Skip to main content

glitch: science and aesthetic

erroneous output of phatica4

I've finally gotten around to doing some tangible work with computer poetry.  Very preliminary stuff: for this round of experiments, my programming contributions are mostly just "glue code".  (But isn't that concept itself at least mildly interesting?)

I've put a preprint/draft online, "X575: a renga writing program in its infancy".

Here's an example poem.  As my coauthor pointed out, "It asks for a lot of interpolation from the reader."

busied for pandas --
the maternity has born
a knack of aircraft

judges of today
nothing but number to bus
to his weatherman

defacing passing --
I went to a crude bearing
the nova of mobs

that pause in the prayer --
this humanitarian --
from the spongy cruise 


My only reason for selecting this particular example (which didn't make the cut for the paper) was that the reference to "the nova of mobs" is a completely inadvertent (well, everything here is inadvertent) reference to Burroughs's (1964) Nova express, and the cut-up method used/contained therein.

That only comes across for a reader with a bit of familiarity with Burroughs "lore" -- which I find to be an interesting glitch.  Burroughs himself writes about the way certain ideas in a cut-up piece of text will jump out to the reader.  Or, perhaps more accurately, the reader will pull those ideas out of the text.

It seems to me that these sorts of glitches put an author/reader/interactant in touch with the medium "more" than a more straightforward text would do.  In "Metaphatics Metaeverything" one of the interesting quotes is about call-in radio programs in Nepal, where, in my opinion, people enjoy "performing" an experience of being in touch with the medium.  So, we're considering not just ritualized exchanges, but also things like radio static or line noise as an important part of the ritual.

"Glitch" as a wider aesthetic seems quite similar.   It also reminds me of Freud's idea of the uncanny.

Perhaps in an evolution of the poetic ideas described in the "X575" paper we could start to home in on the features that make some particular effect glitchful or uncanny.  For example, if the generated text clearly picked a topic, and stuck with it, then it would come across as organized, orderly, explicative and so on.

In the conclusions of my current draft, I've proposed that if the computer explained what it was doing when it was writing a poem, then we could debug that.  As it is, it seems to be quite "random". 

I also drew on some quotes from Waugh on "The poetic function in the theory of Roman Jakobson" (...thanks for making these available by the way!) and I'm specifically interested by the ideas of "a hierarchy of signs ... of ascending complexity, but also one of ascending freedom or creativity," and the idea that a "poem provides its own `universe of discourse.'"

My thought is that these criteria pull in opposite directions: towards complexity, and towards coherence, respectively.  I wonder if these are somewhat encoded in the metafunctions κ and ς?

Update: Agamben, The End of the Poem: Studies in Poetics, p. 79.


Update 2:

"trobar," means "to find" or "to meet" in Catalan; from the Old Provençal trobar, from Vulgar Latin *tropāre, present active infinitive of *tropō, from Latin tropus.

Update 3:
Glitch: from the German word glitschen ("to slip").  20th Century English (US): "Glitches—a spaceman's word for irritating disturbances."
This reminds me of the Serrean idea of deviation:
We do not start out with a relation that is then disturbed or even interrupted; rather, "[t]he deviation is part of the thing itself, and perhaps it even produces the thing."  (Siegert & Winthrop-Young 2007: 32-33)
 Thinking in topological topographical terms, this reminds me of the idea of the "declivity" (downward slopingness) or more fundamentally the "curvature" of a surface:


Seamless topographical map pattern from http://blog.spoongraphics.co.uk/tutorials/how-to-create-a-seamless-topographic-map-pattern

Comments

  1. Poetry in motion --

    From Wikipedia: Torcidas organizadas are formal (or informal) associations of football fans in Brazil in the same vein as Americas Barra brava and European ultras. The name is based on the verb torcer, which means "to root for" but also "to wring" and "to turn". -- https://en.wikipedia.org/wiki/Torcida_organizada

    ReplyDelete

Post a Comment

Popular posts from this blog

Metacommunicative cues

In the previous post on Extra channels I finished with a distinction between diachronic and synchronic metacommunication. In this post I'd like to respond to some comments by the co-author of this blog, Joe, in some of his previous posts, by invoking Jurgen Ruesch's concept of metacommunication . Gregory Bateson was interested in thinking about cybernetics, but didn't seem to feel constrained to think about it using a strictly computational or information-theoretic paradigm, while still being informed by the ideas. This gave him the freedom to talk about ideas like "context", "relationship", "learning", and "communication" without needing to define them in precise computational terms. Nevertheless, he handles the ideas fairly rigorously. (Joe, Phatic Workshop: towards a μ-calculus ) Gregory Bateson and Jurgen Ruesch, among many other notable thinkers, were part of the Palo Alto Group of researchers tasked to apply new methods (a

Extra channels

In the following, I would like to clarify the connection between channel and context and concomitantly the difference between metachannel and parachannel . Paul Kockelman urges us "to notice the fundamental similarity between codes and channels" (2011: 725) but instead of that purported fundamental similarity points out the contrast between them. I argue that context , or objects and states of affairs (Bühler 2011[1934]: 35), demonstrate a closer relationship to channel than to code. This is largely because the first three fundamental relations, sender or subject , context or object , and receiver or addressee , belong to Bühler's original organon model while code , contact and message , which were previously implicit in the organon model, are made explicit as additions to the model by Jakobson (1985[1976c]). Thus the most productive approach would be to pair a component from the original organon model with an additional component in the language functions model.