Weak Evidence but strong priors? Why Children Struggle with Belief Revision in Gradually Changing Environments

Rafetseder et al. (2021) were the first to demonstrate that children have difficulty revising their beliefs when objects morphed from one object (e.g. a rabbit) into another (e.g. a duck). Children aged four to eight reported the second object significantly later than adults. This is surprising, because both children and adults held the same prior belief (“it is a rabbit”) and were exposed to the same contradictory evidence. We conducted three experiments to examine children’s delayed belief revision. In Experiment 1 (N = 58; 3–6-year-olds), we demonstrated that children only struggle to identify the second object in the gradual condition, while displaying adult-like categorical perception when the same images were presented individually outside of the morphing context. In Experiment 2 (N = 86; aged 6–11), we found that only children aged nine years and over showed a similar pattern to adults, and that a global processing style facilitated identification of the second object. In Experiment 3 (N = 47, 5–6-year-olds), when children were explicitly informed of the two possible interpretations of the morphed images, their performance increased, although not to the level observed in the individual condition of Experiment 1. These findings suggest that children of different ages struggle with the morphing task for different reasons. While older children may use less efficient, potentially local, exploration strategies, younger children may not explore at all due to their overconfidence in their initial priors.