functional phonology

[Keith Slater]

I think there's a good reason why generative phonology seems more
functional than does generative syntax.

>From the outset, the generative program in phonology seems to have
tried to account for the same sorts of facts that had been described
in historical phonology.  Underlying representations were (more or
less) based on internal reconstruction, and the goals of the theory
thus had to include the modelling of the sorts of phonological and
phonetic changes that actually occur in human language.

Unfortunately, generative syntax had no such historical model to base
its approach on.  One can only wonder what generative syntax would
look like today, if there had been (for example) a well-developed
approach to grammaticization before Chomsky came along.  Perhaps we'd
see a more "functional" program in syntax, too.

Keith