Event Detection between Literary Studies and NLP. A Survey, a Narratological Reflection, and a Case Study

Narrative structure in fiction relies on the strategic presentation of events, where the ordering and disclosure of information (syuzhet) shape reader engagement and tension. This study outlines a computational model for detecting syuzhet by formalizing event annotation in fictional texts across multiple languages. While automated event detection has been widely applied in domains like journalism and history, its theoretical foundations remain fragmented due to divergent definitions of "event" and domain-specific priorities. We critically synthesize prior approaches, highlighting their methodological and applicative distinctions, and position our model within this landscape. Additionally, we demonstrate its adaptability by comparing it to a storyline analysis framework designed for news, revealing cross-domain utility. Our work offers a flexible computational narratology framework for analyzing narrative progression in both fiction and non-fiction contexts.