• Dr.V.N. Sudheer

  Author

This paper explores how computational stylistics can be effectively used to study and differentiate literary style in British prose of the 19th century using programming environment R. The aim is to find unique stylistic patterns in the chosen works of Charles Dickens, Jane Austen, and Thomas Hardy using the quantitative measures of linguistics. It employs a computational and statistical methodology, which includes the incorporation of corpus using sources in the public domain, and then preprocessing tools like tokenization, normalization, and removal of stop words. Such stylistic characteristics as the frequency distribution of words, type-token ratio (TTR), sentence length, and n-gram patterns were obtained and processed with the help of R libraries tidytext, quanteda, ggplot2. Significant stylistic differences were identified by the use of statistical techniques such as descriptive analysis and clustering techniques. The outcomes demonstrate apparent differences among authors. The corpus of Dickens includes 185,240 words and of which 18,560 are unique words, the average length of the sentences is 21.4 words and it means that the structural complexity is higher. Austen writes 121,870 words, 12,430 distinct words and shorter sentences (average 17.2 words) indicating a tight and conversational manner. The corpus of Hardy contains 142,315 words, containing 15,275 different words and exhibits the greatest lexical diversity (TTR = 0.107). The themes of differentiation are described by frequency of words, analysis of bigrams, and cluster analysis proves the definite stylistic grouping. The paper concludes that computational stylistics is a robust and repeatable approach to a literary analysis, that can fill the gap between the traditional interpretive approach and quantitative approach and has useful applications in English pedagogy and digital humanities.

• PDF