Hadas Kotek will be presenting joint work with Michael Erlewine at the first Ling-Tea of the Fall 2015 semester. This will be presenting a practice talk for Sinn und Beteutung. If you would like to present at a future Ling-Tea, please email Colin Brown!
Coordinates: Tuesday 9/1, 1:00pm–2:00pm in Linguistics 117
Title: Relative pronoun pied-piping, the structure of which informs the analysis of relative clauses
In this talk we argue that restrictive and non-restrictive relative clauses have a fundamentally different semantic interpretation: while restrictive relatives are property-denoting, non-restrictive relatives are proposition-denoting. The evidence motivating this claim comes from a consideration of Relative Pronoun Pied-Piping, concentrating specifically on the behavior of intervention effects inside RPPP in non-restrictive relatives. Intervention effects have been shown in previous work to distinguish regions of Rooth-Hamblin alternative computation from regions where covert movement has taken place (Sauerland & Heck 2003, Cable 2007; 2010, Kotek & Erlewine, to appear, Erlewine & Kotek 2014). We show that RPPP, like pied-piping in wh-questions, is sensitive to intervention effects.We propose that while a movement strategy is used for the interpretation of restrictive RCs, in non-restrictive relatives, relative pronouns are interpreted in-situ within the pied-piped constituent at LF, using Rooth-Hamblin alternative computation. This bring RPPP in line with other instances of pied-piping in wh-questions and focus constructions, allowing for a uniform semantics for pied-piping (cf Cable 2007; 2010). To derive the non-intersective meaning of the relative clause, we propose that the relative pronoun projects a singleton alternative set, corresponding to the RC’s antecedent. Following Demirdache (1991), we treat this antecedent as an E-type anaphor. This allows us to directly derive the proposition denoted by the RC, without first computing the corresponding property, as is done in the case of restrictive relatives. This proposal helps explain two differences between restrictive and non-restrictive relatives in English. First, why non-restrictive relatives can only be constructed using relative pronouns, whereas restrictive relatives also have a that/∅ complementizer option. Second, why RPPP in these RCs can be substantially larger than in restrictive RCs, and why only restrictive RCs are sensitive to islands.