Speaker: Heidi Durresi (MIT) and Giorgio Magri (MIT, CNRS)
Title: Implicational universals with negated consequents in categorical HG and probabilistic ME
Time: Monday, May 1st, 5pm – 6:30pm
Location: 32-D831
Abstract: A sensible strategy to understand a large typology of phonological grammars is to extract the universals shared by its grammars. In this talk, we look at implicational universals between an antecedent phonological mapping and the negation of a consequent phonological mapping. We say that this universal holds of a categorical typology (OT, HG, etcetera) provided every grammar that contains the antecedent mapping does *not* contain the consequent mapping. We say that this universal holds of a probabilistic typology (ME, NHG, etcetera) provided the probability of containing the antecedent mapping is never larger than the probability of *not* containing the consequent mapping. We give a complete constraint characterization of the implicational universals with negated consequents that hold in categorical HG and a complete characterization of those that hold in probabilistic ME. We derive as a corollary that any such universal that holds in ME also holds in HG. We then deploy these characterizations on a number of test cases and show that ME often misses none of the universals captured by HG. We conclude that universals with negated consequents provide a better tool to study ME typologies than the universals with simple consequents investigated in Anttila and Magri (2022).