Subjects and arguments in a-syntax

In this essay, I will examine some murky questions concerning the theory of A’-movement in Italian. I will focus on two main empirical problems: the behaviour of Criterial Subjects (Rizzi 2006) and the syntax of multiple wh-questions in Italian. Both these domain of inquiry pose some questions about the restrictions that A’-movement has to respect and the consequences that these restrictions have on the superficial form of languages. The aim of this study is to show that many idiosyncratic properties concerning these two empirical domains can be explained by a rather simple theory of syntactic computation. The general picture that will emerge is that the syntactic computation in many A’- constructions can be described as a quite simple device, which is limited by (rather) independent interface requirements (Chomsky 1995 and subsequent works) and by some structural characteristics of the left periphery (Rizzi 1997, Cinque & Rizzi 2008). Moreover, it will emerge that the syntactic computation can make use of different strategies to circumvent these restrictions. In Chapter 2, I will go through a long-standing problem concerning the position of subjects in a number of A’-constructions in Romance languages. I will show, that some puzzling restrictions on the position of subjects can be explained adopting two basic elements: a feature-based theory of locality (Starke 2001, Rizzi 2004, Abels 2012) and a quantificational theory of Criterial Subjects (Bianchi & Chesi, to appear). I will argue that it is possible to derive a wide empirical range of data in a number of Romance Languages, from these two basic ingredients. In Chapter 3, I will examine a rather neglected group of Italian constructions, namely Multiple wh-questions. I will show that Italian displays a rather complex behaviour with respect to these structures, much more complex than previously thought. The idea is that 6 the whole pattern of data can be derived by two simple restrictions on A’-constructions. On the one hand, Italian is restricted to have only one position for wh-phrases in the left periphery of the clause (Rizzi 1997); on the other hand, only a sub-class of wh-phrases can be interpreted without being moved, namely argumental wh-phrases (Reinhardt 1997, 1998). I will argue that these two well-known restrictions force the syntactic computation to employ different circumvention strategies, from which the complex pattern of data emerges. I will argue that coordination turns out to be a flexible tool that syntax employs to build multiple wh-questions in compliance with the restrictions mentioned above.