yudantaiteki wrote:

It has occurred to many people, but it's a myth.  Native speakers of a language have rules for grammar in their brain, and some of them are very complicated.  The purpose of linguistics is to figure out what those rules are, and then "pedagogical syntax" attempts to take those complicated rules and distill them into something that learners can make use of.  But the fact that a stated grammar rule is complicated does not mean that it's wrong, however unsuitable it may be for learners.

I beg to differ. The purpose of linguistics is indeed the generation of grammars, but not the discovery of the grammars that people's brains use. A language isn't described exclusively by one grammar, but by many. The purpose of linguistics is to generate grammars with particular characteristics (economy, simplicity, &c.), which aren't necessarily characteristics of our brain-grammars.
(Or at least so I understood from Chomsky's "Syntactic structures". If I misunderstood, someone please let me know. Even if I'm wrong, that's the way I think it should be, leave to brain to the psychologists.)