On reducing language to biology

Abstract

We are concerned with the relation of the psychological level of explanation to biological and formal accounts of the same phenomena. Our position with respect to the relations between different levels of explanation is that constraints operate only under very restricted circumstances. We use data from word recognition and speech perception to show that cogent accounts of the psychological processes involved require the establishment of purely psychological constructs, which are to be judged in terms of their explanatory usefulness rather than their compatibility with formal descriptions, on the one hand, and neurophysiological data on the other. Information processing models appear to be a suitable framework for expressing the psychological constructs. We examine briefly the way in which attempts have been made to account for developmental studies, results in dichotic listening, and neuropsychological work in terms of the architecture of the cortex rather than through psychological accounts of the subject's performance. While functional descriptions of the processes underlying psychological phenomena are immutably psychological, we accept that causal accounts of these phenomena may often best be given in terms of their biology.