Current thinking on language structures

There are multiple language processing areas that are widely but not randomly distributed in the brain. To map these brain areas onto language functions in real time requires knowing not only how they are distributed in space but how they are coordinated in time. Any psychological theory of language processing must include a description of the mental representations involved, computations on them, the timecourse of their operations and interactions. The method of choice for precise temporal tracking of sentence comprehension is the event-related brain potential (ERP) technique. Its application to the analysis of written and spoken sentences of varying structural complexity reveals the critical roles of working memory operations and background knowledge in normal comprehension. Such data show that many of the processes of language include large parts of the brain, are neither modality nor language-specific, are subject to individual differences, and overlap each other in neural and mental space and time, such that analyses of both dimensions are needed to tease them apart so as to understand how we understand.