This paper is a development of material from Honeybone (2001/2002). Versions of (some of) the material discussed here have been presented at the First Old World Conference in Phonology in Leiden, as a talk to the Philological Society in Cambridge, the Linguistics Association of Great Britain in Oxford, and as a job interview presentation in Edinburgh. I am grateful to the audiences there for their comments, and in particular to Phil Carr, Abigail Cohn, April McMahon, and the editors and reviewers for this volume for comments, questions and information at various stages of its development. Of course, this by no means implies that any of the above necessarily share or support my analysis or approach, and any errors in this piece are my own. 2 For some discussion of the notions 'reference variety' and 'non-reference variety', see Honeybone (2001); in line with one tradition of terminology, I often use the terms 'reference' and 'non-reference' in what follows, where other traditions might use 'standard' and 'non-standard'.