|
Surprise-induced learning enhancement in infants and children| old_uid | 14461 |
|---|
| title | Surprise-induced learning enhancement in infants and children |
|---|
| start_date | 2014/10/13 |
|---|
| schedule | 11h |
|---|
| online | no |
|---|
| details | Invitée par l'équipe Perception-Action |
|---|
| summary | Many theories propose that new knowledge is acquired through the accumulation of experiences with the environment. For example, knowledge of objects has been conceived of as the product of learning from actions on objects or observations of objects’ behavior. Here, I consider a reversal of this relationship: learning as a product of prior knowledge. Generally speaking, a learner with limited cognitive resources should allocate these resources “smartly,” directing learning toward objects, events, or relationships about which little is already known. One way to identify such opportunities is to focus learning on situations in which prior knowledge yielded the wrong prediction about the world. In this case, learning should be heightened (relative to situations in which the learner made a correct prediction). Here I review recent experiments with infants and preschool-aged children in support of this hypothesis. Our results show that early learning is enhanced when knowledge of basic principles of object behavior is violated, relative to nearly perceptually identical situations in which no such violations occur. For example, infants who saw a ball appear to pass through a solid barrier subsequently learned about the ball more effectively than infants who saw a nearly identical event in which the barrier stopped the ball. This learning enhancement is specific to the entities that participated in the surprising event, but supports learning about a range of properties. Taken together, this research suggests that children make predictions about entities in the world around them, evaluate these predictions against the evidence, and, when their predictions are wrong, direct their learning resources to revising their prior knowledge. In this sense, core knowledge guides learning. |
|---|
| responsibles | Rämä, Izard |
|---|
| |
|