Quality and Competitiveness in Distributed Poetics

On Wednesday February 15th, 2006 at Penn State University, Charles Bernstein (English Department) and Nick Montfort (Department of Computer Science) organized a reading at the Kelly Writers House. This time, it was under the new name of the "Machine Reading Series", and it was fortunate that the two visitors were "machine" poets themselves. It was also fortunate that I was attending, so I could take some notes. They are linked here.

Loss Pequeno Glazier (Department of Media Study, SUNY at Buffalo) and Jim Carpenter (Computer Science Department at Penn State University) were invited to read and talk about their work. Although the crowd was mostly coming from the literary field, the presence of Jim Carpenter oriented the event towards a technical discussion about the possibilities of language engines. In spite of that, the topic was most of the time drifting back to aesthetics problems such as "What makes a poem worth and/or work?", "What can a prosthetic process bring to a human poet?".

Jim Carpenter brought up two interesting notions: quality and competitiveness. This triggered an interesting debate, quite original among literary scholars: what difference can we make between publishing houses' or poetry reviews' standards (= competitiveness) and the poet's work free from these standards (= quality). To me the distinction appears to be quite artificial as it repeats the traditional dissociation between the commodity/reproduction work and the authentic work of art that does nothing else but flattering the poet. Did not the printed review activities in the 20th century sufficiently demonstrate the experimental potential of reproduction and distribution? And the digital medium allows even more possibilities.

But the core of the debate was really Artificial Intelligence: what Aspen Aarseth calls the Cyborg Author was at stake. It was interesting nonetheless as the cyborg-problematic eventually raises the problem of control over text. Here, it was less a matter of control over writing than a problem of control over reading (How do I react to a poem generated by a machine?). The question of control was transferred from authorship to readership. How do you tell what's worth reading? What is quality? What does it mean to produce something that is competitive? How do you control the cultural field that you chose (or not) as your environment? All these questions were implicit, but not too much discussed.

The two guest poets directed their reading towards this problematic of reading and reception:

1) Jim Carpenter implemented a language engine, the Electronic Text Composition (ETC) project. It has a "geodesic structure" that generates words from the British National Corpus. The paratacic dimension of the program eventually leads to output poetry that stands for pastiche to any literary scholar. It summarizes the history of modern poetry via language experiments on disarticulation and pacification from late romanticism to post-modernism. While, through pastiche, the output poems can become competitive (because they correspond to the review literary standards), the notion of quality is now attributed to the writing of the program itself. The computer program becomes poetic in that sense.

2) Loss Pequeno Glazier, director of the Electronic Poetry Center gave a retrospective overview of his work. He focused on his pieces that experiment new ways of reading. The machine here intervenes as a real-time editor of the poem, disturbing the linear reading by changing the text, allowing semantic jumps. The problem of database was tackled with this other perspective on poetry reading: What value has a poem if it is composed of variable fragments?

This debate was interesting for someone with a little knowledge of the potentialities of electronic textualities. The often-uneasy reaction of the audience (part ironic, part sympathetic) was telltale: how do you recognize a robot production from a human one? How do you value a poem according to this distinction? To me it comes down to the question of the caring for literature in the first place: what draws you into a text - which is the most difficult question perhaps.

Charles Bernstein insisted on the necessity for literary scholars not to forget the notions of database and access, two major properties of the New Media object according to Lev Manovich. He focused on the importance of formats and information processing within the literary world, something that is easily forgotten. Someone told me that it was shocking how English and Linguistic students didn't have a clue how to deal with Internet tools - that might be a caricature, but it also might be true that there is a new form of textuality that we literature students are missing. Pointing in that direction, Bernstein suggested that a poem may be analyzed in terms of data processing. What was the relativity of hermeneutics (several readings of the poem) can be reinterpreted today as a multi-functionality of the poem through the software program.

Camille Paloque-B. (Lyon, France)