Modeling of the Dynamic of Text Extensive Variables in Fiction (Experimental Study in the Material of the Corpus of Years 1900-1930)

The paper is dedicated to the study of fiction text dynamic - namely, it aims to describe its plot development through its extensive variables dynamic (paragraph length and sentence length) and discover a general, "standard" figure of their distribution. The material for the study is based on a subset of 535 texts from the Corpus of Russian Short Stories of the First Third of the XX century - later, texts of "standard" length were chosen. In order to find the dynamic of the text, a statistical method of time series was used. As a result, graphs that show the averaged sentence and paragraph length distribution were created for each 30, 50 and 100 short stories. Finally, the most common patterns among them were found. The research has shown that standard dynamic contours do exist and often follow the pattern of heavy exposition and short resolution. This data can later be used to find the correlation with other characteristics of texts or to classify texts based on their dynamic contour.