The long-term objective of this research is to contribute to the understanding of the output language processing deficits which underlie the symptoms observed in the speech of aphasic patients (both in the classical syndromes and in the 'mixed' cases that constitute at least a plurality of the clinical population). This goal in turn serves two more remote ones: 1) the eventual understanding of normal language processing mechanisms and their brain representations, and 2) the refinement of clinical diagnostic and remediation procedures. The proposed project focuses on certain syntactic problems in the production of conversational and elicited narrative speech in Japanese and English-speaking patients.
The specific aims are: 1) to determine the syntactic structures used by patients in expressing semantic action relations which do not fit the canonical 'animate agent-active verb - inanimate object' schema; 2) to characterize the lexical, syntactic and pragmatic differences across diagnostic categories; 3) to account for variations in the syntactic structures and error types found in conversation vs. elicited narrative; and 4) to arrive at a formulation of the way in which patients' syntactic resources are limited--a formulation which will be equally applicable to English and Japanese in spite of the great differences in the grammars of these languages.l Such a formulation (in psycholinguistic processing terms) can then be further tested against data from aphasic patients speaking other languages. A fifth aim is to make all the raw data available in computerized form for further analysis by other aphasiologists. The findings will be compared to the hypotheses of Bates & Wulfeck (in press) and of Kolk et al. (1985) concerning the relations between fluent and non-fluent output and between conversation and elicited narrative. The data will be examined to see whether the background and preliminary findings about the importance of canonical word order are borne out. If they are, a language production model will be constructed which will have an explicit role for the canonical word order. In addition to the elicited conversation and narrative, full standard test protocols for comprehension and production will be used for individual characterization of the patients, and also neurological reports, in order to address syndrome definition and localization issues. Fifteen Japanese subjects and twenty to twenty-five English-speaking patients will be studied.l The patients will include representatives of the major diagnostic subcategories,k and will also include patients who do not fall into the classical syndromes. All patients will have had single left- hemisphere injuries and be free of complicating factors such as alcoholism; patients with CVA's will be used to the extent possible, and etiology will be treated as a variable. The patients will be matched for age, sex, time post onset, and aphasia type. Ten control subjects will be tested in each language.
