Sentence Comprehension in Aphasia

We show in this chapter that selective aphasic deficits dissociate distinct language processing systems. To this end, we focus on disruptions of lexical access that appear to implicate initial stages of sentence comprehension. The analyses we present turn on comparisons between agrammatic (Broca's) aphasic patients and paragrammatic (Wernicke's) aphasic patients. By the former we mean roughly, that type of aphasia in which complex syntactic structures and closed-class elements tend to be omitted in production and abnormally processed for comprehension purposes (Berndt & Caramazza, 1980; Zurif & Blumstein, 1978). Wernicke's aphasic patients, by contrast, present with an output that features misuse of grammatical devices-prepositions notably-and not telegraphic omission (Goodglass & Kaplan, 1972). Patients within this latter group also have sentence comprehension problems (Goodglass & Kaplan, 1972)-on some accounts, problems of the same sorts as those shown by the agrammatic patients (e.g., Goodglass & Menn, 1985; Shewan & Cantor, 1971). However, arguments for similarity in this respect rest on demonstrations that both groups fail to comprehend complex sentences. Without arguing the matter in detail, it is entirely possible that an apparently similar failure can have different antecedents (e.g., Friederici, 1983; Pastouriaux, 1984). Our group analyses are designed to explore this possibility. A general feature of the analyses described here is that, unlike much work in neuropsychology, we do not translate freely between linguistic and processing theories. We do not distinguish, say, between a syntactic processing module and a phonological processing module simply in virtue of formal linguistic terminology. Rather, although we accept that processing modules are initially defined by linguistic information types, we seek evidence for their existence in terms of fixed and mandatory operating