Do Chimpazees Learn By Imitation Or Emulation?

How best to explain the ability of chimps to learn simple technologies from each other? It is clear that such social learning occurs and is very powerful but do the same processes hold for both humans and chimpanzees? A number of important experiments over recent years suggest that humans actually over-imitate, even at the risk of copying irrelevant detail, and that this is extremely important the more complex, and therefore opaque, the task being demonstrated is. Here Claudio Tennie, Josep Call and Mike Tomasello, of the Max Planck Institute in Leipzig, present results of a so-called "floating peanut" task which showed that chimps learned with the same amount of success whether they were shown the correct solution to the task in a literal sense, or another method which achieved the same result. Emulation, they conclude, may be enough to explain the social transfer of skills. Here is the abstract. PLoS is free, so anyone interested can download the entire paper:

Background

It is still unclear which observational learning mechanisms underlie the transmission of difficult problem-solving skills in chimpanzees. In particular, two different mechanisms have been proposed: imitation and emulation. Previous studies have largely failed to control for social factors when these mechanisms were targeted.
Methods

In an attempt to resolve the existing discrepancies, we adopted the ‘floating peanut task’, in which subjects need to spit water into a tube until it is sufficiently full for floating peanuts to be grasped. In a previous study only a few chimpanzees were able to invent the necessary solution (and they either did so in their first trials or never). Here we compared success levels in baseline tests with two experimental conditions that followed: 1) A full model condition to test whether social demonstrations would be effective, and 2) A social emulation control condition, in which a human experimenter poured water from a bottle into the tube, to test whether results information alone (present in both experimental conditions) would also induce successes. Crucially, we controlled for social factors in both experimental conditions. Both types of demonstrations significantly increased successful spitting, with no differences between demonstration types. We also found that younger subjects were more likely to succeed than older ones. Our analysis showed that mere order effects could not explain our results.
Conclusion

The full demonstration condition (which potentially offers additional information to observers, in the form of actions), induced no more successes than the emulation condition. Hence, emulation learning could explain the success in both conditions. This finding has broad implications for the interpretation of chimpanzee traditions, for which emulation learning may perhaps suffice.