Toddlers distinguish deliberate choices from forced actions when imitating

Evidence

Toddlers who saw that the adult’s arms were free reproduced the action exactly, bending forward to touch the lightbulb with their heads. But toddlers who had seen that the adult’s arms were occupied by holding the blanket simply reached out and touched the lightbulb with their hand. … There must have been some reason why this grown-up bent over and touched the light with her head—her hands were right there! If it seemed to have been a deliberate choice, they would reproduce it exactly. This showed the problem of overimitation in a new light. It was not “mindless” at all, not simply a slavish reproduction of the exact movement, but the opposite—a reasonable, sophisticated insight based on imagining the demonstrator as making rational choices and performing the action as easily and efficiently as possible. — Brian Christian, The Alignment Problem, 948

Categorization

• corroborated: 2 independent passages support this claim — Brian Christian, The Alignment Problem, 948
• XREF: connects to: Children persist with objects that violate their expectations, not merely novel ones. — Brian Christian, The Alignment Problem, 854; Human children overimitate even when they see actions are useless — Brian Christian, The Alignment Problem, 943; Overimitation grows with the ability to model others’ minds — Brian Christian, The Alignment Problem, 950 (+1 more)

Related

• Human children overimitate even when they see actions are useless (similarity 0.89)
• Overimitation grows with the ability to model others’ minds (similarity 0.86)
• Children persist with objects that violate their expectations, not merely novel ones. (similarity 0.79)

Broader Connections

• Deliberate action requires partial but not total foresight of consequences #expands (0.72) — Georg Simmel, on individuality and social forms_…, 785
• Human children overimitate even when they see actions are useless #expands (0.70) — Brian Christian, The Alignment Problem, 943
• Overimitation grows with the ability to model others’ minds #expands (0.60) — Brian Christian, The Alignment Problem, 950