This discussion of vervet monkeys is from a 1984 book about animal communication.
Vervet monkeys have at least three different categories of alarm calls. When a leopard or other large carnivorous mammal approaches, the monkeys give one type of alarm call; quite a different call is used at the sight of a martial eagle, one of the few flying predators that capturesvervet monkeys. A third type of alarm call is given when a large snake approaches the group. This degree of differentiation of alarm calls is not unique, although it has been described in only a few kinds of animals. When ethologists, who study animal behavior, interpret data of this kind, they require proof that variations in animal communication signals convey anything more than information about the communicator's internal state.
The first and relatively simple question is whether the vervet monkey's three types of alarm calls convey to other monkeys information about the type of predator. Such information is important, because the animal's defensive tactics are different in the three cases. When a leopard approaches, the monkeys climb into trees. But leopards are good climbers, so the monkeys can escape them only by climbing out onto the smallest branches, which are too weak to support a leopard. When the monkeys see a martial eagle, they move into thick vegetation close to a tree trunk or at ground level. Thus the tactics that help escape from a leopard make them highly vulnerable to a martial eagle, and vice versa. In response to the threat of a large snake, they stand on their hind legs and look around to locate the snake, then simply move away from it, either along the ground or by climbing into a tree.
Knowing that the monkeys give different alarm calls when they see different predators does not establish beyond a doubt that the calls actually describe the type of predator. When the monkeys, which are usually close to each other, hear an alarm call, each one quickly looks around at the caller. Like many other animals, they are adept at judging the direction in which another animal is looking, so they can easily see what the caller is looking at. This serves much the same function as pointing. When monkeys other than the caller take the appropriate action to avoid the danger, it is difficult to be sure whether they are acting solely on the basis of the call or whether the call simply led them to look at the source of the danger.
To clarify this situation, researchers conducted some carefully controlled playback experiments under natural conditions. The basic idea was to play from a concealed loudspeaker tape recordings of vervet alarm calls when vervets had just seen a leopard, a martial eagle, or a large python, and to inquire whether these playbacks, in the absence of a predator, would elicit the normal response. The experiments required many precautions and refinements. For instance, vervet monkeys come to know each other as individuals, not only by visual appearance but by minor differences in their vocalizations. They might not respond even to an alarm call recorded from one of their own companions if that individual was in plain sight some distance from the vegetation concealing the speaker. In all experiments, the loudspeaker reproduced calls of a member of the group, and the speaker was hidden in a place when the monkeys would expect that individual to be. The experiments had to be prepared with tape recordings of a known member of a well-studied group of vervet monkeys and a hidden speaker located where this individual frequently spends time.
When all these conditions were satisfied, the playback of alarm calls did indeed elicit the appropriate responses. The monkeys responded to the leopard alarm call by climbing into the nearest tree; the martial eagle alarm caused them to dive into thick vegetation; and the python alarm produced the typical behavior of standing on the hind legs and looking all around for the nonexistent snake. Not all ethologists have accepted the straightforward interpretation that the alarm calls convey information about the type of predator. One alternative interpretation is that the alarm calls are injunctions to behave in certain ways. Thus the leopard alarm might mean "Go climb into a tree. But even this interpretation necessarily ascribes three specific types of injunction to the vocabulary of vervet monkeys. Even such postulated injunctions would be more than a simple reflection of the internal state of the communicator.
The development of alarm call behaviour in mammals and birds[2]
Rather than being age adaptive, young might alarm call toa greater number of species than adults because they make moreerrors. For example, Seyfarth& Cheney (1980) showed that alarmcalls normally given by adult vervet monkeys to threatening stimuliare often uttered by young in response to nondangerous stimuli,such as doves (Streptopelia sp.), falling leaves or warthogs, Phacochoerusafricanus, although these overgeneralizations are notentirely random: infants give ‘leopard’ alarms primarily to terrestrialmammals, ‘eagle’ alarms to objects in the air, and ‘snake’alarms to snakes or long thin objects on the ground. As they growolder, vervets restrict their calling to particular predator specieswithin these general classes. Although little is known about theneurological basis of alarm calling (see above), it could be that theincreased vulnerability of young results in nervous systems andbrain structures that are more sensitive to external stimulationthan those of adults (Wiedenmayer 2009). The threshold for callingmight therefore be lower in young, resulting in the production ofalarm calls in contexts that would not elicit adult calling; developmentof adult-like call usage might therefore arise because ofneurological maturation.
Alternatively, adult-like call usage might develop throughincreased experience with predator encounters and the alarm callsof others: individuals could learn to make fewer mistakes andbecome better at conveying the correct information. By hearingothers use specific calls only in certain contexts, young could learnby association. There is also some evidence that adults mightreinforce correct alarm call usage by juveniles. For example, adultvervet monkeys are more likely to give second alarm calls wheninfants alarm call to known predators than to nonpredators (Seyfarth& Cheney 1980), although no data are available on whetherthis leads to more rapid development. Moreover, if reinforcementdoes play a role, it may be inadvertent from an adult’s perspectivebecause they are equally likely to give second alarms followinga correct alarm call by another adult as they are following a correctalarm call by a juvenile (Seyfarth& Cheney 1986). So, while adultresponses may facilitate learning of correct alarm call usage, adultsdo not appear actively to teach infants in this context (Thornton &Raihani 2008).
