面向跨语言典籍文本的语体风格计算研究——以《左传》《战国策》为例  

Research on Style Calculation of Cross Language Classic Texts——Taking Zuo Zhuan and Zhan Guo Ce as an Examples

作　　者：宋旭雯 SONG Xuwen(Library of Changzhou University,Changzhou Jiangsu,213164,China)

机构地区：[1]常州大学图书馆,江苏常州213164

出　　处：《文化创新比较研究》2023年第26期22-28,共7页Comparative Study of Cultural Innovation

摘　　要：为了对比《左传》《战国策》的原本、现代汉语译本和英译本在语体风格上的异同,该文结合语料库和风格计算的方法,在构建大规模跨语言典籍语料库的基础上,采用CRF、BiLSTM、BiLSTM-CRF等机器学习或深度学习的模型处理语料,从词汇、语法、语句三个层面入手,选择能代表语体风格的特征进行统计,利用相关公式计算语体风格度量指标,结合原文比较典籍的语体风格。同样是以叙事为主的历史散文,《左传》《战国策》都是信息量庞大的活跃性文本,并且语言较为松散。不同之处在于《左传》语言古奥,朴素典雅,波澜起伏,含蓄内敛;《战国策》生动形象,辞藻华丽,音律和谐,气势磅礴。In order to compare the stylistic differences between the original,modern Chinese,and English versions of Zuo Zhuan and Zhan Guo Ce,this article combines corpus and style calculation methods to construct a large-scale cross language classic corpus.Based on this,machine learning or deep learning models such as CRF,BiLSTM,and BiLSTM-CRF are used to process the corpus.Starting from the vocabulary,grammar,and sentence levels,features that can represent the stylistic style are selected for statistical analysis,using relevant formulas to calculate stylistic metrics,and comparing the stylistic styles of classics with the original text.Similarly,as historical prose that mainly focuses on narrative,Zuo Zhuan and Zhan Guo Ce are active texts with a large amount of information and relatively loose language.The difference lies in the ancient and elegant language of Zuo Zhuan,which is undulating and reserved.Zhan Guo Ce is vivid and vivid,with magnificent language,harmonious melody,and magnificent momentum.

关 键 词：《左传》 《战国策》 语言风格 风格计算 跨语言 语料库 

分 类 号：G059[文化科学]