Developmental fractionation of working memory

Abstract

Introduction In recent years, neuropsychological evidence has proved to be of considerable value in advancing our understanding of cognitive processes and their organization. A particularly powerful methodology is that of “neuropsychological fractionation” (see, e.g., Shallice, 1979a), which attempts to interpret highly selective neuropathological deficits in cognitive abilities in terms of models of the intact cognitive system in which one assumes damage to specific components. In this chapter I wish to discuss a somewhat analogous method of “developmental fractionation,” which involves studying the cognitive abilities of normal children and which I believe can usefully complement neuropsychological evidence in constraining models of adult function. Since developmental fractionation is not, to my knowledge, particularly widely used, I shall begin by describing what it involves and the general rationale behind it. I shall then go on and describe some research on the development of “working memory” (Baddeley & Hitch, 1974; Baddeley, 1986), which I think illustrates the potential of the method. This work bears directly on the fractionation of visual and phonological components of working memory and on the infrastructure of these components. I conclude by considering some of the mutual implications of developmental and neuropsychological evidence about working memory, and the ways in which they can usefully complement one another. What is developmental fractionation? Although currently rather little used, developmental fractionation has, by psychological standards at least, a long history. Baldwin (1894) took the view that the complex cognitive abilities of adults could be regarded as being “the union of simpler elements” that are assembled during childhood development. He suggested that if one wanted to identify these elements and study them, childhood is a very good place to start looking.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 54 citations exist upstream.