Visual persistence and code selection in short-term memory for letters.

Four experiments were conducted to examine temporal changes in visual and name codes and the relationship between them. Unlike earlier studies, the design included (a) several levels of visual similarity in addition to the conventional same case and different case conditions and (b) physical- as well as name-match tasks to provide direct evidence about subjects' retention of form information. The main results were as follows: First, the presence of multiple typefonts in the stimulus set does not eliminate convergence between the reaction time functions for the physical- and name-match conditions in a name-match task. Second, inclusion of a physical-match task, requiring subjects to retain form information, does not necessarily eliminate convergence in a primary, name-match task. When the secondary task does reduce convergence, however, the results suggest that it does this through its effect on the name-match function rather than on the physical-match function. Third, apart from difficulties with within-cluster typefont combinations, particularly at long interstimulus intervals, subjects can generally retain and use reliable information about letter form over the ISI range (.5--4.0 sec) used in the experiments. The results were thought to be consistent with the conclusion that convergence results from a combination of (a) a nonoptional decrease in the accessibility of the visual code as a function of increasing ISI and (b) an optional increase in the accessibility of the name code. Under ISI conditions at least, convergence cannot be attributed to either generation of a visual code of the alternative form of the target letter or to the absence of relevant form information.