hide
Free keywords:
-
Abstract:
Language serves as a cornerstone of human cognition. However, our knowledge about its neural basis is still a matter of debate, partly because ‘language’ is often ill-defined. Rather than equating language with ‘speech’ or ‘communication’, we propose that language is best described as a biologically determined computational cognitive mechanism that yields an unbounded array of hierarchically structured expressions. The results of recent brain imaging studies are consistent with this view of language as an autonomous cognitive mechanism, leading to a view of its neural organization, whereby language involves dynamic interactions of syntactic and semantic aspects represented in neural networks that connect the inferior frontal and superior temporal cortices functionally and structurally.
Our conceptions of the neural mechanisms of language have developed in tandem with our understanding of the nature of the language faculty as a cognitive system. Initially, research focused on frontal and temporal cortical regions as being involved in vocal production and speech perception, respectively. Since speech is the main medium of language used for communication, it may seem natural to equate language with speech or even ‘acoustic communication’1. This view, however, is too narrow. Speech is just one possible way of externalizing language (with sign or writing being other examples), ancillary to the internal computational system. In addition, ‘communication’ is merely a possible function of the language faculty, and cannot be equated with it. We argue that language is a species- and domain-specific human cognitive capacity (Box 1)2,3,4,5,6. In essence, language is an internal computational mechanism that yields an unbounded array of structured phrases and sentences. These must be minimally interpreted at two interfaces—that is, internal thoughts on the one hand, and externalization via sounds, writing or signs on the other (Box 1)4,5,7,8. Neurolinguistics focuses on the study of the neural substrates underlying the computational cognitive mechanism that lies at the core of human language. From a theoretical linguistic standpoint—that of generative grammar—language is posited to be a process described at a formal level, divided into functionally separable or autonomous components, such as syntax, morphology, and so on. The immediate question of interest that then arises is whether the formal representations exploited in generative grammar correspond to actual brain architecture. We will discuss independent lines of research converging on the result that syntactic processes are in fact independently computed in the brain.
