Wh-questions and relative clauses in Greek agrammatism: Evidence from comprehension and production

摘要

Background: Cross-linguistic studies on the production and comprehension of wh-questions and relative clauses, have revealed selective deficits in agrammatism with better performance observed in (a) subject questions and relative clauses compared to object ones; (b) object non-referential who questions compared to object referential questions; and (c) object what questions compared to object who questions. These selective deficits have been discussed within several neurolinguistic accounts (i.e., D-Linking Hypothesis, Derived Order Problem Hypothesis, and Relativised Minimality), which make different predictions. Limited research on wh-questions and relative clauses has been conducted in Greek agrammatism with inconclusive results, leaving open the question of which linguistic factors affect the agrammatic performance and which account best explains the attested patterns.Aims: The aim of the present study is twofold: (a) to investigate the role of three linguistic factors (syntactic function, referentiality, and φ-features) in the production and comprehension of wh-questions and relative clauses in six agrammatic Greek-speaking participants; and (b) to examine which neurolinguistic account can explain the observed patterns.Methods and Procedures: Two elicitation tasks (one for wh-questions and one for relative clauses) and two picture-pointing tasks (one for wh-questions and one for relative clauses) were used. All tasks targeted the following structures: referential and non-referential (who/what) questions, half with subject and half with object extraction, and relative clauses, half with subject and half with object dependencies.Outcomes and Results: Referentiality had a greater impact οn the production and comprehension of wh-questions compared to the other two linguistic factors under investigation, i.e., syntactic function and φ-feature assignment. Syntactic function did not affect the production or comprehension of relative clauses given that no subject/object dissociations were attested.Conclusion: With respect to wh-questions, the patterns observed in our data can be accounted for within the D-Linking Ηypothesis. The lack of subject/object dissociations attested in relative clauses is attributed to the case mismatch between the head of the relative clause and the relativisation site, and to the high proportion of complementiser omission observed in these structures.