Contextualising (Im)plausible Events Triggers Figurative Language
2026-04-09 • Computation and Language
Computation and Language
AI summaryⓘ
The authors studied how people and large language models (LLMs) judge whether events in sentences are believable or make sense literally or figuratively. They created sets of statements that were either plausible or implausible, using both abstract and concrete words. Their results showed that humans are good at telling the difference between something that is literally true, figurative, or just doesn't make sense. In contrast, LLMs often confused implausible events for figurative ones, showing they understand context less deeply than humans do.
literalnessnon-literalnessplausibilitysubject-verb-objectlanguage modelscontextualizationabstract vs concreteevent tripleshuman judgmentfigurative language
Authors
Annerose Eichel, Tonmoy Rakshit, Sabine Schulte im Walde
Abstract
This work explores the connection between (non-)literalness and plausibility at the example of subject-verb-object events in English. We design a systematic setup of plausible and implausible event triples in combination with abstract and concrete constituent categories. Our analysis of human and LLM-generated judgments and example contexts reveals substantial differences between assessments of plausibility. While humans excel at nuanced detection and contextualization of (non-)literal vs. implausible events, LLM results reveal only shallow contextualization patterns with a bias to trade implausibility for non-literal, plausible interpretations.