The University of Massachusetts Amherst

3. HG with language-specific constraints

Because of their connections to models used in machine learning, weighted constraints offer special opportunities to develop explicit models of constraint learning. In this section of the course, we consider such models, as well as their implications for phonological theory.

Language-specific weighted constraints

Daland, Robert, Bruce Hayes, Marc ?Garellek,? James ?White, ?Andrea ?Davis,?Ingrid ?Norrmann. To appear. Explaining Sonority Projection Effects. In Phonology.

Hayes, Bruce and Colin Wilson. 2008. A maximum entropy model of phonotactics and phonotactic learning.  Linguistic Inquiry 39:379-440. (follow link for software too).

Pater, Joe. 2011. Canadian Raising with Language-Specific Weighted Constraints.Ms, UMass Amherst.

Empirical evidence: formal vs. substantive biases

Moreton, Elliott and Joe Pater. 2011. Learning artificial phonology: A review. Ms, UNC Chapel Hill, UMass Amherst.


Leave a Reply

Your email address will not be published. Required fields are marked *