Gergely Csibra: "Tracking social relations and evaluating potential partners in infancy and childhood"

Professor Gergely Csibra, Central European University, Wien, Austria:
"Tracking social relations and evaluating potential partners in infancy and childhood"

Abstract
Our studies on how children understand third-party social interactions were guided by the following hypotheses: (1) young humans interpret social interactions via an extended version of the “naïve utility calculus” they apply for understanding individual instrumental actions; (2) they conceive social interactions as indicative of underlying social relations rather than of individual traits; and consequently (3) they do not rely on trait attribution for choosing partners for cooperation until middle childhood. While our findings did not support the first hypothesis, they robustly corroborated the other two.

Our experiments found no evidence that infants understand helping as a second-order goal aimed at increasing another agent’s utility. This is in contrast with evidence suggesting that infants (like adults) understand joint actions as guided by minimization of aggregate costs for participating agents. Additionally, we were unable to replicate a well-known preference for helpers in early infancy (consistently with a recent large-scale multi-lab replication attempt). From these findings we concluded that mature understanding of helping actions is not present in infancy and likely develops later from representing simpler forms of collaboration, such as joint actions. Although we did not chart the detailed development of this capacity, we found that, by age 3 to 4 years, children adopt an adult-like concept of helping that relies on utility calculus.

In contrast to the failure to interpret social interactions based on their cost-benefit distribution, we obtained abundant evidence that infants have a principled understanding not only of cooperative actions but also of leader-follower and transfer-based interactions. Specifically, we demonstrated that infants recognize transfer events as instances of giving even in the absence of cues of the recipient’s involvement, and draw rich relational inferences from the observation of such events. They track agents and patients beyond the original transfer event, encode the direction of the transfer, and view reciprocal exchange as consistent with the action. This suggests that infants interpret an action if giving as reflecting an underlying relationship rather than an individual disposition. We found similar reluctance to make dispositional inferences for dominance relations as well.

We directly tested the mainstream conjecture that children select social partners based on dispositional attributions in a foraging game, in which children chose partners from a set of agents whose behavior they had previously observed. Even the youngest children (5-year-olds) recognized the features of the agents’ behavior that could potentially be exploited in a cooperative situation, such as the agent’s speed and its contribution to the partner’s success (i.e., helping). However, children did not rely on these behavioral features in choosing good cooperative partners until about the age of 7 years.

These results confirm that while infants display a sophisticated understanding of social relations and preschoolers conceive helping actions as fulfilling the second-order goal of increasing the utility of someone else’s action, children do not couch these inferences in terms of manifestations individual traits until school age.