Evidentiality as a deceptive function: A cross-linguistic study in English, French, Turkish, and Japanese

This thesis study examined the linguistic structure of lying, focusing particularly on the use of evidentiality–a grammatical marker indexing the information source. In particular, this study investigated how the obligatory (evidential) and the optional (non-evidential) categories of grammatical evidentiality in language impact the production of certain linguistic structures as subtle cues of lying in deceitful versus truthful retellings. In a fully crossed, counterbalanced design, participants (N = 217) from language groups typologically diverse in evidential marking (i.e., non-evidential languages of English and French, evidential languages of Turkish and Japanese) provided written accounts of the events from brief stories under two manipulations: presentation modality (i.e., silent video clips or audio recordings) and veracity (i.e., either truthfully or deceitfully). Narratives were coded for the frequencies of the grammatical and lexical forms covering tenses, negations, and evidential markers, and then, cross-linguistically examined for their distribution across the conditions. The results revealed that lie tellers in evidential languages exploit direct evidentials in their statements–not the indirect ones, suggesting a pragmatic use of evidentials as firsthand accounts. The speakers of non-evidential languages used evidential markers without uniformity in deception conditions, such that perception verbs were fewer in English and affirmation adverb rates were higher in French, except both languages adopted a more negative tone in their lies, confirming the previous evidence. Overall, this study identifies an overlooked grammatical category, evidentiality, as a deceptive cue for the first time and provides a comprehensive approach to deception detection research cross-linguistically.