of the Critical Period
Elissa L. Newport University of Rochester
Variations in learning that appear in different populations exposed to the same external environments provide some of our strongest evidence of internal, or biological, constraints on learning mechanisms. For example, Peter Marler's chapter in this volume describes evidence for constraints on learning obtained by comparing different species of sparrows reared in similar environments; given quite different songs learned in these similar circumstances, we must conclude that there are quite different internal constraints on learning operating in the two species. In a similar fashion, I focus on another type of evidence for innate constraints on learning, also present in the populations Marler described. In particular, I describe evidence from our own work on human language acquisition, comparing members of the same species who differ in the maturational periods during which they are exposed to their learning environments. As in the contrast Marler described, these groups of subjects in our experiments learn language quite differently, suggesting that they, like sparrows, come to the learning situations with quite different internal constraints. In this case, the evidence for innate or biological constraints on learning comes from the finding that, as these constraints apparently disappear or weaken over maturation, the ability to learn declines.
The phenomenon I describe is conventionally termed a critical, or sensitive, period for language acquisition. In general, of course, competence in most domains increases over development: Characteristically, behavioral skills do not worsen over age; rather, they increase. In contrast to this general developmental pattern, domains in which there are critical periods are striking in that, in these domains, there is a more limited period, typically early in life, in which the ability to learn is at its peak, with a declining developmental function after this period. Although I focus primarily on the