How (not) to measure infant Theory of Mind: Testing the replicability and validity of four non-verbal measures

Abstract

A growing body of infant studies with various implicit, non-verbal measures has suggested that Theory of Mind (ToM) may emerge much earlier than previously assumed. While explicit verbal ToM findings are highly replicable and show convergent validity, systematic replication studies of infant ToM, as well as convergent validations of these measures, are still missing. Here, we report a systematic study of the replicability and convergent validity of implicit ToM tasks using four different measures with 24-month-olds (N = 66): Anticipatory looking, looking times and pupil dilation in violation-of-expectation paradigms, and spontaneous communicative interaction. Results of anticipatory looking and interaction-based tasks did not replicate previous findings, suggesting that these tasks do not reliably measure ToM. Looking time and new pupil dilation measures revealed sensitivity to belief-incongruent outcomes which interacted with the presentation order of outcomes, indicating limited evidence for implicit ToM processes under certain conditions. There were no systematic correlations of false belief processing between the tasks, thus failing to provide convergent validity. The present results suggest that the robustness and validity of existing implicit ToM tasks needs to be treated with more caution than previously practiced, and that not all non-verbal tasks and measures are equally suited to tap into implicit ToM processing.