Research On The Chinese Version Of "Chun Xiang Zhuan" Based On Computational Stylistics

Language style is the overall personal style of an individual's written or oral expression and this kind of style can be depicted numerically by the numeral features of words,phrases,sentences,paragraphs and discourse structures.,Computational stylistics is through this data-driven approach to reflect the writer's habits of wording and phrasing,and then clarify the writer's language style.In recent years,the development of computational linguistics has provided a large number of real language data for stylometry-based analysis,which has unparalleled innovation significance for the understanding and the research of language.Chunhyangcheon(Chunxiangzhuan),as one of South Korean's outstanding classical literary classics has been introduced into China since very early and it is fascinated by readers and critics in China,followed by a group of outstanding Chinese translation versions.However,many pervious researches to date have focused on the qualitative study on the original or translation versions.Therefore,this makes significant the present study,a quantitative research on the comparison between two translation versions of Chunhyangcheon.The purpose of this study is to compare the stylistic differences in the four aspects,namely: sentence length,parts of speech,words and punctuations in the Chinese version of Chunhyangcheon translated by Liu Yingjiu(2006)and Xue Zhou & Xu Lihong(2010).In this study,parallel corpora of the original and two translation versions were established,using nlpir word segmentation system to label the parts of speech followed by manual proofreading.Finally,MTmine R and R languages were used for statistical analysis.The major findings are summarized as follows:Firstly,there are significant differences in the sentence length between the two translation versions.Specifically,the average sentence length of Liu Yingjiu's version is relatively longer,and the sentence length distribution is relatively uniform.The sentence length of Xue Zhou & Xu Lihong's version is shorter,but similarly,the sentence length distribution is also balanced.Secondly,considering the parts of speech,Wilcoxon Signed-Rank test showed that there were noticeably differences between the two translation versions in five parts of speech: adjective,noun,state word,auxiliary word and pronoun,with these all higher in Liu Yingjiu's version.The results of the principal component analysis showed that except conjunctions,the usage rate of all the other parts of speech in Liu Yingjiu's version was higher than that of Xue Zhou & Xu Lihong's,which indicated that Liu's version paid more attention to the changes of parts of speech.According to the gini coefficient of random forest,the usage of adverbs was the best indicator to identify two versions,followed by verbs,prepositions and quantifiers.Thirdly,at the word level,the results of type,token and type/token ratio(TTR)showed that two translation versions were rich in words.Although the number of types and tokens in Xue Zhou & Xu Lihong's version were higher than those of Liu Yingjiu's,the TTR was lower.Liu Yingjiu tended to use modal particles such as “a”,“ma”,“ba” and “ne”,while Xue Zhou & Xu Lihong used more adverbs such as “ye3”,“you4”,“jiu4” and “que4”.Finally,in the terms of punctuations,except the period,the frequency of other punctuations used in Liu Yingjiu's version were higher than that in Xue Zhou & Xu Lihong's.The most frequently used punctuation in these two versions was comma and the usage of comma in Liu's version was significantly more than that in Xue & Xu's version.What's more,there were also significant differences in the usage of question marks,exclamation marks,ellipses,semicolons and dashes between the two versions,which were specifically reflected in the fact that Liu's version was more than Xue & Xu's version.Therefore,the translation strategies of these two versions had an important impact on their language styles.Liu Yingjiu adopted literal translation and used more Korean proper words.On the contrary,Xue Zhou & Xu Lihong's version is more readable for Chinese readers by focusing on Chinese expressions.Based on the above findings,it is expected to provide new ideas for future research on translation works.