[-empyre-] visualization as the new language of theory
Thomas LaMarre, Prof.
thomas.lamarre at mcgill.ca
Sat Feb 6 01:18:03 EST 2010
Hi Greg and Simon,
I think this is a really interesting conversation. I really agree with Simon's point about 'text' (and it is often amusing to note how video game theory for instance will cite Barthes on text, mistaking this notion of text for text as data). At the same time, with the proliferation of new techniques for 'data-mining' (a term that makes the techniques sound cruder than they are) I think it important for us to try to work with them as well.
When I previously suggested seeing animation as a mode of experimentation, it was partly to address this sort of impasse. It seems to me that experimentation - as reconceptualized in Stengers via Whitehead and in Latour to some extent - presents us with a process that is between data-mining and meaning-production (or text-generation).
Part of the allure of some of these new techniques of data mining for certain humanists is their aura of scientific accuracy. But humanists tend to have very rigid and narrow notions of what the sciences are and how they operate. On the one hand, we can address this sort of problem with Ian Hacking's brilliant essay 'The Disunities of the Sciences' ( HACKING, I. (1996) "The Disunities of the Sciences," in The Disunity of Science: Boundaries, Contexts, and Power, ed. P. Galison and D. Stump, Stanford: Stanford University Press, pp. 37-74.) For Hacking shows how recent and inaccurate the notion of an underlying unity to the sciences is. On the other hand, we can keep in keep that, although scientists do not necessarily think like Derrida and might protest about his notion of text, scientists in various fields do understand that data mining is but one side of experimentation. Experimenting also calls into play techniques of reading, and scientists are very conscious of these 'secondary' processes. (I hesitate to call them secondary because in fact the data mining and the reading go hand in hand, in parallel.) Here too Stengers and Latour can be very useful for their careful excavation of how scientists deal with the question of representing their 'nonhumans.'
Maybe such an approach can open a more dialogic encounter between these possibilities? Or maybe I am waxing utopian... All the more so perhaps since I think that animation, having become a fairly ubiquitous mode of experimentation, might present such opportunities.
On 05/02/10 3:08 AM, "Simon Biggs" <s.biggs at eca.ac.uk> wrote:
The last line of your post reveals where we differ. You describe the availability of all (or at least many) of our books (whose books?) as an "information sublime". I could go into a critique here of the sublime, but that isn't where the primary difference in our positions is found. It is your use of the word "information", in this context, that I find problematic.
I understand a text (of which a book is an instance) to exist at the point of its being read. I agree with the Derridean assertion that a text exists as an activity, in the process of interpretation. Yes, information can sit on the pages of a book lost on a library shelf waiting to be found or in a forgotten directory on a hard drive (putting aside the idea that the library, computer or network are also all potential texts) but it is in their reading that such information becomes text. However, not all information is text (although some have argued for and revelled in the literary merits of such indexical systems, myself included).
Data-mining is one thing, reading another. I have made a number of automatic reading systems (systems that make and then read texts). I didn't do this because I thought the algorithm involved was actually reading. It didn't enter my mind for a second that a computer programme might be able to apprehend things. Of course not. The computer programme processed the data according to a moderately complex set of rules (grammatic, semantic, semiotic) and in the process exhibited a behaviour that slightly resembled reading. It was, however, very obviously not reading. It is in this failed mimicry that I find the beauty and motivation to make such work.
The idea of data-mining is to me of the same ilk. It is a quotidian process that might reveal unexpected (and potentially valuable) patterns in data which we might treat as texts we can read and in the process come to apprehend something. But to confuse the reading with the data-mining is, in my view, dangerous. It reduces both the reader and the text to data and information rather than meaning and becoming. Lev's (interesting) piece implicitly did that.
s.biggs at eca.ac.uk
simon at littlepig.org.uk
edinburgh college of art
Creative Interdisciplinary Research into CoLlaborative Environments
Electronic Literature as a Model of Creativity and Innovation in Practice
From: Gregory Ulmer <glue at ufl.edu>
Reply-To: soft_skinned_space <empyre at lists.cofa.unsw.edu.au>
Date: Thu, 4 Feb 2010 13:36:35 -0500
To: soft_skinned_space <empyre at gamera.cofa.unsw.edu.au>, Gregory Ulmer <glue at ufl.edu>
Subject: Re: [-empyre-] visualization as the new language of theory
Simon Biggs wrote:
> Hi Sean
> I am aware of the examples you give - but that was not the sort of
> thing I meant when suggesting some form of data analysis of a text. I
> was thinking more about how you could mash-up discourse analysis,
> corpus linguistics and reader reception theory, on the one hand, and
> empirical linguistics and statistical semantic modelling, on the
> other. I imagine it would be a mess so was entertained by what a
> possible visualisation might resemble (a car crash?). If you look at
> you might see something like this. The point I was seeking to make is
> that it seems ambitious to apply quantitative analytical methods to
> the understanding of something as subjective, fugitive and motile as a
> text (or other cultural artefact).
Ambitious, and necessary... Data analysis is an important aspect of
these threads (although my interest includes text). Are you perhaps
stating precisely the challenge to data/text mining, to design or
develop a means of accessing this (connotative?) level of discourse? I
am thinking for example of the Digging into Data Challenge
( I participated with a group that was not awarded one of the few
grants). "What do you do with a million books?" (or umpteen million,
thinking of the Google digitizing project) ... or a million animations?
The challenge relates to my post in October about the possibility of
inventing an approach to Web Ontologies using poststructuralist
ontologies. Lev's work is relevant here of course. As the (more or
less) entire archive of books becomes available online via full text
search, we will be (are) in a condition of the information sublime. Is
the unity between the animation question and text mining found at the
level of database design or Web Ontologies (if all content is digital)?
thanks everyone for the bibliographies
More information about the empyre