There is something very similar to the chicken-or-the-egg question about formal grammar in which meaning is ignored and only the formal properties of the string is discussed.
But would that string exist without production of meaning, without the communicative desire to impart something in an instance of existence?
This is also like asking what is the meaning of life as if life needs to have some universal meaning or purpose before life can arise. There is something a priori about this logic.
This is where Chomskyan linguistics, to me, fails to convince – that there can be an explanation of language without meaning. Whether it be a word, phrase, clause or sentence there will always be two sides to a sign (in the Saussurean sense). Phoneme and basic-unit phonology are different in that they are the building blocks of language and not invested with meaning.
By looking at language and variation in the system is a mathematical exercise that cannot explain the inherent meaning of utterances (which, sadly, it is not trying to explain at all in the first place). For me language is about meaning, and about the limitations a language’s form has on expressing meaning and not the other way around. Syntax should therefore take into account semantics or rather syntax should be studied through semantics.