A New Art Form: Hypertext Fiction

New Art, Old Art

When I taught the sociology of art, I used to give students a choice of research topics on which they might write papers. One, chosen only by the most daring, was to create a new art form and to keep a careful and detailed journal of their efforts. I warned students that they would probably not succeed in creating a new form, and that their journal would probably be the most interesting result of the project.

One student’s experience, as reported in her journal, illustrates some of the problems involved. Mindful of the expense involved in undertaking anything large scale, she proposed to create sculpture from bubble gum, chewing the gum to make it pliable and then molding it into various forms. She soon found that she needed a great deal of gum to make even a small sculpture. Her own jaw got very tired chewing all that gum. So she asked her friends to help by “pre-chewing” gum into pliability. Since her friends were not always there when she needed them, they would chew the gum and then leave it for her to work with later. She then had to “rechew” it to get it into shape for molding. When her friends learned that she was doing this they were nauseated; she had gotten over that. Worse yet, she soon realized that the sculptures she made looked just like any other amateur sculpture she might have made in clay or wax or some other common medium. At this point, as I read her diary, I turned the page and found scrawled in large angry letters across the page “Fuck Becker! How did he get me into this?”

What I had gotten her into was an insoluble problem. When, after all, is an art form new? It is impossible even to imagine an art that is absolutely independent of what went before it, that does not make use of physical resources already in existence (as my student did in using chewing gum); that does not rely on already known social forms for referents (as her work did in referring to the human figure); that does not make use of what audiences and potential audiences already knew to create emotional effects (in her case, creating the effect of disgust by reference to the rechewing of the gum); that does not use some conceptual apparatus that predated its innovations, even if only as something to react against (as her work did in changing, in the end, only one thing about sculpture, the material used to make the work).

You cannot look at a work of art without seeing at once how it is just like all other works of its kind and how it is quite distinctive. Every work extends its own genre and in that sense is making something new. Every work uses, for the most part, devices and materials every other work of its kind uses.

What can, however, be new and different is the organization that arises to see the work to completion, to provide its necessities, and, most importantly, to proclaim novelty, to say “Here is something new, different, and unlike anything done before.” The creation of a new art world makes the work that world creates new, in a sense I will now try to explain through an analysis of the case of hypertext fiction.


Beginning in the mid-1980s, a number of people began to create computer based fictions. Most of these authors took advantage of the computer’s possibilities to write what have been called, generically, hypertexts. (The standard work, if we can speak of such a thing for so new a field, is Bolter (1991a, 1991b). See, also, Joyce (1992), Landow (1992), and Rizk (1990).)

The idea of hypertext fiction is older than the computer, but computer-based implementations of the idea are as new as the personal computer, and have many roots. not just the precomputer attempts I will mention in a moment. It takes more than a new concept to create a new art form so, if we are to understand the development of this new form of literary art, we must explore the development of skills, peer groups, support institutions, and audiences.

The idea and fact of hypertext are sufficiently unfamiliar to require description and explanation. Let us think of a text as made up of units, each unit consisting of a paragraph, a page, perhaps only a word—the size of a unit is variable and arbitrary. In conventional (what we should now, as will become clear, call “linear”) texts, such as books or articles, each unit is connected to at most two other units, the one that precedes it and the one that follows it, the page, paragraph, or word before and the page, paragraph or word after. A straight line runs through the work from beginning to end (which is why the term “linear” is appropriate). The work’s structure is physical as well as conceptual. Paragraphs and pages stand in a physical relation of before and after which cannot be changed without destroying the work’s physical integrity.

In hypertext, each unit (now defined as the computer screen’s contents, including what can be viewed by scrolling) may be connected to many other texts, in such a way that the reader may see any one of many possible successors, either by choosing one, say by using a mouse to click on a word on the screen), or by having the author make the choice by creating “paths” to be followed which depend on previous choices the reader has made. The entire work exists physically in the form of code in a storage device such as a hard disk and in the computer’s RAM (random access memory). Each text can, in principle and in fact, have many preceding texts and many texts which follow it. Rearrangements of the text do not entail any physical disruption of the work’s existence. There is no one linear path through the work. The experience of reading such a text is more like consulting a map, or looking at a painting or photograph, than reading a book. Readers can move in many directions, follow the many paths the author has prepared (just as in a painting we follow the paths the artist has composed for our eye), and in some hypertexts may follow paths of their own devising. All this, as with a painting, leaves the work just as it was found.

We could say that this is merely an extension of older experiments in organizing texts. Footnotes are minimal hypertexts, in which a branch leads from the main text to the note and from the note back to the text. Encyclopedias are typically organized to provide multiple paths from one article to another, the links being indicated by the references at the end of the article (“SEE IMPERADOR JOÃO VI”). Many novelists have experimented with hypertextual forms in physical, print books, e.g. Julio Cortázar in Hopscotch, (1966) with its multiple forking paths, George Perec’s La Vie: Mode de Emploi (Life: A User’s Manual)(1987), which uses the metaphor of a jigsaw puzzle and almost forces the reader into the sort of puzzle-solving activity jigsaws entail, or Milorad Pavic’s’s Dictionary of the Khazars (1989), which takes the form of three parallel encyclopedias deaing with the same historical events. These all push the reader into a meandering, multi-pathed perusal of the text. Indeed, they typically confuse readers who try to read them as continuous linear texts. They all use typographical devices to direct the reader here and there, leaving the physical format of the book just as it has existed for centuries. Given the limitations of the book format, there was no alternative to typographic devices.

Even considering these earlier experiments, we might equally say that being able to move around the work at the click of a mouse has fundamentally changed the relation of reader and writer, who now collaborate to create the kinds of worlds envisioned in Borges’ “Garden of the Forking Paths” (which some hypertext authors, especially Stuart Moulthrop, cite as a legitimation of their enterprise). Readers, it can be said, now make their own books out of the materials the author has prepared, becoming in a real sense co-authors of the work.

Neither argument is persuasive. Computer-based hypertexts often resemble earlier experiments. But, just as often, they clearly have somewhat different roots. Whether hypertext fiction is “old” or “new” cannot be solved by investigating lineage, but rather by describing the kind of world that exists to support one or the other way of creating fiction. Print-based hypertexts depend on the already existing world of print literature, on its institutions, conventions, and audiences. Computer-based hypertexts, on the other hand, have created a new world of cooperative links: new writing tools, new forms, new marketing arrangements, and new readers. “New” and “old” can should be seen, then, as social constructs to be understood sociologically, rather than as literary facts.

Art Worlds

Exploring the problems of “old,” “new,” and the conditions of artistic change requires a conception of what sustains an art form, a conception of an art world which I have analyzed at length elsewhere (Becker, 1986). Though people often speak of art worlds in a loose way, I mean the term in a technical sense, which encompasses the following ideas:

All art works involve the cooperation of everyone whose activity has anything to do with the end result. That includes the people who make materials, instruments, and tools; the people who create the financial arrangements that make the work possible; the people who see to distributing the works that are made; the people who produced the tradition of forms, genres, and styles the artist works with and against; and the audience. For symphonic music, the list of cooperating people might include composers, players, conductors, instrument makers and repairers, copyists, managers and fundraisers, designers of symphony halls, music publishers, booking agents, and audiences of various kinds. For contemporary painting, an equivalent list would include painters, makers and purveryors of canvases, paints, and similar materials, collectors, art historians, critics, curators, dealers, managers and agents, such auxiliary personnel as, say, lithographic printers, and so on.

Why make such a list? Because each of these cooperating links is a point at which the people making the art have to consider how to take into account how the person at the other end of the link will cooperate—what they will and won’t do, and on what terms—or suffer the consequences. They have, for instance, to think about what it will mean to paint a canvas of a size that will fit into a home comfortably as opposed to one that will only fit in a corporation headquarters or museum. Taking the anticipated reactions of others into account, artists can decide to tailor what they do to what others will likely do. They can decide to paint a canvas that is two by three meters because they know it will be easier to find a home for it than one ten by twenty. In the same way, a composer might decide to write a string quartet rather than something for two ocarinas and bassoon, in part because there are many more string quartets than two-ocarina and bassoon combinations. If artists decide not to do what others want, they pay another price. Instead of giving up some of their freedom to choose, they must give up time to do themselves what others might have done for them if they were more cooperative; train others to do it for them; or do without. In each case, the work shows the effects of their choice.

All the people who cooperate in making a work of art do that by using mutually understood conventions. All sorts of aspects of art works are governed by conventional understandings as to how they can be done. Some common examples are: musical scales, which are a conventional choice of just a few from all the tones available; the three act play; the sonnet; the history painting; and so on. Such questions as size and shape, length and appropriate subject matters are all decidable by reference to conventional understandings as to how things should be done. Conventional knowledge is what makes it possible for musicians who have never seen each other to play as though they had known each other for years. It is what makes it possible for knowledgeable viewers or listeners to respond to a painting or musical work. Because you know what ought conventionally to happen, you can be surprised by an innovation which would otherwise be meaningless. It meant nothing special to hear Bob Dylan play electric guitar unless you knew that he had always played acoustic guitar. Using conventions makes it easier for people to cooperate and get the work of art done. Changing or ignoring them makes it harder and lessens the possibility of getting others to cooperate.

An art world, to give a technical definition, consists of the network of cooperative activity involving all the people who contribute to the work of art coming off as it finally does, using the conventional understandings they share. Most work gets made in art worlds. Some does not, whether it is the innovative work of art-world mavericks (e.g., a Charles Ives or Conlon Nancarrow) or the naive work of a Simon Rodia (the maker of the Watts Towers), who never heard of such a thing as the art world and wouldn’t have cared much about it if he had.

As the conditions of an art world’s existence—who gets recruited to the various roles, what kinds of resources are available, what kinds of audiences there are for its works—change, its internal organization and characteristic products change as well. Published fiction, and the organized world that produced it, changed radically when eighteenth-century England developed a new class of literate servants and business people who could read such work and wanted to. The modern novel was born.

One implication of this analysis is that, if we remember that one of the cooperating parties in the production of any work of art is the audience, we can think of a work as coming into existence anew every time someone looks at it, reads it, or hears it. This reminds us, and gives us a way to think about, the fact that the physical object is in a real sense not the whole art work, which is always being reinterpreted. The interpreter helps to create the work’s character as a result.

Art works get their value from art worlds. I don’t mean that art works aren’t agreeable or instructive or edifying or enjoyable,, only that they don’t have these qualities in themselves, but rather as they are commonly interpreted to have them, in a world of like-thinking people. Great works are great to people who know enough to understand them for what they are, as David Hume suggested. And we must remember that art worlds often reinterpret works, finding some valuable that they had thought less so, and vice versa. The works haven’t changed, but their value has.

Perhaps the most controversial thing to be said here is that the quality of a work is not affected by the kind of system it is made in. Good work (generally so recognized) has been produced under every sort of system, including the most vulgarly commercial. Think of the Hollywood film. It is hard to imagine, given the conditions under which movies are made, that there are any good films at all, but we know that there are. Or consider the Victorian English novel, whose authors had to take into account what publishers insisted on if they wanted their works to see print. J.A. Sutherland (1976: 114-16) has shown, for instance, how Thackeray’s Henry Esmond got some of its finest qualities from the intrusion of George Smith, a literate and concerned publisher, who wouldn’t pay the author until he took more care than was his custom.

Perhaps the most important thing to be said is that the participants invest the whole apparatus with an aura of “rightness,” so that this way of producing art seems moral and other ways immoral. Using classical ballet steps is moral and proper, while using more ordinary motions like running, jumping, and falling down is somehow wrong, an insult, a disgrace—to people attached to the world of classical ballet. To adherents of the world of modern dance, of course, it is another story.

These ideas, quickly sketched, suggest what needs to be looked at in understanding the development of an innovative literary form like hypertext.


So questions about the development of hypertext fictions become questions about the development not just of an artistic form but also of the whole world of cooperative activity which makes that form possible. We can begin by asking where hypertext came from. Contemporary fiction has its roots in the long tradition of written fiction, culminating in the story and the novel as the chief contemporary forms. We need not go into that well-known history more than to emphasize that writers in the Western literary tradition have experimented constantly with form and content. Hardly a decade goes by without someone proclaiming that the novel is dead, because all the possibilities have been exploited and there is nothing left to do. But hardly a decade goes by without some substantial new development in format and subject matter being accepted by knowledgeable readers. Forms which barely make sense to conventional readers are absorbed easily by those who are more daring and in time become just as conventional.

Hypertext fictions have grown out of the full range of available styles. Some are more or less conventional novels (e.g., of university life, or of romance). Some—e.g., Stuart Moulthrop’s “Victory Garden” (1991)—are relatively realistic, using the possibilities of hypertext to make it easier to tell a complex story, crisscrossing narrative lines in complex patterns which result fromthe intersection of readers’ choices with constraints the author has created. Others—e.g., Michael Joyce’s “afternoon, a story” (1990)—are experimental, using the same possibilities to create poetic, even mysterious patterns of words, creating seemingly endless loops in which the reader can feel trapped, leaving the the narrative unresolved. Some make use of typographical tricks to hide words inside other words. The transformations of the literary heritage computer-based hypertext make possible have only begun to be explored.

Computer-based hypertext has a second set of ancestors in the world of computer technology. Machines, as Latour (1987. pp. 103-144) has persuasively argued, consist of combinations of humans and technology, in which each is adapted to the other so that the two can effectively act as one. These modes of collective action in wich the actors are both human and non-human, what Latour refers to as “lashups,” make new results possible, provided that the parties are in fact well adapted to one another. The hardware and software combinations which make up a computer allow for the production of texts which are not constrained by the physical requirements of a conventional book. (Keith Smith(1989) provides an analysis of the multiple, largely unused, possibilities of print books.)

Computer programmers, alert to what potential users want, quickly learned that some industrial users wanted to access large bodies of material in a cross-referenced form. These users wanted, for instance, to consult mountains of technical documentation—e.g., the millions of documents which constitute the technical specifications an aircraft manufacturer needs to make an airplane—by going from any document to any of the many other documents which might be relevant—by simply choosing from the possible alternatives on the computer screen rather than leafing through hundreds of pages, books, or files.

Others wanted, when they used a computer, to consult technical documentation in order to know what command to give to accomplish something they wanted to do. Computer manuals, daunting documents even for experienced users, require you to leave the computer and leaf through a large, complex, and confusing printed document. How much better to be able to press the appropriate combination of keys and have the relevant technical “Help” file, part of what is typically called an “on-line help system,” appear on your screen! To access such a help system requires software that lets authors make multiple links between texts, enabling users to follow cross references to whatever degree they find necessary or interesting.

Designers of such computer applications did not intend to help authors write avant-garde fiction. On the contrary. Had it been only the needs and desires of authors of fiction that drove this development, it might never have happened at all. So, until recently, writers of fiction had to adapt what had been constructed for these other purposes to their own uses.

The implementation of the idea most easily available to the average person was Apple Computer’s HyperCard, a program with which people could construct “stacks” of “cards.” Cards could be cross-referenced in such a way that clicking in a particular place on the screen would take you to another card, the destination depending on where you clicked.

Another ancestor is somewhat nearer to fiction. Ted Nelson, in the 1960s, saw hypertext as a tool which would make all information of every kind, in every form, available to everyone (or, at least, everyone who would pay to use the system) at the click of a mouse (Nelson 1977, 1982). Though Nelson did not create a practical implementation of the idea, it grabbed many imaginations. Many people saw possibilities for such a wonder. Scholars could use it to comment on texts, every piece, say, of a classical text being linked to multiple commentaries on it, the whole being readable in a variety of ways (read all of one commentary, then all of another; read all the commentaries on one passage before moving to another; read the whole text and then the commentaries). Some envisioned multiply authored texts linked in multiple ways, readers being able to follow the line of thought of one writer, if they liked, or all the material on a given point or the development of an argument or conversation if that was what they wanted. The idea of a “web,” rather than a book or file of such materials, appealed to scholars. And in some cases, such a web seemed the only practical way to present someone’s work, as in the case of the papers of Charles Sanders Pierce, the American philosopher who left thousands of scattered and unorganized pages, all having multiple connections and relevances within and across pages. In a sense, Nelson’s Xanadu has now been realized, to some extent, in the World Wide Web, which makes all sorts of materials available, often in a form in which readers’ comments can be incorporated into the text, to anyone with access to the Internet.

A final set of antecedents might be computer games, which combine primitive narratives combining fiction with puzzles and hidden dangers to create a world through which the player can wander, seeking some goal—a treasure, or simply to escape from a dangerous situation—through a series of forking paths. The literary quality of these works was usually low (they were, after all, written primarily for teenage boys, though boys of all ages often found them appealing) but they pioneered some conceptual advances.

New Resources and Organizations

Remember that the creation of art works depends on the development of a world that provides artists with what they need to make the works they make. “What they need” includes materials, ideas, traditions, workers, and so on. So the development of a world of hypertext fiction requires more than the idea and more than the development of some computer software. But it certainly requires that, to begin with. The story I’m about to tell suggests some of the stages of the process, but is so truncated as to resemble an origin myth as much as an historical account.

Creating hypertext is a formidable task for a computer programmer, and few serious writers of fiction have the necessary skills, or want to have them. But some authors of fiction, to close in on our topic, saw hypertext as a chance to do something they had dreamed of doing. Michael Joyce, author of “afternoon: a story,” often called the first computer-based hypertext fiction, has said that what he wanted was to create fictions that would not be the same for any two readers, fictions whose construction made it possible for each reader to see the texts that made it up in a different order and arrangement. When he first had this dream, he knew little about computers, and had no idea how such a dream could be realized. He had an Apple II computer, but no program with which to do what he wanted to do.

The creators of the Macintosh, some of whom had a taste for Nelson’s utopian vision, produced HyperCard. But some found HyperCard difficult to learn; in particular, itdid not make the linking of texts easy. Though it could create hypertext of just the kind Joyce and others were looking for, producing such a text required a substantial detour into a kind of computer programming that effectively discouraged most would-be authors of experimental fiction from making the leap.

A kindly foundation executive who shared the vision made it possible for Joyce to spend a year at Yale University, where someone said “You should get to know Jay Bolter, who was here last year; he has similarly crazy ideas.” Bolter (a classicist who is also well-versed in computer programming) and Joyce then spent a number of years creating Storyspace, a program which significantly lowered the technical barriers to writing hypertext fiction.

Storyspace, and here I necessarily fall into the proselytizing language of the computer addict, “allows” you to do the following: create pieces of text of any size, symbolized on the screen by a small box with a title; link those pieces of text by drawing a line between them; control the reader’s access to these texts by specifying that Text B, for instance, can only be read by those who have already read text A and have also chosen a particular word in Text A; make many links, if desired, from one text, so that the reader may return to the same material from many places, each reading being conditioned by the differing materials which have preceded it; make all these mechanisms as clear or as opaque as the author wishes.

The illustrations of typical screens from a Storyspace document (Figures 1 and 2) show the author’s view.

ss screen

Figure 1

The small boxes, each with a title, can contain text. I have opened the box labeled “Lisboa” so that you can see its text, what a reader who entered that space would see in the course of a reading. The arrows between boxes are the links which determine what orders of reading are possible (in this case, there is only one path). Boxes may be nested, as is seen by the smaller box contained in “hypertext.” Figure 2 shows that, when you open up that box, you see the boxes within, and one of them is open to show the text it contains.

ss 2


Figure 2

These oversimplified illustrations suggest what the web of a full text, which would contain many more boxes and links, might look like to the maker

These resulting webs of texts, which can be very simple or extremely complex, can be packaged as a “standalone reader,” which is to say a document one need not have the original program to read. And this is how the major works of hypertext fiction in fact appear.


Writers, having found tools, still had to be able to get their works to readers. They usually wish to reach an audience beyond their own friends, families, and acquaintances, an audience constituted by “the public,” an audience of all those interested and prepared to help defray the costs of creation and production by paying something for the work. That requires some sort of distribution apparatus, whether state or private enterprises, which will produce the works, make their existence known through advertising and publicity, receive and fill orders, keep appropriate financial records, and so on. All quite commonplace, except when they don’t exist, or where those that do exist are not prepared to handle the kind of work you have produced.

Which, of course, immediately became a problem for hypertext writers. Conventional publishers did not “know how” to deal with hypertext books on disks. As the computer person of a major firm told me, their editors didn’t (at least at the time we spoke) know how to edit hypertext links, and their designers design books, not screens. Who might retail the works so produced thus becomes a serious question, which I will take up in a moment.

Mark Bernstein, for his own reasons, had had dreams similar to those of Joyce and Bolter, and was somewhat more entrepreneurial. He founded a company to distribute hypertext software and documents—Eastgate Systems—and undertook to provide the services I have just listedfor the new art form. Eastgate is now the major publisher of such work, being prepared as no one else is (though how long this monopoly will last one can’t say) to do the necessary jobs.

Suppose that hypertexts exist—the authors have written them and the publisher has produced them. You, as a reader, have heard about them and want to read one or two. Where can you buy a hypertext? When hypertexts received their first review in what is surely the most “important” review medium in the United States, the New York Times Book Review (Coover, 1992), the review contained a list of addresses and telephone numbers of the publishers of the works discussed, material ordinarily not found in a review. You don’t need such information to buy an ordinary book. But you cannot simply go to a store and buy a hypertext. What sort of store would you go to: a book store? a computer store? Since more and more computer materials are available on CD-ROM, would you perhaps go to a store that sells recordings? None of these sorts of stores know much, if anything, about hypertexts, have any idea whether their customers want them, or are sure that they should devote shelf space and other resources to them. Since no conventional kind of store stocks hypertexts, customers do not know where to go.


Distribution is, of course, intimately tied to readers’ difficulties with such unfamiliar texts. If I don’t know how to read this document, I will not spend twenty dollars for it and stores will not carry the item because they suspect that sales will be low.

To exist fully, works of art require audiences as well as authors. Work that resembles what is already available finds an audience easily, for it requires nothing of audiences that they do not know how to do. Work that differs from what is current will have predictable troubles.

A striking feature of current hypertexts, which embodies this problem, is that they typically begin with instructions as to how to read them. Introductory screens explain that each piece of text may (or may not) lead to more than one other piece of text. They explain that you can see the “next” piece of text by pressing the Return key or by using a mouse to click on words in the text. They explain that you may indeed see the same passage several times in the course of your reading; this will not be an error or a malfunction of the equipment. They explain that you can save this “reading” (that is, the sequence of choices you have already made) to be taken up again later, as you would “save your place” in a printed text with a bookmark.

Why do readers of hypertexts need explanations? Print books do not come with preliminary matter that explains how to open the book, then how to proceed by turning, or how to insert a bookmark if you wish to remember where you stopped reading. Everyone knows how to do these things. Ivan Illich (1993) has shown that it was not always true that “everyone” knew these things. The paragraph, the page, the table of contents, the index—all these tools which seem so obvious to us now—were not present in the scrolls which preceded printed books, and had to be invented and their use learned. (Claude Fischer (1992) has shown the analogous problems which arose with the introduction of the telephone. The technical problems might have been solved, but the problem of coordinating social action was not. Who, for instance, should speak first when the phone was answered, and what should they say?) Readers have further difficulties, once they learn all this, difficulties which are known generically, in the world of hypertext, as “the navigation problem” (Bernstein, 1991)c.That is, authors of hypertext documents of all kinds know, from bitter experience, that uninitiated readers (and all readers are uninitiated when they read their first hypertext) have difficulties beyond not knowing what to do. They also, and this is much more fundamental and serious, don’t know what to expect and often “get lost.” That is, they complain that they “don’t know where they are” in relation to the beginning and end of a text. They don’t know how to find their way to a point that will allow them to orient themselves to the body of texts that confront them. I had just this experience when I first read “afternoon: a story.” I put the disk into my Macintosh one evening at ten, double-clicked on the icon to open the document, and began reading the story. I met many characters, read many intriguing dialogues, learned the crucial dilemma of the text, which is that the narrator thinks, but is not sure, that he saw his son killed that morning. Engrossed, I read on and on. Some texts were repeated, but the repetitions led to material I had not seen before. In one long section, one word at a time appeared on the screen in a kind of poem. At one in the morning, I realized I had no idea how long this work was, how much of it I had read, how much I had still to read, and so (though it was by now very late) I called Michael Joyce at home. Fortunately, he was still awake. I said, “Michael, I’m reading your story and it’s wonderful, but when do I get to the end? When is it finished?”

I should have anticipated the answer. With a deep chuckle, he said, “Yes, that’s an interesting question, isn’t it?” He went on to say that I might consider myself finished when I lost interest or when I saw no new texts, but neither of those would guarantee that I had read everything he had written. He did not seem to mind that, although many authors would.

I was an as yet untrained reader of hypertext. After this lesson, I understood what I could and could not expect, what pleasures might be derived from a text organized in this way, what the interpretive difficulties would be. But I, after all, was a friend of the author and could call him for an explanation when I ran into trouble. The early audience for hypertext fiction consisted in large part, I think, of people who could get just such lessons from their friends or from teachers who were making these experiments (most of the early authors and theorists of hypertext did teach in colleges and universities). The audience thus grew slowly.

The development of a viable world of hypertext fiction that goes beyond such an intimate world of associates known to each other depends on many more readers learning these lessons than can learn them through private lessons from the already initiated. Such a development occurs when media addressed to a general, anonymous public pass on the news of the existence of such work and instructions as to how it may be read and why it should be read—what you can expect and why you should enjoy it. That did happen when the well-known American novelist Robert Coover published the lengthy account of hypertext fiction, its practitioners, and its distributors I have already referred to in the New York Times Book Review (Coover, 1992). What might once have been called a “coterie” or even a “cult” moved into the larger world, becoming known to people outside the inner circle of writers and devotees.

Coover’s article performed another essential service for the fledgling world of hypertext fiction by providing a critical assessment of its accomplishments, and suggesting critical criteria that might be applied to particular works. Such an apparatus, which announces and applies standards, and creates a forum for the discussion of the work, is an integral part of any fully-functioning art world.

Old and New

The ideas of hypertext fiction might be said to be “old,” in the sense that such conceptions had existed and been triedwell before computer versions were possible.

On the other hand, not very long ago that there were no computer programs to make the composition of multi-linked hypertexts a practical possibility, so the present condition is just as certainly “new.” More to the sociological point, since there were no programs to make hypertext fictions, there were no authors, no body of works to constitute the beginning of a “canon.” no publishing companies to manufacture and distribute those works, no readers who had acquired the skills necessary to read them and who had read a number of them and therefore were prepared to be a “public” for them, no critical writing about that body of work to inform the public about their existence and virtues—in short, no world of hypertext fiction. The list I have just drawn up measures the distance between a concept and a living art form, with all its institutional apparatus. In just the sense that there is now such a world where none was, hypertext fiction is new.


[These references contain, in addition to works cited in the text, other works that can comprise a small bibliography, as of the date it was written, for those who wish to pursue these matters further.]

Bernstein, Mark. “The Bookmark and the Compass: Orientation Tools for Hypertext users.” SIGIS Bulletin 9 (4 1988): 1-12.

Bernstein, Mark. “The Shape of Hypertext Documents: Form and Gesture in Hypertext Rhetoric.” In 36th International Technical Communication Conference in Chicago, 1991a.

Bernstein, Mark. “Deeply Intertwingled Text.” Technical Communication (1991b): 41-7.

Bernstein, Mark. “The Navigation Problem Reconsidered.” In Hypertext/Hypermedia Handbook, ed. Emily Berk and Joseph Devlon. New York: McGraw-Hill, 1991c.

Bernstein, Mark. “Storyspace and the process of writing.” In Hypertext/Hypermedia Handbook, ed. Emily Berk and Joseph Devlon. New York: McGraw-Hill, 1991d.

Bernstein, Mark, Jay David Bolter, Michael Joyce, and Elli Mylonas. “Architectures for Volatile Hypertexts.” In Hypertext 91 in Baltimore, ACM, 1-19, Year.

Bernstein, Mark, Michael Joyce, and David Levine. “Contours of Constructive Hypertexts.” In European Conferenceo n Hypertext in Baltimore, ACMYear.

Bolter, Jay David. Writing Space: The Computer, Hypertext, and the History of Writing. Hypertext edition, on a disk ed., Cambridge, MA: Eastgate Systems, Inc., 1991a.

Bolter, Jay David. Writing Space: The Computer, Hypertext, and the History of Writing. Hillsdale, NJ: Lawrence Erlbaum and Associates, 1991a.

Carlson, Patricia Ann. “Square Books and Round Books: Cognitive Implications of Hypertext.” Academic Computing (1990): 16-31.

Coover, Robert. “The End of Books.” New York Times Book Review, June 21, 1992 1992: 11, 23-25.

Cortázar, Julio. Hopscotch. Translated by Gregory Rabassa. New York: Pantheon, 1966.

Fischer, Claude S. America calling : a social history of the telephone to 1940. Berkeley: University of California Press, 1992.

Illich, Ivan. In the vineyard of the text : a commentary to Hugh’s Didascalicon. Chicago: University of Chicago Press, 1993.

Joyce, Michael. “Siren Shapes: Exploratory and Constructive Hypertext.” Academic Computing (November 1988): 11 ff.

Joyce, Michael. afternoon, a story. Hypertext edition ed., Cambridge, MA: Eastgate Systems, 1990.

Joyce, Michael. “A feel for prose: Interstitial links and the contours of hypertext.” Writing on the Edge 2 (2 1992):

Landow, George. Hypertext: The Convergence of Contemporary Cultural Critical Theory and Technology. Baltimore: Johns Hopkins Press, 1992.

Landow, George and Paul Delany, ed. Hypermedia and Literary Studies. Cambridge, MA: MIT Press, 1991.

Latour, Bruno. Science in Action. Cambridge: Harvard University Press, 1987.

Moulthrop, Stuart. Victory Garden. Hypertext edition ed., Cambridge, MA: Eastgate Systems, 1991.

Nelson, Theodor Holm. Computer Lib/Dream Machines. Redmond, WA: Tempus Books (reprinted by Microsoft Press, 1977.

Nelson, Theodor Holm. Literary Machines. Sausalito, CA: Mindful Press, 1982.

Pavic, Milorad. Dictionary of the Khazars: A Lexicon Novel in 100,000 Words. New York: Knopf, 1989.

Smith, Keith A. Text in the Book Format. Fairport, NY: The Sigma Foundation, 1989.

Sutherland, J.H. Victorian Novelists and Publishers. Chicago: University of Chicago Press, 1976.