Deceptive tactics

At the behavioural level, deception constitutes misinterpretation of situations by one individual as a consequence of the acts of the other one (Krebs and Dawkins 1984; Byrne and Whiten 1988). Deceptive tactics in animals and humans are tightly related with knowledge about what “game partners” see and do not see, that is, with the capacities for perspective taking. Besides, subjects have to understand that " seeing is knowing", in other words, that the visual information that their partner had acquired gave him knowledge of what he had seen, for instance, where the reward was hidden. Such understanding is related to another type of behaviour, knowledge attribution that will be considered later in this section.

Byrne and Whiten (1991) have compiled a catalogue of naturalistic evidence of intentional deception in primates which by 1990 involved over 250 individual observations of such behaviour. Examples include cases when female primates hide from dominant males and suppress their usual copulation calls whilst mating with a subordinate males; diverting attention of dominant individuals by giving false alarm calls; distracting the dominant male with threats of aggression while the other mates with one of his females-and then reversing their roles.

The first experimental attempt to test chimpanzees' ability to deceive was performed by Woodruff and Premack (1979). They let four chimpanzees become acquainted with two trainers, one of whom was kind and helpful to them (“cooperative” trainer), while the other, “competitive” trainer treated the chimpanzees aggressively. During experiment the chimpanzees one by one were placed in such a position that they were able to observe a third person, a laboratory assistant who hided food under one of two containers that animals could not reach. The assistants then left the room, and one of the two trainers came in. Since the chimpanzee was in a cage, it could not reach the cups, but it was able to show the trainer where the food was. If it was the cooperative trainer and the chimpanzee pointed to the right cup, the trainer picked up the reward and gave it to the chimpanzee. If it pointed to the wrong cup, the trainer looked under it and then went out. If it was the competitive trainer and the chimpanzee pointed to the right cup, the trainer took the food and ate it himself. But if the chimpanzee pointed to the wrong cup, the nasty trainer looked under it and then went to the corner to sit and sulk. When this happened, the chimpanzee was let out of the cage and could retrieve the food itself from under the right cup. In this situation the chimpanzee was rewarded if it pointed to the wrong cup. All the chimpanzees needed at least fifty attempts before they could systematically distinguish which behaviour was most appropriate for which trainer. Since it takes such a long time, the question is whether they understand anything about how the trainers think when the chimpanzees point. During their training subjects had ample time to learn that certain responses in the presence of the cooperative trainer always resulted in reward, whereas different responses resulted in reward in the presence of competitive trainer. A discrimination of this sort can more readily be explained as a form of conditioning. However, sets of results obtained in other experiments concerning knowledge attribution in non humans (see details below), indirectly support the hypothesis that the chimpanzees can assimilate the intentional deceptive tactics.

Recently experimental evidence of deception have come from studies on corvids, the birds possessing complex social life and wide traditions of pilfering.

Emery, Clayton and Dally (Emery and Clayton, 2001; Emery et al., 2004) looked at how scrub jays reacted to the possibility that some neighbor would steal their food cache. Some of the jays in the laboratory had criminal histories of snagging food that another bird had buried. The researchers let the “criminal” jays as well as those with clean records hide treats. When the birds stashed their seeds in private, they didn't take an opportunity to move their treasure to a new hiding place. However, if the researchers let another bird get near enough to watch the caching, the criminal jays took the next opportunity to recover the treat and hide it in a different place. The birds with no experience of thievery didn't re-cache.

Bugnyar and Kotrschal (2002, 2004) reported a deceptive behaviour of a common raven (Corvus corax) leading a competitor away from food in a social foraging task. In the experiment, four individuals had to search and compete for hidden food at color-marked clusters of artificial food caches. A subordinate male, Hugin, found and exploited most of the food. As a result, the dominant male, Munin, displaced him from the already opened boxes. The subordinate male then developed a pattern, when the loss of reward to the dominant got high, he would move to unrewarded clusters and open boxes there. This diversion often led the dominant to approach the unrewarded clusters and the subordinate then had a head start for exploiting the rewarded boxes. Subsequently, however, the dominant male learned not to follow the subordinate to unrewarded clusters and eventually started searching for the reward himself. These interactions between the two males illustrate the ravens' potential for deceptively manipulating conspecifics.

Hugin's behaviour enabled experimenters to compare ravens’ misleading behaviour with similar behaviour in other species. First of them was described by Goodall (1986) for a wild chimpanzee that was socially restrained from feeding at an artificial food dispenser. The subject repeatedly induced leaving of the whole group by walking off the site but then circled back and got more food. In addition, Hugin's behavior resembles the results of the experiments with chimpanzees (Menzel 1973 a, b; Hirata and Matsuzawa 2001) and mangabeys (Coussi-Korbel 1994) in which a particular subject was informed on the location of hidden food. The informed subjects repeatedly lost the food to a dominant conspecific and, as a consequence, began misleading the competitor. Ring-tailed lemurs tested with a similar procedure failed to use deceptive tactics consistently although the informed subject apparently did benefit from accidental manoeuvres that distracted the competitor (Deaner et al., 2000). Subordinate domestic pigs, in contrast, showed misleading-like behaviors in response to exploitation by dominants even though these moves hardly affected their foraging success (Held et al. 2002). The authors suggest a parsimonious explanation that misleading of conspecifics is based on the manipulation of the others' behaviour rather than on a consideration of others' mental state.