image / text index

What is an Index Machine?

book machine: index signatures
Jan van Eyck Academie
Ongoing research and experiment

Monday 12 January 2009


Hi. My name is Jack Fisher. I'm a researcher in the design department. What follows in this presentation are fragmentary lecture notes in the form of index cards. I have put together a series of these as a way to begin organizing my thinking around this project. The order of the cards is not definitive. There may be a lack of coherence in what follows but the idea is that I am beginning to enact some of the research procedures my project will develop... so first an abstract of the project:
I will construct an iterating machine which assembles through transcription the discourse of the Jan van Eyck Academie and submits it to indexical experiment in order to produce a collective ennunciation. I should say that I've been reading Deleuze & Guattari's Kafka: Toward a Minor Literature and have found it productive to transpose the vocabulary they develop in explicating Kafka's literary machine to the machinic procedures I am developing in this project.
what is a collective enuniciation? I am defining it for now simply as that enunciation that is processed through a machinic index. I am not entirely sure yet what this means, but the sense of the routes through which a collective can speak is critical to the interest of this project. And the sense in which the index can approach the conditions of a minor literature.
I am building an abstract machine not an actual machine. An abstract machine consists simply of a definition in terms of input, output, and the set of allowable operations used to turn the former into the latter.
This is a first sketch or diagram of the machine's structure. The transcriptions I'll make will form the input. The index is the series of transformations performed upon the transcription. And the output is...well it could be many things. We'll get to this later.
A lot of my recent work has been with transcriptions, in documenting conferences for the American Institute of Architects. My interest in these conference texts has had to do with their position below the threshold of literature of the book. These are texts which have not been submitted to the orders of writing, and as such are available for transformation/reconstitution/rewriting.
I'm going to read a really remarkable passage from Walter Benjamin's One Way Street to begin to indicate my current interest in transcription. The power of a country road when one is walking along it is different from the power it has when one is flying over it by airplane. In the same way, the power of a text when it is read is different from the power it has when it is copied out. The airplane passenger sees only how the road pushes through the landscape, how it unfolds according to the
same laws as the terrain surrounding it. Only he who walks the road on foot learns the power it commands, and of how, from the very scenery that for the flier is only the unfurled plain, it calls forth distances, belvederes, clearings, prospects at each of its turns like a commander deploying soldiers at a front.
Only the copied text thus commands the soul of him who is occupied with it, whereas the mere reader never discovers the new aspects of his inner self that are opened by the text, that road cut through the interior jungle forever closing behind it: because the reader follows the movement of his mind in the free flight of daydreaming, whereas the copier submits it to command.

I am interested in this word command.
My project seeks to develop a mode of text which is not governed by the synoptic orders of a rationalist visualization, but opens itself to commands of contingency and coincidence. If the legibility of the text we read, as if from above, is constructed in the violence of a monologic hauteur, what might it look like reading from the ground?
The Benjamin passage, with its linkage of transcription to submission, gets at what I am calling the ethical interest of my project—the sense in which transcription is the condition for ascetic or ethical formation. I am the scribal agent, the copyist, but also the machine operator.
Along these line, I have become interested lately in what Foucault writes about the hypomnemata as an ascetic technology of the self. He writes, in the last book of History of Sexuality, No technique can acquired without an exercise; neither can one learn the art of living, the tekne tou biou, without an ascesis which must be taken as a training of oneself by oneself... Among all the forms this training took (which included abstinences, memorizations, examinations of conscience, meditations, silence, and listening to others), it seems that writing came quite late to play a sizable role. In the technical sense, the hypomnemata could be account books public registers, individual notebooks serving as memoranda. Into them one entered quotations, fragments of works, examples, and actions to which one had been witness or of which one had read the account...They constituted a material memory of things read, heard, or thought...They also formed the raw material for the writing of more systematic treatises in which were given arguments and means by which to struggle against some defect or to overcome some difficult circumstance. It should be noted that these hypomnemata as personal notebooks were not chronicles of an interior life but rather were comprised of lists, observations, reading notes, etc.
I'd like to indicate a sense of transcription as a condition for disciplinary formation, that is, for the formation of disciplinary knowledge. there are many examples of this where a text which is foundational to a particular discipline began as a lecture or as speech and was later transcribed. my argument might be: it is the very act of stabilizing or fixing speech as a text which crystallizes a field of knowledge as a discipline.
Now I'd like to briefly indicate how the transcription might work. This by the way is a major area of research for this project. The trick is to cultivate proliferating and machinic inputs for all the speech and text produced as Jan van Eyck. And to constantly ask: what is being left out, what is eluding capture?
It is critical in these capturing actions to preserve or to mark the incidental, contingent formalizations already present in the speech of texts. I don't want to collapse every input into, say, a 12 point Helvetica text edit document.
I am interested in experimenting with transcription orthographies, in developing, for example, a typographic means for registering that auditory moment in which understanding or contestation happens.
Ok. This is a project, Ghost in the Machine, which addresses some of these things. It is an adducing or decoding of a machine reading. It formalizes lines of escape through the incidental distortions of a technological mediation. I worked on a text by Gilbert Ryle, The Concept of Mind. I took it from a Google book search and submitted it to an optical character recognition procedure to produce an editable text. I tried to infer the program of how it garbled things, and to visualize it in the moment of reading.
And this is something I did quite recently—a more straightforward operation of converting notes I took at a Jacques Ranciere lecture at the University of Chicago last fall. This was to begin to test some notational strategies in very inclement conditions: didn't know his work, he spoke with a very thick french accent...
The Deleuzian machinic index is an inscription of a line of flight across a language of sense. The indexical procedures I develop will deterritorialize an already deterritorialized language —in the sense that the English of JvE is a vehicular or institutional lnaguage already at a remove from a vernacular/maternal usage.
The index then will push its linguistic materials to a further point of deterritorialization and to the intensity of an extreme sobriety. But a highly libidinal one, which multiplies in formal transformations to all possible states.
These experiments with conference transcriptions submit the static body of a 30,000 word discourse to multiplying formal transformations.
The index will machinically read through the assembled transcription for significant vocabulary and rhetorical categories. It could read for the atomized keyword or for grammatical orders: for imperatives and optatives to produce lists of institutional commands and desires. It could also read for embedded theorizations of the machinic, or order to explicate its own operation. It can mark emerging conceptual linkages across the spectra of transcribed discourses. The index will track —commanded by its social habitat— the language of the Academie.
There is also a notion of the index as a storage card used as a research building block. As cards which can be dynamically sorted by different edge notching schemes. This is fact was the material model for some early theorizing (by Douglas Englebart and Theodore Nelson) about how a universal library might be mechanized...i.e. how a hypertext might be constructed. (There is something here that is worth capturing and commenting on —that what begins as a marginal notation in a book, makes its way (one hopes) to a page in a notebook, from which it might eventually travel to a paragraph in a more fully realized exposition. Or not. Or it lies dormant as a note. The point is to make these notes available, to put them into a machinic circulation.)
In addition to breaking the transcription apart and linking keywords, automatic methods might be devised for the production of new texts. This really interests me—recombinatory strategies to generate new texts or vocalizations from the transcription assemblage. For instance: determine a set of shared words from 2 texts; replace every other sentence in which the word appears in the first with a sentence from the second. Could the machine produce, through its indexical mechanism, a new statement, a rigorous conjoining of two texts? 41-48
This is a project I did in which I interpolated 2 texts; the first, in the three columns, is Gyorgy Kepes's Language of Vision; the second, in the larger type running through the first, is from Vannevar Bush's As We Think, an elaboration of a microfilm machine he termed the memex. My interpolation finds each word about the memex already within Kepes's text, and begins to visualize the linkages between the two.
What new form of the book or of publishing is possible which could begin to assemble the proliferations of a machinic index?
So whatever it is, is is rigorously periodic. And in theory it will include all JvE institutional publications.
It produces parts for a future assembly: sets of orders; lists of wishes; decoding appendix; legend; apologia; bibliography.
And it produces, most importantly?, a set of instructions, so that it might be executed by other operators, on other grounds.