Stefan Schnell received a 4-year SNFS grant for the research project "Reference production and the evolution of person agreement: Corpus-based typological and computational approaches”
In this project we deploy quantitative corpus-linguistic methods to investigate the alternation between referential null forms and overt person forms, and its interaction with the evolution of person agreement across diverse languages. Our investigations target a vexed puzzle in comparative linguistics, namely how the cross-linguistically ubiquitous possibility of null reference (“pro-drop”) can be reconciled with the likewise ubiquitous tendency towards the evolution of person agreement, which presupposes an increased use of overt person forms? Similarly, current models of referential choice leave open a range of questions concerning cross-linguistic variation therein. Our project is the first of its kind to tackle these quandaries through extensive cross-linguistic corpus analyses that implement quantitative methodologies from corpus linguistics, natural language processing as well as phylogenetic comparative methods. In doing so, our project addresses hitherto neglected issues in usage-based approaches to the typology of reference production and grammatical systems, which have only rarely been addressed in first-hand empirical studies of actual language use. It will also contribute to explaining the development of grammatical complexity, and more generally to the areas of language processing and cognitive sciences, as well as human communication and interaction, and their socio-cultural underpinnings.
