Predicting Veracity From Linguistic Indicators

Abstract

Ample scientific research has confirmed significant linguistic differences between truthful and deceptive discourse in both laboratory and field experiments. That literature is reviewed, followed by presentation of an experiment that tested the effects of veracity on a wide array of linguistic indicators and tested which effects were moderated by motivation and modality. A 2 (veracity: truthful/deceptive) × 2 (incentives: high/low) × 3 (modality: FtF/audio/text) factorial experiment revealed that linguistic indicators of quantity, immediacy, vividness/dominance, specificity, complexity, diversity, and hedging/uncertainty were all affected by veracity, and veracity interacted with motivation in the latter four cases. Only personalism and affect failed to differ between truth and deception. Modality also affected language use but did not interact with veracity. Four linguistic indicators together successfully classified 76% of text-based deception and 76% to 78% of truthful responses from text, audio, and face-to-face interaction. The importance of context in predicting linguistic patterns is emphasized.