Helping behaviors rely on explicit but not implicit inferences of affect

Contemporary models of emotion highlight the fundamental function of emotional expressions in eliciting adaptive social behaviors. Critically, past research testing the social effects of emotional expressions has primarily relied on simulated expressions that may not capture the complexity of real-world emotional displays. Across two studies (N=1,534), we examined how naturally diverse vocalizations influence prosocial behavior. In Study 1, authentic videos of crying individuals were synched with typical crying, atypical crying, typical laughter, or no vocalization. When participants were asked to make affective inferences (Study 1A), vocalization type influenced their perceived valence judgments. However, when participants made donation decisions without explicit inference instructions (Study 1B), vocalization type did not impact donations, except when audiovisual signals were affectively conflicting. In Study 2, we examined the role of explicit inferences of affect in facilitating the link between naturalistic expressions and prosocial behaviors. We found that when participants first made explicit inferences about others’ affective states, they donated more money overall and their helping behaviors were more sensitive to vocalization types. However, when participants donated without providing explicit inferences first, their helping behaviors were less influenced by vocal cues and donation amounts were lower overall. Collectively, these findings highlight the role of explicit inferential processing in establishing the link between naturalistic emotional expressions and social behavior.