ABSTRACT
The effect of contradicting an expectation was investigated in a quite different situation by Bruner and Postman in 1949. Expectations may depend on recent or distant experience, including instructions, graphical and semantic context. They may serve to reduce the number of candidates for synthesis or the order in which candidates are considered. A disadvantage of this style of research is that it lifts the perceptual system out of its ecological niche, where it is served by all manner of long-range contextual aids. The central role of expectancies in perceiving and attending has been re-emphasized by J. Hochberg. In the context of a constructive theory like R. L. Gregory's the task is to establish the evidence on which the perceptual system bases its hypotheses. There are theoretical precedents for supposing that perceptual processing involves a preliminary stage at which global characteristics of the input are extracted. The kind of information available at this stage is probably of a low order.
