A by now classical experiment is Faucheux and Moscovici' s (1968) research on the effects of different representations in an experimental game. They told their subjects that they were playing against an impersonal opponent who would react to their own moves either by chance or as nature would. According to their representations of chance and nature, subjects in the chance-condition perceived the moves of the opponent as malevolent and capricious, while those in the nature-condition saw them as more benign, neutral and systematic. The subjects' exploitative behaviour and self-esteem differed systematically.