Conceptual maps of social knowledge support trait generalization

Read the full article

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

People robustly and rapidly infer traits from the social behavior of others. These trait impressions support predictions about future behavior. How do people use this learning flexibly in novel social situations? We propose people use maps of social knowledge (representing similarity between traits) to generalize trait impressions and predict behavior in new domains and contexts, a prediction we test across three experiments. First, in a pilot study, we demonstrate that conceptual similarity between traits predicts situation-specific inferences of trait relevance. In Experiment 1, participants learned about targets (described either as humans or non-human slot machines) whose behavior in a task revealed sociability, and later, made choices about which partners to choose in decision making tasks relevant to other traits. Trait similarity (i.e., between the original context and novel contexts) shaped partner choice, but critically, this relationship was socially specific: participants generalized social learning based on trait similarity when interacting with other humans but not with slot machines. In Experiment 2, participants were given explicit trait labels describing potential partners. Here, we observed that trait similarity continued to influence generalization across an expanded set of traits and task contexts, above and beyond the mere valence similarity of traits. While Exps. 1-2 demonstrate trait-based generalization in decision-making, Experiment 3 demonstrated that trait-based generalization occurs with relative spontaneity. Here, in a false recognition task, perceived similarity between lure and target traits was associated with falsely remembering lure traits. Overall, our results demonstrate the social specificity of traitbased generalization and provide a foundation for future work exploring this process.

Article activity feed