a review of
Elaine Scarry, Dreaming by the Book
by Ellen Spolsky, Bar-Ilan University
[rpt. with permission from Poetics Today 24.1 (2003): 127-37]
Dreaming by the Book. Elaine Scarry. New York: Farrar, Straus, Giroux, 1999. 292 pp.
When I was a graduate student in the 1960s, I asked a neurophysiologist friend how my brain allowed me to understand what it meant for my love to be like a red red rose. "Hmph," he snorted, in scorn of my presumption: "I work on a single neuron in the squid." Even though he is also a prize-winning poet himself, he was not interested in speculating about how the phenomenology of understanding poetry might be instantiated in the brain. Accepting this rebuff, I contented myself with the formal theories of meaning that predominated at the time. Describing a theory as "formal," at least in generative linguistic circles, was and still is a shorthand way of saying that it makes no claims about how the components and processes hypothesized by the model might map onto either the processes of the human mind, as psychologists would describe them, or its physiology or neurology. Theories that do make such claims have recently come to be called theories of embodiment, and it is as a contribution to these theories and to their struggles against formalism that I value Elaine Scarry's book.
The formalist disengagement of levels of description is a methodological reduction that produces a distinction between form and function, usually with an assumption that the two will eventually be reconnected. The advantage of this approach is assumed to be that, even while the second goal is elusive, work can continue on the first. There is an unspoken trust that the explanation of the relationship will somehow emerge from the sheer density of the descriptive work in stage one, but there are always those who will try to hasten its coming. Reaching, immodestly, thus, for answers to the Big Questions about how human heads make meaning, some literary scholars in the 1960s invested in a heretical version of New Critical formalism called stylistics. In these studies, the descriptive categories of linguistics were mobilized to describe formal differences among writers, and
This work had not gotten very far, however, when Stanley Fish (1980a , 1980b) attacked the project with even greater scorn than my neurophysiologist friend had shown. Don't fool yourselves, he told these hyper-formalists, into thinking that you are discovering the connections between form and meaning. Your project is incoherent because it assumes two contradictory positions at once. It begins with an assumption that any intended meaning can be expressed in more than one form, since only if this is true could an author's style be truly a matter of choice. Yet, it must also assume that a given meaning can only be expressed by one form, since only if this is true can it be claimed, as the stylisticians did, that, in describing the author's formal choices, they have identified the intended meaning automatically and objectively.
Fish's critique of the stylisticians' assumptions was itself open to criticism, but it was an important landmark in the ongoing struggle against reductionism in literary theory. It also served to bring into clearer focus the larger issue of determinism and choice in language and thought, already broached by E. D. Hirsch, Jr. (1967) and, in the 1970s, featuring centrally in the language theory of the Yale deconstructionists. In Fish's own work on seventeenth-century poetry (1967, 1972), he did not make the mistake of claiming that he had discovered an objective method of arriving at the meaning of a text. He demonstrated, nevertheless, that he shared the stylisticians' underlying curiosity about how writers lead, even ensnare, readers by the form of their texts. Thus, although there were, no doubt, important differences among the American deconstructionists, the reader response critics, and American critics such as Stanley Fish and Jonathan Culler, all confronted the exclusions enforced by the New Critical focus on the text itself, all revealing how hard it is to live with the constrained ambitions of formalism.
Also during the 1970s, some literary critics became bolder in their demands for the recognition of readers' distinctive experiences as part of the process of making meaning. This demand appeared, for example, in what seemed to many to be an ignorable fringe preoccupation of women: feminist criticism. But, soon, other groups widened the demand for an accounting for difference, a demand that could not be accommodated within a theory that ignored the social contexts and also the explicitly embodied reality of readers. The social critique of Michel Foucault (1969), in its insistence that the power structures of a culture can be read off that culture's evaluation and use of specific texts, argued for the recognition of bodily, class, and gender differences as the grounds for resisting power. An impressive diversity of academic voices was raised in the 1970s to speak of power sharing and the multiplicity of truths. In the practice of the anthropologist Clifford Geertz (1973), for example, in the history of science (see Arbib and Hesse 1986), and in the essays of the population geneticist R. C. Lewontin (1983) there were both a new recognition of the interconnectedness of systems not previously understood as causally connected and a respect for the complexity of those interrelations. It became increasingly possible to question the methodological reductionism that had insulated the study of material texts from the study of the biological producers and consumers of texts or that had generalized them into a single system called the text or discourse.
At the same time, social scientists and humanists began to hear from cognitive scientists about connectionist or parallel processing models of brain function. As distinct from the computational model of mind (propounded by John von Neumann in the 1950s and defended, famously, by the linguist Jerry Fodor  as "the only model we've got"), connectionist models do not, as the digital computer does, take the outcome of one computation as the input to the next. They do not, that is, depend on a linear and hierarchical sequence of ordered processes. Parallel processes within the brain make many computations on many kinds of input data simultaneously, producing competitive patterns of activity, some of which come to dominate the brain's activity, probably because they are the more stable ones and are mutually reinforcing in the production of useful results. Another way of saying this is that, in a connectionist model of the mind, individual neurons do their work by being appropriately connected to large numbers of similar units. They become appropriately connected over time, developing according to the growing experience of the individual brain. The patterns of activity that are reinforced and become established are those that work for the benefit of the individual organism. In the neo-Darwinist terms some cognitive scientists have borrowed, those patterns of neuronal activity that come to dominate are those that fit the organism for the jobs it has to do in order to survive.
Compared to what seemed to most humanists the blatant unsuitability of the binary or digital computational model to ground an understanding of how one reads poems, the connectionist models once again raised hopes that there might yet be a non-reductionist answer to the questions of how the material body (including its brain) produces and understands the unique texts we consider literary. Crucially, the binary models could not recognize just those kinds of shadings and gradations that interested literary readers. A description of the activities of a connectionist or parallel processing model, however, comes closer to describing the actual feel of one's common experience of thinking. This model suggests that one's everyday sense of mental life as being somewhat fragmented, as proceeding from a mix of the habitual and the cognitively challenging, and as producing judgments of varying strengths with varying degrees of security is just conceivably related to the parallel jostling for attention and competition for supremacy which the connectionists hypothesize to be taking place among neurons and among groups of neurons under your hair.
When one begins to understand the complexity of the interactions of the human brain's thirty billion nerve cells and learns that the neuronal connections in the brains of even two cloned fruit flies are not identical, one begins to see why connectionist models of brain activity seem appropriate for what theorists of embodied difference want to explore. The way Gerald Edelman (1998: 42) describes the brain's organization would thus encourage a literary theorist to hope that a closer mapping between poem and brain may yet be possible:
Each brain has uniquely marked in it the consequences of a developmental history and an experiential history. The individual variability that ensues is not just noise or error. . . . It is an essential element governing the ability of the brain to match unforeseeable patterns that might arise in the future.If, in real brains, control is decentralized or distributed among a set of nonhierarchicalized nodes that can, by virtue of this three-dimensional and gradient capability, deal with unforeseen experiences, then it is conceivable that a cognitive science based on this new model would be able to map how experienced readers can understand a sentence such as "Do not go gentle into that good night." Perhaps we need not assume an unbreachable barrier between the way neurons behave and the way people behave. These revisionary sciences might be able to shed light on just how humans produce and understand previously unencountered and apparently ungrammatical or meaningless verbal formulations that are both related to already familiar meanings and are also significantly different‹just different enough to allow the speaker to say something newly valuable because it is different.
The challenge, then, to the computational model of mind parallels a broader challenge within the social sciences and the humanities to the hegemony of hierarchical thinking and converges interestingly with several post-poststructuralist theories of the embodiment of knowledge. These latter are theories that attempt, each from its own direction, to help literary study recover from its inherited Cartesian dualism and also from the deep skepticism engendered by its overvaluation of the abstract mind. From Judith Fetterley's (1978) feminist claim that being a woman rather than a man makes a significant difference in how one reads "The Legend of Sleepy Hollow" to George Lakoff and Mark Johnson's (1980) claim that language is patterned by (for example) the symmetry and verticality of our human bodies, studies of literature and language were waking up to the losses that had been incurred by the abstraction of interpretative strategies away from bodies and by the absorption of all meaning into textuality.
Elaine Scarry's 1985 book, The Body in Pain: The Making and Unmaking of the World, was crucially important to this recovery and to the struggle to find a way of talking about the analogical knowledge bodies provide. Its title is almost enough to make its claim. Our bodies, she asserted, know some things very well: pain, for example. From this foundation we might build a way of using language to analogize among bodily experiences and the ideas we trust, or want to trust, as truths. Scarry's 1999 book, On Beauty and Being Just, argues, then, if only by implication, that humans learn not merely from the pain of torture, the ugly side of the physical world; it is the material nature of human bodies that counts. Scarry now claims that the sensations of physical pleasure experienced in the presence of beauty share with extreme pain the ability to force the suspension of thought and revalue the habitually valued abstraction which is one of the outcomes of routine human mental processing. For Scarry, symmetry, something we know phenomenologically, both visually and kinesically, is the embodied version of justice. Knowing the one is knowing the second.
Again, in Dreaming by the Book, Scarry regales us with her own luscious language transformations, here, of the beauty of her garden and its birds, arguing against dissipating the power of experience itself by carelessly and overly rapidly transforming it into abstraction. Her accomplishment in these three books has been to warn against the loss incurred by the collapse of distinctions along the continuum from the idiocentric experiences of the individual through language to the abstractions of politics and ethics. By her decision to foreground the poetry and the powerful images both from books and from her own lived experience, and simultaneously to subordinate the references to research in cognitive studies, her book asserts that the local and particular must not be buried in theory. Grounded in J. J. Gibson's (1979) interactive approach toward vision, and learning from Stephen Kosslyn's (1980, 1995) recent work on visual imaging to move toward an understanding of the way the imagination depends on everyday visual competence, Scarry considers a group of hypotheses about how literary texts induce us to produce images with unusual vivacity. She thereby nourishes the goal of that chain of literary scholars before her that eventually it will be possible to fill in the dots between lines of poetry and human brains. The project of being able to describe how the architecture of the eye supports the ability of Homer to make us see visions is part of the larger one of asking how bodies (different bodies) analogize from what they know by virtue of having bodies and, ultimately, of understanding how humans fit into and fit themselves to the worlds they live in.
In electing to publish many separable claims about how authors make us see vividly, grouped and summarized as "five formal practices" (239), Scarry indicates, by her adjective formal, that she is more comfortable describing the texts than asserting their meanings. Thus, there is virtually no discussion here of the function of the vivacity that emerges from these formal practices. She nevertheless makes an important generalization about the way texts and brains relate, namely, that good writers are good because they construct their texts so as to recruit the brain's built-in abilities: "imaginary vivacity comes about by reproducing the deep structure of perception" (9). Good writers do not tell us what we need to understand; rather, they know how to tell us what we are able to understand.
Scarry herself worries (256 n. 6) that her argument may be circular, but the problem, I believe, is probably its very high level of generalization and the distance between the generalization and the details of the many examples. The strong version of her claim (which she does not in fact make), based on the connectionist model described above, would hypothesize that the structure of perception can be read from the ways the authors successfully mobilize aspects of it. This is an important claim for literary theorists interested in how the new cognitive sciences relate to their work, because it implies a potentially productive interdisciplinary bi-directionality. Not only might the theories of the cognitive scientists explain to the literary critic how poetry is understood, but, at the same time, the evidence of how literary texts are read should provide evidence to the cognitive scientists about the structure of the brain. Specifying how this works is a project for many hands, and Scarry's methodological decision not to report in detail on the cognitive research that has apparently prompted or confirmed some aspects of her own speculation makes this book unhelpful as a guide to the work that awaits those interested in filling out the hypothesis.
But if it is not a systematic guide, it is a richly suggestive one. For me, Scarry's most provocative conjectures are those concerning the relative ease or difficulty with which a reader can be led to envision what the author is imagining. The variety of suggestions here that human minds grasp some kinds of knowledge more easily than others because of the way those minds are constructed promises, when pursued, to provide just the kind of enrichment that will save the theory from circularity or vacuousness. It promises, that is, to add an important gradience and thus specificity to the claim that the structure of the mind is displayed by the ways writers call on readers to envision. It will also have implications for the perennial literary issues of change and of style and periodization‹just those issues over which Fish tangled with the stylisticians. One might ask how or even whether specific features in the historical and cultural context of a text and its author contribute to its ease of understanding.
Given this potential for understanding the relationship between cognitive and cultural construction, it is unfortunate that Scarry never acknowledges the importance of an individual's visual memory, stocked, as it must be, by both unique experiences and the cultural contents of one's surroundings. In discussing Homer's image of a "surging assembly" (119), for example, she praises the aptness of the poet's images of waves at sea and fields of grain in the winds. Today's readers, however, are surely just as familiar with surging assemblies from watching the nightly news and may, depending on where they live, have rather less-vivid images of the sea or of wind-bowed waves of grain. The question of how readers' reactions to texts differ depending on their own stores of visual memories is not even referred to here. This is particularly odd for someone who seems to be learning so much from her own personal visual life. It has long been known that the functioning of visual processes is influenced by the kinds of visual experience that have been available to an individual visual system as it develops. It would be hard to argue, for example, that computer image morphing, now regularly used in animated advertising, has not encouraged new and easily imagined transformations for many readers. Scarry is clearly entitled to prefer flowers and birds to television advertising (see her Acknowledgments), but her bias has here handicapped her effort to learn about the evocation of visual experience.
Scarry also disappoints by her neglect of the intensive work already done on the way the modular systems of the mind/brain translate among themselves. Clearly involved in the success of the texts that interest her are the way different parts of the visual system are differently engaged by (for example) the language system. Although her own prose shows her to be highly sensitive to the power of language, she does not discuss what she surely knows about the rhetoricity of language, its limits and possibilities. She makes no distinction, for example, between the claim that the movement of light fabrics, a gauze curtain or a silk scarf, is relatively easy to imagine‹presumably because, given the light weight of the fabrics, any stored memory of a curtain (and less obviously) a scarf may itself already represent it in motion‹and the similar claim, as she sees it, that "wheeling" is an easy movement to imagine. The first may indeed be a question of visual memory, while the second may be also a matter of the lexicalization, in English, as a single verb, of just that symmetrical movement she claims is so easy to envision. Similarly, her remarks about the ease of imaging movement from verbal descriptions of stretching, folding, tilting, circling, and skating are empty without reference to the iconic nature of parts of speech. Many verbs refer to movement, and we know this from their morphology: consider the different images that occur to you when you hear the word rock and, then, rocking.
I suppose it is a sign of how stimulating her discussion of imagery is that I found myself producing additional examples or counterexamples throughout the book. One of the most striking in my memory is of a heavy object moving, arising in response to Scarry's claim for the ease of imagining light materials moving. In the last scene of George Eliot's The Mill on the Floss (1979 ), on the flood-swollen river, "some wooden machinery had just given way on one of the wharves, and huge fragments were being floated along." This large object, never clearly delineated, crashes into the rowboat in which the novel's protagonists are attempting escape, and they are killed. "The huge mass was hurrying on in hideous triumph." Although Eliot never even tries to help us imagine the exact shape of the machinery, the verbs float and hurry seem entirely sufficient to stimulate the required picture in my mind, even though they contain no semantic references to shape. I imagine the scene clearly, and Eliot seems to know that it does not matter if the shape I envision is different from the one she envisions.
If much remains to be done in the investigation of how phenomenological transformations or physiological translations among brain modules are accomplished (see Jackendoff 1987), we do not have to start from scratch. On the physiological level, for example, the interconnectivity among vision, language, and kinesis is an issue on which an enormous amount of work has been done in cognitive linguistics and in formal semantics. Word and image studies have already begun to answer questions of what can be well translated, what will be lost, and what added in the transformations. Cognitive cultural studies have begun to suggest that different people differ significantly in how they manage these translations and that these differences may be culturally encouraged.
But, as Scarry's choice of a nonacademic publisher signals clearly, I, as an academic literary theorist with a strong interest in cognitive science, am not her intended audience. Indeed, just those features of her approach which make her book less useful than it might have been to me, namely, her avoidance of the familiar terminology of such academic topics as inference and reference, grammar and rhetoric, word and image studies, and verbal and visual memory, surely make it readable to a wider audience. Scarry's niche is already constructed: she speculates, proposes, and occasionally shocks by her startlingly new connections. She stimulates academics to consider issues that do not fit into the academic containers we already have. Her detailed insights into her own experiences provide the data eventually to be connected to the abstractions of interpretive theory. Whether or not we agree with how she perceives things to be, they suggest an agenda for a great deal of work at many levels from the neurological through the functional and up to the most abstract.
In sum, Scarry seems to be trying, as many literary theorists have been, since Northrop Frye's (1957: 18) demand that literary criticism be more than a "history of taste," to reckon just how far we can engage with the discoveries of science without being distracted from cultivating our own gardens. To me, then, the value of her study lies in her rejection of reductionism and her encouraging us to notice small but significant phenomenological differences. Her hypotheses testify to a healthy respect for both the difficulty of and the necessity of accounting for the material complexity of the interactions of individual readers with valued cultural texts and with the various social/biological systems within which writers and readers grow. This is a book that literary theorists interested in the way human brains and bodies understand literary texts will want to read; and as my lightning summary of the concerns of literary theory over the past forty years indicates, that is more or less all of us.
Arbib, Michael, and Mary Hesse
1986 The Construction of Reality (Cambridge: Cambridge University Press).
Chatman, Seymour, ed.
1971 Literary Style: A Symposium (New York: Oxford University Press).
1982 On Deconstruction: Theory and Criticism after Structuralism (Ithaca: Cornell University Press).
Dolezel, Lúbomir, and Richard W. Bailey, eds.
1969 Statistics and Style (New York: American Elsevier).
Edelman, Gerald M.
1998 "Building a Picture of the Brain," Daedalus 127 (2): 3769.
1979  The Mill on the Floss, edited by A. S. Byatt (London: Penguin).
Feldman, J. A., and D. H. Ballard
1982 "Connectionist Models and Their Properties," Cognitive Science 6: 20554.
1978 The Resisting Reader: A Feminist Approach to American Fiction (Bloomington: Indiana University Press).
Fish, Stanley E.
1967 Surprised by Sin: The Reader in Paradise Lost (Berkeley: University of California Press).
1972 Self-Consuming Artifacts: The Experience of Seventeenth-Century Literature (Berkeley: University of California Press).
1980a  "What Is Stylistics and Why Are They Saying Such Terrible Things about It?" part 1, in Is There a Text in This Class? 6896 (Cambridge: Harvard University Press).
1980b "What Is Stylistics and Why Are They Saying Such Terrible Things about It?" part 2, in Is There a Text in This Class? 24667 (Cambridge: Harvard University Press).
Fodor, Jerry A.
1979 The Language of Thought (Cambridge: Harvard University Press).
1969 L'Archéologie du savoir (Paris: Gallimard).
Fowler, Roger, ed.
1975 Style and Structure in Literature (Ithaca, NY: Cornell University Press).
Freeman, Donald, ed.
1970 Linguistics and Literary Style (New York: Holt Rinehart). [End Page 136]
1957 Anatomy of Criticism: Four Essays (Princeton, NJ: Princeton University Press).
1973 The Interpretation of Cultures (New York: Harper Torchbooks).
Gibson, J. J.
1979 The Ecological Approach to Visual Perception (Boston: Houghton Mifflin).
Groves, Philip M., and Kurt Schlesinger
1982 Introduction to Biological Psychology, 2d ed. (Dubuque, IA: Wm. C. Brown).
Held, R., and A. Hein
1963 "Movement-Produced Stimulation in the Development of Visually Guided Behavior," Journal of Comparative and Physiological Psychology 56: 87276.
Hirsch, E. D., Jr.
1967 Validity in Interpretation (New Haven, CT: Yale University Press).
Holub, Robert C.
1984 Reception Theory: A Critical Introduction (London: Methuen).
1987 Consciousness and the Computational Mind (Cambridge: MIT Press).
Kosslyn, Stephen M.
1980 Image and Mind (Cambridge: Harvard University Press). 1995 "Topographical Representations of Mental Images in Primary Visual Cortex," Nature 378 (6556): 49698.
Lakoff, George, and Mark Johnson
1980 Metaphors We Live By (Chicago: University of Chicago Press).
Lewontin, R. C.
1983 "Gene, Organism, and Environment," in Evolution from Molecules to Men, edited by D. S. Bendall, 27385 (Cambridge: Cambridge University Press).
Love, Glen, and Michael Payne
1969 Contemporary Essays on Style (Glenview, IL: Scott, Foresman).
McClelland, J., D. Rumelhart, and G. Hinton
1986 "The Appeal of Parallel Distributed Processing," in Parallel Distributed Processing: Explorations in the Microstructure of Cognition, vol. 1: Foundations, edited by D. Rumelhart et al. (Cambridge: MIT Press).
Mosteller, Frederick, and David Wallace
1964 Inference and Disputed Authorship: The Federalist (Reading, MA: Addison-Wesley).
Mountcastle, Vernon B.
1998 "Brain Science at the Century's Ebb," Daedalus 127 (2): 136.
1985 The Body in Pain: The Making and Unmaking of the World (New York: Oxford University Press). 1999 On Beauty and Being Just (Princeton, NJ: Princeton University Press).
Sebeok, Thomas, ed.
1960 Style in Language (Cambridge: MIT Press).
2001 Satisfying Skepticism: Embodied Knowledge in the Early Modern World (Aldershot: Ashgate).
2002 "Darwin and Derrida: Cognitive Literary Theory As a Species of Post-Structuralism," Poetics Today 23: 4362.
Suleiman, Susan, and Inge Crossman, eds.
1980 The Reader in the Text: Essays on Audience and Interpretation (Princeton, NJ: Princeton University Press).
Von Neumann, John
1955 The Computer and the Brain (New Haven, CT: Yale University Press).
 See, for example, Sebeok 1960, Dolezel and Bailey 1969, Love and Payne 1969, Freeman 1970, Chatman 1971, and Fowler 1975.
 The only way Fish's criticism could be understood to have hit its mark would be if the meaning of "choice" were trivialized so that an author's "choice" (his or her freedom, indeed), meant simply that either of (say) two possible forms would work just as well, producing no difference at all in meaning. However, the same semantic dynamic that allows language change over time also allows a writer to use familiar words and phrases in new ways and thus to express fine distinctions by differentiating even very close synonyms. It is precisely the ability to exploit this possibility of language that distinguishes an interesting writer from an ordinary one.
 In the anti-Cartesianism of the deconstructionists, the issue of how the form of a text produced its meaning was made to disappear by the denial of the distinction between reader, context, and text. See Culler 1982 on Jacques Derrida, Paul de Man, Geoffrey Hartman, and J. Hillis Miller.
 See, for example, Suleiman and Crosman 1980 and Holub 1984.
 See Feldman and Ballard 1982 and McClelland et al. 1986.
 "Neurons can take on any one of a series of values over a continuum, transistors in digital circuits only a 0 or a 1" (Mountcastle 1998: 30).
 See Spolsky 2001 and 2002.
 The standard reference here is to Held and Hein 1963, demonstrating the importance of individual experience (and motion) to the development of the visual system. Other related experiments demonstrating the importance of experience are summarized in Groves and Schlesinger 1982.