The algorithmic literature of Oulipo

The first of the digital annexes that I'm working on is the canonical Cent mille milliards de poèmes (A Hundred Thousand Billion Poems) by Oulipo co-founder, Raymond Queneau. The text is inspired by combinatorics in its basic functioning, but it was almost immediately digitized by Oulipo as well, using some of the first computers.

The way the text operates is simple: There are 10 sonnets written by Queneau, each of which is composed of 14 lines. Queneau restricts this oetic form even further by adding two more constraints: 1) each line of any sonnet rhymes with the corresponding line of any other; 2) each line is a grammatical unit, meaning there is no enjambment, further allowing Queneau to substitute any one verse for another. In this way, the verses become permutable, and through simple mathematics, one can see that this little volume actually contains 10^14 “potential” poems, most of which were not even written by Queneau, but rather by a combinatory machine.

In my project, I am turning this beautiful physical volume into an interactive digital edition, which I hope will rectify a major problem associated with previous online editions. Reader participation is key here-—this text was literally designed to be manipulated–but most online versions allow the reader to push one button and automatically generate an entire random poem. Queneau’s text, however, was more subtle than that. It was even anti-random, you might say: In the original, a reader could turn one line at a time, or create and dissolve one stanza at a time. That process is what I’m trying to recreate.

A few more thoughts:

1. Learning Python is a challenge, but given my background, a thoroughly enjoyable one. With my knowledge of mathematics and penchant for language-learning, I love practicing such a practical, tool-based language that really makes you rethink the notion of language in general.
2. Computer programming to create digital editions of texts is the perfect combination of my mathematical and literary backgrounds, and I really enjoy how “neat” Python looks with all the indentations. Very logical. Learning how to use Python to create this edition is also helping me understand the text better-—making choices about how the reader should play with the digital edition adds to my analysis of how a reader is called upon to interact with the physical text as well.
3. Programming is tedious, and one slight error can throw everything off. It takes a long time to code things properly, and a great attention to detail. Half of the work is conceptualizing, and the rest seems like busy-work. It’s very hard to reconcile with the idea of humanistic work, which often means long hours in the library, reading carefully in order to come up with a fully-formed analysis. I believe that reconciling these two opposing ways of working is the most important challenge for the digital humanist.