By just a few months of age, infants have expectations about the world around them. In large part this has been revealed by experiments using the violation-of-expectation looking paradigm, which capitalizes on the discovery that infants look longer at events that are surprising than events that are unsurprising. For example, infants look longer at a ball that appears to pass through a solid wall than at a ball that is stopped by the wall. In hundreds of experiments across many domains (including objects, numbers, and social agents), infants look longer at surprising events than at carefully matched non-surprising events. Although this literature has been invaluable in revealing the existence of knowledge early in human life, no research to date has asked why infants look longer at surprising events.
The aim of this project is to understand the nature and function of this behavior. The core hypothesis is that surprise acts to guide young learners to allocate cognitive resources to events for which they had an incorrect or incomplete expectation, thereby increasing learning efficiency. The first series of experiments will identify the functional relevance of the surprise and longer looking that infants exhibit to unexpected events - i.e., they will ask whether surprise leads to better learning in infants (Specific Aim 1). In these experiments, infants will have the opportunity to learn new information following either a surprising or a non-surprising (expected) event, and their relative learning success will then be assessed. The prediction is that infants will show enhanced learning about objects or events that generated surprise. The second series of experiments will ask whether surprise also acts to enhance learning in older children in a non-laboratory setting (Specific Aim 2). Preschool- and early school-aged children will be tested in an informal science setting (a science museum) across a variety of tasks, and their learning following surprising versus expected events will be measured. Testing older children under non-laboratory conditions that more closely approximate children's everyday learning opportunities will allow us to assess the generality of surprise-induced enhancement of learning, and to gauge the kinds of learning that do and do not benefit from surprise. Together, the proposed studies will shed light on the processes that lead to effective learning early in life. This research will help to characterize the nature of an everyday behavior that has been documented across hundreds of experiments but whose functional role remains entirely unknown and unexplored. A better understanding of how surprise guides learning has implications for remediating developmental disorders that have been linked to the deficient detection of unexpected events or outcomes, for understanding how attentional deficits lead to decreased learning in children, and for shaping educational practice.
Decades of research show that young infants look longer at surprising events than at expected events, and older children show analogous behavioral responses to surprising events;however, nothing is currently known about the behavioral significance of these surprise reactions. The goal of this project is to examine the impact of surprise on learning in infants and children, asking whether infants and children learn better following surprising events than expected ones. As several developmental disorders are linked with patterns of poor learning following unexpected or surprising errors, understanding how surprise influences learning in typically developing children may contribute to models for remediating impaired error detection in cases of atypical development due to injury or disease.
|Stahl, Aimee E; Feigenson, Lisa (2018) Violations of Core Knowledge Shape Early Learning. Top Cogn Sci :|
|Stahl, Aimee E; Feigenson, Lisa (2017) Expectancy violations promote learning in young children. Cognition 163:1-14|
|Stahl, Aimee E; Feigenson, Lisa (2015) Cognitive development. Observing the unexpected enhances infants' learning and exploration. Science 348:91-4|