Wednesday, March 16, 2011

The Semantics of Vervet Monkey Alarm Calls: Part II - The Experiment

Last week, I blogged about the semantics of alarm calls in vervet monkeys. This post will focus solely on the ingenious experiment by Robert Seyfarth, Dorothy Cheney and Peter Marler (1980) to test whether vervet monkey alarm calls convey information or if these calls were just an uncontrollable auditory response to predators. Their question was simple. Would vervet monkey alarm calls alone elicit different responses?

A vervet monkey. Photo from Wikipedia.
Here's what the researchers did in the field. They used playbacks of recorded vervet monkey (subjects) alarm calls from concealed speakers. Equal amount of alarm calls for leopard, eagle and snake were used. These alarm calls were recorded from known adult male, adult female and juvenile vervet monkeys in the field. Trials were done when subjects were on the ground and also when they were in the trees. These trials were conducted in the absence of predators to eliminate visual cues from the caller.

Alarm calls were broadcasted in different amplitudes to mimic natural alarm calls. In succession from loudest to lowest amplitudes are alarm calls for leopard, eagle and snake. Subsequently, leopard calls have the lowest pitch while snake calls have the highest pitch. To control for the possible effects of amplitude, the researchers broadcasted alarm calls that do not differ significantly in the amplitudes for all three predators.

Table from Seyfarth et al. (1980). Click on illustration for its original size.
The alarm call playbacks showed two types of responses. First, subjects of any sex and age looked at the direction of the speaker and spent more time scanning their environment once an alarm call was made for more than 10 seconds. The researchers believe that they might be scanning for additional cues from the "caller" and the subject's surrounding.

Second, each alarm calls seem to elicit a distinct response from the subjects. Remember the trials were done when the subjects were on the ground and on the trees? When subjects were on ground, leopard calls were more likely to make them run up into the trees and eagle calls made them look up and run into cover (bushes) Snake calls made them look down. When subjects were on the trees, leopard calls were more likely to make them run higher in trees and to look down. Eagle calls made them look up and sometimes run out of trees. Snake calls made them look down.

From the results, Seyfarth et al. (1980) posit that vervet monkey alarm calls alone do elicit different responses. It's hard to tease out whether these alarm calls symbolize the predator "leopard" or a command "run up tree". However, we can postulate that these alarm calls are rudimentary semantic signals used to warn other conspecific of impending danger. For those that are not familiar with semantics, it refers to the meaning of a symbol, sign, word or phrase. In this case, vervet monkey alarm calls are semantic signals because it conveys a specific meaning.

Here's an interesting video by Robert Seyfarth summarizing his research with the vervet monkeys.

Seyfarth, RM. Cheney, DL. Marler, P. 1980. Monkey responses to Three Different Alarm Calls: Evidence of Predator Classification and Semantic CommunicationScience 210(4471): 801-803.

No comments: