Ellen and Masha in Brain and Language

Is processing subject-relative clauses easier even in an ergative language?

Published at long last, "A subject relative clause preference in a split-ergative language: ERP evidence from Georgian," from a team led by Ellen Lau and Masha Polinsky, with former RAs Nancy Clarke and Michaela Socolof, plus Rusudan Asatiani from Tblisi State University. The paper reports a series of ERP and self-paced reading studies done with speakers of Georgian, presented with relative clauses which are initially ambiguous as to whether they relativize the subject or the object. The studies found "a large anterior negativity and slower reading times when a relative clause is disambiguated to an object relative vs a subject relative," consistent with a robust generalization across languages. The interest here is that Georgian is an ergative language, and recent studies of Basque, also an ergative language have motivated an alternative to the classic generalization, that the preference is not for subject gaps, but for gaps in morphologically unmarked positions, which would in an ergative language includes the direct object of a transitive clause. The Georgian data therefore suggest that this revision is not correct for every ergative language.