/\ ___ ___ ___ ___ __ __ /__\ /__/ / / / / / / / / \/ \ / /__ /__/ / / / ----------------------------------------------------------------- DECEMBER 1990 NUMBER 44 VOLUME 10 NUMBER 10 ----------------------------------------------------------------- Welcome to ART COM, an online magazine forum dedicated to the interface of contemporary art and new communication technologies. You are invited to send information for possible inclusion. We are especially interested in options that can be acted upon: including conferences, exhibitions, and publications. Proposals for guest edited issues are also encouraged. Send submissions to: well!artcomtv@uunet.uu.net artcomtv@well.sf.ca.us Back issues of ART COM can be accessed on the Art Com Electronic Network (ACEN) on the Whole Earth 'Lectronic Link (WELL), available through the CompuServe Packet Network and PC Pursuit. To access the Art Com Electronic Network on the WELL, enter g acen at the Ok: prompt. The Art Com Electronic Network is also accessible on USENET as alt.artcom. For access information, send email to: artcomtv@well.sf.ca.us. *Guest Editor: Abbe Don *Executive Editor: Carl Eugene Loeffler *Editor: Anna Couey *Systems: Fred Truck and Gil MinaMora ART COM projects include: ART COM MAGAZINE, an electronic forum dedicated to contemporary art and new communication technologies. ART COM ELECTRONIC NETWORK (ACEN), an electronic network dedicated to contemporary art, featuring publications, online art galleries, art information database, and bulletin boards. ART COM SOFTWARE, international distributors of interactive video and computer art. ART COM TELEVISION, international distributors of innovative video to broadcast television and cultural presenters. CONTEMPORARY ARTS PRESS, publishers and distributors of books on contemporary art, specializing in postmodernism, video, computer and performance art. ART COM, P.O. Box 193123 Rincon Center, San Francisco, CA, 94119-3123, USA. WELL E-MAIL: artcomtv TEL: 415.431.7524 FAX: 415.431.7841 ------------------------------------------------------------------------------ GUEST EDITORIAL: INTERACTIVE FICTION PART 2 ABBE DON ------------------------------------------------------------------------------ Abbe Don, owner of IN CONTEXT, is an interactive multimedia artist and producer. Her interactive video "We Make Memories," which simulates the way her great-grandmother told stories, has been exhibited nationwide. She has done research with Apple Computer's Advanced Technology Group on the Guides project which investigates the use of narrative and storytelling as a means of structuring and conveying information in large multimedia databases. She was also a guest artist at the Future Fiction Workshop in 1988 and 1990. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - The December issue of ART COM Magazine further explores the theme of interactive fiction, presenting authors who have used artificial intelligence, simulation, and immersion in dramatic narrative theory as the basis for their research. Below, I give an editorial overview of each author's perspective, followed by a representative selection from their longer, in-depth essays. The essays can be read in their entirety via the Art Com Electronic Network on The WELL (Whole Earth 'Lectronic Link) by typing "g acen" at any ok prompt, and entering the "newstand"; or via Usenet in alt.artcom. John McDaid combines narrative theory and ideas about quantum mechanics to formulate a strategy for creating "constructive hypertexts" which "do not proceed from the discovery of hidden content, but rather by symbolic creation." McDaid's essay style, composed of at least one part hypertextual rant, one part highbrow literary reference, and one part tongue firmly planted between brow and cheek, parallels his interactive fiction piece "Uncle Buddy's Phantom Funhouse." ...Jay David Bolter, in a talk at last year's Modern Language Association conference, spoke about the duality of hypertext: the dialectic of node and link, or as he put it, "looking at" vs. "looking through." The reader experiencing the text he says, is "aware of oscillation; [and this is an] explicit measure of interaction." Ultimately, he was arguing, hypertext wants to be both at the same time. Text which can present itself as surface, and yet effortlessly yield through to other levels. Hearing this, I immediately began to suspect this hypermedia duality was the mirror and analogue of the duality of particle and wave, of energy and matter, in physics. What we are seeing in hypermedia is the appearance on the macrolevel (of 2-meter humans and similar Objects) of quantum reality. In this essay, I'll attempt to explain this perspective (and any lack of clarity is purely my own, and should not be attributed to Bolter) and try to link this with a notion of texts as "transformative utterances," occasions for semiotic recombination which serve the function of conscious dreams. As Jung said somewhere, "The psyche seeking transformation yields symbols." It seems to me as if there is a hidden agenda in media evolution which has guided us into stumbling over just this dichotomy... --from "Welcome to the Matrix: From Object Oriented Text to Quantum Indeterminacy," John McDaid (u1475@applelink.apple.com). Jim Gasperini, author of the political simulation game "Hidden Agenda," traces the historical development of interactive "texts" within literature and theatre and then explores early examples of interactive fiction within the computer domain. From there, he postulates a theory of interactive fiction based on "open-ended structural ambiguity" in which a "work becomes more ambiguous, not less the more it is played" suggesting that "through repeated playings, comparing different plots chosen through the same web of potential plots the experience becomes meaningful." Gasperini's essay style and content, as well as his interactive work, differ greatly from McDaid's, yet he concludes with similar observations about the impact of quantum mechanics on the way we structure and represent our experience of the world around us. Though I firmly believe that computer games can be serious works of art, I will not argue that works produced so far manage to do much more than point up the possibility. A serious work of art will not only evoke emotion, it will embody the expression of emotion on the part of its creators. Just how deeply the interactive medium will be able to probe the human psyche and the human condition remains to be seen. Are we witnessing the birth of a new art form, showing promise of accurately expressing the tenor of our times? Or will it be a new hype, a way to sell us the illusion of control, packaging a safely neutered rebellion against the enforced passivity of so much of our current cultural experience? It may well be both. Much of our audience remains unprepared to look beyond the immediate gratification of closed-ended game structures on trivial themes. As the generations growing up with interactive entertainment mature, however, the medium will mature along with them. The medium is very new, and still difficult to work with. For designers it can be a struggle simply to keep the logical nature of our primary tools from dominating our thought processes. All computers really know is how to count to one, after all; everything else is illusion. The challenge is to build works that pretend to some artistic character in a context set by the binary thinking of computer programs--to somehow use logical computers to recreate human fuzziness. Perhaps it is only natural that this form of narrative developed in the age of quantum mechanics. Our physicists tell us that what we take to be the solidity of matter is only a reflection of our limited means of perception... that time may be bidirectional...that things can be true and untrue simultaneously...that it is entirely possible for alternate universes to be created and destroyed. The way the interactive medium plays with ambiguities in the structure of experience may parallel the way we have come to view our universe. Perhaps our cultural imagination is only following, in crude and timid fashion, our vision of the ambiguity of existence as it molds the structure of our dreams. --From "If Brecht Were Alive Today, Would He Be Designing Computer Games?" (Part 3 of a 3 part essay) Jim Gasperini (jimg@well.sf.ca.us). Joseph Bates, director of the Oz project at Carnegie Mellon University, consistently combines the disciplines of artificial intelligence and dramatic theory. Under his leadership, the Oz team is "applying existing artificial intelligence technology to the problem of building dramatic worlds" which are "composed of a (simulated) physical environment, intelligent/emotional agents which live in the world, a user interface and theory of presentation to let one or more humans interact with the world and its agents, and a computational theory of drama which plans and controls the overall flow of events in the world." Oz worlds are intended not only to be realistic, but to be interesting. Often this means giving people the feelings that come with good stories, feelings that arise in part from the structure of plot, such as complication, climax, and resolution. We understand how this can be achieved for static text: an author carefully constructs the text to convey the structure. However, in interactive fiction we do not write out the whole text in advance, and we don't know in advance the detailed sequence of events a reader will experience. Thus, the Oz system must dynamically, and subtly, adjust the behavior of the world and its characters to provide experiences with the desired dramatic structure. This means developing and implementing a computational theory of drama, and using it to guide the behavior of worlds. We see several approaches to developing such a theory. The simplest comes from having the author express a partial order on the significant events of the story (an "abstract plot graph"), explicitly representing that partial order in the system, and using it to drive the character goals and narrative decisions in the rest of the system. This approach would leave almost all the dramatic theory in the mind of the author, with the plot graph serving as a kind of partially ordered program to be executed by the system. A richer approach is to develop a library of abstract plot units and then, as the interaction proceeds, rapidly search abstract plot space for controllable paths that have the desired dramatic structure. We can view this as a kind of abstract adversary search, where we define a set of abstract operators, means for mapping operators into concrete moves, means for recognizing the abstract effects of the users moves, and an evaluation function on event histories (ie, stories) that lets us recognize sequences with "good" dramatic structure. We are working toward implementations of both of these approaches. The former appears to be a relatively easy way to provide dramatic control signals for experiments with the rest of Oz; the latter is a self-contained long term research goal. In both of these efforts, we are drawing on Brenda Laurel's studies of computational versions of Aristotle's Poetics. --from "Computational Drama in Oz," Joseph Bates (joseph.bates@wizard.oz.cs.cmu.edu) Finally, David Graves presents a theory of interactive fiction that builds on the work of both Brenda Laurel and Michael Lebowitz. He emphasizes the importance of creating characters that emulate human emotion through a system of Artificial Personality. ...Almost all attempts to generate behavior in computer-controlled characters have followed a simple stimulus/response model. For each statement you may make to character, there is one response it may display. Giving the same stimulus several times in a row, the character will mindlessly repeat the same response. Clearly, there is room for improvement in creating lifelike characters, but how can we attack such a difficult task? To start, we can borrow a number of ideas and methods from the field of Artificial Intelligence, avoiding the difficult software problems, to produce the >illusion< of intelligent, emotional, motivated characters... --from "Life-Like Characters in Interactive Fiction," David Graves (dag@hpsemc.cup.hp.com). As nearly every author has noted, interactive fiction is still in its infancy as artists and researchers iteratively build on the conceptual and technical frameworks of their predecessors. I hope the dialogue continues to stay open and interdisciplinary! Abbe Don Guest Editor abbe@well.sf.ca.us ------------------------------ MENU OF CONTENTS ------------------------------ 1. AN ART FORM FOR THE INTERACTIVE AGE (part 1 of 3), Jim Gasperini 2. STRUCTURAL AMBIGUITY: THE EMERGING INTERACTIVE AESTHETIC (part 2 of 3), Jim Gasperini 3. IF BRECHT WERE ALIVE TODAY, WOULD HE BE DESIGNING COMPUTER GAMES (part 3 of 3), Jim Gasperini 4. COMPUTATIONAL DRAMA IN OZ, Joe Bates 5. WELCOME TO THE MATRIX: FROM OBJECT ORIENTED TEXT TO QUANTUM INDETERMINACY, John McDaid 6. LIFE-LIKE CHARACTERS IN INTERACTIVE FICTION, David Graves ------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- AN ART FORM FOR THE INTERACTIVE AGE (part 1 of 3) Jim Gasperini Jim Gasperini (jimg@well.sf.ca.us) is author, with TRANS Fiction Systems, of "Hidden Agenda," a narrative simulation of politics in Central America. Chosen role-playing game of the year in 1989 by MacWorld, "Hidden Agenda" has also been used to train diplomats by the Foreign Service Institute of the US State Department. Also with TRANS Jim co-authored the text adventure "Star Trek: The Promethean Prophecy." In recent decades various literary and dramatic artists have experimented with ways to make the role of the audience more active. Now computer technology makes possible the development of a true interactive aesthetic. But grafting interactivity onto earlier forms of narrative is like making a movie with a fixed camera: it fails to take advantage of the essential power of the medium. INTERACTIVITY IN LITERATURE AND THE THEATER The first interactive work I remember seeing myself was a French novel from the 1950's. It originally came in a box, one chapter per page, with instructions about how to shuffle the chapters up and read them in any order. Intrigued by the concept, I hunted down a copy at my college library, only to find the pages firmly bound together like every other book, victim of the indiscriminate efficiency of a library binding service. Since then we have seen such partially interactive works as Julio Cortazar's "Hopscotch," a novel with a number of 'optional' chapters, read or skipped at the reader's discretion. More recently the many short chapters of Milorad Pavic's 'lexicon novel'. "The Dictionary of the Khazars" may be read in a variety of sequences. Numerous series of 'reader-active' books, the first and best known of which is Edward Packard's "Choose Your Own Adventure," ask children to choose different plotlines at the end of each chapter. The theater in this century has seen numerous experiments with the aesthetic distance between audience and performer. As part of a play within a play, Luigi Pirandello placed a simulated audience within the real audience in an attempt to reinforce the emotional link between audience and actors. In his 'epic theater' Bertolt Brecht continually fought the notion that the audience must simply observe an aesthetic object. He searched for ways to break down the veil of illusion, make the experience of theater instructive as well as entertaining, and stir the audience to action in the here and now. Toward the end of his life, when asked whether the theater was still an adequate means for representing the modern world, Brecht replied that yes, the world could be represented on the stage, but only if it was portrayed as 'veranderbar' (changeable). Playwrights and directors continue to experiment with ways to alter the distance between actors and audience. At the theatrical version of "The Mystery of Edwin Drood" the audience is offered the opportunity to select one of a number of endings (a device Brecht had used several decades earlier). The Living Theater, among others, continues the Brechtian anti-tradition into our times. At several current plays ("Tamara," "Tony 'n Tina's Wedding") the audience mingles with the actors, choosing what aspect of the performance to observe at what time. LIMITATIONS Interactivity in book form suffers from a severe technological limitation: the page does not change. The reader's decisions can have only limited impact on the way the story unfolds. Though varying the order in which a story's chapters appear may alter the perceived significance of each chapter, the choice of a particular chapter does not change the way later chapters read, and the author therefore has only limited ability to build contrasting chains of choice and consequence. Interactivity in the theater suffers from different practical limitations. The choice of endings in "Drood," coming at the very end of the play, is more of a gimmick than a true involvement of the audience in the process. While the audience walking around the 'set' of "Tamara" undoubtedly experiences more varied details than it would in a conventional theater, the essential plot structure continues unaffected by anything audience members opt to do. Though these works allow an unusual degree of audience participation, they are not truly interactive. Actors' improvisational workshops, of course, call for a high degree of interactivity. So, in a limited way, do the skits of certain comedians and 'improv' troupes. To give the audience control over the plot development of a full-length commercial production, however, would mean advance preparation for a dauntingly broad combinatorial explosion of possibilities. The enormous expense involved in preparing for all the directions the audience might think to take the play, not to mention the extraordinary demands such a production would make on the actors, works to ensure that the theater remains dominated by linear plotlines and audiences that quietly watch the spectacle. PERFORMANCE IN A COMPUTER GENERATED WORLD Enter the computer. Computer technology allows authors to create elaborate simulated worlds, within which players have considerable freedom of action. The player is given the general outlines of a character and told to improvise in reaction to a simulated environment, simulated props and other characters (simulated or not). It is a performance medium, differing from other performance media in that the audience is also the protagonist. An interactive work must by definition present its player first with some sort of choice, and then with the consequences of that choice. It therefore takes place over time, allowing the construction of a plot. This gives it at least some kinship with narrative. The work may also present simulated characters to encounter, which gives it a similarity to drama, although as we will see later on 'characters' can take unusual forms in an interactive work. Analogies to the theater and to literature are useful, but inexact. By erasing the distinction between audience and performer, stories told using computer technology have made a radical break with both traditions. But what is it about these technologically sophisticated role- playing exercises that put them closer to the short story than to, say, charades? One answer may be found in a closer look at one of the key techniques of narrative: ambiguity. -------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- STRUCTURAL AMBIGUITY: THE EMERGING INTERACTIVE AESTHETIC (Part 2 of 3) Jim Gasperini Jim Gasperini (jimg@well.sf.ca.us) is author, with TRANS Fiction Systems, of "Hidden Agenda," a narrative simulation of politics in Central America. Chosen role-playing game of the year in 1989 by MacWorld, "Hidden Agenda" has also been used to train diplomats by the Foreign Service Institute of the US State Department. Also with TRANS Jim co-authored the text adventure "Star Trek: The Promethean Prophecy." The depth of a narrative work's ambiguity is a good measure of its quality. The richer the work, the more it resonates with ambiguous meanings. Three different 'levels' of ambiguity may be distinguished, two familiar levels and one that is quite new. The first level consists of those ambiguities that appear in a text as it might be read on a printed page or on a screen. A metaphor is an ambiguity: a word or phrase with a surface meaning that also points to something else, meaning at least two things at once. William Empson's celebrated "seven types of ambiguity" in poetry (as discussed in the 1937 book by that title) are all variants on this 'textual ambiguity.' By his definition, ambiguity is 'any verbal nuance, however slight, which gives room to alternative reactions to the same piece of language.' In some forms of writing (technical writing, journalism) the opposite of 'ambiguous' is 'precise.' In creative writing the opposite of 'ambiguous' is 'dull.' When words appear as part of a theatrical performance, another level of ambiguity comes into play: 'interpretive ambiguity.' The same role may be interpreted in radically different ways--one actor may portray Hamlet as indecisive introvert, another as impotently raging victim of Oedipal conflict, a third as suicidal misanthrope. The contributions of the director, set designer, and others also influence the way the audience interprets the work. One way to distinguish a great play from a merely good one is by the degree to which it lends itself to different interpretations, so that actors and audience alike will wish to experience the play again in a different guise. A truly interactive work offers a third level of ambiguity, arising from the role the audience plays in the construction of the plot. This I call 'structural ambiguity.' Rather than building nuanced experience from the meanings of words and phrases, or from dramatic interpretation, structural ambiguity builds meaning from alternate possibilities of choice and consequence played out over time. Two types of this structural ambiguity may be distinguished: 'closed-ended' and 'open-ended.' In the former, certain *apparent* ambiguities are first raised in the player's mind and then resolved. The work becomes less and less ambiguous as the player progresses through it. In the latter, ambiguity grows deeper as the work unfolds. The open-ended type has the greatest potential for richness of meaning, but since most current computer narratives are essentially closed-ended games let me discuss that type first. CLOSED-ENDED STRUCTURAL AMBIGUITY Typical 'twitch' games, including most Nintendo and arcade-style computer games, essentially consist of hand-to-eye coordination learning. The game presents a series of obstacles, sometimes stringing them together into a simple story and sometimes not bothering much with plot at all. The player learns when to zig, when to zag, where to put the falling trapezoid, how to kill the bats, how to deal with endlessly sprouting mushrooms. It's stretching a point to call this process 'resolution of ambiguity,' so let me quickly pass on to firmer ground. A better example of closed-ended ambiguity may be seen in the type of game known as 'adventures' or 'interactive fiction,' stories that proceed according to what I call a 'resistant plot.' The player takes on a role within an imaginary world filled with apparent ambiguities, which it is the player's task to resolve. By overcoming obstacles or solving puzzles--as players discover the hidden usefulness of simulated objects, or how the behavior of a character can be changed--the plot is made to advance. Since the original "Adventure" was invented by a pair of computer scientists named Crowther and Woods in the mid 1970's, the genre has seen considerable refinement in graphical imagery and application to varied subject matter. One problem with resistant-plot stories is that they demand a great leap of faith. The player must spend a considerable amount of time struggling to make something happen, all the while unsure if the end result will be particularly interesting. Many people are understandably reluctant to spend a good chunk of their lives beating their heads against imaginary walls. As a former writer and player of 'interactive fiction,' my current feeling about this part of the computer game genre is best summed up by a line from Voltaire: "life is too short to learn German." Even in the best 'interactive fiction,' once all the puzzles have been solved the plot is revealed in all its naked linearity. A finished 'closed-ended' work is like a punctured balloon, emptied of all ambiguity. There is little reason for anyone to go through it again. OPEN-ENDED STRUCTURAL AMBIGUITY By contrast, an 'open-ended' work becomes more ambiguous, not less, the more it is played. It is through repeated playings, comparing different plots chosen through the same web of potential plots, that the experience becomes most meaningful. This can be most clearly seen in the genre known as 'simulations.' Here we must once again sort out the different uses for a term. Classic computer simulations are designed as serious analytical tools, or as a means for training people in specific tasks. By setting up a model of some real-world system (a power plant, an airport control tower, an all-terrain vehicle) players and designers can examine the workings of that system. Modifications can be tested out on the model before actually being put into practice, potential problems can be identified, trainees can learn their jobs in an environment where mistakes cause no real damage. Narrative simulations, however, are designed with more broadly didactic and expressive goals in mind. By taking on a role within a simulated system, players may explore that system as the author has chosen to model it, and experience the underlying conflicts, powers and constraints peculiar to that role as it is played in the real world. Two recent examples are "Sim City," by Will Wright and Maxis Software, and my own "Hidden Agenda," written with TRANS Fiction Systems. "Sim City" puts the player in control of a city, which begins as a bucolic stretch of riverbank seen from the air. As the player creates residential areas, roads, power plants, and industrial areas, the city quickly comes alive. Little houses start to appear once people (the 'sims') move in. Soon the 'sims' may be seen going back and forth to work in cars and commuter trains. Opinion polls tell you how they feel about your skill at managing their city. They complain about taxes, demand that you spend more on police and fire protection, and plead for emergency aid in the aftermath of floods and earthquakes. "Hidden Agenda" also puts the player in a position of power, this time the President of a fictitious Central American country. Action takes place in encounters with characters, represented by photographic images and bits of dialogue about policy choices. Representing political parties, Army factions, social groups, professions, economic classes, other nations and international agencies, each tries to convince you to follow his or her own policy agenda. If they grow impatient or disillusioned with your leadership they may take action on their own. These two games present the world very differently--one through a graphical representation of a growing metropolis, the other through changing positions taken by characters as expressed in words and background graphics. They are alike, however, in inducing players to feel increasingly responsible and ambiguous about the effect they can have on a virtual culture. Players inevitably measure each city they build in "Sim City" against the variant cities they might have built had they made other choices. The city itself--*your* city, which you may name for yourself if you so desire--becomes a kind of character for whose growth, problems and personality you feel directly responsible. Since you can save the position of any city at any point, you can build two or more variant cities and observe the effects of even minor modifications of policy. In "Hidden Agenda" the player's choices amount to the incremental selection of one plot out of an extremely varied number of potential plots. For example, midway through the game the Presidente can decide to hold elections. Depending on various other choices made along the way, various things happen during the election season. Sometimes events play out like the Salvadoran elections of 1982 and 1984: terrific violence, assassination of leftist candidates, strong right-wing army control of the process, U.S. proclamation of the results as a triumph for democracy. Sometimes instead they play out more like the Nicaraguan elections of 1984: minimal violence, withdrawal of rightist candidates, strong left-wing army control of the process, U.S. denunciation of the results as a fraud. As author I am making ironic points in this obvious example, points that the interactive structure amplifies and sets in context. Each subsequent time the player enters the election campaign, comparisons naturally arise between what happens this time and what happened other times. This serves to deepen the player's awareness of the range of structural possibilities. In both simulations, it is up to the player to decide whether they 'won' or 'lost.' "Sim City" is extremely open-ended: you can follow the progress of your city on through the centuries if you wish. In "Hidden Agenda" your term of office is limited to three years (if you are not overthrown in a coup or voted out of power). Whichever way your term of office ends, the simulation concludes with an encyclopedia entry supposedly written far in the future (the 'Verdict of History'), detailing what you tried to do and what resulted. It is up to you to interpret this Verdict, evaluating your successes and failures according to your own sense of priorities. -------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- IF BRECHT WERE ALIVE TODAY, WOULD HE BE DESIGNING COMPUTER GAMES (Part 3 of 3) Jim Gasperini Jim Gasperini (jimg@well.sf.ca.us) is author, with TRANS Fiction Systems, of "Hidden Agenda," a narrative simulation of politics in Central America. Chosen role-playing game of the year in 1989 by MacWorld, "Hidden Agenda" has also been used to train diplomats by the Foreign Service Institute of the US State Department. Also with TRANS Jim co-authored the text adventure "Star Trek: The Promethean Prophecy." The author of a novel or play, of course, does more than string together sequences of ambiguities. The ambiguous elements must have at least a surface coherence and be informed by some encompassing vision. Can the designers of simulation games do something similar? Can these works be said to be 'authored?' I believe they can. Though most simulation games offer schematic renderings of the essential processes of real-world systems, exactly which processes are incorporated into the model and exactly how they play out is a matter of artistic choice, reflecting the author's understanding and passions. The schematization allows a player to participate in processes speeded up and rendered comprehensible, and allows the designer to impose his or her personality on the work as a whole. Though designers may honestly attempt to model reality with as much balance and accuracy as possible, they inevitably construct works that at bottom reflect their own vision of the world. A term already exists, of course, for imaginative exercises that ask a player to actively perform a role: game. This word already covers so many different activities, however, from charades to baseball to Monopoly, that its use in this context tends to confuse and trivialize. To many people the term suggests something slight, entertaining perhaps but certainly not 'artistic' (unless you wish to consider baseball, say, to be a form of dance with a set, familiar choreography on the theme of conflict, within which individual performers have a certain range of interpretive freedom. I suspect, however, that few ballplayers think of themselves as 'movement artists.') The assumption seems to be that since a game requires the player to take some sort of action it inherently offers less density of meaning than forms which ask the audience to remain relatively passive. This assumption arises, quite naturally, from our long familiarity with many types of games. From earliest childhood we use games to amuse and instruct ourselves, starting with the simplest and moving on to ones of increasing complexity. Whatever the main point of interest--which players win and which lose, how they play the game, how much fun they have along the way, how much money they can make by playing it--most games are designed as elaborately simple sets of procedures for tamed, managed conflict which can be repeated many times. Game designers have usually ceded so much control over how the experience plays out that players can have very little sense of participating in someone else's artistic vision. Yet here I am, arguing that new technologies allow such exceptional authorial control over an interactive experience that a 'game' may be designed with enough precision and depth to be considered a form of narrative art. WANTED: TERM FOR NEW CATEGORY Part of the difficulty is semantic. Once we manage to find a distinctive name for this category of experience, it will be easier to evaluate it on its own merits without having to endlessly distinguish it from other things. Unfortunately, all the words used to describe authored interactive works--game, simulation, interactive fiction, multimedia--are too broad, pointing in too many directions at once (too ambiguous!) Take 'simulation,' for example. If defined very broadly: A simulation of human behavior experienced by watching actors on a stage is called: drama, play, theatricals, the stage, histrionics. A simulation of human behavior experienced by reading words in a book or periodical is called: narrative, literature, novel, story, fiction. A simulation of human behavior experienced by watching patterns of flashing lights on a screen is called: cinema, motion picture, movie, television, video. A simulation of human behavior experienced by interacting with computer technology is called: what? We sorely need a new term or two. Interactive works are not sub-genres of drama, fiction, or cinema--they are a fourth thing entirely. STRENGTHS OF THE MEDIUM Every medium has its peculiar strengths. This one is so new that its strengths have only begun to appear. Since we have only begun to simulate human behavior in interactive works, we have vast amounts of subject matter yet to address. One great strength, clearly, is the medium's ability to present the world from another person's perspective. What does the world look like to a Latin American? What powers come with the role of leader of a small country, what limitations and constraints? Interactive works already enable a player to experience 'a day in the life of an Israeli,' 'a typical Australian neighborhood barbecue,' and 'what it was like to be a French gentleman during the Enlightenment.' So far most of these works have been designed with instructional purposes in mind, offering the player insights into sociology, language, and history. As time goes on, we may begin to see more personal uses of the interactive medium. One game might explore what it is like to be a woman on the edge of madness; another what it is like to be a man in the midst of a mid-life crisis; a third simply 'what it is like to be me.' We will see propagandistic uses: interactive Leni Reifinstahls creating strongly biased simulated worlds. Some interactive artists may decide to use the new medium to address subject matter traditional to narrative art. On the other hand, there may be aspects of human experience that the interactive medium can better address than can more traditional forms. Questions of economic theory and political philosophy, for example, can be more easily woven into a narrative structure in a medium that places the audience in the role of protagonist. Novels, plays and films about men and women in public life usually focus on the tension between the public and the private person. They do so in part because the form (at least as used conventionally) works by inducing the audience to identify with and care about the principal characters. Since the 'audience' of an interactive work no longer sits to one side judging how the protagonist meets various challenges, authors can build emotional resonance on the player's sense of direct responsibility for how those challenges are met. That a work is built with structural ambiguity does not preclude the use of other levels of ambiguity as well. We will soon see much more poetic use of textual ambiguity within interactive works than we have seen so far. As for interpretive ambiguity, this is a natural consequence of the way author and player collaborate in the final production. The effects of this collaboration may already be seen in the tendency of complex simulated systems to behave in ways not explicitly planned by the designers, a phenomenon known as 'emergent behavior.' For example, as certain Sim Cities mature slums begin to appear. This was not planned by the designers, but emerged as the natural consequence of the way some players act within the simulation. Simulation games have the potential to accomplish some of the same purposes Brecht sought to achieve in his experiments with theatrical form. The interactive medium can inform while entertaining, make the audience an active and aware participant, illuminate the political and cultural context within which audience and performance exist. The way the player can choose to examine information from many different standpoints during the course of some games (for example by viewing charts and newspaper reports) is arguably analogous to the Brechtian technique of underlining the significance of events on stage with placards inscribed with statistics. On the other hand, many of Brecht's theatrical innovations were designed to create aesthetic distance, not break it down. Brecht fought against the idea that theater-goers should lose themselves in the characters on the stage, but the appeal of many computer games is precisely that they induce players to lose themselves in a character (or, worse, in a sequence of pointlessly aggressive actions). What Brecht would have thought of the interactive medium is idle speculation. I suspect, in fact, that he would have hated most current computer games. INTERACTING WITH THE FUTURE Though I firmly believe that computer games can be serious works of art, I will not argue that works produced so far manage to do much more than point up the possibility. A serious work of art will not only evoke emotion, it will embody the expression of emotion on the part of its creators. Just how deeply the interactive medium will be able to probe the human psyche and the human condition remains to be seen. Are we witnessing the birth of a new art form, showing promise of accurately expressing the tenor of our times? Or will it be a new hype, a way to sell us the illusion of control, packaging a safely neutered rebellion against the enforced passivity of so much of our current cultural experience? It may well be both. Much of our audience remains unprepared to look beyond the immediate gratification of closed-ended game structures on trivial themes. As the generations growing up with interactive entertainment mature, however, the medium will mature along with them. The medium is very new, and still difficult to work with. For designers it can be a struggle simply to keep the logical nature of our primary tools from dominating our thought processes. All computers really know is how to count to one, after all; everything else is illusion. The challenge is to build works that pretend to some artistic character in a context set by the binary thinking of computer programs--to somehow use logical computers to recreate human fuzziness. Perhaps it is only natural that this form of narrative developed in the age of quantum mechanics. Our physicists tell us that what we take to be the solidity of matter is only a reflection of our limited means of perception... that time may be bidirectional... that things can be true and untrue simultaneously... that it is entirely possible for alternate universes to be created and destroyed. The way the interactive medium plays with ambiguities in the structure of experience may parallel the way we have come to view our universe. Perhaps our cultural imagination is only following, in crude and timid fashion, our vision of the ambiguity of existence as it molds the structure of our dreams. -------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- COMPUTATIONAL DRAMA IN OZ Joseph Bates Dr. Joseph Bates (joseph.bates@wizard.oz.cs.cmu.edu) is director of the Oz project at School of Computer Science, Carnegie Mellon University. He was co- host of the Workshop on Interactive Fiction and Synthetic Realities at the AAAI-90 conference. (Abbe Don) >From working notes of AAAI-90 Workshop on Interactive Fiction and Synthetic Realities, Boston, July 1990. These are informal notes prepared as background material for a lecture given at the AAAI workshop. The Oz project at the CMU School of Computer Science is developing technology for high quality interactive fiction. Our goal is to provide users with the experience of living in a dramatically interesting simulated world that includes simulated people. A variety of researchers on human interfaces are studying virtual or artificial realities. These are computer simulated interactive visual environments that people experience as real. Most existing research concerns issues close to the interface, that is, how to take an underlying simulated world and present it in a convincing fashion. The Oz group plans to use this interface technology as it develops, but our work is on creating rich, deeply modeled underlying worlds. Thus, we study the simulations behind the interface, which we call the deep structure of virtual reality. Our work applies existing artificial intelligence technology to the problem of building dramatic worlds. These worlds are composed of a (simulated) physical environment, intelligent/emotional agents which live in the world, a user interface and theory of presentation to let one or more humans interact with the world and its agents, and a computational theory of drama which plans and controls the overall flow of events in the world. There are clear applications of such simulations to entertainment (interactive fantasy experiences) and to training (e.g., improving interpersonal skills in business). Also, since we think art develops as new media develop, we hope our work will be the basis for one of the first sophisticated knowledge based art forms, using computers as the underlying medium. We believe these simulations, which today require engineering workstations, will run on future consumer electronics products that integrate video and audio with RISC engines, large DRAMs, and digital signal processors. Interactive fiction was the most popular home software in the U.S. in the early 1980's, despite an extremely low level of technical sophistication. We suspect that by taking advantage of known technologies, interactive fiction can develop into a popular and long lasting art form. We think it can serve as a primary motivation for people to own the personal digital systems of the middle 1990's. AN OVERVIEW OF OZ As preparation for the workshop lecture, this document presents an overview of the Oz Project. Our efforts can be partitioned into six areas: physical world simulation, the minds of simulated characters, the user interface with its theory of presentation, theories of drama, the world building environment, and the artistic use of the system, each of which is described below. Oz is built in Common Lisp, with substantial use of the Common Lisp Object System. The system incorporates large amounts of code from other Lisp based AI projects. We develop Oz on Mach/Unix workstations, but do not presently rely on anything beyond Lisp. The Oz group at CMU includes Bates, five Computer Science graduate students, an undergraduate, and a gradually growing collection of users from the English and Drama departments and elsewhere in the CMU community. We are assisted, especially on dramatic theory, by Brenda Laurel. With Brenda and with Margaret Kelso, of the CMU Drama department, we have started studying real life interactive improvisations. These can be viewed as simulations of Oz, and can tell us both about the inherent nature of this art form and about how to build computer based IF (interactive fiction) systems. PHYSICAL WORLD SIMULATION The Oz physical world simulator provides a commonsense model of the physical world. It is an abstract model, thus many aspects of the real world are omitted. Unlike existing interactive fiction systems, our emphasis in not on manipulating objects in the world, but on character and plot. Thus, while we are applying object oriented techniques to flexibly model the world, and while these models could ultimately become quite rich, our requirement is only to provide enough of a physical reality to let authors construct interesting characters and stories. THE MINDS OF CHARACTERS These are the minds of the (non-human) agents that populate worlds. One of the claims of the Oz project is that the mental architectures and real world knowledge bases that have been developed in AI over the last 15 years, while perhaps still too weak for real robots, are well suited to the demands of interactive fiction. Our goal is to draw on the best of these existing systems, such as work from Yale, CMU, and Microelectronics and Computer Technology Corporation (MCC), and interface them to Oz as frameworks for the minds of agents. Once these frameworks have been established, the builders of worlds will use them to construct individual characters. We are currently working on two frameworks: a goal driven reactive planner called HAP and the Prodigy planner. We are developing HAP within the Oz group based on work at MIT by Agre and Chapman and on work at Yale by Firby. Prodigy is a planner/learner being developed at CMU by Jaime Carbonell's machine learning group. We expect to extend both systems using ideas of Wilensky, Dyer, Carbonell, and others to provide some level of social awareness and emotion, in addition to rudimentary intelligence. In addition, we are interested in exploring other systems. The developers of Soar (particularly Allen Newell at CMU and John Laird at University of Michigan) and CYC (Doug Lenat at MCC) have expressed interest in connecting their systems to Oz. We are hopeful that this will occur during the next few years. USER INTERFACE AND THEORY OF PRESENTATION The user interface connects human agents to the simulated world. For the immediate future, we expect this connection will be via natural language text. We have been using software from the CMU Center for Machine Translation to generate text, producing both the narrative description of the world and the textual "speech" of computer modeled agents. We are now developing Glinda, our own generator, with careful attention to the PENMAN work at ISI. In our application, parsing is easier than generation. At present we use a general purpose bottom up parser with a simple grammar and ad hoc semantic and pragmatic analysis. We are considering using instead a word based parser, such as the DYPAR parser originally developed by Roger Schank's group at Yale. These parsers seem appropriate for processing short, syntactically limited, possibly ill-formed input, which is typical in interactive fiction. We have started studying ways to "tune" the natural language generation to provide subtle emotional influence on the human player. In theatre and cinema, extra-semantic influences such as music, lighting, point of view, zooms, and film editing play a significant role in determining viewer reaction. The artistic technique developed in these areas is crucial to their respective media. Our report "Towards a Theory of Narrative for Interactive Fiction" describes results of our initial efforts to find analogous technique for interactive fiction. We are pursuing this research, with the goal of having Oz adjust its style of output to suit the varying dramatic content of the story. Hovy's work on PAULINE is directly relevant to our efforts. As individual character architectures develop, they may well bring their own mechanisms for natural language processing. Where these mechanisms improve the quality of characters, we will probably use them in place of Glinda. Oz presently uses a text interface for two reasons. First, an argument can be made that text based IF is a valid form in its own right, allowing certain kinds of artistic technique, such as narrative, that cannot easily be applied in a VR setting. Second, we feel that our research effort and computing capacity is best spent now on characters, natural language, and dramatic theory. However, as virtual reality interface technology matures and as we develop efficient implementations for inhabited dramatic worlds, we plan to investigate ways of replacing the text interface with facilities for speech, animation, and gestures. We hope this work will be in collaboration with researchers studying each of these areas. We have discussed, but not taken, such steps with Andy Witkin's animation research group. COMPUTATIONAL DRAMA Oz worlds are intended not only to be realistic, but to be interesting. Often this means giving people the feelings that come with good stories, feelings that arise in part from the structure of plot, such as complication, climax, and resolution. We understand how this can be achieved for static text: an author carefully constructs the text to convey the structure. However, in interactive fiction we do not write out the whole text in advance, and we don't know in advance the detailed sequence of events a reader will experience. Thus, the Oz system must dynamically, and subtly, adjust the behavior of the world and its characters to provide experiences with the desired dramatic structure. This means developing and implementing a computational theory of drama, and using it to guide the behavior of worlds. We see several approaches to developing such a theory. The simplest comes from having the author express a partial order on the significant events of the story (an "abstract plot graph"), explicitly representing that partial order in the system, and using it to drive the character goals and narrative decisions in the rest of the system. This approach would leave almost all the dramatic theory in the mind of the author, with the plot graph serving as a kind of partially ordered program to be executed by the system. A richer approach is to develop a library of abstract plot units and then, as the interaction proceeds, rapidly search abstract plot space for controllable paths that have the desired dramatic structure. We can view this as a kind of abstract adversary search, where we define a set of abstract operators, means for mapping operators into concrete moves, means for recognizing the abstract effects of the users moves, and an evaluation function on event histories (ie, stories) that lets us recognize sequences with "good" dramatic structure. We are working toward implementations of both of these approaches. The former appears to be a relatively easy way to provide dramatic control signals for experiments with the rest of Oz; the latter is a self-contained long term research goal. In both of these efforts, we are drawing on Brenda Laurel's studies of computational versions of Aristotle's Poetics. (editor's note: the workshop lecture discussed these matters in greater detail). WORLD BUILDING ENVIRONMENT Once we gather and integrate available AI technologies, thus making dramatic worlds possible, we need to provide some means for "normal people" to construct such worlds. For interactive fiction to develop as an art, many artists must explore it, because their feedback is crucial to guide the technology toward artistically desirable goals. While these artists may be computationally inclined, they do not need to be experienced Lisp/AI programmers. Designing the right tools for a general purpose Oz authoring environment can only come after we have some experience building individual Oz worlds. However, we know that building Oz worlds will be a kind of programming. It will involve creating, accumulating, and reusing large numbers of world parts, such as physical objects and settings, parts of minds (planners, plans, kinds of social knowledge), sets of linguistic rules, and components of narrative and dramatic theories. The Oz "parts" libraries, similar perhaps to the backlots of Hollywood studios, will be large and varied. The libraries will contain mechanisms for modifying and building objects (meta-knowledge) as well as individual objects of special value. We believe that the overall library structure and the processes for building it may be similar to those of the NuPrl system. NuPrl is an interactive environment for mathematicians and programmers to use in semi-automatically creating large bodies of explanations and procedures. It is one of the first and most successful systems of its kind, and has spawned related research in the U.S., Japan, and Europe. Bates was one of the leaders of the PRL project, and based on that experience has designed a successor system, called MetaPrl, which is planned as the basis for the Oz authoring environment. USE OF THE SYSTEM It is very important that the technical efforts of building Oz be guided by the needs of artists building worlds. We have already involved several people from the CMU English, Drama, and other departments in this process. We intend to have the population of Oz users grow as the system develops, by teaching courses to the full CMU community and by making Oz available to the research communities outside of CMU. This widespread use is necessary in practice, since it will require great effort to construct a substantial library of world parts and the effort must be distributed over a large base of developers. But we believe it is even more necessary in principle, to learn the potential of interactive fiction as a new art form and to guide the development of Oz toward reaching that potential. -------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- WELCOME TO THE MATRIX: FROM OBJECT ORIENTED TEXT TO QUANTUM INDETERMINACY John G. McDaid John G. McDaid (u1475@applelink.apple.com) is an instructor in the English Dept. of the New York Institute of Technology. He is currently completing his doctoral dissertation on hypermedia composition at NYU, and is also working on an interactive fiction, "Uncle Buddy's Phantom Funhouse." He is a co-founder of the TINAC collective, a group of writers and teachers investigating hypertext and narrative. "Down among the dancing quanta, Everything exists at once. Up above in Transverse City, Every weekend lasts for months..." -Warren Zevon, "Transverse City" Jay David Bolter, in a talk at last year's Modern Language Association conference, spoke about the duality of hypertext: the dialectic of node and link, or as he put it, "looking at" vs. "looking through." The reader experiencing the text he says, is "aware of oscillation; [and this is an] explicit measure of interaction." Ultimately, he was arguing, hypertext wants to be both at the same time. Text which can present itself as surface, and yet effortlessly yield through to other levels. Hearing this, I immediately began to suspect this hypermedia duality was the mirror and analogue of the duality of particle and wave, of energy and matter, in physics. What we are seeing in hypermedia is the appearance on the macrolevel (of 2-meter humans and similar Objects) of quantum reality. In this essay, I'll attempt to explain this perspective (and any lack of clarity is purely my own, and should not be attributed to Bolter) and try to link this with a notion of texts as "transformative utterances," occasions for semiotic recombination which serve the function of conscious dreams. As Jung said somewhere, "The psyche seeking transformation yields symbols." It seems to me as if there is a hidden agenda in media evolution which has guided us into stumbling over just this dichotomy... The ideal for texts that Bolter describes seems already to exist in our minds. Each idea, embedded in an interconnected holographic space, can serve as both the Node, and then, vanishing into itself and passing on to the Other, as the Link. In physics, we understand that all "particles" are really energy, that "energy" is in fact a "particle" with a different 'spin.' The search for the Grand Unified Theory (GUT) is a search for the invariances, the symmetry which, when shattered, gives rise to the duality of fermions and bosons-- things and force-bearers. Our idea of Idea is in fact such a symmetry-breaking operation, cutting out of the quantum flux of mental process this "thing," which we proceed to label an idea, and which then surprises us when it vanishes into its interconnections. Ideas, in the mind, are active symbols in a Hofstadterian sense. Not "signals," or arbitrary strings of characters to be decoded Chinese-Box-wise, but living entities, each with its own propensities, capable of acting. Each "word" in the mind is a nexus of activity. If you will, the mind is Indeterminate Text in its richest sense. Our mind is this constant being and yielding, the entiteification and recombination, the process of being created and sustained above quantum flux, interfacing back down into web-woven synthesis. But this is the ideal of Indeterminate Text; the actualization of this, in current hypertext schemes, can only be Object Oriented. "It is rather a question of substituting signs of the real for the real itself, that is, an operation to deter every real process by its operational double, a metastable, programmatic, perfect descriptive machine which provides all the signs of the real and short- circuits all its vicissitudes." -Jean Baudrillard, "Simulations" Michael Joyce has created a taxonomy for hypertexts: those which are exploratory, and those which are constructive. An exploratory hypertext, says Joyce, is one which fundamentally recapitulates the models of interaction with previous media, like, say, books. You can poke around in an information space, perhaps making a few notes or building trails, but there is a hidden geometry to the space to which you stand in the relation of discoverer, or interpreter. Hypermedia dictionaries, pre-scripted virtual realities, parser-driven interactive fictions are the paradigms of the exploratory. These texts are object-oriented in a deep sense. Like object- oriented computer languages, they comprise a domain of demons, each awaiting its invocation. But they also deeply replicate the phenomenology of objects in our everyday world: they recapitulate what we know about our world. Well, so what? We return, for a moment, to the world of the quantum, a world where Objects can both "exist" and "not exist," where location is a probability, and where, with sufficient energy and time, improbabilities become manifest. Clearly, our presumptive world, the world of our human-size epistemology extruded into exploratory hypertexts, is not isomorphic with the quantum. However, Joyce's other category, the constructive hypertext, provides an indication of the right direction: Constructive hypertexts...require a capability to act: to create, to change, and to recover particular encounters within the developing body of knowledge....These en- counters, like those in exploratory hypertexts, are maintained as versions, i.e. trails, paths, webs, notebooks, etc.; but THEY ARE VERSIONS OF WHAT THEY ARE BECOMING, A STRUCTURE FOR WHAT DOES NOT YET EXIST. [CAPS mine] The constructive hypertext is the embodiment of Schrodinger's Cat. Constructive hypertexts do not proceed from discovery of hidden content, but rather by symbolic creation. A constructive hypertext (and it must be admitted, there are few examples) is necessarily elliptical, open, and metaphoric. For this reason, we are more likely to find them in the province of interactive fiction or 'narrative' than in the commercial world, for reasons that Elizabeth Eisenstein and Marshall McLuhan would explain by pointing to the linkages between literacy, social control, and capitalism... Pragmatism again. But if we are to look for the leading edges of true hypermedia, we must look beyond the pragmatic. When the protomammals internalized their media ecology, putting a symbolic representation of the world into their brain, they leaped immediately ahead of the presymbolic, associative idea-space of the reptiles. But this technology (and its technological progeny) have at their heart this implicit pragmatism, a pragmatism engendered by the world of sensory experience. If we are to catch glimpses of the future of 'text,' we must look for metapragmatic characteristics: 1) It will not be visual. At least not in the same way that we currently think. As McLuhan said, "[Euclidian] visual space...has the basic character of linearity, connectedness, homogeneity, and stasis." Instead, think about virtual realities in 4-D worlds, or Reimannian geometries. Or fractals. And what of the Hilbert Space which had been spoken of... 2) It will not be linguistic. As Roger Penrose maintains in his book "The Emperor's New Mind," language, localized as it is in the left hemisphere areas of Broca and Wernicke, is inconsistent with whole brain knowing or symbolic cognition. He takes issue with the common assumption that without language, thought is impossible. 3) It will be a reflection of the active constitutory action of mind. William Gibson's book "Neuromancer" highlights this distinction. In "Neuromancer," there are two varieties of digital experience: simstim and cyberspace. Simstim (or sensory stimulation) is the digital equivalent of television: neural implants in the "actor" transmit sensory experience which you "tune in" for the experience of "being there." But the cyberspace cowgirls and cowboys dismiss this as a "meat toy." The real action exists in the Matrix, or cyberspace, which is a "Consensual hallucination...a graphic representation of data abstracted from the banks of every computer in the human system." Not everyone can jack into cyberspace and project their consciousness into the matrix. Like reading and writing, or tv viewing and production, there could be another broken symmetry here, if we allow certain models of hypermedia development to dominate... "The sect...would hide in the latrines with some metal disks in a forbidden dice cup and feebly mimic the divine disorder." -Jorge Luis Borges, "The Library of Babel" Evolutionarily, we are phenomenological pragmatists. Our ideation, our "minding," our language, our Texts, all reflect such habits of mind, and the technologies we externally create to instantiate these ways of minding are all rigorously subject to the contraints of the practical. We are all naive realists and tacit essentialists. "Reality is the shattering of the highest law of motion taught us by experience," says Ernst Cassirer. Quantum mechanics is the shattering of pretty much the REST of the truths we learned poking around in our world of Object-Oriented Childhood. Roger Penrose argues that thought is non-algorithmic and strongly dependent on quantum effects. According to Penrose, "having" a thought is the result of virtual quantum computation, a collapse of superposed thought-functions, and which particular thought we have is a non-local solution to the presenting problem (non-local in the quantum mechanical sense). In the face of this, algorithmic representations (which are the pragmatic progeny of natural language) become sterile replications of received pattern, as unable to produce artificial intelligence as cookie cutters are to produce gingerbread humans. What Penrose's argument implies for hypermedia design is a deep challenge on its most basic assumptions. It seems likely, then (he said, falling into the teleological snare), that the function of consciousness is to become aware of its limitations--factors latent in the anthropotropic media with supply it--and to bootstrap itself through technological augmentation into modes of awareness/consciousness which are enactments, on this layer or level of reality, of the fundamental indeterminacy of the universe and our brains. Our dreams of the Matrix then, are dreams of a TIKKUN, a re-integration of the microlevel and macrolevel. Is it accidental to have choosen such a metaphor as the Matrix, the mother? Is the coincidence in the rise of patriarchy and phonetic alphabets happenstance or a smoking gun? As the ekstasis of intuition was paved over by the tarmac of pragmatic text, so went the model of self and the symmetry of gender. Object-Oriented Artificial Intelligence is nothing more than the latest attempt by the patriarchy to reproduce itself, AB NIHILO, by uttering the Word. It remains to be seen if truly Indeterminate Text can be instantiated macroscopically. Perhaps not by digital systems. They are either/or systems, and while they can model, or approximate indeterminacy, they do not, at bottom, embody it, and therefore seem essentially incapable of manifesting it. Mind, however, seems to have this capability. Perhaps, as Penrose suggests, understanding mind will give us insights into the pathways of the probabilistic. Were we to construct a fictiverse whose laws were those of quantum mechanics, a fictiverse we could inhabit as a metapragmatic consciousness, would we would be capable, in principle, of constructing indeterminate texts about our experiences? Maybe. Perhaps the immobility of the things that surround us is forced upon them...by the immobility of our conceptions of them. -Marcel Proust, "Swann's Way" -------------- A R T C O M / DECEMBER 1990 / VOL. 10(10) / #44 -------------- LIFE-LIKE CHARACTERS IN INTERACTIVE FICTION David Graves David Graves (dag@hpsemc.cup.hp.com) has been doing research and development in Interactive Fiction for the last six years. He is currently collaborating with writers on several IF projects, using a software platform of his own. A common failing of interactive fiction today is that the player is totally responsible for the progress of the plot. Whenever the player cannot solve a given puzzle, progress grinds to a halt. Most interactive fiction works are entrenched in this obstructionist model. There is no plot continuity, and too little dramatic interaction. Our characters are not pulling their own weight. The first step is to take the player out of the role of being responsible for the flow of the plot. In her PhD dissertation, Brenda Laurel explains: "The user of an interactive system may indeed make contributions on the level of plot; however, the responsibility for integrating such contributions into the whole and creating other plot elements that maintain the necessary dramatic qualities in the whole belongs to the system. When that responsibility devolves to the user, first-personness is destroyed, as in the classroom improvisation where the actor must divide his attention between acting and playwriting tasks. By assuming formal control of the action, the system frees the user from playwriting concerns and allows him to immerse himself in the experience of his character." Furthermore, the characters found in most computer games are cardboard cut- outs, without any personality. In her PhD dissertation, Mary Ann Buckles points out this failing in the original Adventure game, which still applies to most games today: "The characters the reader encounters in the fictional underworld have no significance other than to pose a puzzle for the reader. The reader has little emotional involvement with the characters because they, in turn, do not represent any emotional or spiritual facets of human existence." Artificial Personality addresses this need for simulating personality in our characters. Almost all attempts to generate behavior in computer-controlled characters have followed a simple stimulus/response model. For each statement you may make to character, there is one response it may display. Giving the same stimulus several times in a row, the character will mindlessly repeat the same response. Clearly, there is room for improvement in creating lifelike characters, but how can we attack such a difficult task? To start, we can borrow a number of ideas and methods from the field of Artificial Intelligence, avoiding the difficult software problems, to produce the >illusion< of intelligent, emotional, motivated characters. In order for characters to act intelligently they must be able to interpret the state of their environment and apply appropriate behaviors. While applying a problem solving procedure to achieve some goal, however, complications may arise. In traditional interactive fiction, we pass all these problems to the player. Instead of troubling the player character with minor complications, the game software could automatically resolve them. For example, given the command "Drink the beer," rather than having a character complain "The beer isn't open," it could recognize "Open the beer" as an implied subgoal. Any given goal could give rise to a number of subgoals, which may create subgoals of their own. When a character is able to handle low level logistics without being given explicit instructions, he appears much more intelligent. This technique also provides a mechanism for handling tedious logistical details on behalf of the player, who is then free to think at higher levels. No organism's behavior is ever unmotivated. Thus, in order for characters to display behaviors that appear reasonable and believable, they must have their own motivations. These motivations help stimulate the generation of plot. However, without guidance for the plot, chaos is a likely result. To ensure that the generated plot is interesting, the system could have some concept of drama and apply it to the currently unfolding story. Laurel's dissertation gives an outline of an expert system to do just that. "Understanding a story in its totality is a task that integrates natural language understanding and the understanding of characters' goals, plans, traits, and emotions, and utilizes still other techniques for identifying larger patterns of action." This computerized playwright would recognize opportunities for new plot twists and act on them. Clearly, this is a lofty vision, requiring vast resources to implement. Several projects have been successful in creating small expert systems that focus on character behavior and interaction, rather than attempting to recognize and generate plot units. This requires a representation (in software) for the emotional state and goals of each fictional character. Using this model, each character's emotional state and current goals drive the selection of a specific behavior from a large set of possible behaviors. The intensity of the appropriate emotion values is then used to determine the intensity of the expression of the behavior. Even when performing simple actions, a character's hidden emotional state may "leak out." Thus, the way in which a character attempts to accomplish his goals may be influenced by his emotions and those of the other characters. Due to the complexity of the emotional state of the characters, the sub-plot twists are unpredictable, and due to the goals inserted at plot-critical times, the author can control the overall plot coherence and pace. In creating a model of personality and relationships, one must select a manageable set of emotion variables. The magnitude of these variables will define how each character relates to the others. Each of these emotion values tells how one character feels about another. John may love Mary, but Mary may not love John. Conflicting emotions and goals between characters help keep the action lively, giving rise to new goals all the time. In addition to these two dimensional emotions (directed towards other characters), some one dimensional emotion variables may be created, which indicate a character's internal emotional state or mood, or personality attributes that remain constant. Michael Lebowitz, creator of a program that writes soap opera stories, uses one-dimensional attributes such as niceness, guile, physical appearance, and promiscuity. In most interactive fiction products, characters are treated as objects. Most interactions with other characters are limited to making imperative statements to them (giving commands). True interaction with characters is impossible in these worlds because the representation of the world is void of any "interactive media." You cannot talk with them because there is nothing to talk about. In worlds containing only objects, the only topic of discourse is the "object economy" (physical objects that may be manipulated). In order to produce interaction on more human terms, a system must have (1) a rich representation for emotions, knowledge, and beliefs, (2) a rich set of behaviors that are driven by those items, and (3) a rich grammar for communication of knowledge, events, beliefs, and emotions. These subsystems must be fully integrated with each other. One would not want to design-in emotions that cannot influence behavior, or that cannot be talked about using the input grammar. At the center of any Artificial Personality system is an emulation of human emotions. Besides providing new motivation for believable behavior, emotions give the characters a new domain for discourse. They may interact on the levels of physical state, information state, and emotional state. In designing an interactive story, the designer must keep in mind the interlocking dimensions of physical state, emotions, character beliefs, behavior, and communication. One must also keep sight of the vision: characters displaying believable original behavior and engaging in interesting, dramatic interaction. NOTES Brenda Laurel: "Towards the Design of a Computer-based Interactive Fantasy System" (1986). Defines the vision and the technologies required to implement it. Mary Ann Buckles, "Interactive Fiction: The Computer Storygame 'Adventure'" (1985). Michael Lebowitz, "Creating Characters in a Story-Telling Universe" Poetics, 13, 171-194. (1984) -------------------------------- END OF FILE ---------------------------------