The Weekly Newsletter of MIT Linguistics

Syntax Square 10/17 - Nico Baier

Speaker: Nico Baier (UC Berkeley)
Title: Unifying Anti-Agreement and Wh-Agreement
Date/Time: Monday, Oct. 17, 1-2pm
Location: 32-D461

In many languages, phi-agreement is sensitive to the A’-movement of its controller. Some languages, such as Abaza, exhibit ‘wh-agreement’, an effect in which dedicated agreement morphology cross-references extracted arguments (Chung and Georgopoulos 1988). In other languages, such as Tarifit Berber, extracted arguments cannot control full agreement. This is known as ‘anti-agreement’ (Ouhalla 1993). These two effects have previously been treated as distinct. Wh-agreement is viewed as normal result of Agree with a goal bearing a wh-feature (Georgopoulos 1991, Watanabe 1996, a.o.). Anti-agreement is generally taken to reflect a disruption of agreement in the syntax proper (Schneider-Zioga 2007, Ouhalla 1993, a.o.). In this paper, I argue that this traditional wisdom is incorrect and that wh-agreement and anti-agreement are in fact two instantiations of the same phenomenon. Both effects are the result of a phi-probe copying both phi- and wh-features from a goal. Patterns of anti-agreement and wh-agreement arise when partial or total impoverishment applies to the [phi+wh] feature bundle in the morphological component, blocking insertion of an otherwise appropriate, more highly specified agreement exponent.