Selective Social Learning

This research area focuses on how individuals acquire and utilize information about plants over the course of ontogeny. Trial and error learning, in which each individual directly samples different plant species and experiences the consequences, is not an effective way to learn about plants. Given that many plants are poisonous, this strategy could result in frequent illness and perhaps even death. Therefore, we propose that human cognitive architecture contains specialized social learning rules that facilitate the safe acquisition of information about plants from more knowledgeable individuals, and enable that learned information to be used in new circumstances.

Selective Social Learning
© MPI fuer Bildungsforschung



The problem of learning what to eat is of course broader than simply identifying edible plants. However, identifying edible plants is a particularly interesting and informative problem. Humans have relied on gathered plant resources for food throughout our evolutionary history, but plants come in all shapes and sizes and there are no visual features that reliably indicate whether or not they are safe for humans to eat. Therefore, the best solution to this problem is to learn edibility information from more knowledgeable individuals. Our work in this research area investigates how infants use social signals from others to learn about plant edibility, and whether these social signals are used in the same way for plants as for other entities. Our first findings suggest that infants use social information to selectively learn plant edibility (Wertz & Wynn, 2014a) and systematically extend socially learned edibility information to similar-looking plants (Wertz & Wynn, under review). Further, these findings suggest that human food learning systems are sensitive to the particular social information conveyed as well as the type of entity being acted upon. Future planned studies in this project will investigate the scope of selective social learning about plant foods. 


Learning information about a single plant, or a single entity of any kind, only gets you so far. Generalizing one instance of social learning to a broader array of entities is a critical part of the learning process and requires that two entities be recognized as membersof the same category. For example, if an infant learns that a particular apple tree has edible fruit, they can only use the learned information outside of that particular instance if they are able to identify the other apples trees in the environment and correctly apply the learned information to them. However, the processes infants use to categorize plants—and thus appropriately extend socially learned information about edibility or danger beyond the single initial social event—remain unknown. This project seeks to explore this question. First, we are interested in what features infants use to distinguish plants from other types of objects. Our previously conducted studies indicate that infants do not appear to use simple features in isolation (e.g., color). Infants are reluctant to touch green plants, but readily grab green manmade objects. Our current studies aim to identify the features infants use to categorize an entity as a plant, and seek to uncover the features infants use to distinguish one type of plant from another—processes that are critical for making appropriate inferences about plants. 


Wertz, A. E., & Wynn, K. (2014). Selective social learning of plant edibility in 6- and 18-month-old infants. Psychological Science25, 874-882. doi:10.1177/0956797613516145