Learning Set and Hypothesis Theory

Abstract

The impact of set on problem-solving and summaries of the major theories of learning set are reviewed. One model, hypothesis theory, and relevant studies in the area are presented. The function of memory in the model is examined and a local consistency model is described that conforms with the research data. Two competing views of learning are compared and a test of the strength view versus the all-or-none view is discussed in terms of the blank trials law. In an ambiguous situation, trials with no feedback are not found to be equivalent to trials with positive feedback although equivalent performance is predicted by the blank trials law and all-or-none learning. Complete changes in hypothesis and significantly more errors appear in the no feedback condition as compared to a noncontingent positive feedback condition. This supports the prediction of the strength view that hypotheses of equal strength may compete for dominance in ambiguous circumstances.