Toddler’s word learning in virtual social interaction: a study using the gaze crossing paradigm

Eye gaze serves as a key channel for interaction between children and caregivers, which is crucial for child development (e.g., Farroni et al., 2002). A recent study (Lu et al., 2026) has introduced a new paradigm named gaze crossing, which implements real-time mutual gaze in a minimalist virtual space. The study has shown that an abstract visual avatar, which moves in response to children’s eye gaze in a way that faithfully retains the temporal-spatial properties of human eye-movement (a feature termed contingency), elicits more stable attention from the children. The preliminary finding suggests that such differences during gaze-mediated interaction elicited by contingency may further influence the performance of word learning. However, this finding could not elucidate whether the improved word learning outcome can be entirely attributed to the increased quantity of exposure, which was designed to occur only during interaction, or if the experience of contingency as a social cue of agency and/or intentionality uniquely contributed to the improved learning outcome. This registration describes a study that further examines this effect of contingency using the gaze crossing paradigm. It aims to first confirm that toddlers between 15 and 18 months of age can indeed better associate a novel object with a novel auditory word after being exposed to the mapping during screen-mediated interactions that are contingent but show minimal social cues otherwise. If the effect of contingency is confirmed, we will further test if this effect still holds after we control for the amount of exposure.