Colors and their Objective Basis

In 1976, when I delivered the John Locke Lectures at Oxford, I often spent time with Peter Strawson, and one day at lunch he made a remark I have never been able to forget. He said, "Surely half the pleasure of life is sardonic comment on the passing show".  This blog is devoted to comments, not all of them sardonic, on the passing philosophical show.

Hilary Putnam

I will begin by discussing a question from my color-skeptical friend, who recently sent me a beautiful color-illusion. I have been unable to find it online, but the illusion involves a picture crossed by a horizontal white line. From this line, in the center of the picture, a vertical line white line runs straight down. The upper left quadrant of the picture is solid red (the right edge of the red square is just above the vertical white line), and the upper right quadrant is solid green. In addition, there are two white dots or small white circles, a slightly larger dot in the upper half of the picture, above the vertical white line, and a smaller white dot in the lower half of the picture, right on the vertical white line. The bottom half of the picture (the half below the horizontal white line) shows a landscape with  brown hills below a blue sky with clouds. On a second look, one sees that the right side of the bottom half is the mirror image of the left half, i.e., the lower left quadrant is “reflected” around the vertical white line.

The instructions that go with the illusion are as follows: “fix your gaze on the upper white dot and count to 50. The fix your gaze on the lower white dot.” When I tried it, the effect was startling. When I shifted my gaze from the upper to the lower white dot, the hills in the lower right quadrant appeared to change from brown to a verdant green! (This is called a “color accommodation response”.) The look of the lower left quadrant was unaffected, so that the two halves of the picture below the horizontal white line were no longer mirror images—not as far as the colors I seemed to see were concerned!

My friend wonders how “color can be objective in the face of such variation”.

I shall begin with my answer to his question, and continue in future posts with a further discussion of “looks”.

Colors and color-looks

If someone had been standing next to me while I carried out the above instructions, they would not have observed any change in the picture. And indeed the colors on the screen (or on the page, if the illusion had been printed in a book) would not have changed. The “variation” my friend speaks of is not a variation in the actual colors. It is a variation in the look of certain colored regions, not in the actual color of those regions. Similarly, when I turn out the lights in my bedroom at night, my slippers (which I can barely see in the dark) no longer look light gray but almost black. And when I put the milk bottle in the refrigerator, it no longer looks white. Unless the refrigerator light fails to go off and there are little people in the refrigerator to look at it, it doesn’t “look” any way at all. But my slippers are light gray, even when one can’t make that out because the room is too dark, and the milk bottle is still white, even when one can’t see it. By assuming that colors are looks, my friend begged the question for the antirealist side.

But didn’t I beg it as well by assuming there is such a thing as the “actual color”? That is the question to which I now turn.

The objective basis of color discriminations

One obvious fact about colors is that some things are indistinguishable in color. Specifically, a uniformly colored part of a surface, for example part of a wall or of the surface of an object, may “match” a color sample. People with normal color vision will not be able to distinguish any difference in the color of the surface or surface-part and the color of the color sample, say a colored square on a chart provided by a paint company or an interior decorator. And this “matching” relation will persist under the same viewing conditions and for normal viewers, unless the surface or surface-part in question undergoes a physical change.  The obvious, indeed trivial, explanation is that whatever the “nature” of color may be, colors have an objective basis. They are not just “in our heads”. Whether the “objective basis” is the color (the color physicalist position) or is only the explanation of the intersubjectivity of color-judgments is a further question. But that there is an objective basis is, I believe, undeniable. Otherwise, how are we to explain that the traffic lights change their colors as they are supposed to, and as a result of a physical  signal?

By using the word “objective” here, I do not mean only that the properties that cause surfaces or parts of surfaces to match color samples are properties about whose possession or non-possession there is a fact of the matter  (metaphysical objectivity), but they they are viewer-independent, in fact physical, properties of the objects.  They are not fundamental physical properties, or simple functions of the fundamental physical properties, nor are they of interest to physicists who are seeking fundamental laws. But we have a reasonable account of them: they are reflectance types. Here I agree with Alex Byrne and David R. Hilbert, who write (“Colors and Reflectances,” in Readings on Color, Volume 1: The Philosophy of Color, MIT Press, 1997; online at

 http://web.mit.edu/abyrne/www/colors&reflectances.html):

“But is there a physical property that all and only (actual and possible) green objects share? Given our assumption about the general correctness of our color perceptions, the answer is (plausibly) yes. The property that all green objects share is a type of SSR [=Surface Spectral Reflectance - HP]. Very roughly, this property -- call it ‘SSR_GREEN’ -- is the type of SSR that allows an object in normal illumination to reflect significantly more light in the middle-wavelength part of the spectrum than in the long-wavelength part, and approximately the same amount of light in the short-wavelength part as in the rest. Obviously, particular reflectances meeting these specifications -- for instance those of frogs, lettuce, and dollar bills -- may be otherwise very different.

The property green, if it is this type of SSR, is not a particularly interesting property from a physical point of view. Since we only find it salient because our perceptual apparatus is built to detect it, it might be called an anthropocentric property (cf. Hilbert 1987). Alien physicists lacking our visual apparatus would not need to single it out for special attention, unlike the property of having charge e, or spin 1/2. (These aliens might likewise find the visible spectrum no more than an arbitrary segment of the entire electromagnetic spectrum.) But that does not at all impugn the status of an idiosyncratic type of SSR as a physical property that is ‘objective,’ in almost every sense of that protean word. Particular SSRs are not in any philosophically interesting sense dependent on human beings; neither is the type, ‘either SSR_a or SSR_b or SSR_g, ... ,’’ where these particular SSRs seem from a physical standpoint to be a motley collection.”

I would add, moreover, that the property green, and many other color properties, can be discriminated by several non-human species as well, including (after appropriate training) domestic cats.

Why we should identify the SSR-types of which Byrne and Hilbert speak with colors

Sometimes the basis for calling something an X does not support the appellation X. For example, if the basis for calling someone a witch is that her community believes that she causes other people to suffer illnesses and other misfortunes by means of black magic, then the appellation wrong and the basis is insufficient.  If the basis for calling a surface green is that we mistakenly believe it have a property that nothing could possess, greenness, then even if the objects we call green have a common property that prompts our brains to classify it as “green”, the basis is insufficient. But the only serious arguments I have seen to support the idea that nothing “out there” could actually be green are: (1) Russell’s argument in Problems of Philosophy that “the colors we see” depend on the viewer, an argument that conflates colors and looks of colors; and (2) Sellars’ argument that physical properties aren’t what color, solidity, and the like, are “presented as”, an argument that assumes we cannot refer to a property by a term if our mental image  of the property does not resemble the property described by science—an argument that is an obvious descendant of the 17^th century empiricist ideas that our “ideas” are mental images and they refer to what is similar to them. In my previous post I argued that these arguments lead to skepticism not just about colors but about all the properties of objects, and that they depend on untenable requirement on what it takes to refer to something real. If we reject them, as we should, I see no serious reason not to agree with Byrne and Hilbert’s “physicalism” about color.

I know that more needs to be said about “looks”. And I will say more – in future posts.

The Manifest Image is not Wrong
(revised July 15)

In 1976, when I delivered the John Locke Lectures at Oxford, I often spent time with Peter Strawson, and one day at lunch he made a remark I have never been able to forget. He said, "Surely half the pleasure of life is sardonic comment on the passing show".  This blog is devoted to comments, not all of them sardonic, on the passing philosophical show.

Hilary Putnam

In my last post (July 10, ’14), I referred to two views of color that I regard as mistaken; the view that colors are simply not real, and the view that they are dispositions to produce certain sorts of sense data under certain conditions, as well as to the view that I favor, according to which colors are dispositions of objects to affect light in certain ways, ways that in turn affect our eyes and nervous systems and thereby enable us (and O. aegina) to identify and re-identify the objects more easily. Although the second of the views I reject, the view that colors are dispositions to produce sense data [Do octopuses have the sort of sense data we do?], is more charitable to the person on the street than the simple denial that colors are real, the consequences of either view are a wholesale skepticism about the veridicality of what Sellars famously called “the manifest image”. [See http://plato.stanford.edu/entries/sellars/] Recall: In Science, Perception and Reality, Sellars famously wrote:

“Pink does not seem to be made up of imperceptible qualities in the way in which being a ladder is made up of being cylindrical (the rungs), rectangular (the frame), wooden, etc. The manifest ice cube presents itself to us as something which is pink through and through, as a pink continuum, all the regions of which, however small, are pink. It presents itself to us as ultimately homogeneous; and an ice cube variegated in colour is, though not homogeneous in its specific colour, ‘ultimately homogeneous’, in the sense to which I am calling attention, with respect to the generic trait of being colored.”

From the fact that pinkness is not reducible (in the highly restricted way Sellars described) to properties of and structural relations among the atoms of which an ice cube consists, Sellars concluded that the “manifest image”, according to which there are such things as pink ice cubes is wrong, and the scientific image is right.

Moreover, Sellars would certainly have rejected the view that colors are dispositions (in either of the ways mentioned above) on the ground that colors are not “presented to us” as dispositions. Yet Sellars famously claimed (in the very book from which I quoted) that  “in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not”.  This would seem to leave the status of “pink” in limbo, neither in the world nor in the head (colors are not “presented to us” as brain phenomena either, of course). But this is not a post on Sellars interpretation.

What I want to focus on, and what Sellars himself I suppose to have been fully aware of,  is that if you accept his “grain argument” against the very possibility of identifing anything phenomenal with anything scientifically respectable, then solidity is no more real than color. For Newton already knew that particles have “Pores or Empty Spaces between them”, indeed, the proportion of “empty space” to “particles” (which Newton still conceived of as solid) had to be enormous (as Newton himelf calculated). Nor is the criterion for a collection of (today, quantum mechanical) particles being a “solid” a simple one, nor is it free of reference to forces, which are arguably dispositions. In sum, solidity is not “presented to us as” the complex quantum electro-dynamical phenomenon that it is. So, if Sellars is right, there aren’t solid ice cubes at all, never mind the color bit. And, contrary to Davidson’s famous Principle of Charity or any similar principle (something required if there is to be any rationale for interpretation at all!), almost everything the ordinary speaker of a natural language says is false.

Even worse, we cannot really say that at least we know what is strictly true, namely the “scientific image”. For each century seems to produce a radically different scientific image, and we not know what the scientific image of the next century might look like. (For a debate on this, see the very first post on this blog.) Yet the many relevant changes in our scientific image since Newton do not lead scientists to say that gravitation does not really exist; not should it. They say that Newton was right that gravitation exists, but wrong about its nature. And similarly, we can say that pinkness exists, but the manifest image is wrong about its nature. Both the manifest image and the scientific image are, in many respects, approximately right (although the standards of accuracy appropriate to them are, of course, very different).

But the whole Sellars argument depends on the premise that As can’t really be Bs unless there is an explicit (and rather transparent, it seems from his examples) reduction of A-terms to B-terms (terms like “ladder” to terms like “rung” and “fastened to”). I don’t buy that premise. It leads to disastrously skeptical metaphysics and (if some principle of charity in interpretation is essential to making meaning-assignments empirically testable) bad philosophy of language as well. I think the manifest image is largely right, and that what science does is to explain why it is largely right, not show that it is wholly wrong.

Colors and Looks (1)

In 1976, when I delivered the John Locke Lectures at Oxford, I often spent time with Peter Strawson, and one day at lunch he made a remark I have never been able to forget. He said, "Surely half the pleasure of life is sardonic comment on the passing show".  This blog is devoted to comments, not all of them sardonic, on the passing philosophical show.

Hilary Putnam

In a previous post (“Escher Colors”, 6/11/14), I pointed out that that colors do not look the same to all normal people, on any reasonable construal of ‘normal’.  (See also Ned Block’s “Wittgenstein and Qualia”, in Maria Baghramian (ed.), Reading Putnam.)  As I pointed out in that post, colors do not look quite the same when viewed with one eye closed and when viewed with both eyes open, or when viewed with the other eye closed, even when the illumination and the distance and the angle of vision, etc., are held constant. There is a temptation to think that the questions, “Does my right eye see such the color of such and such as it really is?” and “Does my left eye see such the color of such and such as it really is?” make sense, but on reflection one realizes that they don’t. “The color I see”, as Russell would have put it in The Problems of Philosophy, isn’t a color anything could “really be”, for if it were then “The color I see when my left eye is shut” could be the real color of the object, and “The color I see when my right eye is shut” could equally be the real color of the object, but those “colors” are not the same, and the object could not be both colors at once.  Subjective colors are what I called “Escher colors”; phenomenologically it seems as if they are “diaphanous”, as if each eye in turn is seeing the object “as it really is”, but it is impossible for any object then to be both of those ways, and, given that neither way of looking is the way the object looks to all normal people in those circumstances, neither is there any sense in saying one way of looking is “veridical” and the other “non-veridical”. So I argued in that post.

But why not conclude, a friend asks, that colors aren’t real?  The “person on the street” project colors onto the surfaces of objects,  but they aren’t really there. This was, indeed, Russell’s view, and, before him, the view of Hume as well (except that Hume thought that objects aren’t really “out there” either, but that is another story). 

This story about how the person on the street mistakes what is in her own mind for properties of “external things” has not gone unchallenged. In the 18^th century, Thomas Reid, Hume’s great contemporary and rival, argued that the person on the street knows perfectly well the difference between a color and a sensation, and he was absolutely right. We do not think that objects change their color between 10AM and 7PM, although the look of any given colored object changes. We don’t think an object changes color when a cloud passes over the sun, although we are all aware that its look changes. Most people are not aware of the variability of subjective looks, for example, that objects that look bluish green to one person might look pure green to another, and the phenomenon of fine shades of difference in the apparent saturation of colors owing to differences in the pigmentation in (“normal”) macular areas that I mentioned in Escher Colors was unknown to even the optical technician I mentioned in that post, but that a color has a number of different looks depending on the conditions is not news. Reid was right, that the empiricist story makes the person on the street too naïve.

An old story, which is not as uncharitable to the person on the street as the claim that she projects what are really subjective looks onto the surfaces of objects goes like this:  colors are dispositions to produce certain sense data (under appropriate conditions). This story still assumes the person on the street is confused: she thinks that her sense data are looks of objects in an objective sense, the sort of looks that can be recorded with a camera. According to this story, she does not realize that what she calls “looks” are themselves sense data. But she does know that the dispositions to have those looks is a property of the objects, and she does not agree with Russell’s view that the objects themselves are never experienced, but at best “inferred”.

But this story too has problems (which I will talk about in the next post).

My interest in all this is not merely historical. The relations and distinctions between objective and subjective looks, objective colors, sense data (or to use the currently fashionable term, qualia), perception, experience, and apperception, are all topics Hilla Jacobson and I have been discussing for seven years, as we work on a planned book on apperception and consciousness,  and we have eached published things that touch on these topic  (and also have forthcoming unpublished papers on them).  In posts subsequent to this one, I plan to continue my answer to my friend’s question, why I persist in the view that colors are real. But I can  give a preliminary answer: it is nowadays a respectable (and I believe correct) view that many (though not all sorts of) dispositional properties of objects are explanatory and ontologically robust. Properties that enter into explaining the survival value of various phenotypes (and hence of the corresponding genotypes in certain environments) are explanatory and robust in this way. And colors are such properties, as evidenced by the fact that convergent evolution has led very different species (e.g., humans and at least one species of octopus) to develop eyes able to differentially respond to colors. [See Gunzo Kawamura, Kazuo Nobutoki, Kazuhiko Anraku, Yoshito Tanaka, and Masaru Okamoto, "Color Discrimination Conditioning in Two Octopus Octopus aegina and O. vulgaris", Nippon Suisan Gaikashi, vo. 67 (2001), pp. 35-39, https://www.jstage.jst.go.jp/article/suisan1932/67/1/67_1_35/_article]

I don’t believe colors are dispositions to produce particular sense data, as will be explained in subsequent posts. But it is plausible that they are dispositions to affect light reflected from the objects that have those disposition in certain ways, ways that in turn affect our eyes (and those of Octopus aegina), and thereby enable us (and O. aegina) to identify and re-identify those objects more easily.

Of course, if you think that things as intuitively different and colors and dispositions (of the kind just mentioned) just cannot be contingently identical—if you don’t think that science can and does discover non-analytic identities between properties—then you will reject this account of colors as physical dispositions out of hand. But then, I have long argued, you are making a serious mistake. [See my (1970) “On Properties”, collected in Mathematics, Matter, and Method, vol. 1 of my Philosophical Papers (1975).]