A Corpus Pragmatic Approach to Deception Detection in Forensic Contexts
Abstract This paper presents a corpus pragmatic analysis of deceptive language across three simulated forensic contexts: police interviews, border control interviews, and opinion-giving exercises. Whilst most traditional work within deception detection, particularly within applied social psychology, has primarily utilised quantitative methods, this paper instead advocates for a more context-sensitive approach in keeping with the corpus pragmatic methodology. Using a 190, 396-word corpus of 120 participants’ truthful and deceptive responses within three spoken interviews, the analysis uses keyword, concordance, and collocation analysis to identify prominent linguistic patterns within the data. Such patterns include pronoun shifts, specificity in temporal references, and pauses. The analysis thus highlights not only the statistical differences across interviews and across veracity conditions, but also the potential pragmatic reasons why participants use certain linguistic resources. The paper argues that it is beneficial to integrate qualitative interpretation into automated deception detection, offering a transparent framework allowing researchers to verify their findings, and confirm that what they think is happening is actually happening within the data. Ultimately, the paper argues for a more nuanced understanding of deceptive discourse.
Authors
- Mathew Gillings (ORCID: https://orcid.org/0000-0002-6114-0117)
Institutions
- Vienna University of Economics and Business (AT)
Publication Details
- Journal
- Corpus Pragmatics
- Published
- 2026-11-16
- DOI
- https://doi.org/10.1007/s41701-026-00271-3
- Primary Topic
- Deception detection and forensic psychology
- Type
- article
- Field-Weighted Citation Impact
- 0.00