How many memory systems? Evidence from aging.

older (ages 63-80) adults were given procedural, semantic, and episodic memory tasks. Repetition, lag, and codability were manipulated in a picture-naming task, followed by incidental memory tests. Relative to young adults, older adults exhibited lower levels of recall and recognition, but these episodic measures increased similarly as a function of lag and repetition in both age groups. No age-related deficits emerged in either semantic memory (vocabulary, latency slopes, naming errors, and tip-of-the-tongue responses) or procedural memory (repetition priming magnitude and rate of decline). In addition to the age by memory task dissociations, the manipulation of codability produced slower naming latencies and more naming errors (semantic memory), yet promoted better recall and recognition (episodic memory). Finally, a factor analysis of 11 memory measures revealed three distinct factors, providing additional support for a tripartite memory model. The distinction between semantic and episodic memory proposed by Tulving in 1972 has generated a plethora of research, giving the dichotomy unassailable heuristic value. The semantic-episodic distinction has also generated consid-erable debate over its validity. (For summaries, see McKoon, Ratcliff, Dell, 1986, and Tulving, 1984, with accompanying commentaries.) One source of data conspicuously neglected in these discussions is the ubiquitous finding of memory decrements in relation to normal aging. Thus, in the present research, age differences in memory serve as one vehicle for investigating the question of how many memory systems there are. Tulvings (1985) more recent classification calls for three memory systems. If measures tapping different types of mem-ory reveal similar patterns of loss in elderly adults, then the theory of multiple memory systems would be neither sup-ported nor disproved. On the other hand, if three classes of This article is based on a doctoral dissertation conducted at the University of Minnesota. Portions of this research were reported at the meeting of the Psychonomic Society, Minneapolis, November