Markovits H, Savary F
Département de Psychologie, Université du Québec à Montréal, Canada.
Q J Exp Psychol A. 1992 Jul;45(1):133-48. doi: 10.1080/14640749208401319.
Cheng and Holyoak (1985) have proposed that people possess classes of linguistically based schemas that have an internal structure that is determined by pragmatic considerations. They found that when permission schemas ("If you want to do P, then you must do Q") are used in the selection task, the success rate is much superior to what is usually observed. According to Cheng and Holyoak, this is due to the fact that the permission schema is defined by a set of production rules that give the same answers to problems of conditional inference as those of formal logic. In order to test this hypothesis specifically, 160 university students were given one of two tests. The first contained two sets of inferential reasoning tasks, one using a permission schema, the second using a relation of multiple causality. The second test employed the same two conditional relations, but in an appropriate context. The results indicated that subjects did better on the reasoning task with the schema of multiple causality when presented in context, but, as predicted, their performance was much worse on the inferential reasoning task with the permission schema, which generated a higher proportion of logically incorrect responses. These results suggest that contrary to what has been affirmed, permission schemas might not have a logical structure that is equivalent to conditional logic. A second experiment examined selection task performance using the same two relations in context. Performance on the permission schema was superior to that found with the relation of multiple causality. This confirmed previous results indicating that permission schemas do improve selection task performance, but also suggests that this effect is not related to understanding of conditional reasoning.
程和霍利约克（1985）提出，人们拥有基于语言的图式类别，这些图式具有由语用因素决定的内部结构。他们发现，当在选择任务中使用许可图式（“如果你想做P，那么你必须做Q”）时，成功率远高于通常观察到的情况。根据程和霍利约克的说法，这是因为许可图式是由一组产生式规则定义的，这些规则对条件推理问题给出的答案与形式逻辑的答案相同。为了具体检验这一假设，160名大学生接受了两项测试中的一项。第一项测试包含两组推理任务，一组使用许可图式，另一组使用多重因果关系。第二项测试采用了相同的两种条件关系，但放在了适当的情境中。结果表明，当在情境中呈现时，受试者在多重因果关系图式的推理任务上表现更好，但正如预测的那样，他们在许可图式的推理任务上表现更差，产生了更高比例的逻辑错误回答。这些结果表明，与所主张的观点相反，许可图式可能不具有等同于条件逻辑的逻辑结构。第二项实验在情境中使用相同的两种关系检验了选择任务的表现。许可图式的表现优于多重因果关系的表现。这证实了先前的结果，即许可图式确实能提高选择任务的表现，但也表明这种效果与对条件推理的理解无关。
