The first time I clapped eyes on a hypertext, I immediately thought of Roland Barthes' essay, "The Death of the Author." In the age of interactive fiction, the author is not simply dead, I decided; s/he's been quicklimed. I was not alone. Readers encountering MIT's hypertext ASPEN for the first time, as Stuart Brand (1987) noted in The Media Lab, were astounded at "how 'unauthored' a creative work could become" (p. 141). Readers, long the undervalued links in the loop connecting author, text, and public, had at last come into their own as creators of texts in their own right, it seemed. The author is dead; long live the readers.
That was last year. My earliest perspective, as is common in initial encounters with an unfamiliar technology, took in more horizon than foreground or middle distance. After a year of reading and writing with a hypertext application, that judgment now seems a tad premature to me: perhaps not wildly improbable in terms of the agenda for hypertext and interactive fiction over time, but not entirely appropriate for the conditions presently governing hypertext/interactive fiction during its incubation phase.
Although we may already have examples of fiction which suggest an impatience with the restrictions and confines of the print narrative and a readiness to circumvent or even transcend them, we also have an electronic environment alien and inimical to our habitual reading patterns. As a result, the state of interactive fiction, at the moment, is a curious meld of invention and exploration. Interactive fiction, which anticipates the potential reach of interactivity, is coupled with a hasty resuscitation of previously verboten interpretive strategies and a whole lot of bewilderment.
Printed text, like this article, is linear: Words and ideas follow one another in sequence. Although some of the ideas related here may also be linked in another manner--for example, conceptually--those other relationships are constructed only in our thought processes and do not exist physically in the text. Hypertext, however, enables us to display these alternative ways of ordering text and argument graphically and to make them physically present in a text. Although we may be able, as in the hypertext program STORYSPACE, to view an entire text as a graphic map showing connections among places in the text, once we drop into the text itself to read it, the map becomes invisible. The reader is faced with a window containing text, much as s/he would be when using MACWRITE, MICROSOFT WORD, or any other standard writing application. Yet, although the text in these windows may be scrolled, the reader must also make additional decisions about how to move through the text: either selecting words of particular interest or using the directional buttons lining the bottom of the screen that allow the reader to move through the text (left or right, up or down, backward or forward) as it has been mapped out.
Although hypertext has been in existence since the Sixties, only
within the past year or so has it emerged as a viable entity for
the average microcomputer user. Apple's HYPERCARD promises to
yield some interesting entries into the field of interactive fiction
. Another fertile hypertext environment for the development of
interactive fiction, STORYSPACE 5.3, the brainchild of Jay Bolter,
Michael Joyce, and John B. Smith, is slated for publication in
1989. Currently existing in a beta version, STORYSPACE features
a separate Reading Space application permitting read-only access
to text, as well as a reader interface integrated into the writer's
desktop (Figures 1, 2, and 3).
For the writer, the text opens as a graphic map, displaying the position of pieces of text represented by, and contained in, small, box-like windows called "places." Places may be dropped into other places to create hierarchical relationships, or they may be linked by paths. These paths are represented in the map by arrows, or by words or phrases supplied by the writer to define guard fields. Readers may move via paths between places. These paths become default links, accessible with a simple carriage return. Readers may also move through the guard fields defined by logical options. These guard fields will, for example, yield only if readers select a key word or phrase in the text existing within a window, or if readers have already visited a specified place. In some cases, readers may become frustrated at failing to discover yields in the text or at finding themselves in a relentlessly looping segment of text. In such cases, readers can rely on an option that displays the number of links leading from a place, and the names and destinations of these links. Readers can then select and follow one of these links. Should one of these links lead again to an irritatingly familiar textual loop--as occasionally happens--readers can move through the text by using directional commands in the reader interface that permit vertical and horizontal movements. These commands allow readers to circumvent authorial links and, instead, follow the configuration of the text as it is mapped out on the desktop. Movement left, for example, will situate the reader in the place immediately to the left of the previously encountered text, regardless of whether or not this place has been linked to that text by the author. Finally, for the genuinely bewildered and/or frustrated, the reader has the option of quitting the reader interface entirely and of viewing the whole interactive fiction in graphic map form. Readers can then select and read places within the structure discretely at will--or whim.
Novelists have been chafing at the typographic bit since the time of Tristram Shandy, struggling to circumvent the conventions that situated their fictions in the world and their readers within the literary universes of their fictions. Not surprisingly, the two interactive texts written in STORYSPACE--Stuart Moulthrop's "Forking Paths" and Michael Joyce's "Afternoon: a Story"--follow two of the divergent modes already explored in print narratives. Texts like John Fowles' The French Lieutenant's Woman, John Barth's "Lost in the Funhouse," Robert Coover's "The Babysitter," and Jorge Luis Borges' "The Garden of Forking Paths," all use print's linear and sequential nature to explore simultaneity and multiplicity in plotted events. Although occurrences in these authors' plots unfold singly and in order because of their typographic environment, multiple possibilities for developments in the plot and conclusions proliferate. Readers are given the opportunity to experience multiple endings, as in the dual conclusion to the Fowles novel--although one scenario actually does physically end the novel. Ironically, in the very act of manipulating the narrative conventions governing print space, these texts dramatize their inevitable failure to escape those confines. In the Borges short story, time and possibility are literally equated with a labyrinth where an infinite number of plot configurations are not only possible but can be realized; Borges, however, deliberately permits only a single outcome to conclude the story, and the gesture toward liberating the narrative from print boundaries simply reminds us how constricting a typographic framework really is.
Like its print predecessor, Stuart Moulthrop's "Forking Paths" uses the labyrinth to suggest a universe where an infinite number of scenarios may be equally, and even simultaneously, realized. But in Moulthrop's hands, the narrative itself physically becomes labyrinthine. Seen in map view, the text is a dense matrix of crisscrossing strands existing on a single plane; there are no hierarchies in Moulthrop's text, in terms of its physical configuration--no vertical relationships, only horizontal ones. The Borges original has become a hydra. Fully fledged narratives branch out from character names or possible scenarios suggested in the short story, all potentially available for discovery, development, and conclusion in the interactive story--should the reader elect to pursue them. "Forking Paths" offers twelve separate permutations on the ending to the Borges tale, as well as retellings of strands from the original story from alternative points of view; unexpected reversals in character traits and motives; and occasional, playful, metatextual commentary on the nature of interactive text itself.
In fleshing out the Borges story, Moulthrop realizes one of the most obvious potentials for interactive narrative: its ability to provide us with the option to choose, as readers, which narrative scenarios we wish to resolve. In this instance, we are able to exercise a certain amount of intellectual clout. We can decide, for example, that we are not satisfied by a single conclusion and continue reading after we discover Yu Tsun has murdered Stephen Albert. Or, we can decide in one reading that the most satisfactory resolution to the confrontation between English scholar and Chinese spy would involve the two peaceably entering the labyrinth together, and we can later return to the text and find that the hitherto peaceable Stephen Albert has unexpectedly garroted Yu Tsun. For the first time since Plato lamented the immutability of the written word in Phaedrus, readers have open to them the possibility of reexperiencing new narratives in the act of returning to a text. Formerly, a rereading of "The Garden of Forking Paths" would involve the reader suspending certain knowledge of the demise of both Albert and Yu in order to experience again their actions in the early portion of the narrative; with a hypertext environment, rereading of "Forking Paths" promises multiple, although not infinite, differences in a number of readings of the narrative, coupled with the option of choosing which narrative elements to foreground and which to ignore.
If the first approach, based on a print narrative, moves toward providing its readers with greater autonomy in deciding which developments and conclusions are more preferable or probable than others, the second narrative approach, using hypertext, presents its audience with discrete pieces of information and ellipses, requiring a participatory act by the reader nearly tantamount to the creative effort expended by the author. Examples of this mode, like Barthes' The Pleasure of the Text and Julio Cortazar's Hopscotch, resemble textual mosaics. Experience has already been parsed for the reader by the author-figure, but the reader must decide how it is to be reassembled. Accordingly, the order in which Barthes' and Cortazar's texts are configured relies less upon established narrative conventions and more on connections realized by the reader. The Barthes text takes its alphabetic order from what its author conceived--by his own admission--as the most nearly arbitrary system of arrangement possible, while the Cortazar novel (which also may be read conventionally, beginning with chapter one and moving forward) pursues the physical sense of interactivity. The Cortazar novel requires its readers to hopscotch through the novel's typographic space, jumping from chapter seventy-three to chapter one to chapter two to chapter one hundred sixteen to chapter three to chapter eighty-four and so on.
Yet, ultimately, when all is said and done, we read and interpret these texts as we read and interpret their more conventional counterparts. All print-based texts are constituted by events which follow one another in linear sequence. And they are immutable, despite the potential for Cortazar's novel to be reconfigured after a first, conventional reading. Like the written narratives bemoaned by Plato in his Phaedrus, these texts just keep saying the same thing no matter how many times we read them. But, living in virtual space with some of its narrative strands tangibly, physically more marginal than others, Michael Joyce's "Afternoon: a Story," conceived for and written in STORYSPACE, succeeds in realizing the aim of mosaic texts like those of Barthes and Cortazar. The content of certain places within Joyce's narrative acquires markedly different interpretations and force depending upon both the frequency and context in which these places are read. And, as the BACK option in the reader interface often shuttles readers into a different narrative loop from the one in which they are presently situated, the immediate context in which any place is read acquires somewhat the flavor of existential reality. When I encounter a short two-place narrative strand involving a runner collapsing during a marathon sandwiched amid a lunchtime conversation between two characters, it acquires the status of a flashback with some resonance for the quality of the relationship between the two conversants. When I stumble across this short narrative sequence again, in the context of the narrator's anguished recollection of an accident he witnesses whose victims may or may not be his estranged wife and son, the strand foreshadows tragic discovery, suggesting events beyond human control or comprehension. Because the present state of the reader interface precludes an autotracking mechanism for a read-only mode, I cannot easily recover the other context for this particular narrative sequence when I encounter it for the second time--unlike my experience of the Cortazar novel, where paging back and forth to recover the coloration of the text from the first reading is only slightly more complicated than, say, tracking a story through the Sunday Times. I am, in a tangible, physical sense, unable to fully extricate myself from my present experience of the narrative to compare differences in coloration between the passage read in two contexts; I am, instead, obliged to attempt to recover the sense of my early encounter with the text against my second brush with it--not unlike our experiences with sensation and recall in our everyday reality outside fiction.
In truly interactive text, readers navigate through the narrative based upon words--or associations with words--selected from the text before them. Simple movement through relying on the default option, which reaches passages by a simple carriage return, reduces the level of interaction among author-figure, text, and reader to typographic dimensions: The act of reading somewhat resembles listening to a lecture. On the other hand, were navigation only possible through textual yields, the act of constructing the narrative might, arguably, become sufficiently arduous to discourage all but the most stalwart readers from engaging themselves in interactive fictions. We might guess that the combination of both strategies--default and textual yields--coupled with the multitudinous possibilities for links from any given word or phrase and for links from the large number of potential yields in any given window, would prevent readers from attempting to form any expectations for textual configurations. Instead, what is remarkable is the rapidity with which we must begin to identify or assume patterns in narrative configurations in order to comprehend the text before us. Within one or two places of our initial entry into the text, we begin to form expectations about, for example, the presence of default options or textual yields. We can test these expectations by relying on the BACK option, to confirm whether we have moved forward by default or by having matched minds with the author-figure who created the matrix lurking behind the text. If our movements have been frustrated by lack of yields or defaults, we can choose to view potential links between our present place and others, and to make interpretive decisions on the actual content of each place based upon its placement in virtual space. For example, the places in "Afternoon: a Story" that ultimately occupy positions central to our understanding of the narrative are those tangibly central to our reading experience. The single place that comes closest to approximating a closure point in the narrative is linked to nine other narrative strands, making it potentially the most frequently encountered bit of text in the story. Similarly, the failure of the single place to default, making it the sole place in the text without a default, gives the reader pause, just as its failure to branch at all--should the reader have completed one particular link of the nine possible strands--brings the text to a sort of stylized conclusion.
Which brings us, by the long way round, back to the Quickliming of the Author, so to speak. More so than in typographic narratives, authors of interactive fiction act as scriptors for the potential experiences of their readers. In some respects, evidence of their traces is less obvious than it is in a typographic environment. In a hypertext document, authors leave behind multiple traces, alternative, spectral literary worlds. But, at least at present, because of the absence of established literary guidelines or conventions for readers to rely upon in navigating their way through text or making interpretive choices, we find ourselves resurrecting the figure of the Author. The yields we select, the defaults we discover, influence our understanding of the contents of the text we read; in most cases, we realize that we have, unwittingly, made certain interpretive or navigational decisions based upon our apprehension of authorial intention--a discovery which might make every critic since Cleanth Brooks roll over in his or her crypt. An author-figure has, after all, sketched out all the possible worlds that the reader might realize during any number of encounters with the text. The readers must guess at some of the author-figure's intentions, at some point, in order to move forward in a form of shadowplay between reader and author-figure not unlike the game "Mastermind," or the conundrum of the Prisoner's Dilemma. When we arrive at places or textual links that defy the expectations we have formed regarding them--as I did when I encountered an isolated place through the LOCATE option in "Afternoon" with no links attached to it whatsoever--we are obliged to make assumptions about these places/textual links' existence in a larger structure regarding the intended shape of the work. In my case, I assumed that Joyce had used the place to approximate a form of closure resisted in electronic text; the place, however, turned out to be a branch which Joyce had abandoned but neglected to delete.
With the emergence of conventions that will situate readers somewhat less precariously, if bewilderingly, in virtual space, the resurrection of the role of authorial intention in interpretive strategies--long verboten in literary criticism--may prove merely temporary, a glitch, a primitive holding action. After all, the loquacious first-person narrative endemic to the 18th century novel, aimed at cementing its audience into a certain receptive and interpretive stance, faded away once the novel became firmly situated as a form of public discourse. What we can expect