A couple of weeks ago Mark Nelson posted some tweets that caught my attention:
Interactive narrative isn't my research focus anymore, but if I were to restart it, I think the angle I'd take is to try to build computational models from the material in Marie-Laure Ryan's ca. 3-decades body of work, hopefully providing a much more solid theoretical grounding.— Mark J. Nelson (@mjntendency) March 20, 2018
It's cited fairly frequently in interactive-story / drama-management / etc. work, but mostly for a handful of high-level concepts. Think there's lots more as-yet unused in there. Understandably, b/c also heavily in the mode of literary theory and would need translational work.— Mark J. Nelson (@mjntendency) March 20, 2018
Not that this is what I'd suggest as a methodology if you want to build games w/ interactive stories or anything. More of an academic research agenda. Also not really apropros of anything— thought of it today b/c I was reviewing some papers before meeting w/ a PhD student here.— Mark J. Nelson (@mjntendency) March 20, 2018
I’m not sure where that’s going and, as it seems to have been a quick spur-of-the-moment comment, it’s probably not the sort of thing that’s going anywhere in particular. It’s just a quick probe to “mark” a region in conceptual space. But I thought a bit and made a short response, to which Mark replied:
LOTS of translational work, LOTS. It's not at all clear to me that 'translation' is a useful way to think about it, either.— Bill Benzon (@bbenzon) March 20, 2018
Yeah, that's certainly plausible. More of a tweet-level thought than thought-through proposal. :-) Did run across a few papers, e.g. a narratological analysis of time in branching story, that seem temptingly like they could be applied somehow though.— Mark J. Nelson (@mjntendency) March 20, 2018
The fact is, I wanted to say more, but I couldn’t think of anything. I’m still thinking and have a quick comment or two.
Narratology has been within my field of interests since the late 1970s and I’ve read a bit, some articles and book-length expositions, but it never really got me hooked. I became aware of Marie-Laure Ryan's work in the late-1990s when I picked up her Possible Worlds, Artificial Intelligence and Narrative Theory (1991), which, as far I know, is one of the few (only?) books of literary criticism that utilizes some of the technical concepts AI. By that time I had long since made my own peace with computational work on narrative and story-telling and didn’t find much of use to me, which is neither here nor there. But, as I recall, she did note that computational models are so complex that it is difficult to apply them to literary texts; the amount of resulting detail is too difficult to follow. That, I suspect, speaks to the need for “translational work”, and LOTS of it.
But that–level of detail–isn’t all. Early in Pathways of the Brain: The neurocognitive basis of language (1999), Sydney Lamb discusses various types of grammar. One of them he calls analytical. The object of analytical grammar is to describe language products, typically sentences. Chomsky’s approach is typical of analytic grammars. Lamb contrasts this with the neurocognitive grammar he’s been developing, which aims to characterize the computational mechanisms by which language is enacted and understood. As Lamb conceives these matters, neurocognitive grammar is necessarily computational while analytic grammar is not.
My teacher, David Hays (also a friend of Lamb’s), made a similar point by talking about bicycles. One the one hand, we have the blueprints for a bicycle, which detail all the parts and how they fit together. But the blueprints don’t tell you how to assemble a bicycle from its parts. Assembly instructions are quite different.
Well, literary criticism in general, including narratology, is analytic in Lamb’s sense–though informal by the standards of contemporary linguistics. It attempts to provide blueprints for texts, not assembly instructions. And Ryan's work in Possible Worlds seemed pretty much like that despite her uses of computation. This is not simply about detail. It’s about kinds of detail and kinds of intellectual strategy. This is where the real translational work is going to have to be done.
See also: Why Levi-Strauss is Deeper than Propp? (from 2011)
My impression is that narratology has absorbed Propp, but not Levi-Strauss, and the same for computational work on stories.
"Assembly instructions are quite different."
ReplyDeleteI had a class as an undergrad with an extremely good German folklorist. She presented four German 'Jack' tales of a particular type.
I noted that if you placed them together as a group and presented them in a particular order you could considerable heighten the comic effect.
Very clear pattern.
She could not understand what I was talking about, each had a separate Aarne- Thompson classification and therefore the tales had no relationship with each other.
Propp + the historical geographic method. Can be restrictive in the extreme.