Spotting hyperbole in communication: building a scalable framework with linguistic evidence

Abstract

Although the Hyperbole Identification Procedure (HIP) offers operational definitions and procedural guidelines on hyperbole identification, suggesting that hyperbole occurs when the literal meaning of an expression is more extreme than what the context warrants, it does not provide concrete instructions for determining when such a disparity can be considered to occur. This operational ambiguity has hindered the development of clear, replicable criteria. The article proposes an approach that supplements the HIP by constructing linguistic scales from corpus-based lexicographic resources to determine the degree of intensification in lexical items. Building on Peña-Cervel and Ruiz de Mendoza Ibáñez (2017), we operationalise three conditions under which an expression’s meaning can be treated as intensified and, therefore, hyperbolic within HIP. We also recommend triangulation with psycholinguistic databases to evaluate emotional intensity, and conducting contextual verification using domain-specific sources to account for specialized usages of the lexical unit. A reliability test shows that integrating HIP with linguistically derived scales yields consistent and reproducible annotations. We also address the challenges HIP faces with multiword expressions and propose a non-compositional procedure for detecting idiomatic hyperboles. The study concludes by applying the extended HIP to a real-world corpus and presenting detailed case studies that illustrate the analytical and methodological payoffs of the proposed extension. This paper lays a foundation for further research on hyperbole by providing systematic yet flexible guidelines for its identification and emphasizing its role as a key rhetorical device across diverse communicative contexts.