RESULTADO S Y DISCUS IO N
3.1.2. A NAL ISIS INFERENC IAL
experiments with the behaviour of a chimpanzee in a related setting, hi Kohler's experiment a hungry chimpanzee was confined in a cage. A banana was outside the cage just beyond the chimpanzee's reach, and there was a stick lying on the floor of the cage. The chimp made a few attempts to reach the banana through the bars of the cage but was unsuccessflil. In one action, the chimp then picked up the stick and used it to retrieve the banana. This swift solution was in contrast to the trial and error behaviour shown by the animals in Thorndike's experiment. Kohler (1927) asserted that the chimp was able to behave intelligently m this situation because the solution was part of the overall situation (one of the defining features of Gestalt psychology) and so the chimp was able to perceive the solution to the distance problem. Kohler argued that Thorndike's test setting did not permit this sort of intelligent behaviour by the animal as the release mechanism in the cage was hidden and therefore not part of the perceived situation, and in any case its functioning was too complicated for an animal to comprehend.
What is notable though is that Kohler had no objection to Thorndike's assertion that experimental investigation of animal behaviour is necessary. Rather, this debate seems to prove the point Thorndike was making - by controlling the setting for the behaviour it is possible to identify the causes of the behaviour. Different experimental settings between Thorndike's and Kohler's experiments produced different behaviours and further manipulation of these variables was one way of establishing the limits of non-human cognitive abilities.
Stimulus Equivalence
There have been many attempts to demonstrate "human" cognitive abilities with non-human subjects. Numerous attempts have been made to demonstrate stimulus equivalence with non-humans, and to clarify under what, if any, conditions it is possible to demonstrate equivalence with these subjects. Sidman and Tailby (1982) defined the necessaiy properties of an equivalence relation; reflexivity, symmetry and transitivity. Sidman et al (1982) attempted to demonstrate one of these
properties, symmetry, with both rhesus monkey and baboon subjects. The evidence ft'om Sidman et al (1982) suggested that even if non-humans can demonstrate equivalence relations, they do not demonstrate them with the same ease as human subjects. Certainly the procedures used by Sidman et al (1982) which produced equivalence with normally able children gave no evidence of establishing stimulus equivalence with the non-human subjects (see chapter 2 for fuller discussion).
Chapter 3: Cognition in Non-Humans 34
Since Sidman et al (1982) there have been several attempts to demonstrate stimulus equivalence as defined by Sidman and Tailby (1982) with non-humans. Generally these studies have had little success, both with pigeon and primate subjects. D'Amato. Salmon. Loukas and Tomie G 985)
This experiment by D'Amato et al followed on from the work by Sidman et al (1982) testing for symmetry and transitivity with non-human subjects. Sidman et al (1982) had found no evidence for symmetry or tiansitivity with non-human primate subjects. D'Amato et al (1985) tried to verify Sidman et al's results with monkey and pigeon subjects. D'Amato et al found no evidence for backward associations (symmetiy) following conditional discrimination training with monkeys, "If monkeys form backward associations . .. those associations tend to be weak and easily swamped by competing sources of control" (p.41). However, following training of further conditional relations, the same monkey subjects showed convincing evidence of associative transitivity. D'Amato et al concluded that "associative transitivity appears from the present results to be a robust phenomenon in monkeys" (p.44)D'Amato et al then replicated this transitivity experiment as closely as possible with pigeon subjects. However there was no evidence of associative transitivity with the pigeons.
D'Amato et al therefore found evidence for one of the properties of stimulus equivalence, transitivity, with one set of non-human subjects, monkeys. However there was no evidence for symmetry with the same group of subjects and no evidence for transitivity with a different set of non-human subjects, pigeons. Dugdale and Lowe (19901
Dugdale and Lowe (1990) also reported an experiment evaluating symmetry with non-human subjects. Dugdale and Lowe's study is made particularly interesting by the subjects they used. Dugdale and Lowe had hypothesised that linguistic ability, and in particular naming, are necessary in order for a subject to demonstrate emergent equivalence relations, and this might explain why it had proved so difficult to demonstrate equivalence with non-human subjects. Dugdale and Lowe investigated the possibility of demonstrating stimulus equivalence with non-human subjects who had received language training. These subjects were three
chimpanzees who had all participated in an ape-language training programme and who could communicate by touching lexigi'ams on a key board, each of which was
Chapter 3: Cognition in Non-Humans 35
associated with an object, action or location. (Rumbaugh 1977, Savage-Rumbaugh 1984).
Dugdale and Lowe attempted to teach the chimps an AB matching relation and then to test for the emergence of the BA symmetry relation. One subject failed to learn the AB matching task. The other two subjects learned the AB task but showed no evidence of BA symmetry, despite several procedural modifications designed to maximise the chimpanzees chances of success. These modifications included adding identity matching trials to the AB baseline so that the subjects had experience with all the stimuli as both samples and comparisons, and also
providing reinforcement for correct responses on both baseline and test trials. Despite these modifications, neither subject showed any evidence of emergent BA symmetrical relations.
Dugdale and Lowe argued that, given the failure to display one of the properties of equivalence with these subjects, it was likely to be very difficult to display stimulus equivalence with other non-human subjects. Chimpanzees seemed more likely candidates to demonstrate equivalence than more distant non-human relatives such as pigeons and these particular chimpanzee subjects had an unprecedented history of complex training and testing. However, it must be noted that Dugdale and Lowe's procedure failed to establish even AB matching with one subject, and hi their review of stimulus equivalence with non-verbal organisms Dube, Mcllvane, Callahan and Stoddard (1993) suggested that limitations in experimental
procedui es may have contributed to failures to demonstrate aspects of equivalence relations, at least with rats and pigeons.
Both D'Amato et al (1985) and Dugdale and Lowe (1990) showed failures to demonstrate some of the crucial components of equivalence with non-human subjects and other studies have found similar results (e.g. Kendall 1983, Lipkens, Kop and Mathijs 1988, Sidman et al 1982)
Two studies did claim to have demonstrated stimulus equivalence with non-human subjects: Mclntire, Cleary and Thompson (1987) and Vaughan (1988). However, several researchers have argued that it is inappropriate to regard the performances of the non-human subjects on these tasks as demonstrating stimulus equivalence.