by Alexander Mika, ’21 (CLAS)
Kyle Booten was recently hired as an assistant professor in the Department of English in 2020. Booten is a computational poet with research interests that include literacy and media, computer generated texts, and computer mediated texts. In this interview, we discuss Booten’s post-doctoral work at the Neukom Institute at Dartmouth College, the expanding field of Digital Humanities, and his work at UConn.
Creating New Tools for Writers
At the Neukom Institute, an interdisciplinary computer science center, Booten began to research the ways in which digital and print literacy interact, such as the circulation of literary quotations from print texts on social networks. This work then informed a series of projects called “digital progymnasmata.” These progymnasmata (ancient Greek rhetorical exercises) have allowed Booten to develop as a writer and to create in new and exciting ways. Booten explains: “The progymnasmata push me to use more rare or interesting syntax and to move beyond the kind of affects that I might typically strike while writing a poem.”
By using algorithmic media and natural language processing, as well as these progymnasmata, Booten is researching and designing a program that will assist poets in creating more challenging, original works. “Natural language processing can tell us, for instance, what kinds of ways of using language are rarer than others,” Booten explains. “It’s harder to say, ‘Computer, is this a good poem,’ but you can ask, ‘Is this a rare poem? Does it have some kind of linguistic rarity?’” By analyzing the linguistic rarity, Booten hopes that poets will be able to learn more about their own writing practices and habits and be able to examine their work in a new light. Each copy of the poetry assistant software will be made to fit the specific needs of a particular poet. The project is currently in the data collection stages.
One of Booten’s primary goals with this project is to ensure access to such technologies and tools. He remarks, “I can make my own poetry machine because I know how to do it. I’m increasingly interested in questions of access and allowing people who don’t have the time to learn how to code themselves to collaborate and integrate algorithmic logic into the ways that they’re working. I think that’s really important because it also allows more and more people to critique that logic. It’s important to include people from all sorts of different artistic and academic backgrounds.”
Redesigning DH Courses at UConn
Currently, Booten is teaching two courses that he updated and revised: Introduction to Digital Humanities and Writing with Algorithms. With the introductory course, he seeks to expose students to many of the core concepts, or “the bread and butter,” as he puts it, while also using topics with which most students are familiar, like Instagram poetry, as a medium for examining these issues with recognizable language and images. “Computational text processing, I think, is very new to almost every student in the course,” Booten says. “On the other hand, certain other things like literature and social media are spaces where students are also going to have a lot of experience that they can bring into the course. And it’s all about bouncing these different perspectives off of each other. I like grounding new concepts and familiar forms.”
Writing with Algorithms introduces creative writers to coding with Python, a language that allows one to manipulate and generate text. Booten finds that the bridging between writing and programming in this course, the creative and the analytical, has a range of benefits for students: “I think it’s a really important skill for humanities majors to have. A lot of our literary media is always being integrated into algorithmic media. And I actually think that doing something creative with coding is a really good way to get people excited enough about coding to push past some of the inevitable frustrations, because it is frustrating at the beginning.”
Questions of Epistemology and Ethics
With the release of incredibly powerful text generation systems such as GPT-3 (Generative Pre-trained Transformer 3), a system that uses AI to produce text, a number of epistemological and ethical questions have been raised. “You programmed how it will learn,” Booten explains, “but you actually don’t know very much about what it has learned through its training process. And I feel like one of the things that digital humanities Is starting to do and will continue to do is to reckon with that epistemological challenge: how do you think about texts that are so radically inscrutable?”
Since these programs are created by people, biases find their way into the coding, which can have serious implications in everyday life. The advancements in the field prompt important discussions of “algorithmic justice” and questions such as “How do you de-bias algorithms?” “So far as algorithms like this play a huge role in determining all sorts of practical matters in our lives,” Booten notes, “and plenty of scholars have shown in code all of these biases about the world, it becomes that much more important to figure out what it would mean to teach one of these algorithms.”
As Booten works to expand the study of digital humanities at tUConn, one of his keywords is accessibility: his goal is to foster a community in which people share a common literacy and work to develop these skills outside the classroom, engaging in discussions and implementing the coursework into their extracurricular creative projects. He also sees unique possibilities UConn presents for interdisciplinary collaboration. “There is this great constellation of other people across campus in all sorts of different places that are thinking about really similar questions,” Booten remarks. “I’m new to the university but I’m really excited about the English department and the university as a whole; it’s a great place to be doing DH work.”