The illusion of conformity: when independent learning mimics social influence.

Although individuals tested in isolation often show substantial behavioral variation, they tend to behave more uniformly in groups, a pattern commonly attributed to social conformity (ie, a form of social learning in which individuals preferentially adopt the most common trait). However, evidence for social conformity in nonhuman animals remains debated, as group-level convergence can also arise from independent learning, where individuals adopt the trait maximizing their fitness. Such learning can generate either adaptive phenotypic convergence or pseudo-conformity when aligning with the group's average strategy provides a selective advantage (eg, by reducing predation risk). The likelihood of these outcomes is shaped by ecological and social factors, yet their relative contributions remain poorly understood, as most studies rely on simplified laboratory conditions. This study explores the conditions under which conformist-like patterns can emerge solely from individual learning, and how costs of deviating from the group strategy and tradeoffs between different selective pressures influence their evolution. I developed an evolutionary model in which individuals sequentially adjust their exploratory behavior through a trial-and-error learning process. At each learning step, an individual explores 2 small behavioral deviations around its current exploration level and retains the variant that yields the highest payoff under the prevailing ecological conditions. The results show that the likelihood of pseudo-conformity depends on the interplay among ecological conditions, group composition and behavioral flexibility, and is particularly favored when predation risk and the energetic costs of escaping predator attacks are high. Moreover, under some conditions, pseudo-conformity may maintain behavioral diversity at the population level by allowing different subgroups to converge on and stabilize around distinct behavioral phenotypes. Overall, these findings underscore that measuring traits in isolation may misrepresent both behavioral variation and evolutionary dynamics.