> vox

Computers and Poetics

by Robert Pinsky

5 February 1997
MIT Media Lab Colloquium
Cambridge, Mass

Here is the question I proposed for my topic when I was kindly invited to talk to the Media Lab about interactive literature:

Do literary technologies such as narrative, verse, rhyme — the whole traditional range of rhetoric — have figurative analogies or actual usefulness in enhancing the artistic effects of autonomy or responsiveness in interactive programs?

To rephrase the question, is it possible that those old technologies are to be not merely included within the interactive or hypertext work, as still photography is in a way included within the technological means of motion pictures, but in some sense instructive models, as the critical and aesthetic theory of still photography might be included within our understanding of motion pictures?

I'll beg your patience while I begin by very briefly discussing poetry. Any medium has a shape, a nature or genius, a set of limitations that are opportunities that are never entirely defined, always to be further discovered. Nick Montfort has reminded me of Rudolf Arnheim's distinction between the engineer, always striving to eliminate or conceal the limitations of an art and the artist, always striving to exploit those limitations, and turn them to advantage, using the defining genius of the medium, which is also its limitations. My effort will be to make some speculative guesses about the particular genius of computer text works, thinking by analogy with the genius of poetry.

It might seem that I am stretching a point, or trimming my material, when I speak of "the technology" of literature, but in fact that is a term that I use when describing poetry in contexts that have nothing to do with MIT or binary computers. Thinking about the art of verse, I have concluded that poetry is an ancient technology, using the primate grunts of vowels and consonants in the service of memory. The medium of poetry, I have concluded, is the human body: specifically, the column of air inside a single human body, shaped into meaning sounds within the voice-box and the mouth. The ancient technology of poetry is peculiarly intimate: the most bodily of all the arts.

Poetry is the most bodily of the arts. An obvious critique of this definition is — no, it is not: dance, for example, is more bodily.

But the medium of dance is the body of an expert, an artist. The medium of poetry — which I consider essentially a vocal art, but not a performative one — is that single human body, and not necessarily the body of an expert. The poem, for me, is not its sounds as performed by a poet or accomplished performer, projecting passion and personality skillfully for an enraptured audience. Rather, the poem is its sounds as I mumble them by heart while driving to work or in the shower, or as someone says them aloud, from page or memory, for a friend or lover. Cavafy or Dickinson or Ginsberg or Shakespeare live in the way someone's breath follows their orchestrations of vowel, consonant, cadence.

The medium of poetry, in other words, has an ancient, built-in individual scale. Despite nostalgia for the days of Soviet stadiums, the poem embodies both the universalizing, social qualities of language and an individual, personal scale.

This is the first analogy I propose: in an age of powerful, often brilliant mass art — a term often, and I think perniciously, confused with popular art — poetry as I have defined it embodies a resistant power, stubbornly tied to the scale of a few people at a time, a communication that can be called intimate even when the distance — as say, between me in lower-middle-class, mid-twentieth-century, apostate Jewish New Jersey and George Herbert in upper-middle-class, piously Christian seventeenth-century England — is vast.

I think that a similar element pervades the great cultural stirring that surrounds the binary computer and its powers to span distance, to arrange and rearrange information, to respond quickly, to link messages, images, ideas, persons with glorious speed and lordly indifference to old divisions and separations. Within the vastness of the Web and the dazzling speed of the machines there is a loud and proud, cussed individualism. The personal Web page has become a kind of folk art, like the fiddling and piano playing and quiltmaking of the previous century, as expression, it seems to me, of half-conscious resistance to mass art. I think that the interest in interactivity — an interest that sometimes sentimentalizes interactivity — expresses a similar political and cultural longing or nostalgia for individual scale.

This is a profound political matter, one I understand quite tentatively. In the web of electronic communication that covers the globe, at the moment effortlessly crossing all sorts of boundaries and transcending all sorts of regulations, the idea of individualism flourishes at a time when in most cultural areas the concept seems outworn, sometimes even anathematized. Works of art designed for access through the personal computer may exploit this cussedness, this particularity, in unheard of ways. My personal hope is that these works of art act as a contradiction and counterweight to the Clinton Administration's metaphor of the Information Superhighway. The Interstate Highway system is admirable and useful, one of the great accomplishments of our civilization. But as a regulated, Federally funded, restricted-access network, created and maintained at immense profit, it provides a dampening figure of speech.

In my understanding, which is dim as well as tentative, we are in a perhaps brief moment when various immense corporate entities are struggling with one another for control of the information web: because these large creatures are competing for this prize, we small creatures have a certain freedom of it. Conceivably, some portion of that freedom could be staked out or preserved. I can imagine that the first truly great works of art in language to be composed for this still infant medium might define that relatively free area. Such works would enact the relations among language, the binary computer and the human soul in a way comparable to lyric poetry's relation to language, the human soul and the body it inhabits.

I should say that those works may already exist: there is so much varied and intense activity out there that one can skim only a little, fairly randomly. This is an immense world of serious writing that is only beginning to be recognized by print-oriented writers. A couple of years ago, when I published a little essay in the New York Times Book Review about computers and poetry, I was deluged with floppies and e-mails of works — hypertext narratives, multimedia works, interactive poems and so forth, some of it at a pretty high level. Most of it can be called popular art in that it originates in the populace, not in a center of mass distribution power, and some of it begins to approach popular art in the other sense that it has a following not created by any center of mass distribution power.

I have said in passing that interactivity is frequently sentimentalized. I have in mind a paradox that is perhaps by now getting to be well recognized, but it merits sketching-in anyway. In the sentimental or cant notion, the reader of an interactive work has heightened autonomy: one creates a new version of the work each time one uses it. Sometimes, in keeping with this notion, the reader is described as "collaborating," which in the sense of co-laboring is true. But this is a peculiar freedom.

That is, let us assume that for any work's audience the reward is the work: to see the film, hear the music, read the text, to experience the microcosm of a certain work of art. In interactive works, the audience must decide what choices or strategies will produce more of the work. My reward for understanding the text is more text. Perhaps I am prejudiced here by my experience of the game world, but it seems to me that this pleasure is in some ways the opposite of freedom: it is the pleasure of being enthralled.

The pleasure of being enthralled is a considerable one, and perhaps is part of any art-experience. It is most familiar in relation to our great appetite for difficulty: we are enthralled by the difficulty of a chess problem, by a new video game, or depending on our taste by a jigsaw puzzle or crossword puzzle. The reward is to see more, and we struggle pleasurably against the bonds and restrictions and obstacles in our way.

My experience of the computer is the experience of a puzzle that is an aperture. That is, each time I learn to do something new with the computer I try to find my way through a maze — the software or hardware manual, the protocol, the peculiarities of some new application. There is a blend of happiness and frustration, a forceful need to see one's way through the bottleneck. Through the narrow neck or aperture of this maze, I know, a vast world will open. The primitive text adventures like Zork, a category of which my 1984 Mindwheel is a literary example, reproduce this process in their plot: behind a trapdoor as narrow as Alice's rabbit-hole opens an unfolding world, branching out thrillingly into countless corridors and chambers. This narrative action — forcing one's way through a difficult aperture into a vast universe of possibility — seems to be a profound human gesture, and a profound aspect, for me anyway, of the personal computer. Through the innocent rectangle of the monitor one beholds a universe.

But the discovery of that universe is a process that works toward a measure of freedom or autonomy, out through a constriction. The reader of the print text, free to thumb through any part of it, is as a matter of convention free at the beginning: then, as I begin (if the work is successful) to care about characters or outcomes in the work, I am drawn away from my initial freedom toward the constricting focus of the work's formal completion. The print reader is free, and the reward is focus. The computer reader begins caught in an opening aperture and reads or clicks on a journey toward the freedom of one who has found the solution.

But in what order do the parts of the solution come?

This seems to me one of the deepest aspects of art, and a fascinating aspect of works for computer. To put it negatively, I have observed two disappointing categories of work in language for computer. One is the naturalistic novel in which characters like those in a soap opera or in a sedulous, plodding novel of contemporary life go about their adulteries or murder solvings or whatever while one is encouraged to skate around their doings in hypertext. This is like a tour by spaceship of an immense Levittown: we have familiar life-materials arranged in a fluid, but arbitrary feeling order. The second kind of disappointing work is the surrealist poem in which the computer basically turns the crank of the recombinant machine: it is like Guillaume Appolinaire's beautiful book of, say, fifty sonnets, bound in fourteen horizontal strips, with one line per strip. The machine shakes the language-dice and has been taught a syntax into which they can be thrown. Often, the syntax is very ingeniously conceived. Here, we have an impressive order, deficient in recognizable life-materials.

In the Mindwheel days, my programmer collaborators and I talked about narrative rooms (spatial) and narrative scenes (temporal). As a practical matter, the programmers posed philosophical questions: did a scene happen in a room or did a room belong in a scene? The shifts and dazzlements of hypertext, I believe, must reflect some such musing or questioning about reality, or they are merely shuffling the image-deck.

In each of these directions there is potential for art, probably for great art, but thus far I am not sure either that any work has mediated successfully between the nature of the medium and the nature of art.

To speculate a little about that large question, I will again talk first about the conventions of an older literary form, drama.

Aristotle says in his poetics that a work of art imitates an action. I learned this from the great critic Francis Fergusson. Fergusson began as a biologist. He had a Rhodes Scholarship and went from Albuquerque, New Mexico to England, in order to study biology with Huxley. But on his vacation he went to France and saw the Moscow Art Theater, and that was the end of biology for him. The theory of theatre art I am about to describe is Ferguson's, and it applies to any work that occur over time. It is particularly apt to computer works.

Fergusson taught us not to speak of the meaning of a work but about its action. He taught us that a work of art is not a pile of impressive ingredients but something that happens, a kind of vehicle. Aristotle says in the Ethics that a human soul is merely passive and reflexive when in a state only of passion: it is passive, like soft wax taking an impression: feeling threatened or lustful or curious. The soul becomes fully itself when passion joins purpose, in a state of praxis or action — expressed, as the Moscow Art Theatre actors had to be able to describe every scene and speech and hat and gesture, with an infinitive: to save the city, to get the girl, to see the truth. An action is not a punch on the nose but a movement of the soul.

In the Poetics, he says that a tragedy imitates an action. What, essentially, imitates that action? He says that a tragic action can be imitated successfully without dances, without choruses, without poetry, without spectacle. Remarkably, he says that you can have a tragedy without character! None of these is the essential soul of the work, that imitates the action; what imitates the action, he says, is the plot. And he has a great definition of the plot: the plot, he says, is the arrangement of the incidents.

That is, the story behind Sophocles' Oedipus Tyrannos existed: there was a king, and a prophet told him his wife would have a male child who would grow up to kill his father and marry the widow, his own mother. When a male child is born, the king orders a courtier to kill it. The courtier gives the infant to a soldier to kill, but the soldier takes it into the hills and drives a spike through its ankles into a rock and runs away. Instead of dying from exposure or wild beasts, the infant is found by a shepherd, who raises the lad, who turns out to be good at solving riddles, and blah, blah blah . . .

That is the story. In Sophocles's actual work, we see three great doors and the altar of Dionysus; in front of the doors and the altar, people are expressing communal grief and oppression. We do not need to know Greek to see that they are sharing some malaise. From the central door, made tall by special footgear, a regal figure comes limping. With gestures and words he tells the people and their spokesman that he will help them, he will take their suffering upon himself. Then, we see a series of arguments or agons in which the kingly figure presses for information that others grudgingly, bit by bit reveal, until a final blaze of insight and recognition. The action to save the city or to find the truth or to break oneself for the community is visibly repeated: each agon, each speech, is a component vector in the overall arc of that action.

This plot, imitating an action, could be called "linear," I suppose — but in relation to the straight line of the Oedipus story, Sophocles' plot, the course of a few hours' real-time interrogation, is as scrambled, curvilinear and oddly linked as any hypertext experience of, say, Michael Joyce's Afternoon.

It is the dynamism, rather than the literal unity, of this ancient model that attracts me: "imitation," scholars say, is not quite the right English word for Greek "mimesis": something like "identification with" or "assimilation by" would be awkward but more accurate. The work's purpose is to attract our identification or assimilation by presenting a movement of the soul.

It goes without saying that the hypertext author need not be bound by this ancient formulation. But let's entertain it for a while.

If the parts are arranged in an infinite variety of ways, and if we want something nonetheless to happen in Aristotle's sense, the parts must be cunningly devised as temporally modular. The point is not linearity or its opposite but movement — movement of the soul, as in the expression, "I was moved." In life, after all, what looks like an ending — a death, a parting — often is a beginning, and what looks like a beginning — a marriage, a birth — often is the end of something.

Interesting works have already used the card-shuffler or kaleidoscope-tapper of hypertext to suggest a pattern of circularity in reality. This seems to me more promising, and more sophisticated, than merely grooving on the changes. A further step would be to assimilate or imitate not just circularity but an ambiguity of ends: this notion is adumbrated by the many narratives and even games that have incorporated mystery-solving, amnesia, deception: in a process that as Arnheim suggests seems to characterize the making of art, the need to accommodate a limitation — the interchangability of parts in hypertext — has led practitioners toward certain thematic materials: a dreamy circularity or ambiguity in our experience of reality.

Modular parts can of course create a simulacrum of linearity within the circular or atomistic model, by devices that I suppose can be reduced to temporal alarm clocks and spatial trip-wires: rooms and scenes, again, a recurrence that measures how halting and heuristic my thinking still is on this matter.

The only shelter for my personal ignorance — of narratology, of the computer itself — is what I believe is a general unknowing about a vague, exciting future. With that shelter, I'll close these remarks by speculating a little about the genius of this medium and some possible directions. As the poem uses the individual voice, and as the stage uses actors moving and speaking, what does the binary computer use?

It uses, among other things, dialogue — as we call the little boxes when we don't call them menus. And any session of dialogue with a computer is likely to involve as many diverting false starts, gripping divagations, serendipitous variations and fortuitous ambushes as our actual conversations with the rest of the world. Virtual dialogue, you might say, imitates actual dialogue.

Between the chopping-up of mere naturalism, so difficult to attain in the medium, and the mechanical image-confetti, there are materials and modes that intrigue me, and the dialogue is one of them. It seems to me that these materials and modes closely parallel the researches into what may still be called Artificial Intelligence, a term I find congenial: it seems a very good description of what a work of art in language, written for computer, ought to attain: an intelligent artificiality, and an artistic effect of intelligence.

For example, the narrow aperture that leads to a vast realm suggests to me a process of interrogation: a rhythm of tantalizing misunderstanding, garbled implications, half-glimpsed possibilities recombining and refolding as one asks questions and deciphers answers.

This may, again, unduly reflect my experience as a game-maker. But first instincts in an art may be true ones: the chase scene is nearly as old as the film medium, for instance, and sometimes seems about to take it over entirely. Very early computer programs, and not only games, often couched their messages in the first person: "I don't understand that command," said these primitive ancestors of Hal, or "I am still thinking about your previous command.".

A virtue of interrogation is that its conventions in life — the press conference, the interview, the negotiation, the therapy session, the Q and A after the lecture — often involve the distortion, incomprehension, and deliberate misdirection that so easily trickle into hypertext composition. In my work on Mindwheel, I found that the dialogue tables were perhaps what intrigued me most. How might one incorporate a bit of reader input to create the illusion of animate responsiveness in a character.

Demonstrating the game for a friend, I guided him to a room where he confronted a female character. She addressed him in some faintly provocative manner, and having typed in the "look-at" command he found that she was, say — I forget the details — attractive, dark-haired, wearing surgical scrubs. My friend addressed her:

"You look like my mother," he typed. The machine, an antediluvian Kaypro as I recall, clucked over its hard drive, and the character responded:

"I will look," she said to him, "any way you want me to look."

To my pleasure, he looked startled, claimed he got goosebumps, and I got a little thrill myself — the pleasure of feeling that the clunky little game, on that yellow-and-black monochrome monitor, had made a plot that imitated reality: he saw her, he made a remark half-mocking and half-flirtatious, and she had responded in a way germane and unexpected. I will leave off these speculations with that teeny glimpse of a possible art.

original text 1997-02 / posted with the permission of Robert Pinsky 2002-05