Where does the sociolinguistic variable stop?

In 1972 Labov described the basic sociolinguistic question as the one ‘posed by the need to understand why anyone says anything’ (1972:207). Clearly the aim is very different from that of specifying the form of a grammar that generates all and only the well-formed sentences of a language. The goal is a theory of utterances. Moreover, the ‘why’ question can also be read as ‘what for’. What does anyone say anything for? I think we can safely say that this question places sociolinguistic analysis in a functional framework. If sociolinguistics looks for answers to the ‘why’ of saying something, it is seeking functional explanations.