When babies see an object behave in an unexpected way — appearing to pass through a wall or levitating in the air — they are more likely to make a concentrated effort to learn about the object than if it behaved as expected, new findings published in the 3 April issue of Science report.
The study confirms something scientists in the field of infant learning have long suspected but could not prove, that violations of so-called "core" or innate knowledge cue a special opportunity for learning.
"Many decades of research have shown that infants look longer at all different kinds of surprising events," said co-author Aimee E. Stahl, a Ph.D. candidate in psychological and brain sciences at Johns Hopkins University, "but no study had asked why infants produce this systematic behavior."
"Despite surprise being a very helpful tool for understanding what expectations preverbal infants have, no one had asked what the cognitive consequences of experiencing surprise might be," added co-author Lisa Feigenson, a professor of psychological and brain sciences at Johns Hopkins University, during a teleconference for reporters on 1 April.
"Our work shows that infants' early expectations about the world can scaffold or guide their future learning," Stahl said. "This raises exciting questions about whether surprise can be harnessed by parents or educators to shape children's learning. It very well might be the case that probing children's intuitions and predictions in all kinds of settings might help guide their subsequent learning. This is something we are currently investigating in our own research."
Stahl and Feigenson tested their hypothesis that infants use surprising events as special opportunities to learn in a series of experiments with 11-month-old infants.
In their first experiment, they showed infants either an expected event or a surprising event. For example, one group of infants saw a ball roll across a stage and appear to be stopped by a wall in its path. Another group of infants saw the same ball roll across the stage, but appear to pass through the wall. The researchers then taught all the infants something new about the object — that the ball made a specific sound.
"Our work shows that infants' early expectations about the world can scaffold or guide their future learning. This raises exciting questions about whether surprise can be harnessed by parents or educators to shape children's learning."
"Even though all of the babies were equally interested in the events," Stahl said, "we found that infants who saw the expected event failed to learn the new information about the ball. However, infants who saw the surprising event learned this new information very efficiently."
The researchers observed the same learning enhancement for other types of surprising object behavior, such as when babies saw an object disappear and magically reappear somewhere else.
Next, the researchers explored whether infants were ready to learn about absolutely anything after they saw a surprising event, or only about the very thing that surprised them. They again showed infants a ball that violated their expectations by appearing to pass through a wall, but immediately after taught the infants about a different toy. The infants showed no increased interest in learning about this new toy, suggesting that the surprise only enhanced learning for the object that behaved surprisingly.
Finally, Stahl and Feigenson investigated whether babies would try to learn on their own about objects that surprised them. As before, they let babies watch either an expected or surprising event involving a ball, and then gave them two objects to freely explore — the same ball from the event and a toy car. They found that babies who saw the expected event with the ball explored the ball and car equally, but those who saw the surprising event with the ball explored the ball for far longer.
What's more, the babies weren't just interested in the ball generally but were actually seeking explanations for the strange events they saw happen to it, the researchers suggested. In the case of the ball passing through the wall, for example, the infants tested whether the ball was really solid by banging it. (A statistical analysis of the frequency with which the infants banged the ball versus performing another exploratory action, like dropping, revealed that the infants banged the ball much more than they dropped it.)
Feigenson explained that they worked with 11-month olds in this study because they needed infants young enough to be engaged but old enough to learn. "Our hunch is that if we were to ask if surprise boosts learning in younger or older children, the answer would be yes," she said. "We think that this is quite a general aspect of human learning."