A Corpus-based Study On Discourse Analysis Of Netease Open Course

This thesis is a corpus-based discourse analysis of Harvard University’s open courses, taking Justice from NetEase for a case study. It aims at using a corpus in native English classroom discourse from the world-class Harvard University to reveal and verify the core features of cohesion and coherence, contextual features and interaction of discourses. It is expected that findings of the research may provide inspirations for the discourse studies and applications of China’s English classroom teaching and bilingual teaching.This study is based on a self-constructed corpus:corpus of Harvard University’s Justice (CHUJ). Texts collected into corpus come from the subtitles of NetEase open course videos. TXT-format linguistic data are obtained through transferring the format and repeated checking. This corpus contains78,723tokens with12separated classes. With the aid of corpus analytical software AntConc3.2.1, quantitative and qualitative approaches are used to explore the classroom discourse of Harvard University’s Justice by extracting and studying the small words, sorting and observing the high-frequency words, analyzing and interpreting the collocations and clusters of the key words, as well as extracting and studying the discourse management chunks.Main findings are arrived at that classroom discourse of Harvard University’s Justice not only concentrates on its cohesion and coherence, but also highlights the word-using specificity of classroom discourse. In addition, the key words extracted from the context are able to reflect the theme of the course, and the word-using is in accordance with the philosophical issue of Justice by extracting and analyzing the collocations and clusters of the key words. The discourse content tallies with the teaching goals. It can be inferred that heuristic teaching is applied with the use of high-frequency word of "what" to a certain extent. Last but not least, the interaction of classroom discourse is verified by extracting and showing the discourse management chunks.This study is a corpus-based empirical one, which might bring about the problems of subjectivity and one-sidedness towards data interpretation. It is hoped to bring an inspiration for people engaging in similar studies.