Grush et al.’s work could be extended to manifest a broader range of philosophical implications than those they have mentioned; but, as the authors state at the end of their article, this has been left for future philosophical and psychological investigation. Referring to a general theoretical framework of perception such as the enactive approach, Grush and colleagues apply the lessons of their study to sensorimotor enactivism of perception without considering other options such as ecological and active perception approaches as potential targets (Gibson 1979; Ballard 1991; Mossio & Taraborelli 2008; Taraborelli & Mossio 2008). However, since their hypothesis focuses on the nature of perception based on couplings between sensory stimulation and motor activity, it appears justified to focus on the sensorimotor version of the enactive account, which emphasizes an active exploration of the environment determining in this way the content and modality of conscious experience (O’Regan & Noë 2001; Noë 2005).
Sensorimotor theory has been supported by research on sensory substitution (Proulx & Störig 2006) and adaptation in haptic perception, as observed in mirror therapy for phantom limb pain and in the rubber hand illusion (Ramachandran & Rogers-Ramachandran 1996; Botvinick & Cohen 1998). Most relevantly, supporting evidence for the sensorimotor theory of color perception was found in a study on adaptation to half-split colored goggles (left-field blue/right-field yellow), which introduced an artificial contingency between eye movements and color changes (Bompas & O’Regan 2006b; cf. Kohler 1962). These results have left the possibility of similar sensorimotor adaptation to any arbitrarily-chosen colors open. According to the account of enactive vision, sensorimotor principles are fully capable of explaining adaptation to alterations in spatial or color-relevant features of input (Noë 2005). The adaptation can be achieved by resuming constancy through learning a new set of sensorimotor contingencies, i.e., patterns of dependence between sensory stimulation and movements, corresponding to new features of the input. Understanding these dependencies provides the required sensorimotor knowledge that enables perceptual experience.
The experimental protocol of Grush et al.’s study directly refers to an enactive account of color (O’Regan & Noë 2001; Noë 2005; Bompas & O’Regan 2006a, 2006b). The authors’ hypothesis regarding color constancy and phenomenal color adaptation under color rotation is compatible with predictions made by sensorimotor enactivism; the induced adaptation to a remapped spectrum was supposed to imitate a naturally-occurring process of learning sensorimotor contingencies. The results obtained in this pilot study, although not entirely usable and interpretable, may yet provide food for thought to enactive theory, since they offer some interesting insights into supportive evidence and the difficulties that the theory needs to integrate and deal with.
Subjective reports concerning adaptation to color constancy, understood as achieving stability of color experience irrespective of visual conditions, confirm what the enactive theory would expect. This means that when switching between standard visual conditions and color rotation, and at the same time being active in the color environment through altering color-critical conditions such as illumination, viewing angles, or movements, the test persons exhibited temporary disruption of color constancy leading to an immediate change of perceived hues. This is allegedly due to the change of sensorimotor contingencies involved in this experience.
However, when a new set of sensorimotor regularities becomes established, color constancy is resumed, so that the subjects gain the capacity for color constancy under rotation and then come back to normal color constancy when having non-rotated visual input, i.e., the colors that are stable are different in the two conditions. Hence, after a period of time for learning new dependencies, color constancy is restored and the mentioned modifications of visual conditions, such as lighting, have no effect on the phenomenal character of color experience.
An interesting observation and an important point for further deliberation on the development of phenomenal color adaptation is delivered in the subjective report of one of the test subjects, who at the end of his six-day color rotation period suddenly begun to be confused about whether his visual input was still rotated or not, because everything appeared normal. Since he ceased to feel a sense of novelty and strangeness, he was not sure if he was in a situation of (1) normal color vision, or rather (2) adaptation to color constancy under rotation—at least until he explicitly reflected on the colors of the surrounding objects. Although he was evidently in state (2), thus experiencing stability of rotated colors, one may suppose that his confusion about which colors were ‘normal’ in which condition might also indicate the time in which subjects could begin to develop an ability amounting to (3) phenomenal color adaptation under rotation with colors akin to genuine colors in situation (1). Speculations envisioning the occurrence of this adaptation after a longer period than the duration of the current test do not seem completely unjustified. What would be needed here are further studies that not only cover a longer time frame of color rotation, but also focus on searching for a characteristic marker signaling when, within a very smooth transition between (2) and (3), phenomenal adaption under rotation (stage(3)) actually begins. This would be similar to “the feeling of novelty/strangeness”-marker within the transition between (1) and (2), signaling color rotation. The lack of this marker and the occurrence of the feeling of normality would indicate that color constancy under rotation has arisen.
The memory color effect was used by Grush et al. as a method of assessment for phenomenal color adaptation under rotation. It is an effect of processing colors of objects with typical colors that affects the experience of pairings of colors and shapes (Hansen & Gegenfurtner 2006; Hansen et al. 2006). The authors explain the effect by top-down influences of expectations. But it may also be explained by, for example, cognitive penetration of color experience by beliefs (Macpherson 2012) or sensory adaptation through exposure manifesting itself by responding differently to various kinds of objects or co-occurring features (e.g., arrangements of objects’ shapes and their typical colors; Deroy 2013, 2014). All these descriptions express some aspect of the phenomenal liberalism discussed earlier, and as such they seem more or less equally plausible for supporting the proposed reading of phenomenal color adaptation under rotation as adaptation in the cognitive aspects of experience. In standard visual conditions, the memory color effect may suggest that expectations or beliefs about a proper color for a certain kind of objects exert top-down influence on the actual color processing of these objects, their shapes, etc. Thus, the lessened magnitude of the memory color effect under color rotation, as found in the study, shows that the associations of objects with their prototypical colors become weaker and may even get replaced by other associations with new prototypical colors.
This outcome is interestingly combined by Grush et al. with the aforementioned confusion stage (between (1) and (2)) acquired at the end of the color rotation period, when the subject stops having the feeling of novelty and therefore confuses his rotated color experiences with the normalcy felt when perceiving in standard visual conditions. Both of these results imply not only a decreasing strength of the old prototypical color associations, but also the emergence of new associations. Such an emergent set of dependencies is clearly compatible with enactive predictions. The adaptation that took place due to color rotation and that has been demonstrated by the memory color effect appears to be general. This means it is not just a matter of specific associations of colors with particular objects seen during rotation. The adaptation refers to the perceptual system as a whole and its expectations, beliefs, or sensitivity, contributing to a discriminative response to kinds of objects in general. For example, the adaptation might manifest itself as the regaining of a grasp of the way things are colored, as altered cognitive states (cognitive aspects of experience) about what red things generally look like or what red is like.
Obviously the study protocol would have been more plausible if color constancy had been tested in a controlled way with a relevant objective method and not only confirmed by first-person reports. For example, brain imaging techniques would be suitable for detecting temporary changes in perceptual states. Also, comparing the effect with a proper control group, matching the test group for gender, age, and color-related experience (e.g., education, profession), would certainly increase the strength of the findings, providing more evidence for sensorimotor adaptation to color constancy. Because transformations in qualitative experience may be explained in terms of a dynamic model of interdependence between sensory inputs and embodied activity (Hurley & Noë 2003), phenomenal differences between color experiences can be accounted for by different actions. Therefore to exclude the sensorial interpretation, the control group would not be actively exploring their color environment, would not change the rotated visual input through their own actions, and thus according to the enactive theory would not develop new sensorimotor dependencies allowing stable color perception.
For genuine phenomenal color adaptation different results were observed, i.e., the regaining of non-rotated color constancy while using the rotation equipment was not successfully established—subjective reports and objective assessments made with the memory color effect and aesthetic judgments of color-rotated food and people have shown that subjects only started to adapt in late-rotation, at the end of the possible adaptation period. Difficulties in robustly confirming phenomenal color adaptation under rotation are certainly not encouraging news for the enactive view of color. They could even be interpreted as a falsification of this theory. However, according to the investigators, this is still not decisive, and they speculate that the reason for this unfavorable outcome could be the lack of time allowed for relearning the relevant sensorimotor regularities. Indeed, for someone whose phenomenal color qualities remained rotated and did not revert to the genuine color phenomenology, i.e., for whom tomatoes continued to look blue, but did not reappear as red, this may be the case, because perceptual learning, here resulting in action-sensation coupling, is a relatively slow process and its timing varies from one individual to another (Goldstone 1998; Seitz & Watanabe 2005). Such an explanation remains in line with the sensorimotor account of perception and cannot be excluded without further studies. On the other hand, it may be also possible that the development of adaptation under rotation took place unconsciously and therefore was not reported by the subjects.