The Weekly Newsletter of MIT Linguistics

Phonology Circle 3/19 - Rafael Abramovitz and Adam Albright (MIT)

Speaker: Rafael Abramovitz and Adam Albright (MIT)
Title: Deriving allomorphic contingencies
Date and time: Monday, March 19, 5:00-6:30pm
Location: 32-D831

In this talk, we report on two loosely-related things we’ve been thinking about in connection with a handbook chapter on allomorphy that we are working on. First, we consider the analysis of portmanteau in Lakhota (Siouan). Lakhota verbs show both subject and object marking, but the combination of 1sg subject with 2nd person object is marked with a single, distinct affix /chi-/. Data from marking on embedded verbs shows that although /chi-/ indicates a subject/object combination, it must actually be analyzed as 2nd person object marker that is used in 1sg subject contexts. This is consistent with the hypothesis that portmanteau is actually complex contextual allomorphy of both heads (null + special allomorph). Moreover, the conditioning context for /chi-/ appears to be the morphosyntactic (featural) context of 1sg subject, and not the specific 1sg subject marker that it normally replaces. This is consistent with Bobaljik’s hypothesis that outward-looking contextual allomorphy is sensitive to morphosyntactic structure. The conditioning context appears to be non-local (in a higher XP), but evidence from stress and root allomorphy supports an analysis in which the root and both agreement suffixes have undergone head-movement to a single projection. Next, we discuss a case of cross-template blocking found in the Koryak (Chukotko-Kamchatkan) verb, whereby the appearance of the 3PL agreement marker /-w/ blocks the appearance of the general PL morpheme /-la/. Given that /-w/ is more peripheral in the verb-word than /-la/, this appears to be a case of outward-looking phonologically conditioned allomorphy, which is incompatible with Bobaljik (2000)’s hypothesis that vocabulary insertion proceeds cyclically. We argue that a reanalysis of these facts consistent with Bobaljik’s proposal is possible, but that it nonetheless requires some element of countercyclicity in the form of countercyclic application of impoverishment rules.