In this talk I will evaluate different characterizations of the role of Broca's area in receptive language. Specifically, I will compare a purely syntactic approach to one based on Working-Memory (WM). The evidential basis for this evaluation comes from lesion-studies (aphasia) and from functional neuroimaging. Broca's area has been known as neural substrate for processes that establish syntactic movement relations. Until recently, the description of syntactic deficits subsequent to focal brain damage was the sole source of evidence for the identification of the functional role of this cerebral region. I will first review some recent aphasia data that bear on this description, with an eye to its cross- linguistic nature (results from English, Chinese and Japanese will be presented).
Second, I will move to PET and fMRI investigations. Some experiments carried out with these methodologies in recent years have helped corroborate the syntactic view on Broca's area. But there are other results that have led to a new perspective: It has been proposed (most notably by Smith and Jonides) that this region (and more specifically BA 44 within it) supports subcomponents of WM, involved in short-term storage of verbal material. If this latter view is correct, one could imagine Broca's area to have no particularly syntactic character, and that what appears a syntactic role may just be a special case -- a mere instantiation of WM: Since some type of memory is necessary for the on-line computation of movement structures, the very same WM component -- evidenced through fMRI and PET tasks -- may be involved in the on-line analysis of sentences. I will argue that this view is incorrect. I will give one conceptual reason against it, and accompany it with three pieces of relevant data, all coming from recent experimentation on Broca's aphasic patients: Comprehension and grammaticality judgment tests with syntactic constructions that involve intra-sentential dependencies of three different kinds: Head-movement, binding, and a mix between binding and movement. These results will lead to a restrictive view of the functional role of Broca's area: It is home to the part of the language faculty that is dedicated to the computation of movement of phrasal constituents. I will discuss certain predictions that follow from this conclusion, and point out some possible consequences to the theory of syntax.