>"Deep grammar" sounds like a third-hand version of "deep structure,"
>which is part of Chomsky's original, tentative formal model for language
>processing -- the logical (or "deep") structure of sentences is best
>described by what's called a "phrase-structure" grammar, but the
>"surface" structure -- what you hear, when a native speaker speaks --
>is the result of language-specific "transformations" applied to this
>deep structure. In English, for example, relative clauses get placed,
>generally, after the nouns they modify; in Japanese -- or so I understand --
>they often get placed ahead of them. Deep structure just specifies
>the dependence, not the order. But a native speaker of English who learns
>Japanese can recognize that what he's hearing is a relative clause,
>because once he figures out the transformations he realizes he's
>dealing with something he already knows, in a slightly different form.
>
>
I'm not sure this is true though -- Unfortunately I don't speak any non
Indo-European languages; among the Indo-European languages all kind of
similarities exist that allow these transformations to happen and be
fairly valid.
Benveniste tells us that the only linguistic universals are deictic markers (here vs there, this vs that, etc) markers that place things relative the the moment and locus of discourse. Other than that, it's catch as catch can structure language-wise.
Joanna