The varieties of visual persistence: comments on Yeomans and Irwin.

Abstract

In a recent article, Yeomans and Irwin (1985) entered the current controversy about the nature and function of visual persistence, or iconic memory. Specifically, they reported no effect of varying target duration on a standard Sperling (1960) partial-report task, from which they concluded that such a task involves an energyindependent, nonvisible, informational persistence. Moreover, they proposed a two-stage model of visual persistence in order to incorporate these findings with those of other researchers who have reported significant energy effects on some visual persistence tasks. The first stage in their model is that of a brief (less than 300 msec) visible persistence that is inversely related to target energy (luminance or duration); the second stage is that of an informational persistence that can persist from 300 to 600 msec, independent of target intensity, and that may well be nonvisible in character. As Yeomans and Irwin noted, this model shares some characteristics with the alternate models of iconic memory suggested by Coltheart (1980) and DiLollo (1978, 1980). The purpose of the present note is to comment on both the logic of the Yeomans and Irwin procedure and the comprehensiveness of the model they propose with regard to the body ofcurrent empirical results. To anticipate the line of argument to follow, it will be contended that (1) a null effect is a logically weak position on which to base a model-especially given a notoriously complex and noisy experimental task-and (2) there are numerous published results that are inconsistent with the two-stage model favored by Yeomans and Irwin.