Despite the magnitude of its consumption by the general public, online pornography has to date received comparatively very little attention within linguistics. The paper explores the language in video titles on the pornographic site Pornhub using a corpus-driven approach.
A corpus of 17,000 titles compiled specifically for the purposes of the study yielded a word list, which, together with an interdisciplinary literature review, was used to formulate two research questions. With gender as a key variable, the first research question focuses on the linguistic representation of agency by looking at verb objects, while the second examines expressions (nouns and adjectives) used to ascribe roles to the participants.
A quantitative evaluation of the data indicates an interplay between gender and the observed features in both questions. The lexicalisation of agency was found to be more complex than previously thought; nine categories were found, and the role of passive voice differed depending on gender.
As for gender characterisation, descriptions of women's roles are not only more frequent but also more heterogenous and evaluative compared to those of men.