Quantitative analysis of adjectives in the Russian literary corpus of realism and romanticism

Computational analysis of text is an increasingly important approach used by researchers in the field of digital humanities. A much-debated question is whether computational techniques such as text analysis, which is in fact a quantitative approach, is adequate for analysing literary texts, since literature is considered as a type of artistic expression. In the paper we highlight the importance of the application of computational analysis with a study conducted on a corpus of selected Russian literary texts from the periods of Realism and Romanticism. Texts included in the romantic subcorpus are “Eugene Onegin” by Alexander Pushkin and “A Hero of Our Time” by Mikhail Lermontov. Texts that constitute the realist subcorpus are “Anna Karenina” by Leo Tolstoy and “Crime and Punishment” by Fyodor Dostoevsky. The analyzed texts are translations into the Croatian language. The paper presents current methods and approaches used in computational literature analysis. The focus of this research is the analysis of adjective usage in romantic and realist texts, due to the fact that these two literary periods are based on distinctive poetic principles. The texts were analyzed using the programming language “Python”. Part-of-speech tagging was accomplished with an online tagger for Croatian language. Considering that all texts are historical (because they originate in the 19 or early 20 century) difficulties with POS tagging are expected. Results of the research show more similarites in the usage of adjectives between the subcorpora then expected. The paper points out how quantitative methods “borrowed” from the field of natural language processing and statistics can be significant in drawing conclusions about literature and that numbers can be meaningful if interpreted competently.