On the Form of Forgetting

TL;DR: For example, this article showed that the course of forgetting is a curvilinear function of time relative to five other reasonable alternatives (linear, exponential, exponential-power, hyperbolic, and logarithmic).

Abstract: Almost everyone would agree that the course of forgetting is some curvilinear function of time. The purpose of the research described herein was to identify the nature of that function. Three experiments are reported, two involving human subjects and one involving pigeons. The human experiments investigated this issue using recall of words and recognition of faces, whereas the pigeon experiment employed the standard delayed matching-to-sample task. In all cases, the course of forgetting was best described by a simple power function of time relative to five other reasonable alternatives (linear, exponential, exponential-power, hyperbolic, and logarithmic). Furthermore, a reanalysis of Ebbinghaus's (1885) classic savings function showed that it, too, declines as a power function of time. These findings suggest that the form of forgetting is a relatively robust property of memory performance and that its mathematical description, perhaps only coincidentally, matches that of the psychophysical function.