Ekphrasis—poetry to, for, and about the visual arts—provides rich territory for the exploration of literary tradition and genre definition. This talk will consider the challenges and opportunities presented when using topic modeling to consider issues of genre and discourse in ekphrastic poetry. Oriented toward the non-expert, conversation will assume no prior knowledge of topic modeling or social network analysis, but instead will provide a gentle introduction that builds toward an understanding of the potential uses for topic modeling and network analysis to explore large collections of poetic texts. Poetic collections, dense and rich with figurative language, require revising how we interpret topic modeling results, and this presentation will address how changes in interpretation affect the questions we might ask and the assumptions we can make about “topics” generated by latent Dirichlet allocation (LDA)—one type of topic modeling algorithm.
Furthermore, this talk will demonstrate how the story of ekphrasis, a genre of poetry in which contributions by women are often seen as “in addition to” or “outside” of it, changes when we break open the binding of the canon and include poems by women as central to its tradition. The computational strengths afforded by LDA help reorganize critical thinking about ekphrasis as a complex distribution of discourses, and this talk will demonstrate a few possible ways to refocus the aperture of the critical lens to ask enduring humanities questions about visual-verbal relationships from new perspectives.
Lisa Rhody is a doctoral candidate in the Department of English at the University of Maryland, College Park. Her project “Review, Revise, Requery: New Methods for Studying Ekphrasis” was supported by a MITH Winnemore Dissertation Fellowship in Spring 2012. Find her on the web at lisa.therhodys.net or on Twitter @lmrhody.