Variation in valency patterns across Romani dialects is primarily shaped by contact languages

Abstract This paper examines the effects of language contact on valency patterns by comparing Romani dialects with their contact languages. Due to their wide dispersion and extensive contact with diverse European languages, Romani dialects provide an excellent testing ground for exploring the interplay of genealogical and areal factors in valency encoding. Using data from the Romani morpho-syntax database (RMS) and BivalTyp, a typological database of bivalent verbs, we analyzed valency patterns in 43 predicates across 119 Romani varieties and 18 contact languages. Despite their relatively recent divergence (600–700 years), Romani dialects exhibit greater variation in valency patterns than some genealogical groups with a 2000-year history. These patterns align more closely with current geographic distribution and contact languages than with traditional genealogical classifications. The findings suggest that language contact is the primary driver of rapid changes in Romani valency systems, as some dialects resemble their contact languages more than other Romani varieties.