Semantics And Syntax Of Event Classifiers In Mandarin Chinese

In Mandarin Chinese, event classifier is an important component of the system of classifiers. Traditionally, event classifier is use to count actions/activities semantically and to syntactically function as Buyu. This thesis studies the lexical semantics, syntactic properties and distribution of event classifiers within Level-mapping approach, and then provides some ideas different to the traditional ones.Chapter 1 concentrates on the theoretical background. Unlike Transformational Grammar which is derivational, Level-mapping approach is constraint-based. Within this approach, each level of grammar is independent to others. The representations on each level are generated autonomously and are linked by interface rules.Chapter 2 concentrates on the lexical properties of event classifiers, including their semantics and syntactic characteristic. The meaning of an event classifier is to individuate an event kind/subkind. It is the numeral which bears the information of amount. The event classifiers can be divided into two groups with respect to their individuation manners. Syntactically, an event classifier has its own subcategorization frame, in which a event-kind/subkind expression can be the complement and a numeral can be the specifier. Event classifiers is parallel with Entity classifiers on their lexical properties. The only difference between the two sort of classifiers is what are individuated.Chapter 3 discusses the distribution of event classifier phrases(CLVP) in larger circumstances. The CLVPs after verbs can function as objects or predicates but not Buyu. And the CLVPs before verbs function as adverbials or the heads of predicates. Another three kinds of form-meaning mismatch are discussed in this chapter. It is concluded that the solution of these phenomena from Level-mapping approach is better than the solution from Transformational Grammar.Chapter 4 discusses the cross-linguistic distribution of event classifiers and gets us some typological generalizations. These generalizations mean that the many characteristics of event classifiers are determined by entity classifiers and that there is highly tight relationship between objects and CLVPs in the same VPs.