Modularity, Working Memory, and Reading Disability
Ignatius G. Mattingly Haskins Laboratories, and University of Connecticut
One of the many fine things about the research of Isabelle Liberman and her colleagues is that their experimental comparisons of good and poor readers have not been restricted to reading performance. They have also investigated their subjects' speech perception, verbal short-term memory, and linguistic awareness, not only demonstrating clear differences between the two reading groups but also providing evidence that poor readers' problems may be specifically phonological. Brady's current research extends this approach, comparing good and poor readers' performance not only in perceptual and memorial tasks but also in various speech production tasks ( Rapala & Brady, 1990).
In her contribution to this volume, Brady argues that poor readers perform badly in short-term recall tasks not because they fail to employ sophisticated mnemonic strategies, or because they are using a visual or a semantic rather than a phonological coding, but because the phonological processes they employ in "working memory" operate inefficiently. She then considers a more general question: Are these working-memory phonological processes to be equated with those required for linguistic tasks carried out by other cognitive functions? If so, then it is possible that the problems of poor readers are due to a generalized phonological deficiency. Perhaps, she suggests, "the difficulty observed in encoding phonological information is not restricted to memory tasks, but occurs at a more abstract level, whenever it is necessary to create and maintain a phonological representation."