A Computational Linguistic Research On Writer Identity In English Abstracts Written By Chinese, American, And Korean Scientists

Identity is who and what you are. It refers to the features of a person or group that make them different from others. In the field of Genre and Genre analysis, writer identity is not the "state" of being a particular person but a "process". During the process of constructing their identities, writers show the combination of "individuality" and "community" in their academic discourses. This paper studies writer identity in English abstracts written by Chinese, American, and Korean scientists and probes into the genre awareness, concept construction, and language construction during the identity construction process.Based on Hyland’s genre analysis theories, this paper analyses writer identity from the computational linguistic angle using two NLP tools, Coh-metrix and Gramulator. The three corpora of this study include 1007 English abstracts collected from 46 Chinese, American and Korean science journals in the recent 5 years. In the first part of this study, the theoretical framework for analyzing writer identity is established from the three aspects of genre awareness, concept construction, and language construction. Then I used two NLP tools, Coh-Metrix and Gramulator, to analyze the specific and distinctive features in the five dimensions, i.e. narrativity, coherence, syntax, lexicon, and semantics. The features in these five dimensions are further compared and analyzed among native English writers and non-native English writers, as well as between non-native English writers themselves. The main findings of this research are as follows:1) On the genre awareness aspect, Chinese, American and Korean writers share the common identity as scientists. They are fully aware of the informative genre of English abstracts on academic journals and the narrativity in their abstracts is extremely low in Coh-Metrix scores, which shows the community of their professional identity. On the concept construction aspect, scientists from the three countries demonstrate similar Coh-metrix scores in global coherence. This result reflects the stability of cognitive network composed of concepts and relationships in scientific journal articles. The homogeneity of the deep cognitive network also indicates the psychological similarities of human beings in spite of different disciplines. On the language construction aspect, all scientists from the three countries demonstrate similarities on the use of attitude markers. But 10%-20% of the writers used positive attitude adjectives to express the willingness to share their views with readers in order to obtain acceptance from them.2) Chinese, American and Korean scientists demonstrate unique writer identity in their English abstracts. The distinctive features of American abstracts show the combination of creativity and community in writer identity. American writers highlight the subjectivity as researchers and the originality and value of their research. At the same time they interact with the potential readers using engagement markers to obtain common ground and acceptance from their readers. The distinctive features of Chinese abstracts reveal community of Chinese writer identity. They foreground the research findings to the maximum extent and withheld the identity as researchers. They highlight the objectivity of findings and show respect to authorities and peers. The distinctive features of Korean abstracts suggest the recessive demand to show creativity in their writer identity. Koreans scientists show relatively less community in their abstracts compared with American and Chinese colleagues. They use highest frequency of negative structures to show the value of their research and express the willingness to interact with the potential readers. But similar to Chinese scientists, Korean writers also foreground the research findings and avoid showing subjectivity.3) The distinctive features to show unique writer identity among native English writers and non-native English writers, as well as between non-native English writers themselves, are mainly found on local coherence of the concept construction and the language construction aspects.The theoretical and practical contributions of this study consist of the following aspects:1) This paper reclassified the research framework of writer identity and explored the three aspects of identity construction process, i.e. genre awareness, concept construction and language construction.2) The researcher used two NLP tools, Coh-Metrix and Gramulator, to comprehensively investigate and analyze the specific and distinctive features of scientific abstracts. The validation study of the Gramulator is also done to increase the credibility of this research.3) The findings shed light on the understanding and investigation of writer identity among Chinese, American and Korean scientists, which provide some evidence for the future data mining and author identification study.