ABSTRACT
Phonological change begins with changes in pronunciation. The problem of how sound changes become part of the phonology, and the problem of how to impose 'naturalness constraints' on phonological descriptions, are pseudo-problems that have arisen only because modern linguistics, obsessed with economical descriptions, has separated phonology from phonetics. But they are inseparable: the phonology is the phonetics, and phonological change is phonetic change. Scholars like William Labov, Ohala, Janson, and others have documented phonetic change and explored the reasons for such change, but they have not explained how the phonetic changes are implemented in speaker's phonological systems, or how they affect these systems. Conventionalist theory, where phonological rules (like morphophonological rules) must be learned, fails to explain systematically how phonetic motivation is related to the adoption of new phonetic targets, how phonetic rules are acquired, or what it means for these new rules and targets to become phonological.
