Peter Swirski, Hong Kong University
If you suggest some fresh and more ingenious reason for doubting that the author meant what he said, or that what he said had any truth content of the remotest kind, you win ten points by the rules of the modern critical game; you lose ten points if you suggest he said it because he meant it, and twenty if you suggest he said it because it was true.
-George Watson, Times Literary Supplement
The problem of interpretation, central to literary theory and critical practice, is one of the most difficult on anybody's research agenda. As if to illustrate the point, much like professional politicians, philosophers and literary scholars who write on interpretation appear to agree only to disagree. While there have always existed schools of interpretation, today's proliferation of positions, oppositions, suppositions, presuppositions, and propositions on the matter seems to flow from the centrality of the notion of text in critical theory and practice, and the corresponding marginalization of the concept of work. Arguing for a categorical distinction between texts and works, this paper points to an alternative to such textualism professed most frequently, though not exclusively, by scholars of poststructuralist persuasion. With this goal in mind, it discusses in turn the aesthetic distinction between works and texts, the role of intentions in interpretation, the distinction between work interpretations and textual readings, and the role of cognition in literary research.
From Text to Work. What is a work of art? Is it identical with its structure (text)? If not, what is the relationship between the work and its structure? What elements other than textual may be constitutive of an artwork's identity? What is the interpretive relevance, if any, of the artist's intentions in the creation of the work? The merit of various theories of interpretation can be gauged roughly by their success in handling these questions while adhering to commonsensical critical intuitions and common critical practice.
Elsewhere I have defended at length the view that a work of art is an aesthetic structure executed by the artist in a particular art-historical context.1 While not highly regarded in literary theory, this view has traditionally underwritten our interpretive and critical practice, becoming in recent years the subject of intense study by philosophers in analytic aesthetics.2 One thing that this new aesthetic paradigm makes clear is that works of art, and thus works of literature, are not equivalent to their structures (or texts) when considered from the point of view of identity. Building on this interdisciplinary research, I would like to focus on the interpretive aspects of this work-text nonequivalence.
A good way to start may be to examine what features of a work of art are relevant to its aesthetic appreciation. Despite conclusive refutations (see below), a sizeable body of criticism continues to profess that, interpretively speaking, a work of art is equivalent to its structure. On this view a work of art, together with all its aesthetic and artistic features, is not distinct from the physical entity (the text) that we get at by means of purely sensory perception. I will apply the label of "aesthetic structuralism" to this family of views that holds that the evidence for interpretation rests solely on the evidence of the senses. For the aesthetic structuralist, even though artworks may indeed have features other than directly perceptible ones, which can even be of considerable historical or biographical interest, such properties have no bearing on these works' aesthetic or artistic properties. Such aesthetic or artistic properties are judged to be independent, for example, of the work's history of creation, including the significant subset of this history bracketed off by the artist's executive intentions. The inclusion of work-related facts, whether culled from the artist's life or from the history of the work's composition, would on such a view take us outside the perceived "art-ifact," contrary to the structuralist sine qua non.
It may be useful to distinguish here a variant of aesthetic structuralism specific to literary theory that I will call textualism. The label is broad enough to include the activities of scholars who would do away with the concept of work altogether and those who only confuse the interpretation of a work with the apprehension of the text. The change of nomenclature and the corresponding shift in reference is warranted by the fact that in literature straightforward structuralism is not possible. Unlike music or painting, where one can directly perceive at least some aesthetic properties of a work, the interpretive access to a novel or a poem must be preceded by the activities of reading and linguistic comprehension. Whether the purely musical qualities of language, such as rhythm, rhyme, alliteration, onomatopoeia, or perhaps the pictorial qualities in shaped poetry, escape this stricture is an open question worth an investigation in its own right.
Among the most influential textualist statements one must count W. Wimsatt and Monroe Beardsley's "Intentional Fallacy," and Northrop Frye's The Anatomy of Criticism.3 Quite apart from their venerated status as theoretical pillars behind the New Critical practice, these textualist theses remain influential among contemporary theorists of interpretation. The dethroning of New Criticism by structuralism and poststructuralism has not, after all, dulled the edge of textualist avowals from influential thinkers in the field. In Languages of Art, for example, Nelson Goodman continued to expound the textualist theses in a manner remarkable for its conciseness and forcefulness: "a literary work … [is] the text or script itself."4 More recently, an equally uncompromising endorsement of textualist aesthetics was developed by Goodman and Catherine Elgin. There the matter is restated, if possible, in even more unambiguous terms: "works and texts [are] one-to-one correlated, and the work can alternatively be identified with the text."5
What is missing from this strictly structural approach to interpretation? For an answer I turn first to Kendall Walton, who provides a simple but compelling rebuttal to the equation of texts and works.6 Walton discusses three kinds of features of art categories: standard, variable, and contra-standard. In literary works, these correspond roughly to genre properties. In detective fiction, for example, the perpetration of the crime is standard, the vagaries of the plot are the variables, and absence of a solution to the mystery would be contra-standard. Walton's thesis is that aesthetic qualities of works of art are not determined by just structural (pictorial, textual, etc.) features, but by these features as seen through the work's category. To illustrate his point, the philosopher invokes a possible world scenario involving a society that makes guernicas, works of art much like Picasso's Cubist painting, only in various bas-relief dimensions. Next to the three-dimensional guernicas, the same work on a flat surface would seem impoverished in some of its attributes. The contemplation of this scenario suggests to Walton that the standard features of Picasso's work depend on the particulars of the guernica category. He writes that the painting might "seem violent, dynamic, disturbing to us. But I imagine it would strike [the alien society] as cold, stark, lifeless … but in any case not violent, dynamic and vital."7(347).
In the next section I will consider a different set of arguments against textualism based on the creatability of works of art (literature). Here it may be instructive to briefly consider another reason why aesthetic structuralism does not work even in the paradigmatic case of paintings. Aware of the problems of reducing works to their structures, early on Frank Sibley sought to overcome them in "Aesthetic Concepts" by means of the so-called supervenience thesis (its significance is in proportion to the frequency with which it has since been adapted in literary theory). While conceding that aesthetic qualities are not equivalent to pictorial ones-which in itself represents a dramatic retrenchment-Sibley argued that they supervene on the latter, without being reducible to them. An attentive reader will not miss that, much as in the more robust versions of aesthetic structuralism, the supervenience thesis again puts the philosopher in a quandary by assuming that the history of a work's creation is irrelevant to the appreciation of its aesthetic qualities. To rescue his proposal from bootstrapping, Sibley thus invoked the notion of taste, the exercise of which is said to allow contemplators of fine art to map strictly pictorial properties onto aesthetic ones. His entire theoretical edifice thus rests on the notion of taste and its key ability to discern aesthetic-or, as the author calls them, taste-features of a work.8
Are there reasons for rejecting this new version of aesthetic structuralism? I think so. The conclusive one is that the notion of taste necessarily entails an extrastructural knowledge of the art-historical context. In any plausible understanding of the term, taste is not a congenital faculty but an acquired one. To acquire and refine it usually involves a prolonged exposure to artworks as well as an intimate knowledge of many aspects of their history (creation, reception, genre, influence, etc.). The acquisition and exercise of taste is thus an inherently comparative process, in contrast to the aesthetic structuralist's contextless approach. Similar arguments, developed with a more typically literary audience in mind, led other writers on the subject, notably David Daiches, to similar conclusions. Taste, argues Daiches, is always a developed (i.e., acquired) faculty. It is always "the sum of [contextual] discriminations made available by attentive reading in a large variety of literary modes."9
Creating ex nihilo. Although the above arguments reveal the essential drawbacks of the textualist position, they do not go far enough. Genre-categorical knowledge is an important extratextual element that contributes to a work's aesthetic properties, but it is hardly the only one. Further counterexamples to textualism can be found in instances where the work's structure and genre remain the same, but other factors (e.g., art-historical context) occasion a different aesthetic perception.10
One of the better known refutations of the equivalence between a work of art and its structure was developed by Jerrold Levinson in "What a Musical Work Is."11 The author provides, in fact, not one but three separate counters to aesthetic structuralism based on its exclusion of: (1) the event of creation, (2) the art-historical context, and (3) the musical performance means. As the last argument takes us outside the provenance of literature and literary studies, I will not dwell on it here.12 Levinson's argument about artistic creation takes the form of a two-part theorem. Part one proposes that if works were mere structures they would not be creatable since-at least if Platonism is true-structures (such as texts) are types that exist at all times. For example, the text of John Updike's Roger's Version could (in the logical sense) have been written before 1986. As such its structure-type must also have existed prior to the time when Updike composed it. From this premise, part two of the argument concludes that if artists indeed create their works-in the sense of bringing them into existence-works of art cannot be equivalent to their structures.
One conceivable line of objection to this argument would be to question the very notion of artistic creation. Notwithstanding the rhetorical tropes of nihil novi sub sole, such opposition would soon, however, find itself on shaky ground. The creatability of artworks is perhaps the most incontrovertible creed associated with the activity of artists in society. The entire Western tradition in art is rooted in the creation of works that did not exist prior to the artist's activity of composition. In fact, at least some of the status, importance, and symbolic value bestowed on such works owes to the recognition of their having been brought into existence.
For those who may find the Platonism of the first refutation hard to swallow, the second counter to aesthetic structuralism is designed to demonstrate that a work of art acquires at least some of its aesthetic properties by being created in a specific art-historical context. I assume herein that the typically artistic attributes such as skill, originality, influence, complexity, and so forth, are a proper subset of aesthetic ones. One could, I suppose, question this move and argue that the examples and case studies below make the case for differences between the artistic, but not the aesthetic, features of literary works of art. At this point, however, the burden of proof shifts onto the skeptic. The inclusion of artistic attributes in the process of aesthetic appreciation enjoys a venerable tradition in all writings on art and its history, and is supported by the intuitions of most, if not all, writers on the subject.
Speaking of art-historical context, we can discern two groups of contextual factors: those relating to the general environment in which a work is composed, and those specific (one might say idiosyncratic) to the creating individual. The first group comprises pertinent factors from the sociocultural history up to the process of creation, including, but not limited to, the history of the artist's antecedents and contemporaries, or the dominant styles and/or influences. The second group may include the creator's indebtedness to other artists, his current and past repertoire, or style and artistic achievement (the list is, of course, much longer). Levinson's thesis is simple and persuasive: in contradistinction to textualism, different configurations of the art-historical context differentiate the aesthetic and artistic features of works whose structures (texts) might even be identical.
"What a Musical Work Is" fashions a series of scenarios in defense of this thesis, all from the world of music. For us it may be more pertinent to contemplate examples from literature. (1) A work textually identical with Truman Capote's In Cold Blood (1965) but composed by Norman Mailer in 1995 would be aesthetically different from Capote's work. As a Mailer work, In Cold Blood would not longer be contemporaneous with Tom Wolfe's new-journalistic The Kandy Kolored Tangerine Flake Streamlined Baby. It would also be derivative, coming after his The Executioner's Song (1979), which is indebted to Capote's nonfiction novel. As such, Mailer's In Cold Blood would be much less daring, upsetting, disturbing or topical than Capote's work, since it would be created in a literary context which has largely absorbed the breakthrough techniques of the nonfiction novel of the 1960s. (2) Ernest Hemingway's The Sun Also Rises (1926) is held to be a highly original work of literature. Its lean, sparse, and understated prose marks the maturation of a new American style, perhaps never before practiced with such skill and bravado. But a book written in 1946, with the same text as the Hemingway novel, would produce a work that would be stylistically unoriginal within the by then well-established hard-boiled tradition. (3) Mailer's breakthrough war novel, The Naked and the Dead (1948), is strongly influenced by John Dos Passos's U.S.A. (1936). However, a work identical in text with Mailer's but written by Jack London (or even by Dos Passos himself) would hardly have the property of being Dos Passos-influenced. And, written before U.S.A., it would strike readers as highly experimental, which is no longer the case with the more imitative work by Mailer.
These analytically and intuitively compelling arguments reveal, I think, the untenability of textualist claims with regard to interpretation. To forestall potential demurs that the above cases are cooked-up counterfactuals with little bearing on reality, I would like to turn briefly to a real-life example, Stanislaw Lem's early novel, The Investigation (1959). Relying on a nuanced interplay between the genres of detective fiction and the ghost story, the plot follows Lieutenant Gregory of Scotland Yard and his investigation of corpses that disappear from local morgues in circumstances that, despite all odds, point toward resurrection.
Bearing the arguments from this section in mind, it appears that to appreciate Lem's aesthetics we must take into account at least some extratextual factors in addition to what is in the text. Among these are the author's lifelong interest in science, the investigative implications of the second law of thermodynamics, which lie behind the corpses' "resurrections," or his well-documented penchant for breaking genre barriers.13 But the clearest example of sensitivity of a work's attributes to its art-historical context is provided by The Investigation in the context of its literary twin, The Chain of Chance. In this 1976 mystery, Lem returns to the same philosophical concerns that animated his novel from 1959, to the same macabre implications of the second law, developed within the same mixed-genre framework. This calculated move was designed-as he revealed in numerous pieces of correspondence and interviews-to fashion a direct aesthetic and cognitive dialogue with the earlier work.14 No surprise, then, that the solitary seeker of truth, the investigative routine, the red herrings, the eccentric scientist, the air of bafflement and menace, the humanless "crime," return in the companion novel. Given the magnitude of the similarities and allusions between The Chain of Chance and The Investigation, it is therefore impossible to appreciate fully the features of either work in isolation. Questions of originality, complexity, plot development, skill, and other points of artistic and aesthetic distinction cannot be decided for The Chain of Chance on the basis of the text alone, but necessitate a comparison with its earlier literary twin (and vice versa).
A diehard textualist might seek refuge behind the claim that such counterexamples to equating literary works and texts do not refer to the attributes of works-and thus to aesthetic differences between them-but rather to the attributes of their creators. On reflection, however, this objection must be ruled unsound. Although there are properties that are, indeed, equally attributable to writers and to their works-for example, the property of being original or derivative-there is no way of reducing all aesthetic properties to statements about authors. It would make little sense to transfer attributes like "archaic" or "skillfully plotted" from works to their creators.
Before passing on to the discussion of authorial intentions in interpreting literature, here is a quick summary of my theses so far. The aesthetic properties of artworks, including literary works, are to a large extent relational, that is, not contained in the text. For this reason one cannot equivocate between literary works and texts or, in more general terms, between artworks and their structures. For the same reason, the consideration of the relevant art-historical context must figure in appreciation and interpretation of a work's aesthetic and artistic qualities. In fact, as I documented at length in Between Literature and Science, contextual (extratextual) knowledge is indispensable even on such a fundamental interpretive level as grasping the basic story content.15
Intentions in Action. The above theses have several implications for students and scholars of literature. The most important is that, since texts are not equivalent to works, to interpret literature aesthetically we must approach it in terms of literary works, that is, in ways that reflect their creation in a specific art-historical context. An important subset of this context will always be formed by the author's executive intentions, that is, the intentions that actually went into the work's creation.16 It is for no other reason that F. Scott Fitzgerald's biographer, Matthew J. Bruccoli, can claim that, because Fitzgerald was a saver of working drafts, many of his intentions could be recovered by a study of the manuscripts, corrected galleys, and even his marked copy of The Great Gatsby. Like other epistemic domains, intention attribution relies on a potentially fallible, but self-correcting, process of hypothesis formation on the basis of available evidence, as the successful interpretive work of historians, archeologists, cryptographers, or indeed literary scholars, demonstrates.
Even though all this imposes some fundamental constraints on theories of interpretation, especially those inclined to democratize all interpretations in the name of textual jouissance, it is crucial to recognize what these theses do not entail. For one, the appeal to executive intentions does not entail radical intentionalism and the indefensible view that authorial intentions determine the single meaning of the work. Nor does it entail that there always exists the single best interpretation of a given work. The only conclusion so far-albeit one fraught with critical consequences-is that any interpretation of a literary work of art as a literary work of art must be consistent with the best available account of the author's executive intentions.
To accept this conclusion is to readmit authorial intentions into the way we do literary criticism. One must not, of course, confuse this moderate intentionalism with any form of intentionalist fallacy. Interpreting literature in light of the best account of the author's executive intentions is not tantamount to crediting all of the writer's pronouncements about her novels as the most truthful accounts of her intentions. Although artists frequently do communicate their intentions directly by means of interviews, prefaces, letters, and the like, the main body of interpretive evidence will often remain the text itself. There are many reasons for this. The simplest and perhaps most common is that even the best intentions can evolve or even change entirely during the execution of a work. Moreover, complex intentions may be difficult to articulate concisely and the finished work may already be the best expression of its author's design, irrespective of her previous or later elucidations. The author's account can be rendered equally inaccurate by opportunism, forgetfulness, or even lack of explicit awareness of the kind that makes writers frequently report surprise at their own creations. Finally, one need not be confronted with an anonymous medieval palimpsest to recognize that in many real-life situations, such as picking up an obscure paperback in a Club Med resort, the author's extratextual intentions may simply be inaccessible to the reader.
All this is true. Equally true, there is nothing in any approach that grounds interpretation in the author's executive intentions to guarantee that these intentions will always be accessible, homogeneous, or even conducive (in the cooperative sense) to critical interpretation. Certainly, no one claims that for every work of literature one can always determine its author's intentions with complete accuracy. This epistemic truism yields, however, no ground to epistemological arguments against intention attribution. Like all scientific domains, intention attribution relies on a potentially fallible, but self-correcting, process of theory formation on the basis of best available evidence at any given stage of research. Moreover, potential objections of this type would have to be squared with successful reconstructive work of historians, archeologists, cryptographers, or indeed literary scholars themselves whose success relies at least in part on the attribution of intentions.
No appeal to intentions can, of course, settle all interpretive questions. Literary works can carry meanings (e.g., puns, allusions, irony) that did not figure in their authors' executive intentions. According to Robert Stecker, to deal with such unintended meaning critics ought to start "with the conventions, linguistic, cultural, and artistic, in place at the time a work is created. It is very plausible that much unintended meaning can be accounted for by the working of such conventions. For example, it is plausible that it is part of the meaning of a work that it puns on a certain word in virtue of linguistic conventions about word meaning and artistic conventions that puns are artistically significant features of works in the genre in question."17 No less significant, moderate intentionalism allows for failed intentions, that is, in cases when the evidence shows that the author tried to express a proposition P, and the actual result means not P. The intentionalist approach of the moderate variety does not mean that we interpret the work as meaning P, but makes it possible to distinguish between P and not P, which is of aesthetic consequence.
A literary example can highlight the issues at stake. Thomas Disch's brilliant The M.D.: A Horror Story (1991) positions itself-much like Henry James's canonical The Turn of the Screw-on the brink of a ghost story and psychological dementia. The story's protagonist, Billy Michaels, succumbs as a boy to the temptations of an evil deity, Mercury. Armed with Mercury's supernatural powers, Billy grows up to be a distinguished medical researcher and business tycoon, while unleashing evil on his family, friends, and thousands of hapless victims of ARVIDS, a worse-than-AIDS epidemic he sets loose on America. Disch takes great care to present Billy's apparent supernatural acts in an ambiguous light. For one, the magical disasters always have a possible realistic interpretation. Thus, when Billy puts a paralyzing curse on his brother, Ned, the symptoms are identical with those of Bubby Corning, another character in the novel, who is a vegetable owing to natural causes. When Billy curses his grandmother with baldness, she loses her hair during a salon treatment, opening up a possibility of natural causality (e.g., a super-allergic reaction to shampoo chemicals). Halfway through the novel Billy wonders: "Of course, there was Ned, and grandma O. going bald, and the rest, but all those things could have happened naturally."18
People losing teeth, trees failing to succumb to blight, the tragic death of a father, the birth of a monstrously disfigured baby, the rise and spread of an epidemic-all these and other events could, in principle, be explained by natural causes. It would seem that, on one interpretation, all these events ought to be interpreted as natural in origin in view of the author's sustained effort to paint the story in a realistic light. The genre is another pointer in this direction: This self-proclaimed horror story is developed as a bildungsroman and a family saga, in style and tone conforming to their realistic conventions.
Taken separately, the many instances of what looks like magic could thus be explained by coincidence, accident, or genetic or psychological aberration. Yet, considered as a series, their joint likelihood exceeds the realm of probability, shifting the interpretation squarely to the supernatural. The clearest sign of Disch's intention that The M.D. should be interpreted in light of this parallel design, that is, as at once natural and supernatural, comes toward the very end of the novel. Judith, the survivor of the family massacre, learns from a letter left by Billy's long-deceased father that the boy (and his offspring) may have suffered from Huntington's chorea, a genetic malady marked by acute dementia and super-violent behavior. Inserting this startling revelation only a few pages before the end, Disch is at pains to offer an alternative explanation for Billy's (and his son Judge's) apparent visitations by Mercury, for their belief in supernatural powers, and for their pathological tendencies.
The author's intentions can be reconstructed as follows. The novel is meant to be read and enjoyed both as a supernatural story, with Mercury's magic behind Billy's evil reign, and as a realistic tale of psychopathological dementia, with Mercury a figment of Billy's diseased mind and a series of coincidences in lieu of supernatural causality. This ambitious design, which makes The M.D. such a rich and fascinating work, is further reinforced by the genre complexity and playfulness, which characterizes so many of Disch's earlier works (e.g., Camp Concentration, 1968; White Fang Goes Dingo, 1971; On Wings of Song, 1979). Yet, in the course of this complex novel, the attentive reader can detect three minor incidents that do not conform to its author's apparent design. In one example, three family members hear a bell ringing upstairs (Mercury's summons for Billy), although there is nobody there to ring it. Since it is impossible to attribute this event to natural-however coincidental and improbable-causes, through a domino effect the reader is impelled toward a consistently nonrealistic interpretation of the entire narrative.19 The point is that, conceived as a double entendre, the book loses much in aesthetic and artistic complexity if such a parallel interpretation is abandoned. The M.D. may therefore be a good example of a story for which we must first determine the author's intentions in order to appreciate his work's artistic design and aesthetic appeal. At the same time, it demonstrates why we need to separate these intentions from the actual work, which, in some respects, may be inconsistent with them. The inclusion of the author's executive intentions in the interpretation of his work thus provides for a better and more complete interpretation of his work as a work of art.
The theses from the last two sections can be summarized as follows. Literary aesthetic interpretations must be consistent with the executive intentions of the authors who wrote the works in question. In other words, a significant part of interpreting literary works lies in first determining what their authors intended to accomplish in creating them. Again, one must not confuse the relevance of the author's intentions to artwork-interpretive criticism with interpretive determinacy. Work interpretations are open-ended, and no appeal to the author's intentions can ever close a literary work and establish its single best interpretation. On the other hand, there is nothing in this picture to validate unmitigated interpretive relativism. Just as there will always be countless interpretive propositions that are true for all and any accounts of the story, there will be countless propositions that are not true once and for all. It will always remain true that at the end of Ken Kesey's One Flew Over the Cuckoo's Nest Chief Bromden smothers to death the maverick Irishman Randall McMurphy. It will always be false that the Chief is an extraterrestrial raised like a cuckoo by an Indian family, even though nothing in the text directly contradicts it. Put differently, the existence of some interpretations between which no rational choice is possible does not entail that rational choice is not possible between any and all interpretations.
Intentions and interpretations. Gary Iseminger's anthology, Interpretation and Intention, usefully surveys and recasts the persistent divisions of opinion concerning the role of intentions in interpretation. The most influential one, naturally, is the decades-old dispute between the interpretive extremists: radical intentionalists and anti-intentionalists. Over the years these two positions have become almost synonymous with the names of E. D. Hirsh, Jr., and Monroe Beardsley, who, often with remarkable lucidity, staked out a variety of arguments in defense of their views.20 Although there is nothing in moderate intentionalism to resemble the extravagant claims made by the radicals of previous generations, it will be instructive to revisit their positions with a view to the issues at hand.
According to Hirsh, literary meaning surfaces only in the presence of consciousness. On this basis he plausibly contends that intentions directed at understanding the literary meaning of fiction are indispensable to interpretation. The question remains, however, whether the guiding intentions, and thus meaning, are the author's or the reader's. Although Hirsh does not debate this point, his tone suggests that he perceives these alternatives to be mutually exclusive. As a result of his denial of the contribution of the reader, he is led to assert that "the meaning of the text" is identical, or logically equivalent, to the author's meaning.21 Although it may not be immediately apparent, in this concise pronouncement Hirsh compounds no less than three problems, which will be separated below. The first is his identification of texts with works. The second is the compound problem of the (ontological) presence and (aesthetic) relevance of the author's meaning in interpretation. The third is whether there is such a thing as the best or the one correct interpretation of a literary work.
Let us begin with the last point. One reason why we should reject Hirsh's thesis that there is the best or the one correct interpretation of a literary work is that critical praxis across the ages is uniform in belying it. Readers and critics fight endless interpretive skirmishes over the meaning and significance of their favorite passages, be it Iago's last words to Othello, Huck's attitude to Jim, or others. Another reason is that the thesis is articulated in a theoretical limbo- no one has ever proven anything in this respect. All the same, a radical intentionalist could insist that it is not obvious why we must reject it. To show that interpretations do not always converge in all respects and that authorial intentions do not determine interpretations to an arbitrary degree of refinement, we require a methodological, rather than a historical, argument. Here is such an argument: texts are finite in how much they determine, whereas interpretations can be refined without end. Every story is compatible with an infinity of possible worlds, and if we identify each possible world with a different possible interpretation, the number of the latter must also be infinite, instead of one-the best one.22
For an illustration, consider Bernard Malamud's novel God's Grace.23 The last scene depicts the protagonist, a nuclear holocaust survivor, Calvin Cohn, dragged by a sentient chimpanzee up the mountain toward sacrificial death. Cohn's demise has been intimated to him by God, for whose support and forgiveness he had always hoped. In the final scene, defeated and apparently abandoned, Cohn-and with him the reader-is made to reflect on the nature of God's grace. Does the protagonist undergo an epiphany of reconciliation or an agony of abandonment? Is his sacrificial death a symbol of forgiveness, or an ironic verdict on our inveterate sinfulness and hubris? Is Malamud's God grace-ful and compassionate, or grace-less and cruel? There are no evidentiary grounds to believe that the author saw either of these interpretations, or their conjunction, as the correct one, and no appeal to the text can decide the issue.
A similar kind of interpretive indeterminacy appears at one point or another in almost all stories. Is Hemingway's To Have and Have Not an exaltation of individualism, immortalized in the spirit of the last hard-boiled buccaneer, Harry Morgan? Or a testimonial of Harry's failure and an indictment of the limits of rugged individualism in the increasingly bureaucratic and tame America? The same case can be made for Harry's famous last words: "No matter how a man alone ain't got no bloody fucking chance."24 As Wirt Williams writes: "Is this a primitive, intuitive revolutionary declaration? Or a postulation of a primitive, intuitive tragic sense of life? Or both?"25 There are no determinate answers to these questions, and for this reason it makes no sense to speak of the single best or correct interpretation of this novel.
To highlight the second issue-the presence and relevance of the author's meaning in interpretation-I turn to one of the staunchest advocates of anti-intentionalism. Beardsley mounts a serious challenge, echoed since in endless variations, to Hirsh's contention that the meaning of a literary work is identical with the author's. Consider these arguments: (1) there can be authorless texts that are meaningful and interpretable; (2) interpretations can alter after the author's death; (3) works can have meanings unintended by the author. Each shows the nonequivalence between the work's and the author's meaning.
Beardsley's anti-intentionalism is actually an amalgam of two closely knit argumentative strands, which can be labeled ontological and aesthetic. The ontological argument depends on what Beardsley maintains is a segregation of fictional discourse from real life. Acknowledging that in real life the communicative goal of interpretation demands the recognition of the speaker's intentions, he maintains that the literary cousin of real-life language does not behave in this manner. According to Beardsley, even though real people perform various illocutionary acts, such as asserting, interrogating, or satirizing, literary fictions specialize in only representing these acts. Hence, goes the argument, because language in fiction is divorced from ordinary language-on the strength of the fiat that it merely represents various illocutionary acts-critics can dispense with the author's intentions. But this proposed segregation of fiction from real life makes no sense. Few readers fail to recognize that Poe's satire, "How To Write a Blackwood Article," conveys something about the real-life magazine. The crusading tone of the piece as well as available contextual evidence make it clear that Poe speaks at least in part in his own voice, instead of merely representing the illocutionary act of satirizing. In fact, part of the accuracy and piquancy of his satire owes to the recognition that it is the real-life author, rather than his narrative alter ego, who makes these pronouncements.26
The essence of Beardsley's aesthetic argument can be captured by this question: Must aesthetic interpretations be dependent at least to some extent on the author's intentions? The critic himself shows why the answer is affirmative. In the same sentence which proposes that aesthetic attributes arise "out of the ingredients of the [text] itself,"27 Beardsley identifies originality (freshness, novelty) as a deep feature of the text. Much like Hirsh, Beardsley equates texts with works, with fatal consequences for his theses, for, as I have shown earlier, originality is a feature of works, not texts. If, as Beardsley agrees, literature can exhibit the aesthetic quality of being original, apprehending this quality is a matter of going from text to work. And inasmuch as works entail compatibility with the best evidence about the author's executive intentions, Beardsley's attribution of originality is compatible only with the intentionalist account.
The moderately intentionalist stance that some of the work's meaning may significantly depend on the author's executive intentions and the attendant implication that there exist true interpretations of story content have been the subject of a clever attack along Beardsley's lines. Susan Feagin contested it by arguing that, since fictional characters have no referents, we cannot attach truth values to statements about them.28 On her view this precludes speaking of true or untrue interpretations, thus implicitly defending free-for-all interpretive pluralism. This type of objection to moderate intentionalism, however, is easy to repeal by recasting statements about characters and events into statements about the work, perhaps in the following form: the work (W) represents a character (C) or event (E) as having an attribute (A). The disappearance of the reference problem voids the supporting argument.
Textual Games without Frontiers. Although they have been the focus of my attention, aesthetic interpretations (or, as I call them, work interpretations) do not exhaust the range of strategies open to literary critics. The plurality of scholarly aims underwrites a plurality of critical approaches, not all interested in determining what the author intended to do, and succeeded in doing, in creating her work. The evidence is plentiful, in fact, that some critical strategies depart dramatically from this purpose. Professing to dispense with intention attribution, they aim to stimulate critical debate by means of textual readings.29 Textualist theory and practice are, of course, not one and the same. As has been shown in countless examples, even the staunchest critical textualists inevitably use statements parasitic on the author's intentions or the relevant art-historical context.30 This fact alone raises questions about the methodological and pragmatic viability of any textualist theory of interpretation.
But what about actual textual practice? I think it is not unfair to compare textual readings to sampling in music. The sampler also takes a piece of someone else's intellectual and artistic creation and adopts it for his own purpose by amputating it from its original art-historical context. Lest we too hastily condemn such violations of a work's integrity, let us examine whether textual readings can support a legitimate form of inquiry and, more generally, what type of inquiry they could support. The answer to the latter question is provided by the reading strategies popular among contemporary critics. In no particular order those seem to be enhancing the reading value, presenting a personal or just an interesting way of reading, recasting the text in a new (ideological) context, illustrating the tenets of some school of criticism, transgressing canonical boundaries, or even, as Andrew Field seems to attempt, producing an artwork of one's own.31
Undoubtedly, textual readings can support a wide variety of critical aims and strategies and stimulate discussion as effectively as work-oriented interpretations. But, one is compelled to ask, why bother with texts of literary fictions when the critics' attention lies elsewhere? In an interpretive climate that disregards authors' intentions-in-a-context, one might as well generate meanings and readings ex nihilo. Why not dispense with literature altogether, given that textual readings cannot do aesthetic and interpretive justice to the works they purport to analyze? Presumably because to ignore literature would be to ignore a source of cognitively valuable and emotionally enriching models, scenarios, and hypotheses of unsurpassed subtlety, diversity, and complexity. But this ready-made cultural treasury is bequeathed to us by creators who intend to communicate, however imperfectly, with their audience. Indeed, many readers and critics of literature find it crucial to genuinely understand and explain the aesthetic beauty and intellectual complexity of the novels they read. In this light, I must agree with other writers on this subject that little is to be gained by pretending that textual readings, often bounded by only the loosest requirements of plausibility, intelligibility, or interest in the work in question, have much to do with aesthetic appreciation and interpretation of literature.32 If anything, critical textual readings should be owned as such, instead of promulgating a myth that they somehow contribute to the understanding and aesthetic appreciation of works of literature.
Here is a brief example of what I have in mind. In her canonical study, French critic Marie Bonaparte would have us believe she is interpreting Edgar Allan Poe's "The Murders in the Rue Morgue" by taking the following psychoanalytic line. Her starting fiat is that children about one year old are inevitably exposed to a scène initiale of lovemaking of their parental figures, and though they do not grasp it intellectually, this event leaves ineradicable traces in their psyche. Furthermore, because education represses the child's primal sexual instinct, it finds other venues through sublimation. According to Bonaparte, Poe, a precocious child, caught his parents in the act and reenacts his fantasies in "The Murders in the Rue Morgue" by means of the orangutan. For Bonaparte, the whole story hinges on the ape's decapitation of the victim-a symbol of "castration of a woman which is one of the central fantasies of little boys."33 We are told that nothing in Poe's tale is what it seems, since all the characters are only phantoms of little Edgar's real-life sadistic concept of coition (conception sadique du coït). In brief, Bonaparte reads the text of Poe's work as a psychoanalytic roman à clef. The sleuthing critic purports to expose the hidden truth of Poe's repressed subconscious at the expense of interpreting his story as a work of literature. In the light of the theses developed in this article it should be clear that such a reading of the text has nothing to do with the aesthetic and intellectual content of Poe's work.
In the end, anybody is free to do what they please in the academic game of who-is-the-cleverest-dude-around-this-text. In such an interpretive free-for-all, nothing stands in the way of a psychoanalytic deconstruction of Alice in Wonderland as Phallus in Wonderland. Such an off-the-wall reading, through a complex symbolic analysis of the text, might even stimulate critical debate in ways that I don't profess to know. No matter how intriguing or elaborate, however, a textual reading of this kind-one that ignores the work's art-historical context, itself circumscribed in part by authorial intentions-will not be a legitimate interpretation of the English mathematician's work. Nor will it be a viable research contribution, based as it is on an outdated and unsound body of work.
This last remarks brings us at last to the question of literary research.34 How does what we said so far bear on interpreting literature from a cognitive, rather than an aesthetic, standpoint? Guided by cognitive concerns, critics occasionally approach literary works in ways at odds with their authors' intentions. Does it mean that cognitive readings are also constrained only by a minimal dependency on the text? Not necessarily. Ethnographers or anthropologists who use literature to understand the society that produced it may do so without interpreting it in aesthetic terms. Yet at the same time they make inordinate efforts to situate literature in its proper sociohistorical context, in the process forming far-reaching hypotheses about the creative intentions of the authors.
Even in literary studies, where researchers refine and evaluate cognitive models derived from works of fiction, the author's executive intentions may fruitfully guide the reader's response. Departing from the author's modeling strategy is always one of the critic's options, but often at the price of departing from the range of interpretation that the entire artistry (skill, experience, research, etc.) of the author had gone into making. The author's modeling intentions function here as a salient equilibrium point, reflexively (in the Gricean sense) guiding the interpretive moves of the critic.35 The process is partly analogous to aesthetic interpretations, especially in cases when writers pursue cognitive (modeling) goals as part of their successfully realized artistic intentions. The essential difference is that work interpretations always rely on the best available evidence of author's executive intentions. The Modeling Heuristic, which assumes that the author's intentions can guide the reader to the cognitively most valuable interpretation, is an equivalent of a working hypothesis that can be overturned by contrary evidence. Still, even the decision to abandon the Modeling Heuristic demands a prior decision that the heuristic does not apply in a given case, which, in turn, presupposes its prior application. Simply put, to rule out the author's modeling intentions as irrelevant, one need first apprehend them.
Once again, the above proposals do not mean that the Modeling Heuristic will always yield the optimal cognitive reading of the work. In many instances, models that flow from the creator's intentions will turn out to be more illuminating than others. Yet, in pursuing epistemically useful readings, researchers are free to adopt any strategy that will maximize their cognitive gains, including those that depart from the hypotheses implied or even stated explicitly by the author. In other words, the payoff from a cognitive reading depends exclusively on its knowledge content rather than on its correlation with the modeling goals that (reflexively) guided the story's genesis. Cognitive readings-that is, the assignation of a text's explanatory power to a certain aspect of reality-are thus not entirely unconstrained. The merits of competing models and hypotheses can be established in terms of such standard features as explanatory power, economy, lucidity, comprehensiveness, universality, and the like.
It is not impossible to read Kesey's One Flew Over the Cuckoo's Nest as a sociological investigation of the success of lobotomy in exterminating violent tendencies in psychologically volatile types, such as the novel's protagonist. Yet such a reading, although not contradicted by the text, is a paltry account of its modeling potential. In contrast, an interpretation of the loony ward as a social microcosm of USA Inc. is more plausible, and preferable, since it implies and subsumes the other one. It offers a more coherent account of the psychoses, neuroses, and pathologies of the American social machine and, unlike the first, also identifies symbolically the causes of this state of affairs. Yet, even though this cognitive reading can be shown to be a better (more complex, accurate, etc.) model of a particular situation, we could not say that it forms the cognitive meaning of the work, even if only because models can be studied with an arbitrary degree of resolution, thus altering their focus and ultimately the models themselves.
Notes
1. In Peter Swirski, "Interpreting Art, Interpreting Literature," Orbis Litterarum 56.1 (2001): 17-36. In Intention and Interpretation (Philadelphia: Temple University Press, 1992), Gary Iseminger argues for a definition of an artwork that closely echoes my proposals in his Revised Identity Thesis: "a [typical] literary work is a textually embodied conceptual structure, whose conceptual component is [identical to] the structure-compatible with its text-which its author intended [meant] in composing it" (92-93).
Return to article
2. For examples of pertinent work in analytic aesthetics, see the works of Gregory Currie, Jerrold Levinson, Paisley Livingston, and Robert Stecker.
Return to article
3. W. Wimsatt and Monroe Beardsley, "Intentional Fallacy," Sewanee Review 45 (1946): 469-88; Northrop Frye, The Anatomy of Criticism (Princeton: Princeton University Press, 1957).
Return to article
4. Nelson Goodman, Languages of Art: An Approach to a Theory of Symbols (Indianapolis: Bobbs-Merrill, 1968) 209.
Return to article
5. Nelson Goodman and Catherine Elgin, Reconceptions in Philosophy (Indianapolis: Hackett, 1988) 56.
Return to article
6. See Swirski, "Genres in Action: The Pragmatics of Literary Interpretation," Orbis Litterarum 52 (1997): 141-56.
Return to article
7. Kendall Walton, "Categories of Art," Philosophical Review 66 (1970): 347.
Return to article
8. Frank Sibley, "Aesthetic Concepts," Philosophical Review 68 (1959): 421.
Return to article
9. David Daiches, "Literary Evaluation," Problems of Literary Evaluation, Yearbook of Comparative Criticism, vol. 2, ed. L. P. Strelka (University Park: Pennsylvania University Press, 1969) 177.
Return to article
10. The very concept of category requires extratextual knowledge of genre in order to speak of its past, influence, popularity, historical development, extension, etc.
Return to article
11. Jerrold Levinson, "What a Musical Work Is," Journal of Philosophy 77 (1980): 5-28.
Return to article
12. This move is implicitly sanctioned by Levinson, who wants to extend his arguments to plays and novels (22).
Return to article
13. Lem states: "I consider the exercise of taxonomizing (classifications, genealogies) in literature to be a harmful brand of scholastic activity (if it is meant to tell us how to pigeonhole a work), because the most interesting issues happen to be located on the borders of classes, especially the ones we consider sterile (like mules!). That is, it may appear that certain literary species are not to be crossbred, but I benefitted greatly from such crossbreeding" (qtd. in Istvan Csicsery-Ronay, "Twenty-Two Answers and Two Postscripts: An Interview with Stanislaw Lem," trans. Marek Lugowski, Science-Fiction Studies 13 [1986]: 256).
Return to article
14. "The first version does not satisfy me completely, even though it is quite decently constructed and generates a good deal of suspense.... The Chain of Chance is better because it is plausible." In Stanislaw Beres, Rozmowy ze Stanislawem Lemem (Cracow: Wydawnictwo Literackie, 1987) 55. The first version of The Chain of Chance is, of course, The Investigation.
Return to article
15. Swirski, Between Literature and Science. Poe, Lem, and Explorations in Aesthetics, Cognitive Science, and Literary Knowledge (Montreal: McGill-Queen's University Press, 2000).
Return to article
16. See Swirski, "Interpreting Art," and Paisley Livingston, "From Text to Work," After Poststructuralism, ed. Nancy Easterlin and Barbara Riebling (Evanston, IL: Northwestern University Press, 1993) 91-104.
Return to article
17. Robert Stecker, "The Role of Intention and Convention in Interpreting Artworks," The Southern Journal of Philosophy 31 (1993): 485-86.
Return to article
18. Thomas M. Disch, The M.D.: A Horror Story (New York: Berkley, 1991) 234.
Return to article
19. According to Disch's prepublication statement, the novel's 466 pages were culled from well over 700 to reach publishable proportions-a plausible cause of these minor inconsistencies.
Return to article
20. Monroe C. Beardsley, The Possibility of Criticism (Detroit: Wayne State University Press, 1970); E. D. Hirsh, Jr., "In Defense of the Author," Intention and Interpretation, ed. Gary Iseminger (Philadelphia: Temple University Press, 1992) 11-23.
Return to article
21. Hirsh 14.
Return to article
22. In holding this view, I am in complete disagreement with absolutists such as P. D. Juhl, who claims that a "literary work has one and only one correct interpretation." P. D. Juhl, Interpretation (Princeton: Princeton University Press, 1980) 198.
Return to article
23. New York: Avon, 1983. See Swirski, "God's Grace," Encyclopedia of Popular Fiction (Osprey, FL: Beacham, 1998) vol. 11: 5592-601.
Return to article
24. Ernest Hemingway, To Have and Have Not (New York: Simon and Schuster, 1987) 225.
Return to article
25. Wirt Williams, The Tragic Art of Ernest Hemingway (Baton Rouge: Louisiana State University Press, 1981) 114.
Return to article
26. Nöel Carroll demonstrates why, if there is "thesis projection of nonfictional import-whereby actual authors express their views about life, society, morality, and so forth-and a great deal of literary (indeed artistic) interpretation concerns the identification of such theses, then intentionalist criticism has a wide arena of legitimate activity." Carroll, "Art, Intention, and Conversation," Interpretation and Intention, ed. Gary Iseminger (Philadelphia: Temple University Press, 1992) 109. The truth of the starting premise casts further doubt on Beardsley's first argument.
Return to article
27. Beardsley 34.
Return to article
28. Susan Feagin, "Incompatible Interpretations of Art," Philosophy and Literature 6 (1982): 133-46.
Return to article
29. I find an unlikely ally in Roland Barthes, who, in "De l'oeuvre au texte," also separates (although for completely different reasons) interpretations of works from readings of texts. The distinction I propose corresponds in some respects to what Stecker refers to as content-attribution and significance-finding ("The Role of Intention" 486-87).
Return to article
30. See Livingston, "From Text to Work."
Return to article
31. Andrew Field, Nabokov-His Life in Art: a Critical Narrative (Boston: Little, Brown, 1967).
Return to article
32. See, for example, Levinson's "Intention and Interpretation: A Last Look," Interpretation and Intention, ed. Gary Iseminger (Philadelphia: Temple University Press, 1992) 221-56.
Return to article
33. Marie Bonaparte, The Life and Works of Edgar Allan Poe, a Psycho-Analytic Interpretation, trans. John Rodker (London: Hogarth, 1949) 540.
Return to article
34. See Swirski, "Literature and Literary Knowledge," Journal of the Midwest Language Association 31 (1998): 6-23.
Return to article
35. See Swirski, Between Literature and Science, chapter 1.
Return to article