C Plus T As A Necessary Condition For Pro-Drop: Evidence from Code-Switching

The current study investigates the grammatical conditions that allow the occurrence of pro-drop in the discourse. There exist two leading proposals in the literature: the T-head proposals (see Rizzi, 1986; Jaeggli and Safir, 1989; Speas, 1994; Rohrbacher, 1999; Müller, 2006; Camacho, 2013) and the C-domain proposals (see Frascarelli, 2007 and Sigurðsson, 2011). Whereas the T-head proposals maintain that pro-drop is contingent on a “strong” T-head, or strong phi-features, the C-domain proposals argue that pro is interpreted in relation with the closest Topic, located in the C-domain. Preliminary consultant data from code-switching (CS) and results from two CS pilot studies empirically show that both the T-head and the C-domain are involved in the occurrence of pro-drop. In light of these findings, this Dissertation contributes to the discussion by proposing the C and T Hypothesis, by which the availability of a null subject in the discourse is determined via an interplay between the C-domain and the T-head. This hypothesis highlights the value that CS has in shedding light on linguistic theory (see González-Vilbazo et al., 2013; González-Vilbazo and López, 2012, 2013; Ebert, 2014; Koronkiewicz, 2014; Vergara, 2017; Hoot and Ebert, 2018, De Nicolás and Robledo, 2018). In order to see which of the three hypotheses makes the correct predictions, a CS experiment and a data survey of the CS literature were conducted. My analysis of the results suggests that an occurrence of pro is not possible when either the C or the T-head of a linguistic structure does not have the features necessary to allow for pro-drop. Subsequently, I propose an analysis that specifically defines what the roles of the C-domain and the T-head are in an occurrence of pro-drop: the T-head has a facilitative role in pro-drop; it decides whether pro-drop is grammatically possible, and C links a feature bundle with no phonological exponent to its true antecedent in the preceding discourse. Finally, I conclude that the occurrence of a NS in the discourse is no different from the occurrence of an overt subject. Thus, the proposed analysis does not devise extra machinery merely intended to account for the case pro-drop. Instead, the analysis can account for both null and overt subjects in monolingual and code-switched structures.