Computational poetics
Verse forms, poetic language, social production of poetry, authorship attribution.
One of my main areas of work is the analysis of general patterns in verse language, prosody, and structure. I try to figure out ways to access rich formal information, such as rhythm regularity (Šeļa and Gronas 2022), and show the ways it changes over time or shapes poetic communities. In an upcoming paper, written with Mikhail Gronas, we show how simple measures of regularity and complexity allow us to map modern areas of differentiated style production (in Bourdieu’s fashion).
A closely related field is authorship attribution of poetic texts, where differences in formal organization (rhythm, rhymes, etc.) of verse can be used to trace individual habits of poets. Together with Petr Plechác, we worked on a hilarious case of poetic mystification: one contemporary scholar forged a whole corpus of texts of a less-known Russian poet of the 19th century. Using verse features in an author verification scenario, we demonstrate conclusively (in contrast with previous work) that the origins of these texts are highly dubious (Plecháč and Šeļa 2021) .
