Friday March 9th   15:30   Greenberg Room

John J. McCarthy

University of Massachusetts Amherst

Getting to Optimality

This talk will examine a modification of Optimality Theory that incorporates something analogous to the derivations of rule-based phonology. Classic OT's operational component GEN and its evaluative component EVAL do not interact: GEN applies its operations freely to create a wide variety of output candidates, and EVAL decides which one is most harmonic (=optimal). In the modified theory, called harmonic serialism (Prince and Smolensky 1993/2004) or OT-CC (McCarthy 2007), GEN applies an operation only when the result is more harmonic, according to EVAL. From this it follows that, if several operations are involved in getting from the underlying form to the surface, there must be a path of monotonic harmonic improvement from one to the other. If there isn't such a path, "Come to think of it, you can't get there from here".

OT-CC has applications to the problem of phonological opacity, but this talk will focus on its typological consequences. The requirement of monotonic harmonic improvement is more restrictive than classic OT's harmony requirement, and as a result the same constraint set yields a more limited typology in OT-CC than in classic OT. This talk will explore the typological differences.