Mapping syntax using imaging: problems and prospects for the study of neurolinguistic computation

The study of language in the brain is the study of the computations responsible for the structure and processing of language. Hemodynamic imaging techniques can make a contribution to this area, but only if paired with a theory of the underlying formal processes. Statements like 'syntax is mediated by Broca's area' are too coarse (cognitively and cytoarchitectonically) to constitute insights into neurolinguistic computation. Numerous studies of syntactic processing report activation in Broca's area. However, other studies implicate Broca's area in non-syntactic/non-linguistic tasks, ruling out a restrictive linguistic interpretation of this region. This problem results from the failure recognize that 'syntax' is not a monolithic task, nor is 'Broca's area' a monolithic area. Further progress in this area requires linking neuroimaging studies with detailed theories of the linguistic and non-linguistic computations.