Postaccess processes in the open vs. closed class distinction

We present the results of two auditory lexical decision experiments in which we attempted to replicate findings originally presented in Bradley (1978, Computational distinctions of vocabulary type, Ph.D. dissertation, MIT). The results obtained by Bradley were used as evidence for a processing distinction between the open and the closed class vocabularies; this distinction then used as part of an explanation for agrammatism in the comprehension and production of Broca's aphasics. In our first experiment we failed to replicate Bradley's result of frequency insensitivity in the closed class. Our second experiment, however, replicates Bradley's finding that closed class based nonwords (e.g., thanage) fail to induce interference effects in nonword decisions. We argue that our results, together with the various other reported failures to replicate Bradley's frequency insensitivity result, indicate that the open and closed classes may play distinct roles in postaccess phenomena involving the processing of morphological information but that such studies cannot address the question of whether the open vs. closed class distinction plays a role in syntactic processing.