a) An Outline of the Argument

Idealism and its contemporary version, phenomenalism, has two main sources. I have already identified one of these, namely, the distinction between primary and secondary qualities. This division, I pointed out, rests on the argument from physics. But there is also a second source which I have not as yet mentioned and which deserves a chapter of its own. This is the argument from hallucination. The present chapter deals with the argument from physics; the next one, with the argument from hallucination. Finally, there is the principle of immanence which threatens the theory of knowledge with skepticism. It is the basic premise of the argument from the relativity of sensing. In the third chapter of this section, we shall discuss this argument. But an analysis of the principle of immanence will have to wait until we turn to the second main part of this book, namely, to our knowledge of the mind.

The three important arguments just mentioned combine and reinforce each other in strange ways, so as to render succor at times to idealism, at times to skepticism, and at times to both. And whenever one of them seems too weak to destroy realism, another one will jump into the breach.

Modern physics, so it has seemed to many philosophers from Galileo through Descartes to Kant, Sellars, and Armstrong, has chased such perceptual properties as colors, shapes, odors, etc. from the “external” world. This seems to leave us only with two possibilities: Either these properties merely exist in the mind, or else they exist nowhere, that is, they do not exist at all. Berkeley, we saw, opts for the first possibility; Brentano, I have maintained, accepts the second. But to hold that there are no colors, neither in the external world nor in the mind, is, not to present a philosophical position that can be discussed, but to violate common sense to such a degree that argument is made impossible. As Moore used to emphasize, what premises could one possibly hope to share with a philosopher who holds that there are no colors, no shapes, no odors? Philosophical views that deny the common ground from which all disputes must start are not merely false, but are absurd. The view that there are no colors, whether held by Brentano or by Goodman, is absurd (see N. Goodman, “Predicates without Properties”).

Berkeley’s view, on the other hand, is not absurd in this sense, but merely false. Color is indeed in the mind, that is, it is indeed a property of mental sensations. But Berkeley denies, and this is where he goes wrong, that there are apples in the nonmental world which are green. He denies that the color has an existence “outside of the mind.” We need not be misled by the spatial metaphor. Berkeley denies that there are apples as mind-independent perceptual objects. And it is precisely this proposition which a stout-hearted realist must defend. How, then, does physics undermine this proposition?

Things have not changed very much philosophically from the days of Galileo to the days of Gell-Man. Atoms, one used to say, have figure, number, size, and they are in motion or at rest. But they have no color, no odor, etc. Elementary particles (or else quarks, psy-functions, or what-have-you) we now believe, have a number of esoteric qualities like mass, spin, electric charge, etc. But they are still not colored. What these ultimate building blocks of the universe are, and what properties they have, it must be conceded at once, is a matter for the physicist to discover. Philosophical speculation is out of place. Thus I shall not, even for a moment, lock horns with the physicist by claiming that elementary particles are colored. No, elementary particles do have the properties that the physicists have discovered; and color is not among them. But we also insist on, and shall not move one inch away from, the common-sense truth that the perceptual objects around us are colored. Berkeley’s apple is truly green. And we shall rebuff even the most sophisticated attempt by the most famous physicist to convince us otherwise. Apples are green, and no theory proves that they are not. The argument from physics, however, tries. It starts from the true premise that elementary particles are not colored and concludes mistakenly that apples cannot be colored. Obviously, there must be hidden premises. What are they?

Well, since elementary particles are not colored, and since apples consist of elementary particles, apples cannot be colored either. We must assume, in order to keep the argument going, that perceptual objects ultimately consist of elementary particles. This premise is as unobjectionable as the first premise. But it obviously does not suffice to get to the conclusion. Some kind of principle is needed to the effect that things have only those properties which their constituents (parts) have. Let us call a principle of this sort “a principle of reduction.” The idea is that the properties of complex things (of wholes, of structures) are somehow “reduced” to the properties of their constituents.

b) The Principle of Reduction

Sellars, who has devoted much time to this topic, formulates a principle of reduction in this manner: “If an object is in a strict sense a system of objects, then every property of the object must consist in the fact that its constituents have such and such qualities and stand in such and such relations or, roughly, every property of a system of objects consists of properties of, and relations between, its constituents” (W. Sellars, Science, Perception and Reality, Routledge and Kegan Paul, London, p. 27. See also his later “The Structure of Knowledge,” pp. 298-300).

Notice that Sellars speaks of an object which is “in a strict sense” a system of objects. He implies that the principle is only true for certain systems of objects. A certain structure, consisting of pieces of wood, for example, is a ladder, even though none of its parts is a ladder, as he points out. A ladder, a car, or a typewriter is in the strict sense a system of objects, and that is why the property of being a ladder is reducible. Being a ladder, he says, “is [being] made up of being cylindrical (the rungs), rectangular (the frame), wooden, etc.” (Sellars, 1963, p. 26). What Sellars is driving at, I think, is that being a ladder is a complex property. He seems to be reasoning that the ladder can have the complex property of being a ladder, even though none of its parts is a ladder, because the complex property of being a ladder is not really anything in addition to the properties of which it consists. Notice that Sellars slips when he lists the properties of which “being a ladder” is made up of: He mentions being cylindrical and rectangular; but these are properties, not of the ladder, but of parts of the ladder. What we want are properties of which the complex property of being a ladder consists, and these must be properties of the ladder. Let us therefore change the example and talk about Berkeley’s apple. Oscar, we shall assume, is green and round. Let us introduce the expression ‘ground’ as an abbreviation for the longer ‘green and round’. We can then say that the sentence ‘Oscar is ground’ means nothing more nor less than that Oscar is green and round. The “property” ground, one can claim, is defined in terms of the two properties green and round. I put ‘property’ in quotes because, strictly speaking, there is no such property. There is only the word ‘ground’. What there is in the way of property are the two properties green and round. Similarly for the word ‘ladder’. I interpret Sellars to mean that there really is no such property as that of being a ladder; there is only the word ‘ladder’. What there are, in reality, are certain wooden parts, their properties, and the relations among them. To say that this is a ladder is then merely a short way of saying something quite complicated about those wooden parts, their properties, and their relations.

What about the property of being olive green? If we apply Sellars’s principle of reduction to this property, we come to the conclusion that there really is no such color shade “out there.” What there is instead are the various properties of and relations among the ultimate constituents of olive green perceptual objects. There are no colors “out there” because colored objects are in a strict sense systems of elementary particles, and such systems have no properties of their own. Color words are mere abbreviations for complicated expressions describing properties of and relations among elementary particles.

How plausible is this view? Not very. It is obvious to me that the word for a color shade is not an abbreviation of this sort; it is not an abbreviation in the sense in which ‘ground’ is an abbreviation for ‘green and round’. How, otherwise, could Caesar have asked for green olives in his diet? No, ‘olive green’ is an expression for a certain color shade and not an expression that describes the properties of and relations among elementary particles. Sellars agrees with this conclusion. But he arrives at a different lesson. A color shade, he argues, unlike the property of being a ladder, cannot be a complex property: “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.” (Sellars, 1963, p. 26). Since it is not a complex property, the pink ice cube cannot be in a strict sense a system of elementary particles. Thus there is a clash between what Sellars calls the “scientific image,” according to which “out there” are elementary particles without colors, and the “manifest image,” according to which “out there” are pink ice cubes. I see no clash (cf. my “Perceptual Objects, Elementary Particles, and Emergent Properties”). The ice cube of the manifest image is identical with the complicated structure of elementary particles of the scientific image. And the ice cube is truly and literally pink, even though its ultimate constituents have no color. It is simply not true, as Sellars assumes, that complicated structures of elementary particles can have no properties of their own.

What the argument from physics shows, from our perspective, is that the principle of reduction is false: Structures (wholes) have many properties which their parts do not have, and which cannot be reduced, in the precise way described earlier, to the properties and relations of their constituents.

c) Laws and the Complaisant “Identification”
of Properties

There is no such property as ground; there are only the two properties of green and round. In Sellars’s example, there is no such property as that of being a ladder; there are only the properties of and relations among the parts of the ladder. The word “ground” is a mere linguistic convenience, according to Sellars’s view; and so is the word ‘ladder’. The case is clear for ‘ground’. I do not think that it is obvious for ‘ladder’, but I shall not argue. As far as color is concerned, the view is obviously false. There are colors and not just color words. But we must be very careful at this point and must keep apart two quite different notions of reduction or definition. There is, on the one hand, the notion of reduction just discussed, according to which a property is reduced to other properties, if it can be shown that only these other properties exist. But there is, on the other hand, also the notion that a property is reduced to other properties, if its occurrence can be deduced from these other properties and certain laws. A property that cannot be reduced in this sense of the term is often called an “emergent property.” The following quotation gives a precise form to this notion of reduction: “the occurrence of a characteristic W in an object w is emergent relative to a theory T, a part relation Pt, and a class G of attributes, if that occurrence cannot be deduced by means of T from a characterization of the Pt-parts of w with respect to all the attributes in G” (C. G. Hempel and P. Oppenheim, “Studies in the Logic of Explanation”, p. 336).

According to this criterion, the color of the ice cube as well as the color of Berkeley’s apple are not emergent properties, for their occurrence can be deduced from certain laws of physics. Roughly, a surface has a certain color if and only if its atomic structure is in a certain state. It should be perfectly obvious, but apparently is not, that to claim that colors are or are not emergent is not the same as to claim that colors exist or do not exist. What sense could it possibly make to assert in one and the same breath that the occurrence of property P is lawfully connected with the occurrence of properties Q and R, but that there really is no such property as P? The lawful connection between properties presupposes the existence of the properties so connected.

Yet, the discovery of the laws which connect the colors of surfaces with the states of the atoms which form the surfaces has led many scientists and quite a few philosophers to say such things as that “out there, in the physical world, there are no colors at all.” But is it not perfectly obvious that the inference from

(1) A surface has color C if and only if its atomic structure is in state S to (2) C does not exist; only the state of the atomic structure does, is invalid? This version of the argument from physics, too, must somehow fill the gap between (1) and (2) with plausible premises. And since I do not believe that this can be done, I have maintained that the argument is not sound.

It is customary among many contemporary philosophers to say that what the physicist has discovered is that the color of the surface, C, is identical with a certain physical property, P, of the (constituents of) the surface. But the physicist has discovered no such thing. What he found out is, rather, that the surface has the property C if and only if it has P. Since this issue has played such an important role in the philosophy of science—it is at the heart of the dispute over materialism—I shall pedantically ask, firstly: Can you distinguish between a state of affairs of the form:

Property P is the same as property Q

and a state of affairs of the form:

Property P occurs if and only if property Q occurs?

For example, if every green thing in the universe were round, and conversely, would it not be the case that whenever something is green it is round, and conversely, without it being the case that the color is the same as the shape? If you profess not to be able to distinguish between these two states of affairs, then I must confess that I do not know how to proceed. But if you grant the distinction, then I assert, secondly, that what the scientist discovered is a fact of the second kind, an equivalence. He may have looked at a certain surface, determined its color with the naked eye, and then measured the physical state of the surface; to the color pink, for example, he may have discovered in this fashion, corresponds a certain state S₁, while to the color olive green, there corresponds a different state S₂. What the physicist definitely did not do was to compare the properties of the color and of the state in order to determine whether or not they are the same, as one would have to if one were to ask whether the color is the same as the state. The physicist did not try to determine whether or not the state of the surface can be more or less saturated like a corresponding color. No physicist, I believe, unless he has a philosophical ax to grind, seriously contemplates the possibility that the states he studies are the same as the colors he sees. His situation is quite different from the situation of an astronomer who tries to find out whether a certain heavenly body, observed at a certain place at a certain time, is the very same as a body observed at a different place at a different time. This astronomer would ask whether or not the “two” bodies have the same properties, including the same spatio-temporal ones.

I conclude that the physicist discovered that there is a certain lawful connection between the occurrence of colors, on the one hand, and the occurrence of certain atomic states, on the other. Now, it has been claimed, for example by Armstrong, that the discovery of such a law is a prima facie case both for the identity of colors and states and for their equivalence. Both possibilities are open, so to speak. Not really: The second alternative is not a mere possibility, but a fact. We know that colors and states are equivalent. We also know that this equivalence does not preclude the possibility of their identity. But is there any reason to suppose that the identity is more than a logical possibility? Is there any reason at all? It seems to me that, quite to the contrary, there is conclusive evidence that colors are not states, for colors have many properties which states do not have, and conversely. Surely, it is disingenuous at this point to reply: “But if colors are states, then they do have all their properties in common.” It is as disingenuous as is the reply, when the identity of elephants and mice is at stake, that if elephants are mice, then an appetite for peanuts simply is an appetite for cheese.

The “physicalist,” as I shall call my opponent, usually “identifies” the colors with the states. He is hell-bent on getting rid of colors, by hook or by crook, that is, by argument or by decree. What is his motive? Before I try to answer this question, one thing needs to be emphasized. Physicalists often claim to have “theoretical scientific” objections to positions like mine. They presume to speak in the name of science against my philosophical, allegedly antiscientific, view. Nothing could be further from the truth. I warmly embrace the scientific spirit as well as the scientific method. I see no conflict between science and my way of doing philosophy. Nor do I feel any animosity toward the successes of science, an animosity that has sometimes been expressed by an “identification” of science with technology. I therefore know that my opponent’s view rests neither on superior scientific knowledge nor on a deeper admiration for science. It seems to rest, rather, on a metaphysical urge, an urge that is cloaked in the mantle of science. It rests on the urge to be rid of mind and all that goes with it; for what goes with it are colors and all the other “secondary qualities.”

Is it not in perfect harmony with both common sense and physics to say, as I do, that surfaces have colors and that they have these colors because they consist of atoms which are in certain states? I have been told by Armstrong that philosophers like J. J. C. Smart object to my view by claiming how “badly such laws sit with the fundamental laws of physics, which seem likely to be the laws of natural science” (letter from Armstrong). “The laws Grossmann wants to postulate,” Armstrong writes, “would be a sore thumb on the corpus of science, nomological danglers as Herbert Feigl put it. The natural solution for the scientifically minded . . . is that the equivalence marks not a law, but an identity. The perceived color is the atomic structure” (letter from Armstrong). But I do not postulate any laws. I merely accept, in the scientific spirit, the laws which the scientist discovers. It is the physicist, not the philosopher or the common man, who discovered these fascinating laws. And it is philosophers like Smart, Feigl, and Armstrong who misuse the physicist’s discovery in order to justify a strictly metaphysical move, namely, the “identification” of colors with states. It cannot be repeated often enough that what motivates these and other philosophers is not a defense of science against metaphysical speculations, but a metaphysical conviction of their own, a conviction so deep that science cannot reach it.

What about the objection that the equivalence sits badly with the fundamental laws of physics; that it is a nomological dangler? Call it what you wish, dislike it as much as you want to, the fact remains that laws like this are scientific achievements of the highest order. They are shining jewels in the crown of science. For they explain to us the world around us, the world in which we live every minute of every hour of every day, with all its wonders and puzzles, all its delights and horrors. They explain to us why the perceptual world is what it is in terms of its underlying structure. Of course, these laws are not the laws of physics. They are not laws about elementary particles. But it would be an argument with a very tight circle if one were to reason that laws other than the laws of physics are unacceptable, and that they are unacceptable because they are not the laws of physics. Or have we here another metaphysical belief of the physicalist, namely, the belief that there can be no laws other than the laws of physics?

Finally, would it not be simpler just to go ahead and “identify” the colors with the states? Well, the world is as complicated as it is, and to make it simpler than it is, is to misconceive it. Simplicity alone counts for nothing. Otherwise, we may as well identify everything with everything else and be content with contemplating the absolute. The question is not one of simplicity, but one of truth: Are colors the same as atomic states? I argued that the answer is negative. As I see the dialectic, this is not an open question to be decided by considerations of simplicity. The fact is that colors are not atomic states. And this fact points at another facet of the dialectic. If the “scientific minded” physicalist were really scientific minded, he would try to show, in the true scientific spirit, that colors are the same as atomic states in the only way in which we can ever show that a thing A is the same as a thing B, namely, by showing that colors have the same properties and stand in the same relations as states. At the very least, he would ask his physicist friends to do the job for him. But of course he does neither. Instead of rolling up his sleeves and tackling the task of proving that his hunch is correct, that colors really are atomic states, he keeps on talking about “identifying” the two. Astronomers, by contrast, truly scientific minded as they are, did not just talk about “identifying” the morning star with the evening star. They found out that the two heavenly bodies are the same by discovering that the morning star has the same properties and spatio-temporal relations as the evening star.

Berkeley’s apple is green. Nothing in modern physics proves otherwise. That it cannot be green follows neither from the fact that elementary particles are colorless nor from the fact that its surface is in a certain atomic state. It does not follow from the former, because the principle of reduction is false. It does not follow from the latter, because colors are not atomic states. Does it not make sense to adopt our view which so effortlessly combines common sense with the latest discoveries of physics? Must one go on to insist that there is a clash between common sense and physics, not for scientific reasons, but from metaphysical motives?