Memory: task dissociations, process dissociations and dissociations of consciousness
Alan Richardson-Klavehn, John M. Gardiner, and Rosalind I. Java
TASKS, RETRIEVAL STRATEGIES, AND STATES OF CONSCIOUSNESS: A FRAMEWORK
Striking dissociations between explicit and implicit tests of memory have led to an explosion of research addressing the relationship of conscious awareness of the past to observed memory performance (for reviews see Gardiner and Java 1993a,b; Richardson-Klavehn and Bjork 1988a; Roediger and McDermott 1993; Schacter et al. 1993). Despite the vast quantity of data, there is no firm agreement on the theoretical implications of such task dissociations. Two levels of theoretical debate can be identified, the first 'shallower' than the second. The debate at the first level concerns the extent to which task dissociations represent evidence for differences in retrieval strategies and/or memorial states of awareness (e.g. Bowers and Schacter 1990; Dunn and Kirsner 1988, 1989; Challis and Brodbeck 1992; Jacoby 1991; Jacoby et al. 1993b; Merikle and Reingold 1991; Reingold and Merikle 1990; Richardson-Klavehn and Bjork 1988a; Schacter et al. 1989). The debate at the second, 'deeper', level concerns the encoding and retrieval mechanisms underlying the hypothesized strategies and states of awareness. Some theorists, for example, have argued that different manifestations of memory are attributable to the operation of distinct memory systems (e.g. Schacter 1989; Squire 1992; Tulving 1993; Tulving and Schacter 1990). Others argue that these different manifestations are consistent with a process viewpoint (e.g. Jacoby et al. 1989a; Kolers and Roediger 1984; Roediger 1990; Roediger et al. 1989a,b).
We believe that progress on the former issue is likely to be a precondition for progress on the latter one. The present chapter, therefore, reviews and integrates recent data from our laboratory that contribute to understanding of the relationship between task performance, retrieval strategies, and states of awareness. Our data have led us to adopt a framework that makes a threefold distinction between: (a) method of