A quantitative model in dialect subgrouping: The case of modern Wu dialects.

摘要

This thesis is intended to present a new model for the comparative study of dialects with the quantitative measurements, and under this new model, the subgrouping of the Modern Wu dialects, one of the major groups of Chinese dialects, is carried out, based on the syllable initials.;An overview is given on the previous approaches in the literature of the dialectology--namely, the traditional approach, structural approach and generative approach. Each approach has been applied to the case of the Wu dialects to accomplish the task of subgrouping. But none of them can overcome the difficulty of synthesizing the isoglosses, since they ignore the concrete processes of sound changes, and thus fails to determine the degree of associations among a number of related dialects. To study the sound change in process, based on the theory of lexical diffusion, initials a new route for quantitative study into dialect comparison.;Following the methodology in Hsieh (1977), Cheng (1982), Lu and Cheng (1985), Cheng (1986, 1988), Lu (1986a, 1987a, 1987b, 1987c, 1988, 1989, 1990), the thesis proposes a new model, namely, the systematic quantitative comparative model. Since sound change is carried out by means of lexical items, the comparison of quantity of the influenced lexical items in the course of sound change will reveal the degree and scope of sound change in process. Under this new model, the syllable initials of more than 2,700 characters pronounced in 33 localities of the Wu dialects are collected and grouped with the reference of the Ancient Chinese to provide a large database for the systematic comparison. Then, the procedure designed for dialect comparison, which includes correlational analysis, cluster analysis and principal components analysis, is followed to carry out the subgrouping of these 33 localities. The results are presented and discussed.