Do infants form causal expectations?

Lens adapted to this topic: Learned associations may explain early expectations

A serious alternative argues that infant looking and anticipation may reflect learned associations, perceptual heuristics, or limited event representations rather than abstract causal expectations. Associative-learning accounts challenge strong claims of innateness, while toddler studies show that prediction does not automatically produce causal intervention. The evidence therefore supports early sensitivity to regularities, but leaves open whether infants represent causes in a robust, generalizable sense.

• Evidence that infants interpret events causally does not by itself show that causal representations are innate; major reviews explicitly find no good evidence for innateness.
• An associative-learning account argues that infants may acquire distinctions about causal action through experience rather than begin with domain-specific causal knowledge.
• Toddlers readily learn predictive relations but often do not intervene to generate the predicted effect, suggesting a gap between prediction and causal action.
• Toddlers’ causal performance improves when agents, direct contact, or causal language make the relationship more salient, indicating context-sensitive mechanisms rather than uniformly abstract causal reasoning.
• The field contains competing perception-of-causality and constructivist accounts: one emphasizes early sensitivity, while the other emphasizes learning through observation and action.

Featured sourcedoi:10.1016/j.actpsy.2006.05.005harvardlds.org6 sourcesView sources