Abstract
Background: Over the past four decades, narrative theory has increasingly engaged with computational methods, yet the integration of large-scale corpus stylistics into quantitative narrative analysis remains fragmented. This article addresses the gap by demonstrating how corpus-stylistic techniques applied to pre-2020 datasets can systematically reveal patterns of narrative structure, point of view, and character empathy. Methods: A corpus of 150 English-language novels (1850–1999) was compiled from open-access repositories, annotated for key narrative features (personal pronouns, speech-presentation categories, temporal markers, and lexical diversity scores). Computational-stylistic analyses included principal component analysis, regression modeling of reader-response metrics, and diachronic comparison of stylistic variables. Results: Four principal findings emerged: (1) first-person narration increased from 14% to 41% of sampled texts between 1850 and 1999; (2) lexical diversity decreased significantly in twentieth-century prose (β = −0.18, p
Keywords
corpus stylistics, computational narratology, quantitative narrative theory, point of view, lexical diversity, narrative empathy, diachronic stylistics, pronominal analysis