Reading Software Assistant: Spatial and Temporal Reordering of Literary Subplots

The analysis of narrative elements in documents or sets of documents remains an unresolved issue. Working with long texts, such as chapters in literary fiction, presents challenges for handling long-range dependencies and contextual information, even for modern language models, despite their progress in information extraction, natural language processing, and text generation. This study aims to present a method for analyzing literary text as a relevant tool for digital humanities research, specifically for the temporal and spatial ordering of plot elements. The method is implemented through the application of both quantitative methods, such as frequency counts, and natural language processing techniques, as well as graph algorithms to delineate narrative lines. The study also aims to demonstrate, at the chapter level, how certain plot elements (characters and their actions) are temporally situated within the narrative as an attempt to create a profile of event participants wherever they are mentioned. The methods outlined in the study were applied to the text of Mikhail Bulgakov's "The Master and Margarita."