The multiple bases for linguistic structures

A cornerstone of “The cognitive basis for linguistic structures” (CBLS) was to highlight Chomsky’s famous distinction between (1) knowledge of language and (2) how that knowledge is put to use, emphasizing that these need not, and indeed probably were not, one and the same thing—if only because knowledge of language makes no reference whatsoever to the computational steps involved in recovering or producing structured representations from serial input or output, while how language is put to use must explicitly address these essentials. CBLS stressed that the constraints on representations inherited from linguists’ grammars might well be involved in language processing even though the grammars themselves might not be. In this way, CBLS highlighted a fundamental tension between two very different motivations behind the study of language—the first aiming to model external behavior, at heart, a theory that, in the limit, sets out to predict the next thing a speaker will say or a listener will perceive; the second aiming to model knowledge of language, at heart, a theory of a person’s internal state that enables them to acquire any language. How might we reconcile these two quite different motivations?