Cognitive Load as a Key to Lie Detection: A Psycholinguistic Exploration of Pause and Hesitation in True and Fabricated Narratives

Abstract

This study investigates the role of pauses and hesitations as crucial non-verbal cues in deception detection, emphasizing the influence of cognitive load, as framed by the theoretical framework established by Memon et al. (2003). By examining cognitive indicators such as filled and silent pauses, repetitions, slips of the tongue, word omissions, speech rearrangements, sentence incompletions, and drawl, this research sheds light on the mental processes underlying language production in deceptive communication. Utilizing a comprehensive experimental design, we collected narratives from 30 native Persian speakers who recounted both truthful experiences and fabricated scenarios over a one-week period. This design aimed to minimize situational cognitive load and emotional anxiety by allowing participants ample preparation time and enabling them to record their narratives in a familiar home setting, thus isolating the cognitive demands inherent in the act of fabrication itself. Our dataset comprised 60 transcribed video recordings, which were meticulously coded for the specified disfluency types and analyzed using Praat software for phonetic detail where necessary (e.g., pause duration). Statistical analysis via paired sample t-tests in SPSS, with Bonferroni correction for multiple comparisons, revealed a significant increase (p < .006) in the frequency of silent pauses (≥200 ms in duration) during the narration of fabricated stories compared to truthful ones. This finding suggests that the cognitive effort required for deception persists and manifests measurably, even when lies are premeditated and delivered in a low-stakes environment. Despite the controlled setting designed to reduce anxiety and planning demands, our findings underscore the potential reliability of specific non-verbal cues, particularly silent pauses, as indicators reflecting the cognitive architecture of deception. The study emphasizes the importance of integrating analyses of both verbal content and non-verbal behaviours in developing more nuanced and effective lie detection methodologies.