Language and Brain – Facts, Problems, Mysteries

Abstract

The representation of language in the brain is one of the lively topics in recent cognitive neurosciences. The present paper argues that standard models of structures and mechanisms of the brain, which are essentially based on principles of activation spreading in neuronal nets and learning by synaptic strengthening through coincidental activation, cannot account for crucial properties of the Language Faculty and are misguided in central respects. In particular, the apparently elegant notion that working memory is just the activated state of long-term memory is shown to be incapable of accounting for structures and processes of language production and comprehension. Four types of problems are sketched that illustrate the impasses encountered by standard models of activation spreading.