Criticism & Reference: Full Text

The cognitive sciences offer new paradigms for rethinking the relations among culture, linguistic activity, and agency. This essay provides an overview of the "cognitive revolution" and recent work on the brain, and then discusses the dominant strains within cognitive linguistics, arguing for wider applicability to literary theory and criticism.

1. Literary scholars have increasingly taken up interdisciplinary frameworks, methodologies, and pursuits in recent years, despite the difficulties of "being interdisciplinary" cogently outlined by Stanley Fish. Such broad-based critical schools or tendencies as the new historicism, gender studies, and cultural criticism by definition stake out their ground between or across disciplines, and some literature departments have remade themselves as "cultural studies" programs, suggesting that interdisciplinarity is becoming not simply a legitimate option for literary scholars but may be gaining the force of an imperative-one all the harder to resist as university presses attempt to maximize sales with academic titles that can be marketed across disciplinary lines.
2. It might seem initially surprising, then, that those challenging disciplinary boundaries in literary and cultural studies have shown so little interest in cognitive science, the major interdisciplinary initiative marking the convergence of linguistics, computer science, psychology, neuroscience, philosophy of mind, and anthropology. Our widespread lack of engagement with what has been called the "cognitive revolution" is all the more striking given its obvious bearing on theoretical discussions of language, subjectivity, and consciousness, not to mention its status as one of the most exciting and potentially far-reaching intellectual developments of the late 20th century.
3. We hope to provoke such engagement-both critical and collaborative-in the essay that follows. We begin by sketching out some key developments in the constitution of cognitive science as a major interdisciplinary venture, emphasizing the transition from pioneering attempts to describe cognition in terms of the logical processing of coded symbols to more recent efforts to ground cognitive activity in embodied experience. We then look at a series of related issues in cognitive science of special interest to scholars of literature and culture, including categorization theory, nonarbitrary aspects of language, metaphoricity, agency, and the material character of thought. Finally, we survey a number of attempts to date, on the part of cognitive researchers and theorists as well as literary scholars and critics, to forge links between literary studies and cognitive science.
4. The first phase of the cognitive revolution, based initially in linguistics and computer science, successfully dislodged behaviorism as the dominant paradigm within the social sciences. Behaviorism held that only directly observable behaviors could provide a valid basis for scientific study, and that any speculation about the cognitive processes behind those behaviors was insupportable and therefore to be avoided. Reaching beyond the social sciences, the influence of behaviorism was also apparent in the New Critical "intentionalist fallacy"-with its injunction to study only the text as a mode of complex behavior, leaving aside any concern about the author's thoughts or intentions-and in the "affective fallacy" which similarly bracketed off the mental processes of the reader. Another behaviorist axiom, that human behavior is determined by environmental or cultural forces without reference to specific mental functions or constraints, still inheres within several current approaches to literary and cultural studies, such as Marxist and some new historicist criticism.
5. In fields such as linguistics and psychology, however, behaviorism has for some time been largely discarded in favor of research into the cognitive processes of the brain. Noam Chomsky was the first outspoken critic of behaviorism, arguing in the 1950s that the human mind carried a linguistic capacity that was both innate and universal. Chomsky's work focused mainly on syntax, morphology, and phonology, and generally aimed at an almost mathematical precision in the analysis of natural language, leading his critics to charge him with having evaded semantics, and for failing to account for meaning. Yet although Chomsky's theories have remained controversial and have frequently changed over the years, the tenets of innate and universal human linguistic ability have found wide acceptance.
6. At the same time that Chomsky's work was helping to shift the attention of linguists toward the brain and toward the identification of distinct and highly constrained patterns produced by it, computer scientists were beginning to think about the ways in which digital computers might resemble the human brain, be made to simulate its workings, and, ultimately, become "intelligent." Alan Turing, who pioneered the "strong AI" position, argued that a digital computer functioned in ways that were essentially similar to the human brain, and that a computer would therefore eventually come to manifest an artificial intelligence indistinguishable from human intelligence. If a computer were able to hold a conversation with a person, and that person were unable to tell the computer's responses from those of a living person, the computer would pass the "Turing test" and could legitimately be called intelligent.
7. By the late 1950s, computer scientists were actively attempting to build machines that could "think" rationally in ways similar to human cognitive processes. For example, Allen Newell and Herbert Simon constructed Logic Theorist, a program that could prove theorems in logic, while others invented programs that could play chess, solve algebra problems, or manipulate shapes in an artificial universe. Such programs achieved limited success, but proved ultimately inadequate in ways reminiscent of Chomsky's early syntactic theories-they could manipulate symbols that were defined within a logical and finite system, but could not account for the messy complexities of the real world. As philosopher Hubert Dreyfus puts it, "the research program based on the assumption that human beings produce intelligence using facts and rules has reached a dead end, and there is no reason to think it could ever succeed" (ix). Even current AI researchers such as Terry Winograd and Fernando Flores believe that "in spite of a wide variety of ingenious techniques for making analysis and recognition more flexible, the scope of comprehension [of computers with reference to natural languages] remains severely limited" (11). Earlier, philosopher John Searle hit on this weakness in his famous "Chinese room" essay, which argues that machines can never truly understand the symbols that they manipulate with such speed and precision. As he saw it, whereas human beings are able to relate whatever symbols their brains manipulate with meaningful events in the "real world" that their bodies inhabit and constantly negotiate, computers will never be able to do so. Indeed, attempts to have computers perform more complex functions have tended to be least successful wherever "meaning" or knowledge of the world have been most critical.
8. Early cognitive theory thus eventually ran up against the limitations of its overly logical model of the mind and its inability to account for the complex ways in which human beings interact with the lived world. Some AI researchers, accordingly, began to develop computers based more closely on what was being learned about the architecture of the brain, employing more complex "neural nets" rather than simpler and more logical serial processing. Artificial neural networks, however, have as yet nothing like the massive and intricate connectedness of actual neurons, and in general, as helpful as computer models have proved in testing theories about basic cognitive functions, their ultimate usefulness has been essentially paradoxical: their attempts to prove that the human mind functions logically in ways that resemble a digital computer have failed, demonstrating, instead, that human cognitive functioning makes extensive use of "heuristics, strategies, biases, images, and other vague and approximate approaches" (Gardner 385). Concurrently, research in psychology, anthropology, and neuroscience has, in different ways, helped shape a view of the mind which addresses the shortfalls of Chomskian linguistics and strong AI theory.
9. Cognitive scientists now tend to posit a mind that, because of its biological and embodied features, is inextricably implicated with thought processes that seem emotional, metaphorical, and in many ways "illogical" in comparison to the serial logic and separation of cognition and emotion that characterized early cognitive theory. What first suggested alternatives to a strictly logical view of cognition was research-linking linguistics, psychology, and anthropology-that explored the cross-cultural uses of basic color terms. For example, working with the Dani in New Guinea, who possess only two color terms roughly equivalent to "light" and "dark," Eleanor Rosch found to her surprise that despite this limited verbal range, the Dani seemed to perceive colors in ways highly similar to speakers of languages with much more extensive color vocabularies. Rosch's work in the 1970s lent fuel to the cognitive attack on behaviorism, demonstrating that the perception of color represents another area of cognition where innate capabilities help shape linguistic behavior.
10. Rosch (and later investigators) further suggested that the division of colors into categories does not operate in accordance with classical theories of categorization. As John Taylor explains, classical logic posits categories that are clearly bound, equal in status, and composed of necessary and sufficient features that are discriminated along binary lines (2324). In terms of color, this would mean that something categorized as "red" would necessarily share all of the same features (in this case, a particular range of wave-lengths of light) with all other reds; a given color would either be "red" or "not-red;' the difference between "red" and "notred" would be clearly defined, and all reds would be equally red. Rosch found instead that the human cognitive system categorizes not by a list of features, but on the basis of a prototypical member (the reddest red). While the membership of that red in the category would be clear, other instances of red would be less clearly "red." This kind of "radial" category, anticipated by Wittgenstein's theory of "family resemblance;' has fuzzy boundaries, with some instances of reddish-blue or bluish-red that cannot definitively be placed in either the red or blue camp. Prototype effects were also found to structure categories like "birds," "vegetables," and "furniture."
11. George Lakoff (a linguist) and Mark Johnson (a philosopher) have extended these insights to posit cognitive functions that emerge from the human experience of living in a body and that are extended metaphorically (in radial categories) to classify and characterize even the most abstract concepts. One of these basic spatial schemas is "containment," which structures not only thinking about our location in space but also the way we express concepts in language-e.g. the widespread metaphor that words "contain" ideas. Thought and its representation in language, then, are shaped not solely or even primarily by cultural forces but also by the innate and universal physical parameters of our bodies and brains, as we attempt to make sense of and successfully negotiate what surrounds us, in nature and culture. Neuroscientists working both at the level of brain cells (neurons) and with larger brain structures and functional neuronal groupings (like the hippocampus, the amygdala, and the visual cortex) have arrived at a similar view of categorization and of the basic interrelatedness of mind, brain, body, and environment. Gerald Edelman, for example, explicitly presents his theory of neuronal group selection as a "biological underpinning" for Lakoff and Johnson's work on linguistic and conceptual categories (246-52). In turn, in arguing for a fundamentally embodied, anti-dualistic account of the mind-brain on the basis of neurological observation and experimental study, Antonio Damasio cites Rosch, Lakoff, Johnson, and Edelman as all having arrived at the central notion of a "body-minded brain" along converging routes (223-44).
12. The implications of these newer cognitive-neuroscientific theories of the mind for our currently dominant literary theoretical paradigms are potentially as revolutionary as was Chomsky's challenge to behaviorism. Consider, for example, how prototype theory crucially departs from Saussurean linguistics, the taking-off point for much post-modern theory. Saussure put forward two basic propositions about meaning: 1 ) that linguistic signs are arbitrary with respect both to the connection between phonetic form and meaning, and that between meaning and the world; and 2) that language is a closed system within which concepts are constituted out of their difference from other terms in the system, without reference to any extra-systemic reality. Rosch's work, however, suggests that linguistic meaning (as opposed to a given signifier/signified pairing) is not arbitrary, but is instead "motivated" by experiences of perception and embodiment.
13. According to Saussure, the signifier RED is arbitrarily linked with the experience of perceiving the color "red," which is itself based on an arbitrary (and culturally determined) division of the spectrum. According to cognitive linguistic theory, the link between the word RED and the color is still, of course, arbitrary, but the experience of redness is not; it is, instead, based on the fact that a specific range of wavelengths of light excite a particular set of cells in the human retina in a certain way. All human beings thus experience the prototypical color red in essentially the same way, and whatever their name for it, the concept of "redness" is motivated by physical experience. Furthermore, according to Saussure, RED is defined as red because it is not blue or green; the spectrum is thus carved up arbitrarily in pieces which are defined by their differences from all other relevant terms in the linguistic system. According to cognitive theory, however, a red is red because it resembles the prototypical red. Nor can red be definitively separated from blue, although there will be a prototypical blue from which it can easily be discerned. Thus, cognitive theory posits concepts that are not simply determined by the symbolic order in which they exist; instead, meanings are formed by an interaction of the physical world, culture, and human cognitive systems.
14. Other prevailing paradigms challenged by cognitive semiotics can be found within post-structuralist theory, which at first glance would seem to have close affinities. The Derridean critique of Western rationalism, for example, seems to parallel the cognitivist claim that rational thought does not comprise the natural functioning of the human cognitive system, just as Derrida's deconstruction of binaries can be seen as compatible with the cognitive insight that categorization is not structured by binary opposition, but by a gradient of resemblance to a prototype. When Derrida attributes to Western culture the view that "metaphoricity is the logic of contamination and the contamination of logic" (Dissemination 149), his argument overlaps with the cognitive linguist's sense that the metaphoric basis of human cognition places it at odds with most systems of formal logic. Cognitive linguistics, however, also challenges or modifies some post-structuralist concepts. Derrida, for instance, assumes Saussurean postulates of arbitrariness and definition by difference, even though he questions Saussure's phonocentric bias (Of Grammatology 2773). However, if meaning is not arbitrary but motivated, then slippage of meaning is at least partially bounded or anchored by its physical motivations. In addition, when Derrida argues that "there is nothing outside the text" he can be seen as postulating what amounts to a particularly sophisticated version of the behaviorist refusal to consider the systems and processes that produce behavior, whereas it is precisely these systems and processes that cognitive linguistics wants to investigate.
15. Post-structuralism and cognitive science stand to gain from critical juxtaposition since each has blind spots which the other may usefully illuminate. F. Elizabeth Hart has recently argued that cognitive linguistics can provide a more thoroughly materialist approach to language, meaning, and culture than can a Derridean post-structuralism still haunted by "residual formalism" ("Matter, System" 5). Hart suggests that this formalism has been carried over into the work of cultural materialist and new historicist criticism and prevents such criticism from acknowledging the complex and reciprocal relationship of culture and the embodied mind in forming the human subject. The critical paradigm which is most directly challenged by cognitive linguistics is thus that informed by Foucauldian assumptions about human behavior, for while cognitive science does give a large role to the shaping power of culture, it also stresses the role of innate mental dispositions and innately constrained cognitive procedures that hold across cultures and historical eras.
16. The Dani encountered by Rosch, for example, with no word for "red," will nevertheless point to the same prototypical red as the best example of the color, and will identify roughly the same range of shades as belonging to that category. Linguist-anthropologists Brent Berlin and Paul Kay found that cultures develop words for colors according to a universal pattern: if a culture has two terms, they will designate focal black and white, a third term will always be red, and a fourth term will be yellow or green, followed by blue, then brown, then gray, orange, pink and purple in no particular order. This suggests that while colors may exist in nature as an undifferentiated spectrum of light, the human perceptual system, even across cultures, will divide them in predictable ways. The number of color terms a given culture possesses is, of course, determined by cultural factors, and in general, the more complex the concept or topic, the more important are culturally determined elements in shaping its contours. Still, while research on categorization of concepts more complex than color suggests that there are striking cultural differences in the nature of prototypical examples for each category, it also demonstrates that these categories are structured in similar ways across cultures. Cognitive science, in short, does not rule out the role of cultural construction but promises more flexible and productive methods for studying the multiple and complex intersections of culture and human cognition as they influence and shape each other.
17. Competing accounts of human and authorial agency offered by different cognitive scientists at times support, and at times counter post-structuralist assumptions. Some computer scientists especially have theorized a human subjectivity that is as fragmented as that posited by Lacan (or by the contemporary neo-Lacanian Slovoj Zizek), most notably Marvin Minsky's conception of the mind as a "society" of multiple agents working in parallel fashion without centralized control. Similarly, the philosopher Daniel Dennett has devoted himself to debunking the common illusion of a homunculus or "little man" in the brain who both watches and directs conscious activity in a theater of the mind, proposing instead a model of contending agents or "demons" who produce a continually disrupted "stream of consciousness" from "multiple drafts"-a model that Dennett himself has compared to David Lodge's "faithful paraphrase" of deconstructionist theory in his academic satire Nice Work (Spolsky 39).
18. Other computer scientists and cognitive philosophers, however, leave room for the "emergence" of agency through the massive integration of neural activity. Neuroscientists seem especially willing to posit an integrated cognitive system, although they also clearly acknowledge a brain that has many different parts with functions that are to some extent separately localized. Edelman allows for a "personal individuality" that emerges from developmental and social interactions (167), while maintaining a more than Freudian insistence on the unavailability of much mental activity to conscious introspection and an acknowledgment of the presence of the "other" in our linguistically-based consciousness (146). Damasio, while denying that "all the contents of our minds are inspected by a single central knower and owner," nevertheless argues that "our experiences tend to have a consistent perspective" which is "rooted in a relatively stable, endlessly repeated biological state" (238). Damasio further revives something like a notion of "will" when he argues that rational decision-making functions in the brain are powerfully influenced by emotion. Cognitive scientists thus acknowledge a partitioned subject but also find suggestive ways to discuss our persistent belief that we possess integrated selves.
19. Cognitive theory also provides an incentive to rethink the Marxist views on the materiality of language and its role in material culture. Such criticism has to a great extent based its theory of language in Saussure and has moreover adopted a behaviorist refusal to consider cognitive processes. Marxist linguists like David Silverman and Brian Torode argue that "linguistic communication consists in the transmission of immaterial ideas or concepts from one person (speaker or writer) to another (reader or listener) by means of material signs," and since those immaterial ideas are located (by Saussure) in the brain and "the brain is unavailable to the researcher," then "its content, conceptual or otherwise, remains mysterious, and can only be the subject of speculation or arbitrary assumption" (3). It is no longer necessary or advantageous, however, to "black box" the brain in this fashion. The cognitive neurosciences have made the brain increasingly available to researchers and almost all cognitive theorists believe that the mind is essentially material since, as Kosslyn and Koenig put it, "the mind is what the brain does" (4). Damasio, Edelman, Lakoff and others have emphasized the ways in which human cognition is deeply influenced by its materiality and embodiment. Materialist criticism may thus need to begin to consider the implications of the brain as the material site where culture and biology meet and shape each other. Althusser, indeed, leaves room for such consideration when he argues that it is important to distinguish desire from "the biological realities that support it (exactly as biological existence supports historical existence) but neither constitute, nor determine it," even if that supporting role remains unexamined in his work (213n6).
20. Finally, materialist studies of literature will profit from acknowledging the brain as the material site where language, culture, and the body meet and form each other. Foucault's influential redirection of questions about the author's cognitive processes to questions about the place of authorship in material culture should be supplemented by a more thorough materialism that recognizes a stage of existence and circulation of discourse within the author's material body. This recognition would take the form of scrutiny of texts not only for traces of ideological formations or cultural systems, but for traces of cognitive process as well. Literary studies will also need to pay new attention to various writers' attempts, through history, to imagine, understand, and represent their own cognitive processes. Critics will have cause to reopen topics that will help to redraw the boundaries between body and culture: the effects of psychoactive drugs, for instance, or the physical manifestations of emotion, or the mental effects of physical experiences such as pain.
21. While much remains to be done, for over a decade now a handful of scholars working in literature departments, and a few others in computer science and cognitive psychology, have in fact been moving toward the convergence of cognitive science and literary studies largely without reference to one another's work. We might begin with Norman Holland's The Brain of Robert Frost (1988), which he claims as the "first book to bring to bear on literary criticism and theory the revolutionary discoveries of cognitive science and recent research into the brain" (vi), although Frederick Turner had included several chapters on literature and the brain a few years earlier in Natural Classicism (1985) and Revuen Tsur had been working in the field of "cognitive poetics" since the 1970s. Holland, wellknown for his psychoanalytical approach to reader-response theory, makes a strong case for replacing the social constructionist paradigm of mind relied on in most literary theory with the "more powerful psychology" emerging from cognitive neuroscience (6, 13). Tracing the social constructionist paradigm back to the "stimulus-response" model of early 20th-century behaviorism, Holland is especially persuasive in proposing a more active and fine-grained sense of the reader's work of cognitive processing in reader-response studies, and he rightly emphasizes the critical role of"experience"-and thus of social context and genetic development, history and biography-in neuroscientific accounts of the mind (7-8).
22. Holland's desire to retain central elements of psychoanalytic theory along with new insights from cognitive science, however, produces a book which does not fully realize the potential of a brain-based psychological literary criticism; his psychological portrait of Frost, for example, reads very much like conventional Freudian psychobiography (16-42). Tsur's weighty and wide-ranging Toward a Theory of Cognitive Poetics (1992), which synthesizes twenty-five years of research, is characterized by an even greater sense of eclecticism; in attempting to develop new approaches to versification, figural language, emotive effects, and critical response he draws upon early Gestalt psychology, Russian and Prague school formalism, psycholinguistics, the New Criticism, and structuralism as well as cognitive science. If the "cognitive" elements of his theory often fail to emerge cleanly, Tsur offers a wealth of leads and insights and some of his findings seem especially useful-such as his cross-linguistic theory of poetic caesurae (134-39), which overlaps suggestively with Frederick Turner's "universal" theory of the poetic line (61-108).
23. Two books more pervasively conversant with cognitive linguistics and categorization theory, Mark Turner's Reading Minds (1991) and Ellen Spolsky's Gaps in Nature (1993), provide readable, engaging, and sophisticated examples of how literary scholars can both profit from and contribute to work in the cognitive sciences. Turner, who has collaborated with Lakoff, delineates a "cognitive rhetoric" which ties literary experience to our everyday (but largely unnoticed or unconscious) activities of conceptualization, categorization, and, especially, invention and interpretation of metaphors and other rhetorical figures. Turner's rich sense of connectedness among the language-producing human brain, the body that includes it, and the "physical environment that it must make intelligible if it is to survive" ( 17), as well as his informed speculation on the likely neural substrates of such adaptive, creative linguistic and literary activity, give his work a strong sense of conviction. Some literary scholars, however, have raised objections to his call for a "new common ground" and large-scale reconstitution of the "profession of English" on the basis of cognitive linguistics, a field that he himself characterizes as still speculative and "embryonic"(vii-viii, 20).
24. Spolsky is less interested in explicating the extraordinary productivity of ordinary linguistic competence than in theorizing how the mind's inherent "cognitive instability"-arising from the gaps and incommensurabilities among various brain functions and the fundamental neuronal plasticity that keep categories fuzzy and concepts fluid-makes possible the literary innovation we find in "culture's most powerfully imaginative texts" as well as in our own attempts to rewrite literary history (2, 4, 7). For Spolsky, supplementing cultural materialism with a "materialist, biologically based" model of the mind-brain helps account for innovation within what might otherwise seem a linguistic and cultural prison-house: resistance to "reigning cultural patterns" stems from the brain's stubborn materiality, the friction between functional modules or neuronal groups dedicated to divergent sensory and motor domains (206). Her illustrations of such resistance from feminist criticism and ecriture feminine are of particular interest given how little attention to gender has been paid by cognitive approaches to date, although her distinctions between "associative" thinking, "presocialized feelings," and feminist modes of discourse on the one hand, and traditional (male) "logical/deductive" modes on the other, seem open to the critiques directed at other forms of "difference" feminism (202-04).
25. Two researchers in computer science, Jerry Hobbs (in Literature and Cognition) and Herbert Simon (in the target essay for Bridging the Gap, a special issue of the Stanford Humanities Review) bring their own disciplinary perspective to bear on issues of literary theory and criticism. Having spent years prodding machines to decode natural utterances like "we're having our new colleague for dinner Thursday" without invoking the "Urban Cannibal" script, they are in a good position to charge literary theorists with portraying human text processing as more difficult and ambiguous than need be. Their position of relative detachment can produce unexpected and refreshing insights on such matters as intention (how is the text produced by a pocket calculator "intentional"?) and context (what rules do we tacitly follow in delimiting the potentially limitless set of contexts for a given text?). Both Hobbs and Simon, however, fail adequately to account for the emotive aspects of literary production and reception and, as Mark Turner points out in his cogent response to Simon, they ignore to their cost that "a human brain resides in a human body in a human environment" ("Cognitive Science" 111). A few of the other responses in Bridging the Gap assail Simon and cognitive science in general with little or no sense of the diversity of cognitive approaches or the debates among various cognitive theorists and neuroscientists. Such wholesale resistance underscores that considering a given cognitive theorist's work without some knowledge of the larger cognitive enterprise can all too readily lead to incomprehension. Scholars who would begin the important work of gauging the usefulness of neural and cognitive models for literary studies would do well to explore various tendencies before choosing any one for specific critique, or, for that matter, endorsement.
26. Some narratologists have begun turning to AI and other fields within cognitive science in an effort to rejuvenate their field. For example, Gary Saul Morson, in his foreword to a recent reissue of Roger Schank's Tell Me a Story, credits Schank (an important AI pioneer) with having "reinvented" the discipline of narratology (ix). Morson finds particularly refreshing Schank's emphasis on the "amazing web of ordinary and surprisingly complex responses that comprise daily life"-an emphasis that aligns Schank's work on narrative and intelligence with Mark Turner's recent work on parable and "narrative imagining" (Literary Mind 4-5), both viewing story-telling as a pervasive and highly adaptive aspect of mental life, tying our most quotidian acts of problem solving to literary creativity. Morson also identifies, quite intriguingly, an area of common ground between Schank's work on the pervasiveness of story in social discourse and the dialogic theory of discourse developed by the Bakhtin group (xix-xx, xxxi). Similarly, Marie-Laure Ryan finds in AI (as well as in "possible worlds" theory) an urgently needed "new source of ideas" for narratology: she fundamentally recasts, for example, conventional narratological notions of framing or embedding by appealing instead to the model of "stacking" employed in AI, showing how reconceptualizing narrative as a "computer language" can account for the complex metafictional effects notoriously resistant to traditional narratology (3, 175).
27. Another promising site for new interdisciplinary work has been mapped out by cognitive psychologists, who share with computer scientists a vocabulary of "scripts" and "story grammars," "scenarios" and "text processing," and who have begun to apply their considerable knowledge of language comprehension to specifically literary topics. Richard Gerrig and Raymond Gibbs have both produced books that emphasize the continuities between literary experience and everyday linguistic production and comprehension. Gerrig, in Experiencing Narrative Worlds, and Gibbs, in The Poetics of Mind, draw on hundreds of experimental studies which decompose and clarify the cognitive work performed by readers in discriminating, interpreting, remembering, and paraphrasing literary and nonliterary texts alike. In the spirit of Lakoff, Johnson, and Turner, both Gerrig and (especially) Gibbs argue that conventionally "literary" topics like narrative and figurative language are central rather than peripheral to understanding the basic workings of human cognition, revealing what Gibbs calls the "poetic structure of mind" (2). While both psychologists draw on literary theory and criticism in filling out their "poetic" accounts of mental life, however, neither has much to say about possible material substrates or additional evidence from neuroscience, although as early as 1978 Howard Gardner and Ellen Winner had noted the value of such evidence, for literary scholars and philosophers as well as for cognitive psychologists ("Development").
28. David Rubin, in contrast, does consider neuroscientific evidence in Memory in Oral Traditions, his groundbreaking and inclusive study of what the ability to memorize various forms of oral poetry reveals about mental functioning and how, in turn, the nature of human memory shapes elements of such poetic forms on many levels, from alliterative word pairings to large-scale narrative structures. Summing up and refining a whole tradition of empirical work on memory and generously providing more information than is strictly necessary about the principles and methods of cognitive psychology as a kind of ongoing tutorial for the student of literature, Rubin has provided one of the most important and useful examples to date of how cognitive science can help illuminate traditionally literary concerns.
29. Other literary scholars adapting empirical methods from the cognitive neurosciences to reconsider or reopen literary questions include David Miall and several contributors to a recently published volume on Empirical Approaches to Literature and Aesthetics (Kreuz & MacNealy). In an empirical study of "literariness," Miall argues for regrounding readerresponse criticism on the "reasonably clear neuropsychological evidence" emerging from studies by Damasio and others on the interrelation of cognitive and affective functions: he outlines a research program for elucidating the effects of literary foregrounding and other anticipatory elements of the reading process in terms of what is being learned about affective anticipation and links between the prefronal cortex and the limbic system (294-95). Miall cites work-in-progress by Johan Hoorn (later published in Empirical Approaches) on "psychophysiology and literary processing" that has established, through measuring event-related potentials with EEG (electroencephalogram) recordings, the salience of the "deviation" effects in poetic language theorized by Slavic formalist critics. Tom Barney, another contributor to the volume Empirical Approaches, attempts to bring work on stylistics (particularly metrics) into line with the psycholinguistic research into phonetics, and W. John Harker argues that cognitive research into attentional behavior in text processing can help account for felt differences in reading literary and nonliterary works.
30. Yet however intriguing the preliminary results, and however timely the acknowledgment that findings from scientific research can help delimit and redirect literary investigations, empirical studies should by no means be thought of as the only, or even the principal, way for students of literature to profit from work on the brain and cognition. Elaine Scarry, in one of the most impressive essays in literary criticism yet to engage with cognitive neuroscience, draws on work by Stephen Kosslyn and others on mental imagery to reopen the old question of the "vivacity" of the literary imagination: why do images evoked by the reading of certain literary texts seem so vivid, given the flatness of mental images in general? Her methodology relies almost entirely on the sort of rich introspection and literary tact that scholars of literature, with years of attentive reading and a good deal of insightful published analysis to build on, are uniquely trained to bring to textual interpretation, qualities that in fact are highly valued by Hobbs and other cognitive scientists. Scarry herself, however, is now working with Kosslyn to design empirical tests, using magnetic resonance imaging (MRI) scanning, of her hypotheses concerning literary devices and mental imagery (Magner), suggesting that the intuitions of trained readers can provide models for confirmatory research in controlled settings.
31. The dizzying progress of cognitive and neuroscientific research over the past twenty years has produced a large number of models and findings of pressing relevance to such literary topics as prosody, narrative poetics, reader response, figural language, genre theory, and imagery. In addition, theorists can be expected to feel the pressure to begin rethinking received notions like the arbitrariness of the linguistic-conceptual field, the sociocultural constructedness of the self, and the bracketing-off of the neural substrates of mental activity. Indeed, literary scholars like F. Elizabeth Hart have already begun looking to the "cognitive linguistics" of the Lakoff group, for example, for its disclosure of a "middle space in the dichotomy between determinacy and indeterminacy" bedevilling current literary theory and for its capacity to "demysify figurative language" and describe its effects in compelling ways unavailable both to structuralist and deconstructionist positions ("Cognitive Linguistics" 22). There is also an opening for what might be called a "neural historicism," which would explore how the peculiar structure and workings of the human brain may enable cultural innovation over time and offer revisionary accounts of the representation of mind and mental processes in literary-cultural history along new lines suggested by the frameworks and models emerging from the cognitive neurosciences. Conversely, recent work in cognitive rhetoric and the "poetics of mind" presents an invitation to literary scholars and theorists to bring their particular expertise to bear, as collaborators, critics, or both, on the further development of cognitive science. It is worth recalling that one of the more fruitful distinctions in 20th-century literary studies, that between metaphor and metonymy as opposite poles of both linguistic and literary activity, was formulated by the linguist and poetician Roman Jakobson as a contribution toward the development of early neuropsychology. Those willing to cross disciplinary lines in the age of cognitive science stand to contribute similarly productive models and to elicit a whole range of new meanings in the texts, practices, and historical and cultural configurations that we study.*