Sunday, April 20, 2014

We may believe the red rose is red

In my History of Early Modern Philosophy, I tell the story of how the cutting-edge metaphysics of 400 years ago became the common sense of today.

Two pieces of common sense:

  1. We may believe that the rose is red, but this belief is strictly and literally false.
  2. Strictly and literally, nothing really is red.
These express the view of colors propounded (in different versions) by Galileo, Descartes, and Locke.

It seems only fair to confront common sense with the metaphysics of today.

Antirealism about ‘secondary qualities’ derives its plausibility from early modern accounts of mental content. We acquire our concept of red from the contents of our perceptual experiences: ‘adventitious ideas’ (Descartes), ‘ideas of sensation’ (Locke). From these arise thoughts of red, as well as the meaning of ‘red.’ Both the sensory and the cognitive content are thus subjective, logically private, and transparent to the subject, who is infallible about them.

What happens when we combine the commonsense view of red with current views on mental content and concept formation? (And add in the arguments of philosophers like Barry Stroud, Jonathan Ellis, and John Campbell.)

We clearly run afoul of the Private Language Argument. A meaning, a concept, must follow a rule. But ‘purely private rule’ is an incoherent notion. As Wittgenstein said: “…(W)hatever is going to seem to me to be right is right. And that only means that here we cannot talk about ‘right’.”

We also get into trouble with Twin Earth-style arguments, from which Putnam concludes, “Meanings ain’t in the head!” They are not subjective, nor are we infallible about them.

It conflicts with externalist and representationalist accounts of mental content according to which we have no conscious access to our ‘ideas.’ Gilbert Harman asks, “Does your experience represent this redness by being itself red at a relevant place, in the way a painting of a ripe tomato might represent the redness of the tomato with some red paint the appropriate place on the canvas?” Harman concludes that there is no ‘mental paint.’ Our perception and thought represent the world as being a certain way.

Consequently, when we see something as red, this seeing must be a connection to the external world. What is its object? Some property of the rose, evidently: a structural property of the surface by which it absorbs certain wavelengths, or the dispositional property to do that. Or a dispositional property to represent an object thus and so. Or a simple, transparent property supervening (somehow) on physical properties.

Surely the natural candidate for being the referent of our concept red is that property.

But then our belief that the rose is red is true, not false.

Some things really are red.

The remaining issue, then, is what kind of property being red is. If it’s a structural property of surfaces, or even a dispositional property based upon its physical structure, then arguably colors are physical.

This would fit with the early modern denial of colors as having no function in scientific explanations. If they’re physical, they would.

But this moves a bit too fast. Colors function in explanations not because of purely physical interactions with the visual systems of animals, but in virtue of how they represent things. Color explains things not in terms of structural properties but in terms of how objects appear.

This would make color a real property of objects, but not a physical one.

Thomas Pyne
Department of Philosophy
Sacramento State


  1. Tom, thanks for this. This is how I see it. There is a range of behavior than can be explained partly by reference to how the world appears to an agent. And it is surely a fact that the world appears redly, loudly, softy, sweetly to said agents. But what the moderns discovered is that these explanations have very limited explanatory value; they are what we now call folk explanations, not scientific ones.

    The distinction between primary and secondary qualities is fundamentally a recognition of the fact that the way the world appears to us requires explanation in terms of properties that are not themselves appearances. It is only folks who say that we believe the rose is red, that the rose appears redly to us because it actually is red. This can be an adequate explanation for some practical purposes (say, to determine whether unusual lighting conditions might account for its appearance) but it can not be an adequate scientific explanation because redness, and sensory appearances generally, are crude, though functional representations, of the external world. To use Metzinger's analogy, appearances are the outputs of a user operating system. They are a highly simplified way of representing the world that permits us to interact with it successfully- if all we are trying to do is use it.

    The truth is that we still do not know what makes the rose appear redly, just like we still do not know how the universe was created. (I understand that lots of folks need it to be otherwise, and that is just too bad.) But what we do know is that appearances don't explain appearances. How the world appears, and why it appears the way it does, will ultimately be explained by references to properties that are not sensory in nature, not themselves appearance, but that part of reality that makes appearance possible. Nobody who holds this view needs to say that appearing isn't real. In fact, it is just an erroneous attribution, because, for the most part, anything we are trying to explain we stipulate to be real. All they are saying is that appearances don't themselves explain anything in the scientific sense.

    Usually disagreements between you and me turn out to be more appearance than reality, and I don't know how much of what I just said you would disagree with. But I think it might conflict with what seems to me to be an assumption you make in the above. You say:

    "Consequently, when we see something as red, this seeing must be a connection to the external world. What is its object?"

    This seems like you are asking in what sense we are attributing redness to the physical world. The answer is that we are attributing it in the perfectly innocent sense characteristic of our folk understanding of it. But we are capable of attributing this property to the physical world when we are behaving as folks do, and at the same time realizing that it is scientifically inaccurate. This is no different than the fact that we continue to speak of the sunrise or the passage of time. It just seems to me that you sometimes speak as if you are requiring us to take these attributions seriously, when what you are really doing (helpfully, in my view) is revealing the ontology of the manifest image, the commitments we would have to accept if we were foolish enough to insist that it s ultimately accurate. The explanatory superiority of the scientific image is, of course, not a matter of informed dispute.

    1. Leaving the existence of appearances as much of a mystery as the origins of the universe seems a terribly high price to pay. It’s also unnecessary.

      It seems indisputable that appearances do, after all, function in scientific explanations. Insects are drawn to flowers not by any structural property of the surfaces but by how those properties make the flowers appear. Dogs are drawn to objects not by some purely chemical transaction between a substance and their olfactory systems, but by how the chemical composition of the substance makes it smell. Substitute a different substance with the same smell and the dog will respond the same. So appearances do play a role in explanations.

      A purely physicalist explanation of the behavior is still possible, of course: you can go disjunctive. Instead of appealing to appearance, correlate the animal’s response, a response which is normally caused by a transaction with physical property ‘F,’ instead with a transaction with disjunctive physical property ‘F or G’. The problem is obvious: such explanations become purely recursive (‘F or G or H or…’) leaving us no closer to understanding why the animal responds as it does, or being able to predict, for any new substance whether it will respond that way or not. Jerry Fodor was candid about the inadequacy of such explanations. (Unless, course, you cheat and sneak the concept of ‘smelling the same’ in through the side door.)

      We must then make some place for non-physical concepts like representational similarity, representational accuracy, etc., in our explanations. But these make appeal to how objects appear.

      So, yes, we must take these secondary quality attributions seriously; they form an ineliminable part of explaining Things That Actually Are So.
      I’m not disputing the explanatory superiority of the Scientific Image, though I am disputing its scope. Disciplined reasoning regarding phenomena, using evidence, observation, and theory construction applying concepts appropriate to the realm in question are not limited to the physical sciences. Indeed I argued that, strictly speaking, neo-darwinian evolution by natural selection is not a physicalist theory. Additionally, in your blogpost asking for examples of moral discovery to put beside scientific discovery, I proposed as an example Adam Smith’s discovery of the existence of a market in labor. That was the discovery of something real, though not physical.

    2. Tom, thanks for the thoughtful response. I think I am inclined to disagree with just about every claim you make in your second paragraph. I don't think we know any of those things to be true at this point. It's quite a leap to say that because dogs will respond to very different chemicals in the same way that there is no purely chemical explanation of their behavior. It just means that we must show that the different chemicals have the capacity to produce the same olfactory reactions, which of course happens.

      I agree with you that we forego understanding by going disjunctive as you say rather than take advantage of our ability to engage in abstraction. And I also think that doing so commits us to the existence of what are commonly called emergent properties, which are substrate independent. chess games, traffic and computer programs are all real aspects of my world, and so, of course, are appearances.

      But the first three concepts involve no epistemic sacrifice, whereas appearing does. Appearances are fundamentally approximations, the brains rules of thumb for representing the deluge of sensory information it receives to itself. We really don't understand yet either how the brain creates appearances or how they function causally. But I think your view, which commits you admirably to specific predictions, is very likely to be shown false. The same appearances will be associated with very different behaviors because they are ultimately only crude representations of the causal determinants of behavior. We already know that appearances lag temporally far behind our physical reactions to stimuli. You have already jumped out of your skin by the time you feel the fear aroused by a snaky looking rope. They simply can not have the role that our folk understanding of mental phenomena assigns to them.

      You may be right that I am exaggerating by placing the mystery of the origins of the universe and the mystery of the function of appearances in the same category, but I'll stand by it. It isn't for lack of interest that our understanding of the former far outstrip our understanding of the latter.

    3. Tom,

      I've been thinking about your remark, "We must then make some place for non-physical concepts like representational similarity, representational accuracy, etc., in our explanations."

      In Aristotelian language, seeing is an actualization, but looking is a process, just as winning a race is an actualization whereas running the race is a process. We can appreciate the difference if we remember that we can be halfway through running the race, but it would sound odd to say we are halfway through winning. Similarly, looking may take some time, but seeing doesn’t. Aristotle makes this kind of point in the Metaphysics, 1048 b, 30-34.

      A dualist in the philosophy of mind is likely to say that neurobiologists have had more success with process than with actualization because, they say, actualization is not physical. That explanation of the difference in success is going too far. Seeing is still physical. It’s just a more difficult problem than looking.

  2. Common sense is what the average person would claim to know. Here are some examples. Sand is not good to eat. You can’t push a car up a hill with a feather. Some roses are red.

    I don’t believe it’s common sense that “nothing really is red.” Many great philosophers believe nothing really is red, but the average person thinks many things are red; they think red is an intrinsic property of an object, not a relationship between them and the object, nor a phenomenon that occurs only in their mind. The average person believes roses are red even when no one is looking at the roses. We should think seriously about the possibility that the average person is correct, while the great philosophers are incorrect.

    Tom, I agree with you that Descartes, Locke, Berkeley, Hume, and Kant are to blame for upsetting our naive realist belief that roses are red while replacing it with the claim that we do not see the real world as it is, but rather see our private sense data or our ideas or impressions or representations or whatever. The great philosophers of old tried to tell us we are cut off from the external world and can know it only indirectly or not at all. Thank goodness we have philosophy instructors today who set things right and point out that merely because scientists can give a causal explanation of some of what is going on starting with the light from the rose and ending with our mind or brain, it’s a mistake to conclude that we aren’t really seeing a red rose. It’s just like the mistake of concluding rocks aren’t solid because scientists have revealed to us that rocks are made of atoms which have lots of empty space within them. Rocks are solid and roses are red, despite the science.

    I can dream of a red rose. Dreams are a form of conscious experience, but during this experience I don’t see a dreamed-red-rose, a rose that exists only in my mind. I don’t see anything at all.

    Yet color phenomena are real. Even illusions and hallucinations are real phenomena that are amendable to scientific study. Nevertheless, scientists know very little about the phenomenal experience of color or about any other conscious experience. What they do know is that we can use our eyes to take in information from the external world and use that information to build a mental image of a red rose and to acquire the propositional belief that the rose is red.

    The scientists know a little more detail about this. They know that normally when we perceive a red rose, the rose has to be reflecting certain wavelengths of light, those light waves have to reach us and be detected by our eyeball, and these eyeball events have to cause a series of neuronal processes that eventually result, in some mysterious way, in our having a conscious experience of seeing a red rose. The scientists definitely know that none of our neurons are red. They also know that brains cause minds, and that Cartesian dualism is dead, but scientists do not know very much else. There are many Nobel Prizes yet to be won in neuroscience.