Neuropsychological reasons for a transformational analysis of verbal passive

Experimental findings from the neuropsychological study of language are used to evaluate grammatical theories. Data from tests of sentence comprehension by agrammatic aphasic patients reveal a pattern of selectivity that is used as a means for assessing competing syntactic hypotheses. In particular, it is argued that agrammatic linguistic behavior provides evidence for a transformational analysis of verbal passive. This is so because the patients failed to comprehend properly verbal passives and structures involving Wh-movement, but succeeded in tasks that required the proper comprehension of lexical passive. Grammatical theories in which such a performance pattern can be readily described - namely, theories that allow for generalizations such as Move-alpha - pass a test of psychological reality in that they are compatible with patterns of language breakdown observed following brain damage. By contrast, syntactic frameworks in which all passives are derived by lexical rules cannot account for this performance pattern without the introduction of ad hoc devices. Neurolinguistic findings are thereby shown to restrict the class of biologically feasible grammars.