Sunday, March 29, 2015

Color and the dress

Does the Great Dress Debate of 2015 suggest that colors are subjective? If we can’t agree on the color, does it follow that color is in the eye of the beholder?

For those who use their time more productively than in following internet bunfights, here’s the story:

I profess color realism, so I claim the dress has a color – two, actually.

It’s blue with black stripes.

However, the debate does provide insight into how we see the colors that really are there.

The argument for color realism is long, so consider the following ‘Ten Theses of Color Realism’ nailed, unargued-for, on the cathedral door.
1. An account of Everything That Is So must include not just physical things, but how they appear. To deny this is so costly that no consideration scientific or philosophical is powerful enough to compel it.
So eliminativism is not in the running: it takes too much of Neurath’s Ship apart at one time.
2. If there were no living creatures there would be no appearances. But this does not imply that appearances are somehow derivatively real, like shadows or reflections.
If there were no water there would be no living creatures either, but we’re not like shadows or reflections.
3. The concept of color incorporates two elements: one having to do with appearances, the other with whatever it is about the object’s possession of the color property that makes it appear that way. So our concept of color is of an objective property, the nature of which is revealed in its appearance.
This twofold feature is the source of the ‘queerness’ arguments for error theory (see 5 below). Those arguments presuppose that all properties are physical, thus begging precisely the question at issue.
4. Reductive physicalism regarding color founders on a stark contrast with other reductive programs. Given the state of scientific knowledge, if water were not H2O, there simply is no other candidate. We would be completely at a loss. But for color there is a whole cat’s chorus of candidates, none strikingly more plausible than the others, all agreeing on which property goes with which color. Choosing one as the property necessarily identical with color would be arbitrary. Therefore none are.
General agreement on the science combined with non-converging disagreement on the reduction shows that, short of a paradigm change in physics, the nature of color will not be settled by science.

It’s philosophy.
5. Error theory, the standard view since Galileo, falters because it fails a presupposition necessary for stating the view. Error theory cannot supply an explanation, meeting current externalist standards, for how we can have formed the concept of ‘red’ in the absence of anything red.
If a philosophical position must be statable, error theory is not in the running either.
6. A thing’s color supervenes on its physical composition. Since different physical arrangements may yield the same color, the supervenience relation is weak. It’s also not ‘physicalistically kosher.’ Physically indiscernible things could be different colors in other possible worlds with the same physical laws as ours, but different psycho-physical bridge laws.
To use Terence Horgan’s term, it’s not a case of ‘superdupervenience.’
7. An explanation for a thing’s appearing red is that is, indeed, red. We can misidentify what color it is, of course, but getting it wrong presupposes that there is something to get right. This is relevant to the Dress Debate.
8. Facts about color are facts about an autonomous realm over and above the physical realm. In this it resembles evolutionary biology. The explanation for why I am seeing red in a given case cannot be given in physical terms any more than we can explain in purely physical terms why one phenotypic trait confers fitness and another doesn’t.
9. The color realm is autonomous because no consideration intruding from outside the forms of explanation peculiar to that realm can compel us to retract a color judgment. Coming to believe that we are being exposed to 650 nanometer electromagnetic radiation does precisely nothing to weaken our conviction that the object is red. This suggests a high degree of autonomy.
10. We can know objective truths about color by observation: red is more similar to purple than either is to green. We can formulate autonomous laws of color that express necessities, and even support counterfactuals. The laws will prescind from any physical base. Again, in this it is similar to the realm of evolutionary biology.
An essay by psychologist Steven Pinker provides a brief explanation of the relevant autonomous laws of color.[1] Notice that it is couched in explicitly psychological terms: “lightness constancy,” “color constancy.” Both are instances of representational constancy, for which there is no physicalist account. Pinker speaks metonymically of “the brain” achieving representational constancy instead of the visual system, but we shouldn’t be confused.

Pinker’s conclusion is congenial to the color realist: The color of an object can appear a different color under “extreme” conditions, conditions removed from those our visual system evolved to handle. We can then become badly confused over what color the object really is.

To be a realist regarding some realm is to resign oneself to imperfect epistemic access to it. That our access to real things, including colors, is fallible and imperfect was never a thesis in dispute.

Thomas Pyne
Department of Philosophy
Sacramento State



  1. Tom, thanks very much for this. Each time you talk about it here it becomes more interesting.

    Early on you point out that appearances are not derivatively real, denying that they are like shadows. But I'm not sure exactly what your point is here. You don't mean to deny that appearances are subjective right? After all, as you see it, colors are not appearances, they explain appearances.And you can make the same point about subjectivity. Our subjective states are objective features of the world. I think I have another question, but I would need to be sure what I'm saying here is ok, first.

  2. Randy,

    Yeah, ‘appearance’ is a difficult notion to get around, given how early modern philosophy has corrupted us. I know I have trouble getting around it.

    Anyway, let me try and explain.

    ‘Appearance’ here is not as the early modern philosophers conceived of it: logically private, subjective, purely mental content.

    I’m borrowing the usage from John McDowell (as I understand him at any rate). For ancient philosophers the ‘world of appearance’ was just the external objective world. It was not a separate realm of facts internal to each of us. If the appearances of things can trick us sometimes, it’s that external world we’re wrong about. It was Descartes who invented the world of appearances as a separate realm of internal facts.
    It’s this early modern notion of ‘appearance’ that we assume, and that causes the mischief.

    The appearance of something is an objective feature, one we can have access to if suitably situated.

    Appearances are subjective only in the sense that, if you have a particular faculty, an external object will appear to you in a certain way. If not – you’re color blind or a Martian, say – then it won’t.

    Appearances are objective in the sense that agreement is more than merely inter-subjective. Agreement results from shared access to the same feature of the object.

    What kind of feature is an appearance?

    If there were no creatures in the world with perceptual or cognitive faculties then things would not have appearances. But that’s similar to saying that if there were no points of view on the world, nothing would be to the left of anything else. However whether something is to left of something else from that point of view is an objective fact. We can know that fact even if we don’t have that point of view.

    We don’t generally conclude from something’s being dependent on a given relation to something else that it is consequently only derivatively real or epiphenomenal. The facts of evolutionary biology depend on (supervene upon) physical facts, but it doesn’t make them any less objectively real. Likewise the appearances things have depend on their relation to perceivers and thinkers, but it’s only that early modern prejudice that allows us to think that appearances are therefore not objectively real.

    Can we say that, “If there were no creatures with color vision in the world then nothing would be red”? Forming a conception of what features would be present in such an ‘independent world’ is, as Barry Stroud has pointed out, a very slippery task. We might think that aliens who can’t smell are not missing out on anything ‘real,’ but that aliens who can’t do arithmetic are. We could think that the independent world lacks smells, but does contain numbers of entities whether we’re observing it or not. The only way we could say this, however, is if we use a prior conception of what features are there anyway. That is, we must presuppose some features in order to frame the question about what features there are.

    I think one common set of such prior conceptions should be examined.

    The usage of ‘appearance’ at work here is closer to the common meaning of the term: “That liquid has an oily appearance.” Here the term refers to a feature out there in the world that is common to all our observations. It doesn’t designate some micro-property of the liquid, itself unobservable, that produces in each of us a private state that somehow (magically?) we all give the same name.

  3. Tom, thanks, that's very helpful.

    Do you think the kind of objectivity you attribute to appearance is unavailable to someone who believes that the way things appear to us is determined by a complex physical interaction (roughly,between our nervous system and the external world)? As someone who is inclined to roughly that view, I would also say that appearances are objective in the sense that the objective features of that interaction determine the appearance; and counterfactually so if no perceivers actually existed.

    I am not actually sure that you even reject this view. I know you reject physicalism, but you seem to associate that with the view that appearances aren't real, and that's a view I've never really understood. The kind of physicalism I am attracted to would imply only that the way the world appears is fully determined by physical processes (where physical is understood by reference to what physics studies, not a common sense corpusculariainism.) Is there any part of my view that I need to reject in order for you and I to be in rough agreement (even if we might not use quite the same vocabulary)?

  4. Randy,
    You say, “…(T)he way things appear to us is determined by a complex physical interaction (roughly, between our nervous system and the external world.)?

    That seems right.

    It seems that we also agree, roughly, on what physicalism means. I would put it thus, borrowing from Barry Stroud: To be ‘physical’ means to be describable or explicable in the terms and concepts of physics.

    Where things get sticky is with the word ‘determined.’ It seems that neither of us thinks that appearances are identical to complex physical interactions. I don’t understand you to be a reductive physicalist. Appearances must then supervene on physical interactions in some fashion without being physical interactions themselves.
    So the issue then boils down to what kind of supervenience is involved. As I indicated in the blogpost, I think the color is an autonomous realm, a part of the autonomous realm of psychological properties and psychological facts. It has its own laws. Those laws could be the same even in worlds where the physics is different from the actual world. There could still be appearances – in fact things could be red – in worlds where the physical laws are different from ours. (Within limits, of course!)

    That’s a ‘weak’ supervenience relation. It’s also not physicalistically ‘kosher,’ since it’s not limited to our physics.

    It’s similar to evolution in this respect. It seems possible for evolution by natural selection, with its autonomous laws, to go on in possible worlds where the physical laws are quite different from ours. In fact it seems possible for evolution to proceed in a reality that is not physical at all, like a virtual entities in virtual world.

    Colors are determined by complex physical interactions in the actual world. That’s a contingent fact about color in the actual world. How they’re realized in other possible worlds is a different story. So they’re not physical properties.

  5. Tom, you're right, I'm not a reductive physicalist. I really like the openness of your approach, too. Philosophy that cuts off possibilities unnecessarily is bad philosophy. So the idea that redness is a property that could result from radically different physical interactions is worth preserving until we have strong reasons to reject it. I confess to a suspicion that the more we learn about our actual world the more we might have a basis for a stronger supervenience relation. More generally, I confess to a suspicion that the more we learn about the world, the more we will find that conscious minds probably can't result from anything other than carbon chemistry. But suspicion is all it is.

  6. Replies
    1. Yes, I'll wear that, though perhaps not entirely comfortably.

      l think at the end of the day I am still not entirely clear about your explanatory apparatus. When you say that something appears red because it is red, the objectivity of color you wish to preserve seems to lie in its status as an explanans, the task of which is to explain why things appear the way they do. But it's clear that you also wish to preserve the idea of appearances themselves as objective, and not just the kind of objectivity that anyone can give them, by acknowledging that, after all, appearances do really occur. So it sounds to me as if you are saying that redness is an objective appearance that also explains the way things appear.

      Looked at from a slightly different angle, I don't know how far you go in rejecting our Cartesian notion of subjectivity. It's one thing to reject the idea of a purely private object of which we have incorrigible knowledge- which I reject as well- and another thing to reject the idea of qualia, the subjective experience that results from a distinctively first person perspective. I don't reject this. I think the first-person perspective is real and that it provides the framework for explaining how we can think of some experiences, like dreams and hallucinations, as reflective of external reality and others, not. That strikes me as what Descartes got entirely and quite revolutionarily right. I think any account of color vision has to preserve all this. Galileo, Locke and Pinker could all agree to say (if they were willing to be flexible about language) that something appears red because it is red. But if they did they would insist that redness itself would not be an appearance or a way of appearing, but a physical/dispositional property of the external world. Similarly, of course, with temperature and any other property that we have sensory modalities for detecting.

      Which raises another interesting question about your view, which is how consilient a framework it is with respect to appearances generally. You focus on vision, but do you apply this way of thinking comfortably to the the perception of heat and sound? When we say that something feels hot because it is hot, the objective sense of "hot" is today uncontroversially associated with mean kinetic energy of the molecules. Heat itself is not an appearance and it does not have the qualitative properties we associate with the subjective sensation. Do you think this thinking is similarly flawed?

      And now (since this is after all, a blog) I will throw caution to the winds and go completely ad hominem on you.. It seems to me that in many ways what you are doing- incredibly well by the way- is trying to provide us a better understanding of how the ancients actually saw the world. I'm sure you and I agree that in a certain sense there is no going back, but I think you are bothered more than I am by the insultingly cartoonish way that we characterize it now that we feel we have surpassed it so completely. That strikes me as totally admirable. But I think it also carries the danger of mistakenly believing they got something right that we no longer appreciate. This strikes me as a priori extremely unlikely. The fact that a theoretical belief arose from something as fundamentally ignorant as the ancient mind is, for me, a reason for thinking that it is probably wrong. We would and should be incredibly surprised to find out that a view they had about the nature of disease, or reproduction or the heavens or the origin of the universe were correct. So I think that's the proper stance to have about their philosophical views as well. Not that they are just obviously mistaken, but that the burden of proof is on anyone who thinks our understanding of the world is diminished by their absence. And I mean no disrespect to ancient people when I say that. And it is in no way to deny that our current conceptual framework is shot through with inconsistencies, I suspect that will always be the case as long as our knowledge is growing

  7. Here’s a list of readings which are the source of the ideas in the blogpost:
    Broackes, Justin, “The Autonomy of Colour,” in Byrne, Alex & Hilbert, David, Readings on Color, Vol. I, MIT Press (1997) 189-225.
    Campbell, John, “A Simple View of Color,” in Byrne and Hilbert (1997), 177-190.
    Ellis, Jonathan, “Colour Irrealism and the Formation of Colour Concepts,” Australasian Journal of Philosophy 83:1, 53-73.
    Gert, Joshua, “What Color Could Not Be: An Argument for Color Primitivism,” Journal of Philosophy 105:3 128-155 (2008).
    Watkins, Michael, “A Posteriori Primitivism,” Philosophical Studies 150: 123-137 (2010).
    The analogy between color and evolution as autonomous realms was my own. Evolution may be a ‘naturalist’ theory (as opposed to ‘supernaturalist’?). But because facts about evolution only weakly supervene on physical facts, acceptance of evolution offers no aid and comfort to the physicalist.