Assessment of meaning and the detection of unreliable signals by vervet monkeys

Free-ranging vervet monkeys, Cercopithecus aethiops, who had learned to ignore playbacks of one type of call by an unreliable signaller subsequently also ignored playback of an acoustically different call by the same individual if the calls had similar referents. Such transfer did not occur if either the identity of the signaller changed or if the two calls had different referents. After repeatedly being played an intergroup call in the absence of other groups, vervets also ignored an acoustically different intergroup call given by the same individual. The monkeys did not transfer habituation, however, if the new call was given by a different individual. In tests where the referents of two calls were different, the monkeys also failed to transfer habituation across call types. Vervet monkeys who had learned to ignore an unreliable leopard alarm call did not later ignore an eagle alarm call, even when the signaller remained the same. Results suggest that vervets, like humans, process information at a semantic, and not just an acoustic, level. Humans make judgments about the similarity or difference between words on the basis of an abstraction, their meaning. By contrast, it is often assumed that animals respond to signals not according to some abstract feature such as meaning but according to the signals' physical features (e.g. Morton 1977). Is this dichotomy valid? One method for determining how group-living animals assess the meaning of calls is through their ability to detect anomalous or unreliable signals. For example, if signallers occasionally attempt to deceive others by using multiple signals with the same information (Andersson 1980), selection should favour the ability of recipients to compare signals on the basis of their meaning, and to transfer information about the reliability of a signaller's calls from one context to another. We investigated whether vervet monkeys, Cercopithecus aethiops, were capable of transferring information gained in one context to another context by repeatedly playing recordings of either the intergroup call or the alarm call of one individual to establish that individual as an unreliable signaller. We then tested whether subjects attended to this 'unreliable' individual in other contexts by playing recordings of acoustically different calls whose referents were either similar to or different from the habituating stimulus. Subjects' responses to 'unreliable' individuals were compared with their responses to group members whose calls had not been played previously.