According to proponents of enactive perception, perceptual experience amounts to relevant behavioral skills (O'Regan & Noë 2001; Noë 2004, this collection).
To be a perceiver is to understand, implicitly, the effects of movement on sensory stimulation […]. An object looms larger in the visual field as we approach it, and its profile deforms as we move about it. A sound grows louder as we move nearer to its source. Movements of the hand over the surface of an object give rise to shifting sensations. As perceivers we are masters of the sort of sensory dependence […]. We spontaneously crane our necks, peer, squint, reach for our glasses, or draw near to get a better look (or better handle, sniff, lick, or listen to what interests us). The central claim of what I call the enactive approach is that our ability to perceive not only depends on, but is constituted by, our possession of this sort of sensorimotor knowledge. (Noë 2004, pp. 1–2)
The enactive approach correctly predicts that there will be adaptation to certain kinds of spatial distortion to visual input (Noë 2004). The idea is that if perception is a matter of learning sensorimotor contingencies, then though these contingencies can be disrupted via altering the spatial features of the input, the new contingencies can be learned (they are just a different set of contingencies, after all) and when this happens, perceptual input will seem normal again.
Noë (2004) boldly claims that not just spatial features, but even color phenomenology might be explained on enactive principles.
Our ability to perceive [a] wall’s color depends on our implicit understanding of the ways its apparent color varies as color-critical conditions vary. At ground, our grasp of these dependencies is a kind of sensorimotor knowledge. We can distinguish two different kinds of sensorimotor dependencies […]. Crucially, the perceptual experience of color depends on the perceiver’s knowledge of both kinds of sensorimotor patterns.
Movement-dependent sensorimotor contingencies are patterns of dependence between sensory stimulation, on the one hand, and movements of the body, on the other […].
[O]bject-dependent […] sensorimotor contingencies […] are patterns of dependence between sensory stimulation and the object’s movement, or the object’s changing relation to its environment. (Noë 2004, pp. 129–130)
Accordingly, our protocol speaks as directly to the enactive account of color as inverting prisms speak to an enactive account of vision’s spatiality. Notice that the technological apparatus of color rotation comes into play after both sorts of patterns have manifested (with the exception of eye movements, which happen after the rotation is effected, though this fact does not change any of the contingency patterns). What this means is this: suppose that there is a particular way that a red surface changes its reflectance properties both as we move around it (movement-dependent), and also as it changes relevantly with respect to the environment (object-dependent). Call this pattern of changes Pattern R. And suppose that the same is true for a blue surface, meaning that it has a different, but characteristic pattern of changes we can call Pattern B. To get a specific example, let’s suppose that the red surface gets brighter when it gets angled upwards, but the blue surface does not (the details of these patterns don’t matter for purposes of illustration, all that matters is that there are such patterns, and that they differ for different colors). Whatever the patterns R and B are, they occur whether anyone is wearing rotation gear or not. But after the rotation gear is involved the surface that is behaving according to Pattern R will be presented to the subject with stimulation from the blue part of the spectrum, and the surface that is behaving according to Pattern B will be presented with stimulation from the green part of the spectrum. For instance, the subject will see the apparently blue surface getting brighter as it is angled upwards, which is Pattern R, because the apparently blue surface is actually red, and hence behaves according to Pattern R.
So the question is: after experience with red surfaces, which behave according to objective patterns appropriate to red surfaces (Pattern R), but are presented through the goggles with blue parts of the spectrum, will these surfaces start to look red again? This is clearly the prediction that is made by the enactive theory of color, since to appear red just is to behave according to Pattern R on this theory, and the surfaces that are being presented with light from the blue parts of the spectrum are behaving according to Pattern R. The enactive theory makes this prediction for both color constancy and color phenomenology. We will discuss color constancy first.
We did not test color constancy in any controlled way, but the subjective reports are quite unmistakable. Subject RG noticed that upon first wearing the rotation gear color constancy went “out the window.” To take one example, in normal conditions RG’s office during the day is brightly lit enough that turning on the fluorescent light makes no noticeable difference to the appearance of anything in the office. But when he turned the lights on after first donning the gear, everything had an immediate significant change of hue (though not brightness). He spent several minutes flipping the light on and off in amazement. Another example is that he also noticed that when holding a colored wooden block, the surfaces changed their apparent color quite noticeably as he moved it and rotated it, as if the surfaces were actively altering their color like a chameleon. This was also a source of prolonged amusement. However, after a few days the effect disappeared. Turning the office light on had little noticeable effect on the color of anything in his office, and the surfaces of objects resumed their usual boring constancy as illumination conditions or angles altered.
Subject JK reported the same thing: an initial period in which the apparent colors of objects shifted widely with changes in illumination conditions or viewing angles, followed after a day or two with the restoration of color constancy such that those same changes had no effect on apparent color.
There was one difference, though, between RG and JK. While RG’s perceptual system gained the capacity for color constancy under rotation, he never lost color constancy in normal conditions. After the first few days of intermittently wearing the gear, objects had stable apparent colors whether he wore the gear or not. Though of course the colors that were stable were different in the two conditions. JK (who wore the gear continuously for six days) also gained color constancy under rotation, but lost it for normal conditions, as was apparent at the end of his trial when he removed the gear. Indeed, almost immediately after he removed the gear and was seeing things without rotation for the first time in six days, he spent several minutes flipping a light switch on and off and marvelling as the apparent color of everything in the room changed at his command (while no one else in the room noticed anything). So while RG’s visual system became, so to speak, bi-constant, because he switched back and forth between rotated and non-rotated visual input, JK’s visual system, because it was exclusively rotated for six days, gained color constancy under rotation, but temporarily lost normal color constancy. Normal constancy returned for JK within a few hours after he stopped wearing the gear.
These results are precisely what the enactive theory of color constancy would predict. Initially the kinds of patterns of color-relevant change exhibited by objects in the environment was different from what the visual system had come to expect, both in terms of changes in environmental conditions generally (e.g., switching on a fluorescent light), and movement specific changes (e.g., walking around them or rotating them in hand). The sensorimotor contingencies changed, and as a result color constancy was disrupted. But after a period of time during which these new dependencies were, presumably, learned, color constancy was restored.
A more convincing protocol would be one in which there were control subjects whose visual input was rotated, but who were not active in their color environment. Such a protocol would be difficult to implement. Wearing rotating equipment for six days is quite difficult. If you were then to disconnect visual input from overt behavior on top of that, this would become extremely burdensome for test subjects. This could be done either by having the video camera not moving at all, or moving randomly; or by recording the video from a rotated subject as they are actively exploring their environment, and simply play that video back to control subjects, so that their visual input would not change at all as a consequence of their own actions. But even though there was no control of this sort in our protocol, it is safe to say that the proponent of an enactive theory of color constancy should be encouraged by this result.
But what about color adaptation? Did red surfaces start to look red again? The results here are less encouraging for the enactive theorist. With one interesting and suggestive observation to be discussed shortly, we found nothing that suggested color adaptation. As assessed by subjective report, stop signs continued to appear blue, the sky green, and broccoli red throughout for both subjects.
Though this is not the result that the enactive theorist would hope for, it isn’t entirely conclusive. First, it may have been the case that a protocol of longer than six days would have resulted in phenomenal adaptation. Six days may simply not have been long enough to learn the new relevant sensorimotor contingencies. This is certainly possible. But it should be noted that all other sorts of visual adaptation (to spatial inversions, spectral filtering etc.), including our own result with adaptation to color constancy, occurred in less than six days.
Second, the gear does in fact introduce a lot of artefacts besides just the change in presented colors. Artefacts are introduced by the digitization of the image and its presentation through LCD goggles. So proponents of the enactive theory of color needn’t jump off a building just yet. They may maintain that because of these artefacts, the process of relearning the needed sensorimotor contingencies was somehow short-circuited.
But there are a couple of considerations that suggest that the result is not so easily dismissed. First, whatever artefacts the gear did introduce were not such as to make any difference to normal color vision. If one wears the gear without color-rotation, things appear to be in their normal colors. This seems to suggest that whatever patterns of change account for our ability to see normal things in their normal colors, they are not significantly compromised by whatever artefacts the gear introduced. This would seem to remove some of the motivation from the suggestion that adaptation did not occur because of artefacts introduced by the gear. Second, the gear clearly maintained enough information about patterns of color change that adaptation to color constancy occurred fairly quickly. Again, this suggests that much or all of whatever is important in patterns of change relevant to color perception is preserved by the gear. If it was not, adaptation to color constancy should not have occurred. The one thing not preserved by the rotation gear was the sensorimotor-independent feature of which retinal cells were stimulated when various surfaces were in view. And that one feature seems to be the best candidate for the determinant of apparent color, given that everything else changed but apparent color did not.
We mentioned above that there was an interesting pair of events that, while not quite amounting to phenomenal adaptation, are at least very suggestive. On two occasions late into his six-day period of wearing the gear, JK went into a sudden panic because he thought that the rotation equipment was malfunctioning and no longer rotating his visual input. Both times, as he reports it, he suddenly had the impression that everything was looking normal. This caused panic because if there was a glitch causing the equipment to no longer rotate his visual input, then the experimental protocol would be compromised, and the value of his days of sacrifice in the name of science and philosophy would have been significantly diminished.
However, the equipment was not malfunctioning on either occasion, a fact of which JK quickly convinced himself both times by explicitly reflecting on the colors that objects, specifically his hands, appeared to have: “OK, my hand looks purplish, and purple is what it should like under rotation, so the equipment is still working correctly.”
Prima facie there seems to be a clear difference between i) a tomato looking “normal” because it now appears phenomenally to be red; and ii) a tomato looking “normal” because though it appears blue, one is now used to tomatoes appearing blue, that is, blue no longer appears unusual. JK’s situation was a case of (ii), but the lack of a sense of novelty of strangeness made him briefly fear that he was in a (i) situation. He described it as a cessation of a “this is weird” signal.
Even though (i) and (ii) seem to be quite different, the phenomenon is suggestive. It indicates that there is definitely a stage in which the subject requires explicit reflection to discern (i) from (ii). This might lead one to speculate that this stage might signal an early stage of genuine color adaptation (we will discuss this further in the final section).
But it could also be an initial stage of a very different possibility, one discussed by Noë himself:
[…] the strongest [inverted spectrum] arguments ask us to consider the possibility in the first person. At stage 1 I am inverted. At stage 2, I get used to the inversion. I realize things now look color-inverted compared to the way they used to look, and I use this knowledge to guide my correct use of words. I get really good at acting normal. At stage 3, I suffer amnesia and forget that things ever looked different. The point of this thought experiment is that it suggests a reason to believe that things are now different with me with respect to my color experience, even though I am now unable to report those differences. (2004, p. 94)
While JK never suffered from amnesia, his two episodes suggest that it is possible to at least go some distance down precisely this path.