Previous chapters have suggested that the human mind is equipped with core phonological knowledge – a system specialized for the computation of phonological structure. This chapter examines what brain mechanisms mediate phonological computation and evaluate their presumed genetic underpinnings. While the findings suggest that a neural phonological network certainly exists, they cannot determine whether this network is specialized for phonology. An answer to this question hinges on how specialization is defined, and, more generally, how cognitive explanations are linked to neuroanatomical models. Existing neuroanatomical models presently lack an explicit account of that link. I thus conclude that specialization, in general, and the hypothesis of core phonology, specifically, can be presently evaluated primarily at the functional, cognitive level. Neural data can be profitably correlated with functional findings, but they can rarely falsify functional hypotheses concerning specialization.

Individuating cognitive functions: functional specialization vs. hardware segregation

At the center of this book is the question of specialization: Are human minds equipped with a system specialized for phonological patterning? The previous chapters present several observations that are consistent with this possibility. We have seen that distinct phonological systems share design principles that distinguish them from nonlinguistic systems, that knowledge of grammatical universals is evident even when they concern structures unattested in one’s language, and that the capacity for phonological patterning emerges spontaneously, in the absence of a model. Not only are phonological constraints universal and possibly innate, but they are also demonstrably distinct from nonlinguistic pressures, most notably, the phonetic pressures governing the processing of aural stimuli and their production. Functional specialization, however, should be further mirrored at the neural level. If the mind has a specialized computational system dedicated to phonological patterning, then one would expect this special “software” to require a specialized brain “hardware” that mediates phonological computation. The brain networks that support phonological computation could potentially present another test for the specialization of the phonological mind.