Faculty member (and triple alumna) Susana Béjar will be giving an invited talk for our department on Friday, January 22. Susana earned her Ph.D. from our department in 2003, and has since been a postdoc at McGill and an assistant professor at both our Scarborough and St. George campuses. Her research interests are centered on syntax, particularly the relationship between structures and features. She has worked on agreement, concord, Case-licensing, phi-features, and more, and is also interested in typology, allomorphy, and language acquisition. Her talk, "Copular agreement is nominal inflection: Evidence from unrestricted persons in specificational clauses", will begin at 3 PM on Friday and be followed by a reception in the department lounge.
Specificational sentences present a puzzle for the literature on person agreement restrictions. They are widely agreed to be inverted structures, where the 'notional' subject (italicized you in (1)) surfaces low in the clause, in a non-canonical inner subject position (Moro 1997, den Dikken 2007, inter alia).
(1) [[The person I want to see] is you]
Herein lies the puzzle: cross-linguistically, inverted structures generally share a striking feature: the low subject position is a restricted one; 1st and 2nd person pronouns are often excluded from appearing there and to the extent that they can appear, they are famously only able to control number agreement, not person agreement. Prominent accounts of these patterns relate the restrictions to a generalized pattern of licensing failure that arises when an agreement domain must be shared by more than one DP (Anagnostopoulou 2003, Bejar and Rezac 2003). Specificational clauses would seem to fit the bill, as they are non-transitive clauses that introduce two DPs in the agreement domain of T. Yet, across languages, 1st and 2nd person pronouns appear to be uniformly available in the inner subject position; in many languages, they also control person agreement on the copula. With special attention to Georgian, I argue that this striking departure from what are otherwise robust generalizations signals a fundamental difference between the syntax of copular inflection and that of verbal inflection; specifically, I propose that the proper locus of copular agreement is not a φ-probe on T but rather a φ-probe on the head/label (D) of a complex DP (below T) that introduces both the notional subject and the (inverted) copular complement. Cyclic Agree (Bejar & Rezac 2009) determines that the probe on D invariably agrees with the notional subject in specificational structures, due to the special syntax of the complement DP which, for independent reasons, must be a (sometimes reduced) free relative clause (cf. Higgins 1973, Heycock and Kroch 1996). The complement, inverted, moves to Spec, TP; valued D cliticizes to T, giving the appearance of T-agreement with the inner subject. Support for the analysis comes from various domains, including a challenging Georgian agreement pattern, generalized connectivity effects in specificational sentences, nominalization patterns and the cross-linguistically common paradigmatic correspondence between forms that realize copular agreement and those that realize possessive morphology (den Dikken 1998, 1999 Bernstein & Tortora 2005, Oxford 2014). The analysis opens up new possibilities for explaining other recalcitrant puzzles in the syntax of copular agreement systems.