Topic Identification In The Discourse Of English For Science And Technology

The issue of discourse topic identification arises from an attempt to cope with the non-native English reader's difficulty in reading the EST (English for Science and Technology) text. Nowadays, international communications play an important role for the development of a country, especially the communication on science and technology. And it is widely acknowledged that English is the international language. Yet a reader may feel hard in comprehending the EST text. It is to help a reader get the central idea of the text that this thesis makes an attempt to analyze the EST texts with the use of RST (Rhetorical Structure Theory). The thesis hypothesized that: 1) the EST text, as an informative text genre in contrast to the literary text, is suitable to be analyzed with RST; 2) the topic location in the tree diagram of EST discourse accords with the intuitive understanding of what the topic is. The analysis result satisfies the two hypotheses.The thesis proceeds as follows:(i) Topic issue is a hot topic in linguistic field. The persisting interest is certainly an indicative of the importance of this area. Yet it is still doubtful whether the terms as used have acquired a precise meaning. Based on the widely acknowledged "three levels of grammatical study" (Fan, 1996), this thesis establishes a new way to approach topic from three perspectives: 1) sentence topic on the level of syntax, 2) discourse topic on the level of semantics, and 3) the speaker's topic on the level of pragmatics. This thesis focuses on "discourse topic" (the central idea of a text). In the informative text discourse topic appears as an explicate element. What this thesis concerns is the way to identify the topic of EST text. According to Scott (1997), the discourse topic can be identified in ways similar to bottom-up and top-down processing. It is preferable totake the top-down perspective for the longer written discourse.(ii) The Rhetorical Structure Theory (RST), one of the theories for macro-structure discourse analysis, is of particular relevance in top-down perspective here. One of the most significant contributions of RST analysis is to locate the topic of the text in the text tree diagram. RST analysis is to break the text down into text spans, and to find an RST relation that connects each pair of spans until all pairs are accounted for. RST uses a number of predefined rhetorical relations. These relations are defined functionally, in terms of what their intended effect on the reader is. A tree diagram of the text is presented after RST analysis. And the topic lies in the nucleus span where no . more vertical line goes downward from the top level of the tree diagram.(iii) Three texts of EST are taken randomly from the English journals, which are in 200—500 words. They are all written by native English speakers, with Text 1 from BBC English, Text 2 from The Futurist and Text 3 from Science Digest. Its readability is moderate in difficulty, which is suitable to be read by the university students in higher grade. The texts are typical readings on science and technology that any common reader may take up. An RST analysis is to present the tree diagram of the text in which the topic is located. Then, an experiment is designed to conform the scientific status of the above-mentioned method. The subjects are four postgraduates who major in English and have passed the TEM-8. They are asked to identify the topic of the texts after reading. As a result, the topics being located in the tree diagram agree to the results of the experiment.This thesis has briefly reviewed what has been done in the literature of topic studies and has confirmed a new way to identify topic of EST text through RST analysis. The attempt to identify discourse topic through a formally structure analysis is to bridge the meaning and form of language, which has always been the focus and nodus of linguists. Thus, it not only has theoretical significance but sheds light for macro-discourse analysis as well, even may attribute to the computer discourse processing a lot.