Whither Leads the Poem of Forking Paths?

Robert Kendall

When a lifeline of words is dangled for an instant before meaning about to go under, or when some desperate insight pulls a knife on language, what happens next is poetry--that extraordinary product of extreme circumstances in which every verbal action has to count. A line of poetry is a walk along a high ledge, and one false word can mean a plunge to the prosaic parking lot below. Nothing less than "the best words in the best order," to quote Coleridge's no-nonsense definition of the art form, will get the reader to that rooftop where the view of the world is broadened. (Samuel Taylor Coleridge, Table Talk, 1827)

At least that's the story from one side of the tracks. In another part of town Dada cutups, automatic writing, found poems, and procedurally generated texts have settled in. For decades they have flirted with or embraced the arbitrary to challenge the institutions of aesthetic convention, the top-down social values those conventions represent, and even the repressive overlordship of the conscious mind. Meanwhile the deconstructionist renovation project moves from one neighborhood to the next, putting the whole notion of authorial intention increasingly under suspicion. Is the pursuit of the perfectly wrought poem becoming as ideologically unconscionable as the hunt for elephant ivory?

The hyperpoem, that recent arrival from the cold northern capitals of information processing, would seem to add more weight to this question. It can go the old-timers one better at the extra-authorial sweepstakes. No matter what part chance may play in fashioning a linear poem for print, the author still gets "final cut," ultimately determining what goes down on paper. Hypertext, on the other hand, is enduringly indeterminate, a permanently provisional enactment of virtual pages left to flap in the winds of reader interaction. A poem in hypertext may be finished but never fixed.

Reading hypertext isn't the same as drawing shredded text from a hat, but the rules of both games come from similar-looking handbooks. The author may carefully plan out each intersection in the network of hypertextual pathways but can't anticipate all the possible combinatorial routes through the text. The reader/traveler of the branching and intersecting text trails may carefully choose the way at each turning, deliberately weighing each option, but can't know the full significance of each choice until after it's made. Author, reader, and Lady Luck each have a finger in the pie.

Literary hypertext seems to reverberate with the sound of the door slamming shut on some traditionally important literary ideals. When it comes to ideological territory, however, new art forms can be more ambulatory and less inclined to burn their bridges than one might suspect. By way of demonstration, let's explore the rather paradoxical-sounding notion that hypertext could be the only means of actually exorcising the arbitrary from the poem--that the pursuit of "the best words in the best order" may inevitably lead the poet down the hyperpath. Let's start by taking the notion of textual multiplicity and holding it up against the nature of linear poetry.

Cut to Poetry Nature Preserve. Here the author traditionally tracks poetry through the wilderness of imagerial and philosophical underbrush that sprawls in the mind. Our well-thumbed Baedeker tells us that the creative process is one of finding the "right" route through this inner landscape--the route that is mapped out by the finished poem. The poet's successive drafts will evidence the many wrong turns that mark this journey of discovery. The writer will lay down the pen only when the process seems to have successfully arrived at something. When we read and contemplate the finished poem in print, it will be as if we're retracing the poet's steps through this inner landscape.

Now the hypertext theorists step forward to point out that when we follow the linear flow of words down the page, there is often not just a single route. Look at all those unmarked intersections in the form of ambiguities, allusions, and levels of symbolism. They lead in many different interpretive directions at once. Don't we, in fact, judge a poem largely by its ability to get at places that aren't accessible by the main road of direct statement? If, as Umberto Eco puts it, the poetic effect is "the capacity that a text displays for continuing to generate different readings, without ever being completely consumed," shouldn't we welcome anything that can deepen its textual multiplicity, such as, oh, say, hypertext? (Umberto Eco, Reflections on the Name of the Rose, 1983)

Sounds logical. But when a poem resonates in different ways, sending variable frequencies of meaning like sonar into the various nooks and crannies it's trying to plumb, do these multiplicities always suggest true multilinearity? A short poem may exude multiple meanings, yet these may be mere passing lanes on a single road progressing from the beginning to a rhetorical punchline at the end. The linearity of such a poetic specimen may be an inviolable part of its nature and even a necessary foil to any subtextual digressions, giving them something to push against. A creative impetus can end up nailed to the page in a particular verbal form simply because out of all its many possible verbal incarnations, that one happens to be the best--the most resonant, the most far-ranging in its different implications. Multilinear variants would just water down the broth.

On the other hand, when the multiplicities multiply beyond a certain point, we reach the outskirts of the hypertextual. Printed poems have been known to go overtly in several thematic directions at once--like dull roots stirred with spring rain--or circle around their subjects in unpredictable orbits. (Modernism, anyone?) Particularly as a poem grows longer and more complex, perhaps subdividing into semi-independent episodes, the relationships among its components become more multifaceted and less clearly linear. Instead of a steady progression toward the closing line there may be an irregular shifting toward and away from various motifs as the poem unfolds. Each new line may build upon the line immediately preceding it or upon something that came up many lines back.

A certain challenge arises in circumstances such as these. The poet must realize a conceptual structure that's supported not by a single developmental thrust, like a lone pillar, but rather by a complex of interconnected thematic elements, like a profusion of intersecting walls. Yet however multidimensional this thematic framework may be, the poet must reconcile it with the single dimension of a column of print. This isn't always a painless task. It means settling on specific orderings and juxtapositions of material that establish or nourish particular relationships and leave other potential connections dangling in the breeze. It means deciding how to chop up many a thematic thread in order to maintain it throughout the poem by dropping the fragments periodically into the linear flow. It means leaving some threads truncated when there's no good way of splicing a continuation into the textual rope. Finally, it means bestowing the conclusive weight of The Ending upon whatever comes last, even if that weight could equally well serve other portions of the text or if the poem would be better off without it altogether.

The poet labors to ensure the best possible effect from the way the printed poem's many strands interweave and its many parts interrelate in the temporal reading experience. Success depends partly upon the poet's diplomatic finesse in arbitrating conflicts among the poem's many competing agendas. Certain individually appealing possibilities must be sacrificed for the greatest common good. Logic and intuition are on the advisory staff, but when push comes to shove, sometimes it takes an arbitrary hand to get things done. For example, suppose a group of episodes could conceivably be ordered in a number of different ways. Suppose each potential ordering has a strong structural argument in its favor, yet none asserts itself as clearly superior. The poet may be quite likely here just to leave things in the order in which they happened to have been written.

The more complex and expansive a fixed linear poem becomes, the further it may lead the poet into the territory of compromise and sometimes arbitrarily determined structures. What if a poet feels a little uneasy in this territory? Hell, just chalk it up to the limitations of language and keep pushing on. Or chalk it up to the limitations of linearity . . . and reach for the computer.

Hypertext doesn't demand that the multifarious structural possibilities inherent in a poem be pared away to accommodate the sheath of linearity. The poem's multiplicities can stand unashamed as the structural framework itself. Interrelationships among the poem's parts become actual links rather than implied connections. The painted-on trompe l'oeil doors become real, letting the reader open them to explore alternative juxtapositions and interweavings. Images and themes expand over the once-closed linear borders, gaining room to stretch and breathe, perhaps setting out to explore what could otherwise only have been glimpsed from the highway. The wheedling voice of compromise falls silent as all the fruitful configurations are laid on the table at once.

Or does it? Is the hypertext author really eliminating a reductive and sometimes arbitrary linearizing process or just foisting it onto the reader? After all, the reader can't directly absorb the full panoply of hypertextual multiplicities, but can only cobble together an individualized linear reading that encompasses some of them. The reader doesn't have the author's intimate knowledge of the poem's materials, but in fact knows next to nothing about what hasn't yet been read. From this position of ignorance, how can the reader avoid making structural decisions that are not only arbitrary but sometimes downright bad? All the navigational options available at any given moment may have equal structural validity on a purely local level, but this doesn't guarantee that the cumulative result of all the reader's choices--that is, the reading itself--will be cohesive and satisfying.

In practice, being proffered an inspired array of reading possibilities is one thing, turning them into an inspiring reading experience is another. Any time spent with the many literary hypertexts now available for public consumption from Eastgate Systems or on the Web will reveal that coaxing a meaningful read from such work is a sometimes messy business. The writing may seem to wander at times. One may reencounter passages with annoying frequency, while there's no guarantee that key episodes won't be missed altogether. Looking for the continuation of a thematic or narrative thread, trying to return to an interesting section, or simply trying to find unread material can be a process of wrong turns, backtracking, and struggling to remain on one's cognitive feet. As a reading progresses, one must inevitably spend more and more time roaming tracts of familiar text in search of something new, and it can be damnably hard to know when one has finished a hypertext.

One's initial foray into hyperreading may suggest that the sumptuous theoretical garb of the multiply realized poem is in danger of looking rather shabby in the cold light of actual practice. Yet, if one is patient enough with a good hypertext, the beauties of its protean structure will emerge. As nodes recur more and more frequently and reading and rereading gradually blur together, one can sort out seemingly redundant recurrences of nodes from those that put old text in a revealing new context and clarify structural interrelationships. From the reading/rereading that actually unfolds, however flawed it may be in its details, one can derive some sense of the more nearly ideal readings that are contained within the piece and the ways these complement one another. One can in retrospect "edit out" the missteps, redundancies, and ill-placed episodes as the overall image of the work congeals in memory.

Some of the imperfections in a hypertext reading can even be deemed an integral part of the multidimensional texture, the byproduct of an almost tactile experience of feeling one's way through the work. For example, a reader might "try out" several links from the same node by selecting one, backtracking, selecting another, and so on, before settling on a route to follow. This may leave the reader less with the sensation of having sullied the flow of the reading than with a feeling similar to having run a curious hand over a physical object. Some sense of resistance to the onrushing stream of reading can be a benefit. Without it, how can the reader really grasp that there is more to the hypertext than the virtual linear text that is generated from it in real time?

Finally, the reader's capacity for mental reconstruction can help transcend the work's real-time flaws altogether. It can spawn a mental image of geometric relationships among the linked parts of the poem. With a short or simply structured hypertext, such efforts to conceptualize the work spatially can be quite successful, especially if the writer helps out by providing a graphical map depicting the nodes and links in a geometric pattern. A reader can navigate a piece as if exploring a building and its grounds, not just putting together a reading but also creating a mental model of the text's architecture. Many of the work's multiplicities can in this way be perceived simultaneously in the mind's eye.

With a complex work, however, any underlying geometric structure may be incomprehensible in its labyrinthine entirety. Key structural relationships may resolve themselves into geometric outlines, but even these may seem flimsy and ambiguous. Efforts at decoding a theoretically pristine hypertextual geometry can therefore go only so far. There's also a drawback to our other method of cognitive reconstruction, the mental editing technique: Carrying a bottle of whiteout around in your head can simply grow tiresome after a time. Must a hypertext be deciphered like a set of blueprints or corrected like a bad galley proof? Can't it simply be read?

We might reply that if it could simply be read, the reader would lose the deconstructive benefit of being compelled to reconcile the artificial purity of art with the messy realities of life. We might say that hypertext poetry is uniquely real simply because the very process of actualizing it forces it out of the penthouse of High Art Towers and into the mean streets of an imperfect world.

This wouldn't silence the unsettling questions, however. Must we approach hypertext as we would the efforts of a talented but immature writer, making accommodations for the faults and rough edges, trying to appreciate the qualities that often seem but half formed? Must we make concessions to the medium's limitations and therefore apply lower artistic standards to hypertext than to printed poetry, as we would when judging, say, synchronized swimming rather than dance? If so, does this mean hypertext is intrinsically inferior to fixed linear writing?

I think all it means is that hypertext is younger than linear writing. What we see today is just the beginning of a development that will likely take many years to come to full fruition. We're witnessing the birth of an entirely new kind of writing, one that has no close parallel in human history. The newborn I'm referring to is not hypertext per se but the computer programming through which it is realized. Though we often loose sight of this, computer code is a species of writing. Programming languages actualize procedures and principles rather than signifying objects or concepts. But their goal, like that of the natural languages we speak and write, is at least partly to embody human thought.

Hypertext poetry is an effort to integrate these two branches of language to form a richer, deeper medium of expression. The hypertext system complements the expressive capabilities of natural language by representing the dynamics underlying the words. The words are the final product of emotional and intellectual processes, an artifact of something in the author's head. The computer interface can embody a thing more primal in character--the impetus behind the poem's text. The configuration of navigational options and constraints can parallel the laws of nature that first brought the poem's metaphorical world into being and then seem to operate within that world.

Readers of hypertext don't actually create their own texts, though this has become a popular trope for discussing the medium. A hypertext poem is not to be confused with a blank word-processing screen. Rather, hypertext readers take the reins of a creative process that is uniquely the author's but also no longer the author's. The interface is a harnessing of a drive toward verbal expression, an enactment of the inner guidance system that leads the author in the quest for the text.

The success of the interactive reading depends, therefore, upon how effectively the writer's artistic impulses and vision are encoded in machine language--quite a different matter from encoding the products of these in verbal form. Needless to say, the poetics of programming are less developed than those of the more familiar variety. As an expressive medium, executable binary code has managed to reach a state of development roughly equivalent to writing when it was preserved on unwieldy clay tablets, understood by few, and incapable even of representing the entire vocabulary of spoken language. Just as no one could have foreseen the full flowering of written literature from such inauspicious beginnings, we can only guess at what interactive writing may hold in store.

As hypertext progresses through the school of hard knocks and its mechanisms become more accurate surrogates for the creative process, the quality of the hypertext reading experience will increase commensurately. A more street-smart hypertext will carry much of the author's highly developed sense of the work's overall structural possibilities, letting readers tap into this. They'll have more direct and intuitive control over the big picture. They won't be confined to interacting with the text on a purely local level, merely following one link after another and hoping it all comes together in the end. They'll be able to alter the structural principles governing the layout of nodes and links, interacting directly with the forces at work in the writing.

A more street-smart hypertext will solicitously monitor the reader's progress and anticipate the sorts of choices that its human charge would probably want to make at any given moment. It will know not just which sections can be effectively juxtaposed--that is, how the author has strung nodes together--but also how the significance and appropriateness of these juxtapositions will change in the context of the differently shaped readings in which they may figure. As the reading progresses, it will reconfigure not only the road signs that confront the reader but also the roads themselves whenever necessary to keep things moving. It will ensure that each step the reader can take will lead to something effectively complementing the text that has already been traversed. Flaws in the reading need not be an issue, because the author's (binary-encoded) judgment will be at the reader's beck and call to eliminate them before they can arise.

Hypertext is still settling into its new literary quarters, hoping to stay awhile. Gaining long-term acceptance will ultimately depend on how well it finds its way around among the many demands and expectations of the writing/reading world. This will become easier for it as writers (and programmers) become more proficient at working directly in a new artistic medium: possibility. Mastery of this medium will give rise to a hyperpoetry that is secure in its uniqueness yet as resilient as any of the genres of linearity--a hyperpoetry that is malleable to the nth degree yet produces readings that are all equally polished in detail, solid in overall structural integrity, and revealing in their differences from one another. Then, perhaps, we may grow used to thinking of poetry in a new way: as the best words from the best possibilities.



Robert Kendall's electronic poems have been published by Eastgate Systems, The Little Magazine, and Saint-Gervaise Genève. His book of printed poems, A Wandering City (Cleveland State University Poetry Center, 1992), won the CSU Poetry Center Prize. He teaches an on-line class in hypertext poetry and fiction for the New School for Social Research.
To explore a practical attempt at hypertextual street smarts, as they're hypothesized here, see the author's long hypertext poem A Life Set for Two (Eastgate Systems, 1996).




Copyright © 1997 ebr and the author. All rights reserved.

riPOSTe:ebr@uic.edu