Polymers, Paranoia, and the Rhetorics Hypertext

Stuart Moulthrop

Of what lasting benefit has been man's use of science and of the new instruments which his research has brought into existence? They are illuminating the interactions of his physiological and psychological functions, giving promise of an improved mental health.
—Vannevar Bush, "As We May Think" (1945)

He used to think of the world of language ... as a vast and perpetual conflict of paranoias.
—Roland Barthes, The Pleasure of the Text (1973)

The development of interactive writing systems has now proceeded far enough to engage humanist academics and theorists, who were largely absent from the technical phases of the creation. Now that hypertexts and hypertext systems are becoming common, academics are rushing to theorize these new communications tools, often in the name of rhetoric. But this is no easy task. Those who would create a rhetoric for hypertext must be prepared to throughly reconsider their subject: as Jay Bolter points out, forms of electronic text production "threaten the definitions of good writing and careful reading" established by the culture of the printing press (2). The rhetoric of hypertext requires substantial re-thinking.

Print endows writing with stability, authority, and artifactual integrity. Books in their highest form, as Alvin Kernan instructed, are "ordered, controlled, teleological, referential, and autonomously meaningful" (141). But it is difficult to find these attributes in hypertext. Concepts like order, control and teleology seem of limited application in "non-sequential" writings, which after all can embrace elliptical and even contradictory expressions (Nelson, 5/2). Likewise, since hypertexts favor a plurality of discourses over definitive utterance, they are less likely operate denotatively or referentially, as an exclusive description of reality, than as a form of simulation, exploring hypotheses and contingencies. And of course it is impossible to regard any interactive text as "autonomously meaningful" since the text achieves expression only through interaction with its user/respondents. The functional and experiential bases of hypertext differ substantially from those of the book.

Nonetheless, most rhetorics of electronic writing proposed to date still orbit the aging star of Gutenberg technology. We may renounce particular print conventions like univocality and singular sequence, but we maintain the culture of the book as our benchmark. Patricia Carlson, for instance, argues that any electronic information system must be able to "duplicate [the] functionality" of print (129). Such appeals to traditional values display a certain nostalgia: the more our communications technologies change, the more we insist on their continuity with older forms of expression. This reaction is understandable enough: as people make their first acquaintance with new technologies, they need reassurance that innovation will not utterly invalidate the cultural forms to which they are accustomed. But there comes a point at which it is no longer productive to discuss electronic writing as if it were a simple extension of print. The rhetoric of hypertext may still be in its early days, but we have already reached that point.

To their credit, hypertext rhetorics proposed so far are founded on experience with actual projects and systems. But the writings on which these essays are based often are digital incunabula, somewhere in the evolutionary scheme between books and true electronic texts. Carlson, for instance, bases her "rhetoric of hypertext" on a technical training system that emphasizes predefined pathways and hierarchical structures, features inherited from the printed manuals from which her hypertext was assembled. George Landow formulates his "rules for hypermedia authors" during development of Intermedia, a teaching resource considerably more flexible and complex than Carlson's, but with its own allegiance to the book. Among other radical innovations, Intermedia collapses "the rigidly hierarchical distinction between a main text and its annotation," bringing into question basic concepts of textual identity (81). Yet in spite of this conceptual advance, Intermedia remains at least in early applications an appurtenance of print culture, a source of enrichment or context-building for students primarily concerned with traditional texts.

It hardly seems surprising, then, that these first hypermedia rhetorics describe "ordered" and "autonomously meaningful" discursive objects. Landow's rules emphasize the utility and intentional coherence of the text. He directs designers' attention to points of "arrival and departure," arguing that they "must decide what readers need to know at either end of a hypermedia link in order to make use of what they find there" (82). To determine what users "need to know," the designer must assume some unified or teleological understanding of the text. In much the same vein, Carlson contends that "hypertext presentations need to preserve contextuality and... provide a means by which the user can see the fragment's place in the coherent whole" (129). The complexity and multiplicity of the hypertext, a complex system of documents and exhibits, reduces at some conceptual level to a unified body of information, a training manual or a course text. While it is true that this material could not be presented as effectively in a book, the difference is of degree not of kind.

For Landow and Carlson, the most important aspect of hypertext design is integration, the assembly of parts into a meaningful whole. Their prime directive is well expressed in E.M. Forster's famous injunction, "Only connect"—the "connections" in this case being hypertext links. Rhetorics of coherence or itinerary (arrival/departure) theorize hypertext as a network of terminated segments, each joined to another. Since every discursive unit necessarily connects, the system in all its heterogeneity constitutes a closed circuit. To bring in the metaphor of "topography" so common in electronic writing, the hypertextual network forms an enclosed territory, a hortus conclusis or "garden of forking paths" (see Bolter; Moulthrop).

By invoking such a constrained notion of text construction, early rhetorics of hypermedia have confused two kinds of extension in writing, which we might call volume and space. Bodies of text came to be called "volumes," Bolter notes, because they were traditionally stored on rolls or volumina (85). "Volume" thus indicates an objectified domain, a dimension of writing limited by the boundaries of an artifact. Volume is the extent of the text expressible in ells of papyrus, pages of print stock, or tracks of mylar storage. It is that linear and hierarchical measure invoked in expressions like "stated above," "see below," or "go to first card." But writing (or indeed any language) extends itself not just in artifactual volume but also in a mental or imaginary dimension of space. Phenomenologists like Wolfgang Iser contend that the act of reading engenders a non-sequential, "virtual work" in where the reader's subjective response intersects with the objective specifications of the text (21). As Joseph Frank has shown, modern literature exploits this aspect of textuality, creating thematic structures that can be fully understood only when held in simultaneous mental suspension or juxtaposition (99). Because this "spatial form" of writing transcends such material categories as page order or textual boundaries (see Douglas's discussion of Durrell in this issue) it cannot be measured or circumscribed in the same way as the volume of the text.

Roland Barthes attempted to differentiate writing as volume, which he called "the work," from writing as virtual space, or "the text" (73-74). He posited a theory of discourse that would no longer conceive of texts as discrete properties or territories. He conceived instead of a polyvocal, intertextual "social space of writing" (81). Texts produced in this communal space would be heterogeneous networks where the language of the author would be joined to those of readers, commentators, critics, biographers, and so on. But Barthes was dreaming: clearly such social texts cannot exist in print—books are not and never can be open discursive networks. No matter how assiduously we attempt its deconstruction, print retains closure and coherence. Even subversive anti-books like The Medium is the Message or Glas remain definitive objects, expressions of a unified master discourse. Printed writings may imply or evoke a writing space, but they can do so only from within their allotted volume. The book is always in essence a territorial form of writing, an attempt to fix boundaries for thought or expression.

Hypertext incunabula, with their emphasis on "coherence" and well-defined boundaries, attempt to impose this territoriality on the electronic environment. Inside the text-as-territory the reader or interpreter becomes an explorer: She is invited, as Michael Joyce explains, to "[transform] a body of knowledge to meet [her] needs and interests" (11). But this "exploratory" conception raises an ideological problem. When we think of electronic texts as territories, Greg Ulmer cautions, we exhibit "the will to power in knowledge," a desire to master discourse by establishing our control of its frontiers (5). Exploratory hypertexts serve this impulse nicely. Despite the appearance of "interaction," they are just as "autonomously meaningful" as books. Or to be more accurate, like books they also preserve the discursive autonomy of their authors. The user may "transform" the textual body by following alternative paths or linkways, but that body is protean: it retains its fundamental identity under all transformations. After all, a good exploratory hypertext is teleologically coherent. The maze may have many permutations, the circuit many switchings, but in all of them the user still circulates through the same mechanized volume, working toward some predefined end.

Still, not all hypertexts are exploratory and not all theories of electronic writing must proceed from assumptions of utility and coherence. We need not remain trapped in volume when a true "writing space" is available. Electronic writing systems enable—literally—an unbinding of the text. Diane Balestri observes that even a transitional and print-based technology like word processing brings fundamental changes to the production of writing, replacing "hardcopy" with "softcopy," a fluid and malleable text-in-flux (17). Softcopy documents differ from their hardcopy counterparts in that their digitality is preserved even as they reach the screen and become alphabetically analog. Written for display and not printing, they are destined not for output in read-only form but for provisional recording in read/write electronic storage. These document can thus never be "closed" against further extensions, either by the original writer or by others. They are infinite process, not definitive product. Thus they are always open to expansion and intervention, by other authors as well as by their originators. They may therefore provide (at least potentially) an opening to a "social space of writing."

Hypertext lets us make good on this opening. In moving from hardcopy to softcopy we recognize that the volume of electronic writing is infinitely expansible, though we still conceive of the text as a linear or summary form of expression, a coherent file or document. Hypertext, being "non-sequential," allows us to move from our notion of volume (however expanded) to a true textual space—"writing space" as Bolter calls it. As he points out, hypertext exploits "the computer's capacity to designate any unit of text as a new element in an expanding vocabulary of signs" (60). Hypertexts need not be considered unitary documents. We can conceive of them instead as multiple discursive sequences, not documents but "docuverses" (Nelson 2/9). The model of discourse implied here is not a bounded territory but a dynamic and expansive system, a structure much like the social space that Deleuze and Guattari call "rhizome." Rhizome, an alternative model for social and discursive formations, is an entity "made only of lines: lines of segmentarity and stratification as its dimensions, and the line of flight or deterritorialization as the maximum dimension..." (21). Hypertext offers the possibility of a "deterritorialized" writing, a conception based not on circuitries of arrival/departure or "exploration," but instead on pure extension or "the line of flight." This idea represents the logical next step after softcopy: a form of writing that does not attempt to duplicate the functionality of print but instead sets out to explore new possibilities for written communication.

Joyce calls this next step "constructive" hypertext, a scheme in which the body (or activity) of writing is considered not a subject territory but a scene of creation: a "version of what [it is] becoming, a structure for what does not yet exist" (11). Constructive hypertexts extend not just in material volume, but also into the deterritorialized space of electronic writing. They are not closed books but open ranges, discursive improvisations that grant no one the last word. The shift in metaphors from "exploration" to "construction" signals an important difference. The reader or "user" of an exploratory text can "transform" the discourse by subjecting it to a range of manipulations predetermined by the text's designer. But the operator of or on a constructive hypertext (the terms "user" and "reader" will not do) has a much greater degree of freedom. She enjoys "a capability to act: to create, to change, and to recover particular encounters within the developing body of knowledge" (11). She is no longer an explorer, discovering what was left for her to find, but a constructor or co-creator, a belated but equal participant in an unfolding "social" text.

Owing to the dissemination of microcomputers and hypertext software, writers, teachers, theorists, and other humanists can now begin to explore the possibilities of constructive hypertext. The interwoven network of fictions and essays published in this issue represents a limited first experiment in constructive writing (see "Writing on the Hypertextual Edge"). Because this work is still a hybrid of print and electronics (a disk inside a bound volume), it also belongs unmistakably among the digital incunabula. Though the composing phase of the project did at some point involve a true constructive hypertext, everything the reader will find here exists in read-only or exploratory form. So this experiment, like Carlson's and Landow's, remains in a Gutenberg orbit. But we have tried to push that orbit as far as possible into ellipse, orienting ourselves not back toward volumes, territories, or "electronic books," but outward to the final frontierlessness of writing space. Since earlier speculations on hypertext rhetoric have been limited by their dependence on print concepts, it seems fair to ask if this attempted divergence affords any new perspectives on the nature of electronic writing. Can we propose a rhetoric of constructive hypertext?

Any such rhetoric must differ crucially from the rhetorics proposed for exploratory hypertexts: it would presumably be founded not on coherence and order but on instability and "chaos," an understanding of structure not as an imposition from without but as spontaneous development from within (see Prigogine and Stengers). It would have to describe hypertext linkage not in terms of positive constructions or artifacts—arrival/departure or cause/effect—but as Terry Harpold suggests, in terms of absence and negativity—gaps, openings, and fissures (see both "Threnody" and his essay in this issue). In short, coming to terms with constructive hypertext might entail seeing the text not in Carlson's terms as a "context" that cements fragments but rather as itself a fragment of context, not a functional unity but as the limit case of the unifying function. If exploratory hypertext is the place where things come magically together (Only connect...), then perhaps constructive hypertext is the space in which things fall apart, flying outward in lines of deterritorialization.

All of this might make brave literary theory, but does it in fact help us understand how to read and write electronic texts? This was among the issues we meant to engage in our experiment with constructive hypertext. One particularly revealing product of this engagement is John McDaid's meditation, "The Planes," which began as electronic text (a separate hypertext meant to be merged into Michael Joyce's WOE) and was later reduced to print. In introducing his piece, McDaid sets out to describe the interface between the definitively expressed printed text and the network of potential and actual hypertext connections from which it has been abstracted. His problem thus maps onto our more general concern: how to describe hypertext as an interplay of reading and writing, structure and dynamism, volume and space? McDaid offers a highly suggestive trio of models:

...Hypertext, I suspect, is an enactment of these fundamental metaphors of connection: gluons, mitosis, polymerization.

Gluons, according to one theory of subnuclear dynamics, are the hypothetical force-bearing "particles" which moderate interactions among quarks. This gluon force has the peculiar property of increasing with distance; trying to pry a proton apart... only increases the "number" of gluons binding them together. Add enough energy, and instead of freeing the quark you pass the threshold for the creation of a new particle. The infinitely extensible text. In mitosis... cellular DNA replicates, separates, and aggregates itself into two child-cells. The infinitely... repeatable text. Polymers arise as monomers (simple molecules) gang up to form long, interesting chains, whose characteristics can differ in important ways from the free monomers in isolation. ...The infinitely emergent text. (See this issue.)

McDaid's metaphors let us simultaneously differentiate and reconcile the structured and the deterritorializing aspects of hypertext. "The Planes," a set of prose variations-on-a-theme, represents "the naked monomer" abstracted from WOE, a densely cross-connected or "polymeric" hypertext. We can consider McDaid's text entirely as a particular structure with its own consistency and coherence, an essay on the fear of flying and the more ominous "fear of the Created." But at the same time we must also recognize that this artifact of print is a fragment or reduction.

"The Planes" exists on many planes. Behind or above or around the linear or two-dimensional text lies an "infinitely extensible... infinitely repeatable... infinitely emergent" hypertext whose dimensions are many. Like the subchain of a polymer or a strand of DNA or the theorized substructure of the quark, every chip off this old hypertext announces and instantiates a range of possible assemblies. In encountering McDaid's text, we must think not just of its particular contents but also of the way its "planes" intersect with other discursive fields (for instance, with the references to sex aboard airliners in both WOE and Izme Pass). This writing after all is putatively constructive, an intersection of creative processes both prior and inchoate. The structures we do see always imply structures—and unstructurings or restructurings—that are not yet apparent. Thus in contemplating constructive hypertext, we might well find ourselves in the grip of a certain "Puritan reflex of seeking other orders behind the visible, also known as paranoia" (Pynchon, 188).

"Have I read too much Pynchon?" McDaid asks rhetorically. His answer is a foregone affirmative, but we might demur. Arguably anyone interested in the rhetoric of hypertext could use an acquaintance with Thomas Pynchon's works, if only because Gravity's Rainbow provides us with an essential tool for understanding complex systems: the cognitive strategy called paranoia. This state of mind, Pynchon observes, constitutes "the onset, the leading edge, of the discovery that everything is connected, everything in the Creation, a secondary illumination—not yet blindingly One, but at least connected...." (703). This is not exactly clinical psychosis (the syndrome Pynchon has in mind is chemically induced), but rather a transcendental enlightenment, another way of knowing (para-nous ), or a "secondary illumination."

It should be evident that paranoid enlightenment contains a strong parallel with exploratory hypertext, a system of expression in which everything is indeed connected. But Pynchon's concept of paranoia has even greater relevance for constructive hypertext, for it gives us a useful way of understanding both connections and disjunctions, both links and gaps—both the present structure and the numinous presence of "what does not yet exist." Paranoia as Pynchon defines it is a liminal condition, a threshold state where the mind encounters the "leading edge" of a higher knowledge. Because the paranoid perceiver is still rooted in reality and history, his cognition depends in part on rational constructs—connections and correspondences, lines of influence and causality. This condition of paranoia, the stage of plotting and counterplotting, maps onto the exploratory aspect of hypertext. Indeed, every exploratory hypertext is a kind of informational conspiracy.

But in Pynchonian paranoia as in hypertext, the perception of structure is never a simple matter. Paranoids produce not just delusions but delusional systems, structures of compound association that attempt to embrace "everything in the Creation." Here we have a clear parallel with constructive hypertext—infinitely extensible, repeatable, and emergent, with an inexhaustible latency of other orders. If the exploratory text is a conspiracy, then the constructive text is a kind of lunacy, a riotous proliferation of discourse (something like Pynchon's novels themselves, only moreso). As Pynchon describes it, paranoia is a dynamic process, a relentless approach to a "next higher assembly" of meaning. But this progression, though apparently end-directed, is in fact endless: for all its dynamism, paranoid cognition never produces definitive results. The heightening of consciousness characteristic to paranoia is asymptotic. Its overture to universal connectedness remains "a secondary illumination, not yet blindingly One," constantly approaching but not quite attaining an axis of unified vision. Most of Pynchon's novels operate in just this manner, generating a plethora of clues but resolving none of their primary mysteries. Critics have suggested that the fictions turn out this way because they embody the paranoid epistemology they describe: Pynchon affirms structure and significance while denying ultimate meaning (see Hite).

If paranoia actually produced cosmic understanding, then we would have to call it mystical instead of liminal; and since we do not wish to argue that hypertext is a form of religious experience, there would thus be little point in bringing paranoia to bear on hypertextual rhetoric. But paranoia as Pynchon imagines it does not carry the perceiver all the way to transcendental enlightenment; it brings her only to a borderline or threshold. The paranoid's "secondary illumination" reveals a region where causal analysis and holistic awareness converge, and that paradoxical convergence is the natural element of hypertext. Consider then a paranoid conception of hypertext, an approach that portrays the encounter of user and text in terms of fundamental liminality or doubleness, a tension between the structure that exists and a metastructure that either does not yet exist or is not apparent.

What does this liminality mean for the rhetoric of hypertext? How can an approach to structure as fundamentally transitional or contingent be brought into accord with rhetoric's attempts to provide rules for effective expression? Might the whole project of rhetoric be impossible for constructive hypertext? Indeed, if we naively pursue the rhetoric of hypertext or hypermedia in hopes of defining a discursive territory, we are likely, as Ulmer warns, to fall victim to our own will-to-power. Constructive, post-Gutenberg hypertext seems likely to subvert or overturn any system of rules established for it. The time is probably not far off when we will see the first academic handbooks (or hyperhandtexts) offering authoritative principles for linking, navigation, and collaborative construction. Projects of this kind should probably be greeted with suspicion. When definitive rhetorics are applied to hypertext they will no doubt operate as they have to date, steering a radically divergent technology back toward the familiar categories and territories of print culture.

On the other hand, suppose we adopt a broader (and more historically accurate) definition of rhetoric: not a project of positivistic reduction, but a genuinely philosophical inquiry into changing conventions for communication. A dynamicrhetoric of hypertext, informed by McDaid's notions of textual "plasticity" and a liminal or "paranoid" understanding of the limits of structure, could provide a useful way of thinking critically about constructive texts. Such a dynamic approach would necessarily have a strong interdisciplinary character, combining strategies from semiotics, political theory, cognitive science, and phenomenology, among other fields. It might find a fairly high place on the agenda of cultural studies or other new disciplines with an interest in social semiotics.

But the rhetorics (or politics) or constructive hypertext cannot be a new game played by the same old academic rules. It almost goes without saying that such a rhetoric would itself be produced as constructive hypertext—which means that its notion of authority would be very different from those with which we are familiar in rhetoric and literary theory today. So there may finally be something recursive about hypertextual rhetoric. Barthes saw his critical practice of "writerly" reading lapsing ultimately back into its own subject: discourse on the text would itself be only "text... and textual toil" (81). In similar fashion, the rhetoric of hypertext may turn out to be inseparable from constructive process that motivates it: in the final analysis, the only rhetoric of constructive hypertext may be constructive hypertext itself.

Works Cited

Balestri, Diane Pelkus. Softcopy and hard: wordprocessing and writing process. Academic Computing (February, 1988): 14+.

Barthes, Roland. From work to text. In Josué Harari (Ed.) Textual Strategies: Perspectives in Poststructuralist Criticism. Ithaca: Cornell UP, 1979. 73-81.

Bolter, Jay David. Writing Space: The Computer, Hypertext, and the History of Writing. Hillsdale: Lawrence Erlbaum Associates, 1990.

Carlson, Patricia Ann. The rhetoric of hypertext. Hypermedia 2 (1990):109-32.

Delany, Paul and George Landow (Eds.). Hypermedia and Literary Studies. Cambridge MA: M.I.T. Press, 1991.

Deleuze, Gilles and Félix Guattari. A Thousand Plateaus: Capitalism and Schizophrenia. Brian Massumi (Trans.). Minneapolis: University of Minnesota Press, 1987.

Frank, Joseph. Spatial form in modern literature. In Michael Hoffman and Patrick Murphy (Eds.) Essentials of the Theory of Fiction. Durham: Duke UP, 1988. 85-100.

Harpold, Terence (1991). Threnody: psychoanalytic digressions on the subject of hypertexts. In Delany and Landow (1991), 171-84.

Hite, Molly. Ideas of Order in the Novels of Thomas Pynchon. Columbus: Ohio State UP, 1983.

Iser, Wolfgang. The Act of Reading: A Theory of Aesthetic Response. Baltimore: Johns Hopkins UP, 1978.

Joyce, Michael. Siren shapes: exploratory and constructive hypertexts. Academic Computing (November, 1988): 10+.

Kernan, Alvin. The Death of Literature. New Haven: Yale UP, 1990.

Landow, George. The rhetoric of hypermedia: some rules for authors. In Delany and Landow, 81-104.

Moulthrop, Stuart. Reading from the map: metaphor and metonymy in the fiction of forking paths. In Delany and Landow, 119-32.

Nelson, Theodor H. Literary Machines. Edition 90.1. Sausalito: Mindful Press, 1990.

Prigogine, Ilya and Isabelle Stengers. Order Out of Chaos: Man's New Dialogue with Nature. New York: Bantam, 1984.

Pynchon, Thomas. Gravity's Rainbow. New York: Viking, 1973.

Ulmer, Greg. Grammatology hypermedia. Postmodern Culture 1(1991): 1-19.

[Back to the Writing on the Edge special section, 1991 index.]