Comprehension of reversible sentences in Turkish agrammatism

Abstract

The present study tests the hypothesis proposed to explain the agrammatic comprehension deficit in terms of the loss of traces in syntax, namely the Trace-Deletion Hypothesis (TDH; Grodzinsky 1995, 2000a, 2000b), against data on Turkish agrammatism.A picture-matching task was designed to assess the comprehension abilities of seven agrammatic and seven non-pathological subjects. Test stimuli consisted of canonical SOV, OSV, truncated passive, objectless active sentences and sentences with object relative and subject relative clauses. The second path of this study explores various syntactic accounts available in the literature on Turkish linguistics concerning the types of sentences employed in the picture-matching task and discusses the predictions of the TDH given a specific syntactic account.A One-Sample Test and a Pairs Samples Test were employed to determine statistical significance. The results showed that the OSV type was comprehended at-chance, while all the other sentence types were comprehended better than chance, but not perfectly. When the results are interpreted in terms of the TDH, it is observed that the base-generation analyses for passives and relative clauses in Turkish are granted support, whereas the only sentence type which is comprehended at-chance, namely the OSV, must be formed via movement. Finally, it is argued the TDH has important theoretical and empirical shortcomings which sheds doubt on the validity of an analysis assuming it and that an approach more sensitive to the role of morphological elements in comprehension would give a more satisfactory account of the Turkish facts.