A Multi-dimensional Approach To Reeister Variation In Mandarin Chinese

Traditionally, the study of registers and register variation falls on the scope of sociolinguistics and are usually intuitive on a qualitative basis. With the rapid development of computer science, corpus emerges as a promising tool to analyze language from a quantitative perspective.Douglas Biber (1988) firstly proposed the Multi-dimensional approach to the research on register variation, which aims to simplify the complex interactional pattern of large numbers of linguistic features within the corpus on a large scale based on quantitative test, factor analysis in particular. It tries to elicit the co-occurrence pattern of linguistic features and the communicative functions they are to perform as an integral. While Multi-dimensional approach has been employed to explore language variation in and across English, Somali, Nukulaelae Tuvaluan, Korean and Spanish, little research has been done on register variation in Mandarin Chinese on a full scale. Yet Chinese as the mother tongue mostly used around the world makes the research on its register variation necessary. This study aims to investigate register variation in spoken and written Mandarin Chinese from a macroscopic perspective. In the present research, we construct a structurally-built corpus composed of sixteen Chinese spoken and written registers. There are1000texts of1000-word in length for spoken and written respectively. Eighty-eight linguistic features have been marked and their frequencies counted in the corpus. The dataset is examined using Multi-dimensional approach to gain a comprehensive picture of the co-occurrence patterns of linguistic features. Five underlying dimensions of register variation are extracted and identified after factor analysis. They are interactive vs. informational discourse, literary vs. non-literary concern, colloquialized expression with subjective emphases, situation-dependent reference&emotional concern, persuasion and argumentation vs. non-persuasive and non-argumentative concern, respectively. Linguistic features clustering in each dimension are firstly interpreted in a tentative way according to the communicative functions they might perform. The interpretation is further justified by computing the factor score for each register on each dimension as supplementary evidence. From macroscopic and quantitative perspectives, a distinctively gradual change among spoken and written registers of Mandarin Chinese on each dimension has been proved in the current research.