The students in my Science Fiction Film course (English 232: Film in Focus) are screening the last film of the spring semester this weekend. This will make 16 in all, and they range from Fritz Lang’s 1927 classic Metropolis and 2001: A Space Odyssey from 1968 to contemporary cutting edge films like Bong Joon-Ho’s Okja from 2017 and Janelle Monáe’s 2018 masterpiece Dirty Computer. This group of students has been such a delight to work with as they’ve engaged and explored each new film with curiosity and rigor, bringing insights to the regular critical responses they write as well as class discussions. Now, as the course nears its conclusion, I’m doing a road test with a new assignment I just devised inspired by some of the reading and research I did during my sabbatical last semester.
This new assignment aims to challenge students to exercise their capacities for poetic imagination and algorithmic coding. Right now we inhabit a time and space where algorithms run recommender systems that suggest the next things people might like to stream or buy based on data extrapolated from what they’ve watched and how they’ve rated these films and series. Relatedly, Netflix and other services are using these algorithms to make plans for the next projects to invest in producing to stream. The codes that do this work are written and refined by people, and their effectiveness is measured by how well they work for people. It’s an exciting zone of human beings and machines collaborating—in other words, it’s a perfect place for students at a liberal arts college taking a class on science fiction film to explore.
Specifically, each student is assigned to generate 30-40 keywords for one of the films we’ve studied. The keywords might be obvious things—like the names of actors, directors, or composers as well as the settings and plot points such as alien contact or time travel—or latent things, like using a match cut at a pivotal moment (as in the most famous match cut in all cinema when the first section of 2001 jumps into the far future). They need to think of the keywords as things that would be useful to enter into a film database for the recommender systems to rake and build combinations of affinities across films.
As a second part of the assignment, each student is also assigned to compose 5 original haikus about the same film. Some haikus must be about the film as a whole, some about characters, and some about one of the key techno-scientific developments that drives the narrative. Essentially, this part of the assignment is a form of analysis, just as the coding portion is. Students act as critics by breaking the film into parts and then thinking through those parts and how they fit together to interpret the film and to communicate about the film to other people. What’s more, the haiku is a sort of poem algorithm with a formal code to follow, though the output is expected to have different impacts than a recommender system. These poetic algorithms function through compression, association, line breaks, sonic pyrotechnics, rhythm, pace, and activation of various senses.
I’m eager to see what the students produce and especially what we can deduce together after a class period of performing their haikus aloud and sharing their keyword datasets. If you’ve got a favorite science fiction film and you’d like to write some haikus and/or parse it into algorithm-ready keywords, I’d love to invite you to share them here in the comments.