Copying others can be a way of getting along, not just getting facts
Most research on animal social learning has treated strategies like conformity as ways to acquire accurate information about the environment. A 2024 review argues that this focus has overlooked a second payoff: behaving like others can itself bring social advantages, and non-human animals may deploy behavior matching strategically to increase affiliation [1]. In other words, the same act of copying can serve two purposes at once, and the social one has been understudied in animals.
The evidence for this comes from several lines of work reviewed together: homophily (associating with similar others), imitation, and rapid facial mimicry. Across these, the review finds evidence suggesting that behaving like others affords social benefits to non-human animals [1]. The claim is framed as a proposal about adaptive function, not a proven universal purpose, so the social payoff should be read as a plausible and supported possibility rather than a settled rule.
For a newcomer, adopting the group's ways may be a route to acceptance
The clearest case is immigration. The review examines conformity in dispersers, animals that leave one group and join another, and suggests that forgoing one's own personal information or preferences in favor of the new group's may be a strategy to facilitate social integration [1]. Here the learning is not mainly about which food or route is correct; it is about signaling membership in a group that already has established ways.
This reframes conformity as potentially social rather than purely informational. The review proposes that informational and social functions of conformity apply to humans and animals alike [1]. It also raises open questions, such as the conditions under which animals use informational versus social conformity, and what role uncertainty plays for immigrant individuals [1].
When does learning count as social rather than informational?
A separate line of thinking asks what makes social learning distinctively social in the first place. One account proposes that learning is distinctive when the input mechanisms, meaning perceptual, attentional, and motivational processes, are biased or tuned to a particular channel of social information [2]. That is a claim about how the machinery of learning is wired toward other agents, not a direct claim about social payoffs.
Putting these together, the social side of learning can show up in two places: in the motives and mechanisms that tune an animal toward others [2], and in the consequences of matching others, such as affiliation and integration [1]. The review's proposal that animals may choose social over informational learning under some conditions, and its question about the role of uncertainty, remain open research questions rather than resolved findings [1].
What this does not settle
The social-benefit account is a proposal built from reviewed evidence on homophily, imitation, mimicry, and disperser conformity, and the authors themselves frame it as a perspective meant to generate new research questions [1]. It does not establish how often animals learn for social rather than informational reasons, or how the two motives trade off. The conditions under which social conformity dominates, and the role of uncertainty, are explicitly left as questions for future work [1].
Sources used in this answer
The forgotten adaptive social benefits of social learning in animals
A review argues that conformity and behavior matching in non-human animals can provide social benefits such as affiliation, and that dispersers may adopt a new group's behavior to facilitate social integration, with open questions about when animals use informational versus social conformity.
What's social about social learning?
Social learning is proposed to be distinctive when perceptual, attentional, and motivational input mechanisms are biased or tuned to a particular channel of social information.
The Hazards and Benefits of Condescension in Social Learning
In a sequential social learning model, mild condescension (perceiving peers' private information as lower quality than it is) improves outcomes, while too much condescension or anti-condescension worsens them.
Dynamic social learning in temporally and spatially variable environments
The paper examines dynamic social learning in temporally and spatially variable environments, analyzing how individuals adapt learning strategies as conditions change.
Steering the herd: A framework for LLM-based control of social learning
The paper presents a theoretical framework in which an information-mediating planner such as an LLM controls information in social learning to shape social welfare or agent behavior.