2021 Abstract

Tang Xianzu is a representative of legendary Chinese operas in the Ming Dynasty, also known as“Oriental Shakespeare”.In the past, some scholars studied the style of Tang Xianzu’s opera from the perspective of literature, aesthetics and drama, but there were rarely studies on the analysis of the style of Chinese classical opera using computer quantitative methods. This article introduces computer language processing methods to analyze the style of ChuanQi-style opera works. This research uses the surviving five Chinese operas and lyrics of the“Tang Xianzu Opera Collection” published by Shanghai Ancient Books Publishing House in 1972 as the corpus. First, the Jieba tokenizer is used for word segmentation to generate a dictionary. Secondly, the delta method is used to set the contrast. The text is compared with other texts, and the vocabulary similarity between the texts is compared and analyzed by calculating the difference between the average value of the text feature vectors. Finally, the TF-IDF method is used to analyze the rules of text feature vocabulary. The results show that the delta method uses “The Purple Jade Flute” as the control text, and the delta values ​​of“The Legend of Purple Hairpin” , “Dreaming Under the Southern Bough”, “The Handan Dream” and “Peony Pavilion”are between 1.0098404795952909 to 1.7722406232197077.Which indicating that the styles of“The Purple Jade Flute”  and“Peony Pavilion”  are quite different. The style differences between the other three works and “The Purple Jade Flute”gradually become smaller in the order of “The Handan Dream”, “Dreaming Under the Southern Bough”and “The legend of the Purple Hairpin”. Using the TF-IDF method to analyze those four articles, I find the first 20 characteristic words in Tang Xianzu’s opera such as“princess(郡主)” , “Wash the yarn(浣纱)” , “comfort(安抚)”. The style of the words ranges from gorgeous to elegant, such as word in “The Purple Jade Flute” like“beautiful Jade(璚)” and“gorgeous(绮)”  are more gorgeous than the other four works; the characteristic words also reflect the change of Tang Xianzu’s thoughts. The words “marriage(红丝)” and “faithless(薄幸)”in the“The Purple Jade Flute” mainly reflect love. The words“Military(军事)” and “Soldier(金兵)” in“Peony Pavilion” began to express political ideas. In the later period, “King(国王)” and “Ambassador(大使)”in “The Handan Dream” and “Dreaming Under the Southern Bough” were significantly more politically colored. “Dreaming Under the Southern Bough”and“The Handan Dream”, such as “a Buddhism goodness Guanyin(观音)”, “Chan Master(禅师)” again express Buddhist thought. The frequency of characteristic words such as “Hell（阴司）”, “be reincarnated(超生)”  and“Dreaming(黄粱)”  in“The Handan Dream”  shows that Xianzu’s operas are surreal. In summary, this article explores the Delta combined TF-IDF method as a quantitative analysis method for text style research, which has strong feasibility and has a certain reference value for opera style research.

Keywords: Tang Xianzu opera; delta method; TF-IDF method; text analysis