A New Apology For Literature
Francis Steen, Department of English, UC Santa Barbara
Not for distribution -- comments welcome

A Scandalous Practice

Contemporary literary studies are focused on the task of coming up with fresh interpretations. [1] What could the text mean -- to me, to some subset of readers, to the author? In what multiple contexts could it be placed? What agendas did it serve at the time of composition or of publication -- and which ones could it serve now? We are free to apply formal criteria, treat the text as an aesthetic object, or as a piece of devious propaganda. This search for new readings is a somewhat paradoxical enterprise, since it is made possible precisely by the assumption that literature does not have a determinate meaning. At the same time, the force of a particular interpretation hinges on the critic's being able to persuade the reader that it is legitimate, by one criterion or another. I'm not suggesting is not suggesting that a work of literature is so ill-defined that it means nothing in particular, only that the kinds of meanings that people find interesting, that readers and critics abstract from the works, don't tend to converge towards some single, unified, consensual, well-defined interpretation. The very prospect of such a convergence, of a literary work having a well-specified and determinate meaning, even if it is attractive, is inherently absurd. One might imagine that a reliable method to "decode" literary works were discovered or invented, and it turned out, say, that Don Quixote means ---- yet the sentence can scarcely be completed, even in the imagination. Who would this meaning be for, in what context would it be made, and who can keep me from making up another? There is simply no conceivable exclusive meaning we could assign to Don Quixote without lapsing into just the species of monomania that the main character so engagingly succumbs to.

The problem is not, as some critics have suggested, that a literary text lacks referentiality, that it is not about the real world. Literary meanings incorporate in complex if indirect ways what is real -- real psychology, real people, real economics, real places; in any event, reality can scarcely claim to have only one meaning. The impossibility of conceiving of a single literary meaning can more plausibly be attributed to the multiplicity of our purposes, what we want from a literary work, what we use it for. We see this brought into relief in the paradox we encounter when we try to come to terms with cognitive models of interpretation. While cognitive linguistics and the cognitive sciences more generally are focused on the descriptive act of determining how it is that a reader or a listener in theory and in practice is able to construe a given text or utterance in a particular way, the literary critic approaches a text looking for new meanings, assembling interpretations and perspectives that have never existed in another mind before, that in a radical sense is not empirical. At the same time, in seeking to justify any reading, the critic must nevertheless appeal to the text, to some general if novel principles of interpretation, to psychological and historical processes that in principle lend themselves to empirical research. Constraints can be invoked to delimit the meaning: the interpretation, of the letters is typically stable and reasonably uncontroversial, and you can look up words in a dictionary. Yet there is no principled limit to the kinds of constraints that are invoked or violated: nothing to stop an enterprising critic, for instance, from insisting that the letter e must be subverted and read as an a, or that "cloud" really means "mother." The relation of words to meanings, as Wittgenstein pointed out, is regulated not by an objective correspondence but by various language games; in literary studies, there is a premium on coming up with new games, or sustaining games that explores certain questions in interesting ways, even if these games in other contexts appear grossly incoherent. The overall result is that meaning multiplies beyond any hope of control, not merely by accident, but by conscious effort and design.

The deliberate proliferation of literary interpretations raises a delicate issue: how do we explain to ourselves, our colleagues in other disciplines, and all those whose support we in various ways depend on or attempt to enlist, what it is we are doing? Without a theory that vindicates our seemingly perverse practices, there is the danger of a spreading professional embarrassment (Liu, 1984), even if we're having fun. A tempting solution -- the temptation of postmodernism -- is to take the career of meaning to an extreme, as though it were inherently and in all contexts never possible to nail down anything. Yet this is clearly not the case: in many disciplines as well as in everyday situations, people are perfectly capable of conveying meanings with exactly the degree of precision they desire. Language games can be strictly defined and yield perfectly unambiguous interpretations. The attempt to generalize the approach to meaning that is adopted in literary studies has if anything served to hold us up to ridicule and pitifully discredit our worthy pursuits. It puts us in the position of inverted Don Quixotes, haughtily riding through town insisting that everything is a fiction.

Perhaps it is time for a new apology for poetry, that is to say, for literature in general. We need to find a way to talk about what we do that doesn't strike the inhabitants of the villages of La Mancha as fantastic and laughable. It is tempting to conclude they just don't understand us, but in such a general matter as literature it is surely better to assume, as Hobbes suggests, that common sense is very equitably distributed, since nobody complains of having less of it than anyone else (Leviathan). What we need to mount a successful defense, however, is nothing less than a thoroughgoing revision of a some of the cherished basic tenets of contemporary literary theory. Let us begin with the perennial notion that meaning inheres in a system of signs.

Two Modes of Interpretation

In "Structure, Sign and Play in the Discourse of the Human Sciences," Jacques Derrida acts as midwife to the birth of the then "as yet unnamable" and now familiar face of postmodernism. It can only proclaim itself, Derrida ominously writes, "under the species of the nonspecies, in the the formless, mute, infant, and terrifying form of monstrosity" (1126). The target of his attack is structuralism, and all of Western philosophy propped up behind it, dedicated to the delusion of a fixed center in the fabric of thought, an eidos, "consciousness, God, man, and so forth" (1118). This construction, which in reality exists only within a system of signs, is assigned being and presence, underwriting the soundness of the structure of knowledge and restricting the free play of signification. The welcome if slightly scary message is that meaning is now finally loosed from its moorings. Faced with the intriguing insight that "language bears within itself the necessity of its own critique" (1120), we are left with a ceaseless and wide-eyed task of reinterpretation. The point is that we get from Derrida's contentions, even if they seem slightly outlandish, is precisely what we want: the legitimated freedom to interpret. On the other hand, I will argue the price we are paying for it is too high.

"Structuralism," Machin and Norris note, "offered criticism its last chance to make a science out of theorizing literature" (1987: 1); poststructuralism was above all a rejection of science. Derrida's monster, even if largely ignored by other disciplines, soon became the pet of literary scholars. Armed with its deconstructive fangs, literary studies shied away from what smacked of empiricism or rationalism alike, turning everything into a "text," adding quotation marks, hedging, erasing, having a ball. Meanwhile, linguistics, anthropology, and philosophy of mind have increasingly been drawn into cooperation with biology, psychology, neuroscience, and artificial intelligence in a loose convergence of the cognitive sciences. Literary critics have given and still give scant attention to the revolution taking place in our own front yard, the subject of what it means to be human -- in Pope's words, "the proper study of mankind." What is it exactly that keeps us from participating? It is, I suggest, the conviction that what we do is radically different and probably incompatible with what scientists of any color are doing. [2] The interesting question is, in what sense radically different?

In Allegories of Reading, Paul de Man makes a distinction between "cognitive and performative language," associating the former with structuralism, with its emphasis on "grammars" and "codes" (ix) and the latter with rhetoric, or figurative language. "The existence of grammatical structures, within and beyond the unit of the sentence, in literary texts is undeniable, and their description and classification is indispensable," he writes. "The question remains if and how figures of rhetoric can be included in such a taxonomy" (7). His own reply is they cannot: "Rhetoric radically suspends logic and opens up vertiginous possibilities of referential aberration.... I would not hesitate to equate the rhetorical, figural potentiality of language with literature itself" (10). In support of this he justifiably invokes "a great number of antecedents," citing, unfortunately, only Monroe Beardsley's claim that literary language is characterized by being "distinctly above the norm in ratio of implicit to explicit meaning" (Beardsley, 1973: 37). Implicit meaning need not be figural; more damagingly, one of the most solid results in cognitive linguistics is that all language, ordinary, pedestrian, as well as specialized and technical, is massively metaphorical. Figures of rhetoric are even being slotted into reasonably orderly taxonomies (Lakoff, Lakoff & Turner, Sweetser). [3] Even if de Man is identifying a key phenomenon -- namely the necessary failure of structuralism, conceived as a search for the hidden codes of interpretation -- he is mistaken about is source; the notion that metaphor makes literature possible by destabilizing meaning is a gigantic red herring.

Yet the monster must somehow be fed. Derrida tosses it "man and humanism," similarly invoking "two interpretations of interpretation":

The one seeks to decipher, dreams of deciphering a truth or an origin which escapes play and the order of the sign, and which lives the necessity of interpretation as an exile. The other, which is no longer turned toward the origin, affirms play and tries to pass beyond man and humanism, the name of man being the name of that being who, throughout the history of metaphysics or of ontotheology -- in other words, throughout his entire history -- has dreamed of full presence, the reassuring foundation, the origin and the end of play. (1125)

Yet humanism is hardly a history of deciphering a truth which escapes play; one of its founding documents is The Praise of Folly by the arch-humanist Erasmus. Derrida's grim assessment of mankind's "entire history" is further undermined by the fact that it is itself based on one of the really endemic fallacies of Western thought, the notion that meanings are disembodied.

The Rise of Cognitivism

A crucial challenge facing complex cognitive systems is that of accurately tracking the different flows and types of information -- what comes from the senses and what comes from memory, what did I think and what did someone else tell me, what did I dream and what did I experience, what did I think I saw and what can I infer must actually have happened. These processes are known by the collective name of source monitoring and are receiving increased attention by cognitive scientists. "Constructive and reconstructive processes that interpret, embellish, transform, and synthesize experiences," Mitchell and Johnson write, "are powerful engines for comprehension and creativity, but the potential cost is distorted memories and beliefs" ("Source Monitoring," 179).

In relationship with others, the task becomes exponentially more difficult: what did she say and what did she mean, what did she want me to think and what was she actually thinking, what does he think she wants and what does she really want, what did he say and what did someone else claim he said. Some of the terminological and conceptual difficulties and confusions that exist in linguistics stem from the high degree of complexity of the kinds of informational structures that underwrite day-to-day as well as specialized linguistic communication. "I shall try to avoid confusions," Saussure writes, "which could be serious ones" (Saussure's Third Course, 75a). It is not a trivial task to produce an accurate account of what is going on even at the rather basic level of source attribution, leaving aside the question of how the different kinds of information are being processed. In the following, I approach the disagreements surrounding the nature of meaning-making from this rather limited perspective.

Saussure defines the linguistic sign as composed of an acoustic and a conceptual aspect, which he labels signifier and signified. This terminology is used to indicate the source: the acoustic component of a sign -- say, the sound sequence "bull" -- originates in the physical event of enunciation, which creates a set of sound waves that are detected by the ear; the conceptual component -- say, the visual image of a bull -- originates in the perception of some object. In the linguistic sign, however, there is no physical event. The sound is only remembered, as when one reads silently to oneself. Saussure calls the signifier a "sound image" and thinks of it as a representation that has the power of "speaking to the imagination" (Saussure's Third Course, 76a). As Locke [4] and Aristotle [5] before him had noted, the name of something and the thing itself are not inherently linked; their association is conventional and in some regards "radically arbitrary" (ibid., 93a). For Saussure, however, what is named is not the actual thing but rather the concept of it; the composite linguistic sign is wholly mental. This initial analysis is already quite problematic. It is not at all clear, for instance, what a concept is, how it is produced, and how it relates to the world; nor does it seem self-evident that the way to think about speech sounds is to imagine reading to oneself. On the other hand, Saussure does not deviate very far from common sense; his theoretical leaps appear to stem from a desire to establish linguistics as a distinct discipline.

When Saussure describes the conceptual component of the linguistic sign, for instance, he argues it is intimately tied to the acoustic image; the two are like the sides of a sheet of paper (General Course, 113). He is at pains to explain, however, that concepts exist independently of words; their existence in that case, however, is the subject of "pure psychology" and not a topic for linguistics. Similarly the study of the actual physical qualities of the sound belongs to "pure phonology" (ibid., 113). His insistence that language -- the abstract system, langue -- is "a system of signs" (Saussure's Third Course, 92a) is picked up by Benveniste. He argues, incongruously, that the association between words and concepts is precisely not arbitrary but necessary: when I produce the sound image "bull" I cannot help recalling the concept bull. What is arbitrary is the link to the actual bull, "that one certain sign and no other is applied to a certain element of reality, and not to any other" (Patterns, 46). Clearly, this is very confused; the link between word and concept is commonly fast and automatic, yet this does not amount to a necessity. What Benveniste is attempting to establish, however, is that language is thought; there are no concepts outside of language. Linguistics, in turn, is "exclusively a science of forms" (44).

It is here cognitivism and postmodernism meet up and part ways. "The Cognitive science that started fifty years or so ago," Fodor writes, "more or less explicitly had as its defining project to examine the theory -- largely owing to Turing -- that cognitive mental processes are operations defined on syntactically structured mental representations that are much like sentences" (The Mind Doesn't Work That Way, 1). Once language became synonymous with thought, mental processes could be conceived as interactions of signs, a view Derrida inherited from the structuralists. Postmodernism broke the promised collaboration between structuralism and cognitivism, declaring that if meaning emerges in the operation of "difference" between signs, it must always be "deferred." Even as he and his heirs rejected structuralism, however, they retained the underlying cartesian dualism that still characterizes cognitivism (and more broadly, analytic philosophy): thought is language and language is a system of signs -- a 'text' -- that "bears no determinate relation to the world" (Holt, "New Model High School;" for a detailed analysis, see Lakoff and Johnson, Philosophy in the Flesh, 248-66 and 440-68). Where cognitivism adheres to a "methodological solipsism" (Fodor, 1982), Derrida proclaims "there is nothing outside of the text."

If there is an epochal turn in the history of Western Philosophy of the kind Derrida invokes, a far better candidate than the supposed disappearance of the center is the slowly growing realization that meanings are not disembodied signs and thought is not a process of symbolic operations. This realization is particularly important for literature, which lends itself especially poorly to a purely semiotic analysis.

Reclaiming Literature

Since we are readers and critics of literature, we might want to cast around for an understanding of signification, meaning, and imagination that accords more closely with these experiences. Consider perception: when you observe a landscape, say, or a friend's face, there is a distinction between what is outside and what is inside the mind -- a distinction worth tracking. Yet this distinction is itself a construction, not an actuality. "To perceive the world is to coperceive oneself," writes the ecological psychologist J.J. Gibson. "The supposedly separate realms of the subjective and the objective are actually only poles of attention" (Ecological Approach to Visual Perception, 116). The mind-world unity is uniquely ours -- an Umwelt or life-world of our specific kind, a human world of colors, textures, artifacts, faces. The work of the senses is presented to us in consciousness, at that point already mixed with memory and desire.

When Don Quixote spots the windmills on the plain, we as readers must recall the appropriate presentations from our personal and idiosyncratic memories and re-present them: we see, in the global workspace of the mind (Baars, 1997), the line of mills, their canvas sails thrust out. At the same time, we reconstruct the scene from the perspective of the protagonists: Don Quixote, looking for an enemy, exclaims that "fortune is arranging matters for us better than we could have shaped our desires ourselves, for look there, friend Sancho Panza, where thirty or more monstrous giants present themselves." How do they present themselves? We don't even need to imagine giants: it is enough to realize that Don Quixote's desires, as he notes himself, are bent on heroic deeds. The windmills present what Gibson called the affordances of monsters; in the absence of the real thing, windmills will do fine; in fact -- truth to tell -- they do rather better than the real thing. In the mind's quick and unadumbrated simulation, the function of the giant -- large, powerful, and intent on killing -- is recruited by the perception of the inocuous mills. Concealing the delight is a vague intuition that Don Quixote is doing the right thing: it is heroic to attack the windmills; Sancho Panza is just "not used to this business of adventures." This concession does not diminish the absurdity of the event: Sancho is right, the low-level perceptual processes are not compromized by belief, and even Don Quixote must admit a sage "has turned these giants into mills."

To model literature, then, we need a notion of perception as creative, yet in a way that continuously negotiates reality, belief, and desire. Our minds are anchored in a creative conjunction of the senses and the world, a functional unity honed through millions of years of organic evolution. Perhaps surprisingly, we see this most clearly in cases where the conjunction is inappropriate, in visual illusions. Even if we know, say, that the two lines are parallel, or of equal length, our perceptual systems doggedly persist in presenting them to us as angled and unequal. Pylyshyn argues convincingly that this persistence in the face of our better knowledge argues -- pace Bruner's highly influential thesis -- that our intentions and beliefs do not radically determine what we see in the first instance (Bruner, "Social Value and Need;" Pylyshyn, "Is Vision Continuous With Cognition"). We have something of Sancho Panza in us: at a basic perceptual level a mill is a mill, whatever you might want it to be and however you might want to treat it. What Pylyshyn calls the "cognitive impenetrability" of perception rescues us from the cartesian chasm: our senses, at once physical and mental, at a basic level underwrite the unity of mind and world. Yet this poses no limits to our creative imagination: Don Quixote's monsters are just as real as they need to be, and can be spotted as such if one looks for them, even if in the end they look like windmills.

Consciousness, following Baars, functions as a global workspace, allowing sensory presentations to be accessed by conceptual processes (Theater of Consciousness). These presentations can be and are parsed -- into objects, events, concepts at multiple levels of analysis -- and selectively stored in memory. They can be recalled, voluntarily or involuntarily, and re-presented in consciousness -- a view echoing Locke, who argued that concepts -- he calls them ideas or conceptions -- spring from experience (Essay, 59). Conceptual representations are thus intimately and non-arbitrarily linked to sensory presentations, though tracking the details of these relations is no a trivial task. Presentations, in turn, are systematically linked to the real world through the intricate functional design of the senses. In Rosch's view, concepts do not "stand for" things in the world in a referential fashion; instead, they reestablish the unity of inner and outer, functioning as a "bridge" between the subjective and the objective poles of existence:

Since the subjective and objective aspects of concepts and categories arise together as different poles of the same act of cognition and are part of the same informational field, they are already joined at their inception. They do not need to be further joined by a representational theory of mind.... Concepts and categories do not represent the world in the mind; they are a participating part of the mind-world whole of which the sense of mind (of having a mind that is seeing or thinking) is one pole, and the objects of mind (such as visible objects, sounds, thoughts, emotions, and so on) are the other pole. Concepts -- red, chair, afraid, yummy, armadillo, and all the rest -- inextricably bind, in many different functioning ways, that sense of being or having a mind to the sense of the objects of mind. ("Reclaiming concepts," 73)

What, finally, does it mean to say meanings are not disembodied signs? Signs can refer to representations -- for instance, a letter can refer to a sound, or a concept -- because in a particular language game, we can define this to be the case. A sign can be defined to refer to any mental or physical activity that is under voluntrary control -- "x" can mean "when I present the symbol x, run to the bus." Or it can mean, "when I present the symbol x, imagine a red apple." Running to the bus is not a sign, the red apple of the imagination is not a sign; they are both kinds of activities. These activities cannot be strictly segregated into "material" and "mental;" rather, they have a material or somatic pole and a mental or semiotic pole (cf. Bohm, "A New Theory"). Saussure might recognize this a "psychologizing" of language: the sign is again only the signifier, conventional and arbitrary; the signified, however, is precisely the psychological and material reality he wanted to keep out of linguistics.

So why does this matter? The embodied nature of concepts matters for two reasons. First, it matters for literary theory. If concepts are embodied, we can meaningfully ask the question why literature alone should have "sufficient power to melt into our affections, to incorporate itself with the blood & vital juices of our minds" (Wordsworth, Essay on Morals). A theory of literature should minimally be compatible with the subjective phenomenology of reading; the notion that language is a system of signs misses the mark by a wide margin.

Second, it matters for cognitivism. If concepts are embodied, then there is no reason to think that the mind is like a computer. Computers perform syntactic operations on symbols, but -- with the possible exception of certain special cases -- the mind does not perform syntactic operations on concepts. To the extent that thinking involves concepts -- that is to say, meaning -- computers do not think, though they may model various aspects of thinking.

The Scope of Fiction

After discrediting classical cognitivism along with postmodernism, it might seem that I'm suggesting all we need to do to set literature on a productive track is to ally with what is known as embodied or active cognitive science -- a tradition that reaches backwards to Merleau-Ponty, Dewey, and Wittgenstein and includes George Lakoff, Mark Johnson, Eleanor Rosch, Francisco Varela, Eve Sweetser, Mark Turner, and Gilles Fauconnier. Yet even if the explicit argument of the postmodern revolt fizzles when its own source in a central fallacy of Western Philosophy is uncovered, the basic motivation and grounds for the revolt remain, paradoxically, valid. The challenge of postmoderism, I contend, cannot be said to be fully met by second-wave cognitive science. The two modes of interpretation persist: while the focus of a scientific investigation, including those motivated by the cognitive sciences, is to discover the facts of the matter, the focus of literature and literary theory simply is not. Facts matter, but they are subordinated and act as raw materials to the play of the imagination in the scandalous generation of new readings. Derrida's monstrosity may be no more real than Don Quixote's giants, but it is no less necessary. Literary studies is not a science, and for good reasons. But what are these good reasons?

Let us admit that literature is a paradoxical endeavor, seemingly wasteful, pointless, illogical, at worst irrational -- indeed, Don Quixote is the embodiment of the madness of reading, and on what basis, with what mandate, do we label him a poor reader? His comprehension is beyond reproof. What do we think we're doing when we silently read, or in that antecedent metaphor of reading, gather around in friendly company, "Telling of Tales of pleasure & of witt," as Cavendish puts it in the frontispiece to Assaulted and Pursued Chastity? Consider some paradoxical instances from early novels and tell me what is going on. What is Aphra Behn doing when she presents Lord Grey, the champion of a Whig cause and rebellion she detests, the seducer of his a prominent Tory's daughter, as the epitome of lovers? "At least for my own satisfaction," she writes to Condon, a loyal Tory and the dedicatee of her scandalous Love-Letters, "and that I may believe Silvia truly happy, give me leave to fancie him such a person as yourself." Silvia truly happy? What truth can rest in the imaginary love bestowed on a fictionalized character? What satisfaction can Behn derive from pretending Grey is like Condon, even as she admits he is not, fancying a Philander, a figment of her imagination, that is neither one nor the other? Why does she -- and we along with her -- engage in these absurd behaviors? Consider Margaret Cavendish, complaing that "though I cannot be Henry the Fifth, or Charles the Second, yet I endeavour to be Margaret the First; and although I have neither power, time nor occasion to conquer the world as Alexander and Cæsar did; yet rather then not to be Mistress of one, since Fortune and the Fates would give me none, I have made a World of my own" -- is not this madness? (The Description of a New World, 1). What is the purpose of being in charge of nothing? For surely we cannot be expected to believe her blatantly fictive world exists for real.

These writers are rightly apologetic: they are clearly recreational liars. Yet they accomplish nothing, other than their own and our pleasure. The information they convey is either wholly erroneous or -- which is more pernicious -- true in shreds and patches, as in the case of Behn's incestuous story, yet so seamlessly conjoined with lies that the poor reader can trust nothing. Defoe's pretended autobiography Moll Flanders, some say, contains useful information about London low-life around the turn of the eighteenth century, yet it must not be overlooked that it is false in each and every concrete particular. In his famous apology, Sir Philip Sidney proposes that literature does not lie because it does not affirm anything, yet this is difficult to reconcile with the facts. Defoe, for example, manifestly claims that Moll "was Twelve Year a Whore, five times a Wife" (Moll Flanders, 1); Cavendish even less plausibly claims that her heroine put an end to "inhumane Sacrifices" in a city ringed by crystal walls (Assulted and Pursued, 241); Lennox claims Arabella believed one of her father's menial servants was "a Person of sublime Quality" (The Female Quixote, 31), reasoning from romance novels.

I must apologize for a momentary madness: all this is clearly going too far. If authors make claims, these claims hold only under very constrained circumstances, and readers normally have no problems tracking what these are. The scope of truth in fiction is in the first place the fictive world itself, bracketed by pretense (see Hamburger, Ohmann, Searle, Cohn, and Genette on the logical status of fiction). Within the simulation that the reader constitutes with the author's help, Moll was indeed five times a wife; no parallel claim is necessarily put forward that Moll ever existed in the real world. Arabella's mistake is precisely to misjudge the scope of fiction: what holds in a romance novel cannot be assumed to hold in reality -- that is to say, within the fictive world in which she lives. Constrained within its proper scope, fiction can safely make claims without lying.

The notion that the claims of literature can coherently be taken to be valid within the simulated world that the reader produces on the basis of the text may appear to resolve the central paradox of literature. Yet such a tidy demarkation is in turn delusive. Even if literary worlds in certain basic ways are decoupled from reality, there are multiple backstrung connections. If literature really had nothing to do with the real world, we might not find it interesting: relevance and timeliness are well-known motivators of reading. It is also a fact that readers routinely get affected by what they read, in ways that cannot necessarily be sharply distinguished from the effect of reading on Don Quixote and Arabella. What we need is a new set of conceptual schemas for modeling the relationship between the literary and the real. It is here, I suggest, that it becomes useful to invoke the paradigm of play -- in its simplest form, the spontaneous and cooperative play of children.

Literature as Play

In the summer of 1999, I participated in a cross-disciplinary conference in Santa Barbara, Imagination and the Adapted Mind. [6] Linguists, cognitive scientists, philosophers, psychologists, and anthropologists met with literary scholars in an attempt to break some common ground. For several years, I had been trying to understand the literary within the context of evolutionary and cognitive theories, working with the evolutionary psychologists Leda Cosmides and John Tooby, the neuroscientist And Turken, the cognitive linguists George Lakoff and Eve Sweetser, and the literary theorists Paul Hernadi and Mark Turner. I had begun to develop the notion that literature involves rich mental simulations and that these simulations don't serve any immediate purposes; we engage in them because they are experienced as rewarding in themselves.

The sheer impracticality of literature suggested that a different mode of mental function had to be proposed for the humanities. Whatever cogntive scientists came up with, it was clear that these could not in principle satisfy literary critics: critics were not looking for a single interpretation, or an understanding of how readers reached one; the very purpose and motivation for their activity was new readings and applications within new contexts. Fiction reaches into territories we cannot otherwise inhabit; its purpose may be a certain kind of training. In my talk at the Santa Barbara conference, I hypothesized a learning mode, designed by evolution to solve certain kinds of problems. In the audience was an educational psychologist from Colorado, Stephanie Owens; we discussed possible kinds of evidence for this thesis. After the conference, we began a collaborative project on one of the simplest and most ubiquitous of children's games, that of chase play.

In his Letters on the Aesthetic Education of Man, Friedrich von Schiller invokes a power to "annul all constraint" and "set man free both physically and morally" (421): he calls it the "play-drive" (Spieltrieb). Its object is beauty, or living form; its activity is man's essence: "he is only fully a human being when he plays" (422). Huizinga's Homo Ludens brings the metaphysics of play down to the level of social history. To understand literature as play from a cognitive perspective is a rather different enterprise, one that is continuous with the attempt to situate thought and feeling as embodied processes. Play is one of those complex adaptations that seem transparent and simple to us, like vision or object manipulation; it is ubiquitous among mammals and -- as I return to in subsequent chapters -- is associated with a distinct set of cognitive capacities and motivations. In young children's chase games we witness a game that has likely been played for tens of millions of years (Steen & Owens, "Pretense as an Adaptation for Learning").

My favorite video segment from our research shows what the parents say was the child's first chase. Waddling along, the child turns her head and sees her father following behind her. In a sudden, magical moment she freezes and then smiles, fixing her eyes askance on her father, giggling excitedly. A world has opened; she has entered a cognitive mode where the normal rules just don't apply. Running at top speed, she stops to make sure he is pursuing; as if under great pressure, in a knot of excitement, she squeezes out a "Daddy... Get me!" He comes stalking after her, arms spread menacingly, fingers crouched, eyes wide -- and yet smiling, laughing; he knows well how to play his part. She flees but stumbles, her excitement nearly paralyzing her; he lifts her up and she runs on a few more steps. Finally, she tumbles and lays pinned to the ground, gasping. Her loving father gets down on all fours and opens his mouth, digging into her soft underbelly. She giggles uncontrollably.

There was no evidence of mental representations; the child did not appear to respond to an internal image recalled from memory. At most there was what Derrida calls "the species of the nonspecies, in the formless ... and terrifying form of monstrosity" (1126). The child's monster is the terrifying thing you run away from, that wants to eat you, giving you this very special thrill that the children signal in a piercing glissando at the top of their little voices. Derrida's final sentence evokes the same excited shiver of a non-descript terror that lies at the heart of play. Naturally, there is no real monster; there is only an absent center, a creative act of pretense, that motivates and organizes play.

The child is not interpreting her circumstances in any normal sense of that term; she is not making sense of what she sees in order to accomplish something. Instead, she is playing -- using Gibsonian affordances of her environment to trigger a pretend world in which everything takes on a different hue. Just as the sails of the windmills become in Don Quixote's eyes the arms of giants, so her father's eyes and arms become the features of the monster; the child reads creatively, in metaphors. The game defines a privileged domain where the child reader is in charge: any adult that oversteps the limits of the game is quickly put in his place.

The chaser, then, is in some ways like an author: he provides the cues that allow the child to create the world of the chase. It is not that the fleer is manipulated by the chaser: the relationship is cooperative and the chaser will adjust his behavior to fit the fleer's responses. The goal of the chaser, as of the author, is to please his audience, the fleer. This he does by providing the bare cues of a scenario the fleer constructs: a stalking gait, raised arms as if ready to grab, wide open eyes, and yet, a reassuring smile: this is just play. It is remarkable that the child, less than two years old, is able to construct the pretend space and to track it. Yet the source monitoring is not flawless. Even a loving father, if he sustains the role of wolf too long, may scare his child; she finds it confusing not to be reassured by an affirmation of the secure reality that frames and protects play. In this, too, we may see an antecedent of the persistent confusion at the edges of literature, the madness of fiction thematized in Don Quixote and its heirs.

In this view, rumors of the death of the author are somewhat exaggerated, yet they are not entirely groundless. The paradigm of play suggests the author may accept that what he creates is for the benefit of the reader and that it will be modulated and used to suit the reader's purposes. [7]. Yet at the same time, in the spirit of cooperation, a reader that successfully enlists an author and engages him in a collaborative game will have a much more sophisticated, elaborated, and satisfying experience than a reader who must (or choose to) read against the grain. The author-god is dead, but the collaborative author lives, lifting from the reader part of the burden of constructing and maintaining the fictive world, freeing her resources for a unique and personally appropriate emotional and imaginative elaboration.

Towards a New Anatomy of Criticism

The model, then, suggests that literature is a form of play -- or, to be more precise, a set of discursive practices that invite the interpretive mode of play. It is a cognitive account in the sense that it describes mental processes, but these are not syntactic operations on signs, they are embodied processes of presentation, simulation, learning, and action. The consequences for literary theory are rich and must be unpacked slowly. The following terms and concepts are the most salient constellations for guiding a cognitive vessel on the treacherous waters of fiction.

The Organizational Mode. Literary works allow and invite us to switch into a distinct cognitive mode, the organizational or learning mode. In contrast to our normal executive mode of functioning, which is designed to motivate and help us to control our environment, the learning mode is designed to intervene in the organization of our being in the world.

Safe Environment. The organizational mode requires a safe environment where no executive demands -- dangers, duties, desires -- are made on us. Our basic needs -- food, rest, warmth -- must have been met. The organizational mode relies on surplus resources of time and energy.

Pretense. The key innovation of the organization mode, pretense is a metaphorical mapping where the inferences generated are deliberately controlled to optimize the learning situation. This requires sophisticated but luckily unconscious capacities for keeping track -- see source monitoring.

Playful Enjoyment. The spontaneous enjoyment we feel in play requires -- indeed, tolerates -- no extrinsic motive: the organizational mode is motivationally independent of the executive mode.

Suspension of Identity. Under the right conditions, we voluntarily suspend our normal identity by setting aside or inhibiting the memories, intentions, and emotions that guide our actions in the world and entering the organizational mode. This frees up the cognitive resources -- working memory, possibly a distinct agent memory, emotional engagement, inference systems -- of our executive agency.

The Virtual Agent. The resources freed up by the suspension of the executive agent are utilized to construct one or more virtual agents, of differing degrees of elaboration, by adopting the history, capacities, and purposes of fictive characters. These agents consequently inherit the cognitive structure of our executive agent.

Simulation. Virtual agents act in simulated worlds, that is to say worlds that are constituted by the reader's personal memories and assembled into a simulated environment. The simulation -- in the form of visual images, sounds, activities, thoughts, emotions, and so on -- exists in consciousness, although many of the properties and entailments of the simulated entities will typically remain unconscious.

Imaginative Immersion. The virtual agent tracks the information it needs in order to act, such as what it is trying to accomplish, what it already has accomplished, what obstacles it is faced with, and what resources it can muster. In the simulated world, the construction of a virtual agent allows the reader to experience and respond to the situation as if it were real, a subjective experience we describe as imaginative immersion.

Source Monitoring. The organizational mode includes cognitive strategies, such as the construction of a virtual agent, that require multi-level tracking of the sources of experience in pretense, perception, rememberance, and imagination. These systems are fallible and can be manipulated.

Possibility Spaces. The virtual agent and the simulated world are created in order to explore some of the vast set of possibility spaces that are defined by the reader's capacities. In our executive lives, only an infinitesimal proportion of these possibility spaces can actually be realized and inhabited.

Strategies. A possibility space is explored by elaborating and practicing strategies of being -- attitudes, goals, behaviors, feelings, and thoughts -- in a range of different simulated situations. The goals that directs the strategies are typically derived from the executive mode: what we want as virtual agents is informed by what we want as real agents.

Structural Learning. The result of a simulation is structural learning, a largely unconscious process affecting what Siegler (1991) calls the "distribution of associations." Strategies practiced are associated with or assigned a scope (the conditions under which they may be utilized), a probability of success or failure (the likelihood that a properly performed strategy will lead to the desired outcome), and a confidence level (the degree of expertise the reader assigns to his own ability to execute a strategy).

Roles. As they are learned, sets of strategies are organized into roles, or roles may function as nodes around which strategies are organized. Once a role is learned, new strategies may be added to this role with little effort.

Realism. Unless the text specifies otherwise, the reader generally assumes, for the sake of coherence and simplicity, that the simulated world inherits or reproduces the causal and relational characteristics of the real world. At the same time, since the simulated world is motivated, designed to explore certain possibility spaces that are unavailable in the reader's real life, aspects of reality can be systematically altered in order to accomplish this task.

Collaborative Worldmaking. By providing the organizing elements of the simulation -- a memorable general schema of the settings and the actants of the plot -- the author reduces the working memory requirements of the simulated world, permitting the reader to construct, elaborate, and sustain more complex simulations. A particularly creative author will give the reader meaningful access to strategies or to entire possibility spaces unknown to or unexplored by the reader.

In sum, the organizational mode hypothesis suggests that literature keeps alive possibility spaces; it reminds us in an experiential if simulated manner of our potential as human beings -- even in situations where there is no hope of actually realizing much of this potential. It allows us to understand others as acting within similar possibility spaces but occupying very different locations within them; by making their actions comprehensible, they may turn our judgment towards kindness and compassion (Martha Nussbaum, Poetic Justice). Finally, literary works can give us new ideas for being in the world that can be actualized in our lives, in ways that we may be only dimly aware of ourselves.

The organizational-mode hypothesis has the further implication that the prototypical relation between author and reader is one of friendly cooperation: the act of imaginative immersion is undertaken in a cognitive mode that is oriented not towards control but towards learning. This is of course not in principle a surprise -- as readers, we are grateful to authors that transport us imaginatively to places we had not the ability to experience or even imagine on our own, and as writers we attempt to create worlds for our readers that are enjoyable and meaningful for them. However, much of contemporary literary theory treats the author as an adversary by default; the organization mode hypothesis suggests instead that the default condition is cooperative.

Nevertheless, the hypothesis also opens for the abuse of the organizational mode: the enjoyment we experience in fiction is seductive, and since structural learning is largely unconscious, literature can be utilized to convey an executive message about the real world under cover of play. In the following chapters, I examine the processes sketched out above in a series of concrete examples. I discuss not only the intricacies of cognitive operations but also the ways in which these operations can be and were enlisted for social and political purposes.

I should make it clear that I do not view the conceptual framework I have just presented -- and will continue to unfold -- as the truth about literature. Literature is not children's play, and the particular features of children's play we have observed in our research represent a small subset of the total (Owens & Steen, Steen & Owens). In addition, the relations between literature and play are themselves complex. The spontaneous play of children provide a cognitive antecedent, a basic set of skills and attitudes from which the complex cultural expressions of literature develop. At the same time, the particular forms that literature takes are culturally embedded and historically specific. The view that literature derives its basic cognitive architecture from play permits the humanities to talk to the cognitive sciences without losing their own voices.

Notes

[1] The term itself is problematic. In "Against Interpretation," Susan Sontag argues interpretations distract from the literary experience, in that they focus on what literature says rather than on what it does. Yet the target of her critique is a narrow, structuralist conception of interpretation:

Of course, I don't mean interpretation in the broadest sense, the sense in which Nietzsche (rightly) says, "There are no facts, only interpretations." By interpretation, I mean here a conscious act of the mind which illustrates a certain code, certain "rules" of interpretation. (546)

In place of hermeneutics, she writes, "we need an erotics of art" (550). Sontag's insistence on the primacy of the sensual and experiential in art has not lost its force. Even if she is right, however, the fact remains that literary scholars continue to generate new readings, and my present purposes are descriptive rather than prescriptive. Nor is it clear that she is right, once we give up her narrow definition of interpretation: if we view literature as experiential, as she suggests, it is not possible to draw a wedge between what it says and what it does; they are both aspects of interpretation.

[2] There are notable exceptions to the assumption that scientific approaches are incompatible with literary modes of thinking. N. Katherine Hayles' notion of the "post-human," for instance, allies postmodernism's rejection of the "metaphysics of presence" with the disappearance of a coherent, unified self or consciousness in cybernetic theories of cognition (How We Became Posthuman).

[3] See my review of Lakoff and Johnson's Philosophy in the Flesh, "Grasping Philosophy by the roots."

[4] "Names are but the arbitrary marks of conceptions..." Locke, Essay, quoting from his Second Letter to the Bishop of Worcester, p. 5.

[5] "The limitation 'by convention' was introduced because nothing is by nature a noun or name -- it is only so when it becomes a symbol." Aristotle, On Interpretation.

[6] The program of the conference is available at http://www.anth.ucsb.edu/projects/esm/IAM/.

[7] As Laurence Sterne writes in Tristam Shandy, "no author, who understands the just boundaries of decorum and good breeding, would presume to think all": "The truest respect which you can pay to the reader's understanding, is to halve this matter amicably, and leave him something to imagine, in his turn, as well as yourself..."

Works Cited

Aristotle. On Interpretation. Trans. E. M. Edghill. Available from Björn's Guide to Philosophy, <http://www.knuten.liu.se/~bjoch509/works/aristotle/interpretation.txt>, September 2000.

Baars, Bernard J. In the Theater of Consciousness: The Workspace of the Mind. New York: Oxford University Press, 1997.

Benveniste, Emile. Problems in General Linguistics. Trans. Mary Elizabeth Meek. Coral Gables, FL: University of Miami Press, 1971 [1966].

Bohm, David. "A New Theory of the Relationship of Mind and Matter." Philosophical Psychology 3 (1990): 271-286.

Bruner, Jerome. "Social Value and Need as Organizing Factors in Perception." American Psychologist 1 (1946): 241.

Cavendish, Margaret. Assaulted and Pursued Chastity, in Natures Pictures Drawn by Fancies Pencil to the Life. London: Martin and Allestrye, 1656. ESTC R35074.

----------. The Description of a New World, Called the Blazing World. London: Maxwell, 1666. ESTC R31744.

Culler, Jonathan. “Beyond Interpretation.” The Pursuit of Signs: Semiotics, Literature, Deconstruction. Cornell University Press, 1981. 3-17.

Defoe, Daniel. The Fortunes & Misfortunes of the Famous Moll Flanders, &c. Written from her own memorandums. London: Chetwood and Edling, 1722. ESTC T70314. Available from Gutenberg at <ftp://uiarchive.cso.uiuc.edu/pub/etext/gutenberg/etext95/mollf10.txt>

de Man, Paul. Allegories of Reading: Figural Language in Rousseau, Nietzsche, Rilke, and Proust. New Haven, CT: Yale University Press, 1979.

Erasmus, Desiderius. The Praise of Folly. Translated by John Wilson, 1668. Ann Arbor, MI: University of Michigan Press, 1958.

Fodor, Jerry. "Methodological solipsism considered as a research strategy in cognitive psychology." Husserl, Intentionality, and Cognitive Science. Ed. H.L. Dreyfus and H. Hall. Cambridge, MA: MIT Press, 1982.

-----. The Mind Doesn't Work That Way: The Scope and Limits of Computational Psychology. Cambridge, MA: MIT Press, 2000.

Gibson, James J. The Ecological Approach to Visual Perception. Boston, MA: Houghton-Mifflin, 1979.

Hayles, N. Katherine. How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics. Chicago, IL: University of Chicago Press, 1999.

Hobbes, Thomas. The Leviathan. London, 1651. Available from Oregon State University at <http://www.orst.edu/instruct/phl302/texts/hobbes/leviathan-contents.html> (1660 edition).

Huizinga, Johan. Homo Ludens: A Study of the Play Element in Culture. New York, Harper & Row, 1970.

Lakoff, George. Women, Fire, and Dangerous Things. Chicago: University of Chicago Press, 1987.

Lakoff, George, and Mark Johnson. Metaphors We Live By. Chicago: University of Chicago Press, 1980.

Lakoff, George and Mark Turner. More Than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press, 1989.

Lennox, Charlotte. The Female Quixote, or, the Adventures of Arabella. London: A. Millar, 1752.

Liu, Alan. 1984. "The Power of Formalism: The New Historicism." ELH 56 (1989): 721-71

Locke, John. Essay Concerning Human Understanding. 6th ed. London: Churchill and Manship, 1710. An edited version of the same edition is available online from ITL, Columbia University, <http://www.ilt.columbia.edu/academic/digitexts/locke/understanding/title.html>, September 2000.

Machin, Richard and Christopher Norris (eds.). "Introduction." Post-structuralist Readings of English Poetry. New York, NY: Cambridge University Press, 1987.

Mitchell, Karen J. and Marcia K. Johnson. "Source monitoring: Attributing mental experiences." The Oxford Handbook of Memory. Eds. Endel Tulving, I. M. Fergus Craik, et al. New York, NY: Oxford University Press, 2000. 179-195. Abstract.

Nussbaum, Martha Craven. Poetic Justice: The Literary Imagination and Public Life. Boston: Beacon Press, 1995.

Owens, Stephanie A., and Francis Steen. "Chase Play: The Structure of a Neglected Form of Physical Activity Play." Under submission.

Owens, Stephanie A., and Francis Steen. "Does Chase Play Train Predator Evasion?" In preparation.

Plumb, John Harold. The Growth of Political Stability in England 1675-1725. London: Macmillan, 1967.

Pylyshyn, Zenon W. "Is Vision Continuous with Cognition? The Case for Cognitive Impenetrability of Visual Perception." Behavioral and Brain Sciences 22. 3 (January 1999): 341-423. Available from the author's home page at <http://ruccs.rutgers.edu/faculty/ZPbbs98.html>, September 2000.

Rosch, Eleanor. "Reclaiming Concepts." Reclaiming Cognition: The Primacy of Action, Intention and Emotion. Ed. Rafael Nùñez and Walter J. Freeman. Journal of Consciousness Studies 6. 11-12 (1999): 61-77.

Saussure, Ferdinand de. Course in General Linguistics. Ed. Charles Bally and Albert Sechehaye. Trans. Wade Baskin. New York, NY: Philosophical Library, 1959.

Saussure, Ferdinand de. Saussure's Third Course of Lectures on General Linguistics (1910-1911): From the Notebooks of Emile Constantin. Ed. Eisuke Komatsu. Trans. Roy Harris. New York, NY: Pergamon Press, 1993.

Schiller, Friedrich von. Letters on the Aesthetic Education of Man. In Critical Theory Since Plato. Ed. Hazard Adams. Rev. ed. New York, NY: Harcourt Brace Jovanovich, 1992.

Siegler, Robert S. "Strategy choice and strategy discovery." Learning & Instruction 1. 1 (1991) : 89-102.

Sontag, Susan. "Against Interpretation." The Critical Tradition. Ed. David H. Richter. New York, NY: St. Martin's Press, 1989.

Steen, Francis F. "Grasping Philosophy by the Roots." Philosophy and Literature 24. 1 (2000): 197-203.

Steen, Francis, and Stephanie A. Owens. "Pretense as an Adaptation for Learning." Under submission.

Sweetser, Eve, and Gilles Fauconnier. "Cognitive Links and Domains: Basic Aspects of Mental Space Theory." Spaces, Worlds, and Grammar. Ed. Gilles Fauconnier and Eve Sweetser. Chicago, IL: University of Chicago Press, 1996.

Sweetser, Eve. "‘The Suburbs of Your Good Pleasure:’ Co-orientation of Metaphorical Mappings in Literary and Everyday Language." Zeitschrift für Semiotik, in press.

Turner, Mark. The Literary Mind. New York: Oxford University Press, 1996.