A study of agrammatism with special reference to Hebrew

Abstract

The aim of the thesis was to test empirically Grodzinsky's account of agrammatism. Grodzinsky's account is based on Chomsky's Government and Binding theory and it claims that the comprehension deficit in agrammatism is due to the deletion of 'trace' present in passive and relative clauses. English and Hebrew speaking patients were tested. The experiment exploited a special feature of the Hebrew language in which it is possible to construct passive sentences without trace. In addition to passive and obJect relative clause sentences, other sentence types were also used. The results did not support the trace deletion hypothesis of Grodzinsky. An alternative version of his hypothesis, according to which sentences that require coindexation between two elements in the sentence are difficult for agrommatic aphasics did obtain support. The results also suggested that reversible sentences are particularly difficult for agrammatic patients. Grodzinsky's account also claimed that in agrammatism governed prepositions are impaired and ungoverned prepositions are preserved. In order to test this part of the theory an indepth case study of a Hebrew speaking agrammatic patient who never used prepositions in her spontaneous speech was carried out. The study tested the hypotheses of Grodzinsky and Friederici and it concluded that Grodzinsky's hypothesis according to which governed prepositions are impaired and ungoverned prepositions are preserved is not supported by the evidence. Meaningful prepositions as Friederici suggested, were more likely to be produced in certain tasks. Although this also cannot explain the total omission of prepositions of this patient. In addition to the preposition case study, the patient's ability to deal with the Hebrew verb system was investigated. Both the preposition and the verb study suggested that in agrammatism it is not the principles of Universal Grammar that are violated but the particular features of individual languages.