Current NSF grant: Computing constraint-based derivations

The project “Computing constraint-based derivations: Phonological opacity and hidden structure learning” (NSF BCS-424077), started September 1, 2014, and will last for 3 years. Gaja Jarosz, John McCarthy and Joe Pater are directing the grant, and the other senior personnel are Robert Staubs (UMass) and Mark Johnson (Macquarie). The project plans include the creation of a publicly accessible database on instances of phonological opacity, development of accounts of opacity in Harmonic Serialism and comparison with other approaches, extensions of computational tools for working with derivational versions of OT and Harmonic Grammar (“OT-Help 3”), and research on the learning of hidden structure in probabilistic frameworks, including learning of derivations.