Messing with chimpanzee minds

On the subject of ape tool use, Andrew Whiten and colleagues have an interesting experiment in Nature this week (9/29/05).

Here is the experiment in a nutshell:

Our experiment bridges the gap between population-level studies of wild apes and one-to-one social learning experiments by (1) extending the experimental approach to the group level, (2) focusing on ape-to-ape transmission, and (3) using a powerful 'two-action' methodology. In this approach, individuals see a given task completed using one of two possible techniques, allowing the extent to which their own subsequent behaviour matches the demonstration to be systematically measured. We studied three groups of chimpanzees: a control group exposed to a new task with no expert present, and two experimental groups, each supplied with a familiar, conspecific expert trained to solve this new task in a different way. Unlike previous attempts to study traditions using a single experimental group, our three-group design allows us to measure the extent to which two quite different techniques are copied sufficiently well to become traditions, with the control condition identifying baseline levels of individual discovery (Whiten et al. 2005:737, citations omitted).

Using this procedure, the experimenters introduced a device that would vend food to the chimpanzees. The device could be worked in either of two ways: by using a stick to lift a hook, or by using the same stick to poke a flap. The workings of the device inside are not visible from the outside, although both lifting and poking are always available to the chimpanzee using the device.

The question is, when chimpanzees learn extractive foraging techniques, how much of the learning is direct imitation of the techniques they see others doing, and how much is emulative learning by individual experimentation?

There are basically two options here: either the chimp uses the device the way he saw another chimp doing it, or he experiments with the device himself and figures out the other method. The first option is imitative learning: the chimpanzee copies not only the goal, but also the actions leading to the goal. The second option is more emulative: the chimpanzee copies the goal, but figures out its own way to attain the goal.

The experiment found that the chimpanzees predominantly used the method they saw another group member using:

In the Poke group, all tool users adopted predominantly the Poke technique. In the Lift group, the first six chimpanzees to succeed adopted the Lift method predominantly. However, chimpanzee JL then discovered both the Poke and Lift techniques, and continued to use both of them (Fig. 2b). Two other chimpanzees in this group then acquired both methods, while two adopted only the Lift method and four only the Poke method (Whiten et al. 2005:738).

The experiment also found that chimpanzees are conformists: even the chimpanzees who learned both techniques tended to use the technique that was most common in their group.


Whiten A, Horner V, de Waal FBM. 2005. Conformity to cultural norms of tool use in chimpanzees. Nature 437:737-740. Full text (subscription)