Effects of stimulus contingency in infant-adult interactions

Two experiments manipulated contingent aspects of adult face-to-face displays using a TV-replay procedure, after Murray and Trevarthen (1985), to examine social-releaser and social-expectancy models of infant social interaction. In Experiment 1, 5-month-olds and their mothers interacted by viewing each other over closed-circuit television. One group of infants (n = 14) viewed a replay of the mothers' behavior during a 90-s period, interposed between two normal 90-s interaction periods; a control group (n = 14) received normal interactions during ail three periods. No group differences were obtained, which could be explained by either model. In Experiment 2, a stranger interacted with the infants to eliminate maternal familiarity as a factor. During three 2-min periods, 4- to 6-month-olds (n = 32) were presented with live-contingent, televised-contingent, and televised-noncontingent face-to-face interactions in one of four different orders. Infants gazed less at TV-noncontingent than either live- or TV-contingent displays, and there was evidence that noncontingency had a long-lasting effect on gaze. They were equally attentive to both contingent displays, but smiled more to live than to televised displays. An order effect revealed that smiling depended on the nature of both the current and prior interaction. These results are more complex than would be predicted by the social-releaser model. Taken together, the results of the two studies show evidence that infants form expectancies about interactions based on past experience (in Experiment 1, long-term experience interacting with the mother, and in Experiment 2, the initial period of interaction with the stranger). Violations of these expectancies appear to have a lasting effect on visual attention, but only a temporary effect on smiling.