Not Worth Mentioning? A Pilot Study on Salient Proposition Annotation

2026-03-28

AI Summary

This study addresses the lack of fine-grained annotation and quantification of propositional salience in natural text. It introduces, for the first time, a graded notion of salience at the propositional level and develops an operational annotation framework. Human annotators applied this framework to a small-scale, multi-genre corpus, and inter-annotator agreement was evaluated to validate its reliability. The work further investigates the relationship between this salience measure and nucleus centrality as defined in Rhetorical Structure Theory (RST). The contributions include a viable annotation scheme for propositional salience, empirically demonstrated annotation consistency, and preliminary evidence linking salience scores to discourse nucleus centrality, thereby offering a novel pathway to bridge abstractive summarization and discourse structure analysis.

Abstract

Despite a long tradition of work on extractive summarization, which by nature aims to recover the most important propositions in a text, little work has been done on operationalizing graded proposition salience in naturally occurring data. In this paper, we adopt graded summarization-based salience as a metric from previous work on Salient Entity Extraction (SEE) and adapt it to quantify proposition salience. We define the annotation task, apply it to a small multi-genre dataset, evaluate agreement and carry out a preliminary study of the relationship between our metric and notions of discourse unit centrality in discourse parsing following Rhetorical Structure Theory (RST).

Problem

• proposition salience
• extractive summarization
• graded salience
• annotation
• discourse centrality

Innovation

• graded salience
• proposition annotation
• extractive summarization
• Rhetorical Structure Theory
• salient entity extraction