Ergative Construction In Chinese-A Funtional Approach

This dissertation investigates the ergative phenomenon from the perspective of functional grammar. It begins with a comprehensive description of different approaches in the studies of ergativity, and subsequently ventures an improvement of the ergativity analysis in the functional grammar. Based on this improvement, a new system of ergativity analysis is established and thereafter applied to deal with the ergative constructions in Mandarin Chinese.The study of ergativity was originally centered on "ergative languages", which, typically represented by Dyirbal, has attracted the attention of linguists with its unique structural characteristic. Such languages are characterized by an identical case form assigned to the objects of transitive verbs and to the subjects of intransitive verbs. The subjects of transitive verbs are in a different case. In English and other "non-ergative languages", the subjects of both transitive and intransitive verbs are identical in case form and the objects of transitive verbs are in a different case.However, there exists in English and other "non-ergative languages" a phenomenon that some the object of some transitive verbs can function as its subject when they are used as intransitive. The existence of such constructions widens the scope of ergativity. And the study of "ergativity in non-ergative languages" has begun to gain its momentum.The ergativity concept in functional grammar is not about "ergative language". Halliday considers "ergativity analysis" as complementary to "transitivity analysis" and holds that ergativity analysis can be applied to any process. In a transitivity analysis the processes are viewed from the perspective of whether an action is extended to another participant, whereas in an ergativity analysis they are viewed from the perspective of whether a process happens all by itself or it is triggered by some outside factor. According to Halliday, a process with an outside trigger is "ergative", otherwise it is "unergative".We propose that the key difference between ergative construction and transitive construction is whether the outside factor (Agent) is indispensable in the process. In a transitive process, the Agent is the doer of an action that extends to a Patient. The process is impossible without the Agent. In an ergative process, the outside factor is only a trigger, the core participant is the Medium. Therefore, a process in which the outside factor (Agent) is not indispensable is ergative; if the outside factor is indispensable, it is transitive; if there is no outside factor, it is intransitive.Based on the above, we propose a new classification of the processes. We hold that all processes can be classified into ergative and unergative processes. The unergative processes can be subdivided into transitive and intransitive. In ergative processes, if the outside factor appears, it is Full Ergative Process; if the outside factor does not appear, it is Agentless Ergative Process. If the appearance of an outside factor alters the meaning of an Agentless Ergative Process, then it is a Marginal Ergative Process; otherwise it is a Proper Ergative Process. Our definition of Marginal Ergative Process originates from Halliday's "Middle Voice". According to Halliday, A middle construction has a Patientive subject but the verb is in active voice, which is identical with Agentless Ergative Process. But when an outside Agent appears in the subject position and thus puts the Patient in the object position, the meaning changes, which is different from a Full Ergative Process. Based on this observation we believe that the middle construction should be a process in between ergative and unergative.Applying this new ergative analysis system, we look into the ergative constructions in Mandarin Chinese.In Chinese, Proper Ergative construction always appears with a verb that has perfective suffix -le or -guo or a verb that indicates stative action (with verb suffix -zhe ), Marginal Ergative construction always appears with Modal verb constructions (neng, keyi, etc.) This feature is coherent with Trask's diachronic study of Ergative languages in which the origin of ergativity is dichotomized into ergative as perfective and ergative as passive.