| Summary: | <p>This thesis reports a series of experiments which study various aspects of perceptual coding processes in normal adults. Perceptual coding involves several distinct stages of analysis, which it is possible to separate by suitable experimental manipulations. The importance of "functional" concepts and models, in discussing the nature and relationship of these stages, is a recurrent theme which is illustrated both by review and by experiment.</p><p>The opening section of the Introduction examines the relationship between "recognition" and "discrimination" and distinguishes the concepts of "norminal" and "functional" as they are applied to human information processing. Following a brief discussion of "active" and "passive" theories of recognition, the nature of information theory is elaborated and evaluated with particular emphasis upon recent criticisms and its ability to model discrimination performance. The theory is found to be inadequate in several important respects - in particular in its failure to provide a viable account of the functional basis of information processing.</p><p>The following section discusses and criticises the concepts of serial and parallel processing which are frequently employed in modelling the way in which pattern discriminations are executed. Although the terminology is useful in certain instances, it is suggested that it presents difficulties which appear to multiply with the increasing conceptual refinement demanded by new data. Several logical and empirical assumptions upon which the dichotomy rests are made explicit and considered. A number of these are found to be questionable, to an extent which makes it difficult to maintain and heuristic value for the original distinction.</p><p>The necessity for functional models is further emphasised in a review of data obtained from some well-tried experimental paradigms. Three are selected on the basis of their relevance to the chapters which follow, as well as for their own intrinsic interest.[see text for continuation of abstract].</p>
|
|---|