Memory for Expectation-Violating Concepts: The Effects of Agents and Cultural Familiarity

Previous research has shown that ideas which violate our expectations, such as schema-inconsistent concepts, enjoy privileged status in terms of memorability. In our study, memory for concepts that violate cultural (cultural schema-level) expectations (e.g., ‘‘illiterate teacher’’, ‘‘wooden bottle’’, or ‘‘thorny grass’’) versus domain-level (ontological) expectations (e.g., ‘‘speaking cat’’, ‘‘jumping maple’’, or ‘‘melting teacher’’) was examined. Concepts that violate cultural expectations, or counter-schematic, were remembered to a greater extent compared with concepts that violate ontological expectations and with intuitive concepts (e.g., ‘‘galloping pony’’, ‘‘drying orchid’’, or ‘‘convertible car’’), in both immediate recall, and delayed recognition tests. Importantly, concepts related to agents showed a memory advantage over concepts not pertaining to agents, but this was true only for expectation-violating concepts. Our results imply that intuitive, everyday concepts are equally attractive and memorable regardless of the presence or absence of agents. However, concepts that violate our expectations (cultural-schema or domain-level) are more memorable when pertaining to agents (humans and animals) than to non-agents (plants or objects/artifacts). We conclude that due to their evolutionary salience, cultural ideas which combine expectancy violations and the involvement of an agent are especially memorable and thus have an enhanced probability of being successfully propagated.