Identifying Stylistic Features and Conceptual Integrity of the Book of Verse ‘On Early Trains’ by Boris Pasternak by Methods of Computational and Interpretive Linguistics

While Pasternak’s oeuvre has been thoroughly studied in terms of style evolution, there are still works, the stylistics of which has not been fully explored. One of such works is the book “On Early Trains”, which comprises the separate cycles written in different periods of time, and can be considered therefore as a complex combination of the themes, concepts and stylistic approaches. Thus, the current study aims at identifying the specific stylistic traits, themes and concepts that characterize the book. In the current work, we use the methods of computational and interpretive linguistics, including thematic modeling applying Latent Dirichlet Allocation (LDA) (Blei et al., 2003) and interpretive content analysis. We conduct the study on the corpus of poems derived from the book “On Early Trains” with the use of various text analysis programs, namely AntConc and Voyant Tools to characterize the book individual style. In order to identify the main themes of the book and their distribution across the cycles, the Latent Dirichlet Allocation (LDA) and keywords extraction is performed. On the basis of the extracted data, the interpretive content analysis was conducted to describe the conceptual sphere of the book. As a result, the detailed characterization of the book including the individual style of the text and its conceptual integrity is provided.