On the processing of object relative clauses

Abstract

This study investigates the effect of featural mismatch in the processing of object relative clauses (ORCs). A self-paced reading experiment is reported in which the matching of formal and semantic number features of the head noun and the RC subject is manipulated. The predicted full mismatching intervention/interference effect was not obtained but collective nouns gave rise to the shortest reading times in both the subject and the RC verb regions. A follow-up suggests that it is subject-verb agreement in both the RC and the main clause that is vulnerable to semantic interference.