Introduction

The issue of the abstractness of phonological representations seems to be consistently avoided in the literature generally. There is an assumption in many papers that phonetic implementation is not a problem and it is, therefore, not addressed. (This is instanced in the many contributions to Goldsmith, 1995, some of which I have already referred to in the previous chapters; this collection of papers has been chosen for particular scrutiny, firstly because I reviewed it some time ago, Lodge, 1997, and secondly it was intended to be a collection of up-to-date seminal papers on phonological theory.) On the other hand, there are specific references to the matter; for example, Harris & Lindsey (1995) and Spencer (1996: 141-4), though in both cases it is assumed that phonological elements should have a direct and seminal phonetic interpretation. In a footnote Clements & Hume point out that their place features ‘labial, coronal and dorsal have specific phonetic correlates just as other features do’ (1995: 301). These are listed in the text (Clements & Hume 1995: 252 (7) and, revised to include vowels, 277 (41)), as follows: