基于Zipf分布拟合的藏文字词发展演变研究

Study on the Evolution of Tibetan Words based on Zipf Distribution Fitting

字词是自然语言中音义结合体的最小单位,因此研究字词的演变有利于剖析一门语言的语法演变规律,有利于研究字词本身的语法属性。Zipf定律是已被证实的所有语言都共有的一个词频分布定理。文章以藏文第3次厘定为时间节点,选取厘定前后不同历史期的26篇文献著作,对其中的字频和词频进行统计分析,并与同时期的汉文、英文文献字频和词频进行比较研究。实验结果显示,藏文词频和汉文、英文一样符合Zipf分布,但藏文字频在厘定前后有较大的差异,即厘定前的文献《兄弟教诲录》和《罗摩衍那》字频分布比汉文《诗经》更符合Zipf分布,碑文和敦煌藏文文献字频也较符合Zipf分布,而厘定后的字频不符合Zipf分布。此结果表明,藏文在发展过程中,存在由单音节词向多音节词演变的语言文字演变过程,这与汉字的演变规律相似。此外,根据厘定前后的高频字词统计分析可知,古今藏文词缀和虚词也存在一定的变化,而且它们的演变规律也反映了被修饰成分的实词的变化规律。

Word is the smallest unit of phonetic combination in natural language,therefore studying the evolu⁃tion of words is beneficial to analyze the grammatical evolution of a language and to study the grammatical prop⁃erties of words themselves.Zipf law,which is a word frequency distribution law,has been verified being charac⁃terized by all language in common.In this paper,setting the 3rd determination of Tibetan as the time-separation point,the frequency of word selected from 26 Tibetan literatures written before and after this time-point was stat⁃ically analyzed,and meanwhile,words in contemporary Chinese and English literatures were statistically ana⁃lyzed as well.Our results show that the frequency of Tibetan word follows Zipf distribution,however,there is a significant difference in word frequency between the Tibetan literatures written before and after the selected time-point,i.e.the word frequency distribution of literatures written before the selected time-point,such as"Brothers Didactics"and"Ramayana",more comfort to Zipf distribution than that of the Chinese literature,i.e."Book of Songs".And the word frequency of the studied inscriptions and the Dunhuang Tibetan Buddhist scrip⁃tures also approximately comfort to Zipf law.However,the word frequency of literatures written after the selected time-point does not comfort to the Zipf distribution.Our results imply that in the course of the historical develop⁃ment of Tibetan,there is a transformation of language from monosyllabic to polysyllabic words,which is similar to the evolution of Chinese words.In addition,according to the statistical analysis of high-frequency words,it can be seen that there are certain changes in affixes and false words of ancient and modern Tibetan,and their evolution directly reflects the changes in the modified components of the content words.