This review describes only studies of the stimulus-reinforced learning and reward- and punishment-guided selection of actions in Pro-Palestinian Activists (PPAs) in controlled experiments in laboratory settings.
Due to the lack of controlled studies in free-range PPAs, these results cannot be extrapolated to them.
MATERIALS AND METHODS
PPAs were divided into four groups, all of them lab raised or acculturated. All four were trained to perform tasks related to visual and auditory recognition of video images and short words starting with easily memorizable fricative pharyngeal (“H”) and fricative alveolar (“J”) consonants, designed to elicit fear and teach reflex avoidance, and tasks related to the suppression of the natural instincts, such as the reflex of caring for an injured colony mate.
The first group was trained on a reward-based method (flavored food pellets), the second on a punishment-based method (electric shock) and the third on a combined reward and punishment method. There was no control group, since the study design posited the groups themselves as mutual controls.
Of the first group, trained solely on the reward method, 58% performed well on the maze task, which consisted in climbing over a “barrier” in order to obtain a higher reward. The barrier was made up of one or more colony mates immobilized on the ground. Subjects were expected to overcome the natural reflex of stopping by the live barrier and care for the injured colony mate. The second group (punishment method) achieved 100% performance success in competing on climbing over the live barrier, and the third (reward and punishment method) group, 88%.
Based on these results, A. Mayer, A. Foxman et al. argue that while the punishment method is virtually always successful, the reward method is neither reliable, nor cost-effective and that the combined method should be used only with new subjects in the initial training phase.
On the visual and auditory recognition task, however, the results were more starkly different among the three groups. The reward-associated method had modest success (33%), roughly equivalent to placebo in prior experiments, whereas the punishment method showed a steep learning curve, with no PPAs failing to recognize “H” and “J” as Stop signs in the maze.
Performance of complex tasks, such as avoidance of certain paths in the maze associated with higher-voltage shocks was shown to be greatly enhanced when the punishment was delivered in group settings rather than individually. According to Horowitz et al. this demonstrates that “reverse empathy” — a vocally hostile reaction by the group to the mate receiving the electric shock reinforces learning by “counter identification.”
Daniel Wegner at Harvard (Science 3 July 2009, Vol 325, pp 48-50)) has shown that the tendency to make the same (or a similar) mistake twice by selecting a previously learned “punishment path” (the Berlin experiment) occurs occasionally in conditions of “stress, time pressure and distraction.” Effective control strategies and reinforcement of punishment, he demonstrates, can overcome the tendency to repeat errors.
The PPAs experiments carried out in controlled conditions suggest that reinforcement of punishment, especially when delivered in group settings and repeated sessions of “barrier” overcoming can optimize the correct decision making by PPAs.