Morphology & Syntax Workshop
|When||December 07, 2018 11:30 AM - 1:20 PM |
|Where||Harper Memorial Library, Room 141 |
|Event Website||https://voices.uchicago.edu/morphologyandsyntax/ |
|Contact Information||Linguistics Department |
|Description||Barbara Citko (University of Washington) |
Binarity of Merge and Structural Syncretism
Barbara Citko & Martina Gračanin-Yuksek
While syntactic research on multidominance has led to a number of important theoretical and empirical insights involving the properties of multidominant (MD) structures, the linearization and interpretation of such structures, there is less of an understanding of how different types of multidominant structures interact, and what this interaction might tell us about the properties of Merge itself. In this talk, we seek to remedy this gap. We assume the basic architecture of the grammar in which there are two types of multidominant structures: Internal Merge (Move) creates MD structures in which one of the mothers dominates the other (as in (1a)), and Parallel Merge creates MD structures in which neither mother dominates the other (as in (1b)). Furthermore, both Parallel Merge and Internal Merge can apply to a single object (as in (1c)).
(1) a. Internal Merge/Move b. Parallel Merge c. Parallel Merge + Internal Merge
We examine the interaction between Parallel Merge and Internal Merge, and show that it reveals constraints on Merge that are not detectable in configurations that do not involve multidominance. We contrast Right Node Raising (e.g., John wrote and Mary reviewed a new article.), which we take to involve Parallel Merge, with ATB wh-questions (e.g., What article did John write and Mary review?), which we take to involve both Parallel Merge and Internal Merge. Based on the data from English, Polish and Croatian, we show that Right Node Raising tolerates mismatches that are not possible in ATB questions (namely those that violate the Parallelism Constraint of Williams 1978), and we derive this contrast between RNR and ATB movement from a constraint on Merge we dub BiCoM (Binarity Constraint on Merge), which prevents Merge from relating more than two positions within a single derivation. We then show that Merge can sometimes relate three (or more) positions, but only if two of them are identical (‘structurally syncretic’). Thus, structural syncretism can rescue otherwise impossible ATB derivations, just like morphological case syncretism can rescue ATB derivations with conflicting case requirements on the fronted wh-pronoun. We conclude by examining the consequences of BiCoM and structural syncretism for the effects of embedding (i.e., why do the non-parallel cases improve when the subject is embedded?) and the vacuous movement hypothesis (i.e., why do the non-parallel cases improve if the subject is in the first conjunct, as opposed to in the second one?)?
|Categories||Conferences/Lectures, Workshops |
|Persons with disabilities who need an accommodation in order to participate in this event should contact the event sponsor for assistance. |
Back to main calendar