Social learning about places: observers may need to detect both social alarm and its cause to learn

It is widely established that social alarm signals trigger learning about discrete stimuli present at the same time. Such learning facilitates, for example, acquisition of responses to novel predators and has the functional advantage that individuals avoid exposing themselves to a potentially risky situation. Avoidance of potential danger might equally apply to learning about risky places, but would require social alarm signals to trigger learning about contextual cues, rather than discrete stimuli. Here, we tested this hypothesis by analysing the behaviour of experimental observer Indian mynahs, Acridotheres tristis, both before and after they had watched demonstrator mynahs showing alarm behaviour at a foraging site where observers were accustomed to feeding. To isolate changes specifically attributable to the behaviour of demonstrators, we compared this groupʹs post-training behaviour with that of a control group, which watched social companions foraging at the feeding site. Unexpectedly, we found no evidence that experimental observers became more wary of the feeding site after observational training relative to control observers, suggesting that social alarm signals do not trigger learning about the location in which an alarmed individual is observed. In light of previous work in our laboratory showing that Indian mynahs become more wary in a place in which they have observed a predator attack on a social companion, we suggest that social learning about places may require observation of both social alarm and its cause.