Hypertext: the convergence of contemporary critical theory and technology. By George P. Landow. Baltimore: the Johns Hopkins Press, 1992. xii + 242 pp. $48.50 hardcover, 15.95 paperback. The subtitle of this book describes a meeting, or attempted meeting, between critical theory and the new interactive, manipulable, interpolable, computer text. The thesis is that both partners in the meeting "argue that we must abandon conceptual systems based upon ideas of center, margin, hierarchy and linearity and replace them with ones of multilinearity, nodes, links, and networks." (2.) I believe that it might be more accurate to describe hypertext in Landow's book not so much as one of the partners but as the venue or locale of a somewhat different meeting, between traditional literary scholarship and poststructuralist critique, with implications that are both fascinating and disturbing. George Landow is evidently a fine literary critic and scholar of Victorian literature, a good teacher, a sensitive reader with a refined historical sense, and a decent human being, who has found and used hypertext to do some interesting things in his classroom at Brown University. I am not convinced that his own critical sensibility has been much assisted by the computer, however. The generation of literary academics that grew up before the computer essentially had to use their own brains to do hypertextthis was a large part of their training. What is happening now is that this expertise is being turned into an "expert system," through the hypertextualization of literary texts, so that students with less cultural background will be able to take a sort of crash course in literary scholarship. Perhaps the traditional literary education, in which the brain was hypertextualized, was richer and deeper, but the new electronic concordance/reference systems are certainly better than nothing, and may give students in the humanities more time for other activities, like the study of science, in which they are often woefully ignorant, and of works of literature, now frequently replaced by works of critical theory. Landow's book is a good introduction to the kinds of hypertext that can be used in literature classes. The description of his own hypertext version of Tennyson's In Memoriam is especially valuable. In his discussion of the Victorian sermonist Henry Melvill he shows that hypertext can be valuable for traditional scholarship as well. He quotes Vannevar Bush, Walter Ong and Alvin Kernan to good effect, and has intelligent things to say about the canon, recognizing that it was never as graven in stone as either its enemies or its defenders believedin other words, the canon is really a nonissue. He writes interestingly on how the process of writing itself is altered by hypertext, on the new ease of collaboration, on intellectual property and on the new dangers of exclusiveness once a hypertext "canon" is established. But the core of the book is the meeting between Landow, the hypertextualized literary scholar, and the deconstructionist and poststructuralist theorists to whom he goes for advice and 1
confirmation. The meeting itself is like the encounter in The Silence of the Lambs between the Jody Foster character, a courageous and decent FBI trainee, and Hannibal Lecter, the brilliant insane cannibal/psychiatrist whose help she seeks in solving a mystery. The movie itself may in fact be an allegory of the experience of the humanities student entering university literary studies: to uncover the mystery of the text we go to school with its new custodians, the "cannibal lecteurs" who like to deface literature, or bite the faces off its dead authors. This may sound extreme; but it is no more so than the lecteur's selfdescriptions. Landow quotes Gregory Ulmer quoting Derrida, perhaps the greatest Lecter of them all, on the subject of the "lexia" or "mourceau," the decontextualized citation, the "bit, piece, morsel, fragment; musical composition; snack, mouthful" (and we might add "bite" or even, now, "byte") that is the meat of the critic. This "mourceau" [sic; from "mourir"?], says Derrida, "is always detached, as its name indicates and so you do not forget it, with the teeth." (9.) "Mors" is "jaw." As Ulmer says, "The organ of this new philosopheme is the mouth, the mouth that bites, chews, tastes. . . The first step of decomposition is the bite." (8.) What are these teeth that bite? Landow quotes Ulmer's explanation, that they are the "quotation marks, brackets, parentheses" (and we might add, the virgules or "slashes") so dear to the theorists and so rife among the titles of MLA papers. What is it that is bitten into pieces, into bits or bytesor, in the practice of the media, "soundbites"on this island of the Doctors Morceau? It is, as Foucault says, the individuality of the writer (74), or as Said says, the human subject (75): in other words, the author's personal expression or face. When we sup with the devil we had better use a long spoon; and when we go to school with poststructuralists, they had better be muzzled. Landow, like most of his academic contemporaries, has been insufficiently briefed about the more problematic habits of his mentors; and like the Jody Foster character he is partly responsible for enlarging them in the world where they can have more old friends for dinner. Could she not solve the mystery without being helped by Lecterand without allowing him to escape? Speaking as a poet, and as a translator of the great poet Miklos Radnoti who was murdered by the Nazis, I must confess to being an interested party, and declare my political solidarity with all victims of the lecteurs. Radnoti describes the poet's mission, at a time when the likes of Paul de Man were challenging all notions of truth, in his poem "O Ancient Prisons": O peace of ancient prisons, beautiful outdated sufferings, the poet's death, images noble and heroical, which find their audience in measured breath how far away you are. Who dares to act 2
slides into empty void. Fog drizzles down. Reality is like an urn that's cracked and cannot hold its shape; and very soon its rotten shards will shatter like a storm. What is his fate who, while he breathes, will so speak of what is in measure and in form, and only thus he teaches how to know? He would teach more. But all things fall apart. He sits and gazes, helpless at his heart. (Foamy Sky: the Major Poems of Miklos Radnoti, trans. Zsuzsanna Ozsvath and Frederick Turner, Princeton University Press, 1992) This poem could well stand for the poet's predicament today, when Foucault undermines the poet's authority, Said the poet's subjectivity, Jameson the poet's political integrity, Derrida the poet's very existence, and Rorty the notion of truth for which the poet speaks. The only thing a poet can do in his or her defense is reveal the distortions of language practiced by the enemies of poetry. Let us therefore return to Landow's thesisstatement, quoted at the beginning of this review, and look more closely at the terminology he has borrowed from his poststructuralist mentors. Let us pay especial attention to the terms he uses throughout the book for what is retrograde and oppressive: "linear," "hierarchical," and "central." A close examination will show how profoundly incoherent these terms are as they are used by literary theorists, and thus how profoundly subject to question the whole critique that is based upon them, including large areas of feminist and multicultural studies. Let us begin with "linear." As far as I can see, the word is used both by Landow and most contemporary theorists in three totally different and often contradictory senses, without any awareness of the difference or contradiction. The first sense is "like a geometrical line" with the connotations of "spatial" as opposed to temporal and processual, and "orthogonal" as opposed to curved or differently angled. Of course a geometrical line is completely reversible: it is the same from B to A as it is from A to B (whereas a temporal sequence, say March 1992 to September 1992, is by definition irreversible). Pure logic or inference, which is often called linear when it is free from selfreflexion, is similarly reversible; there should be nothing in the conclusion that is not in the premisses. Some hypertext theorists, like Nancy Kaplan, are fond of contrasting graphics (as a politically positive medium) with text; of course in this sense, the spatial as against the linguistic/auditory, graphics are much more "linear" than text.
3
The second sense is "like a causal sequence in time," with the connotations of "deterministic" as opposed to free, and "without branches or alternatives" as opposed to branched and choiced. The inductive conclusions of science, and the progress of technology, are often thought of by humanists, inaccurately, as being linear in this sense. Induction is actually a process of branching hypothesisconstruction coupled with experimental testing of the branches, and technology, far from taking a single track, always branches into more and more diverse forms in order to enlarge the marketplace. But even without the contradictions inherent in this use of the term "linear," our humanists must somehow grapple with, or repress, the profound contradictions between the first sense of the word and the second: between the timeless reversibility of the first and the temporal irreversibility of the second; between the perfect economy of the first and the entropic running down of the second; and between the endlessness and beginninglessness of the first and the ends and beginnings of the second. We might also add that those humanists who say with Jameson that "everything is 'in the last analysis' political" (32.) are subscribing to a view of history where everything that happens is nothing more than the direct result of power, a view which is profoundly linear in just this sense. The third sense of "linear," as theorists so carelessly use it, is "like a story," with connotations of "making sense," "intentional," "controlled," "closed" and "irrevocable" as opposed to aleatory, random, uncontrolled, open and revocable. What makes stories and narratives what they are is precisely the feedback relationship between later and earlier eventsearly events partly cause later events, but are partly caused by them in the sense that they are brought about by persons seeking to achieve the later event; and early events make sense in the light of later ones. This characteristic of storiestheir reflexiveness and feedbackis rather precisely definable in terms of Chaos science as "nonlinear." Stories only make sense as stories, moreover, if they contain, at least in implication, a widelybranching set of alternative futures at every moment, whose selection or rejection makes the protagonist's actions interesting and significant. Thus we may define "storylike" as "free," and thus "linear" in this sense could also mean "free," much to the confusion of the theorist. Stories have climaxes and epiphanies and irreversible peripeteias (which Chaos scientists would call bifurcations or catastrophes) in which the old levels of organization are broken through and transcended; thus we could call them revolutionary by nature, and if that is what our theorists mean by "linear," then maybe we should hang on to the linear for dear life. Stories, by their closure, their ending, and thus their removal of themselves from the future, make room for the Other in ways that no "openended" form can, and thus they are essentially more empowering to a reader, and more politically liberating. I need not describe in detail the contradictions between the third sense of "linear" and the first two; they should be obvious by now. The second set of confusions sown in the minds of our literary investigators by their cannibal mentors concerns the use of the term "hierarchy," especially when contrasted with "network." It 4
becomes clear that the word as it is used simply means "what is bad," because any other meaning it might possess turns out to be an objective property of the world, the fundamental condition of any form of organization. The more hierarchicallyorganized an organism, the more it is able to support feedbacks within its own structure, and thus the more autonomousselfrulingit is, and the more free. It is the hierarchical organization of the organs, limbs, and nervous system of an animal that give it mobility and autonomy; if we were to dismember it, it would not become more free (though indeed it would be liberated from the confining shape of its body), but less. In human institutions the most highly organized form of government is modern democracy, which has a richly complex and multileveled hierarchy of control and responsibility (unlike feudalism or despotism, which usually have only about four levels or less, or totalitarianism which has really only two!) The usual cliche which contrasts hierarchy with democracy, a cliche that Landow unthinkingly repeats (174), is thus totally in error. Tyranny hates hierarchy. Why else should the first act of any totalitarian government, right or left, be to dismantle the existing hierarchy? The difference between democracy and more primitive hierarchical systems is that democracy has achieved sufficient hierarchical complexity to include subsystems, such as the vote and a constitutional legal system, that are designed to legitimate the assignation of individuals to positions within the hierarchy, and to render it flexible, transient and porous. Thus to invite us to "abandon conceptual systems based upon ideas of center, margin, hierarchy and linearity and replace them with ones of multilinearity, nodes, links, and networks" is tantamount to an invitation to abandon mammalian and vertebrate forms of life and replace them with polyps and slime molds, or to abandon life itself and replace it with inorganic matter, or to abandon matter and replace it with energy. Landow admits as much: "the lexia. . . associates with whatever text links to it, thereby dissolving notions of the intellectual separation of one text from others in the way that some chemicals destroy the cell membrane of an organism: destroying the cell membrane destroys the cell; it kills." (53) In other words, poststructuralist theory is a sort of intellectual AIDS on the biological level, or hydrogen bomb (which turns hierarchicallyorganized matter into a simple web of energy) on the level of physics. Networks are all very well, but they are in themselves only the most primitive and inflexible form of hierarchy, one with only two levels. When they are allowed to evolve they will, as chaos theory shows, develop highly organizedand often highly flexible and responsivestructures of their own. The infant human brain is something like a network of neurons; but it immediately begins to organize itself into the Hebbcell circuits which carry thought and memory. The cells of a fetal animal or plant, at first an unhierarchical mass, differentiate themselves into an organic division of labor. Even the interacting nodeandweb ensemble of photons in the first moment of the Big Bang clustered and condensed into more hierarchical forms. In computer science, the
5
whole point of an artificial neural network is to develop a hierarchy of weights and preferences when asked to identify a given object. One of Landow's main pointsand he is not of course alone in thisis that traditional book text is somehow more hierarchical (and linear, of course) than the new hypertext medium. In fact, one could argue exactly the opposite: that hypertext, because it offers more levels of reference than traditional text, is much more hierarchical in both its invitation to the writer and in its actual organization. It sets up an immediate hierarchy, with those elements of the hypertext which make most references and are most frequently referred to at the top, those containing some references to and from other texts in the middle, and those that are singlereference dead ends at the bottom; the more references a lexia gets or makes, the more valuable it is, and the higher in the hierarchy. Of course poetry and belleslettres have always found ways, through allusion, rhyme and meter, coherency of image, allegory, echoes and foreshadowings, to deepen, hierarchize, and enrich the text; but they have always been subject to the discipline of a print format that in its physical form reduces every word to equality with every other, and must be made to organize itself through its sonic and semantic content, rather than its visual form. Anyone not prejudiced by poststructuralist shibboleths would instantly recognize Landow's fascinating reproductions of hypertext screens as being far more hierarchical visually than ordinary text. Politically, fixed book text led to egalitarian, legalistic and levelling forms of government; technologically it led to the uniformity and dehierarchizing of the assemblyline. Hypertext, then, is promising and exciting for exactly the opposite reasons from those that are usually offered; and this huge mistake, fostered by the blunders of poststructuralism, will dog hypertext until we throw it off. And here we encounter the third illusion that Landow allows himself to be persuaded by: the notion that somehow we can escape the distinction between center and margin into a "decentered" world. The claim is, of course, that the bad old hierarchical linear world of text distinguished center from margin and privileged the former over the latter, whereas the liberating revolution of hypertext breaks down the distinction and the privilege, or at least allows any element of a system to assume the role of center according to the needs of the moment. The problem is that the capacity of attention is impossible without the distinction between center and margin, and that is it is hard to imagine what value a text or a hypertext might have if not in some way to direct attention. The reason why one would buy a book or a piece of hypertext is to have one's attention directed to something that it would not have been directed to otherwise, so that a new object of importance would occupy the center of one's attention, consigning other objects to the margin. A library could not be used in a truly noncentered "democratic" fashion, for the reader would make no inegalitarian distinctions in importance between the books, the catalogs, the shelves, the readingtables, the floor and the ceiling; and the same thing goes for a 6
hypertext document. Attention is the whole work of consciousness. Living organisms evolved from the directionlessness of the amoeba, through the twodimensional radial symmetry of the starfishes and the axial directedness of the vertebrates, to the highly focussed attention of human beings, who in attending to something must necessarily, as Michael Polanyi points out, attend from everything else. The claim that a centerless and marginless document empowers the reader leads logically to the proposition that a book of blank pages, or an empty computer chassis, would do so even better, and no book or computer at all better yet. Although we might thank a person who offered us the world thus unmediated for the Zenlike wisdom of his or her minimalism, we might object to paying the twenty dollars for the blank book or the fifty for the empty disk. Nor is the case helped much by the claim that a hypertext document is temporarily centered at any given moment. So is the visual field when the eyes, without an object of attention, wander about the room. In his defense of his position, Landow quotes Richard Rorty's warning against the search for overarching or underlying truths, which Rorty stigmatizes as an attempt "to close off conversation by proposals for universal commensuration through the hypostatization of some privileged set of descriptions." (182.) But how does this plausiblesounding idea differ logically from the assertion that the leaves and twigs would grow better if they were hacked free of the single trunk that connects them? Rorty's procrustean alternatives are summed up in the assertion that it is our job "to keep the conversation going rather than to find objective truth." Again, the logic of this statement feels sound until we ask the innocent question, why we should not do both. It is only if objective truth is closed and simple that it could possibly threaten open conversation; and even so, some might prefer an honest silence to a lying discourse. But perhaps objective truth is in fact more wildly open than any artificiallymaintained conversation could ever beand this is the challenge that science currently poses to the humanities. To do Landow justice, there are signs that he himself suspects the motivations of the poststructuralists to be not be entirely trustworthy. He suggests (53), for instance, that because hypertext is a literal test of many of the principles so abstractly proposed by the theorists, that they might find it embarrassing and even threatening to their power. But Landow does not follow up on such insights, because, I believe, he feels that his mission of promulgating the literary use of hypertext is best served by an alliance with the fashionable and dominant ideology. However, I believe that a better case for literary hypertext could be made, which might go something like this. The deep intellectual goal of all cybernetic research is the emulation and improvement of the operations of the human brain. In other words, the new computer technology, including 7
hypertext, is an early study and approximation of artificial intelligence. Now it so happens that the traditional arts, including pictorial art, music, and especially writing, are already very sophisticated forms of artificial intelligenceprograms designed to be booted into and run on human brains, "meat computers" as some have cruelly called them. Under the influence of a work of literature, a reader experiences within him or herself the thoughts, feelings, and even some of the intellectual and imaginative capacities of a (perhaps dead) writer. A great writer is able to do what amounts to a "core dump" into a text, so that the reader can see the banks of the nineteenthcentury Seine or the shores of archaic Ithaca through the eyes of a past protagonist. (Perhaps it is precisely this commandeering of the reader's mental hardware that poststructuralists find so threatening, so challenging to their own little moral worldsand thus they try to bite off their authors' faces.) But if this analysis is correct, there is a very exciting role for hypertext, both as a critical tool and as a field of creative art: that is, as the convergence of traditional and contemporary artificial intelligence research. Both cybernetic intelligence theorywhich has bogged downand contemporary literary theorywhich as we have seen is mired in poststructuralist assumptions could use the new ideas that might result.
Dallas
8
Frederick Turner University of Texas at