Multiple memory systems: what and why.

Abstract Evidence from a variety of domains converges on the view that there are multiple learning/memory systems, but there is no clear understanding of what these systems are, and why they should exist. I review an hypothesis about multiple memory systems postulated by O'Keefe and Nadel (The hippocampus ac a Cognitive map, Oxford: Clarendon Press, 1978), in particular our assertions about the nature of two kinds of systems (the locale and taxon systems), and what differentiates them from each other. I concentrate on our assumption that systems differ in terms of the type of information they process, rather than in terms of the temporal duration over which they function. Our "content-driven" dichotomy is contrasted with several other current views, none of which appears to explain extant data, or provides clear predictions about when one or the other kind of system would be engaged. Finally, 1 briefly consider why such distinct systems might have evolved.