Can memory color effects be explained by cognitive penetration?

ABSTRACTOrange heart shapes are commonly perceived as slightly reddish, which is an example of the memory color effect (MCE). Given that the MCE is a modulation of visual memories of typical colors of familiar objects, it can be considered to be a top-down effect. Whether cognitive penetration can explain MCEs has been actively debated since Macpherson argued that the belief that hearts are red alters orange perception. This paper aims to provide a credible explanation of the MCE that is consistent with scientific studies. I first prove that Macpherson’s theory is not properly supported by the psychological evidence that she provides. I also show that her model contradicts a report on the neural correlate of the MCE. Next, I examine color constancy, which enables the perception of colors as constant under various illuminations, as a possible noncognitive principle for the MCE. The relationship between them, however, is not well established. Using statistical analyses of scenes around us, Purves suggests that perception is determined by the most frequent percepts. I argue that Purves’ view better explains the noncognitive aspects of visual memory and is consistent with the finding of the neural correlate of the MCE.KEYWORDS: Memory color effectscognitive penetrationnoncognitive penetrationvisual memorycolor constancywholly empirical theory AcknowledgementsI thank anonymous reviewers for their valuable comments and suggestions.Disclosure statementNo potential conflict of interest was reported by the author.Notes1. MCEs are also known to be triggered by artifacts or brand logos in specific colors, such as Nivea, that are familiar to consumers (Kimura et al., 2013; Witzel et al., 2011).2. Here, CP means a causal influence of propositional attitudes on perception. Although Pylyshyn (1999) defines CP in terms of semantic or logical influence, I follow Macpherson (2012, 2015), Siegel (2012), and Stokes (2012). In general, frontal activations, which indicate top-down effects, are favored as evidence for CP, although there is a noteworthy exception (Firestone & Scholl, 2016).3. Siegel (2015) has a similar view.4. Macpherson (2012) expects that the reddish-orange heart case successfully defuses the typical modularist response that interprets alleged cases of CP as a noncognitive shift in attention. Color perception differs from our experience of ambiguous figures, such as the Necker cube, since color has no spatial location for attention. She further denies the possibility of penetration into judgments. If the reddish-orange heart is an example of the misjudgment of accurate perception, it is difficult to elucidate why perceivers made such a brute error. Considering that perceptual experience and judgment usually go hand in hand, it is challenging to explain why perception was not penetrated, while judgment about it was. Gatzia (2019) also indicates that since subjects spent enough time and paid enough attention, it is implausible that they made incorrect judgments about accurate perceptions. Although it may still be arguable whether perceptual judgment was penetrated, I assume that Macpherson’s claim about a strong link between perception and judgment is legitimate primarily for the sake of argumentation, unless strong empirical evidence against it is provided. A more significant objection to her is to consider Delk and Fillenbaum (1965) to be a case of noncognitive penetration.5. Meanwhile, Vandenbroucke et al. (2016) reported activation in frontal regions beyond early visual cortices. This may suggest that object knowledge, which is propositional, contributes to MCEs. Their experiment is quite complex, including sixteen stimuli in multiple dimensions (red or green color, shapes of typical objects, and shapes of nonobjects). Bannert and Bartels (2018), in a later study, consider such results to be voluntary or explicit effects, involving “conscious and attentive object imagery,” which are different from “automatic (and likely unconscious) association of memory colors to objects” (p. 3665). They seem to suppose that experiments using grayscale photographs of typical objects show implicit effects.6. Macpherson (2012) seems to admit this problem (p. 54).7. Macpherson (2012) also mentions a case in which the sound of an alarming clock is integrated into a part of a dream (pp. 53–54). This suffers from the identical problem.8. The two steps of Macpherson’s model may not be so closely related, as they employ different neural pathways, each of which takes a different time course (Gross et al., 2014).9. Among various suggestions for a neural correlate of color constancy mentioned earlier, adaptation at the retinal level or V1 is too early for any type of top-down effect.10. Gatzia (2019) reviews the history of the discussion of MCEs in relation to color constancy (Section 4).11. However, the link between perceptual learning and color constancy seems to be underdeveloped. The mechanism for color constancy may not match that for perceptual learning (Gatzia, 2019).12. For example, an orange foreground appears yellowish when the background is red, whereas the same orange appears reddish when the background is yellow.13. The title of their article is “Memory modulates color appearance”.14. When subjects are forced to memorize thousands of object images, their memory performance is expected to decline. It is the conceptual distinctiveness of objects, rather than perceptual distinctiveness, that interferes with memory performance. From this, Konkle et al. (2010) suggest that conceptual structure is more crucial to visual long-term memory.15. While subjects focused on mental math calculations as a primary task, secondary words were provided. When the problem set was difficult, secondary words were less attended to. However, even unattended words were remembered above by chance.16. This is analogous to the discovery of iconic memory which lasts no more than a second (Sperling, 1960).