When an African vervet monkey spies a snake in the grass, it understandably screams. The obvious similarity to a human scream of terror led ethologists for decades to assume the vocalizations carried only the same entirely emotional meaning. But in 1967, Thomas Struhsaker, then of the University of California, Berkeley, called that obvious conclusion into question. He described three distinct vervet calls and showed that the call a particular monkey gave depended not on how startled it was, but on what kind of predator it saw. The degree of fear or surprise further modified these calls, but Struhsaker could recognize the three calls by sound and by the way other vervets reacted.

A vervet gave one call when it saw a snake. Other members of the troupe then stood on their hind legs and scanned the ground. A second type of call always followed the sighting of a leopard. Other members of the troupe immediately climbed to the smallest branches of nearby trees, safe from the heavy leopard. Lastly, a vervet called yet a third way when it saw a martial eagle cruising the sky. Clinging to the outer branches of a tree, or standing tall in the grass would leave the monkeys vulnerable to this attack. Instead they climbed the tree, but stayed near the trunk, deep in the tree, or alternatively dove into dense bushes.

Vervet monkeys live in close-knit groups in forests and open areas. Because the groups remain stable over long periods, communication between individuals seems valuable to any individual. Struhsaker’s discovery of three types of calls that triggered three responses, and that furthermore could be modified by degree, argued for a level of communication beyond a mere shout of fear. But perhaps the responses of troupe members represented merely “monkey see, monkey do.” Hearing a snake-call, a monkey might look up, see the calling monkey standing and scanning the grass and just follow suit. Robert Seyfarth, Dorothy Cheney and Peter Marler, all at Rockefeller University at the time, began a series of attempts to manipulate the vervet communication system to answer the question “Do vervet calls contain information?”

Seeing an animal’s call as an attempt to refer to an object, rather than as a simple emotional reaction was a radical departure from past thought, so the team took great precautions to avoid pitfalls in their research. They recorded many alarm calls and arranged to play them back to a vervet troupe. They spent time with the monkeys, allowing the troupe to become accustomed to their presence. They took care not to play an alarm call when the monkey recorded was in clear sight and obviously not alarmed. They waited until the monkeys were quiet and had no real dangers about. Lastly, they filmed the responses to playbacks. Those categorizing the behavior seen on films did not know what calls had triggered the behavior.

The monkeys responded to playbacks as they had to the original calls. The snake-in-the-grass call caused troupe members to stand up and scan the ground. The leopard-call sent them to the farthest reaches of tree limbs and the eagle-call triggered a retreat into bushes or the middle of trees. The first monkey to respond, at least, could not be imitating, because there was no one to imitate. And the troupe’s response related to the call itself, not to a real danger, because there was no danger. These results, while forcing a dramatic change in ethnologists’ assumptions about animal calls, did not assign meanings to the calls. A snake-call could mean “snake,” or “look on the ground,” or may not even have a “meaning” in the sense that human words have meaning.

In a separate experiment, Cheney and Seyfarth played recorded alarm calls of young vervets. Adult females responded by looking at the infant’s mother, not at the infant. This behavior implies they can both identify an infant by its distinctive call and can understand the relationship between mother and offspring.

Cheney and Seyfarth went on to analyze the more subtle vervet calls called grunts. With practice, humans could detect differences in the grunts, but previous researchers had assumed these differences to depend on different situations, not on different meanings. Cheney and Seyfarth tested this assumption, again with playback experiments. Vervets responded differently and consistently to different grunts, no matter what context the researchers chose for playback. When they heard grunts recorded from a dominant male, the vervets moved away from the speaker. On the other hand, when they heard grunts of a subordinate to a dominant male, they never moved. Again, the experiments do not assign a meaning to the grunts. But they do show that the grunts convey some meaning; they are not all equivalent.

Next: Macaques