4. Discussion
This study provides evidence that abstract attributions of intentional agency engage automatic reorienting of covert spatial attention in both preschool-aged children and adults. Children’s most rapid, automatic social behaviors do not appear to be limited to overlearned responses to specific, perceptually defined stimuli. The present results suggest that these behaviors can also be informed by children’s rich, conceptual appraisals of the entities that they encounter. The effects we observed were not due to globally heightened attention in the Socially Contingent condition. Children attended equally to each condition’s familiarization video, and adults were even slightly more interested in the NonContingent familiarization video; moreover, both ages contributed an equal number of acceptable cueing trials across conditions. Nor can perceptual learning within the study session account for this finding, as both the entity’s rotational motion and its non-predictive nature were constant across conditions. Rather, subjects were only cued by the entity’s turn when they viewed it as an intentional agent whose orientation provided a meaningful directional, and perhaps communicative, signal. This finding both supports and extends previous research on the detection and representation of intentional agents. Young children who view an entity as an intentional agent may interpret its behavior as goal directed (Johnson, Booth, & O’Hearn, 2001; Shimizu & Johnson, 2004) and produce voluntary social responses toward it, such as gaze following, imitation, and helping (Beier & Carey, 2014; Johnson et al.,2001; Johnson, Ok, & Luo, 2007; Johnson et al., 1998; Kenward & Gredebäck, 2013; Movellan & Watson, 1987). The present results suggest that in addition to these deliberate behaviors, representations of intentional agents also guide children’s most rapid, automatic social behaviors, such as the cueing of covert spatial attention. This occurs spontaneously (because subjects received neither instruction nor demonstration of the entity’s directionality) and reflexively (because subjects were explicitly told that the entity’s turns did not predict target locations).