A Syntactic Study On Empty Functional Categories

Empty categories (ECs), proposed by Chomsky during the late 1970s and early 1980s, are an important issue in the theory of generative grammar. The research on ECs is one of the focal topics in current theoretical linguistics since it is quite enlightening and helpful for making further efforts to determine properties of syntactic and semantic representations and the generative rules of them.So far, most researches on ECs are involved in the field of lexical categories, but those in the functional categories field are few, even no research has been made at home. Hence, within the framework of Chomsky's ECs, based on Radford (2000,2004,2006), this thesis attempts to show that the phenomena of ECs exist not only in lexical categories field, but also in functional categories field, e.g., empty tenses (Ts), empty complementisers (Cs) and empty determiners (Ds). There are both empirical and theoretical evidence for the existence of empty functional categories. The thesis attempts to explore that the assumption that empty functional categories exist in languages is supported both empirically and theoretically, and to make further efforts to analyze the explanation power of three common types of empty functional categories (empty Ts, empty Cs and empty Ds) and their significant contributions.According to the Split INFL Hypothesis presented by Pollock in 1989, we often use [+tense] and [-tense] to express the characters of tensed clause and tenseless clause. In fact, INFL, a part of sentences, is replaced by T (tense) in MP. This thesis claims that we use the definition of empty category when we are constructing the theory of T. The tenseless clauses exactly imply the phenomenon of empty category, so we can infer that the T here is an empty category. The analysis of empty T enables us to arrive at a uniform characterization of the syntax of (finite and infinitival) clauses as TP structures headed by a T with a V or VP complement. In general, the assumption of empty Ts helps us to optimize and unify the X'-theory, and makes the X'-theory have more explanation power and become more systematic.Chomsky (1995,1998) makes an assumption in recent works that all canonical (i.e., normal) clauses are complementiser phrases (CPs), and many linguists support his point. Based on it, we prove the assumption of the existence of empty Cs, and demonstrate it via some illustrations. The assumption of empty Cs not only is the forcible argument in favor of the operation of syntax, but also enables us to make a uniform explanation for the interrogative sentences.Building on the terminology suggested by Jane Grimshaw (1991), we can infer that N has an extended projection into determiner phrase (DP) in the same way as V has an extended projection into IP (i.e., IP is an extended projection of V, while DP in an extended projection of N). According to Radford (1997), we can draw a conclusion that the bare nominal structures are DP structures just as the infinitive clauses are IP or TP structures, and both of their heads are ECs. Based on Abney's (1987) DP hypothesis, it is concluded that nominals modified by an overt D are DPs, bare nominals are DPs headed by an empty D and all nominal and pronominal arguments are DP constituents headed by an overt or empty D constituent. This thesis concludes that the assumption of empty Ds has two important significances. Firstly, it enables us to unify the structures of all the nominal phrases. Secondly, revealing the uniformity of the nominals and the sentences leads us towards the goal of arriving at a unitary characterization of the nominals and sentences, as well as unifying the theory of Universal Grammar (UG) and making it briefer.This thesis aims to systematically summarize and analyze the phenomena of the empty functional categories, show that such ECs do exist in certain constructions and the assumption for the existence of empty functional categories is supported both empirically and theoretically, and make further efforts to explore the EC's syntactic and semantic representations as well as each generative rule. Furthermore, it can help us to explain more language phenomena, to perfect the theory of ECs, and to demonstrate Chomsky's UG.