If subjects are given control over the hue of an image and asked to adjust it until it appears grey scale there is an interesting effect. If the image is of an object with a salient prototypical color (such as bananas, which are saliently and prototypically yellow), subjects will judge that it is grey scale when in fact the hue is slightly in the direction opposite to that of the standard color. So, for example, an image of a banana will be judged to be grey scale when it is in fact just slightly periwinkle. This is the memory-color effect (Hansen et al. 2006). One possible explanation of this effect is that when the image actually is grey scale, subjects’ top-down expectations about the usual color make it appear (in some way or another) to be slightly tinted in that hue. So when the image of the banana is actually completely grey scale subjects judge it to be slightly yellow. The actual color of the image must be slightly in the direction opposite yellow (periwinkle) in order to cancel this top-down effect and make the image appear grey. This is the memory-color effect.
We expected that after a period of wearing the gear this effect would diminish or even reverse. The reasoning was that if the rotated experience either just disrupted the association of the objects with their prototypical color, or even established a different prototypical color, the effect would be compromised.
Stimuli used in our trials were images of a banana, tomatoes, broccoli, a fire engine, a school bus, a stop sign, and a Starbucks logo. We wanted examples of natural objects as well as artefacts with strong color associations. We also used squares and circles, which have no obvious prototypical color, as controls. The hue of the initial image presentation was random, and subjects were to adjust the hue until the image appeared completely grey scale.
There were technical issues with RG that prevented his data from being usable. But subject JK’s results were in fact what we expected. Pre-rotation, JK’s results were normal. He judged the stimuli to be grey when in fact they were, on average, 3.5% saturated in the direction of the opponent color. JK was quite consistent with this except for one stimulus condition, broccoli, which he actually exhibited the opposite of the expected pattern. He judged it to be grey when it was about 1% saturated in its normal color. The effect was robust, and we have no idea why. We suspect that it was some sort of artefact connected to that stimulus being presented through the rotation gear, but we don’t know for sure.
Post-adaptation JK’s assessments were, on average, about 0.5% saturation, again in the direction of the opponent color. Meaning that he still exhibited the effect, but its magnitude was lessened. This could mean one of two things: i) the rotation protocol disrupted the usual associations of colors and objects, and so all stimuli ended up being treated by his perceptual system just as controls, that is, as objects with no salient associated color; and ii) JK partially adapted to objects’ new prototypical colors. Unfortunately the result we got, in which assessments were on average very close to grey, is consistent with both. The alteration was consistent with both a movement towards grey and a movement towards the canonical color, since both are in the same direction from a spot opposite the canonical color. But one consideration that speaks in favor of (ii) is that the broccoli stimuli, post-adaptation, did not move towards grey, but in fact were judged to be grey when they were even more green than in the pre-rotation trial. That is, the broccoli stimuli judgments moved not in the direction of grey, but in the direction of canonical color, and by about the same amount as the other stimuli moved in that direction: 2.5%.
Our result in the memory-color effect and the two occasions in which JK panicked are consistent. Both effects would seem to result from a re-aligning of the salient prototypical color of objects that have a salient prototypical color. The adaptation of the memory-color effect suggests that the experience of being color-rotated lessened the extent to which top-down effects associated certain objects with their actual prototypical colors, and perhaps even started associating them with new, different prototypical colors. And the lack of the “this is weird” signal when viewing his purple hand also suggests that the old prototypical look of his hand was being supplanted with a new prototypical look. And it also suggests not just that the old prototypical color association was being lessened, but that a new one was emerging. What failed to look weird was his purple hand—not just a hand in any non-flesh color, but in purple. We did not test this, but it seems quite unlikely that had his hand suddenly looked bright red he would have similarly experienced a loss of the “this is weird” signal. This is speculative, but it at least suggests, consistently with the broccoli effect discussed above, that the result we saw with the memory-color effect was not just a loosening of the old associations, but the emergence of new ones.
Moreover, the adaptation of the memory-color effect appears to have been general. The items we used for testing in the memory color effect fell into two groups: first, there were those items, such as Starbucks logos and bananas, which were such that items of that type were observed at least sometimes during the period of rotation; and second there were others, such as fire trucks and baby chicks, which were not observed by the subject under rotation. If what was being altered by rotated experience was just the specific associations of colors with experienced objects, then we should have found different results for these two groups of objects. Bananas and Starbucks logos would be subject to adaptation with respect to the memory-color effect, and fire trucks and baby chicks would not. But this is not what we found. We found the memory-color effect was impacted for all tested objects, even those that had not been seen during rotation.
This suggests that the effect was general, meaning that the adaptation was manifested not as an alteration in some part of the perceptual system concerning its expectations about what bananas or other specific objects look like. Rather, the alteration appears to have concerned expectation about what yellow things generally look like. To put it in dangerously loose and anthropomorphic terms, some part of the system started cranking up the periwinkle knob when objects known to be yellow, like baby chicks, were seen. If this is indeed a memory effect (psychologist do, for some reason, call it the memory-color effect), then it suggests that some part of the system knew that baby chicks were supposed to be yellow, but was beginning to misremember what yellow was like.
Recall where we left off in the last section. There was a suggestion to the effect that for a subject who underwent an inverted spectrum procedure, a cessation of the “this is weird” signal, together with a loss of memory of how things used to look might result in that subject’s inability to report differences between their current inverted experience and their prior non-inverted experience. Such a subject would of course verbally report that their phenomenology had adapted (or, if they were more reflective, they might also admit the possibility that their phenomenology was still inverted, but that their memory was failing to make this fact apparent to them). To indulge in some wild speculation, the general nature of the adaptation to the memory-color experiment suggests that something along these lines might possibly happen if there was a longer adaptation period. We will return to this in the final section.