One area of disagreement is the relevance of contrast to the architecture of grammar and the course of phonological derivations. The Toronto school, for example, advocates delaying the introduction of noncontrastive information, possibly until the application of gradient processes of enhancement and coarticulation in phonetic implementation. Through its morphosyntactically defined levels, Stratal OT allows redundant features a choice of entry points into the derivation, but does not necessarily ban allophones at any level. Classical OT goes further: it encodes contrast in the constraint hierarchy, but does not afford any means of segregating contrastive from noncontrastive specifications in either the input or the output. Finally, exemplar theory appears to deny contrast a privileged position in the formation of its fuzzy categories: other factors such as social indexicality may in principle carry just as much weight.

There is also no agreement on how contrast should inform the theory of representations. Many phonologists have postulated a universal inventory of distinctive features enabling the expression of all the phonemic contrasts that may occur in the languages of the world. Typically, phonological processes have been assumed to operate on features belonging to this set. But this entails that the categorical phonology of a language L can refer to a feature F only if there exists at least one language, not necessarily L, in which F is contrastive. Is this true? Scholars such as Steriade deny this traditional claim. And, of course, important consequences follow from the choice between unary elements and binary features. (And there are also attempts to model contrast licensing through acoustic distinctiveness rather than through phonological features.)

Difficult questions also surround the relationship between phonemic contrast and phonetic substance. On the one hand, the viability of an opposition is significantly constrained by phonetic factors such as perceptual discriminability: Stevens's quantal theory, for example, emphasizes the importance of regions in phonetic space where a small articulatory change produces a large acoustic effect. On the other hand, established contrasts appear to guide phonetic implementation: /s/ is likely to have a larger range of realizations in a language where it is the only coronal fricative phoneme than in a language where it contrasts with /theta/ and /esh/. How is this two-way relationship between phonemic contrast and phonetic substance to be negotiated? In particular, is the relationship largely emergent (as in Boersma's bidirectional framework), or does the grammar include devices dedicated to the evaluation of contrast (the P-map, dispersion theory, etc)?

Finally, the nature of contrast itself is in need of elucidation. Oppositions may be more or less marginal, and so the phonologist faces delicate choices when contrasts arise only in certain morphosyntactic environments (cf. Harris's notion of 'derived contrast'), or in certain subsets of the vocabulary, or only in consequence of the irregular application of erstwhile mandatory processes. What are the implications of incomplete neutralization and Labovian near-merger for the grammar of contrast? Is it true that fine phonetic detail can be lexically specific, and, if so, how should this be accommodated in our understanding of contrast?

Our invited speakers will provide different perspectives on these fundamental questions.