In the Zones
Hypertext and the Politics of Interpretation

Stuart Moulthrop
February, 1989

This essay appeared in Cynthia Selfe and Gail Hawisher's collection, Evolving Perspectives on Computers and Composition Studies in 1990.


    Tomorrow's hypertext systems have immense political
    ramifications, and there are many struggles to come.
                                            -- Ted Nelson

 

     Separations are proceeding.  Each alternative Zone speeds

     away from all the others, in fated acceleration, red-shifting,

     fleeing the Center.

                                            -- Thomas Pynchon


For the last twenty years the technology of writing has been overdue for a paradigm shift (if not a red shift). When humanists first realized the power of electronic computers in the sixties, they began to discover that the dominant mode of textual organization, the bound volume, is not necessarily the best way to organize expression. The book is not the ultimate random access device. Neither were decks of punched cards or spools of magnetic tape, but these methods did permit a major advance over the printed page -- segments of discourse no longer had to be bound into a fixed sequence. Storing text as a matrix of information or "random-access database" enabled readers to retrieve and assemble the text as they pleased. Final production of the text thus became "interactive," the succession of its parts determined by reader response.

In electronic writing systems, the allusive and elliptical forces inherent in prose were no longer constrained by pagination and binding. Drawing on the speculations of Vannevar Bush, Douglas Engelbart, and others, Theodor H. Nelson proposed a system of "non-sequential writing" for which he coined the term hypertext (Nelson, Dream Machines 12). A hypertext is in some ways like an encyclopedia, a collection of writings through which the reader is free to move in almost any sequence. But unlike a printed encyclopedia, the hypertext does not come to the reader with a predefined structure. The 'articles' in a hypertext are not arranged by title or subject; instead each passage contains links or reference markers that point toward other passages. These markers may be words in the text, keywords implied by the text, or special symbols. Invoking the link, by typing a phrase on a keyboard or sending some indication through a pointing device (or "mouse"), brings the indicated passage to the screen.

At minimum, hypertext automates and simplifies the reader's task in moving through a complex, non-linear document. It eliminates the distractions of page turning and volume hunting, permitting access to information in fractions of a second rather than fractions of an hour or day. This economy of effort could substantially alter the pace and scope of intellectual exchange (and not necessarily for the better); but the time saving aspect of hypertext is not the real source of its importance. Hypertext offers not simply to streamline our access to writing, but to transform the way we produce and organize bodies of text.

Since all elements in a hypertext system are subject to connection, it becomes harder to separate one discourse from another. In printed works notes and bibiliography give writings outside the current text a presence on the page, but that presence is metaphoric. Hypertext abolishes this metaphor: the other writings actually become present when the reader activates a link. Hypertext thus offers to revise our notions of definitive discourse. It seems to move us in the direction of Roland Barthes' "writerly" text, defined as "that social space that leaves no language safe or untouched, that allows no enunciative subject to hold the position of judge, teacher, analyst, confessor, or decoder" ("From Work to Text" 81).

But such radical transformations of writing do not happen overnight. The idea of hypertext was received with enthusiasm in computer science circles and by the early eighties a number of experimental systems had been created. Nonetheless, few humanists either understood or took much interest in the new technology. Hypertext remained the province of computer scientists and commercial software developers until the mid-eighties, when personal computers became sufficiently powerful to support small-scale hypertext systems. For the first time people not initiated in the mysteries of mainframe computing and programming languages could work with multiply linked electronic text.

As it proceeds this latest hypertext boom is likely to affect every field of writing, but it seems likely to produce the greatest controversy in academia. Hypertext systems are already in place in a number of professional schools, where they serve essentially as highly efficient information retrieval systems, and plans have been announced for ambitious linked-text projects involving humanities departments. In the humanities, however, where discourse is never separable from polemic and the status of 'established fact' is always dubious, the advent of hypertext will probably provoke a certain amount of conflict.

Hypertext assumes a system where diverse and even antithetical statements coexist within a single structure, each capable of emerging in the act of reading. The system thus presents the ideal figure for its own future -- for the notion of hypertext is itself a site of convergence for opposing ideas about text, authority, and the social function of writing. Humanistic hypertext projects may very well come under the sway of competing ideologies with sharply different approaches to the organization of knowledge. At present it is too soon to tell which of these ideologies will dominate and what "reading" of hypertext will ultimately emerge. It seems time, however, to begin thinking about the political implications of hypertext in the academic world.

Despite its recent adoption by corporate vendors of high technology, the hypertext concept owes much to the critique and dissent that came out of the sixties and seventies. Consider the case of Ted Nelson, pioneer of hypertext and developer of "Xanadu," one of the most ambitious electronic text projects yet conceived. Nelson describes himself as "a rogue intellectual," and a "social critic" (Literary Machines 2/10). He sees his new writing system as part of a broad social movement whose tendency is to decentralize authority and empower individuals. In Nelson's view, the development of personal computers was a crucial blow to the power structure of the computer world. He claims to have foreseen ten years ago that "the new desktop computers would deeply threaten... mainframe computing in general; that there would be a new spirit of freedom and initiative among the users thus liberated...; that this would bring an explosion of all kinds of computing uses the computer centers had suppressed... uses they could now no longer co-opt and control...." (Computer Lib 163). Nelson is a complicated figure who cannot be tagged with a single ideological label, but he regularly adopts a rhetoric reminiscent of the New Left, here pitting explosion and liberation against suppression and co-opting. Nelson seems to mean the term "computer revolution" more literally than many of his younger followers.

While Nelson was planning out his new technology for writing, literary theorists were busy deconstructing traditional methods of interpretation. Informed by an intellectual politics in many ways similar to Nelson's, post-structuralist critics envisioned a user-centered literature. "The goal of literary work (of literature as work)," Roland Barthes wrote at the end of the sixties, "is to make the reader no longer a consumer but a producer of the text" (S/Z 4). In his critique of the "readerly" text, Barthes produced a distinction which has great value for anyone trying to understand the difference between hypertext and the textuality of the printed page. This is the opposition of "the work," the object of traditional literary study, to "the Text," the new field of discourse which Barthes sought to open. "The work" is a defined body of writing, a bound volume marked with an author's name, sanctioned and validated by tradition. Against this idea of "classic" literature Barthes set "the Text," a web of language that links the "work" to other discourses, including works of other writers and critics, reader responses, and even non-literary documents. "The Text," Barthes wrote, "is read without the father's signature. The metaphor that describes the Text is also distinct from that describing the work.... The Text's metaphor is that of the network...." ("From Work to Text" 78).

Ted Nelson's Xanadu system, in which the whole of recorded discourse would be woven into one enormous matrix, in many respects literalizes Barthes' new understanding of "the Text." On a basic level at least, every hypertext is an instance of Barthes' text-as-network, and this convergence of French theory with American technology suggests an implementation of hypertext without absolute hierarchies, where the "patriarchal" functions of author, editor, and critic are abolished. Until now such discursive anarchy has existed only in theory, but that has not diminished its speculative appeal. No doubt J.F. Lyotard had such an arrangement in mind when he proposed in the late seventies that the public be given "free access to the memory and data banks." In Lyotard's view, a universal information franchise might neutralize the influence of political coercion, reconfiguring the world as an open system capable of infinite innovation (67). If one were to model an academic hypertext on Lyotard's projection, one might very well envision that system as a kind of "Democracy Wall," a huge library-cum-bulletin board whose users would be free to forge connections and publish theses, where ten thousand flowers could happily bloom.

Yet such a system has distinctly troubling implications -- not so much because it would lead to an unchecked spawning of discourse, but because the realpolitik sense of "let ten thousand flowers bloom" is quite sinister. In Mao's China, that dictum was an invitation to self-impeachment. By momentarily opening the system, the dominant faction identified its most dangerous opponents, then silenced them the more effectively.

Perhaps nothing so Machiavellian would occur on western campuses, but there could nonetheless be serious dangers in ostensibly "open" hypertext systems. Hypertext could become a technological updating of Barthes' "Operation Margarine," the scam in which the power structure subjects itself to trivial critique in order to pre-empt any real questioning of authority. Even in the most liberal of academic settings, the development of hypertext systems is not likely to evolve toward the people's republic of information proposed by Lyotard. It will almost certainly stay much closer to traditional channels of discourse, the university library and the university press. But if these models are followed too closely, academic hypertext could betray the antihierarchical ideals implicit in its formulation. It could merely routinize the intellectual charisma of the late sixties.

At the heart of this problem lie the essential design issues of "filtering" and "navigation." In its raw state, any system that "links continents of knowledge" would overload its user with choices. An undergraduate encountering Heart of Darkness might want to know something about the geography of the Congo, the history of Conrad's travels in Africa, and the iconography of light and dark in western literature -- but she probably would not want to visit these discursive places before exploring the world of Conrad's narrative. Hypertext demands a mechanism that affords users flexible, filtered access to information, something that has been described as "broadcatch" (Brand 42-45). Once filtered for relevance, information has to be put before the reader in a way that preserves both the coherence of a present line of thinking and the freedom to explore alternative pathways.

Proponents of "expert systems" seek an answer to these problems in semi-intelligent software that can anticipate its user's needs. But this approach is not without its perils, some of which were inadvertently demonstrated in a promotional videotape released by Apple Computer in 1986. Part of Apple's presentation shows a 21st-century "knowledge navigator," a computer-generated talking head, catering to the wishes of its user, a white male college professor. This "intelligent software agent" is remarkably good at what it does, but therein lies the difficulty: some of the transactions in which the navigator collaborates are distinctly unsettling. With the help of his software agent, the human user makes an intrusive call to a female colleague and breezily enlists her to lecture his class on a subject he has not bothered to read up. When the agent calls up a display predicting massive drought in sub-Saharan Africa, its user appears more impressed by the ingeniousness of the computer simulation than he is with the fact that the shifting colors on the screen spell slow death for thousands. Though the professor seems the real villain of the piece, his electronic agent seems little better than a yes-thing. The lesson seems to be that a "knowledge navigator" is only as good as the company that keeps it.

We are probably better served in the short run by human navigators, who can better challenge our biases, and we theirs. Vannevar Bush, who in 1945 sketched out the first automated hypertext system, proposed to call these workers "trail blazers," specialists "who find delight in the task of establishing useful trails through the enormous mass of the common record" (Bush 104). Trail blazers would combine the skills of intellectual historians, biographers, cognitive psychologists, and archivists. They would follow the documentary track by which discoverers came to their findings, trails which would be recorded in great detail by Bush's "Memex" system. Such efforts would no doubt greatly contribute to the advancement of knowledge, particularly scientific knowledge, but the system does raise certain problems. Bush's informational community is overwhelmingly hierarchical: Memex as Bush envisioned it was clearly intended for the upper echelons, not for "the rest of us." Accordingly trail blazers are at best acolytes in the priesthood of invention and ordinary readers are excluded from the mysteries altogether.

Ted Nelson imagines a different role and character for the navigators of Xanadu. He proposes not a priesthood of pathfinders but a "subculture of intellect" whose members would frequent the local "Silverstands," Nelson's answer to the Golden Arches. Like virtually everything else in Nelson's futurology, the name for these navigators, the "Xanadu Hypercorps," is a trademark. Indeed, members of the Hypercorps would be de facto workers in a service industry: "They will not be people who can program or repair a computer; rather, like the stewards and stewardesses of the airlines, they will know how to make users comfortable" (Literary Machines3/17).

But the Hypercorps is itself a bit disturbing. Nelson says that it would be a "cult" of information hackers, "a subculture of generalists who act more like trivia freaks or D[ungeons] & D[ragons] players." Nelson unfortunately leaves this comparison incomplete in its local context -- more like game players than what, one wants to ask. While it is fairly clear that Nelson wants to contrast his freewheeling generalists to academic specialists, a more problematic axis of comparison also presents itself. The Xanadu navigators would be more like players than like employees. Members of the Hypercorps would not draw wages or commissions, and though they could earn royalties by publishing their research within the system, they would still have to pay for any information they accessed. Since their work would be primarily synthetic, forging connections between existing ideas and discourses, they would be likely to access far more data than they produced, suggesting negative cash flow. More to the point, members of the Hypercorps would not be compensated for the time they spent making other users "comfortable" with the system. Nelson's navigators could end up donating not merely capital but valuable expertise to a profitmaking enterprise. That enterprise would be benevolent in its intent and laudable in its social function, but it would create at least an appearance of exploitation.

The Hypercorps is supposed to offer a "populist" alternative to the "stuffiness" of traditional institutions. Presumably the military echo in the name is a playful irony on the model of "Peace Corps." Yet Nelson specifies that "the Xanadu Hypercorps is expected to be an unusual and elite group," suggesting an esprit de corps that is not exactly egalitarian. Nelson foresees that the Hypercorps will offer "a social system with its own status ladder (highest are the travelling generalists), a promise of 'education' to reassure parents with -- it's better than pinball, right? And cheaper by the hour." Whether he calls elect navigators generalist or generalissimo, and whether he calls their activities education or play, the system Nelson describes is clearly not "populist." If Nelson's navigators are ranked on a status scale, then Xanadu is at heart a hierarchical organization, with those near the bottom knowing less about the system, and thus having less power to influence it, than those on top the pyramid. In spite of Nelson's manifest good intentions, some of the old computer center mentality seems to have escaped repression.

Xanadu was conceived not as an experiment in political economy but as an innovative writing system, so these objections are in some sense unfair. The importance of Nelson's work on the design and technical specifications of Xanadu far outweighs his questionable performance as an ideologue. But the re-emergence of exploitation and hierarchy in Nelson's view of the future carries a powerful lesson: unless the problems of navigation and filtering are handled carefully, even the most liberal and benign hypertext systems can begin to resemble bureaucratic and even oppressive structures.

But suppose the designers of a hypertext project were motivated not by Ted Nelson's commitment to individual initiative, but by a concern for preordained architectures of knowledge. Let us assume that the system's designers took their brief from E.D. Hirsch's declaration that:

As the universal second culture, literate culture has become the common currency for social and economic exchange in our democracy, and the only available ticket to full citizenship. Getting one's membership card is not tied to class or race. Membership is automatic if one learns the background information and the linguistic conventions that are needed to read, write, and speak effectively. (22)

It probably seems a long way from Nelson's textual populism to Hirsch's notions of required cultural literacy, but hypertext may offer a shortcut between these two frames of reference. A link-driven reading system could be a very formidable tool for teaching "background information" and "linguistic conventions," especially where this information is defined (as Hirsch defines it) in terms of literary and social history. Early on in his essay, Hirsch refers to cultural literacy as "the network of information that all competent readers possess" (2). We cannot have long to wait before someone converts Hirsch's list of "What Literate Americans Know" into an electronic database. The idea of hypertext will probably recommend itself strongly to defenders of core curricula and "mainstream culture."

There would perhaps be no harm and much benefit in a "cultural literacy" hypertext, as long as users were free to browse beyond the confines of "mainstream culture." Even if the Hypercorps were replaced by a Commissariat of Culture, there would still be opportunities for dedicated individualists to bend the rules and broaden discourse, since electronic publishing creates ideal conditions for the circulation of 'unauthorized' writing. Hypertext might in fact help the cultural literacy movement achieve its goal of enriching American education while protecting it from overprescriptiveness.

But one also has to consider a darker alternative. If it is possible to imagine academic hypertext as an open system, one can also posit a hypertext project governed by arbitrary rules, hostile to innovation, and largely opaque to its users. Such a system, conceived perhaps on the model of the "learning machines" of previous decades, might present itself as a hypertextbook containing all the information a student 'really' needs to know. The student's paths through this information would be limited to a narrow range governed by binary evaluations: 'if the reader understands this concept then proceed to the next one; otherwise repeat.' Such a system would be a monolith of integrated knowledge. Its navigators (if there were any) would not be trail blazers or generalists but park police and drill sergeants. The system would not be hyper-"Text" in Roland Barthes' terms, but hyper-"work" -- a code of received knowledge designed to engrave its idea of order upon the reader.

But these speculations deliberately emphasize the negative. Hypertext will no doubt fail to produce a utopian anarchy of discourse, but if we take care in designing our systems, it need not rigidify into an empire of signs. We urgently need a social theory of hypertext, a set of principles that account for the text not as an autonomous object but as an evolving transaction between readers and writers. Essential contributions to this theory have already been made by humanists interested in the rhetoric of hypertext, such as George Landow and Jay David Bolter. However, there has yet to be full discussion of the "immense political ramifications" Nelson saw in his textual revolution. As a first step in that discussion, here are three proposed principles that might be incorporated in the design of future hypertexts.

  1. Hypertext is not an object but a system. This is simply a reiteration of Barthes' distinction of "work" from "text." Designers of the next generation of hypertexts should not set out to produce textbooks for a sequence of core courses. Hypertext is not a definable artifact like a bound volume, it is a dynamic, expansible collection of writings whose contents will change from moment to moment. It is nothing at all like a book, only a bit like a library, and much more like the university itself in that it is shaped both by inherited resources and current contributions. Though part of any system will probably need to be permanent, it is probably better not to depend too heavily on a framework of canonical text or definitive discourse.
  2. Hypertext is a medium. Every hypertext project should support writing as well as reading. The function of the hypertext system is not simply to disseminate information but to create better conditions in which people can exchange, develop, and evaluate ideas. Users of an academic hypertext should be able to explore links between different areas of discourse, but they should also be able to create links of their own by posting their writings within the system. Since it is therefore a medium of public communication, the hypertext system should be administered by a representative body -- ideally a committee composed of all classes of user: students, faculty, and other members of the community.
  3. Hypertext must be diverse. The first two principles raise a crucial problem: how can an academic institution support both free expression and a meaningful structure of knowledge? If anyone can change the system, and if both students and faculty have authoring privileges, how are necessary privileges and distinctions to be preserved? The solution is simple enough -- the hypertext system must not be a uniform matrix but a heterogeneous network of textual spaces. The best hypertext proposals will provide for separate zones of discourse, where different conventions of access will apply. One zone, which may be an outgrowth of the university library, would only allow access for reading, providing an archive for information which the supervisors agree should be retained in permanent form. Other zones might allow linking and writing privileges only to one class of user, for instance students enrolled in a certain course or major, while still others might allow anyone to contribute. But the division of the hypertext into zones should not constitute balkanization or gerrymandering. While opinions may differ on the intellectual value of discourse in one domain or another, no zone should be subjected to special editing or censorship and all should be open on at least the reading level to any member of the community.

Systems incorporating some version of these ideas, particularly the notion of independent zones, should be able to steer between the extremes of informational anarchy and despotism. After all, this model imitates the social structures that already promote the free exchange of ideas in academia. But while a pluralistic hypertext might resemble the current system of seminars, forums, conferences, and journals, it would still have significant advantages over the present community of discourse. The "zones" of intellectual exchange today resemble Thomas Pynchon's centrifugal vision of postwar Europe. They are separate societies moving along distinct ideological and methodological paths, fleeing the center. By linking these zones together, a pluralistic hypertext system could create new opportunities for communication and empowerment, both within and among disciplines.

There is no denying, on the other hand, that hypertext is also going to upset the stability of intellectual institutions, perhaps more violently than any of the political upheavals of the late sixties. As Ted Nelson points out, there are many struggles to come: struggles over access to information, design of curricula, and definition of canons and methods. These struggles antedate any technology, even the technology of writing itself. The kind of allusive, associative communication that hypertext fosters has always been at the heart of humanist scholarship, but it does not follow that hypertext merely augments or amplifies the status quo of discourse. Hypertext not only makes textual linkages more numerous and effective, it changes the social uses to which those links are put. Since it creates a world where discourse can no longer be limited to isolated "works," where students' questions have the same textual presence as professors' glosses, hypertext will probably force us to reformulate our notions of intellectual authority. This may be a daunting prospect, but there is really very little to fear. The medium of hypertext favors plurality over singleness, movement over rigidity, and connection over isolation. These values have sustained humanistic discourse in the past, and if we do not neglect them now, they should guarantee its future.


Selected Bibliography

Balestri, Diane.

"Softcopy and Hard: Wordprocessing and Writing Process." Academic Computing February, 1988: 14+.

Barthes, Roland.

"From Work to Text." In J. Harari (Ed.), Textual Strategies: Perspectives on Post-Structuralist Criticism. Ithaca: Cornell UP, 1979.

Beck, J. Robert and Donald Z. Spicer.

"Hypermedia in Academia." Academic Computing February, 1988: 22+.

Bolter, J. David.

Writing Space: The Computer, Hypertext, and the History of Writing. Fairlawn NJ: Laurence Erlbaum Associates, forthcoming [published 1991].

Brand, Stewart.

The Media Lab: Inventing the Future at M.I.T. Viking, 1987.

Bush, Vannevar.

"As We May Think." Atlantic Monthly July, 1945: 101-108.

Conklin, Jeff.

"Hypertext: An Introduction and Survey." Computer 20 (1987): 17-41.

Drexler, K. Eric.

Engines of Creation: The Coming Era of Nanotechnology. New York: Anchor Books, 1987. 217-30.

Hirsch, E.D.

Cultural Literacy: What Every American Needs to Know. Boston: Houghton Mifflin, 1987.

Joyce, Michael.

"Siren Shapes: Exploratory and Constructive Hypertexts." Academic Computing November, 1988: 11+.

Landow, George P.

"Changing Texts, Changing Readers: Hypertext in Literary Education, Criticism, and Scholarship." Reorientations: Literary Theory, Pedagogy, and Social Change. Ed. Bruce Henricksen and Thais Morgan. Urbana: University of Illinois Press, forthcoming [published 1990].

Nelson, Theodor H.

Computer Lib/Dream Machines. 2nd ed. Redmond, WA: Tempus Books, 1987.

Yankelovich, Nicole, Norman Meyrowitz and Andries van Dam.

"Reading and Writing the Electronic Book." Computer 18 (1985): 15-30.

 

URL: http://iat.ubalt.edu/moulthrop/essays/zones.html

 

 

[BACK]