This chapter identifies what the authors believe to be the core computational primitives that underlie phonological knowledge. It presents evidence from the domain of cognitive neuroscience that attempts to investigate the nature of the neural correlates of these primitives. Phonological grammars provide a framework for how speech sounds are represented and the nature of the various combinatorial operations they undergo in mapping between lexical representations and their surface-forms. Traditionally, generative phonology has asked two questions: (1) what representations subserve phonology (e.g., features, segments, syllables) and (2) what procedures map between surface forms and memory representations. The only language-specific circuits motivated by classic phonetic and phonological theory are those needed to represent language-specific phonetic and phonological primitives in long-term memory. The chapter focuses on the nature of phonological representations, the primitives that have been investigated to date in the cognitive neuroscience literature.

This chapter focuses on linguistic operations, and discusses the possible candidates for primitives of syntactic computation and the state of the art with regard to their possible neurobiological correlates. From the perspective of functional neuroanatomy, empirical investigations into the processing of long-distance dependencies have, perhaps unsurprisingly, focused largely on the role of Broca's region. The chapter discusses the issue of sequencing (i.e. word order) more generally. The positive correlation between dependency distance and left inferior frontal gyrus (lIFG) activation is one of the best investigated phenomena in the cognitive neuroscience of language. Many of the existing neuroimaging findings on syntactic processing appear to be more parsimoniously explained in terms of more general cognitive mechanisms (specifically: cognitive control). The chapter discusses "Merge" as the most promising current candidate for a neurobiologically implemented primitive of syntactic computation, and describes approaches that can bridge the gap between grammar and processing.

This chapter offers a satisfactory account of morphological processing within the overall language system from a neurobiological perspective. It discusses the computational primitives in morphology and their possible brain correlates. The chapter reviews research on the two morphological distinctions that have attracted the most attention in the literature on the neural bases of morphological processing: regular versus irregular morphology and inflection versus derivation. It argues that neither of these two oppositions appears suited to explaining how morphology is organized in the brain. The chapter offers some more positive suggestions regarding the neural representation and processing of morphology. It also argues for the primarily relational role of morphology, which serves to distinguish it from the combinatory nature of syntax, and for the importance of distinguishing between purely formal and interpretively relevant relations in this regard.