a) The Main Argument

Berkeley shows ingeniously how to avoid the skepticism within the heart of Cartesianism. But he fails miserably because he accomplishes this miracle only by succumbing to the idealism that lies at its soul. His strategy is very simple. Firstly, he denies the existence of Cartesian ideas, that is, of representational mental things. This is the point of the Introduction of the Treatise. His attack on Cartesian ideas appears there in the form of an attack on abstract (general) ideas. Secondly, he renames Cartesian sensations (and images) “ideas.” And, thirdly, he identifies perceptual objects with bundles (collections) of Berkeleyan ideas, that is, with bundles of Cartesian sensations. The problem of how the mind can know external objects disappears! In accordance with the principle of immanence, the mind knows its own (bundles of) sensations; and since these sensations are the perceptual objects of the external world, the mind automatically knows the external world. We might say that Berkeley, like the Tradition, rejects representationalism. According to the Tradition, what is in the mind, the essence of the thing known, does not represent anything, but is the very thing known or, rather, is its essence. It is true that what is in the mind is not identical with the substance without the mind, but only with its essence or form. But everything knowable about the external substance is truly in the mind. Berkeley, of course, has no material substances. Oscar, the apple, is for him merely a bundle of properties. Thus there is no distinction between matter and essence (form). All of Oscar, not just his essence, can therefore be in the mind. But what is in this sense “in” the mind, is mental. Oscar is therefore a mental thing. This is Berkeley’s idealism.

A while ago, I said that either shape and color are both properties of the wax, or else neither is, and they are both mere sensations in the mind. Of course, I believe that the wax has a color and a shape. Berkeley, on the other hand, accepts the second alternative. He acknowledges joyfully what the Cartesian tries to hide, namely, that idealism is inevitable, once colors are relegated to the status of secondary qualities. Berkeley’s argument is as simple as it is powerful:

(1) A material object is nothing but a bundle of sensible properties.

(2) Sensible properties are nothing but ideas (read: sensations).

Therefore:

(3) A material object is nothing but a bundle of ideas.

It is the power of this argument that made it all but impossible for modern philosophers to mount a defense of realism. There even exists what may be called a philosophical “superstition” to the effect that Berkeley is irrefutable. I call this a superstition because it seems to me to be as obvious as anything can ever be in philosophy that Berkeley’s argument is proven to be unsound by the fact that the conclusion is false. But we cannot be content with merely pointing out that apples do not exist in minds. We must show precisely where and how Berkeley’s argument goes wrong.

Two simple steps lead from Descartes’s professed realism to Berkeley’s defiant idealism. In the first place, Berkeley attacks the notion of material substance, using ammunition provided by Descartes himself as well as by Locke and others. And, in the second place, he argues that whatever speaks for the contention that colors are mere sensations in minds, also shows that shapes are such sensations. In short, he argues that all sensible properties are mere sensations.

b) The Attack on Material Substance

I mentioned earlier that the Tradition has a problem about how individual things as individuals are known (perceived). What is in the mind, we saw, is the form or essence of Oscar. But this essence is the same for all apples. What distinguishes Oscar from other apples, therefore, must be his matter. But this matter cannot be known, for it does not enter into the mind of the knower. Hence, Oscar, this particular apple, cannot be known. Platonists, by contrast, have precisely the opposite problem: They have a difficult time explaining how the mind knows, not the particular, but the universal, the form. According to this tradition, what comes through the senses is the particular, what belongs to the world of becoming. What is sensible, to say it concisely, is what is in space and/or time. How then do we know the forms, denizens of the world of being? The Platonist usually invents a special mental faculty in answer to this question. The mind has two eyes, he claims, the eye of the senses and the eye of understanding (contemplation). Thus while the Aristotelian has to explain how we know particulars, the Platonist has to explain how we know universals. But let us return to Berkeley or, rather, to Descartes.

Descartes’s solution to the Aristotelian problem, we can now see, is “Platonistic” in kind: he simply postulates a special faculty of the mind, the understanding, which is supposed to acquaint us directly with material substances. Sensible properties, he holds, are known through the senses; the material substances in which they inhere, on the other hand, are known by the understanding. But this solution will not do. Descartes is mistaken: there is no such faculty as what he calls “understanding.” If there were, then we should be able to recognize the piece of wax as the same, even if we have not continuously observed it. But we are not. Assume that I show you a white billiard ball, a while later show you another white billiard ball, and then ask you whether the one I am now showing you is the same as the one I showed you earlier. Obviously, you can only guess. But if Descartes were right, and if there existed a special faculty which “grasps” the material substance directly, as it were, then you would be able to tell whether or not the one billiard ball is the other, irrespective of how similar they may look. It is simply a fact that we can only recognize perceptual objects as the same by means of their properties and relations to other things.

But one may also draw quite a different conclusion from Descartes’s wax example. One may argue that it shows that there can be no such thing as a material substance. If there were such a thing, one could reason, then one would have to be able to recognize it directly, as Descartes claims we can. But we have just seen that we do not, as a matter of brute fact, recognize perceptual objects directly, but only through their properties and relations. Hence there are no material substances. In this manner, the wax example contains a powerful argument against the existence of material substances. Berkeley, however, attacks the notion of substance from a different direction. He follows in the footsteps of Locke. He simply asserts that we are never acquainted with material substances; that we perceive nothing but sensible properties (see, for example, the beginning of the first dialogue of G. Berkeley’s Three Dialogues between Hylas and Philonous). Nor is he the only philosopher who holds this strange view. Locke, we remember, speaks of material substances as things we do not know. Reid claims that it is a metaphysical principle that “the qualities which we perceive by our senses must have a subject, which we call a body, . . .” (Reid, Essays on the Intellectual Powers of Man, p. 650). And Hume asserts that there is no impression of a material substance; and since there is no impression of it, there can be no idea of it either. It seems to be a dogma of the empiricists that the mind does not perceive individuals but only properties. And yet what could be plainer than that we see apples as well as their colors, that we perceive individual things as well as their properties and relations? Both the Tradition and empiricism fail because they deny this fact.

In regard to the question of whether or not material substances (particular things) are knowable, we must make several distinctions. It must be admitted that individual things cannot be directly recognized, but only through their properties and relations. We can recognize a white billiard ball as the same as one we have seen earlier because it has a small red spot on one side. Or we can be sure that it is the same because we have observed, uninterruptedly, its spatio-temporal path. But we cannot directly recognize it to be the same. In this sense of ‘knowing’, individual things cannot be known. But this is not the important sense of the word. There are two such important senses. According to one, to know an individual thing is to know things about it. It is to know what properties it has and in what relations it stands to other things. We know an apple, in this sense, when we know what shape and color it has, how much it weighs, what kind of apple it is, where it comes from, who sold it to us, etc., etc. It seems quite obvious to me that we know individual things in this fashion, for we can know all of the things I just mentioned about Oscar.

It has been said that to know things about an individual means to know things about it when it is considered in isolation, stripped of its properties and relations, as a mere, naked, particular thing. And since we cannot know what an individual is when we consider it bare of all of its properties and relations, one has argued, we cannot know individual things. But is it not silly to ask what properties an individual has when considered devoid of all properties? If we adopt this notion of what it means to be known, then it follows that nothing whatsoever can be known. Properties, for that matter, can then not be known either; for what is a property when considered devoid of all properties and relations? No, we know individual things just as well as we know other kinds of things, for we know their properties and relations. We even know some of their categorial properties which distinguish them from all other kinds of existents. An individual thing, for example, stands in temporal relations to other individual things; no other category is temporal in this sense. This is one of the fundamental theses of my ontology.

In another important meaning of the word, we know an individual thing when we are acquainted with it. In this sense, I know the Eiffel Tower, but do not know the Kremlin; I know the color shade olive green, but a blind person does not. I believe that it is in this sense of acquaintance that Berkeley denies knowledge of anything other than sensible properties. At the beginning of the first dialogue of the Three Dialogues, Hylas concedes without the slightest protest what is quite obviously untrue, namely:

that sensible things are those only which are immediately perceived by sense. You will further inform me whether we immediately perceive by sight anything besides light and colors and figures; or by hearing, anything but sounds; by the palate, anything besides tastes; by the smell, anything besides odors; or by touch, more than tangible qualities.

(Berkeley, 1965)

We would want to protest that we immediately perceive by sight such things as apples, and that apples, though they have colors, are not colors, and though they have shapes, are not shapes. We want to object that we can smell roses, taste lobsters, and touch lips, and that none of these things is a property.

It is easy to turn the table on Berkeley. Since you do not deny that there are apples, we may ask, how do you think we are acquainted with them, if not by sight, touch, etc.? He may reply that we are acquainted with them because they are nothing but bundles of sensible properties, and we are, admittedly, acquainted with such properties. But this answer will not do. What is a bundle of sensible properties? There are only two plausible answers: a bundle is either itself a sensible property or something other than a sensible property. But it cannot be the former for ontological reasons; a property must be a property of something, and an apple is not “of something.” (Berkeley, I take it, does not believe that everything is a property of the absolute!) And it cannot be something other than a sensible property, for then we could not be acquainted with bundles of sensible properties, according to Berkeley’s own view. To put it bluntly, Berkeley contradicts himself when he holds both that nothing comes through the senses but sensible properties and also that we are acquainted in perception with such individual things as apples.

Nothing, but nothing, can shake our conviction that we perceive individual things as well as their properties and relations. Nor can we doubt even for a moment that we know what kinds of properties and relations individual things have. On this matter, as at many other crucial junctures of our inquiry, common sense rules supreme. The Aristotelians and the modern empiricists are both wrong. But the Platonists do not do much better, as we shall see later. They go wrong by postulating a mysterious faculty, whether it be called “contemplation” or “eidetic intuition,” which is said to acquaint us with universals. Here, too, we take our stand with common sense. We see not only individual apples, denizens of the world of becoming, but also their properties and relations, citizens of the realm of being. Perception, “the senses,” acquaints us with individuals as well as their properties and relations. How can perception accomplish this miraculous feat? By being propositional! Both Aristotelianism and Platonism in regard to perception can be overcome if we hold fast to the fundamental insight that what we perceive are such states of affairs as that this apple, Oscar, is green. In perceiving that Oscar is green, we perceive both the individual apple and its universal color. This is the most important thesis of our theory of knowledge: Perception is propositional.

c) The Identification of Sensible Properties
with Ideas”

I believe that the first premise of Berkeley’s main argument is false. But the argument would still be devastating if the second premise alone, that a material object is nothing but a bundle of sensations, were true. It is this second thesis, to the effect that sensible properties are nothing but sensations in minds, that opens the door to idealism. A mind-independent individual thing whose properties are not the sensible properties we know, is no more an apple than Kant’s Ding an sich is. A realism worth defending cannot be the pitiful view that there are nonmental things “we know not what.” What motivates this kind of last-ditch realism, one suspects, is nothing more than a fit of defiance. The idealist has already won the argument, but one last gesture is left: prove to me, one challenges, that there really is nothing at all out there. But realism does not consist in this futile gesture. Realism holds that there are dogs and apples, clouds and mountains, elements and electrons; that these things have an existence distinct from being perceived; and that we know by means of perception and inference from perception what some of their properties and relations are.

Sensible properties, contrary to what Berkeley claims, are properties of perceptual objects; they are not sensations in minds. The apple before me is green; and it would still be green if nobody looked at it or had ever looked at it. If there were no minds, sunsets would be red, clouds would be white, and the sky would be blue. What is “out there” is not just “extension,” but color as well. Berkeley, of course, accepts the Cartesian belief that color is a mere sensation. But he also insists that the so-called primary qualities are not any different from color, as far as their mental status is concerned. He argues that whatever speaks for the mental nature of color speaks for the mental nature of primary qualities as well. In order to defend realism, we must resist the very first step of the dialectic. Berkeley has got it right: What holds for color, holds for shape as well. But he is wrong about color: It is not a mere sensation in a mind. What we must refute is what I shall call “the argument from physics.”

Berkeley also alludes to another argument (see, for example, par. 14 and 15 of his Treatise Concerning the Principles of Human Knowledge). I shall call it “the argument from the relativity of sensing.” Let me remind you of Locke’s version of it, involving three buckets of water, each one having a different temperature. However, Berkeley very astutely remarks that this argument does not prove that warmth is only a sensation in a mind, but rather that we cannot know the true degree of warmth of the water or the true color of apples. The argument from the relativity of sensing is really an argument, not for idealism, but for skepticism. It is a version of the skeptical argument which I outlined earlier. In this particular form, the question is not how we know that certain objects correspond to certain ideas, but how we know that certain properties correspond to certain sensations. This version naturally arises from the original argument when one identifies, in the Berkeleyan vein, ideas with sensations. Properly analyzed, as I said, the argument is not an argument for idealism, but rather an argument for skepticism.

Berkeley avoids skepticism in regard to perceptual objects by paying the price of idealism. But skepticism is a hydra with many heads. Having cut off one of these heads, Berkeley is immediately confronted with another. According to Berkeley, there are other minds, and the question immediately arises of how these minds are known. It is evident that other minds are not in our minds as apples presumably are. The principle of immanence once again steps forward and announces the imminent arrival of skepticism! Berkeley can only close his eyes and refuse to acknowledge defeat. Other minds, he says, are known not by way of ideas (sensations), but by way of notions. These notions in our minds are not identical with the things (minds) of which they are notions. Rather, they represent those things. But this kind of representationalism, as we have seen, must fall before the skeptical argument. We can now appreciate the full extent of Berkeley’s failure: skepticism, even at the price of idealism, has merely been postponed. In order to escape from it for a second time, Berkeley would have to pay the ultimate price: he would have to embrace solipsism.

I mentioned earlier, in connection with Descartes’s classification of color as a sensation, that there is a problem of precisely how sensations are in minds. If a sensation is an accidental modification of a mind, and if olive green is a sensation, then it follows that the mind of someone with an olive green sensation must be olive green. But how could a Cartesian mind, this extensionless substance, possibly be colored? On the other hand, if sensations are not accidental modifications of minds, in what specific relation do they stand to minds? This problem becomes particularly urgent in Berkeley’s philosophy, for according to this philosophy, not just colors but whole apples are supposed to be in minds. Berkeley clearly sees that there exists a problem. Sensations (his ideas) cannot be properties of minds. However, he does not solve the problem. He merely declares (in par. 49) that color, extension, etc., are in the mind “not by way of mode or attribute, but only by way of idea.” Of course, this does not answer the question of how an idea is in the mind.

But, as I just said, not only colors but whole apples are supposed to be in Berkeleyan minds. And this complication raises a host of further questions. The apple in Berkeley’s mind is said to be a bundle of ideas (sensations). What, precisely, is a bundle? To what category do bundles (collections) belong? If we had a plausible answer to this question, we could perhaps figure out how such bundles are supposed to be in minds. There is one plausible answer, as I pointed out earlier: A bundle is a complex property, consisting of the properties which Berkeley identifies with sensations. But if the bundle that is Oscar is a complex property, what could it possibly be a property of, except the mind in which it exists? It will not do to say that it is not a property of anything, for a property which is not a property of anything is an ontological absurdity. But if a bundle is not a (complex) property, what else could it be? It cannot be a substance, for, according to Berkeley, there are no material substances, and it is not a mind. The bundle that is Oscar is thus neither a substance nor a modification of a substance. It must therefore belong to a category unknown to the Aristotelian (and Cartesian) tradition. And herein lies the importance of Berkeley’s philosophy: Unlike Descartes’s, it does not merely try to improve on the Tradition within the Traditional ontology, but completely breaks with that ontology. After Berkeley, the ontology of substance and modification of substance is supplanted by an ontology of whole and part. (For a contemporary version of the latter, see Parts and Moments: Studies in Logic and Formal Ontology, B. Smith, ed.)

d) Sensations Distinguished from Sensible
Properties

The question of how sensations are in minds depends on a prior question: To what category do sensations belong? Cartesians, as well as many other philosophers, seem to have taken for granted that sensations are properties. A certain shade of color, olive green, is said to be a sensation. But this color is obviously a property of things. Therefore, the sensation must be a property. And, indeed, what else but a property could a sensation be, given the framework of a substance-modification ontology? It could hardly be a mental or a material substance. Nor does Berkeley’s bundle ontology open up new possibilities. But sensations are not properties. To believe that they are is one of the most fatal mistakes of modern philosophy; a mistake that has pervaded most theories of knowledge from Descartes to G. E. Moore. We cannot possibly hope to escape from the twin scourges of skepticism and idealism until we have exposed this mistake. And now, at the beginning of our inquiries, is as good a time as any to distinguish sharply between sensible properties on the one hand and sensations on the other.

If the visual sensation (sense impression) which we under normal circumstances experience when we look at a green apple were identical with the color of the apple, then the sensation would have to be a property. The so-called sensation green would be (a) a property, insofar as the color of the apple is a property, and it would also be (b) something mental, insofar as a sensation is supposed to be something mental. It would follow, please note, that the apple has a mental thing as a property. This conclusion should suffice, in my opinion, to discredit the identification of the color with the sensation. But be that as it may, I hold that the sensation is not a property at all, neither of the apple nor of anything else, but is an individual thing. This sensation, like all sensations, has a certain temporal location. It is also spatial, unlike some but not all other sensations. But in my ontology, these are characteristics of individual things: What is in time and/or space is an individual thing. When I look at the apple, I experience, among other things, a certain individual sensation. The apple causes me to have this visual sensation. How does the color green (a certain shade of green) enter into the picture? Under normal circumstances, when I look at a green apple, my visual sensation is also green. What I experience is quite literally a green sensation. The sensation has this color in the same sense in which the apple has it. By the way, the sensation is also round, just like the apple. Some sensations have shapes, just like apples. Under normal circumstances, therefore, when we look at a green apple, both the apple and the sensation are green. Under these circumstances, the apple and the sensation literally share a common property. By a “green sensation” or a “sensation of green” we may therefore mean two quite different things. We may mean, firstly, a sensation which is caused by a green object. Or we may mean, secondly, a sensation which is green. That these are not the same thing can be seen from the fact that a sensation caused by something green need not be green. If you wear tinted glasses, a green apple may cause you to have a grey sensation. Nor is it impossible for something which is not green to cause you to experience a green sensation. If you wear green glasses, a white wall causes you to experience a green sensation.

One way of summing up my view is to say that there is no such thing as a “sensation green.” There is such a thing as a green sensation (that is, a sensation which is green) and there is also such a thing as a sensation of green (that is, a sensation caused by something green). If we make these important distinctions, namely, the distinction between a sensation and its properties and also the distinction between the properties of the sensation and the properties of the object which causes the sensation, many of the traditional confusions can be avoided.

Under normal conditions, when you see a green apple, in what sense is the color green in the mind? This color shade is literally in the mind because it is a property of something mental, namely, of a sensation. But this sensation is in the mind in quite a different way, namely, by being experienced by a mind. Thus a color is in the mind in the sense that it is the property of something, a sensation, which is experienced by a mind. We shall have to explain this matter in greater detail later, but it should already be clear what our answer is to the question of whether or not colors are mental. This question, we can now see, misses the mark. In our ontology, it makes little sense. The apple is nonmental; it is a perceptual object. The sensation, on the other hand, is a mental individual. Some individuals are mental, others are not. But colors are not individual things; they are properties of individual things. As such, they are neither mental nor nonmental. Or, if you prefer, they are both mental and nonmental. It does not matter what we say as long as we keep in mind that they are not individual things, and that they can be properties of both mental and nonmental things. What holds for color, holds for other sensible properties as well. Roundness, for example, is a property of both the apple which you see and the visual sensation which you experience. Does this mean that the mind is green and round when you see something green and round under normal conditions? It all depends on what you mean by “the mind.” Since the mind is not identical with any one of its sensations, it does not follow from the fact that any one of its sensations can be green, that it can be green.

The color of Oscar is not identical with the visual sensation which we experience when we look at Oscar. The color and the sensation do not even belong to the same category: the first is a property, the second is an individual thing. Yet, philosophers, as I have already said, are prone to identify the two. Reid, for example, tells us that when smelling a rose, the agreeable odor which he experiences is a sensation. But he is not only experiencing this sensation; he is also perceiving an external object:

. . . and the object of my perception in this case, is that quality in the rose which I discern by the sense of smell. Observing that the agreeable sensation is raised when the rose is near, and ceases when it is removed, I am led, by my nature, to conclude some quality to be in the rose, which is the cause of this sensation.

(Reid, p. 243)

Reid is a hair’s breadth away from idealism. In order to cling to a realism of sorts, he “postulates” a property of the rose which somehow corresponds to (causes) the experienced sensation. This property is not the “agreeable odor” which he experiences. The rose does not have this odor. It has some mysterious property, a property which is only known (by description!) as the property which causes a certain sensation. Skepticism waits around the corner. The familiar pattern emerges: Idealism avoided means skepticism embraced. We escape from the horns of this dilemma because we do not identify the smell of the rose with a sensation in the mind. According to our view, when smelling the odor of the rose, we experience a sensation which has a certain odor, namely, the odor of the rose. We smell the odor of the rose and we experience a sensation with the same odor. No mysterious property hides behind the familiar odor.

Even near-contemporary philosophers make the mistake of identifying sensations with their properties. G. E. Moore, for example, makes this mistake in his Some Main Problems of Philosophy (pp. 48, 52, 54, and 67). It is clear that his infamous question of whether or not sensations (sense-data) are parts of the surfaces of perceptual objects must appear in a different light as soon as we distinguish between a sensation and its properties. When I look at Oscar, is the green sensation which I experience part of his surface? Obviously not, for this sensation no longer exists when I turn my head. When I turn away, it is the sensation which disappears, not a part of the surface of the apple. Is it then the color green which is part of the surface of Oscar? Obviously not, for the color is not a spatial part of the surface of the apple. Rather, it is a property of that surface. The color of the sensation is the same as the color of (the surface of) the apple.

Or consider what Russell says in connection with the argument from the relativity of sensing. Commenting on Berkeley’s conclusion from this argument, Russell states:

Our previous arguments concerning the color did not prove it to be mental; they only proved that its existence depends upon the relation of our sense organs to the physical object—in our case, the table. That is to say, they proved that a certain color will exist, in a certain light, if a normal eye is placed at a certain point relatively to the table.

(B. Russell, The Problems of Philosophy, p. 41)

But colors, according to Russell, are universals and hence exist timelessly (Russell, p. 100). Thus the existence of a color could not possibly depend on normal light and a certain eye’s position. What does so depend on the light, the position of the eye, and on many other things as well, is the existence of the corresponding sensation. More precisely, it is the existence of a sensation with certain properties. Under different circumstances, a different sensation with different properties would exist. Russell has carelessly identified an individual sensation with its properties.

Let us return to the main line of our inquiry. The Cartesian philosophy, I have claimed, contains the seeds both of idealism and of skepticism. The seed of idealism is the distinction between primary and secondary qualities; the seed of skepticism is the principle of immanence. We have seen how Berkeley manages to avoid skepticism by embracing idealism. But surely “houses, mountains, rivers, and, in a word, all sensible objects” do exist outside of minds. Is there no other way to avoid skepticism? And, leaving skepticism aside, what is the proper way to guard against idealism? Perhaps a look at Reid, brilliant critic of Berkeley and Hume, will show us the way.