Valency (mis)matches in Baltic languages: contact, prefixation, and genealogy

Abstract

Abstract This study investigates valency patterns in Latvian and Lithuanian using parallel film subtitles. The data reveals about 2.5 valency matches for every mismatch, with most matches stemming from parallel monovalent and transitive constructions. Systematic mismatches are attributed to five principal factors: (i) differences in case inventories, (ii) idiosyncratic uses of shared cases, (iii) divergent transitive verb inventories, (iv) the role of parallel prefixation, and (v) correspondences with major contact languages. I test three hypotheses, using quantitative methods. First, cognate verbs do not yield significantly more valency matches than non-cognates. Second, parallel prefixation robustly increases matching likelihood, regardless of prefix cognacy or function. These findings suggest the genealogical effect on valency rests on shared strategies of role marking and word formation rather than direct root inheritance. Third, when Baltic mismatches occur, Lithuanian more often aligns with major contact languages than Latvian does. This can be explained by the long-term command of the titular language by Lithuanian minorities, which is not the case for Latvian minorities. The study underscores the importance of a usage-based approach to the examination of valency distributions and points to methodological directions for future work.