LAUSR

Abstract Perceivers tend to strongly agree about the basic trait information that they encode from faces. Although some research has found significant consistency for social inferences from faces viewed at multiple angles, disrupting configural processing can substantially alter the traits attributed to faces. Here, we reconciled these findings by examining how disruptions to configural processing (via face inversion) selectively impairs trait inferences from faces. Across four studies (including a pre-registered replication), we found that inverting faces disrupted inferences about particularly human-relevant traits (trustworthiness and humanness) more than it did for a trait relevant to both human and non-human animals (dominance). These findings contribute to emerging research linking configural processing to the humanization of social targets, helping to provide a clearer understanding of how visual cognition may moderate perceptions of humanness.