metaleptic world

I’ve done no work in the last couple of weeks. None. Been either too busy, or too mentally and emotionally exhausted. Tomorrow a very welcome houseguest arrives; I’ll probably get to do some work over this visit, strangely enough. This is why I like writers! They can amuse themselves for an hour, easy.

You got your wish, François: I am unable to sleep because of metalepsis. In writing up Marie-Laure Ryan’s plenary talk of this past May, I’ll be going on my sparse notes, so please forgive and correct me if my recollection is wrong. I have a habit of quoting speakers directly when taking notes, but later do not have any way of figuring out which are the actual direct quotes, and what I paraphrased. (Typing quotation marks takes up too much time, when trying to record thoughts expressed at normal speaking speed.) Consequently, anything enclosed in quotation marks below refers to the notes file, and may or may not be a direct quote from the speaker. In other words, blame me not them.

Let’s start with another paper, given at the same conference by Wendell Piez (who is brilliant and articulate and one of my academic role models, which is remarkable because his day job is not in academia). Wendell spoke last in his session, drawing together the previous two papers: Susan Brown’s talk on the Orlando Project, and Willard McCarty’s paper on Ovid’s Onomasticon. In his discussion of the different approaches to encoding taken by these two impressive projects, Wendell drew a distinction between proleptic and metaleptic encoding. The context was deep (not merely thick) semantic markup of texts: a well-like nesting of markup elements as a modelling tool for literary knowledge. Any other knowledge too, I suppose, but in these papers people were mostly talking about information structures present in literature. One of Wendell’s central questions was "the question of whether markup is really very good for modelling on the deep level at all." From the notes (edited for readability):

Prolepsis anticipates (modulates, plays against) a future signification (e.g. dramatic irony). Metalepsis recovers (interprets, commandeers) a past signification. (Bloom: poets revising each other). Applied to markup languages, both are kinds of “generic” or descriptive markup. Prolepsis looks forward to processing (W3C Recommendations tag set, frex). Metalepsis looks toward description of an artifact or presupposed existence or state (TEI, frex). Both "propose theories of the text," except one stands as a representation while the other enables a media format. Most e-text projects include aspects of both.

We are ceasing using the computer as simply an indexing tool, and beginning to apply algorithms to our data, so that information could be revealed that cannot be seen on index cards. This is what Orlando and Onomasticon have in common.

"Transformative" criticism?

These projects are impressionistic. They do not take the author or the designer out of the tool. Many of these projects are very collaborative in nature. We take advantage of the networking opportunities of the computer.

Perhaps deep markup is not the core issue here. Not the modelling technique, but the context in which the modelling is happening is the most interesting thing.

Roland^HT, then, is completely metaleptic. Commandeering and reinterpretation if ever I saw it. I’m blithely changing almost every constituent work’s medium, nullifying some potentially crucial semantic information. I’m picking up on themes and imagery and (by way of hyperlinks? or a database? the possibilities are many and exciting) hurling them into other contexts. Or perhaps hurling the reader on the back of a theme or image. Literally by the hand, since in most cases it is the hand which activates the link which in turn transports Reader to another point in a text. (Here and elsewhere, I use the term text loosely.) Like Ryan’s Scheherezade (see below), I remain on level zero, completely outside my worlds.

Here’s where I run into trouble. In her plenary, Marie-Laure* treated narrative levels as stacks. Level zero is the actual world, with the author and the reader present in it. Then, narratives are stacked or nested (is there a difference, here? help…). I (the author) tell you (the reader) about a girl who tells stories. That’s level one. The girl tells a story about a pauper who sneaks into a king’s palace. That story is level two. The pauper gets caught, and in order to not get punished spins some yarn about a dead mother’s ghost asking him to accomplish a task. That yarn, in turn, is level three.

It is not clear to me whether, to Ryan*, metaleptic narrative is equivalent to stacked or nested narrative. If it is, then Roland^HT is not metaleptic. The stories present in it, the constituent works, aren’t stacked. For the most part, they’re on the same level; and would require not deep but sprawling markup.

But a larger question is whether it matters at all. Metalepsis is a cool word, and a handy one for describing a process which "recovers… past signification." But does it matter whether this specific project is metaleptic or proleptic, or if both then to what degree?

I think it will matter later on. My biggest conceptual problem right now isn’t how I’d like to go about analyzing the corpus; it is how I’d like to present it. I’ve no idea what would be best for a user interface, and can’t really talk about it with anyone until the corpus is marked up, or at least until there’s a markup/organizational system in place which is comprehensive enough to allow for future additions of corpus works. When the interface or a prototype is ready for scrutiny, then it might be interesting to look at [pro|meta]lepsis in the project. This issue will be native not to the data, but to its specific presentation, and studying it might be useful in thinking about electronic modelling in the humanities in general.

Other notes from that plenary, presented without comment (but feel free to correct, interpret, or further comment):

Rhetorical metalepsis: keeps narrative levels separate. Example in patinting: Magritte, La Condition Humaine. Example in narrative: The Arabian Nights (Sheherezade is not part of the world of her story, remaining on a separate level from it).

Ontological metalepsis: opens up a passage between narrative levels that results in their interpenetration. Example: Woody Allen’s short story [Title]. Also, Julio Cortazar, Continuity of Pacts. Elements (usually characters) migrate between two levels of a story while remaining part of each level. Exmaple in visual art: Escher, Drawing Hands.

D. Hofstadter talks about metalepsis (though not in these terms) in Godel Escher Bach.

Tangled hierarchy: when what you presume are clean hierarchical levels take you by surprise and fold back in[] a hierarchy-violating way. [rhetorical metalepsis]

Strange loop: when, by moving upwards (or downwards) through the levels of some hierarchical system, we unexpectedly find outselves back where we started. [ontological metalepsis]

Ground level [zero] is immune to metaleptic contamination.

Incomplete (containing true propositions which cannot be proved) vs. contradictory (when some of the propositions are true and false at the same time, in violation of the principle of non-contradiction) systems

This sentence is false. <--- the sentence is both used and mentioned

What do you do when you have a system, and one of its components states “I am not a component”? You can exclude it, and thus have an incomplete system; or retain it, and have a contradictory one.

Baudrillard’s view of the 4 stages of the image:

1. it is the reflection of a basic reality
2. it masks and perverts a basic reality
3. masks the absence of a basic reality
4. bears no relation to any reality whatsoever: it is its own pure simulacrum.

*Argh. I dislike the nebula of not knowing whether to refer to people by their first or last name.