Tracing the shapes of adventure: a subject-based interpretation of emotional arcs

Abstract Narrative, a common practice in human life, has long been a focus of humanities research. Advances in natural language processing have enabled us to delve deeper into narrative universals based on large-scale data. This study examines the emotional modes in 1,222 adventure stories from Project Gutenberg. Emotional arcs were extracted from these books, revealing six primary modes of emotional changes. To interpret these emotional modes, we analyzed their alignment with the subjects of the corresponding stories. The findings highlight the reliability of data-driven methods in uncovering universal emotional patterns in adventure narratives. Moreover, the study demonstrates that the subjects of stories are often consistent with their emotional modes. This consistency enables us to interpret and compare these modes. Our study offers valuable insights into narrative universals by enhancing the interpretability of emotional modes.