Whamit!

The Weekly Newsletter of MIT Linguistics

Syntax Square 2/14 - Colin Davis

Speaker: Colin Davis (MIT)
Title: English Possessor Extraction and Linearization
Date and time: Tuesday February 14, 1-2pm
Location: 32-D461
Abstract:
Received wisdom tells us that in English, wh-movement of a possessor requires pied-piping of the whole DP containing that wh-possessor, as in (1). This falls under the Left Branch Condition of Ross (1967), describing a ban on moving the leftmost element of a nominal phrase in languages like English. Unexpected in light of this generalization is the fact that for some English speakers, wh-movement of just the possessor, stranding the DP it originated in, is also possible. This is possessor extraction (PE), as in (2):

  1. [Whose fat cat] do they think [t is cute]? (Pied-piping)
  2. Who do they think [[t’s fat cat] is cute]? (Possessor extraction)

A classic example of a PE language is Hungarian (Szabolcsi 1984), some others are Chamorro, Tzotzil, and much of Slavic. However, English has never been recognized as a PE language as far as I know, though in the course of a study of child English Gavruseva & Thornton (2001) get some adult English PE data, and take it to be a production error. I show that to the contrary, English PE is a productive and interestingly constrained phenomenon. An example of such a constraint is the fact that PE out of an in-situ object is impossible, as in (3). PE out of the embedded object in (3) can be rescued, however, if the residue of the DP where the possessor was born is pied-piped/moved to the edge of the embedded clause, as in (4):

  1. * Who do you think [John likes [t’s cake]]? (No PE from object in-situ)
  2. Who do you think[[t’s cake] John likes t]? (PE from pied-piped object)

In the context of a Cyclic Linearization framework (Fox & Pesetsky 2005), I argue that some movements independently necessary for coherent linearization in PE contexts are in conflict with a PF constraint which, roughly speaking, requires adjacency between a (moving) possessor and the saxon genitive ‘s at the phase level. My ambition is to show that the quirks of English PE are an automatic consequence of this tension and the methods of its resolution. I argue that the unique pied-piping in (4) and a family of similar examples is one way of resolving this tension, while in some structures there is no possible repair, ruling out PE in those contexts.