The Weekly Newsletter of MIT Linguistics

Ling-Lunch 10/13 - Barbara Citko

Speaker: Barbara Citko (Visiting Scholar, University of Washington)
Title: In Search of MDDs (Multidominance Diagnostics)
Location: 32-D461
Time: Thursday, Oct 13, 12:30-1:45p

Many existing multidominant (MD) proposals have focused, either directly or indirectly, on various types of coordinate structures, such as right node raising (Wilder 1999, De Vries 2009, Johnson 2007, among others), across-the-board wh-questions (Citko 2005, in press, De Vries 2009, among many others), gapping (Goodall 1987, Kasai 2007, among others), questions with coordinated wh-pronouns (Gracanin- Yuksek 2007), determiner sharing (Citko 2006, Kasai 2007). However, the presence of coordination is not a reliable diagnostic of MD for two reasons. First, there exist coordinate structures that do not involve MD (simple coordinate structures with no ellipsis or movement whatsoever), and second, there are many non-coordinate structures that can be (and have been) analyzed in a MD fashion. These include (but are not limited to): free relatives (Van Riemsdijk 2006, Citko 2000, in press), parasitic gaps (Kasai 2007), serial verb constructions (Hiraiwa and Bodomo 2008), amalgams (Kluck 2008), comparatives (Moltmann 1992), discontinuous idioms (Svenonius 2005). Thus, given the fact that multidominance is independent from both coordination and ellipsis, the search for reliable diagnostics of a MD structure continues.

The intuition that all MD approaches build on is that the shared element has to simultaneously satisfy the constraints imposed by it by the two elements between which it is shared. In other words, it must match these two elements. In this talk, I explore the nature of this matching requirement with an eye towards developing a set of diagnostics for a MD structure. More specifically, I address the following three questions:

A.What kind of matching do MD structures require?
B.What kinds of mismatches do MD structures tolerate?
C. Why are different types of MD structures appear to be subject to different matching requirements?

Drawing on data from Polish, I compare the matching requirements in four types of arguably multidominant constructions: two coordinate ones (right node raising and across-the-board wh-questions) and two non-coordinate ones (parasitic gaps and free relatives), focusing on the differences between them and ways to account for these differences.