Anik de Ribaupierre
Although the definition of working memory (WM) varies among researchers, there is a clear consensus with respect to its function. WM is defined as a system, or as a set of processes, that serves to process and maintain temporary information for use in other cognitive tasks. Thus, in contrast with earlier concepts of primary memory or of short-term memory, WM implies an active mental manipulation in addition to a temporary maintenance of information; that is, it is considered as a process-oriented construct. Moreover, WM is generally considered to be a complex system with a limited capacity, constrained both by limitations in the amount of activation that can be distributed and by limited attentional resources available to activate and maintain task-relevant information while inhibiting task-irrelevant information (e.g., Miyake & Shah, 1999; Richardson, 1996). There is an important controversy with respect to the question of whether WM constitutes a structural entity, i.e., a memory system, such as defined by Schacter and Tulving (1994), with its own, specific processes, or whether it combines a set of processes shared by other psychological functions. Thus, Baddeley (1986, 1992; Baddeley & Hitch, 1974) defines WM as a multicomponent system (in contrast with a unitary view of short-term memory), consisting of three subsystems: two slave systems—the articulatory loop (which is responsible for the maintenance and processing of verbal material) and the visuospatial sketchpad (VSSP, responsible for maintaining and processing visuospatial information)—and the Central Executive (whose main functions are coordination and planning). In contrast, a number of other models suggest that WM designates only an activated subset of long-term memory, and/or call for the same processes as other cognitive tasks.
For instance, Moscovitch (1994) proposes that WM would be more appropriately labeled as "working with memory.'' Ericsson and Kintsch (1995) also speak of long-term working memory.
However, this controversy around the issue of an underlying (neural) system specific to WM is stronger in the field of adult cognitive psychology or neuropsychology than in developmental psychology. Developmental approaches have generally focused on the capacity of WM more than on its architecture (e.g., Ribaupierre & Bailleux, 1994) and have either simply ignored this controversy and merely considered that WM is complex and calls for a combination of storage and processing processes, or have argued that WM indexes an activated subset of information (e.g., Pascual-Leone, 1987; Pascual-Leone & Baillargeon, 1994). In the latter case, WM appears as a compound of different processes that are also at work in other cognitive tasks, most of which can be considered to be attentional in nature (e.g., Ribaupierre, 1999). Although I adhere to this second perspective, I will not further discuss this controversy here especially because, once again, it does not seem central in developmental psychology, which has essentially focused on changes in capacity with age and on possible factors accounting for this change. However, I would like to argue that it is preferable to speak of performances in WM tasks rather than in WM per se, thereby stressing that such tasks call for a number of processes that are also at play in other cognitive tasks.
Was this article helpful?