Hyper-raising under control

Abstract

Hyper-raising (HR) is a cross-clausal dependency where the DP subject of a finite complement clause surfaces as an apparent argument of the matrix verb (Alboiu and Hill, 2016; Carstens, 2011; Ferreira, 2004; Fong, 2017; Halpert, 2019; Lee and Yip, 2020; Nunes, 2008; Ura, 1994, 1995). HR is puzzling because it involves apparent DP raising from a finite complement, even though the driving mechanism of raising - the non-finiteness of the complement clause - is absent. Nonetheless, the majority of the existing literature argues that HR is, like raising, A-movement. In this thesis, I present novel data on HR from Brazilian Portuguese that contradicts these analyses. Specifically, I show that HR displays signature properties of a control dependency. Thus, I argue that HR is not an instance of A-movement, but an instance of control into a finite complement clause. I demonstrate a technical implementation of the Agree Model of Control (Landau, 2000, 2004, 2006, 2007, 2008, 2010, 2013) in the context of HR in Brazilian Portuguese that accounts for the relevant data while also providing novel predictions and extensions.