To Program a Prose Machine


Nanni Balestrini, Tristano, copy #10750 (Verso, 2014).

In order to program a poetry machine, one would first have to repeat the entire Universe from the beginning—or at least a good piece of it.

— Stanislaw Lem [1]

"All directions are of equal importance." This is the second sentence in the second paragraph on page 88 of my copy of Nanni Balestrini's 1966 novel Tristano, #10750. You cannot read this novel, unless I lend it to you, as each of the 10,000 copies Verso publish this month contain different iterations of the same text.

When composing Tristano, Balestrini used a computer algorithm to shuffle the sentences of the ten paragraphs which comprise each of the ten chapters. The exact methodology is not clear, but it was likely similar to the process he used for an earlier computer-manipulated text, Tape Mark 1 (1961). For that work, snippets of Lao Tzu, Michihito Hachiya, and Paul Goldwin were divided into fifteen short phrases and then remixed combinatorially by an IBM 7070 and a program comprising 322 punched cards to create short texts, each a unique sequence of ten elements.

Nanni Balestrini, flow chart for Tape Mark I (1961). Published in Jasia Reichardt, ed. Cybernetic Serendipity: The computer and the arts (Studio International, 1968).

In the case of Tristano, his process allowed for 109,027,350,432,000 different possible variations. A single variation of Tristano was published in Italian in 1966, its text "a mixture of original prose and text borrowed from guidebooks, atlases, newspapers and other artistically marginal sources," according to the publisher. For the text to make any sense at all post-scramble, it had to be, as Balestrini’s neoavanguardia comrade Umberto Eco writes in his forward, "'prepared,' like pieces of lego, each already designed to fit together with other pieces in multiple ways." As a result, there are a limited number of set-pieces, and all proper names are "C," leading to such delirious formulations as, "In the month of July C went on a trip up the river on the ship as far as C and on his return he decided to abandon C." This sentence also serves as a neat summing up, as near as I can tell, of Tristano's "plot," which is ostensibly based on the Arthurian legend of Tristan and Isode.

The "plot" isn't really the point, but the most readable and pleasurable outcome of Balestrini's game, at least for #10750, are its accidental mini-narratives. Some examples:  

This being done we hoisted jib and mainsail kept full and we start boldly out to sea. Twenty minutes later we climbed on board. Vomiting over the side leaning on the ropes. Dark blue of the panorama. Ten seconds.

A long thin rivulet of water slowly advances on the asphalt. She moves slowly under his body. The woman answered no certainly not.

Languidly undulating surfaces lack of watercourses the frequent outcrops of rocks that emerge from the fine layer of red earth which nonetheless supports rich crops. The scar on her stomach was visible in the faint dusk light. I'm so happy you came. Let's try another position.

Each mini-narrative dissolves back into the overarching form, usually quite quickly, although #10750 features one which lasts as long as a single paragraph save one sentence. The frustration this generates highlights formless chaos as the pre-history of every narrative, as well as the unholy formal dance of writer and reader required to create meaning.

Other moments of coherence take the form of meta-commentary interspersed throughout, such as:

You could even start from another episode and obtain a slightly different story. Though the question is rather irrelevant.

The provisional nature of the assemblage of the materials from different sources not connected together by integration but by association.

The other possible interpretations are endless but at the moment this is the only reality that belongs to us.

Tristano is still, at least nominally, a novel, one where the voice and temporality can change not only every line but within every line. Thus, ascribing these gnomic pronouncements to some "authorial voice," or taking them to be "statements of intent" or "ironic self-criticism," would be inadvisable. It is tempting to compare Tristano to hypertext fiction, which seems to be undergoing something of a resurgence with Twine, an open-source tool for telling interactive, non-linear stories. And both do indeed seem to be interested in extracting and making visible the "rules" which govern modern and post-modern lit, breaking narrative down into its consituent elements. However, hypertext fictions places great value on "exploring" the possible sequences of these elements, while each iteration of Tristano is fixed, concrete. The computer has already explored; we merely have the path.

Instead, Balestrini's experiment focuses on attacking the twin myths of the creative genius and culture as property. As Eco writes, "The creative man will not, then, be he who has deduced something new ex nihilo, but he who has identified it, by intuition, by trial and error, by chance…" Balestrini's algorithm might limit the scope, but this work of creative identification is still shared between the author and his readers.

Tweet from Brian Droitcour, November 6, 2013.



[1] Stainislaw Lem’s short story "The First Sally (A), Or Trurl's Electronic Bard" (collected in The Cyberiad, 1965).