Blurb:

Harmonic serialist inflectional morphology incorporating MinSat raises the question of how standard instances of disjunctive blocking (or disjunctive ordering) in morphology can be accounted for, where it would seem that it has to be the exponent that maximally satisfies faithfulness constraints that must be selected. I argue that this is also done gradually, in a series of optimization steps: Based on an initial morphological array (Kager (1996), Masaró (1996)) containing all the potential exponents for a given slot, the derivation selects an exponent in accordance with MinSat, and discards it again in the next optimization step in favour of an even more faithful exponent, and so on, until the most specific exponent satisfying Ident constraints remains as the optimal candidate on which the derivation converges. This approach necessitates two simple assumptions: First, there always is an elsewhere marker in a morphological array. And second, extended exponence scenarios differ from disjunctive blocking senarios only in that the additional, newly selected exponent removes the earlier, locally optimal exponent in the later case, but not in the former one. (See Müller (2017a;b) for arguments for an operation of structure removal in syntax.)