Closed- and open-class lexical access in agrammatic and fluent aphasics.

Bradley, Garrett, and Zurif (Bradley, Computational distinctions of vocabulary type. Unpublished doctoral dissertation, MIT Press; Cambridge, MA, 1978; Biological studies of mental processes, MIT Press, Cambridge, MA, 1980) have suggested that closed-class word access is normally mediated by a different route than the open-class one, and that the loss of this closed-class route might account for agrammatism. However, in an earlier study (Gordon & Caramazza, Brain and Language, 15, 143-160, 1982) we were not able to confirm a meaningful difference between closed- and open-class word frequency responsiveness of the type Bradley (unpublished dissertation, 1978) had seemed to find in normal subjects. We have now done a direct comparison of closed-class frequency sensitivity in agrammatic and nonagrammatic aphasics, to directly test Bradley and colleagues' hypotheses and to avoid some of the experimental problems with between-class frequency comparisons. We find that closed-class words behave similarly whether or not the subject is agrammatic. Therefore, the differences between agrammatic and nonagrammatic aphasics must arise at a deeper level (or levels) of lexical processing than the one probed by the frequency sensitivity effect.