Screenshot: "STATE OF nature," a digital poem, uses agnet-based modeling and natural language processing to tell the story of the eve or twiligh tof a civilization. According to the non-deterministic algorithms, people have children, steal, kill, create and use simple tools, give gifts, and invent religions. Virus-like, these beliefs mutate over time as the civilization grows and language spreads.

This summer, Kyle Booten, Ph.D. candidate in Education with a designated emphasis in New Media, explored the fundamentals of Python programming through digital poetry with his undergraduate students. Meeting for six short weeks at the Berkeley Center for New Media, “Poetry and Technology: A Digital Verse Lab” students worked together in groups to produce works of digital poetry. Students came from a wide range of majors, spoke various languages, and most had little or no programming experience. Digital poetry, which can include computational elements to derive its structure or its content, is an excellent introduction to programming. “It’s a medium with a high ceiling and a very low floor,” Booten explained. “There’s an elegance to writing these simple programs that make things that are beautiful and weird. With just a few lines of code, you can begin creating poems, but you can also make things that are incredibly technical.”

Though Booten has taken several graduate courses at the School of Information and uses Python in his scholarly workflow, his interest in programming stems from his background in creative writing. “I took my first programming class during the last semester of [my undergraduate education], and it was for a breadth requirement. I did an MFA in poetry, and I’ve always really enjoyed the technical aspects of verse, like practicing writing a really good iambic line,” Booten recounts. At the I School, Booten attended “Information 155: Introduction to High-Level Programming” and “Information 256: Applied Natural Language Processing”. Booten suggested that the same tools that researchers in computational linguistics and the social sciences use to gather a large corpus of text and examine patterns can be used for creative purposes as well. Natural language processing (NLP) tools, which use a variety of automatic methods to reveal the internal structure of language, were a perfect fit for Kyle’s interests in language. “I wanted to know how computers ‘read’ things,” Booten explained. “Once you learn how a tool like Python ‘reads’ poetry, you can also learn how Python can write poetry.” Mimicking Kyle’s own introduction to programming, students received a crash course in Python by working through Natural Language Processing with Python, a book that teaches programming using examples from a popular text analysis library. With the help of IPython notebooks, Booten was able to prepare interactive lessons with annotated code. Students worked through exercises in groups, adapting code to their own projects and poetic forms.

In addition to acquiring technical skills, Booten’s students also touched upon historical artistic movements, such as surrealism, and played with them in digital contexts. During one class session, Booten introduced the surrealist game of exquisite corpse, in which a few participants pass a piece of paper around to assemble a poem or image. Booten asked his students what it might look like if poems were authored by dozens, hundreds, or thousands of authors, and presented David McClure’s “Exquisite Haiku” project as an example. In the lab, students took their own approach to “crowdsourced” authorship, using Python to mine certain Twitter hashtags and generate couplets with part-of-speech taggers.

Touching on N. Katherine Hayle’s essay, “Hyper and Deep Attention”, Booten insists, “Poetry has much to offer in the age of the frantic.” Booten presented Nick Monfort’s “Taroko Gorge,” an infinitely generated digital poem, as an example. Coded in Python and JavaScript, the poem’s code creates trochees and spondees that draw upon the author’s original collection of words. Presentness and contemplation are built into the structure of the poem, which endlessly generates lines at a slow, meditative pace, but offers no way to scroll up and view lines once they pass out of view. Since its creation, several poets have remixed Monfort’s “poetic engine” to create generative poems on a variety of topics. “The internet wants us to be addicted, to scan, to click, to jump from place to place, and forget,” Booten comments. “Poetry is slower; it’s something you meditate upon, contemplate, and internalize. Thinking about poetry is urgent in the context of digital media.”