Because most research on language processing in aphasia has been carried out in English, there is a serious confound in our theories, between universal processing mechanisms and language-specific content. Current explanations of patterns of selective dissociation in aphasia (e.g. grammar vs. semantics, algorithmic vs. heuristic processes) may reflect particular facts about English which do not generalize to normal or asphasic speakers of typologically-distinct languages. After presenting pilot data suggesting that there are important differences among aphasic patients from different language groups, we propose three years of cross-linguistic research on sentence processing in aphasia. The proposal is based on 10 years of cross-linguistic research by the co-investigators, with normal adults and children, and on an ongoing program of research with aphasics at the Boston V. A. Besides testing existing theories of dissociation in aphasia, the research tests the """"""""competition model"""""""", a theory developed to account for cross-linguistic differences in sentence processing among normals. In the monolingual core of the project, we will test patients in four language groups: English (a strong word order language), Italian (weak work order), Hungarian (strong case inflections) and Serbo-Croatian (weak case inflections). Patients from four diagnostic categories will be tested: agrammatics (Broca's), paragrammatics (Wernicke's), fluent anomics, and patient controls without neurological damage, matched for age and educational level. Several aspects of sentence processing will be investigated: free speech (in a structured interview), sentence comprehension (including aspects of word order, grammatical morphology, semantic constraints, and prosody), and sentence production (in response to pictures and films designed to tap different semantic and pragmatic constraints on lexicalization and ellipsis, word order and morphology, pronominalization, definite and indefinite reference). We will also develop a single-word naming and comprehension test with items familiar to all the cultures involved in this study. The results of the proposed research should contribute to international standardization of assessment techniques. In addition, some innovations in data analysis (structural equation modelling) will be applied to aphasia research here for the first time.
