Two weeks ago, the schedule for our conference arrived, appropriately by electronic mail. When I examined it, I was delighted by the number of papers and presentations devoted to hypertext. It seemed to me that the Computer and Writing Conference was again leading the field of composition and rhetoric--this year by alerting the field to the importance of hypertext and other forms of electronically linked communication. However, I also feared that with all these excellent presentations (some reporting experiments and others featuring demonstrations of working computer systems) my own anecdotal remarks about hypertext and collaboration would seem anticlimactic at best, at worst jejune. From this apprehension there arose a strong desire to practice a kind of evasion, to avoid pontificating on the issue of electronic collaboration that we as a conference are seeking to address. The question became: How could I honorably evade?

It happened that I was reading a book called The Great Railway Bazaar by Paul Theroux (1975). Those of you who have read this subtle travel book will remember that the author Theroux journeys in a vast circle from London to India to Japan and back through the Soviet Union, by train wherever possible, including a 6,000 mile trip from Vladivostok to Moscow on the Trans-Siberian Railway. Theroux makes himself the central figure of this quasi-fictional narrative. He begins:

Ever since childhood, when I lived within earshot of the Boston and Maine, I have seldom heard a train go by and not wished I was on it. Those whistles sing bewitchment: railways are irresistible bazaars, snaking along perfectly level no matter what the landscape. . . . If a train is large and comfortable you don't even need a destination; a corner seat is enough, and you can be one of those travelers who stay in motion, straddling the tracks, and never arrive or feel they ought to. Anything is possible on a train: a great meal, a binge, a visit from card players, an intrigue, a good night's sleep, and strangers' monologues framed like Russian short stories. It was my intention to board every train that chugged into view from Victoria Station in London to Tokyo Central; to take the branch line to Simla, the spur through the Khyber Pass, the Mandalay Express . . . the Orient Express, the North Star, the Trans-Siberian.

. . . I sought trains; I found passengers. (Theroux, 1975, pp. 1-2)

I wish I could just continue to read to you from this book. But I can hardly claim to be the kind of reader who could turn such a reading into a performance. And even if I were and if I did read to you for the next thirty minutes, I would probably become guilty of copyright infringement. (And the issue of copyright in the late age of print deserves a lecture all its own.) So let me just read from the final page:

All travel is circular. . . . After all, the grand tour is just the inspired man's way of heading home. And I had learned what I had always secretly believed, that the difference between travel writing and fiction is the difference between recording what the eye sees and discovering what the imagination knows. Fiction is pure joy--how sad that I could not reinvent the trip as fiction. . . . I had worked every day, bent over my rocking notebook like Trollope scribbling between postal assignments, remembering to put it all in the past tense.

Gladly I boarded the train for London--correction: I am now leaving Harwich. . . . On my lap I have four thick notebooks. One has a Madras water stain on it, another has been slopped with borst, the blue one (lettered, in gold, Punjab Stationery Mart) has the ring from a damp glass on its front, and the red one's color has been diluted pink by the Turkish sun. These stains are like notations. The trip is finished and so is the book, and in a moment I will turn to the first page, and to amuse myself on the way to London will read with some satisfaction the trip that begins, Even since childhood, when I lived within earshot of the Boston and Maine, I have seldom heard a train go by and not wished I was on it. . . (Theroux, 1975, p. 342).

Theroux's mention of his notebooks seemed important to me. Here, at the very end of his story, the author was revealing his technology. He was surely not being entirely honest in suggesting that his notebooks constituted his book, that the book was nothing more than those notes and needed no revision after he arrived home. The last page, in which Theroux began to write in the present tense, reminded me of Tristram Shandy, who tried to keep pace with his life in his written memoir and yet fell further and further behind. It occurred to me that I might do with a computer what Theroux had done with his notebooks. I would get myself a portable computer, put a hypertext program on it, go off on a long journey by train, and make a hypertextual record of my journey. Because Theroux had already done Asia, I proposed to stay in Europe (where the trains were more punctual and physically less demanding). But I would go Theroux one better. As far as I could tell from The Great Railway Bazaar, the author never voluntarily showed his work to others. It was a private journal recorded in the appropriate technology of pen and paper. Mine would be a communal hypertext: I would turn my fellow travelers into collaborators. I would explain to them what hypertext was, show them what I already had, and then get them to tell me their story, which would become part of the text. This we would link to other stories in a growing web of connections.

Those of us who are enthusiastic about hypertext have argued that the technology of the link can redefine genres and bridge traditional gaps in the academic community. My peripatetic hypertext would bridge the gap not only between fiction and nonfiction, as Theroux's book had done, but also between individual writing and collaboration. It would challenge the distinction between the oral and the written: Hypertext might be the ideal medium for oral history. Finally, the hypertext that I proposed would also straddle genres associated with the social sciences and those associated with the humanities. It would be something like what Sherry Turkle (1985) had done in her book, The Second Self: Computers and the Human Spirit, where she had watched children playing with computerized toys and recorded and analyzed their reactions. So I would watch people making a hypertext and could report in the hypertext itself on my experience of watching.

For a moment I felt the way many students feel when they first read Walden. They want to rush out into the woods and build a cabin--until they begin to consider the realities of living through a New England winter in a cabin. In my case, I came to see that I would look like some absurd mendicant Buddhist, carrying around my sacred electronic object. Can you imagine a German businessman--and they are, I think, still overwhelmingly male--sitting patiently in his first-class compartment on the Intercity from Frankfurt to Düsseldorf as I explained what hypertext was and how it worked on this little Macintosh Powerbook? He would be particularly pleased to learn that hypertext would inevitably lead to the breakdown of the rigid hierarchies of Western economic and social practice and give us instead a society of temporary, associative networks of affiliation. After patiently sitting through that explanation, would he really be likely to contribute even a link or two?

Of course my project was not serious: I meant it simply as a conceit to introduce this speech. But when I floated the idea to a colleague, he began to ask questions. And his penetrating questions persuaded me that my proposal for hypertravel was not entirely frivolous; it was instead a kind of thought-experiment, a way of envisioning how hypertext fits into our literate culture. The trip I had in mind would not be an evasion at all--at least not a successful evasion. Nor would it be an escape from the challenge posed by the computer as a technology for writing, a challenge that I think all of us face who are trying to promote electronic writing in this late age of print. We can call it the challenge of narrative.

The problem, if it is a problem, is that we love narrative. We love a good story told by a master storyteller in the order that the storyteller dictates. We love the authoritative voice of a powerful author. We love those qualities of traditional printed literature that postmodern theory undermines and that hypertext apparently tries to do away with. Even today, really popular fiction is highly linear. I mean books such as Jean Auer's Stone Age romances, the epics of Danielle Steele, and the thrillers of Tom Clancy and Stephen King. The lure of narrative is powerful for sophisticated readers as well: The Great Railway Bazaar is a fine example. The Great Railway Bazaar is highly anecdotal, but in another sense it is an ultimately linear book in which the narrator travels in a single, sweeping line from London to Japan and back. What could be more linear than a trip by train? On a train you cannot diverge from a path that, if not always straight, is definitely narrow. It is probably too much to call The Great Railway Bazaar an allegory of writing and reading. And yet it does seem that, in his journey by rail, Theroux is trying to act out or actualize the experience of linear narrative. Because every instinct of his narrative style is anecdotal and descriptive, he needs the train to keep him literally on track. He can indulge himself in describing the landscape as it passes by his window because the speed of the train assures that the landscape will always change. Furthermore, the landscape will change relatively smoothly: The train assures continuity as the narrator moves from place to place. Occasionally, Theroux has to take an airplane where no railways run, for example from Saigon to Tokyo. And then he and his readers are forced to jump out of one cultural context into another. The smooth linearity of train travel is replaced by the discontinuity of the airplane. I think it is the hypertextuality of air travel that Theroux hates: Air travel violates his sense of a narrative line.

By that "line" of argument, if I were to travel around with a hypertext and portable computer in conscious imitation of Theroux, I should go by plane. In any case, I would need to discover some way to embody the metaphor of travel in the text itself, and my metaphor could not be the linear progress of the train. What hypertext forces us to do is to redefine narrative--to find an analogue in the electronic medium for narrative line and authorial control in the traditional medium of print. Hypertext is not nonlinear but rather multilinear. For each single reading of a hypertext is a linear experience. The reader is confronted with one paragraph, one frame after another and inevitably has the sense of a narrative, often a confused narrative. A narrative that seems to be carried by more than one voice. A narrative that seems to change direction abruptly (as a train ride does not do). We need to understand how this multiplicity of voices, how these abrupt changes, can work both to extend the tradition of narrative and to expand it.

It is the need to reinvent narrative in hypertext that makes hypertextual fiction so important. The first major hypertexts have been by and large fictions--Michael Joyce's AFTERNOON, A STORY and WOE, Stuart Moulthrop's VICTORY GARDEN, and John McDaid's UNCLE BUDDY'S PHANTOM FUNHOUSE. [1] Each of these fictions handles in its own way the conflict between the multiplicity of hypertext and the linearity of the reading experience. In VICTORY GARDEN, Stuart Moulthrop offers a mixture of voices and genres: first and third person narrative fiction, excerpts from other books, fiction and nonfiction, and quotations from televised broadcasts. The result is a web of intersecting, narrative strands in which a strand often continues for about four or five screens. And yet much of VICTORY GARDEN obeys an Aristotelian unity of time: Much of story takes place on one single night. Michael Joyce's WOE is a different narrative experiment. Instead of becoming more ramified as you read, WOE becomes more linear. If you persist, you are given fewer and fewer choices for branching. The hypertext eventually collapses into a linear story that is read screen by screen. WOE is proof that a conventional printed novel is simply a degenerate hypertext.

There are also important examples of nonfiction. There is PROJECT PERSEUS, an interactive library of materials on Greek literature and culture. There are George Landow's pedagogical hypertexts: One centered on Dickens and another on Tennyson's poem "In Memoriam." Landow is very much aware of the importance of the reader's linear experience of the hypertext. It is he who first discussed the need for a rhetoric of departures and arrivals--that is, the need to prepare the reader as she or he departs from one location in a hypertext and to situate the reader as she or he arrives at a new location. (Landow, [1992], explores hypertext both from pedagogical and literary-theoretical perspectives in his excellent book Hypertext: The Convergence of Contemporary Critical Theory and Technology.)

But what about other kinds of nonfiction, in particular scholarship in the humanities and social sciences? Here it is harder to point to examples. Scholars have been slow to embrace hypertext, perhaps because their own passion for narrative line and authorial control is so strong. In scholarship, the love of narrative expresses itself in the demand for a consistent line of argument with an implied narrative voice. When I speak about hypertext before groups of traditional scholars, I often suggest that electronic writing will make possible a new kind of essay, one in which the scholar is not committed to a single linear argument but can offer multiple explanations or even contradictions. In other words, instead of fixing on a single explanation of, say, the French Revolution, a historian could write a hypertext in which several, possibly contradictory, explanations are offered. No attempt would be made to reconcile the contradictions. In this way, the hypertextual essay would reflect the complexity of the event itself. When I suggest this, someone invariably comes up after my talk to explain that it is the job of the historian to provide an explanation, to give the scholarly story an ending.

That was certainly my experience in my own field of classical studies. In fact, what a classical scholar tries to do is to make a linear text out of a hypertext. Suppose you are trying to interpret or theorize about some aspect of Greek society or Greek literature. All the evidence is hypertextual. That is, it consists of pieces and small strands: passages from Greek authors; archaeological or historical evidence; and tightly focused arguments taken from secondary literature. All this floats around in a scholarly space inviting you to attempt new syntheses. And what is a new synthesis? It is new linearization, a new reading that lines up these pieces in a certain order and gives them a new significance. A scholarly reading promises closure: It promises to put to rest some topic that may have a long history of dispute or interpretation. And, of course, in classics all topics have long histories of dispute. Unless new evidence turns up, these disputes are over conflicting readings of materials already available.

On the other hand, we find in traditional scholarship a great unwillingness to accept anyone else's story because any one story, if accepted, threatens to organize and therefore close off a field. Many scholars are afraid of closure. Some are so defensive that they end up living with only the hypertext of the evidence--with the literary texts themselves as hypertexts. Their knowledge of Homer or Greek tragedy (or Milton, Austen, or Trollope) is so thoroughly networked that it constitutes a shared and unwritten hypertext among scholars. Yet one could also argue that the unwillingness of many literary scholars to accept new interpretations is an unwillingness to give up the sanctity of the primary sources. To interpret Homer is to violate the linearity of the Iliad and Odyssey. For that is what an interpretation does: It tears the poem into topical pieces and rearranges the pieces in the narrative order that suits the intepretation. The conservative scholar tries to protect the linearity of the primary source against competing linearizations, the story that Homer tells against the retelling of contemporary criticism.

Yet every writer (at least every writer for print, and the scholar is after all such a writer) is lured by the desire to tell the story her or his way--to appropriate the Iliad, The Warden, or the French Revolution and retell it in a way more compelling than all the previous retellings. Electronic writing in general calls that act of retelling into question. It suggests that the writer need not and perhaps cannot reduce a complex network of evidence to a single line of argumentation.

So whether fiction or nonfiction, successful hypertexts appropriate and extend the idea of the narrative line and voice. And, to repeat, that is exactly what we must do: extend, modify, and reinvent narrative if we are going to forge a writing space appropriate to the new technology. And if we are now on the verge of thinking the unthinkable, that is, teaching hypertextual writing to our students, then we must find appropriate pedagogies for defining and fostering new electronic narrative styles.

I have yet to reach my theme of community and collaboration. I am going to do so by invoking the now familiar link between narrative and community. Many have argued that narratives, or perhaps narrative strategies, define communities. For example, in his book Technopoly, Neil Postman (1992) laments the fact that we as a society have lost our narratives, our shared stories that tell us who we are, where we have come from, and where we are going. In fact, it is not that we have really lost such stories--not that we have forgotten the stories of the Pilgrims, the Civil War, or the American Frontier. Rather it is that our society can no longer accept these stories as unambiguous expressions of national character. They offer a narrative that seems unimportant or even prejudicial to some part of society. The question now is whether there are any narratives that can speak to the experience of everyone in our society. Or whether, as many claim, every univocal story creates a cultural inequality, for the story must elevate the narrator and her or his group at the expense of other groups; it must place someone at the center and others at the margin.

Many of our current cultural struggles are really over this question: Who gets to tell her or his story? There was a controversy last year over a Smithsonian exhibit called "The West as America: Reinterpreting Images of the Frontier." The exhibit included paintings such as Irving Crouse's "The Captive," in which an Indian man sits cross-legged and gazes in a disinterested way at a captive white woman who lies sleeping on the floor of his tepee. An article in Newsweek (1991, May 27) states that the catalog claimed that in "The Captive," the artist was unconsciously expressing the white "fear of miscegenation" (p. 70). In general, the traditional view of the American frontier "extolled progress but rarely noted damaging social and environmental change" (Evan, 1991, p. 70). Senator Stevens, of Alaska, was outraged, and Daniel Boorstin, the conservative librarian at the Library of Congress, wrote that it was a "perverse, historically inaccurate, destructive exhibit. No credit to the Smithsonian." (Evan, 1991, p. 70; see also Kimmelman, 1991). Apparently, the paintings themselves were not the problem. What upset the conservatives was the catalog and perhaps the captions that went with the pictures. The problem was the revisionist history that the writer of the catalog wanted to tell. The revisionists had one story, the conservatives another, and they both wanted to tell that story with the same pictures. Their fight really was a fight over text--fixed, printed text.

Suppose that this exhibit had presented its catalog with two sets of notes, one with traditional interpretations of the pictures and one with the revisionist interpretations. Suppose that the museum-goer could shift at will back and forth between the two. In other words, suppose the catalog had been presented as a hypertext. There are already museums in which visitors can rent audiocassettes to guide them through the exhibit. Why not give visitors a portable computer and let them experience both stories? Like my railroad hypertext, this is not a entirely frivolous suggestion: It would show that hypertext with its multiple narratives can provide an answer to the current dilemma of our culture. For if we have not one story, but two or many, then the appropriate technology with which to tell those stories is not print, but hypertext.

Hypertext is not only the best compromise for our fragmented community; it is also itself a metaphor for community at the end of the twentieth century. What we have is not a single, organized, regulated community, but a collection of communal interests groups held together by lines of assocations that cross and recross. McLuhan's term global village has recently enjoyed a renaissance. The phrase has remained popular since the 1960s, but its current popularity comes from a different quarter. In the 1960s, McLuhan was avant garde, and his notions of the global village and the retribalization of the human race through electronic media seemed dangerously left wing. The counterculture of the 1960s has become the advertising copy in the 1990s: Corporations such as AT&T and CNN now celebrate the global village. Their acceptance of the phrase should perhaps make us suspicious. If in fact we want a rustic metaphor for community in the electronic age, the bazaar is better than village. A village suggests the shared values of a preindustrial society. The image of the bazaar captures more effectively the jostle, the incompatibilities, and the confusion of this second generation of electronic communication.

The image of the bazaar brings me back to Theroux and reminds me that his title The Great Railway Bazaar is an oxymoron. What could more antithetical to the idea or the economic fact of the bazaar than the railroad? Trains cut bazaars up and put them out of business. In Europe and North America, the coming of the railroad meant that goods could be shipped and sold hundreds of miles from where they were produced, not in bazaars but by chains of wholesalers and retailers. In Asia, however, this process was never completed in colonial days. The bazaars remain. Theroux's train journey is a single path through a bazaar of Asian culture: He intersects briefly with dozens of fictionalized lives, adds in each case an episode to his collection, and moves on. The train trip itself, by which he intersects with so many narrative lines, assures the relentless linearity of his final story. That is, the train physically and fictionally pulls him out of any narrative before he can get too far off track, too far into someone else's narrative. Perhaps that is why Theroux, who is a very conventional writer in some ways, likes trains.

I would only add that the train came of age in the 19th century, which was also the highpoint of mechanical print technology. Is there not a relationship between the relentless linearity of the Victorian novel and the technology of the rail? In any case, the bazaar is surely the appropriate metaphor for the community defined by electronic writing. Because electronic writing, too, is an eclectic and constantly changing combination of elements. In electronic writing, one has to sacrifice the orderly sense of traditional community in order to achieve the spontaneity of the bazaar.

The metaphor of the bazaar also suggests noise and distraction, not so much alternative order as disorder. But this aspect hypertext and electronic writing in general need not imitate. There need to be opportunities for the electronic writer to leave the bazaar behind--to withdraw, isolate herself or himself in the act of writing. Isolation is not in itself always a bad thing. Remember that all technologies of writing and reading isolate. At some point in the writing process, the writer must separate himself or herself from the flowing world of oral communication and enter into a reflexive relationship with a writing tool. That was surely the effect of the classic instance of a transition from oral culture to written culture: The transition in Archaic Greece. This was a transition from the oral poetry of Homer to the written lyrics of poets such as Sappho. Homer's poetry was the narrative of a shared mythology; it was an apparently objective narrative, in which the narrator's voice and persona seldom intervened. Sappho's poetry was intensely and obviously personal. The persona of her poems isolated herself in order to write a text that she would then share with her readers. But we need not go so far back, writing with pen and paper or a typewriter for publication in print is also isolating. And so is the experience of reading a printed book.

It is ironic that when traditional humanists condemn the new electronic technology they often conjure up the image of the computer hacker working in isolation from others: The hacker whose social skills, like her or his muscles, have atrophied from long hours at the video screen and who prefers the predictability of computer programming to the dangerous ambiguities of human relationships. Yet humanists who paint this picture of the computer hacker do not seem to remember that the printed book (and before that the manuscript) exercized the same fascination for scholars that the computer does for hackers today. Lorenzo Valla and Poggio Bracciolini were hackers of the Renaissance. We read tales of Renaissance hackers who were so excited by the prospect of rediscovering the classics that they could hardly tear themselves from their books long enough to empty their bladders. And of course the Germans in the 19th century raised hacking to a science, which they named Geisteswissenschaften. The scholar surrounded by books is isolating herself or himself just as much as the computer programmer--and in the same way, by entering into a world of verbal and numerical symbols.

The story is the same with any technology of reading and writing. As writers, we isolate ourselves, ultimately, in order to communicate. And as we at this conference surely understand, computer technology, too, is a writing technology. [2] So the computer is not only a channel for communication it is also a technology for representation. It is the reflexive relationship between the writer and the writing technology that makes representation possible. What is exciting about the computer is that it is such a supple and innovative space for representation. We can put symbols in the computer's writing space and set them in motion; we can let them collide with other symbols. We can program symbols to operate apparently autonomously for long periods of time. These aspect of electronic writing are often best done alone.

This is emphatically not to say that collaborative writing is somehow inappropriate in the electronic environment. Collaborative writing is of growing importance in our literate culture, and presentations at this conference make it clear that computer can be a uniquely effective tool for collaboration. I can cite Christine Neuwirth's and David Kaufer's innovative PREP system, Fred Kemp's extensive work on computer-based collaboration, Claudia Bernett's discussion of "collaborative ghost writing," and so on. The conference also features presentations on electronic mail and computer conferencing, which are, of course, forms of collaboration. (In the rapidly growing literature on computers and collaboration, I would simply mention the recent collection edited by Handa, 1990.) A further demonstration of the power of collaboration in electronic writing is the experiment in collective hypertext conducted at this conference by Michael Joyce, Nancy Kaplan, and Stuart Moulthrop. It begins with a session using of real-time electronic conversation. The various lexia (individual contributions) will then be linked together into a hypertext. The links will connect lexia that address the same issues by responding, confirming, and disagreeing with other lexia. These links will also constitute narrative paths through this multiple text and suggest strategies for reading the complex document that emerges from this experiment. The final document (if we can speak of finality in a hypertext document) will be the record of both individual work and collaboration.

Collaborative computer systems should permit writers to work both alone and together. Each writer needs at times to be alone with his or her textual ideas, and computer writing systems should provide that facility. The systems also need to provide a common work space or a shared structural vocabulary in order that writers can coordinate their work--so that each writer can be conscious of the place of her or his contribution in the whole evolving text. Perhaps we should envision a process in which writers come together, negotiate a common vocabulary, separate and work in isolation, and then come together again to see how well their efforts coincide. And this alternation may repeat itself many times. In the case of an electronic group (such as Megabyte University), the process is potentially endless.

Having managed to touch on the themes of collaboration and community, I have reached Siberia in my journey, and the question is how to get home. Let me try quickly to cover the same ground in reverse by looking once again at the relationship between narrative and collaboration. By definition, collaboration requires sharing, and collaborative writing requires something to share--not just a text but a way of constructing or construing that text. It occurs to me that traditional linear narrative is a device that we use to make a text into a shared object; to make sense of a text for others. If we have read a book and want to communicate that experience to a friend, we typically tell our friend what the book was "about." The Great Railway Bazaar is about a man who travels through Asia by train, and whose ostensible purpose is to gather material for a book about a man who is traveling through Asia by train. That is the story; that is what I can tell others, who have not read it. But in my memory, the book is not a single story and it is not about one thing. It is instead a hypertext shot through with associative links. When I find myself thinking about the book, I do not trace the story. Instead, I wander through the anecdotes as associations suggest themselves.

Our traditional notion of community is also that of finding a line, a theme that organizes and sets priorities. That certainly seems to be what Postman (1992) has in mind when he speaks of a shared story that defines a community. Does collaboration then require that all the collaborators share the same story? Does collaboration turn a hypertext of possible stories into a single univocal narrative? But this would mean that there is no possibility for truly hypertextual collaboration, which is surely not what we want to say. What I think we want to say is that hypertext will require a changed sense of the goals at which collaboration might aim. The result of collaboration in hypertext ought to be a text with multiple narrative lines, reflecting the multiple perspectives of the collaborators. Indeed, each collaborator may create multiple perspectives, which the others may share only in part. In the medium of print, multiplicity is a problem because a printed document is supposed to have continuity of tone, style, voice. Hypertext should be a better environment for collaboration precisely because it does not impose the same rigid requirements of continuity. However, if writers are using the computer to collaborate on printed documents, they must still confront the demand for narrative consistency. But when the writers are using the computer to build a hypertext that will be read asa hypertext on the video screen, then they can collaborate in ways that do not require an artificial consistency or consensus.

It may be that the definition of community as a shared, univocal narrative is really determined or at least influenced by a tacit appeal to print technology, which is the technology that records and validates such a narrative. Hypertext implies a new sense of community, a new tolerance for multiple and even conflicting narratives. As a strategy for writing, hypertext can contribute to building communities, but the communities it builds will be less cohesive than the ones that we have known. And, I think, in the electronic age, we should try always to speak of communities in the plural. We must resist the pull of the single story; and this is not easy to do. We each want our own narrative to prevail. In a world of electronic writing, each community must realize that its story cannot subsume and control the stories of other communities. Electronic reading and writing means, above all, holding things in suspension, deferring conclusions, resisting closure. We must be prepared to be perpetual travelers. Even Paul Theroux got tired of railroad travel in the end. We leave him on the final page happy to be rereading the story of his trip, the linear narrative that puts everything in its place. That kind of intellectual rest and recuperation is perhaps one thing that hypertext will not give us.

Jay David Bolter is a professor in the School of Literature, Communication, and Culture at the Georgia Institute of Technology in Atlanta.

1. AFTERNOON, A STORY; VICTORY GARDEN; and UNCLE BUDDY'S PHANTOM FUNHOUSE are all published by Eastgate Systems, 134 Main St., Watertown, MA 02172. WOE is available on a diskette accompanying Writing on the Edge 2,2 (spring, 1991).

2. This talk was presented as a keynote speech at the 1992 Computers and Composition Conference in Indianapolis, Indiana.

Evan, T. (with Glick, D. & Clift, E.). (1991, May 27). Time to circle the wagons: A senator attacks the Smithsonian's 'politics.' Newsweek, p. 70.

Handa, C. (ed.). (1990). Computers and community: Teaching composition in the twenty-first century. Portsmouth, N.H.: Boynton/Cook.

Kimmelman, M. (1991, May 26). Old west, new twist at the Smithsonian. The New York Times. p. H1, H27.

Landow, G. (1992). Hypertext: The convergence of contemporary critical theory and technology. Baltimore: Johns Hopkins University Press.

Postman, N. (1992). Technopoly. New York: Knopf.

Theroux, P. (1975). The great railway bazaar: By train through Asia. Boston: Houghton Mifflin.

Turkle, S. (1985). The second self: Computers and the human spirit. New York: Simon & Schuster.