Learning from watching alarmed demonstrators: does the cause of alarm matter?

Many social-learning opportunities expose animals to the behaviour of conspecifics, but also to causes and consequences of those behaviours. Attending to information over and above social behaviour per se may provide a strategy by which the reliability of social information is ensured. Earlier work in Indian mynahs, Sturnus tristis, has demonstrated that observers become more wary in a location in which they are accustomed to foraging after they have viewed a conspecific undergo a ‘predator’ attack at that site. We determined whether observation of both an alarmed demonstrator and the cause of the conspecific’s alarm (capture by a human) were critical to such observational learning. Experimental observers watched a demonstrator mynah display high levels of alarm in response to pursuit and capture by a human, while control mynahs watched a demonstrator express a similar level of alarm to a threatening nearby human, but visual access to the human by observers was blocked. Analysis of observer behaviour at the feeding site both before and after observational training revealed that experimental observers remained wary at the feeding site after training relative to before, relative to control observers that became far less wary, strongly suggesting that both social and causal information were important for observational learning. This result contributes to the growing body of empirical evidence that use of social learning is modulated by a rich variety of contextual information that may help ensure that its use is adaptive.