The Evolang conference in Barcelona officially begins tomorrow, but there were a number of “pre-conference” workshops and public lectures today. At one of them a neurologist, Friedmann Pulvermüller, reported that the view of language as the product of an enclosed, encapsulated system within the brain cannot be maintained. This statement posed a direct challenge to Noam Chomsky who argues that syntax and semantics developed as a separate system which “interfaces” with other parts of the brain. (See: Chomsky’s Theory of Language Origins).
Arguing from the axioms of neural networks, Pulvermüller said the brain’s language, actions, and perception systems are heavily interwoven. Learning to speak words (or gesture linguistically) depends on associated connections made between nearby neurons and also with associations established between distant regions of the brain. Chomsky might retort that the language developed originally for thinking and only needs these other connections to externalize these thoughts, but experimental evidence refutes that objection. The assemblies of neurons that fire for particular words are linked to perceptual and motor areas as well. So, when the word “knee” fires an assembly in the linguistic cortex it also triggers activity in the leg motor area as well.
Word learning and recognition activity in the brain follows closely with the predictions of associationists and behaviorists; however, even the most radical believers in innate linguistic knowledge assume that the vocabulary must be learned. The real split is over sentence structure, with associationists saying that grammatical relationships are learned through the probabilities while others insist the relationships are rule-dependent. Pulvermüller has found evidence in favor of rules (reported in the Journal of Cognitive Neuroscience, abstract here). This argument had seemed settled 50 years ago when Chomsky rejected behavioral accounts of grammar; Pulvermüller’s data now adds neurological confirmation to the matter. But just as Pulvermüller seemed to be handing Chomsky his victory, he pointed out that the presence of rules does not mean they are innate. Neuronal circuits can develop that are analogous to syntactical rules of language. At this point the question remains open and the fundamental issue of whether some linguistic rules are innate remains open, even at the level of brain circuitry.
Comments