With roots in indeterminate and asemic writing in print, the term "generative poetry" is used to refer to any born-digital poetic project that uses code, algorithm, or other indeterminate means to generate poetic texts. In generative poetic works, a program or algorithm generates a poem or series of poems based on a lexicon or set of lines developed by the authors. This generation may run once, for a fixed period or a fixed number of times, or indefinitely, depending on the project. Alison Knowles's and James Tenney's "A House of Dust" is one of the first generative poetic works, using a standardized syntactical structure and a set lexicon of words and phrases as variables to allow the poem to continuously generate its own randomized stanzas. Generative poetry calls into question the role of the author, the division between human and machine learning, and the tension between material and ephemera in the digital poetic text.
This glossary entry was derived from David Jhave Johnston's workshop, "From Dust to Deep Learning: Introduction to Generative Poetry," held in Toronto, 21 October 2017.