cortical linguistics Kortikale Linguistik                    DEUTSCH

Günter Kochendörfer

Cortical Linguistics

Abstract

Chomskyan generative grammar is (by definition) not a possible model of syntax in the brain. The same is true for more recent syntactic theories such as radical construction grammar. Modern tools of brain investigation such as PET, fMRI, and ERP are not yet able to provide information precise enough to answer many of the important questions in this field.

The rapidity of computational processes in the brain requires massive parallel processing and therefore the identity of memory structures and processing units. As a consequence of this requirement, there are difficulties in adopting rules for processing components. On the other hand, syntactic competence is not like a lexicon with whole sentences as entries. Connectionist models in the tradition of Rumelhart & McClelland, Elman, and others, which some linguists consider a possible solution, have not yet been constructed with sufficient complexity to mirror the observable syntactic facts.

Nevertheless, linguistics cannot work without syntax, and, in the post-generativist era, attempts must be made try to gain some insight into the nature of syntactic representations and processes in the brain. This part of the project of "Cortical Linguistics" aims at developing new proposals for understanding syntax, based on information regarding the possibilities of neural functioning, consequences resulting from other components of language processing (phonology, lexicon, semantics), and observable characteristics such as the possibility of repairs in speech. In this investigations, simulation models are used everywhere as essential tools.

Last update 9/15/2007