Noë on color

Question: Noë writes (143): “[T]he account of color I offer here treats color as irreducibly phenomenal, but it avoids the circularity of saying that red things are those that look red.” How does he claim to avoid the circularity? Does he succeed?

Noë maintains that his view preserves the phenomenality of color while avoiding the circularity given above. It may be worth pointing out that there seems to be an unavoidable and trivial circularity in any account of experience. One always has to admit, for instance, "Red things are such that we consider them to be red," whatever our assessment of things as red amounts to. [Levin—At one point he calls such observations “innocuous”.] It is hard to see how a phenomenal account of color could avoid relying on the appearance of red; it's equally hard to see how a comprehensive account of color experience could be merely phenomenal. If one rejects qualia as having no explanatory bearing, one tends to look elsewhere for explanations, seeking a theory that bridges color-experiencing and -descriminating capacities in the subject (taking into account individual variation) with physical properties and causes in the object and its environment. Noë, presumably, expects that such a bridge theory would neglect the phenomenality of color. The sensorimotor profile theory he develops instead, I will suggest, neglects such objectivity as we can hope for. I agree that qualitative reports emanate from a subjective look-and-feel, but the qualities themselves are not public, and reports of them form the merest starting point in the search for objective explanation. Explanations of ineffabilia by appeal to ineffabilia will not lead anywhere. In this sense, there is no “irreducible phenomenality” worth preserving at the cost of objectivity. The reported variations in qualitative experience in the research Noë cites only triangulate on our notion of color to the extent that a fuller story can be told, insofar as variations in sensorimotor and contextual factors ultimately can be related to the physics, biology, and cognitive science relevant to both subject and object. There is no need to seek a richer yet still fundamentally phenomenal characterization of color, as Noë tries to do. Research on sensorimotor profiles and the like does serve a purpose; it lays out a taxonomy of observation circumstances and their corresponding qualitative reports. This is useful, even though it multiplies explananda, because it makes the explanatory project clearer and more robust. What it doesn't do, independently, is any of the explanation.

Noë's project is motivated by his acceptance of the ineffability of color experience despite his dissatisfaction with dispositional theories that use qualia as a way of avoiding circularity. He claims a disanalogy between experiencing shapes and experiencing colors. The experience of viewing a coin at an angle has two distinct aspects, whereas the experience of a color has only one. The coin has an elliptical look and a cylindrical or disk-like actual shape, whereas a color seems to have, and to be, just its look [Levin—or a function of its looks]. The empirical results Noë cites are meant to build a case for a contextual and sensorimotor account of color perception that puts it on par with shape perception yet retains the essential phenomenality of color. I don't believe there is such a disanalogy. Noë gives an account of our ability, for instance, to interpret the color of a variably lit wall to be uniform despite splotches of color caused by the dimness or unevenness of the lighting. This strikes me as entirely analogous to the orientation of the coin and the look we interpret as that of a coin viewed from an angle. Noë's point seems to be that all the colors on the wall—all the looks—are themselves colors, whereas the same phenomenality does not hold for the coin viewed at an angle. Even though the elliptical look of the coin is itself a shape, Noë is loath to relinquish a privileged phenomenality for color. He says, “The shape of a thing is independent of its look or feel in a way that the color of a thing isn't independent of its look.” Yet looks of shapes are phenomenally individuated, perhaps as finely as are looks of colors. Looks of shapes are shapes, just as looks of colors are colors. [Levin—still, I suppose he would reply, the shape of an object is a characteristic that can be given intrinsically, independent of its looks. A shape is not a shape-appearance profile. I see you soon get to this same point.]

The word 'shape', for better or worse, is used in both a phenomenal and an objective sense. Noë's misstep, as I see it, strangely enough, may originate in the tendency to preserve a phenomenal sense of 'shape' when talking about the actual or objective shape of the coin. When we consider the coin independently of any particular act of perception, we claim for it an actual shape. This usage of 'shape' should strive for objectivity, yet it carries with it the intuition that the true shape of the coin, under the right circumstances, might be experienced as a phenomenal shape that happens to conform in every respect to the coin “as it really is.”1 I'm not sure the use of “true shape” or “objective shape' does carry this intuition with it. I'm not sure that '”the box is a cube” carries the suggestion that there is some way of looking at it that yields its cubicalness. When we consider a color independently of perception, on the other hand, there seems to be no place for an actual color—all colors are experienced. In order to bridge this apparent gap, Noë appeals to the notion of an exhaustive 'sensorimotor profile' intended to serve as something like a schema for actual color, which effectively gets defined as the range of alterations of apparent color (or colors) in various circumstances. One aspect of this approach is that it does not overcome the supposed disanalogy between color-experience and shape-experience. The actual shape of a coin is not the sum of all viewings of it; it is a geometric description (perhaps including the coin's dimensions, although size, strictly, is not related to shape). For Noë, preserving the disanalogy is presumably "a feature not a bug" of his theory, since he hopes to preserve a distinctive phenomenality for color.

Does Noë avoid circularity? He seems to be in a position to replace, “Red things are those that look red,” with something like, “Red things are those that exhibit the following range of sensorimotor color profiles:…” This involves a multitude of profile-sentences of the form, "This red thing is such that it looks brown in situation S1 (under dim fluorescent light, say), orange in S2 (under a vapor lamp, say), etc.," along with a theory that maps changes in lighting, perspective, location, and so on to changes in 'look'. The theory will have to individuate the physical characteristics of different sorts of colored thing, because different materials often seem to share the same color in one circumstance but not in others. I believe this ties Noë's account of color to a full-blown physical theory of color perception, encompassing colored objects and their physical properties, causal interactions among light sources and reflective or translucent media, and the structural 1 This is impossible, of course, not just for the subtle reasons Noë mentions but also because we cannot view opposite sides of opaque objects, and because there is no 'right' perspective. At best we can stipulate a 'canonical shape' (see below) as the apparent shape of an object when viewed from a certain angle, under certain conditions. and cognitive 'intervention' of the perceiving subject. The shift of color profiles from self-contained phenomenal schemata to intermediate steps on the way to a physics of color perception tends to push Noë's theory away from its phenomenal point of origin. [Levin—That is so.]

Under a full physical theory of color perception, the usage of 'color', 'shape' , 'look' , and what I have in mind for 'objective' would fall out as shown in the examples below, which also suggest how the apparent disanalogy between color and shape might be reconciled. Note that there is still room for what we might call “canonical looks” of shaped or colored objects, that is, 'standard' phenomenal views stipulated to be normal within, say, a certain statistical tolerance based on reports—the red of the tomato viewed in sunlight comes to mind. These are to be distinguished from “objective qualities,” a near-oxymoron intended to capture the physical features of objects that permit them to be experienced in more or less consistent ways by different observers, given a suitable account of perceptual faculties. Canonical looks are arbitrary, though guided by empirical data; objective qualities, meanwhile, may be thoroughly non-intuitive, depending on the science involved.

'look':

1a) The look of a coin viewed at a 45-degree angle is elliptical.

1b) The look of a cherry-red wall at twilight is a splotchy field of colors ranging from dark red to brownish- grey. The look of each patch of different color is just that color.

“canonical look”:

2a) The canonical look of a coin's shape is defined as the look of the coin viewed orthogonally to the plane of one of its sides, i.e., circular. You might have to add distance; the look of a coin viewed along a line normal to its surface from five miles away will not be canonical. How about at arm’s length?

2b) The canonical look of a cherry-red wall is defined as the look of the wall under broad-spectrum floodlights that illuminate it without perceptible shadows or other variations in intensity, saturation, etc., i.e., cherry red.

“objective qualities”:

3a) The objective shape of a coin is that of a flattened cylinder (say) 350mm in diameter and 2mm in height.

3b) The objective color of a cherry-red wall is the RGB value #551110, or (perhaps better) the reflected wavelength gradient of the material out of which the paint is made, which reflects light at 630nm at an angle of 90º.

A final note on virtuality: Noë says that a colored object will change its apparent color even as it is being viewed. If the reason for this, as he suggests, is the inescapable 'virtuality ' of some constituents of any experience, an incompleteness of conscious apprehension that holds “holographically” (i.e., at any scale), then experiences of shape should be expected to exhibit, in their own way, this incompleteness and fluctuation of aspect. The question seems empirical and testable. Even if shape-experiences were found not to fluctuate as color-experiences do, it would speak to the robustness or precision of the structures and mechanisms for geometric perception, not to a fundamental disparity between color and shape regarding their status as perceptions caused partly by objective features of objects, partly by sensorimotor or other contributions of the perceiver.