Title: The uncooperative child: Three studies in which children refused to confirm my perfectly reasonable predictions
I will present three projects, united only by the fact that in each case, I was wrong. The first study investigated word production in toddlers. Prior work suggests that preverbal infants are sensitive to what others do and do not know, and selectively point out information (e.g., the location of a cookie jar) unknown to their audience. Consistent with this (and rational speech act models), I hypothesized that even at the one-word stage, toddlers might selectively refer to information unknown to their parents. I was wrong. Across two large-scale online experiments, we found that children’s one word utterances were equally likely to refer to novel information and information in common ground. In a second study, we looked at children’s understanding of the value of social attention. I was sure that four and five-year-olds would understand that people prefer attentive to distracted social partners. Instead, across seven experiments, children failed to infer a preference for attentive social partners or identify interventions that would hold others’ attention. This was the case even though children distinguished the attentive and distracted agents, and made accurate inferences about social preferences and the epistemic consequences of differential attention. In a third study, we looked at whether children understand multiple mutually exclusive possibilities. I was sure they would; spoiler alert: they did not. Four-year-olds treated trials where there were two mutually exclusively possibilities indistinguishably from trials where there was a single deterministic outcome. Children did not reach adult levels of performance until age eight. I will end with some reflections on intellectual humility.