a) The Scholastic Background

Descartes’s “new way of ideas” breaks decisively with the Aristotelian tradition. But it also leads inevitably to both skepticism and idealism. It is ironic that the supreme attempt to refute skepticism should make skepticism all but inevitable. And it is tragic that the modern features of Descartes’s philosophy should unavoidably lead to skepticism and idealism. The tragedy lies in the fact that modern philosophers seem to be faced with an inescapable dilemma: either they have to reject Descartes’s innovations, or else they have to pay for embracing his revolution by succumbing, sooner or later, to skepticism and/or idealism. What precisely is responsible for this malaise of modern theories of knowledge?

According to the Aristotelian tradition, a perceptual object, for example, an apple, is a combination of matter and form. This combination, the so-called substance, has certain properties, the so-called accidents of the substance. For example, the apple may be green. When one perceives the apple, let us call it “Oscar,” a complicated causal process takes place. The gist of this part of the traditional theory is that the apple, through a medium, causes a material impression on the sense organ; in our case, on the eye. This impression is considered to be a material image of the perceived object. Next, the so-called active intellect goes to work on the material image and extracts from it the substantial form of the perceived object. As a result of this mental process, the form of Oscar comes to exist in the mind of the perceiver. The perception of Oscar thus simply consists in that his form exists in a mind. The form is then both “in” the apple and “in” the mind. One and the same thing, to emphasize, thus exists both in the mind and in the apple: the form, as it is usually put, exists materially in the apple and immaterially in the mind. Some philosophers conclude that the form has two kinds of existence: it has material existence in Oscar and mental existence in the mind of the perceiver.

There are three parts of this analysis which I wish to stress for future reference. Firstly, the Tradition assumes that in perception the mind does not “reach out” to make direct contact with Oscar. It is directly related only to what is in it, namely, the form of Oscar as it exists mentally. I shall call this all-important assumption “the principle of immanence,” and I shall sum it up in the slogan: The mind knows (directly) only what is in it. This principle, I shall argue, is the source of the skepticism invited by Descartes’s philosophy.

Secondly, even though the mind can only know what is in it, the mind nevertheless knows the form of Oscar; for this form is literally in the mind. The mind, therefore, has really no need to “reach out” and touch the apple. The form of Oscar comes to the mind, so to speak. Let us call the form of Oscar, as it exists in the mind, a “concept.” In having the concept of Oscar in his mind, the perceiver knows Oscar. The relationship between the concept and its object, between the form as it exists in the mind and as it exists in the substance, is one of identity. I shall call this “the thesis of identity.”

Thirdly, the matter of this particular apple, Oscar’s matter, does not exist in the mind of the perceiver. Now, some philosophers of the Tradition have held that it is the matter of a substance which materially distinguishes between one substance and another. If so, then it is Oscar’s matter which accounts for his individuality. It is Oscar’s matter which accounts for his “thisness.” But since his matter never exists in a mind, we cannot perceive Oscar in his individuality. It follows that we cannot really perceive the individual apple Oscar but merely his form.

b) Knowledge by Way of Ideas: The Inevitability
of Skepticism

Descartes’s epistemological revolution consists in a rejection of the thesis of identity. What is in the mind when you perceive Oscar is not his form, but something else, namely an idea of him. This idea is neither identical with any property of Oscar’s nor with the property of being an apple. This idea, far from being identical with the form (essence) of Oscar, does not even resemble Oscar’s form. The Cartesian idea, unlike the concept of the Tradition, is not in any sense, shape, or form like the apple. This rejection of the identity thesis follows straightforwardly from Descartes’s conception of the mind as a substance which is essentially different from material substances. Nothing which is “in” a material substance can also be “in” a mind, and conversely. Mind and matter have nothing in common.

But Descartes, like almost every other philosopher of his time (and many philosophers of our time), accepts the calamitous principle of immanence. He agrees with Arnauld that “we can only have knowledge of what is outside us through the mediation of ideas in us” (A. Arnauld, L’Art de penser; la Logique de Port-Royal, p. 63). In Leibniz’s words, human souls “perceive what passes without them by what passes within them” (The Leibniz-Clarke Correspondence, p 83).

It is the combination of the rejection of the thesis of identity with an acceptance of the principle of immanence that condemns Descartes’s philosophy to fall before the skeptical argument. Since the mind can only know what is in it, it can only know the idea of the apple. Knowledge of Oscar, if it is possible at all, must therefore be mediated by this idea. But since this idea of the apple is no more like Oscar than the word ‘Oscar’ is, Oscar can only be known as the thing which is represented by the idea. The relationship between the idea and its object is neither identity nor resemblance. Let us call it “representation.” What kind of relation is representation? Some of Descartes’s opponents claimed that representation is inconceivable. Descartes’s view, they said, is simply unintelligible. Surprisingly, some Cartesians agreed with this verdict. Representation, they conceded, is a miracle. Yet they preferred the miracle to a rejection of their revolutionary new insight that ideas are not like their objects. The idea of Oscar, they kept on insisting, is not at all like Oscar (cf. R. A. Watson’s The Downfall of Cartesianism). An old way of thinking thus clashed in their minds violently with a new-found conviction. We shall see that the representation relation is not at all inconceivable. It is perfectly reasonable to assume that there exists a unique and indefinable relation between every idea and its object. The idea of Oscar, for example, is an idea of Oscar, and of nothing else, because it represents (intends) Oscar and nothing else.

No, the problem with Cartesianism lies somewhere else. The principle of immanence is the true culprit. It is this principle which provides the skeptic with his decisive argument against Cartesianism. The argument goes as follows (see, for example, John Sergeant, Solid Philosophy, Asserted against the Fancies of the Ideaists . . .). In order to know what a given idea I₁ represents, we must be able to compare the idea with what it represents. But in order to compare I₁ with an object O₁, so that we can find out whether or not O₁ stands in the representation relation to I₁, we must obviously know both the idea and the object, but in addition we must know the object other than by means of the idea. However, according to the principle of immanence, we cannot know an object other than by means of an idea. Therefore we cannot compare an idea with an object and hence cannot find out what objects correspond to which ideas. We do not know, therefore, what objects our ideas represent. And thus we do not know what there is in the way of an external world.

Since I believe that the skeptic’s argument is valid, let me reformulate it in a different way in order to bring out what I take to be its fundamental structure. The question is again: What object is represented by a given idea I₁? We can only answer this question if we can establish a fact of this sort: I₁ stands in the representation relation to O₁. This fact must somehow be presented to the mind. But in order to have this fact present to the mind, both I₁ and O₁ must be present to the mind. But, according to the principle of immanence, O₁ cannot be present to the mind; for the mind can only be presented with what is in it, and O₁ is not in it. Hence we cannot find out what object I₁ represents. The situation is quite different when it is a question of relations between ideas. If we want to know whether or not I₁ stands in the relation R to I₂, we can apprehend the fact that I₁ stands in R to I₂; for we can have present to the mind both I₁, and I₂ (and, we have assumed, also R). But in the case of I₁, and an object, the object can only “be known by description” since, according to the principle of immanence, it cannot be present to the mind. But such knowledge is useless in our case. If we ask whether or not I₁, represents “the object which I₁ represents,” the answer is obvious but does not tell us what specific object I₁ represents. And if we ask whether or not I₁ represents “the object which I₂ represents,” the answer is obviously negative.

I have tried to give as clear a version of the skeptic’s argument as I can, because it has also been applied to sensations and has been accepted in that context by philosophers from Locke to Russell. It alleges there, as we shall see, that we cannot know the true degree of warmth of a bucket of water or the true shape of a table top. All we can ever know are our own sensations of warmth and of shape.

In whatever context the skeptic’s argument may appear, it is clear from our analysis how we must attack it. We cannot give up Descartes’s brilliant insight that ideas do not resemble their objects. The idea of Oscar is not at all like Oscar: it has no color, it has no shape; it is neither hard nor soft; it has no smell, nor has it size. The idea of Oscar is as different from Oscar as anything can be. Yet it represents Oscar. There is a unique (intentional) relation which connects every idea with its object. This part of Cartesianism is sound. What has to be abandoned is the principle of immanence. It is not true that the mind only knows what is in it. Quite to the contrary, as we shall see: the mind never knows what is in it; it only knows what is before it! As I see the dialectic, skepticism can only be averted if we refute the principle of immanence. In order to refute it we shall have to unearth the motives for the principle and attack these motives. And we shall have to show how a rejection of the principle agrees with the phenomenology of perception. This is the main task of the second section of this book.

Cartesianism, by combining a rejection of the thesis of identity with an acceptance of the principle of immanence, leads to skepticism. But this combination is almost always part of a representational theory of knowledge. Representationalism, therefore, almost always falls prey to skepticism. But Cartesianism also leads to idealism. In this case, the fault lies not with the Tradition but with modern science. Cartesianism, as has often been remarked, has one foot in the Tradition and one foot in modernity. Since it retains the principle of immanence, it is too old-fashioned to escape from skepticism. Since it embraces the distinction between primary and secondary qualities, it is too modern to avoid idealism.

c) Secondary Qualities: The Inevitability of Idealism

The distinction between primary and secondary qualities, it should be emphasized, is not an essential part of the Cartesian revolution. Rather, it is forced upon philosophy in general and the Cartesians in particular by the scientific revolution of the 17th century. Galileo, for example, makes the distinction in these words:

Now I say that whenever I conceive any material or corporeal substance, I immediately feel the need to think of it as bounded, and as having this or that shape; as being large or small in relation to other things, and in some specific place at any given time; as being in motion or at rest; as touching or not touching some other body; and as being one in number, or few, or many. From these conditions I cannot separate such a substance by any stretch of my imagination. But that it must be white or red, bitter or sweet, noisy or silent, and of sweet or foul odor, my mind does not feel compelled to bring in as necessary accompaniments . . .

Hence I think that tastes, odors, colors, and so on are no more than mere names so far as the object in which we place them is concerned, and that they reside only in the consciousness. Hence if the living creature were removed, all these qualities would be wiped away and annihilated. [The Assayer, in The Philosophy of the 16th and 17th Centuries, R. H. Popkin, ed. (New York: The Free Press), 1966, p. 65]

One can hardly say it more clearly: colors exist only in minds, and if there were no minds, colors would not exist. Before there were minds, ferns were not green, sunsets were not red, and the sky was not blue! What an invitation to apply modus tollens! If it were true that colors exist only in minds, then the sky wouldn’t be blue without minds. But the sky would obviously be blue, even if there were no minds. Therefore, it cannot be true that colors exist only in minds. I am perfectly willing to rest my case against the distinction between primary and secondary qualities on this argument. But, as usual, the philosophical situation is more complicated; though, I think, not more sophisticated, and we must return to Descartes to fill in the rest of the story.

Extension, according to Descartes, is a modification of material substance and, as such, is known by ideas. This means among other things, I take it, that the shapes of things are known by way of ideas. Shapes are therefore represented by ideas which do not resemble them. Color, on the other hand, is not a property of material substances. It does not exist in the external world at all; it is merely a sensation in minds. Sensations, in distinction to ideas, though they are caused by material substances, do not represent such substances (or their properties). The color olive green, a certain determinate shade of green, is merely a sensation in the mind. Thus the mind contains both ideas and sensations, and both ideas and sensations are caused by material substances. But ideas, unlike sensations, are not merely caused, they also represent.

I think that this is an accurate description of Descartes’s view. But we must realize that his view is less clear than this brief description may make it appear. For example, both sensations and ideas are said to be in minds. But how are they in minds? What kind of relation do they have to minds? As far as ideas are concerned, an answer is close at hand: Ideas are modifications of mental substances. They are “in” mental substances in precisely the same way in which shape is “in” a material substance. The relation between a mind and its ideas is the relation between a substance and its accidental modifications. But what about sensations? If olive green were a modification of the mind, then the mind would have to be olive green! But surely minds are not colored. Nor, of course, are they hot or smelly. Notice that this problem does not arise for shape. Shape is not a property of the mind. It exists in the mind only as an idea. Notice also that a similar problem already arises in the Tradition. The form or essence of the apple Oscar, as we saw, is supposed to be present in the mind of the person who perceives Oscar. But it is clear that it cannot be “in” the mind in the same way in which it is “in” Oscar. By being in Oscar, it makes Oscar what Oscar is, namely, an apple rather than, say, a snake. But by being in the mind of the perceiver, it does not make the perceiver into an apple. Returning to Descartes, we must assume that while ideas are modifications of minds, sensations are in some other, not further specified, way in the mind. But this implies that we cannot give a univocal explication of what it means for a mind to know what is in it. If to know an idea is to be modified by it, then sensations must be known in some other fashion. Or else they cannot be known at all.

Another difficulty appears in connection with Descartes’s famous wax example in the second Meditation. According to our interpretation of Descartes, a certain idea, I₁, represents a certain shape, say, squareness. Now, in connection with the wax example, Descartes claims that we recognize that the piece of wax before us is the same as the one we saw earlier, even though all of its sensible properties have changed, including its shape and color. Here he seems to be treating shape and color alike as sensations in the mind, known, as he says, “by means of the senses.” But what is then known, not by means of the senses, but by way of ideas (by the understanding)? If we reject everything that does not belong to the piece of wax, Descartes asserts, then nothing is left “but something extended, flexible, and movable.” What is this notion of something extended? Is this the notion of something square, or of something rectangular, or of something round? Obviously not, Descartes answers, for he would not conceive clearly and truthfully what wax was, if he did not think that this bit of wax were capable of receiving more variations in extension than he had ever imagined. Thus, when Descartes conceives of the wax as something extended, he does not conceive of it as having this or that particular shape.

The wax example seems to show that Descartes treats particular shapes exactly like particular colors, namely, as sensations. But he also says that we know the piece of wax as something extended, not by means of the senses, but through the understanding. How can we reconcile these two views? Well, it could be that particular shapes are sensations, but that the property of having shape is not a sensation but is known by means of an idea. According to this possibility, what the wax really and truly has is not this or that particular shape, but the property of having shape. The distinction between primary and secondary qualities is then, not, as we have assumed, the distinction between squareness and olive green, but the distinction between having shape, on the one hand, and being square or olive green, on the other. However, this interpretation runs immediately into difficulties. If the distinction between primary and secondary qualities is the distinction between having shape and squareness, then it must also be the distinction between having color and olive green. And then being colored would be just as much in the material substance as extension, that is, as the property of being shaped. And the same would hold for all other sensible properties. Galileo’s distinction would be destroyed. Nor does this interpretation make much sense on other grounds. What sense could it possibly make to assert that a material substance has the property of having shape, but does not have any particular shape, that it is colored, but does not have any particular color?

One thing is certain: Descartes’s wax example throws our neat distinction between shapes as primary qualities and colors as secondary qualities into a cocked hat. Perhaps the facts are rebelling against theory, as so often happens in philosophy. This is a most pleasing thought. For is it not preposterous to believe that the wax really has a shape but has no color? Does not the wax change its shape just as much as its color when it is held to the fire? No, the reason for Descartes’s distinction between primary and secondary qualities cannot be found in enlightened common sense. We know of course where it comes from. Physics tells us that the “atoms” of which the wax consists have shape but have no color. Scientifically minded philosophers conclude, therefore, that “out there,” in the nonmental world, color is not to be found. And since it is not “out there,” where else could it be but “in” the mind?