Modeling Event Salience in Narratives via Barthes’ Cardinal Functions

This work adopts Barthes’ definition of event salience and proposes several unsupervised methods that require only a pre-trained language model that outperform baseline methods and finds fine-tuning a language model on narrative texts is a key factor in improving the proposed methods.

Abstract

Events in a narrative differ in salience: some are more important to the story than others. Estimating event salience is useful for tasks such as story generation, and as a tool for text analysis in narratology and folkloristics. To compute event salience without any annotations, we adopt Barthes’ definition of event salience and propose several unsupervised methods that require only a pre-trained language model. Evaluating the proposed methods on folktales with event salience annotation, we show that the proposed methods outperform baseline methods and find fine-tuning a language model on narrative texts is a key factor in improving the proposed methods.