This half-century is where combinatorial literature and software finally merge. Mainframes could now execute the production rules that print could only suggest, while the Oulipo gave recombination a rigorous theory of constraint.
On the Manchester Ferranti Mark I, Christopher Strachey wrote a program that generated love letters from templates and word lists — arguably the first work of digital literary art.[1] Its slot-and-fill structure is the direct ancestor of context-free grammar generators.
Theo Lutz produced "Stochastische Texte" by recombining vocabulary from Kafka's The Castle under simple logical rules, one of the earliest German experiments in computer-generated poetry.[2]
Founded in 1960, the Ouvroir de littérature potentielle treated constraints as engines of possibility. Raymond Queneau's Cent mille milliards de poèmes (1961) printed ten sonnets on sliced pages so any line could pair with any other, yielding 1014 poems — a physical grammar machine.[3] Queneau and mathematician François Le Lionnais theorized this as "potential literature": the work as a system that produces texts.[4]
"A potential literature... furnishes forms for the good use one can make of literature."
By the 1980s, recombinant fiction like the program Racter (1984) and a wide ecosystem of Markov and template generators made grammar-based text routine.[5] The pieces that define modern tools — vocabularies, rules, weights, and substitution — were all in place.