This talk examines the interaction between phonology and morphology, and also syntax, in the realization of nominal inflection (gender/class and number) in dialects of Catalan. The two phenomena that will be considered can be analyzed fairly easily under a morphous approach to morphology while they are problematic for a-morphous approaches.

Ablaut is discussed as a problem case for approaches to exponence, using Contemporary Standard German as an example. It is shown that the multiple gradations of German strong verbs are but manifestations of a rather uncomplicated system. There is a small number of ways to make up ablaut forms; these types of formation are identifiable in formal terms and, what is more, they have definite functions as morphological markers. On this basis, gradations may be collected into inflectional classes that are defined in strictly synchronic terms. At the same time, these classes continue the historical seven classes known from reference grammars.

I will argue that morphologically-driven phonological mutation is the result of segment coalescence - when two segments merge for phonological reasons (e.g. /ai/->[e]). The motivation for 'mutating' coalescence is often contained in the morpheme itself: it contains one or more unacceptable surface segments. Mutation therefore follows from independently motivated processes: coalescence and absolute neutralization. Two implications of the theory proposed are that no morpheme can be smaller than a segment, and that there are no special mechanisms for mutation and phonological polarity.

Although the core of OT is compatible with virtually any idea about morphological exponence, most of the literature seems to implicitly assume a mixed model where the exponence of morphemes consist of (i) phonological stuff (features, segments, mora's, etc.) and (ii) constraints which are specific to morphemes or groups of morphemes. In this tutorial, I discuss some of the discussion in the literature concerning both assumptions, in particular, the principle Consistency of Exponence, and the desirability of indexed constraints. Furthermore, I pay some attention on assumptions concerning the interaction between morphology and phonology more generally. I conclude by taking a brief look at the Trommer Table, i.e. the table of difficult cases of morphological exponence that Jochen Trommer compiled in his proposal for our current network: which is the 'mainstream' idea within OT on the different elements of the Trommer Table?

This talk outlines some of the evidence for a particular view of the mapping from syntax to exponence, arguing for a particular order in the operations involved (assignment of morphological case, assignment of agreement features). Special attention is given to the analysis of agreement in Chukchi, including the Spurious Antipassive construction (Hale/Halle 1998, Spencer 2000, Bobaljik & Branigan 2006). I offer here a slight revision of the account offered in Bobaljik & Branigan 2006, and focus on the implications of the construction for an understanding of the kinds of manipulations on feature structure, prior to the operation of rules of exponence, that must be countenanced.