Rhetoric Mining: A New Text-Analytics Approach for Quantifying Persuasion.

This article presents rhetoric mining, a novel text-analytics methodology that quantifies persuasion by combining qualitative rhetorical analysis with automated sequence-alignment algorithms adapted from computational biology. The approach identifies complex rhetorical moves—such as ethos (personal or cited expertise), hedging (certainty or uncertainty), and types of evidence—in large text corpora by detecting semantically equivalent word sequences, enabling scalable analysis beyond traditional small-sample rhetorical studies. The method is illustrated through an application to stock-pitch arguments in the Motley Fool online investment community, where rhetoric mining effectively distinguishes persuasive and trustworthy pitches with high confidence, outperforming standard and state-of-the-art natural language processing classifiers. The authors argue that rhetoric mining offers a new interdisciplinary lens for analyzing intentional language use in business decision making and suggest broad potential applications in areas such as consumer reviews, financial reporting, marketing, and social media analysis.