefs, knowledge, and familiarity with conventions. His artistic choices are based upon these assumptions, conscious or unconscious, and to a certain extent, his artistic success will depend on their accuracy. Demby's The Catacombs, for instance, takes place during the early sixties, and the novel achieves its sense of impending doom only if the reader knows that John F. Kennedy will be assassinated when the events of the novel reach 22 November 1963. Had Demby assumed that his audience would be ignorant of this historical event, he would have had to rewrite his book accordingly. Since the structure of a novel is designed for the author's hypothetical audience (which I call the authorial audience), we must, as we read, come to share, in some measure, the characteristics of this audience if we are to understand the text. Just as the implied author is often a person ethically superior to his ftesh-and-blood counterpart, so we are often forced to call upon the "best part" of ourselves when we join the authorial audience. But most novelists, even if they do call on our better selves, will only call upon those moral qualities which they believe the actual audience has in reserve, just as they try not to rely on information which we will not in fact possess. For most novelists are concerned with being read and hence try to minimize the distance between the actual and authorial audiences. There are, of course, exceptions. Some writers, such as the Joyce of Finnegans Wake, appear not to care about actual readers at all; others, such as John Barth, have intentions which are so subtle and complex that they can only write for an authorial audience which they know to be, at best, but a tiny portion of their actual audience; and Vladimir Nabokov appears to derive an almost sadistic satisfaction from knowing that his authorial audience is intellectually well above his actual readersalthough it is possible that Nabokov in fact writes for an authorial audience quite dose to his actual readers but writes in order to make that authorial audience feel intellectually inadequate. 13. Few critics have paid much attention to the actual audience. perhaps because it is much harder to gather evidence about real people than it is to discuss the implications of a text. Notable exceptions are Slatoff (With Respect to Readers) and Holland (The Dynamics of Literary Response and 5 Readers Reading). Critical Inquiry Autumn 1977 127 But even if an author makes a serious attempt to write for the "real people out there," the gap between the actual and the authorial audience will always exist. And since all artistic choices, and hence all effects, are calculated in terms of the hypothetical knowledge and beliefs of the authorial audience, this gap must be bridged by readers who wish to appreciate the book. The greater the distance-geographical, cultural, chronological-between the author and his readers, the more of a challenge this is likely to provide. If historically or culturally distant texts are hard to understand, it is often precisely because we do not possess the knowledge required to join the authorial audience. Topical allusions, in particular, lose their clarity through time. (What will our grandchildren make of Philip Roth's Our Gang?) But even such things as the belief structures of a society must often be "explained" to the reader before he can fully understand the text. Anna Karenina's anguish, for example, might well appear ludicrous to a contemporary student for whom divorce is the social norm. Liberal arts education, to a certain extent, provides the relevant information so that we can join various authorial audiences, and so that the rhetoric of various authors may have its impact; many footnotes do much the same thing. (3) Since the novel is generally an imitation of some nonfictional form (usually history, including biography and autobiography), the narrator of the novel (implicit or explicit) is generally an imitation of an author. He writes for an imitation audience (which we shall call the narrative audience) which also possesses particular knowledge. The narrator of War and Peace appears to be an historian. As an historian, he is writing for an audience which not only knows--as does the authorial audience-that Moscow was burned in 1812 but also believes that Natasha, Pierre, and Andrei "really" existed, and that the events in their lives "really" took place. In order to read War and Peace, we must therefore do more than join Tolstoy's authorial audience; we must at the same time pretend to be a member of the imaginary narrative audience for which his narrator is writing. I4 Whether they think about it or not, this is what all successful readers do when approaching the text. I5 14. Gerald Prince has been working in a similar vein; see his excellent "Introduction a l'etude du narrataire," Poetjque 14 (1973): 178-96. He develops the idea of "na1Tataire" (narratee), the person to whom the narrator is addressing himself. The na1Tataire, however, is someone perceived by the reader as "out there," a separate person who often serves as a mediator between narrator and reader. The "narrative audience," in contrast, is a role which the text forces the reader to take on. I think that my analysis, centering on an activity on the part of the reader, more successfully explains why certain texts evoke certain responses. 15. The pretense involved in joining the narrative audience is rather different from what Frank Kermode calls "experimental.assent" in The Sense of an Ending (London, New York, and Oxford, 1967), pp. 38-40. If I understand it correctly, "experimental assent" is an activity on the part of the actual audience through which it relates the novel to reality, 128 Peter J. Rabinowitz Truth in Fiction One way to determine the characteristics of the narrative audience is to ask, "What sort of reader would be implied if this work of fiction were real?" or, even better, "What sort of person would I have to pretend to be-what would I have to know and believe-if I wanted to take this work of fiction as real?" Normally, it is a fairly simple task to pretend to be a member of the narrative audience: we temporarily take on certain minimal beliefs in addition to those we already hold. Thus, for a while we believe that a woman named Anna Karenina really exists, and thinks and acts in a certain way; or, on a broader scale, that Yoknapatawpha County and its inhabitants really exist. Occasionally, a novel's demands are somewhat greater: in 1984, the narrative audience has knowledge of a series of "facts" about future world history. (To whatever extent the narrative audience possesses "knowledge" of nonexistent facts, these facts must, of course, be provided by the text itself. An author or narrator may allude to actual historical events and expect his audience to understand without explanation; but he cannot, and usually will not, expect the narrative audience to catch unexplained allusions to nonhistory.) Sometimes, however, we must go even further, and pretend to abandon our real beliefs and accept in their stead "facts" and beliefs which even more fundamentally contradict our perceptions of reality. In much science fiction, for instance, the narrative audience accepts what the authorial audience knows to be false scientific doctrine. And the process can become more complex still. Jules Verne's From the Earth to the Moon has obviously lost much of its impact as science fiction now that moon voyages have become a part of our lives. If we wish to read it and get anything like the intended effect, we must first, as authorial audience, pretend not to believe in moon travel so that we can then, as narrative audience, pretend to be convinced that it is possible.I 6 accepting the novel if it turns out to be "operationally effective," neglecting it otherwise. The pretense is closer to the "willing suspension of disbelief," except that I would argue not that disbelief is suspended but rather that it is both suspended and not suspended at the same time. In this article, I am not really concerned with the actual psychological processes by which a specific reader performs this act. This subject, however, is treated in a fascinating, if controversial, fashion in Holland's The Dynamics of Literary Response. Holland starts out with the same observation as mine: that we both believe and do not believe a literary text. But since he is concerned with the psychological actions of readers (particularly with their unconscious fantasies) rather than with the conscious audience roles implied by a text, his resulting categories (intellecting reader/introjecting reader) differ markedly from mine. 16. Here I differ signihcantly with Prince ("Introduction"). He asserts that, unless the text specifies otherwise, it is addressed to a "degre zero du naTTataire"-a narratee with minimal knowledge (primarily of language and logic) and no ethical or social character. Everything else that he knows must be indicated in the text. The narrative audience, in contrast, is much like ourselves, with our beliefs, our prejudices, our hopes, fears, and Critical Inquiry Autumn 1977 129 If we fail to pretend to be members of the narrative audience, or if we misapprehend the beliefs of that audience, we are apt to make invalid, even perverse, interpretations. For instance, the narrative audience of Cinderella accepts the existence of fairy godmothers (although the authorial audience does not share this belief). A reader who refuses to pretend to so believe will see Cinderella as a neurotic, perhaps psychotic, young woman subject to hallucinations. Although there are as many narrative audiences as there are novels, they tend to fall into groups, the members of which are quite similar. We don't have to shift gears very sharply to move from War and Peace to Gone with the Wind, different as those novels are in quality. Sometimes, however, a novelist is able to create a startling tone or mood by demanding a narrative audience which is unexpected or unfamiliar. 17 Franz Kafka's Metamorphosis is a good example. What is striking about this novel is not simply its fantastic premise, which is no more fantastic than the basic premises of Alice in Wonderland. Nor can the peculiar quality of Kafka's tale be explained purely in terms of the characters' odd reactions to Gregor's transformation. What strikes me as most curious about this book is the unusual-and for its time, perhaps unique-nature of the narrative audience. In Alice we are asked to pretend that White Rabbits wear watches, that Cheshire Cats fade away, and that numerous other miracles can take place. This is readily done by joining a narrative audience of the sort which is well known from fairy tales. In Metamorphosis, however, we are only asked to accept the single fantastic fact that Gregor has been transformed into a gigantic beetle; in all other respects, the narrative audience is a normal, level-headed bourgeois audience. Furthermore, we are asked to accept this without surprise; contrast the matter-of-fact opening of Metamorphosis with the equivalent passages in Alice or in Gogol's "The Nose," where the narrative audiences are openly warned that the events portrayed will be strange and unusual. There is no doubt that this curiously contradictory role that we are asked to play-half mundane, half fantastic-contributes greatly to the novel's disquieting tone. No other work that I can think of (except for those which imitate Kafka's technique, such as some of Marcel Aymes short stories) demands that we join a narrative audience which is at once so far from the authorial audience and yet so close to it. expectations, and our knowledge of society and literature-unless there is some evidence (textual or historical) to the contrary. Prince's method would never clarify the Pale Fire dispute; such clarification, as we shall see, can only occur if we assume a narrative audience with more knowledge than that explicitly called for by the text. 17. Father Ong, too, argues that much of the character of a literary work stems from the roles it demands of its readers. Once again, however, his analysis is weakened by his failure to distinguish authorial from narrative audience. 130 Peter J. Rabinowitz Truth in Fiction 3. The Relationship between Authorial and Narrative Audiences When Walker Gibson and Father Ong speak, respectively, of "mock readers" and "fictionalized" audiences, they make no distinction between authorial and narrative audiences. 18 One can agree that both the authorial and the narrative audiences are "fictions" (neither of them exists in the flesh), but they are fictions in radically different senses. When speaking of the authorial audience, we might more accurately use the term "hypothetical" than the term "fictional"; for as I have suggested, most authors, in determining the authorial audience they will write for, will try to come as close to the actual audience as possible. For to the extent that an authorial audience is invented, footnotes or other explanations will be required before the text can work as intended. Thus, while some authors such as Barth are forced, because of the subtlety of their intentions, to idealize and write for an audience they know does not exist (or does not exist in significant numbers), few authors intentionally strive for such a situation. The distance between authorial and actual audiences, in sum, may be inevitable-but as Tolstoy argues in What Is Art?, it is generally undesirable; and authors usually try to keep the gap narrow. The narrative audience, on the other hand, is truly a fiction; the author not only knows that the narrative audience is different from the actual and authorial audiences, but he rejoices in this fact and expects his actual audience to rejoice with him. For it is this difference which makes fiction fiction, and makes the double-leveled aesthetic experience possible. As we shall see, the author plays with this distinction and builds much of his effect on it. Similarly, the reader's act of joining the authorial audience is not really a pretense in the same way that joining the narrative audience is. As good readers, we usually try to become the authorial audience as much as possible. Thus, the authorial audience of Hemingway's For Whom the 18. In fact, almost all critics who discuss "the reader" are discussing a hybrid form which crosses the lines I have set up. For example, !ser's discussions (The Implied Reader) of the reader's discoveries are really studies of the experiences of the narrative and authorial audiences combined. Only toward the end of his book does he suggest a duality in the reader-a duality which stems from the differing personalities of the author and the reader. This intriguing notion is not developed, however; throughout the book, he treats the reader as a unity. So does Stanley E. Fish. In Self-Consuming Artifacts, the "reader" seems to be a complex combination of three audiences: at least two actual audiences (the current, informed actual audience, with Fish himself as representative, and the historical audience at the time the work was written), the authorial audience, and-when he is writing about fiction-the narrative audience. Indeed, a primary difference between Fish's model and my own is that his is horizontal (he is concerned with the progress of a unified reader through time) while mine is vertical (I am concerned with distinguishing the different levels on which a reader operates simultaneously). Gerald Prince's distinction ("Introduction," p. 180) between leeteur reel, leeteur virtuel, and narrataire (even though it does not, as I have pointed out, deal precisely with differing roles played by the reader) seems much more subtle. Critical Inquiry Autumn 1977 131 Bell Tolls knows quite a bit about the Spanish Civil War. If, as actual readers, we do not possess this information, we will ideally do our best to acquire it. There are, however, two circumstances under which we may pretend to join the authorial audience. First, we are often lazy; thus, we may decide to forego a research trip to the library and instead simply pretend to understand Hemingway's politics. Underlying this pretense, however, is the knowledge that we could really understand the book if we wished; and it is clearly a qualitatively different sort of pretense than the pretense of believing that Robert Jordan, Maria, and Pablo all exist. Second, there are cases where the authorial audience differs from the actual audience not simply because it has additional facts at its disposal (as in Hemingway) but because it operates under assumptions which fundamentally contradict our normal way of thinking. Through research, we may gain a better understanding of the beliefs of another culture. But no matter how much research we do, we are unlikely actually to believe what the authorial audiences of the Iliad or From the Earth to the Moon believe. To the extent that our joining of the authorial audience is a pretense, we are that much less likely to receive the work's intended effect. But the pretense involved in joining the narrative audience does not interfere with a novel's effect; it is, on the contrary, an essential and desirable element in it. And as we have seen, the reader who does not realize that this is pretense-the reader who looks up Humbert Humbert's murder trial in the newspapers-has clearly missed the point. Obviously, the narrative and authorial audiences are closer together in some novels than in others. I would argue that the distance between these audiences is a major element in any novel's structure. Indeed, I suspect that one could develop a classification of narrative literature stemming from a theory of "realism" expressed in terms of this distance. Surely, in the most "realistic" novels (War and Peace is a paradigm), the narrative audience is asked to accept very little-and usually nothing which in any way contradicts the fundamental beliefs or experiences of the authorial audience. To believe that Natasha exists, for example, in no way contradicts the authorial audience's general experiences-it is not at all improbable that such a person should exist and act as she does. At the extreme end of realism, narrative and authorial audiences are so close as to be almost indistinguishable, as in Henry Miller's Tropic of Cancer. When the distinction between the two disappears entirely, we have autobiography or history. In antirealistic, or fantastic, novels (Alice in Wonderland, for example), the narrative audience is asked to take on a great deal more-beliefs which, like belief in the White Rabbit, do moreover contradict the very beliefs and experiences of the authorial audience. Obviously, the wider the gaJr-the more unusual or outrageous the beliefs that the narrative audience is required to take on-the greater the effort required to 132 Peter J. Rabinowitz Truth in Fiction bridge it. Thus, we become more conscious that the novel is doubleleveled and that we must employ "pretense" to become involved in reading it. This, in turn, increases our awareness of the novel as art, and tends to diminish our direct emotional involvement in it. The metaphor of distance may suggest the possibility of quantitative measurement; but this is obviously deceptive, for such measurement is difficult, if not impossible, especially since the distances occur along many axes at once. How, for instance, can one compare the distance the novelist causes by asking us to believe that a man can walk through walls (Marcel Aymes "Le Passe-Muraille") with the distance caused by asking us to believe that a single mad genius named Fu Manchu secretly controls the politics of half the world? Yet while such a theory of realism would not be able to assign a quantitative "realism factor" to every novel, it would nonetheless have interesting consequences. Most importantly, it would escape from the notion of "verisimilitude," a notion which is misleading because it theoretically measures the novel against the "real" world but in fact only measures the novel against the world as perceived by the current actual audience. Since perceptions of the world change with history, the "verisimilitude" of a novel is less a quality inherent in the novel itself than a fluctuating value dependent upon the beliefs of the particular critics who read it. But by considering the distance between narrative and authorial audience rather than the more traditional distance between narrative and actual audience, realism can become grounded in its proper historical context. The realism of a novel will then depend upon the beliefs (or the author's perception of them) common to the audience at the time when (and in the class of readers for whom) it was written, rather than upon the beliefs common to the audience by which it is now read. The works of Horatio Alger, according to this notion, are extremely realistic, despite their almost comic mimetic inaccuracy, because they conform so well to the beliefs of the audience for which they were written. Similarly, with this definition of realism, we can view realism outside the confines of "schools" or "traditions." It is thus possible to see realism in a novelist like Alain Robbe-GriIlet, even though his works hardly fall into the Tolstoy-James tradition. His novels may require considerable imagination and even creativity from the authorial audience, but at least The Voyeur and Jealousy do not require much pretense before experienced modern readers can join the narrative audience. Our notion of audience distance is also suggestive when considering the didactic power of fantastic novels. It would seem that the greater the distance between authorial and narrative audiences (the less realistic the novel in our new definition), the less impact a moral lesson learned by the narrative audience is likely to have on the authorial audience. To use a Tolstoyan metaphor, the greater the distance, the less the possibility of "infection" from one to the other. Thus, for instance, I wonder about the Critical Inquiry Autumn 1977 133 power of the utopian novels of H. G. Wells. As narrative audience, we are asked to accept all kinds of scientific absurdities; once having done this, it is not difficult to accept the necessity of socialism as well. But in our role as authorial audience-where we see the novel as a novel-our great distance from the narrative audience takes its toll. Just as we recognize the obvious fictionality of our pretended scientific beliefs, so we are apt to abandon our newly found social beliefs as well. Since the narrative audience's beliefs in the naturalistic novels of Zola are not so far from those of the authorial audience, we are less apt to be conscious of their fictionality. It is thus all the more likely that the lessons learned as narrative audience will stick for the authorial audience too-and, unless the authorial audience is very distant from the actual audience, for the actual audience as well. The act of joining the narrative audience is not, of course, the ultimate step in literary interpretation-rather, it is the first and most elementary step. It is an essential step, however, and many novels fail to make an impact because they are unable to make their readers join the narrative audience. Such novels are usually deemed "unconvincing." The ability of a novel to convince is related to, but not identical to, its realism as I have defined it above. The realism of a novel depends on the distance between authorial and narrative audiences. The novel is more or less convincing, depending on how skillfully the novelist navigates us across that distance, and on how likely we are to be standing on the dock-in the authorial audience-when the trip begins. Most unconvincing novels are unrealistic, although Horatio Alger's realistic epics no longer persuade. And with skill, a novelist like Kafka can make a highly unrealistic novel entirely convincing: we make the initial leap of imagination and remain with him throughout. Only after having joined the narrative audience can we begin to study the meanings of a work of art. Thus, we cannot even perceive the moral dimension in Ayme's "Le Passe-Muraille" until we pretend to accept its initial "scientific" premise. A person who merely insists that no one can walk through walls can hardly respond to the hero's situation as the author intended. But the act of accepting this premise does not in itself lay bare the story's moral dimension. Decisions as to ultimate meaning or meanings, symbolic significance-in fact, any decision about the work as art-are made by the authorial and actual audiences, and not by the narrative audience. I do not wish to imply that in order to become members of the narrative audience, we must pretend to accept everything that the narrator tells us. There are unreliable narrators. 19 An unreliable narrator, 19. See Booth, The Rhetoric of Fiction, pp. 158-59 and passim. 134 Peter J. Rabinowitz Truth in Fiction however, is not simply a narrator who "does not tell the truth"-what fictional narrator ever tells the literal truth? Rather, an unreliable narrator is one who tells lies, conceals information, misjudges with respect to the narrative audience-that is, one whose statements are untrue not by the standards of the real world or of the authorial audience but by the standards of his own narrative audience. It would, for example, be trivial to argue that Dostoyevsky's Underground Man is unreliable because his statements are often false. He is unreliable because even after we have pretended to accept the beliefs of the narrative audience (that he, Liza, and the dinner party actually do exist), we still find that much of what he says--particularly when he analyzes motives--contradicts the other elements of his framework. In other words, all fictional narrators are false in that they are imitations; but some are imitations of people who tell the truth, some of people who lie. The narrative audience believes the narrator is a real, existing historian. But it does not automatically assume that he is an accurate historian any more than in reading a work of history we automatically assume the author to be accurate and truthful. It is probably in these terms that we have to understand William Demby's distinction between "real" and "fictional" in the passage quoted above. 20 4. The Fourth Audience The idea of unreliable narrators brings us at last to our fourth audience. This is the audience for which the narrator wishes he were writing and relates to the narrative audience in a way roughly analogous to the way that the authorial audience relates to the actual audience. This final audience believes the narrator, accepts his judgments, sympathizes with his plight, laughs at his jokes even when they are bad. I call this the ideal narrative audience-ideal, that is, from the narrator's point of view. Thus, inJohn Barth's End of the Road, the authorial audience knows that Jacob Horner has never existed; the narrative audience believes he has existed but does not entirely accept his analyses; and the ideal narrative audience accepts uncritically what he has to say. Similarly, in the Jason section of Faulkner's The Sound and the Fury (each section of which has a different ideal narrative audience), the ideal narrative audience believes that Jason has been victimized and sympathizes with his whining misery, although the narrative audience despises him. 20. Assuming that The Catacombs is a novel as it claims to be. The work is deceptive, however, and the implied author (indeed, the "real" author as far as I can tell from the little I know of Demby) and the narrator are all but indistinguishable. As a result, the authorial and narrative audiences of the frame novel (there is also a novel-within-thenovel) are almost indistinguishable, and we have the type of extreme realism found in Tropic of Cancer, if not an actual autobiography. It is always possible, therefore, that Demby means "true" in a more literal sense: that is, true on the level of the authorial audience. Critical Inquiry Autumn 1977 135 As a general rule, the distance between authorial audience and narrative audience tends to be along an axis of "fact," either "historical" or "scientific." That is, the narrative audience believes that certain events could or did take place. The distance between the narrative audience and the ideal narrative audience tends to lie along an axis of ethics or interpretation. The ideal narrative audience agrees with the narrator that certain events are good or that a particular analysis is correct, while the narrative audience is called upon to judge him. Much of the problem-and most of the ｪｯｹｾｦ＠ reading irony comes from sorting out these three levels, and feeling the tensions among them. But just as there are extremely realistic novels where the narrative and authorial audiences are indistinguishable, so there are non-ironic works where the ideal narrative audience is virtually identical with the narrative audience. This sort of ironic tension between audiences is found in nonliterary arts as well, even-surprisingly-in certain types of music. Obviously it cannot occur in non-imitative music. In a Bach Prelude and Fugue, for instance, there may be a gap between the authorial audience (which knows how to listen to several contrapuntal lines at once) and the actual audience (which often does not). But that gap is clearly undesirable, and when it exists, it diminishes the music's impact. In any case, there is nothing here which is analogous to the narrative audience. From the point of view of audience dynamics, however, imitative music is considerably more interesting and intricate. What, in fact, happens when an audience listens to the storm music in Liszt's Les Preludes? Is there an authorial audience which is listening to these sounds as pure musical sounds and a narrative audience which pretends to believe that it is actually hearing a storm? Or is there a single audience which is hearing music which it relates metaphorically to the sounds of a storm? I do not know the answer to this question, although I suspect that my model might shed some light on the traditional controversy over program music. There is one subtype of imitative music, however, where our literary model surely holds: music which is imitative of other music. A good example is Mozart's A MusicalJoke (K. 522). Why do we react to this as music which is good but comic, rather than as music which is simply bad? For there is no doubt that by normal canons of eighteenthcentury style the piece is quite dreadful. What we have here is a musical equivalent of an ironic fictional narrative. Mozart has created a fictional persona, the incompetent composer. The authorial audience knows that this piece is, in fact, by Mozart, and therefore knows that its violations of convention and taste are intentional. The narrative audience, on the other hand, pretends to believe that this piece is in fact written by some hack composer of the time and pretends to believe that the piece is an unintentional failure. Finally, this mock composer has written for an ideal narrative audience, an audience which will listen to his music and, not noticing its crudities, will think it is brilliant. Listening to this sort of 136 Peter J. Rabinowitz Truth in Fiction music-like reading Jason's narrative or Swift's "A Modest Proposal"-is intellectually demanding since we must play all these parts at once. A similar interplay between audiences is at work when we listen to Charles Ives imitating the sound of a second-rate marching band (as in The Fourth of july), or when we listen with ironic pleasure to the absolutely horrendous fanfares which pop up in Offenbach's La Belle Helene. 5. Two Types of Ambiguity and the Problems of Pale Fire Since we cannot read a novel properly until we have joined the narrative audience, reading problems can occur when we have difficulties in discovering precisely what are the characteristics of the narrative audience. Usually, this can be determined more or less by common sense and familiarity with literary traditions; occasionally, the author will have to take special steps to avoid confusion. The purpose of "John Ray's" introduction to Nabokov'sLolita, for example, is partly to provide the "factual ground work" for the narrative audience. It tells the reader that the narrative audience knows that Humbert Humbert, Lolita, and especially Clare Quilty existed and that the murder and trial actually did take place. As members of the narrative audience, then, we are free not to believe certain details in Humbert's account-but we are not free to believe that Quilty is a figment of Humbert's paranoia. Occasionally, however, there are more difficult problems. It is clear that the evidence in Frankenstein informs us that the narrative audience accepts the scientific possibility that a man may create life in a laboratory. But does the narrative audience of Dostoyevsky's The Double accept the scientific possibility of a double? This ambiguity in Dostoyevsky's novel is radically different, both in its structure and in its effects, from the more usual kind of ambiguity found in novels such as Lord jim. In this more familiar form, the ambiguity exists within the narrative audience: the narrative audience itself is unaware of where, exactly, the truth lies. This ambiguity may relate to ethics (was Lord Jim a coward?), to motives (why does Quentin Compson commit suicide?), to facts (was Owen Taylor, in The Big Sleep, murdered?), or to anything else which may normally be questionable. The effect, however, is "internal." That is, since we are confused on the level of the narrative audience, it is possible to ponder this ambiguity "within the world of the novel." Particularly if the ambiguity relates to ethics or motives, the more we ponder it, the more deeply we get into the characters. The second kind of ambiguity, found in novels like The Double, is far less common and usually accidental. It occurs when we are faced with an ambiguity about which of several narrative audiences we are to acceptalthough each potential narrative audience may itself face no ambiguity. Critical Inquiry Autumn 1977 137 Thus, for example, we know that the ideal narrative audience believes that Golyadkin has a double. But we do not know whether the narrative audience accepts the scientific possibility of a double (in which case Golyadkin is being persecuted-at least in part-by forces external to himself) or denies such a possibility (in which case Golyadkin's double is but a mad projection, a result of his paranoia). In The Double-where the ambiguity is probably accidental-the ambiguity makes it difficult to determine how we are to read the book on the surface; but, curiously enough, it makes little difference to its ultimate meaning. Even if, in our role as members of the narrative audience, we assume that the double actually does exist, in our simultaneous roles as members of the authorial and actual audiences, we will interpret it as a symbolic representation of some inner force anyway. Much of the controversy over The Turn of the Screw arises from precisely this same structural source. Once again, the beliefs of the ideal narrative audience are clear; but where does the narrative audience stand? Is James' narrative audience for this novel the narrative audience traditionally implied by gothic novels, one able to believe in ghosts? Or is it closer to the narrative audience of Portrait of a Lady? James (unlike Bram Stoker, for instance, in Dracula) has made no effort to tell us what are the beliefs of his narrative audience; and since we are used to viewing James as a "realist," we have difficulty accepting the surface meaning of his tale. Perhaps my critical bias is showing here, but I find that in this novel, as in The Double, it ultimately makes little difference which reading one chooses. Dracula is, in fact, a novel about vampires; but even if the narrative audience of The Turn of the Screw accepts ghosts, the authorial audience is apt to treat them as in some sense symbolic when called upon to interpret the novel. The ambiguities of Pale Fire are similar but more difficult, and I would like to conclude by showing how my model, by revealing Nabokov's use of this second kind of ambiguity, explains the novel's persistent ability to baffle readers and critics. Clearly, the novel is largely "about" the relationship between the poem and the commentary, between the poet and his critic. But the subject of a novel is not its substance, and to state a subject is not to explicate it; and while nearly every critic agrees that this is, in fact, one of Nabokov's major concerns, there has been no agreement on just what Nabokov is saying about this relationship. For one cannot discuss the relationship between two things until we know what those two things are, and as soon as we try to examine Shade and Kinbote in any detail, we are confronted by those plaguing problems of "fact" which characterize the Pale Fire controversy. For example, does the narrative audience believe that both Shade and Kinbote actually exist-as most critics seem to believe--or does it believe that one has invented the other, or that a man named Botkin has 138 Peter J. Rabinowitz Truth in Fiction invented them both? Clearly, until we know the answer to this question, we cannot adequately discuss the question of whether Kinbote's commentary is totally "wrong" (as Richard Kostelanetz suggests when he says, "Kinbote's characteristic fault is missing the point"),21 or whether, to the contrary, Kinbote has actually understood Shade's poem, as Stegner and others have argued. And until this question is answered, no statements about the nature of the relationship between commentary and poem are liable to be worth very much. And does the narrative audience agree with the ideal narrative audience that the country of Zembla really exists? Is the relation of poem to commentary the relation of "invention" to "history" (even if somewhat distorted history) or the relation of invention to invention? Again, we cannot make any meaningful statements about the nature of the connections between the writings of Shade and Kinbote until we have answers to these questions. There are other thorny problems as well. Is Shade's poem any good in the eyes of the narrative audience? Does the narrative audience see it as "competent but pedestrian" (as suggested by Claire Rosenfeld)22 or are we to treat it as Andrew Field does-an important poem in its own right?23 Who really shoots whom at the end of the book, and why? I wish that I could answer these questions; but I cannot. I think we are now, however, in a position to understand why these kinds of problems have arisen in this novel and why, in fact, they can never be solved. For convenience, let us limit ourselves to two of the questions raised above. (1) Does the narrative audience believe that John Shade is a real poet, or does it believe that he is a figment of Kinbote's imagination? (2) Does the narrative audience believe that the country of Zembla exists? As we have seen, the novel does not provide answers to these questions. This in itself may not seem remarkable. There are many novels where facts remain unknown to the narrative audience. The narrative audience of Camus' The Fall never learns whether the events described by the narrator are true or not; the narrative audience of Robbe-Grillet's Jealousy never knows whether A. has been unfaithful or not. But the ambiguities of Shade's and Zembla's existence do not exist for Pale Fire's narrative audience. They are external to the world of the novel. We can see this as soon as we start to reconstruct the nature of the narrative audience. It is, first and foremost, an intellectual audience which is interested not only in literature, but also-since the novel is an imitation of a critical treatise-in literary criticism as well. From the broad range of references made by Kinbote, we can further assume it to 21. Richard Kostelanetz, "Nabokov's Obtuse Fool," in On Contemporary Literature, ed. Kostelanetz (New York, 1964), p. 432. 22. Claire Rosenfeld, "The Shadow Within: The Conscious and Unconscious Use of the Double," Daedalus 92, no. 2 (Spring 1963): 342. 23. Field, Nabokov, p. 106. Critical Inquiry Autumn 1977 139 be well read, familiar not only with Robert Frost but also with some of the more obscure passages in Timon of Athens. It follows from this implicit sophistication that the narrative audience is familiar with the basic facts about contemporary art and politics and, if given the name of a country or a famous American poet, would be able to tell whether it is real or fabricated. The narrative audience, therefore, either knows that John Shade is a great American poet, or else knows that Kinbote has invented him and his poem; it either knows that there is a country named Zembla, or else knows that Kinbote has invented it and its history. The narrative audience cannot be undecided on these questions. 24 Here then lies an obstacle which prevents us from even the most superficial understanding of the text. Although the narrative audience has facts about Shade and Zembla at its disposal, we do not know what they are, and are hence totally unable to join the narrative audience. There is nothing which serves the function performed by John Ray's introduction in Lolita. The situation is similar to that faced by readers of The Turn of the Screw and The Double-but in Pale Fire, the ambiguities seem intentional, and there is no simple way for the authorial audience to resolve them. As I suggested earlier, the moral ambiguities of a novel like LordJim, which exist on the level of the narrative audience, have the effect of immersing us more deeply in the world of the novel-they get us more involved as narrative audience. The ambiguity of Pale Fire has an entirely different--even opposite--effect. Since this ambiguity is perceived only in our capacities as authorial and actual audience, it makes us more aware of the gap between authorial and narrative audience, and hence of the novel as art, as construct. It is thus difficult to get involved in Pale Fire as narrative audience, and for many readers, including myself, the book is generally unmoving, witty and brilliant as it may be. This exaggerated artificiality and remarkable use of our second kind of ambiguity is quite consistent with Nabokov's general aesthetic, but it makes Pale Fire a frustrating novel to read, and in some respects an impossible one. It is not simply that the novel raises difficult philosophical questions, as The Brothers Karamazov does; it is rather that we can't tell precisely what issues the novel does address. For as I noted above, the central concerns of Pale Fire are not whether Shade and/or Zembla exist. Rather, they arise from the relationship of poem and commentary. But the novel's ambiguity prevents us from knowing just what that relationship is. Thus, we may say vaguely that Pale Fire has something to do with the nature of imagination, the nature of criticism, and the relation 24. Although it would be quite impossible, given Prince's methodology (see n. 16), for us to make the same assumptions about a narrataire. In a situation like this, his methodwhile not inherently "wrong"-seems to have little connection with the way we actually read books. 140 Peter J. Rabinowitz Truth in Fiction of truth to illusion. Yet until we know whether or not Shade and Zembla exist, we cannot know, with any more specificity, just what the novel is doing with these subjects-what questions it is asking, what solutions it is proposing. If both Zembla and Shade exist, we have one novel, probing one set of problems; if Zembla does not exist, but Shade does, we have an entirely different novel, with another set of problems; if ... How then is one to read the book? The only way, I suppose, is to make an arbitrary choice about which narrative audience one wants to join-or to read the novel several times, making a different choice each time. As in a game, we are free to make several opening moves; what follows will be dependent upon our initial decision. Simply with respect to the questions suggested above, we can generate four novels, all different but all couched, oddly, in the same words. And as we begin to ask further questions-Has Shade invented Kinbote? Is the poem a good one in the eyes of the narrative audience?-the number of possible novels begins to proliferate at a geometric rate. 25 At present, this model of the four audiences is less a complete theory of fictional structure than a suggestion for an approach which needs further study. The total use of the scheme-just what questions it can answer and what insights it can bring-has yet to be determined. But it does seem a promising way of looking at literature-it does provide a useful critical vocabulary for speaking of fiction, which in turn affords a new perspective on the multifaceted nature of literary experience. It would be useful, no doubt, in examining such works as Robbe-Grillet's In the Labyrinth or Doris Lessing's Briefingfor a Descent into Hell, to mention only two novels where the problem of truth underlies the reader's difficulties in interpretation. And it would certainly help us sort out the complex ironies of novels with multiple narrators, such as Lermontov's A Hero of Our Times or Faulkner's As I Lay Dying. Although I have talked primarily about fiction, my method can be adapted to drama as well. It would surely provide a new angle on the problem of aesthetic distance posed by contemporary theater, especially those attempts (for instance, by Pirandello or more recently by the Living Theater) to break down barriers between actors and viewers, and hence, between narrative and actual audiences. In music, it might help us better understand program music, espe25. There is at least one novel of infinite regress among the many Pale Fires. Suppose the narrative audience knows that the famous poet John Shade has invented a character named "Kinbote." It will then treat the text before it not as a critical commentary but as a "novel." But if it is a novel, it in turn has its own narrative audience-Narrative Audience 2. That narrative audience, in turn, must decide whether Shade and Kinbote both exist-and if it decides that Shade has invented Kinbote, it too will view what it has in front of it as a "novel," with its own narrative audience-Narrative Audience 3. This narrative audience, in turn ... Critical Inquiry Autumn 1977 141 cially that type in which the musicians, in performing the music, pretend to be other performers performing other music. (This curious technique may not cause particular difficulties when we listen to Don Giovanni's serenade or the interpolated dance music at his parties, but it produces serious analytic problems in its more avant-garde manifestations: Berio's Recital I [jor Cathy], for instance, or Peter Maxwell Davies'Missa Super "L'Homme Arme.") And, to return to literature, my model would cast some light on the eternal problem of the role of the reader's beliefs. By making distinctions among actual beliefs, authorial beliefs, narrative beliefs, and ideal beliefs we can talk about this issue with considerably more clarity and specificity than has hitherto been possible. In short, this model is at the very least a new handle by which we can grasp familiar but perplexing problems. How far it enables us to carry them remains to be seen.