古全浊声母今读的聚类分析和主坐标分析

Present pronunciations of the ancient complete voiced initials in the existing Chinese dialects in the perspective of the cluster analysis and principal coordinates analysis

以26个汉语方言点古全浊声母的今读为样品,用NTSYS进行聚类分析和主坐标分析,并讨论相关的问题。结论是:(1)聚类分析结果不一定能直接作为一种合理的分类方案;(2)古全浊声母演变的特定方式跟大方言之间不存在一对一的关系;(3)汉语方言基本特征库建设的困难主要在于转写特征时如何保证足够的专家干预以及设计特征时如何把表面相似现象分开。

With 26 samples,this paper adopts the cluster analysis and principal coordinates analysis with the help of NTSYS for the study of the present pronunciations of the ancient complete voiced initials in the existing Chinese dialects. It has obtained the following conclusion： 1） the result of the cluster analysis is inadequate as a reasonable classification scheme; 2） there is no one- to- one relationship between the evolutionary ways of the ancient complete voiced initials and the main groups of Chinese dialects; 3） the major difficulties in the construction of the database for Chinese dialectical characteristics are the approaches to the sufficient expert intervention and the distinction of surface similarities.