var news early June
- Time to start thinking about Gnoetry Daily Volume 2! (following up Volume 1 from last year.) I recommend a release date in early December – Just In Time for Christmas!!! JYNX said he was interested in editing, which is cool with me; speak now if you got a problem with that.
- shout out to the Computational Linguistics for Literature workshop this Friday (co-located with NAACL). Check out the details and the papers. Of particular interest are Kao and Jurafsky’s “A Computational Analysis of Style, Affect, and Imagery in Contemporary Poetry” (non-avant- poetry, natch) which claims “the most important indicator of high-quality poetry we could detect was the frequency of references to concrete objects” (!?!); and Brooke et al’s “Unsupervised Stylistic Segmentation of Poetry with Change Curves and Extrinsic Features” which attempts to automatically detect voice shifts in Eliot’s “The Waste Land.”
- I recently came across Talan Memmott’s dissertation “Signifying Strategies in Electronic Literature” from last year (available off of a Malmo U web page, though I sometimes have problems downloading it) which talks a bit about charNG and codework parenthetical insertions! Apparently charNG “is more interesting for it interface than the poetic qualities of its output” and “the purpose of charNG seems less intended for the analysis of language than an exploration of the procedural properties that allow for such an analysis to take place through computation.” I guess that’s right, since the reason I wrote charNG (as described on a netpoetic posting) was to get a feel for character n-gram generation. Although: I don’t really see charNG as something that generates final-draft code, but rather code that I can look through for interesting sequences I can put together. For example, for “all preceding numbers” I generated approx. 2400 non-space characters, of which I selected approx. 263 non-space characters for the final version, i.e. only using 11% of the originally generated text.
Regarding codework parenthetical insertions, Memmott says:
“Addad’s text mapping does not produce the same sort of embedded critical, philosophical, variable and subjective text that Mez is best known for. What are reproduced are the most banal formal aspects of Mez’s technique – the bracketed insertions – and though the mapping is interpretative, the output lacks the mastery and prerogative of a text by Mez herself.”
True, true, all of it true!!! Though in my defense I gotta say that parenthetical insertions are something I use in conjunction with other techniques (such as selected character n-gram sequences, as in the “Decline and Fall” pieces) rather than by themselves.
Anyway, I’m totally not hating on Memmott, check out his dissertation, which covers loads of other writers!