In response to Kevin’s question about the Align-Head /W-S interaction in HG:

1. I should have pointed out that it’s not entirely clear that the HG result is problematic. That’s the line I take in the Cognitive Science paper – see section 3.1 (this is also the section that talks briefly about the HG-HS connection).
2. My best shot at doing some Align-Head effects without gradient alignment is in section 1.2 of this handout.

As a follow-up to the point Karen raised at the end of class, I made two OT-Help files, one with Ident-Nas, and one with Max-Nas. I use capitals for nasalized vowels. Try each of these out in parallel HG (“Find HG solution”), and try to understand why Karen is right – the strange pattern is impossible with Max-Nas. To think about it, you might find it useful to go to the “comparative view”, and remember that the sum of the weights preferring a winner has to be greater than that for the constraints preferring any loser. This should allow you to construct a set of inconsistent inequalities. You might also want to think about whether a parallel theory with only Max-Feature constraints is viable, and whether the answer changes in serial OT/HG. We’ll talk about this on Tuesday.

Remember to come to class Thursday with a discussion point for each of the three papers. We’ll also need to plan for the following Thursday, so think about what might be useful to advance our understanding of HG and HS.