17 February 2010

digital storytelling, Zola, experimentalism, & the scientific method

I managed to make two panels at CHAT today before I had to dash back to Duke for class -- one on audience size in digital projects, the other on digital storytelling. I won't rehash them, since others have already done that better than I could. I encourage you to visit these posts, or follow the #uncchat twitter feeds.

Instead, after a brief summary, I wanted to follow up on a few mental sparks lit during the "Soundbyte" session on storytelling. The panel has two titles on the CHAT website: on one page, "Scientific Method and Narrative Form"; on the other, "Transforming Narratives." Both descriptions pit the "electronic literature expert," embodied in Kate Hayles, against the "computer scientist," Michael Young, who works with the Liquid Narrative research group at NC State. The panel itself intended to "explore the intersections and the opportunities" of such a pairing.

Young started the conversation, using analytic theories from cognitive science and linguistics to construct a computational, generative model of narrative. As I tweeted earlier, his approach to storytelling reminded me very much of Nick Montfort's talk to HyperStudio on designing his new interactive fiction programming language: like Montfort, Young is interested in building comprehensive systems for users to engage in, interact with and act within stories, themselves conceived as an unfolding series of events (i.e., plotlines) constructed by an (external) author-agent. As Young underscored, this approach to storytelling is governed by the "clash between control and coherence," as users must be given the freedom to make choices within a nonetheless scripted narrative.

Hayles provided a counterpoint to the kind of causal, plot-driven method of Michael Young. In several mini-close readings of three pieces of digital fiction -- Michael Joyce's Twelve Blue, Judd Morrissey's The Jew's Daughter, and Kerry Lawrynovicz Girls' Day Out -- she pointed out that that visual metonymy of "threaded" details (Twelve Blue), the complex temporalities of textual palimpsests (The Jew's Daughter) and fragmented subtexts (Girls' Day Out) all offer complex, non-causal methods by which narrative can emerge in digital fiction. She concluded by asking whether any of these stories could be described as algorithmic or generative, and whether audiences would accept these more nuanced narrative strategies.

I think the differences between these points of view -- which is really the difference (for me personally -- this isn't a value judgment) between what's interesting about literature, and simply what isn't -- came out in one of the questions. Someone asked: when talking about causality and literature, where does the scientific method fit in?

Michael Young said he was inspired by the scientific method -- that he, like other scientists, takes a theory, builds a model of that theory, then runs human subjects through a controlled experiment to see if it validates his hypothesis. The challenge, he added, is to design a model that is controlled enough that you can attribute the result to the theory itself, and not outside factors.

Now this is fascinating to me. I recently read Émile Zola's novel Germinal, alongside Claude Bernard on experimental medicine and Zola's essay "The Experimental Novel" in which he explicitly positions his work as a novelist in terms of producing a scientific experiment -- that is, of setting up a controlled, contained environment in which characters act out their own socially-determined possibilities. The results are nothing less than a total scientific understanding of human psychology:

[T]he novelist is equally an observer and an experimentalist. The observer in him gives the facts as he has observed them, suggests the point of departure, displays the solid earth on which his characters are to tread and the phenomena to develop. Then the experimentalist appears and introduces an experiment, that is to say, sets his characters going in a certain story so as to show that the succession of facts will be such as the requirements of the determinism of the phenomena under examination call for.

We are making use, in a certain way, of scientific psychology to complete scientific physiology; and to finish the series we have only to bring into our studies of nature and man the decisive tool of the experimental method. In one word, we should operate on the characters, the passions, on the human and social data, in the same way that the chemist and the physicist operates on inanimate beings, and as the physiologist operates on living beings.

In a widely-read essay on the differences between realism, which narrates, and Zola's form of naturalism, which mechanically describes, Lukács criticizes Zola's methods as merely accumulating detail for the sake of some external (and, although he doesn't use the word, experimental) purpose, at the expense of the "inner life" of literature. He writes:

When the artistic literature of a period does not provide actions in which typical characters with a richly developed inner life are tested in practice, the public seeks abstract, schematic substitutes.

In failing to provide an adequate narration of the "inner life" of the human, Zola's experimentalist naturalism inadvertently spawns "abstract, schematic substitutes," by which it seems Lukács means detective novels and hardboiled pop fiction. I wonder, though, if we can't map this distinction onto the differences articulated by Young and Hayles, in a panel of course called (on at least one site) "Scientific Method and Narrative Form." If Young's computational model of storytelling is the new Zola-ian experimental fiction -- and I think his own description of his work as a "controlled experiment" makes the analogy fairly solid -- then the fictions Hayles describes are the new "abstract, schematic substitutes": a literature sunk deep in an obsession with words, with language and the materiality of letters. It's perhaps no surprise that digital storytelling splits along these lines. Although we tend to think of Young's form of algorithmic literature as arising from video games and more recent forms of "interactive" narration, it may have a longer lineage to turn-of-the-century schools of naturalism, narratology and the push to scientize literary fictions as experiments played out by constructed human subjects, as a means of sparking understanding in socially-determined reading subjects. In fact, programmable interactive fiction systems may just be the latest stop in the search for a perfectly controlled, and perfectly controllable, narrative microcosm.

So why does this matter?

First, none of my comments should be taken as a criticism of the kind of work that Young (or someone like Montfort) does. I deeply appreciate the ways in which digital storytelling models merge hardware, software, structure and text into a perfect little machine driven by user experience. It's a thing of beauty that, unlike the novel, never forgets its own medium.

At the same time, these methods are not new. Placing past methods in conversation with the present helps delimit the boundaries of the conversation by probing our own assumptions. For instance, a longer reading of Zola alongside interactive fiction systems would perhaps illuminate the nuances in a term like "virtual worlds" or "simulations," both narrative models which I think Zola would shy away from in favor of a more determined notion of experimentalism -- that is, of seeing his narratives as controlled experiments which assemble and operate on literary objects with the explicit goal of furthering our understanding of the human subject. What can the machinery of Zola's novels -- very like video games or film in their accumulation of detail and excessive attention to environmental conditions -- teach us about the built-in hypotheses in our digital storytelling models? And (the question that I personally find more interesting) what literary methods escape the scientizing of literature? What forms of narrative exceed the experimental method?

In her response to the question on the scientific method and narrative -- the question that started this tirade of mine -- Hayles pointed out that experiments often begin with a state of initial confusion, in which the observer is bombarded with information that must be sorted out, not unlike reading a novel such as Gravity's Rainbow. The payoff only comes after a significant amount time, and only emerges through nuanced and multilayered textual strategies. This strikes me as right, and perhaps provides a way of bringing these two divergent strands of thought back together. Only through the act of interpretation can stories, interactive narrative environments, or scientific experiments find, and make, meaning in the world.


eastgatesystems@mac.com said...

Are you coming to FUTURE OF DIGITAL STUDIES next week at the University of Florida? http://www.english.ufl.edu/fds/

Whitney said...

Unfortunately, no; I was considering submitting a proposal, but this semester is packed for me and I simply can't afford to do anymore traveling on a grad student's stipend.. will be following tweets on it, though!