Field experiments with wild primates reveal no consistent dominance-based bias in social learning

Abstract

Directed social learning suggests that information flows through social groups in a nonrandom way, with individuals biased to obtain information from certain conspecifics. A bias to copy the behaviour of more dominant individuals has been demonstrated in captive chimpanzees, but has yet to be studied in any wild animal population. To test for this bias using a field experiment, one dominant and one low-ranking female in each of three groups of wild vervet monkeys was trained on alternative methods of opening an 'artificial fruit'. Following 100 2 demonstrations from each model, fruits that could be opened either way were presented to each group and all openings were recorded. Overall, the dominant females were not attended to more than low-ranking females during the demonstrations, nor were their methods preferentially used in the test phase. We conclude that these monkeys show no overall bias to copy high-ranking models that would lead to a high-ranking model's behaviour becoming more prevalent in the group than a behaviour demonstrated by a low-ranking model. However by contrast, there were significant effects of observer monkeys' rank and sex upon the likelihood they would match the dominant model. Additionally we found that the dominant models were more likely to stick to their initially learned method than were low-ranking models.

Wild primates copy higher-ranked individuals in a social transmission experiment

“…Our results offer strong evidence that new box opening techniques spread through social transmission in groups of wild vervet monkeys, consistent with previous findings using two-option designs 27,44 . We found that (i) observing one option increases the probability of learning it, (ii) individuals who learned one option are 31x are more likely to subsequently learn the other option, and (iii) females tended to socially learn twice as fast as males.…”

“…The single performance bias towards higher rankers does not explain the overall bias from higher rankers to lower rankers because we found evidence of a social information use bias, with observations of higher-ranked individuals having a greater effect on the rate at which behavior is socially learned. This is in accordance with an earlier study 27 , in which it was found that the dominant models' demonstrations did not elicit more observations than the lowranking ones, but interestingly they reported high-ranking females showed a significant bias to copy the dominant female. In that study the authors used a two-option design and trained two models of differing ranks in each group: one was the highestranked female and the other one was a female of mid to low rank.…”

“…Field experiments reported multiple social-learning strategies in wild vervet monkeys in which infants processed food in the same way as their mother 23,24 , females-the philopatric sex-were preferred as models than males 25 and dispersing males conformed to the local foraging norm, abandoning their native preference 26 . By contrast, wild vervet monkeys showed no preference to copy a highranking adult female compared to a lower-ranking one 27 . Capuchins have been found to copy older individuals with more successful fruit opening techniques 28 while in vervet monkeys, males, but not females, were more likely to copy the male model receiving a higher payoff in a two-action learning task 29 .…”

“…An individual might be copied more than other potential demonstrators because it performs the behavior more (performance bias), and/or because its actions are more likely to be observed than the actions of others (attention bias). For example, Older and higher-ranking capuchins have been found to be more frequently watched by conspecifics while cracking nuts 30 ; vervet monkeys have been reported to show a greater attention to female models 25 but not to dominants 27,31 , and juveniles have been reported to pay more attention to their maternal relatives when foraging 31 . The performance and attention bias effects multiplied together determine the rate at which an individual is observed.…”

Untitled

“…neuroscience (Woods et al 2011;Mendell et al 2014), social learning (Botting et al 2018) and sexual selection (Borgeaud et al 2015;Rodríguez et al 2015), and these studies require a clear understanding of genetic boundaries and connectivity between populations, using both neutral and adaptive markers. Vervet monkeys also often find their way to primate rehabilitation centres following human primate conflicts (Wimberger et al 2010) and re-introduction of these animals to suitable recipient populations without considering the provenance of the rehabilitated animals should be discouraged.…”