The Study Of Verb-Object Construction In Machine Translation

Translation is a complicated process. The first thing to do in translation is to understand the meaning of the text in source language (SL) by means of semantics, syntactic and pragmatics. Apart from that, it is also necessary to under the knowledge of SL; the knowledge of the target language (TL);the contrastive linguistics of SL & TL; the general knowledge and the knowledge of the field concerned and the cultural knowledge of both languages. But no matter how difficult it is, our computer engineer and the linguists still keep on their hard work to enhance the performance of the machine translation system so as to eliminate the language barriers. It is the only way to survive in this hi-tech era and the speed of globalization. It is hoped that there will be a day that language will no longer be a problem in communication among people in the world.However, in view of the current popular online machine translation systems, they are all far from satisfactory especially in Chinese-English machine translation. It is not only due to the design of the machine translation systems; one of the main reasons rests on the inadequacy in the study of Chinese grammar. Under this circumstance, we would like to pick up an important construction in Chinese language, ie verb-object construction (V-0 construction), to have a detailed study.The first objective of this thesis is to re-categorize the V-0 construction. We are going to look into this construction from both intra-language and inter-language prospective and then we'll re-categorize different types of V-O constructions under 4 different groupings. They are core grammar, sub-core grammar, sub-peripheral grammar and peripheral grammar. For those which fall into the core grammar group will be those V-O constructions having strong analogical ability and high accuracy in word-for-word translation. Those fall into the peripheral grammar category are V-0 constructions which are not possible to perform the analogy and also word for word translation. The rest of the V-0 constructions which lie between the core and the peripheral grammar belongs to the sub-core or sub-peripheral groupings. The second task of this thesis is to help establish various dictionaries or data bases based upon the above categorization. It is hoped that these dictionaries can eventually help improve the efficiency and efficacy of the machine translation systems.There are altogether 8 chapters in thesis. Chapter 1 is an introduction of the study. Firstly, we'll have an overview of the categorization methods suggested by various scholars in V-0 construction. The advantages and disadvantages of these groupings are discussed. Then, we'll introduce Chomsky's core grammar theory in Universal Grammar, the Construction Grammar theory of Goldberg & Jackendoff and the theory of Core and Peripheral Grammar in re-categorization of V-O constructions suggested by Dr WU Zhen-guo. All these theories formed the basis of our study to establish a new grouping system for V-O constructions and also set up clear cut rules for the verbs and objects which commonly co-occur in order to facilitate Chinese-English machine translation as far as possible.In Chapter 2, there is an attempt to distinguish core grammar and peripheral grammar by means of the theories of UG and CG as well as the acquisition of first and second language. Also, we are going to study the types of peripheral grammar formed such as special words/phrases, constructions formed by agglutination, special constructions formed by special rules and regulations. Apart from the above mentioned, we will also focus on the importance of re-categorization of V-O constructions in language study, language teaching, edition of dictionaries and information management.In Chapter 3, we will study the methods to distinguish core grammar and peripheral grammar in V-O constructions. In Chapter 4, the results of 3 online machine translation systems in translating V-O constructions will be brought up for detailed analysis. The inadequacies will be carefully studied.After studying the problems in V-O construction translation by the three online machine translation systems, in Chapter 5 and 6, we are going to look for solutions. The first thing to do is to put the different types of V-O constructions into 4 different categories as mentioned above and then we can establish rules and regulations for translation if required. For example, we can edit a dictionary particularly for the idioms with V-O constructions as they are being categorized under the peripheral grammar. This type of V-O constructions is not possible to be handled by word-for-word translation as these constructions are usually having a concrete meaning which cannot be understood by studying the surface meaning of the words. Or, we can set up a base with translation rules for retrieval during the Chinese English translation process. In Chapter 7,we are going to imitate the running of the machine translation system and implement the rules and setting as we discussed in Chapter 5 and 6 in order to have a preliminary test of the feasibility. Then we will have a conclusion of what we have found and established. Also, we will point out the inadequacy of our study for further improvement in the near future.