February 20, 2008
That was a significant amount of detail about Tale-Spin, more than I will offer about any other system described in this book. I hope it gave some sense of the type of undertaking involved in creating even a first-generation story system. There’s much more going on — at the levels of character and story — than in something like Eliza/Doctor or a standard computer RPG. Further, it illustrates how a computer system that seeks to generate representations of human behavior can be built as an operationalization of theories about human behavior.
But it’s also worth noting that the story produced in our Tale-Spin example wasn’t a particularly strong example of fiction. While Tale-Spin creates character behavior, this behavior doesn’t necessarily take the shape of a traditional story. This is something to which I’ll return later. For now, I want to consider what it means to say that Tale-Spin produces fiction at all.
A shorthand definition of fiction might be a listing of its most familiar forms: “novels and short stories” or perhaps “novels, short stories, theater, television, and movies.” But such definitions won’t be much help to us if we are interested in thinking about emerging forms. At one point the novel was an innovative new form of fiction, and would have been left out of such a definition. Now there are various innovative digital forms of fiction emerging. In order to think about new forms, we need a more principled definition of fiction.
In response to this, a number of digital media theorists have begun to work with definitions of fiction emerging from possible worlds theory and modal logic. These are among a wider group of literary approaches derived from philosophy and linguistics, and this particular strand began to establish its own identity in the 1970s. Marie-Laure Ryan, in her Possible Worlds, Artificial Intelligence, and Narrative Theory (1992) traces the lineage of this work to the late-1970s essays of a number of writers.8
In philosophy the notion of possible worlds is traced back to Leibniz, whose position is perhaps most widely remembered as that devoutly held by Dr. Pangloss in Voltaire’s Candide. Leibniz argued that there are an infinity of possible worlds in the mind of God — and that one of them, the best possible, is the one in which we live. Leibniz’s position is engaged with the most elevated concerns of philosophy and theology, but that is not the only place we find possible worlds. Closer to home, we imagine possible worlds when we speculate (“I wonder if the lunch special will be Lemongrass Tofu”), when we wish (“I hope Jane gets that promotion”), when we plan (“We’ll get there before the doors open, so we have great seats”), and so on. Possible worlds, and modal logic more generally, are tools for thinking through versions of such questions: What is possible? What is impossible? What is necessary, or contingent? The work of a number of thinkers has, especially since the 1960s, returned philosophers to consideration of these topics, well after the time of Leibniz.
A simple attempt at applying possible worlds theory to fiction might propose that non-fiction texts refer to our world, the real world, while fictional texts refer to alternative possible worlds. However, as Ryan points out, there are a number of problems with this. For example, this formulation does not provide a way of distinguishing between fiction, errors, and lies — all are statements made in reference to alternative possible worlds. Also, at a more complex level, there is the problem that further alternative possible worlds are continually embedded into both fiction and non-fiction. For example, both fiction and non-fiction describe the unrealized wishes and plans of the people who appear in them. Ryan’s work responds to these issues by identifying a further element, beyond the creation of an alternative possible world, that is necessary for fiction.
Ryan considers the constituent move of fiction not simply the creation of an alternative possible world but the recentering of discourse to that world — so that indexical terms such as “here” and “now” are understood to refer to the alternative possible world, and terms such as “actual” themselves become indexical. Further, for Ryan fiction not only creates an alternative possible world, but also a system of reality, a universe. This is necessary because the alternative world of a fiction may also have many alternative possible worlds emanating from it, and each of them may have further alternative possible worlds (as when one character speculates as to the plans of another character). Obviously, parallels with the operations of Tale-Spin were already at work in the description above.
Beyond the field of story generation (an interest I share with Ryan) a number of authors have used possible worlds theory as a way of grappling with emerging forms of fiction in digital media. For example, Jill Walker Rettberg, cited earlier for her discussion of quests, is a digital media theorist interested in many forms of fiction with a strong textual component (including email narratives and weblog fictions) as well as some without text as a primary component. Her Fiction and Interaction: How clicking a mouse can make you part of a fictional world (2003) is an examination of interactive digital fictions that somehow include the user as a character in their alternative possible world. She introduces the term “ontological interaction” in the course of describing how users are included in these worlds, building particularly on the notion from Thomas Pavel’s book of “ontological fusion” (between our actual selves and fictional selves when engaging with fictional worlds) and Kendall Walton’s theory of how we use fictional representations as props in a game of make believe. (She cites Walton’s 1993 Mimesis as Make-Believe.) Walker notes that this approach provides a vocabulary for discussing the common generation of fictional experience in works “as disparate as installation artworks, interactive narratives and computer games” (31).
In a related vein, Half-Real (2005) — by game developer and theorist Jesper Juul — takes computer games as the primary objects of study. Given this, Juul is not interested in “fiction” as a category of artworks, but as something that is an element of many games. He is interested in understanding how games project fictional worlds: what kinds of worlds, how players are cued to imagine them, and how those worlds relate to the games’ rules. Juul notes that game fictions, like those of traditional literature, are incomplete. Just as in Hamlet we know little of the world outside the castle and its immediate vicinity — but on some level assume it is embedded in a fully-detailed world partially filled in by knowledge from our own world and other texts — so in the game Half-Life we know little of the world outside the complex where it is set. However, unlike most literary fictions, the worlds of many games are, in Juul’s terminology, “incoherent” (which is one of the things that limits Juul’s interest in discussing games in terms of narrative, as opposed to fiction). These are worlds in which significant events take place that cannot be explained without discussing the game rules, such as the many games that feature multiple lives and extra lives without any element of the game fiction that points toward reincarnation. Juul describes a number of ways, going far beyond traditional novelistic techniques, that games cue players to imagine possible fictional worlds, including: graphics; sound; text; cut-scenes; game title, box, and manual; haptics; rules; player actions and time; and rumors (133–138).
Given that authors considering the concepts of fiction and digital media from differing directions have found the possible worlds approach to fiction fruitful, how can it help one understand Tale-Spin, and story generation more generally? First, of course, it can clarify that Tale-Spin’s operations produce fictions. The psychological action noted during George’s speculations about Arthur is revealed as precisely the stuff of fiction when seen through the lens of possible worlds theories. Second, this approach provides a way to think about fictions that is purely in terms of structure. This is important because Tale-Spin, like most story generation projects, is concerned with producing the events of stories — and very little with how those events are presented to the audience. In literary theory this is sometimes referred to as the distinction between “story” and “discourse” (or, in some more specialized contexts, “fabula” and “sjuzhet”).9 In fact, Meehan was so little concerned with the presentation of Tale-Spin fictions that Mumble — the natural language generator used to turn Tale-Spin’s CD expressions into English — was cobbled together in a single day. (However, as I will discuss shortly, the problems with Tale-Spin’s output run deeper than Mumble’s stilted constructions.)
Most importantly, seeing fiction in this light can help us identify the core operational logic at work in Tale-Spin: planbox-based planning. This method of planning, as outlined by Schank and Abelson, operates by projecting potential behaviors that change the state of the world — possible worlds — which launch further projections, which launch further projections, finding a “chain or lattice” of worlds that may lead from the current world to one in which a goal is satisfied. The Tale-Spin system certainly contains other operational logics, such as those governing movement, character relationships, and so on. But Tale-Spin does not begin with a complete virtual world that characters can move across. Instead, movement happens only when required by plans, and the world is fleshed out only to the degree required by plans. Similarly, Tale-Spin does not generate a complete set of interpersonal relationships when characters are created, but only fleshes out the connections between characters are required by plans. So while other operational logics may contribute to the fictional worlds of Tale-Spin, it is the logic of planbox-based planning that is central. It creates the profusion of imagined worlds that define Tale-Spin’s fictions and drives the work of the system’s other logics.
8Ryan particularly notes Lucia Vania, Umberto Eco, David Lewis, and Lubomir Doležel — followed by the 1980s books of Doreen Maître and Thomas Pavel. If we look at the first chapter of Pavel’s book (Fictional Worlds, 1986) he sketches a similar picture of Doležel and Eco’s contributions, while also noting the importance of Ryan’s early essays on the topic.
9Or, to put it another way, a fiction’s events and their portrayal.