Theoretical explorations of the Bruner-Potter 1964 interference effect

Abstract

Bruner and Potter (1964, Science, 144, 424–425) showed that identification of objects which were gradually brought into focus became progressively worse as initial levels were made more blurred. They attributed this interference effect to subjects' development of erroneous hypotheses about the object which interfered with correct perception. In a series of five experiments, we tested their claim that erroneous perceptual hypotheses are responsible for the interference effect against an alternative explanation of the effect. This alternative explanation was generated from a connectionist model of item perception whose success in simulating the effect was produced by transient activation in perceptual structures. Experiment 1 replicated the Bruner-Potter effect in picture fragment completion by showing that performance was worse with an ascending method of limits than with a fixed level of presentation. Experiments 2 and 3 tested the erroneous hypothesis explanation of Bruner and Potter by requiring subjects to generate hypotheses at each level in the ascending condition. Subjects given corrective feedback in Experiment 3 did not show a reduction in interference from the ascending condition compared to subjects given no such feedback in Experiment 2. Experiments 4 and 5 tested the model's transient activation explanation by having subjects solve math problems during the intervals between presentations in a spaced ascending condition. In both experiments, this ascending spaced condition was immune to interference, showing performance equal to the fixed level condition.