Literary Style Quantitative Analysis Based on NLP: Taking Dickens and Austen as Examples

Abstract

: Literary style research has always relied on researchers’ subjective perceptions and qualitative descriptions, lacking objective quantitative evidence. In order to make up for this methodological limitation, this study introduced natural language processing technology to conduct a cross-field quantitative comparative analysis of the literary styles of Dickens and Austen, two representative writers. The study constructed a digital corpus of representative works of the two writers, and systematically extracted quantitative features from four dimensions: vocabulary richness, syntactic complexity, emotional distribution and topic model. Through methods such as word frequency statistics, keyword analysis, syntactic analysis, sentiment calculation, and LDA topic modeling, the systematic differences in language use between the two writers are revealed in a data-driven manner. This study aims to use accurate data to verify the core assertions about the styles of the two writers in traditional literary criticism, thereby proving the effectiveness and scientificity of the NLP measurement method as a useful supplement to traditional literary research, and providing an empirical path that can be used as a reference for literary criticism from the perspective of digital humanities.