Computational methods in gender linguistic research: Distributional semantics and discriminative learning
As part of the research forum, Dr. Dominic Schmitz (Heinrich Heine University Düsseldorf) will give a guest lecture on Monday, December 1, 2025, at 4 p.m (room ES 24, Johannisstraße 12-20). Anyone interested is cordially invited. The talk will be given in English.
Computational methods offer, alongside corpus-based and experimental approaches, an additional quantitative perspective for empirical research in gender linguistics. Accordingly, they have increasingly been employed in recent years to systematically investigate linguistic structures, meanings, and biases (e.g. Schmitz, 2024; Sökefeld & Amaral, 2025). This talk introduces two central domains within this methodological area: distributional semantics, which models word meaning using high-dimensional vectors (Boleda, 2020), and the discriminative lexicon model, which conceptualises language processing as a learning process linking form and meaning (Baayen et al., 2019). The two computational implementations of the discriminative approach – naive discriminative learning and linear discriminative learning (Baayen et al., 2011, 2019) – are discussed alongside selected techniques from distributional semantics. Drawing on concrete studies, the talk illustrates how these approaches can be used to model and theoretically interpret questions in gender linguistics on a data-driven basis
