Ha. I'd forgotten that. It's been at least 15 years since I read it, I think.Quoting lovejuice (view post)
Yeah, I have the same problem with Chomksian linguistics in general: he says that he's talking about concrete, innate physical systems in the brain, but his actual theories are only models of usage. In order to make any connection with the brain, he'd require a mapping from his symbolic systems to those of cognitive science, and finally, to the actual processes in the brain. The underlying argument seems to be that if we assume that an innate system S1 exists in the brain, and an isomorphic mapping f:S1->S2 exists that maps the system in the brain onto the system S2 that models language usage, then the first system and the map together induce the model of usage. Hence, from Chomsky's assumptions, we can say that there is a model of usage, and that is the level on which his axiomatic linguistic theories operate. Conversely, we can say that if no such model of usage exists (i.e. if no such model can accurately model all human usage), then Chomsky's assumptions are wrong. Historically, this was basically Chomsky's path in his criticism of Behaviorist linguistics. However, the success of a model of usage really says nothing at all about the innate processes in the brain, if any, or the map between them and the model of usage (in particular, what are the uniqueness properties of this map?). At this point, the system S1 and the map f are not merely black boxes, but operationally meaningless; they have not been discovered or formulated, and they play no role in Chomsky's actual theory, only in his justification of using the theory.
Sorry, I'm tired and frazzled. Not sure if this makes sense.