Another Look at Looking Time: Surprise as Rational Statistical Inference

Decades of developmental research have capitalized on the fact that infants are surprised (i.e., look longer) at some events but not others. Differences in looking time have been considered to be a reflection of perceptual discrimination, or a reaction toward witnessing a violation of prior expectations. Here, we provide an overview of a new perspective on infant surprise that examines the underlying cognitive processes that drive this response. We suggest that looking time may reflect sophisticated statistical inference, and we review empirical evidence and computational modeling results from several recent studies to support this conjecture (Kidd, Piantadosi, & Aslin, ; Piantadosi, Kidd, & Aslin, ; Sim, Griffiths, & Xu, ; Sim & Xu, ; Teglas etal., ). We also discuss how our view relates to other new developmental research on surprise and learning (Stahl & Feigenson, , ) and outline some suggestions for future research. Surpriseoperationalized as looking timehas a long history in developmental research, providing a window into the perception and cognition of infants. Recently, however, a number of developmental researchers have considered infants' and children's surprise in its own right. This article reviews empirical evidence and computational models of complex statistical inferences underlying surprise, and discusses how these findings relate to the role that surprise appears to play as a catalyst for learning.