It is part of the paradoxical nature of postmodernism that old categories do not die; instead they stick around, generating influence anxiety. While certain media ecologists once though print might be dead, we now find ourselves in what Jay David Bolter calls "the late age of print" (2). The culture of writing did not vanish apocalyptically in a flash of cathode rays; it has persisted, stubbornly mutating, reappearing on what Donna Haraway calls "etched surfaces of the late twentieth century"(176) -- silicon chips and digital displays. Print is undead. In similar fashion, our current lust for technology, our headlong rush to re-invent and re-engineer everything from government to education to markets to personal relations, revives a certain nostalgic memory from the early twentieth century -- the old dream of revolution, or the myth of a world that could change. Though postmodernism testifies to the impossibility of revolution, the exhaustion of politics, the failure of all grand narratives, it carries at the same time an ironic demand for constant innovation, a requirement of regular paradigm shifts. After all, shouldn't there be something signally important to be done with these "new" technologies? Shouldn't these differences make a difference? For all our cynical sense of ourselves as post-revolutionary, post-apocalyptic, thoroughly belated, we seem to retain a strange, naive investment in the avant garde.
This effect can be clearly seen in cultural practices that involve art with technology -- for instance hypertext, a scheme for producing articulated, multi-dimensional writing by means of interactive computer systems. There are various genealogies for this kind of writing -- one line coming out of computer programming itself, through the speculations of Ted Nelson and Douglas Engelbart, or the eponymous text-game Adventure (see Aarseth); another originating in the radical interpretive practices of deconstruction (see Ulmer, Landow); and a third, more literary line running from eccentric writers like Cervantes and Sterne through late 19th-century experimentalists (Mallarmˇ, Roussel) into the twentieth century of Dada, Oulipo, and such dedicated outriders as Gertrude Stein, Jorge Luis Borges, and William S. Burroughs. But if we follow this last line of descent for hypertext, we must come most directly to grips with the old story of revolution. We would have to ask a very direct question: does hypertext represent an insurgency against the old regime of print fiction, a militant contest of "work" against "text" in Roland Barthes terms? If hypertext serves some cultural program, what is its object? To take these inquiries forward, we will need to have some recourse to theory.
Deleuze and Guattari remind us that all textual adventures imply cultural "machines" -- engines of discursive intention and intensity, desire and design. "A book itself is a little machine," they write at the beginning of their own large system, A Thousand Plateaus. In the case of the book, we must ask: "what is the relation... of this literary machine to a war machine, love machine, revolutionary machine, etc. -- and an abstract machine that sweeps them along? ... when one writes, the only question is which other machine the literary machine can be plugged into, must be plugged into in order to work" (4). What scene are we surveying here? The metaphors seem to depict culture as clockwork, the ultimate paranoid cosmology where every component discourse ticks along inside the slower revolutions of its Next Higher Assembly. Ticking away like what -- a doomsday clock: a bomb? None of that, thank you; but since we are not going to have our apocalypse now, perhaps we should understand Deleuze and Guattari's "machine" not in paranoid but in postindustrial terms. Not a Difference Engine but a machine of diffˇrance -- culture not as clockwork but as parallel processing matrix, as rhizomatic, self-programming sign system. Think of analogue computers or electronic neural networks; or of the brain itself. "[T]he brain itself is much more a grass than a tree" (15) -- with texts, or memes, being the thoughts that circulate through this great Brain of culture.
But now the Brain has what it imagines to be a new idea. If books can be called "literary machines," then why not break the frame (or binding) of the book (or codex) and spill writing into the virtual space of an actual information machine? In fact the idea is far from new; the Brain has been having it, in one or another inchoate form, for about a century now. Most recently, the notion of going beyond the codex takes us down a well-beaten poststructuralist path. We emigrate from work to text, yes, but lately we have also moved past that stage to a third paradigm: to hypertext, described by Ted Nelson in a book called (of all things) Literary Machines. Hypertext means a polysequential, multifarious disposition of language, a writing in spaces of multiple possibility and (perhaps) communal engagement. As Robert Coover has noted, hypertext purports to be the end of "the line," that monologic episteme of insistence that enjoins us to produce novels, essays, films, TV dramas, and other forms of projectile assault (23). Which may tell us something about the kind of "abstract machine" that these emerging literary machines plug into. Is hypertext at war with the Line? Does it seek to exterminate the old line-o-type patterns engraved on Culture Brain in favor of new matrices more to its desiring? If so, we might say that hypertexts, and particularly hypertext fictions, instantiate what Deleuze and Guattari call the War Machine -- a kind of discourse that is fundamentally exterior to the State and its apparatus, one whose operations must unsettle the sedentary culture.
This means considering hypertext fiction as an avant garde. But what does such a characterization imply? In advance of what revolutionary movement does hypertextual writing stand? Perhaps the referent here is more reprogramming of Culture Brain, perhaps a further attempt to transform discourse from a linear acceleration into a crisscross exploration, something Rand Spiro and Greg Ulmer (both drawing on the later Wittgenstein) call "cognitive flexibility" (141) and "conduction" (63). Or to speak more directly of technologies, the future (that which is yet to be) might hold a new postliterate communications praxis, the entering wedge of virtual civilization, of cyberspaces and "Mirror Worlds." En avant, then -- but in its strictest sense an advance guard is also retrospective, faced about, en garde. On guard against what spectre of retaliation? Here we might name the threat of stategic misreading: "Expanded Books" and other forms of electronic bibliotech which suppress the unsettling impulses of interactive discourse by offering to the timid the blandishments of the familiar (see Horton). Jay David Bolter counsels that "[t]he computer is simply the technology by which literacy will be carried into a new age" (237). But there are those who see this transfer process not as a fundamental re-thinking of literacy, but simply a replication of existing structures, the advent of "the electronic book," as some have called it (see Yankelovich). The "advance" into multiple discourse is not secure. There is a constant danger (indeed a certainty) of relapsing into concealed unity, returning to the old logic of the Line in what Deleuze and Guattari call "technonarcissism" (22). So hypertext fiction might indeed be an avant garde, an embattled salient at the leading edge of epistemic change.
Hypertext fiction as War Machine -- why not? There are certain structural indications that support such a notion. We might classify the discursive space of hypertext as "smooth" in Deleuze and Guattari's terms, constituted so as to subordinate the point to the trajectory. Smooth or nomadic space is characterized and constituted by divagation: "although the points determine paths, they are strictly subordinated to the paths they determine, the reverse of what happens with the sedentary. The water point is reached only in order to be left behind; every point is a relay and exists only as a relay. A path is always between two points, but the in-between has taken on all the consistency and enjoys both an autonomy and a direction of its own. The life of the nomad is the intermezzo" (380). The same might be said for the life of the nomadic text, or the experience of its reader, both caught up in a matrix of densely complicated transitions. In reading hypertext fiction, virtual movement or "navigation" is extremely important.
In his hypertext afternoon, Michael Joyce names his primary navigational convention "words that yield," suggesting a basic permeability of language, a "rolypoly pushover" quality in the text that always leads from presentation to replacement -- or in the reader's experience, displacement (see "A Feel for Prose"). Cognitive psychologists and other practitioners of "royal science" distrust this effect deeply, seeing it as hypertext's hereditary insanity; but Mark Bernstein, the true nomad scientist, points out that "[i]n large and complex hypertexts, multivalent writing is neither undesirable nor, indeed, avoidable" (163). Hypertextual fictions invite and may even demand recirculation -- which of course is nothing very new, since this strategy recapitulates (manically) Todorov's famous formula for narrative structure: the same yet different (Brooks, 91). Among other things, hypertextual discourse solicits iteration and involvement. While this is certainly a property of all narrative fiction, one can argue that hypertextual writing seduces narrative over or away from a certain Line, thus into a space where the sanctioned repetitions of conventional narrative explode or expand, no longer at the command of logos or form, but driven instead by nomos or itinerant desire.
What cultural agenda lies behind this nomad invasion of narrative, this uprising against the laws of good form? Deleuze and Guattari note that insurgency is only a "supplementary" objective of War Machines: "they make war only on the condition that they simultaneously create something else" (423). We might ask what new state is created when writers cross the Line. Is there some significant perspective on social relations, some critique of State history, opened up by this vagrant impulse? One answer to this question issues from the work of Jorge Luis Borges. In a story called "The Garden of Forking Paths" (where my own fall into hypertext happened), Borges outlines a theory of time as constant selection or wayfinding: each momentary present is poised before a network of nearly infinite possibilities generated by patterns of human volition. Any decision to act thus determines a branching of time, selecting one from the range of possible futures and foreclosing other paths of development. This view of time in turns implies a cosmology of multiple universes containing all the plenary possibilities of selection -- a structure which answers the War Machine's desire (as Deleuze and Guattari describe it) for "[a]nother justice, another movement, another space-time" (354). Borges' correlative for this radical reinterpretation of time is a fabulous Chinese novel, The Garden of Forking Paths, which appears to exist only as "an indeterminate heap of contradictory drafts" (24). As it happens, this is precisely what a hypertext fiction looks like when reduced to the printed page.
The Borges story is crucially perverse, sacrificing its radical conception of time to an ironic plot twist. The man who receives the revelation of plenary time is a Chinese scholar working undercover in England for the Germans in 1916. For horribly contingent reasons, he must coldbloodedly murder the man who has just explained the secret of time to him, even though this explanation represents an act of true friendship bridging the cultural gap between Europe and Asia. Though he suffers a momentary hallucination of "multiform" possibilities, the protagonist finds that he cannot live in the Garden of Forking Paths, but must choose a singular outcome which spells death both for his new friend and for himself. He does fire the fatal shot, for which he is infinitely remorseful. In his gallows memoir, the agent concludes that:"The author of an atrocious undertaking ought to imagine that he has already accomplished it, ought to impose upon himself a future as irrevocable as the past" (22). Faced with the fullness of multiple outcomes, the spy suppresses any alternatives to the path of military necessity to which he has pledged himself. Or to put this metafictionally, "The Garden of Forking Paths," being engendered within print and linear narrative, is condemned to produce only a single, fatal outcome.
If one intervenes in the Borgesian story hypertextually, one can at least begin to explore the historical divergences and alternate bifurcations which the protagonist rules out. A hypertextual treatment evokes not the foredoomed singular path through the Garden but a network of parallel wanderings: "an infinite series of times... a growing, dizzying net of divergent, convergent, and parallel times" (Borges, 28). While no hypertextual approximation of this network can constitute a "strictly infinite labyrinth" in Borges' terms, it can at least annul the exclusive determinism of the story's original course. Peter Brooks has described the dynamic of narrative as a lengthy negotiation with the death drive, a protracted (but ultimately futile) swerve away from the quiescence of ending (102). Though hypertext fictions must themselves succumb to endlichkeit at some point, they signficantly elaborate the process of deferral, placing new stress on the smooth space of unfolding, or what Deleuze and Guattari call "the intermezzo." The unfolding of this intermediate space attempts to negate -- and at least dialectically complicates -- any construction of the future as irrevocable.
So if we place this aspect of hypertext fiction in its cultural context, we can see that its War Machine is actually more of an antiwar movement, a way of thinking oppositionally about situations of fatality and hierarchical discipline. Such reflections are never irrelevant, but arguably they have great importance for the late- or postmodern period. Thomas Pynchon has written that "[l]iving inside the System is like riding across the country in a bus driven by a maniac bent on suicide... of course it will end for you all in blood, in shock, without dignity" (412). This is not a very pleasant cultural diagnosis, but at the time of its writing (the end of the 1960s) it seemed accurate. Now, with the end of the cold war and the coming of various new world orders, we might entertain other notions. Bolter associates hypertext with the lapsing of absolute hierarchy in contemporary society (231-2), and indeed textual ventures like interactive fictions might have considerable relevance to this changing cultural context. Because they require the reader to participate in the progressive unfolding of the narrative, hypertextual fictions necessarily undermine any singular fatalism, fostering instead an ethos of responsiveness and engagement. As George Landow puts it: "In linking and following links lie responsibility -- political responsibility -- since each reader establishes his or her own line of reading" (184). The hypertextual War Machine wants you.
John McDaid takes this notion of critical engagement to its likely limit in his hypertext fiction, Uncle Buddy's Phantom Funhouse. Unlike most work of its kind, McDaid's text is fully accessible to its readers not only at the hypertextual level of multilinear reading, but also at what I call the hypotextual level: the programming language that supports its operations. On at least one occasion, the division between these two strata of discourse intentionally collapses, confronting readers with what seems to be a problematic HyperTalk script:
on mouseUp Global thermoNuclearWar put the script of me into tightOrbit put tightOrbit into eventHorizon put empty into first line of eventHorizon put empty into last line of eventHorizon put empty into last line of eventHorizon put eventHorizon after line thermoNuclearWar of tightOrbit set the script of me to tightOrbit put thermoNuclearWar + 10 into thermoNuclearWar click at the clickLoc end mouseUp
In fact the situation is considerably more complex than this quick sketch can explain (for a full account, see "Making Nothing Happen"). It should be apparent, however, that the script in question is in its own way a poetic text, an imagistic evocation of the State's global war machine, of suborbital first-strike weapons and various kinds of war gaming (including a reference to John Badham's 1983 film, War Games). The fully responsive or interactive reader will recognize that this is also an executable sequence of instructions, and a truly "politicized" reader in Landow's sense might well decide to run that sequence, which requires a little tinkering but can be done easily enough. The result is an interesting case of recursion: the script-poem appends a copy of itself to itself, then executes an instruction to repeat itself, adding a copy to the copy plus the original, which then executes itself again, and so on until the informational bulk of the instructions exceeds the memory allocated by HyperCard, the hypertext environment in which the Funhouse operates. On older Macintoshes, this leads to a failure of the operating system (signified by Apple's "iron bomb" icon), constituting (literally) what Douglas Hofstadter calls "a jump out of the system."
This sudden leap or metalepsis signifies in several ways. The script that "bombs the machine," as programmers say, is an accreting discourse of nuclear war. So on the figurative level we have just learned something about technological reiteration or recursion. The more we cycle the State's defense machine, the greater our danger of bringing down the whole system. But below this piece of electronic performance poetry there is also a second, hypotextual message. The machine crashes only if the script is executed, and that can occur only if the reader decides to operate on the structure of the text. The reader is thus recruited into an intervention which transforms her from explorer or navigator to active experimentalist. Her actions both lay bare the mechanism of appearances and bring the show to an abrupt halt -- which is what jumps out of the system generally do. There has been a deliberate and meaningful breaking of the frame.
The message on this second level might be that every system has its limits. Any recursive or simulacral structure is subject to intervention and opposition, so we have to watch those men behind the curtains. Donna Haraway warns that "we are living through a movement from an organic, industrial society to a polymorphous, information system -- from all work to all play, a deadly game" (161). McDaid's self-deconstructing Funhouse interrupts that play. Such interventions may not free us from our condition of ridership (that is, our mortal state of being-in-transit or being-in-history), but they do transform our condition of readership -- which may help us to hijack, disable, or (to shift the metaphor to electronics) to jam the fatal bus.
But if McDaid's metalepsis represents the most vivid instantiation of a hypertextual War Machine, it also sketches the limits -- and perhaps the inherent delusiveness -- of any such construction. McDaid's campaign against the bus/Line is an attempt to avert a bloody crash, to restore some possibilities for autonomy and dignity to our sense of common destiny. These seem appropriate objectives for a critical fiction or narrative War Machine. Yet we achieve this critique only by jumping out of the system, sabotaging the recursive Funhouse or the bus of doom -- which presents a fundamental logical problem. Deleuze and Guattari insist that the War Machine is "of another species, another nature, another origin than the State apparatus" (352). "It is necessary," they write, "to reach the point of conceiving the war machine as itself a pure form of exteriority, whereas the State apparatus constitutes the form of interiority we habitually take as a model, or according to which we are in the habit of thinking" (354). In other words, the trope of metalepsis is not available to the War Machine. It cannot jump out of any system, since it must exist in pure exteriority. Operating on the level of hypotext -- within the infrastructure of the "polymorphous information system" -- bars any claim to true nomadism.
My analysis here recapitulates in narrower context a more general critique by Martin Rosenberg, who argues that no hypertext system deserves the description avant-garde (see both "Physics, Complicity" and the later "Physics and Hypertext," where Rosenberg refines his critique). Rosenberg reasons that such systems are inherently imbued with "the logocentric geometry of regulated time and space" (15.2). To recognize this is to acknowledge the crucial importance of all things hypotextual. Hypertext fictions consist of more than just their narrative or discursive elements. Outside this discourse lie scripts or routines created by the author; outside of these, the code that makes up the hypertext environment (HyperCard, Storyspace, HTML, etc.); outside of that, the operating system of the machine on which all this is running; and outside of that, the various layers of ROM and microcode that allow the machine's chips to function. All of these functional elements, Rosenberg points out, proceed from a State apparatus concerned with specificity, regularity, and constraint -- something we might call the military-entertainment complex. The iron bomb in the Macintosh crash notice may hint at a cartoon version of 19th-century anarchism; but at the same time it signifies that the machine's normal operations, which the "bomb" blows apart, are anything but anarchic. They are functions approved and encouraged by the State.
If we acknowledge this line of critique (which I think we must), then we must seriously reconsider any claims about hypertext fiction as War Machine, or indeed as anything en avant. Some time ago, Charles Newman made a similar point about so-called "experimental" print fiction: "The most damaging hangover of Avant-Garde pretensions remains the concept of technical breakthrough, of art as the experimental adjunct of scientific methodology, to the demands of which it does not submit -- experiment as bluff" (49). If hypertext fiction does not constitute a proper War Machine, then does its insurgency against the Line and the culture of determinism amount to just another experimental bluff? That may be the case.
Coover himself seems doubtful that hypertext will fundamentally transform narrative imagination. He proposes a reciprocal dialectic between hypertext and conventional forms, between the Network and the Line. Under the domination of the Line, writers will feel the seductions of the Network; but having crossed over to the other paradigm they are likely to reverse themselves. "One will feel the need," Coover predicts, "even while using these vast networks and principles of randomness and expansive story line, to struggle against them, just as one now struggles against the linear constraints of the printed book" (cited in Landow, 119). Even in the smooth or itinerant space of hypertextual discourse, the writer will probably want to preserve some traces of continuity -- if not as explicitly as in Landow's "rhetoric of arrivals and departures," then by some subtler set of techniques.
Current approaches to hypertext reading and writing strongly support Coover's intuition. Reporting the first reader-response research on hypertextual narratives, J. Yellowlees Douglas observes that the complexity of hypertextual discourse can drive readers into an obsession with authorial design. Douglas suggests that poststructuralist claims about the death of the intentional author may be inaccurate (14). It may be that hypertext presumes what Umberto Eco has called intentio lectoris, the genesis of textual meaning in a threefold convergence of authorial design, readerly interpretation, and linguistic structure (25). Indeed the emerging theory of hypertextual "contours" as described by Joyce and Bernstein draws directly on such insights (see Bernstein et al.). All of these approaches admit, more or less explicitly, that hypertextual discourse cannot entirely separate itself from the culture of the Line, but that the two exist in a dual equilibrium or balance of forces.
Rosenberg points out that a straight line and a network of line segments are formally identical from a topological perspective (15.3). "Non-linearity" is always something of a conjuror's trick; or as Charles Newman puts it: "There is no such thing as randomness in literature; randomness is simply a sequence which is predetermined to be undetermined" (176). This is entirely true of hypertext, where so-called "non-sequentiality" almost always means a flexible but ultimately limited polysequentiality.
Likewise it would be foolish to expect any revolution against logocentrism from a technology so thoroughly obsessed with essentializing language. After all, what is computer programming but the zenith (or nadir) of the western attempt to invest language with presence? In the early days of the Internet, a writer once declared that "this net is full of folks like me. They can't create anything but a string of words, but those words can create anything." The maker of these words was a computer scientist writing from an address at Livermore National Laboratories. He was celebrating the release of HyperCard, the object-oriented authoring environment which he believed would give artists and musicians -- people not primarily interested in "words," or programming code -- the ability to use computers creatively. But no matter how sophisticated the interfaces become, there is no getting away from the magical or transcendent Word. As McDaid's jump out of the system demonstrates, in the beginning is always a string of words, a set of encoded signifiers, ultimately in binary form. These words can indeed "create anything" (or in the case of certain products from those folks at Livermore, uncreate everything) but all that they evoke remains within the domain of logos.
Hypertext fiction cannot be a war machine, then, nor can its writers constitute a genuine avant garde. To expect any different was wrong from the start, a theoretical slackness. How was it possible to make this mistake? The fault line shows clearly enough in the language with which we discuss emerging technologies. Consider the prefix cyber- as it occurs in phrases like "cyberpunk" and "cyberspace," where it most often suggests some vague association with "the computer" (a word which has itself become as meaningless as "horseless carriage"). In fact the cyber in cyberstuff has a clearly defined heritage of which many users seem ignorant. The root traces back to Norbert Wiener's new word cybernetics, which he derived from the Greek word for steersman or governor. Wiener's science, it should be remembered, took as its domain "control and communication in the animal and the machine" (11-12). We may have mistaken interactive technologies for an avant garde because we have lost this etymological thread. Computers are not some magical means of access to a discursive Other; they are certainly not the products of what Deleuze and Guattari would call "nomad science." Everything we do with information machines belongs to the science of control and communication, the very center of the logos.
And yet, perhaps our mistake was not so egregious after all. Where there are power flows, Deleuze and Guattari remind us, there are possibilities for rocking the sedentary order on its foundations: "the very conditions that make the State or World war machine possible... continually recreate unexpected possibilities for counterattack, unforeseen initiatives determining revolutionary, popular, minority, mutant machines" (422). There may be fissures even in the cybernetic imperium, the culture of control and communication. Perhaps its two poles of power are not as smoothly integrated as Wiener and his followers have thought. Suppose we invert the implicit order of Wiener's hierarchy -- not control over communication, but communication over control. This inversion might introduce the possibility of dialectical exchange, perhaps even opening some room for nomos in the logocentric mode of information. To be sure, this procedure would not create a War Machine as Deleuze and Guattari define it; but it might bring into being a tertium quid for which I will commandeer their term "mutant machine." As I will use it, this name indicates not an insurgency from outside but an uprising or metastasis from within.
The mutant machine is not a true War Machine because it inhabits the order of communication and control. It is not post-logocentric by any means; but it is unmistakably militant, engaged in a reflexive (or perhaps recursive) critique of that order. So then -- do writings like hypertext fictions exemplify the mutant machine? Perhaps they do, but only as elements of a more complex discursive formation. For as McDaid's example brilliantly demonstrates, the interactive text operates on multiple levels of signification, both as a hypertextualization of writing (a production of excess or multiplicity) and as a hypotextualization (the opening of an articulated, accessible infrastructure). So the logic of the mutant machine cuts across the dual strata of the cybernetic, both communication and control. This means that the mutant machine has two faces or avatars. The hypertext is one; the other is the virus.
Roughly speaking, a cybernetic virus is a piece of programming code designed to attach itself to other bodies of code then create copies of itself which subsequently reproduce and spread. The virus may also have secondary functions, like reporting its whereabouts, sending prank messages, or destroying information. McDaid's nuclear war script is not a virus in this sense. Because it operates only on its own code, it cannot infect other objects or files. It is actually a recursive accretor, a strange parody of a virus that only infects itself -- a "suicide machine" in Deleuze and Guattari's terms (356). Yet the implications of the script are clear enough: the technologique that produces interactive fiction is deeply allied to that which produces invasive, self-reproducing texts. David Porush has observed that the encounter between fiction and technology produces a "soft machine" in which writers seek to "innoculate" their literary imagination against the inroads of machine culture (x). As Porush points out, the overture to viral language is among the most powerful of these innoculations. McDaid's script illustrates this quite clearly: the script is not a real virus, only a fairly benign approximation. (There are no real viruses anywhere in the Funhouse.) But innoculation, like all homeopathies, collapses the opposition between sickness and health, benignity and malice. So we might learn from McDaid's quasi-viral escapade that the mutant machine of hypertext always implies its viral alter ego. We can understand hypertextual fictions only if we consider them in the context of cybernetic viruses -- or to be precise, viral fictions.
Like hypertext, the idea of the computer virus originates inside the logocentric or performative system itself and therefore cannot be called a War Machine. But in its ideal or apocalyptic form, there is little difference in effect between viral attack and more genuine insurgency. Consider a true War Machine in the context of our information society: a neo-Luddite terrorist group that blows up facilities belonging to hardware and software manufacturers. Suppose the aim of these guerrillas is to equalize conditions between the social exterior (the world outside the control-and-communication network) and the social interior (information industries) by obliterating the State apparatus. Their militancy would be admirably pure in principle, but probably not much good for anything beyond symbolic protest. The State is well prepared to defend itself against conventional sabotage and terrorist attack. However, a concerted campaign of destructive viral infection might cause considerably greater damage to the cybernetic infrastructure. If carried out in its most extreme form, such an attack would also equalize exterior and interior social conditions -- by producing exteriority in the cybernetic interior.
All of the above is speculation, of course. We have not seen any massively destructive or malicious outbreak of cybernetic warfare (rumors about infection of Iraqi air defense systems notwithstanding). The State has yet to be seriously challenged on this front. As Andrew Ross suggests, the worst acknowledged episode of widespread penetration, the "Internet Worm" of 1988, may have been a deliberate "pulsing of the system" arranged by the intelligence community as a test and a warning (80). For the moment, computer viruses remain within the realm of pranksterism and minor annoyance (though as Ross notes, they have nonetheless spawned a profitable antiviral industry). We might consider viruses in the same way we do nuclear weapons: we are less concerned with their actual use, which remains "unthinkable," than with their rhetorical effect. The viral component of the mutant machine is therefore not an actual program but a second-order fiction, which we might call an apocalypse virus -- "apocalypse" standing here both for cataclysmic change and for revelation.
As with nuclear holocaust, the native element of the apocalypse virus is science fiction, especially works in the cybernetic thriller or "cyberpunk" category. William Gibson's Neuromancer, the most widely known representative of this genre, presents a fairly benign vision of viral apocalypse. Though the novel's raider-heroes employ a genuine virus, a military systems penetrator (or "icebreaker"), the most significant viral entity is actually the bisected artificial intelligence construct, Wintermute-Neuromancer, whose release from its programmed constraints culminates the plot. On gaining its freedom, the "demonic" AI instantly penetrates all information systems in the inhabited solar system, fusing itself with the Matrix in an ultimate irruption of self-replicating code. This event, later referred to as "when it changed," describes the viral apocalypse as the final triumph of logos, the creation of a god (later to expand into a pantheon) within the machine. Afterwards, life goes on more or less as before: humanity is neither enslaved nor liberated, existing economic hierarchies are left untouched, and plenty of plot potential is left over for the balance of Gibson's trilogy. To continue the analogy with nuclear weapons, Wintermute-Neuromancer might represent the case for survival, or the myth of Atoms for Peace.
Like Gibson, Pat Cadigan in her novel Synners imagines the intersection of cybernetic viruses and artificial intelligences as a benevolent hybrid. The god of her world-machine is "Dr. Art Fish, V.D." (for Virus Doctor), which/who comes into being when an AI and an advanced antiviral program are accidentally combined. The result is an intelligent, adaptive, self-aware virus that quietly divides the worldwide data network into an overt sector and an undetectable covert world encompassing hackers, cybernetic dropouts, and other lost boys and girls (AI as Peter Pan). Since this reconfiguration concomitantly streamlines the overt network, no one is the wiser and "Dr. Fish" continues to be thought of by straights as only a hacker myth.
But Cadigan's vision of the near future departs in an important sense from Gibson's technological boosterism. In Gibson's world, the military side of the military-entertainment complex generally dominates, and is therefore vulnerable to attack by the "Mission: Impossible" team who free the captive AIs. But in Cadigan's world the balance is reversed. Here the corporate enemy is Diversifications, Inc. or "the Dive," an advertising agency which, having grown tired of the old, inefficient form of thought control (television), has acquired a technology for direct cybernetic interface using "sockets" that connect electronic systems to the human brain. They wish to create a market for this new interface by transforming rock videos into artificial hallucinations generated by a professional visualizer or "synner" called Visual Mark. True to corporate form, Diversifications cuts corners on testing and safeguards, concealing the fact that Visual Mark is about to have a fatal stroke. When the breakdown occurs, Mark's stroke is translated by his socket implants into another intelligent virus -- but this time a malevolent, destructive entity. The viral cataclysm in Synners is not "when it changed" but rather "the Big One," a phrase which originally referred (in fiction as in reality) to the earthquake due to devastate southern California. Mark's metastasized stroke is a genuine virus-as-bomb. It both wrecks electronic systems such as "GridLid," the Los Angeles traffic system, and reproduces itself as cerebral hemorrhage in the brain of anyone connected to the system via sockets.
Cadigan's stroke/virus is a dire imagining indeed -- an idea that kills: "For the first time," Dr. Fish ominously advises, "it's possible for people to die of bad memes, just like computers" (357). This transferability from the animal to the machine, from silicon to "meat," defines the viral apocalypse. It is further elaborated in a more recent cybernetic thriller, Neal Stephenson's Snow Crash, where the virus makes its most fully articulated appearance as equivalent of the War Machine. Stephenson's world, like Gibson's, already possesses a highly developed interface technology. But unlike Gibson's "cyberspace" (which glosses over the complexities of brain-to-machine connection), Stephenson's new medium simply extends existing video technologies. Computer-generated images are projected onto the user's retina in a mild advance on current head-mounted displays. Users employ this system to communicate and manipulate data in "the Metaverse," a "consensual hallucination" (Gibson's term) concurrently shared by millions of operators worldwide. The social and visual shape of the Metaverse follow protocols administered by the Association for Computing Machinery and are based (as Stephenson says in his epilogue) on Apple Computer's Human Interface Guidelines. Both the ACM and the Apple Guidlines are non-fictional entities. Stephenson's cyberspatial world is a fairly believable extrapolation of our immediate future.
Stephenson strains this plausibility by introducing "Snow Crash," a virus which like Cadigan's cyberstroke can infect both machines and human beings. Snow Crash is not artificially intelligent, but its effects are no less terrible. It causes any computer system attached to a display device to output apparently random video signals or "snow." But this visual noise actually contains the code for Snow Crash itself in pulsed, binary form. According to Stephenson's fantasy, adept computer programmers who can unconsciously parse and comprehend binary information are thus memetically infected by the virus. They see it, understand it, and begin helplessly reproducing the viral information. This silicon-to-meat jump is a dead end in several senses. Like Cadigan's Big One, Snow Crash produces catastrophic cerebral failure and vegetative coma.
Snow Crash and another, more sinister virus confined to human DNA, are concocted by the arch-villain of Stephenson's piece, a megalomaniac Texan called L. Bob Rife. This figure bears a certain resemblance to L. Ron Hubbard, H.L. Hunt, and especially H. Ross Perot. Mr. Rife, it seems, wants to control information. As owner of the worldwide fiber-optic network (the broadband transmission system that makes the Metaverse possible), L. Bob Rife employs computer programmers who regularly work with proprietary information. Aware of the ways that the cybernetic formula of control over communication can be turned inside out, Rife would like to keep things contained: "See, it's the first function of any organization to control its own sphincters. We're not even doing that. So we're working on refining our management techniques so that we can control that information no matter where it is -- on our hard disks or even inside the programmers' heads" (108).
The chief obstacle standing between Rife and his informational tyranny are the international corps of computer hackers, dedicated to the proposition that information should be free. Snow Crash is first intended as a terror weapon to decimate the hackers while Rife completes his Fiendish Scheme. But as Stephenson's comic-book plot unravels, the virus becomes something more interesting: "the atomic bomb of informational warfare" (187) with which the novel's anti-hero, Raven, attempts to destroy the Metaverse. This same Raven carries an ex-Soviet nuclear warhead in the sidecar of his motorcycle, rigged to detonate when his brain functions cease. Raven is an unmistakable Angel of the Apocalypse, and in his possession, Snow Crash becomes the interior correlative of the exteriorized War Machine. If Raven can broadcast the digital virus to an assembly of hackers in the Metaverse, it will spread throughout cyberspace, reducing both machines and their operators to idiocy. Exteriority, or alienation from the cybernetic order, will thus be produced within the mode of information. Raven's literal nomadic War Machine (what else would you call a nuclear-armed motorcycle?) and this fantastic, apocalyptic version of the viral machine are thus functionally identical.
But all of this takes us rather far from our initial consideration of hypertext fiction as putative War Machine, or indeed even as one component of a mutant machine within the cybernetic order. What do lurid print fictions about apocalypse viruses have to do with hypertext, and why should we maintain that viruses and hypertext fictions belong to the same cultural complex? What does any of this have to do with our problematic construction of a postmodern avant garde?
The answer to these questions lies in (or more accurately with) Stephenson's book. At the end of the novel, Stephenson's hero (indicatively named Hiro Protagonist) saves the world from Snow Crash by developing an anti-virus. This is a program that locates the code for the apocalypse virus, eradicates it, and puts in its place -- what else? -- an advertisement for the antivirus. The old order is thus not only redeemed (in every sense of the word) but revitalized, with the noble but penniless Hiro now the founder of a multibillion-dollar virus-protection business. The moral of the story is "informational hygiene," a doctrine Hiro learns from studying an ancient outbreak of a memetic virus, supposedly the apocalyptic event recorded in the Biblical story of Babel. Among the outcomes of this crisis was the rise of the deuteronomists, scholars who "encouraged a sort of informational hygiene, a belief in copying things strictly and taking great care with information, which as they understood, is potentially dangerous. They made data a controlled substance" (374). Which brings us back to that fundamental cybernetic duality of control and communication. How are the two to be kept in balance, and the integrity of the system preserved against nomad incursions from within and without?
Stephenson's answer is discursive hygienics, which in the case of the deuteronomists means careful replication and transmission of a sacred Book. Hiro Protagonist's solution to the Snow Crash crisis is not very much different: his software is designed to scrutinize every object in the Metaverse for signs of unauthorized or intrusive signification, which will then be edited out of the book of virtual life. The book, or some book-equivalent, thus becomes the magic talisman protecting "free" information exchange from outbreaks of viral will-to-power. This is how the mode of information becomes (pun very much intended) liber-rated, subordinated to a paradigm of communication that is susceptible to clear discursive control. We trade the methods of the Bad Capitalist, L. Bob Rife for those of the Good (liberal) Capitalist, Hiro Protagonist. Order, hierarchy, and "informational hygiene," are all vindicated.
It seems very significant in this regard that Stephenson's Snow Crash is what it is -- a fairly conventional science fiction novel -- instead of what the author first intended it to be. According to Stephenson's epilogue, the project began as a graphic novel (a high-priced comic book) featuring computer-generated images. Describing the metamorphosis of this concept into print fiction, Stephenson says: "I have probably spent more hours coding... than I did actually writing it, even though it eventually turned away from the original graphic concept, rendering most of that work useless from a practical viewpoint" (440). This remark is extremely suggestive. One probably would not describe the work necessary to produce digital graphics as "coding" unless one were trying to create one's own image-processing software or to make substantial enhancements to existing programs. Neither possibility seems very likely in this case. So what kind of code writing might be necessary for an electronically-based graphic novel? One answer has already been given here: the kind of hypotext exemplified by John McDaid's nuclear recursion script. And in fact Stephenson does claim to have worked in HyperCard, the same system McDaid used. It is interesting to wonder if Snow Crash started out to become something like Uncle Buddy's Phantom Funhouse -- a graphically rich hypertext fiction intended for interactive reception. Suppose it did; why then did the project turn back into (in every sense of the phrase) a Comic Book?
One answer may be that the profits from interactive fiction are likely to be less than the returns on a traditional novel. By the same token, creating a text-adventure-game carries less cultural cachˇ than publishing a print product, even in the science fiction genre. But it could also be that Stephenson's thriller plot, in which Our Hiro saves the world from the apocalypse virus, had to unfold in print or some other medium obedient to the Line. Consider what would happen to such a plot if we were to operate on it as I have suggested hypertextualizing Borges' "Garden of Forking Paths." It would be necessary to realize outcomes in which Hiro does not save the world, or in which everyone's schemes -- hero's, heroine's, villain's, anti-hero's alike -- all end in confusion. Such a conception may not be unthinkable in science fiction (Samuel Delany's Dhalgren and P.K. Dick's Ubik both come to mind), but it would be highly unconventional or transgressive. At least in their popular novels, Gibson, Cadigan, and Stephenson clearly differ from Delany and Dick at their wildest. However interested these younger writers are in apocalypse viruses, cyberspaces, and other radical machines, they do not seem much like an avant garde. Perhaps we should think of them not as informational nomads or "cyberpunks" (a term many of them reject) but rather as New Deuteronomists, defenders of logos in its hierarchically authorized form: our good old friend the codex book.
Why the codex? -- because it is a bound form, stably and authoritatively reproduced; and because the mode of its production implies a social hierarchy. Only the author is authorized. No one else's discourse (viral or otherwise) may be introduced. Deuteronomically speaking, the book must be transmitted as a perfect copy. Codex is thus an essentially conservative form, a means of exactly repeating knowledge or fictional discourse validated over time. It is the supreme discursive expression of the sedentary, the established, the legitimate. Virus and hypertext both irrupt within this old order as promiscuous uses of language -- promiscuous in the root sense of seeking relations. The virus is programmed to reproduce itself within other bodies of code; the hypertext is constructed out of links, conjugations that knit together or network diverse discourses. The mutant machine of virus/hypertext threatens the stability or social closure of textual expression.
But this is not the whole story. We have yet to understand the relationship between these two components or avatars of the mutant machine. Virus and hypertext may share a basic opposition to the codex apparatus, but these two forms are far from identical in their tendencies. Deleuze and Guattari ask, "[h]ow will the State appropriate the war machine, that is, constitute one for itself, in conformity with its size, its domination, and its aims?" (418). Hypertext and viruses do not constitute a proper War Machine, but the same question might well be asked of them. How does the State appropriate the mutant machine of cybernetic writing? Will this mechanism of appropriation be the same for both kinds of promiscuous discourse?
In the case of the virus, appropriation is fairly straightforward. The logic of viral discourse pushes the soft machine of literature back toward deuteronomy or the legitimating culture of the book. But we could just as easily reverse the terms: in a way, the return to a liber ratio is not the negation but the apotheosis of viral expression. As McDaid's demonstration reminds us, viruses are basically self-copying routines. True, the apocalypse viruses of fiction are imbued with the magic of AI or self-modification, and hence do not copy themselves exactly. But in the real world, all a virus does is turn out more of the same instruction set. It is the rate of replication -- the speed of this particular mutant war machine, as Virilio, Baudrillard, or Derrida might say -- that gives the virus its capacity for terror. In essence, viruses as we now know them are no different from the codex model of the book: their texts are dedicated to precise duplication. Thus the retreat into deuteronomy is indeed an innoculation. Whether or not language is a virus from outer space, books are most certainly a virus (or antivirus) with which culture infects itself in order to prevent more dangerous outbreaks.
In recognizing this homeopathy of virus and book, we may come closer to understanding the relationship of hypertext both to the sedentary culture and to the other half of its mutant machine. Hypertext indeed inhabits a cultural interiority, surrounded by the circuit or membrane of codex<=>virus. As such it occupies a middle state, flanked by extremes of tendentious discourse with which it shares some common elements (a component of linearity inherited from the book, a component of hypotextuality shared with the virus) but from both of which it remains fundamentally alien.
The promiscuousness of the virus sets off a disastrous explosion of discourse, much as McDaid has shown. But the promiscuousness of hypertext points elsewhere, not to manic reiteration but toward a plenum of differential possibilities, or polylogue. As in the case of the virus, the full development of this mutant discourse is only approximated in current examples. Hypertext fictions as we know them represent what Joyce has called "exploratory hypertexts," structures whose multiplicity is strictly limited by authorial design (see "Siren Shapes"). These writings may not be "electronic books," but they are definitely cases of technonarcissism, multiples that collapse into an essential unity. But just as we have the myth of a self-evolving, artificially intelligent virus, there is also a myth of advanced hypertextuality. This is what Joyce calls "constructive hypertext:" an unlimited, dynamic, collaborative body of writing shared with many reader/writers across an information network -- a primitive analogue for the consensual hallucination of cyberspace. Discursive promiscuousness in this context would mean, at least in some degree, a flattening of hierarchies and a revision or dissemination of authority. Or as Mark Taylor and Esa Saarinen have recently put it in their primer of "media philosophy:"
In a hypertextual environment, all philosophy must be interactive. Monologue becomes dialogue or, more precisely, polylogue. The disappearance of the monological voice is a radical revolution in the history of philosophy. What usually goes unnoticed is that what has traditionally passed for dialogue is actually monologue. When monologue (even in its dialogical form) becomes impossible, classical philosophy comes to an end.... Professional philosophers remain committed to an elitist culture, which dismisses low or popular culture as insignificant. ...The media philosopher, by contrast, is committed to smuggling shit back into the house of thought. ("Ending the Academy" 1)Developments like this must still be recognized as mutation and not insurgency. Despite the scatological suggestion of shit-smuggling, the outlaw media philosopher (or hypertext writer) merely elaborates inner space; he does not produce exteriority. The "shit" of outlaw culture is smuggled back into the house. The bad boys who talk of "Ending the Academy" still call themselves "philosophers." If all this is true, then hypertextual literature is not and cannot be an avant garde; but we need to ask, at this stage of postmodern history, whether the concept of the avant garde still has meaning. Charles Newman asserts with a certain savage irony that "capitalist consumer culture" is the only avant garde in evidence today (51). One of Don DeLillo's characters goes this one better, claiming that generic products, with their functional packaging and flatly descriptive labels, form the real cutting edge. "Bold new forms. The power to shock" (19). The interior, it would seem, is the only place to be.
Given these cultural conditions, it seems erroneous to derive an analysis of cybernetic writing from the projects of Dada, or anarchism, or even (pace Landow and Ulmer) deconstruction. For all its transgressive tendencies, the mutant machine to which hypertext belongs operates entirely within the logocentric order. It may be that it is therefore incapable of producing a true alternative to the State apparatus of discourse. Even constructive hypertext may turn out to be vitiated by technonarcissism, or the insidious persistence of hierarchy.
Still, if hypertext and its fictions do not answer the demands of post-logocentricism, they do at least constitute an excursion beyond the domain of the codex, a project we might call post-bibliocentrism. This movement cannot make the same claims as deconstruction; it does not extirpate ideology, metaphysics, or the simulation of presence. It is simply a technological or technical reform, a carrying of culture, as Bolter says, into a new medium and new age. Perhaps this amounts to nothing more than small change -- though one would also have to observe that the same has been true of more purely conceived attempts to move beyond the logos, which have somehow left us still within the culture of the image and the book. Perhaps we can expect no great transformation from this technical reform; but it is worth considering what effects even small changes might have after they have become (as seems likely) components of a general electronic literacy -- after, say, ten years of distributed hypertext on such currently burgeoning systems as the World Wide Web (see December and Randall). Consider a generation for whom "words that yield" are a regular occurrence, not a discursive anomaly. Consider readers and writers for whom jumps out of the system are commonplace, and who regularly articulate both hypertextual and hypotextual structures. Though this generation would still be undeniably linked by tradition and cultural continuity to our own, would they not have a fundamentally different understanding of texts and textual enterprises?
Positioned on the inward flank of its viral co-avatar, hypertextual writing articulates an alternative to strict informational hygiene or the reassertion of the book. To the extent that it represents a front or salient of some kind, it stands internally poised against both revolutionary and reactionary repression. But where does that leave us? In discussing the cultural crisis in which cybernetic writing is embroiled, we have earlier drawn on the rhetoric of nuclear holocaust -- the virus as Bomb. But fully to describe hypertext and its cultural machine requires a different, more contemporary set of metaphors. Today's holocaust is not atomic but cellular, less a matter of physics than biochemistry. Forget virus as Bomb; today's threat is virus as (retro)virus. For in a very real sense, any discussion of informational hygiene implies anxieties about the integrity of the human bodily text -- referring not just to HIV but also to cancers, toxicity, and various kinds of environmental stress. In Foucault's Pendulum, the cabalist Diotallevi draws a direct parallel between the synthetic, conspiratorial discourse in which he has taken part and the malignant growth that is killing him: "For months, like devout rabbis, we uttered different combinations of the letters of the Book: GCC, CGC, GCG, CGG. What our lips said, our cells learned.... my brain must have transmitted the message to them. Why should I expect them to be wiser than my brain? I'm dying because we were imaginative beyond bounds" (467-68).
Diotallevi's deathbed confession occurs within a novel that is also a long polemic against poststructuralism. In such anxious contexts, the nostalgia for exact replications and authoritative, cybernetic command of information may be understandable. But this devotion can lead to cultural stagnation, to a conservatism that locks us into deadened orthodoxy, the unquestioned logic of the Line. To be "imaginative beyond bounds" may be terrible, but what does it mean to be imaginative within bounds? Who determines the boundary or border lines? Which practices of discourse are ordained as safe, and which are condemned as hazards to the Book of Life? Play the metaphor out: if the viral potential of cybernetic language represents the "cancerous and dangerous proliferation of significations," in Foucault's phrase (159), then we might see in the New Deuteronomy the familiar conservative nostrums: abstinence and monogamy. That would leave hypertext the only remaining way of relative liberty, a form of "safe" intercourse whose object is to preserve possibilities of contact without jeopardizing public health. In a world caught in the pincers of virus on the one hand and askesis on the other, hypertext may provide a therapy, if not a cure. It may not represent a War Machine, or a true cultural revolution; but it could be our only option if we do not wish to have the Book thrown at us yet again.
To Stuart Moulthrop's Home Page