Memory for Color

The experiments concerning the effects of color on memory are a continuation of former studies on the influence of sensory information on memory (cf. Zimmer, 1993). In these experiments we had observed that the color of line drawings was not automatically remembered, and even if color was studied intentionally, memory was poor. Additionally, an incongruent color did not impair episodic object recognition if color was changed from study to test. In contrast, size and orientation were remembered even if they were not intentionally encoded, and changing these features impaired episodic object recognition even though the features were declared irrelevant. In implicit memory tasks a repetition effect occurred and this effect was uninfluenced by changes of color.

In recent experiments we investigated these effects further. We followed three lines of research:

1. We tested the influence of color on object recognition in order to check whether color is actually used during identification.
2. We investigated whether the color effects reported in some implicit tests (Wippich & Mecklenbräuker, 1998) are still present if explicit memory is absent.
3. We tested color effects in explicit and implicit memory tasks under enhanced color influences.

This research was supported by a DFG-grant under Zi 308-2 and partially under Zi 308-4 in context of the Research Group on Binding (FOR 448).

Color in object recognition: In word-picture and picture-word matching paradigms we were able to observe an influence of inadequate colors on decision times. Exemplars from categories with prototypical colors were matched more slowly if they were colored inadequately (Zimmer, 1998). Hence, color is used during object identification, so in principle it can cause implicit memory effects.

Episodic recognition of color: We explored which factors influence explicit memory for the colors of objects (more precisely, we used pictures of objects). We presented line drawings, silhouettes and multiple-colored pictures, and we tested recognition of these colors. We observed that color memory increased over these three types of pictures, whereas the importance of an explicit intention to remember color during study decreased (Zimmer & Steiner, 2004). We explained these effects by the assumption (a) that color has to be actively bound to the object because it is a separable dimension in perception, and (b) that this binding requires attention which is provided either voluntarily due to an explicit instruction or involuntarily due to the salience of color.

Implicit memory for color in object recognition: After we had observed that the probability of color encoding is maximized with multiple-colored objects, we used these objects in an implicit memory test. Colors were either identically repeated or changed from study to test. Explicit memory was influenced by color changes although color was neither intentionally encoded nor relevant for episodic recognition. In implicit memory, we observed a clear repetition effect for repeated pictures irrespective of their color. Only in a very difficult perceptual condition - using colored line drawings embedded in line segments of different colors with brief exposure durations - the additional match of objects' color enhanced identification performances (Zimmer & Steiner, 2003). We take this as further support for our hypothesis that explicit memory relies on different memory entries (episodic tokens) than implicit memory. Implicit effects should be caused by temporary changes of perceptual types representing the invariant features of a typical category exemplar.

Color in an implicit color test: In a color preference task, Wippich & Mecklenbräuker (1998) instructed their participants to assign arbitrary colors to black-and-white figures (or their names) at their own discretion. The color could be selected from a small set of colors. Prior to this decision, some of the objects had already been presented in one of the colors. The authors reported that people more often chose an old color for a picture than one of the 'new' alternatives. We used the same procedure, but we brought explicit and implicit memory into opposition, so we could test whether this effect also occurred if participants had no explicit memory for the color. In the absence of explicit memory we observed no implicit color preference (Zimmer, Steiner, & Ecker, 2002). Implicit color preferences could be observed only if an explicit access to the item's color was possible, as well. Obviously, these indirect memory effects in the implicit task are based on the same memory traces that are used in the explicit memory task.

In order to gain more insight into the role of color for object identification, and in order to detect implicit memory effects in this process, we are currently running additional experiments using electrophysiological methods. In one experiment, for instance, participants are to identify colored objects in arbitrary, typical or inadequate colors, and we test whether respective processing differences are reflected in event-related potentials. Additionally, we vary the level at which the objects have to be classified (basic level or specific level, e.g., bug versus lady bug).