Corpus approaches to deception detection: a case for interdisciplinary collaboration

Activity: Talk or presentationScience to science


Research from the field of psychology has shown that there are significant differences between the language of a truth teller and that of a liar, and investigating these differences in more detail may point towards so-called ‘cues to deception’. In this talk, I will offer an overview of past, present, and potential future work within the field of deception detection, spanning both forensic psychology and linguistics alike. I will first discuss earlier works, which have mainly consisted of hand-coding interview transcripts, before moving on to later studies which have tended to utilise LIWC (Pennebaker et al., 2001) as a form of computer assistance.

However, whilst LIWC was introduced to deception detection research with the hope of standardising analyses and reducing researcher subjectivity/bias, it has produced its own unique set of problems. The tool decontextualizes language by stripping away its grammatical coherence, and uses taggers which rely on simple dictionary-based algorithms (Lorenzo-Dus and Kinzel, 2019; Gillings, 2021). The present talk will therefore focus on how current practice could be refined from a forensic linguistics perspective, in particular by applying a corpus linguistic methodology, to explore language use in context - both quantitatively and qualitatively. To do this, I will draw upon case studies from my own and others’ work, and offer some thoughts on how the field may move forward through collaboration, rather than opposition.
Period5 Nov 2022
Event title3rd Indonesian Community for Forensic Linguistics Conference
Event typeConference
Conference number3
LocationBandar Lampung, IndonesiaShow on map
Degree of RecognitionInternational