Causative Constructions In Mandarin Chinese

Basing on Minimalist Program, I study many causative constructions in Mandarin Chinese, such as BA Constructions, DE Constructions, Verb-copying Constructions, Resultative Constructions, Ditransitive Constructions and Dative Constructions. However, the notion construction has no independent status in my framework.To economically derive these narrow causative constructions, I argue that the word order is [S h [C H]] in natural language, which is that lexical heads are head-final and functional heads are head-initial, and the merger order of argumental elements are determined not by thematical hierarchy but by functional head. What's more, thematical hierarchies advocated by many different scholars might be derived in different stages of language computation. So, functional heads play important roles in my theory. I propose a Functional Category Hypothesis, which says functional heads not only motivate movement, but also determine merger, including argument selection and semantical role assignment.To derive the word order in natural language, I propose Same Direction between Adjunction and Movement Hypothesis. This hypothesis is a useful tool, which can help me detect some effects of earlier researches proposed by some linguistics. We use the hypothesis to find that DE is higher than V and NONG is not a causative verb (cf: Gu, 1992). These findings active me to discover some important phenomena behind them. For example, when DE is higher than V, then DE must be in a light verb position, and then it must refer to 'become', based on Predicate Decomposition. For DE has phoneticalisation and a syntactic position, and its meaning is 'become' according to Wang Li, so I argue that 'become' is a syntactic notion (to avoid confusion, I use Bee to represent it) and, so, Bee is a functional head. Also I argue that Causative is a syntactic notion and Caus is a functional head. Caus is phoneticalised by BA in Mandarin Chinese. This can be proved by the behaviors of BA Constructions and other evidences proposed by Lu and Song, among others. There are some advantages, when Caus is treated as syntactic notion, for example, it can select as causer any arguments which can posit in [Spec, CausP], such as agent or patient of matrix verb, or even a event. Because Caus selects an event as its argument, the Verb-copying Constructions might be construed easily. Because it can select agent as its argument, agent and some other causer can be posited in the same place at outer merge. For some other reason, agent can move out to a higher functional head, this can explain many syntactic behaviors.Frankly, semantic theories such as Lexical Conceptual Structure, Event Structure and Predicate Decomposition active and guide my research. But I don't want to establish syntactics above semantics. I want to keep syntax more automatic. So, in my research, Cause and Become and many other functional heads are treated as syntactic notion.