Some analyses of lying establish a necessary connection between lies and semantics: the content of a lie needs to be semantically expressed by the liar’s utterance. This necessary connection between lying and semantic holds the potential to strengthen and improve Michaelson’s (2016) proposal of using studies on lying as a supplemental source of data for theorising in semantics. However, several surveys on laypeople’s application of the concept of lying appear to preclude this possibility. Such surveys suggest that laypeople judge deceptive additive particularised conversational implicatures—definitely not a semantic phenomenon—as lies. If those surveys are reliable, they disrupt the necessary connection between lying and semantics and, with it, a promising methodological avenue for semantics. We present the results of a survey that counters the trend of those experiments. Our findings show that laypeople do not classify additive particularised conversational implicatures as lies. Moreover, we provide a diagnosis of why the results of previous empirical surveys yield the opposite conclusion. Our results provide a partial but significant defence of the possibility to overcome the limitation of Michaelson’s proposal to use studies on lying as a source of data for semantics.
