The chapters discussed so far attempt a negotiation of the relationship between hypertext and other domains, whether the classroom or the nature of the mind. In the chapter by Moulthrop and Kaplan, in contrast, we find a totalizing drive typical of postmodern theorists bent on polarizing the evidence and creating an asymmetical system of agency in which all virtue resides in the textual phenomenon currently in view. In the present case hypertext is promoted in order to reify conceptions of print and its users, and to announce the final and inexorable triumph of textual multiplicity.

The author of the first chapter in Part III prepares the ground by appealing to a notion of self-consciousness in reading: in Afternoon, notes Johnson-Eilola, readers become aware of their reading process through the need to make choices (p. 209). This initiates us into seeing that a similar conscious benefit would flow from hypertext-based deconstructive practices for students. A critical commentary has the same weight as the literary text it is about, so in hypertext a deconstructive reading is able to merge with the text it is based on (p. 209). This facilitates the decentering of text on which Derrida insists (p. 210). Literary texts are deprived of univocal meaning, and hypertext becomes the most effective vehicle to bring this to the surface (p. 211). It is as though we are confronted with only two alternatives as readers: either complete absorption in the text (pre-hypertextual), or the totalizing deconstruction of resistance and infinite deferral that is best represented in hypertext.

These implications are worked out explicitly in the account of Moulthrop and Kaplan. They celebrate the multiplicity implicit in Borges's story, "The Garden of Forking Paths." In the ideal fiction of Ts'ui Pên described in the story, time does not exclude one event because another has happened: events occur in parallel, and the actor in the novel, in Borges's words, "chooses -- simultaneously -- all of them," so that he has "diverse futures, diverse times" (p. 229). This is an entertaining fable, but the story shows that it is not literally attainable. The protagonist Yu Tsun must decide, and he does so by committing the murder he had set out to do earlier in the story. Borges's plot places emphasis on the necessity of choice: Yu kills; he in turn is hanged. Moulthrop and Kaplan, in contrast, set out to subvert Borges's story by creating all those alternative paths left unrealized by Borges in their own hypertextual version of the story. This maneuver is an instructive one: "With hypertext," they claim, "the range of options broadens, allowing narratives that at least approximate Yu's vision of infinite pathways" (p. 229). Hypertext thus appears to evade the necessary contingencies of time. In fact, a moment's thought shows the impossibility of simultaneous pathways existing in hypertext, except potentially; an actual reader must choose, and will read events in one order or another. The maneuver thus only mimics "intertextuality, polysemy, or difference" (p. 235): the praxis of hypertext reading is necessarily linear, an unfolding in time. The formal properties of hypertext thus cannot be said to "exactly invert those of print" (p. 235): this is an illusion.

Moulthrop and Kaplan invited students to produce their own responses to the new Borges-based hypertext. They make a particular issue of one student, Karl Crary, who decided to add his response to the hypertext story itself in the form of additional nodes (p. 231). Their theoretical claim, in brief, is that although Crary attempted to offer a critique of the hypertext story, his writing was actually absorbed by the hypertext to which he contributed. "In this medium, there is no way to resist multiplicity by imposing a univocal and definitive discourse . . . In the space of hypertextual writing, anything that arises will be merged, gathered into the network of polyvalent discourses" (p. 235). This conclusion, however, does not have the theoretical force that Moulthrop and Kaplan attribute to it. Crary made two elementary errors, neither of which is critical to the understanding of hypertext. First, his categorization of the nodes into four types is clearly deficient: describing a fourth type as "complete digressions" allows Moulthrop and Kaplan scope to categorize Crary's own nodes as belonging to the hypertext. A simple reclassification would avoid this, such as "Critical commentary." Second, Crary failed to signal the presence of his own nodes (as far as one can tell) by any marks of design or style -- again, an easily corrected feature.

The implication of Moulthrop and Kaplan's argument is that any node linked to an existing hypertext becomes a part of that hypertext, which is clearly absurd. Thus, the links available within the document you are now reading or within the course web site mostly point to related documents within the same discourse domain. I have taken trouble to design each course-related document with the same look and feel so that a reader can tell what virtual space they are in. But links that point outside the course bring up documents with a different look and feel, ensuring that readers will be in no doubt that they have moved into another domain. Hypertext may at times subvert distinctions, but Moulthrop and Kaplan's account makes hypertext something completely amorphous which, if true, would mean an end to all discourse. As the old proverb puts it, The sea and the gallows refuse no one.

But Moulthrop and Kaplan, just as they deploy Crary as a pawn in their larger game, systematically violate the distinctions and properties of the domains against which they set hypertext. By denying value to pre-hypertext practices they engage in a kind of texual terrorism, and like a Star Chamber or Committee of Public Safety they distort evidence and suppress witnesses in order to precipitate the verdict they want. For example, the textbook Reading Texts is praised for denying absorption and promoting interaction during reading (p. 225), a comment that might puzzle some readers until it is realized that the only readings authorized by Moulthrop and Kaplan are self-conscious, interrogative ones. At the same time they note how the textbook "retains the 'idealist' model of the text as perfected product" (pp. 225-6), their quotation marks around idealist showing that this word is already under erasure, with one foot in the tumbril, used here only to mark our progress towards the revolutionary enlightenment of hypertext. Similarly, the textbook is said to encourage "resistant readings through critical independence," or the "strong" readings that resist "the text's seductions" (p. 226).

And just as texts must be resisted, so all earlier textual practices become reified: that is, a stereotyped account of them is placed on view for our derision. Thus we encounter such characterizations as "the realm of self-validating truths and decrees" of pre-hypertextual writing (cited p. 206); or the printed book that "creates a bias toward hegemony and monologue" (p. 227); the reading process analysed by Iser that is "internal and passive," a purely mental event which hypertext externalizes (p. 221); the previous practice of literacy that "serves the interests of individual authority, monologic discourse, and linear argument" (p. 221). Broadening their scope, Moulthrop and Kaplan also attack the education and publishing system, accusing them of conspiracy. The capital investment made by publishers in "literature" supports educational practices that are obliged to treat a text as univocal in order to preserve the investment. Since hypertext is said to oppose this "strategy of containment," it threatens "the orderly and autonomously meaningful text" with its "unvoiced assumptions" (p. 222). Moreover,

Restrictions on the creation of 'authorized' writing produce an economy of scarcity. If only a few texts survive to be disseminated, it is much easier to concentrate and control literary value, and, of course, this economy benefits those who live by study as well as those who live by sales. (p. 223)

Thus print is said to produce "conditions [that] favor singular and definitive discourse -- the production of a literature devoted to property, hierarchy, and a banking model of culture" (p. 236); this is the consequence, they repeat, of "the literary institutions we have inherited from the history of print" (pp. 236-7).

These and similar arguments are put forward with no acknowledgement of the wide range of values associated with several centuries of book-related practices. They totalize by decree, annihilating all previous values in favour of hypertext multiplicity: they act in the academic sphere somewhat as the Jacobins in the political, who made a point of executing intellectuals such as Lavater, or as the Khmer Rouge, who murdered anyone who had a higher education. Fortunately, academic arguments of this kind are paper tigers: their cavalier treatment of the history of reading means that they can have little influence. The prospects for hypertext are both more problematic and more interesting than proponents like these seem willing to understand.