The Relationship And Classification Of Sino-Tibetan Languages

The classification of languages and dialects has always been one of the debatable problems. Since the traditional language classification method, adopting the qualitative analysis with experience, cannot describe the distant genetic relationships of languages quantitatively, the credibility of the relevant conclusions should be reconsidered. The purpose of this paper is to provide a scientific determination and classification of the genetic relationship between the languages and dialects of Sino-Tibetan family with the advanced methods such as the computational method, the molecular anthropological method and the etymological statistics, according to the principles of computational linguistics. The paper systematically and comprehensively evaluates the genetic relationship between the languages and dialects of the Sino-Tibetan family by describing the genetic distance of languages with the distance-and-feature-based method, and thus to make a scientific classification. The mathematical tree diagram adopted in the paper could catch the overall relationships of all the languages, and also find some problems of theoretical significance, such as predicting the changing speed and substitution direction of various words and working out the separation time of Chinese languages and dialects. The significance of this study lies in the establishment of the specific mathematical model for the classification of Chinese languages and dialects, and thus providing the basis for the comparative research of the genetic relationship and geological distribution distance between the system tree based on the language genetic information and that based on the biological genetic marks, which constitutes the pioneer research of the Chinese linguistic information processing with the combination of computational linguistics and genetics.With the etymological statistics and the molecular anthropological method, the paper mathematically classifies the Sino-Tibetan languages and dialects and describes their genetic relationship. The distance relationship is indicated by the branch lengths of the constructed trees, which shows the clustering of the languages and their separation in hierarchies. In addition, the time depths of the languages in the family are estimated and a hypothesized process of the formation of the language family is given.The research finds that different lexical categories in a language vary in changing degrees. Even the changing degrees of the basic vocabulary should not be hypothesized as of the same value. Although the word list of Swadesh has been successfully applied in thousands of languages in the world (Rosetta Project), it still cannot be testified by Chinese languages. This paper claims that the 100 words in Swadesh's list can be used as the classification standard of the Sino-Tibetan languages and dialects, and thus contributes to linguistic theories.As is shown by the classification results, the research has corrected some traditional false classification and thus provides a more scientific method and conclusion to settle the long-lasting dispute within the traditional classification, for instance, the classification of She; the languages of Bai, Qiang and Tujia being of independent language branches; Lin Gao being genetically closer to Li and Zhuang, as contrasted with the traditional classification which considers that Li Gao is only close to Zhuang but is of remote relationship with Li and they appear to belong to different language braches.From the structural change of the language evolution tree diagram, the paper proposes to reconsider the traditional family classification framework within the western historical linguistics, which arbitrarily classifies language sub-families, language branches, languages and dialects on the level of only one language. This kind of method is too simply and outdated to tell the hierarchy and overall structure of languages. From the tree diagram, it is shown that language branch level is not fixed but is in a dynamic state and is of multiple levels, that is to say, there is not only one language branch, but rather every separation point in the tree diagram amounts to the position of"language branch". The language evolution tree diagram can improve the traditional classification theory and thus reflects the relationship between the language levels and the whole better.With the computational method of the average value of the reservation rate per thousand year, we have worked out not only the separation time of each pair of languages, but also that of every separation points in the tree diagram, and the conclusion is consistent with the latest research findings of the archeology and the anthropology. Our results differ from the existent viewpoints of many linguists. This research intends to demonstrate the special features of the evolving process of the East-Asian languages and peoples and also the limit of the historical comparative method.It is considered in the paper that Austronesian family lived in the large area around Southern China, and began to separate from 4000 B. P. and expand to Taiwan and the Araucaria islands along the East-south coastal area or the West-south– Middle-south peninsula.The research also finds that the Southern Chinese dialects have the"regional common tradition", that is, the mutual communication of the various regional cultures in the area plays an important role in formation of the common features of the region. The"Han-people-centered theory"in the linguistic mainstream research has been outdated.The traditional historical linguistics mainly depends on the systematized experience principles based on the expert knowledge to reconstruct the"imagined"ancient languages, but cannot be testified directly. Compared with the traditional classification, although with different materials and methods, the research provides the roughly same results with the traditional classification, and finds out the deficiency of the traditional qualitative classification. The new method differs from the traditional linguistics in that it can be derived by formulas and be applied and testified repeatedly. It provides a sample of mathematical classification to settle the long-lasting academic dispute caused by the traditional classification, and its successful examples and research methods can also be extended to the classification of more languages and dialects.