Literature has the unique power to let us experience more than life. Its language draws readers into embodied, emotional, and perceptual engagement that can rival real life experience. In this talk, I present two studies that investigate this evocative power: first, by examining where humans and computational models diverge in sentiment judgments on literary texts, and linking these divergences to features such as concreteness, imageability, and interoception; second, by introducing embedding-based approaches for measuring imageability and visuality in sentences, enabling a scalable analysis of literary immersivity.Together with other contributions from narrative analysis, this line of studies can shed light on how narrative immerses readers and potentially expand their experiential horizons.

Speaker:

Yuri Bizzoni is a Senior Researcher in Computational Linguistics at Aarhus University, where he works on computational narratology, sentiment and emotion modeling, and large-scale analysis of literary style, and interactive tools for story generation. He is interested in how computational methods can deepen our understanding of reader engagement, literary quality, and the emotional impact of texts.