Brief Familiarization Primes Covert Imitation in 9-month-old Infants

Brief Familiarization Primes Covert Imitation in 9-month-old Infants Bennett I. Bertenthal (bbertent@indiana.edu) Ty W. Boyer (tywboyer@indiana.edu) Department of Psychological and Brain Sciences 1101 E. Tenth St. Bloomington, IN 47405 USA Abstract Previous research reveals that 9-month-old infants who passively observe an experimenter search repeatedly for a toy in the Piagetian A-not-B error task covertly imitate these actions and manually search incorrectly when the toy is hidden in the B-location. Two experiments tested whether infants would also search incorrectly if the experimenter was replaced by a pair of mechanical claws or if the experimenter performed less familiar actions. Although infants did not commit the search error when tested directly without any familiarization to the novel actions, a significant majority of infants committed the search error following two minutes of familiarization with the actions performed on the A trials. These results converge to suggest that infants’ brief experiences with observing actions will facilitate the activation of a corresponding motor representation. Furthermore, the specific process by which this facilitation occurs varies with the similarity between the observed action and its motor representation. Keywords: Imitation; action observation; infancy; learning; motor simulation; A-not-B search error. Introduction During the first year of life, virtually every situation affords infants the opportunity to learn something new about themselves, other people or their surroundings. Much of this learning occurs through the observation of actions produced by others as well as the self. There is currently a range of opinions concerning the extent to which infants’ action representations derive from experience vs. innately specified core knowledge (Csibra, Gergely, Biro, Koos, & Brockbank, 1999; Woodward, 2009). Nevertheless, regardless of theoretical perspective, there is general consensus that infants’ action representations are enriched and elaborated over the first year. One intriguing explanation for how infants learn about goal-directed actions is that the perception and understanding of these actions is developmentally related to self-produced experience with these same actions (Rakison & Woodward, 2008). For example, it is now well documented that infants interpret actions as goal directed by five to six months of age, which is roughly the same age at which they begin to successfully reach for distal objects (Bertenthal & von Hofsten, 1998). Likewise, 9-month-old infants who understand the referent of a point are likely to also be able to point at distal objects (Brune & Woodward, 2007). At 10 months of age, infants who are capable of pulling a cloth to retrieve a toy are more likely to understand the means-end structure of a hierarchical action (Sommerville & Woodward, 2005). In spite of considerable evidence supporting a developmental relation between the perception and production of human actions, the vast majority of studies are correlative, and thus not informative about the underlying mechanisms. In order to draw firmer conclusions about the causal relation between perception and production of goal-directed actions, it is necessary to experimentally manipulate either the perceptual experience or the motor experience of infants and test how this manipulation affects performance in the other domain. A few recent studies manipulating infants experience with reaching have now been conducted. Although 3-month-old infants are not yet capable of reaching and grasping distal objects, they can learn to retrieve these objects by swiping at them with ‘sticky’ mittens (Needham, Barret, & Peterman, 2002). When infants were given practice with these sticky mittens prior to testing their perception of others’ mittened reaches as goal- directed, they interpreted the reaches as goal directed rather than as simple movements (Sommerville, Woodward, & Needham, 2005). By contrast infants given practice following the test did not interpret the reaches as goal directed. In a related study, Sommerville, Hidebrand, and Crane (2008) compared the effects of active vs. observational experience on 10-month-old infants’ ability to identify the goal of a novel means-end task and reported that infants were more likely to understand the means-end task after receiving active as opposed to observational experience with the relevant action. These results suggest that active, but not passive, experience facilitates infants’ learning about goal-directed actions. It is difficult to know whether the preceding conclusion generalizes beyond the specific paradigms that were used. Hofer, Hauf & Aschersleben, (2005) tested 9- and 12- month-old infants and reported that they interpret actions performed by a mechanical claw as goal directed, but the younger age group was successful only after observing an experimenter demonstrate the operation of the claws. In this case, observational learning was sufficient to change infants’ understanding of whether or not the claws acted in a goal-directed fashion. Moreover, Daum, Prinz, and Ascherleben (2009) report evidence that appears to directly challenge the findings of Sommerville et al. (2008). – The ability of infants to perform a hierarchically organized action was not predictive of their understanding of someone else performing the same action. These inconsistencies between studies are quite likely attributable to differences in the cognitive demands of the task (cf. Daum et al., 2009).