The paper discusses patterns of person agreement in Mehweb Dargwa. The focus of the paper is constructions with dative subjects where person agreement can be controlled by neither the dative subject or absolutive direct object. This constitutes a violation of Bobaljik’s conjecture about the role of morphological case in agreement. The paper shows that person agreement in dative subject constructions is possible only under condition that both the dative and absolutive NPs are first person arguments.
The constructions čto nazyvaetsja ‘what is called’ and to, čto nazyvaetsja ‘that what is called’ share properties with both parentheticals and relative clauses, especially transparent relatives. The construction čto nazyvaetsja ‘what is called’ is extremely close to parentheticals and highly lexicalized. The construction to, čto nazyvaetsja ‘that what is called’ patterns in some respects with relative clauses, e. g. it is more frequent if the designation is a NP and it is avoided in postposition to the designation. During the 18th–19th centuries the construction čto nazyvaetsja ‘what is called’ acquired some common features with parenthetical clauses. For instance, the lexical constraints became stricter, and non-NP designations became more frequent. The data allows to suggest that the observed constructions constitute intermediate stages on the cline of change from relative clauses to parentheticals.