Knowledge of Primary and Secondary Qualities

Knowledge of Primary and Secondary Qualities

We normally think we know the shapes and colors of things: I know, for example, that the cup in front of me is cylindrical and blue. But reflection casts doubt on this commonsense assumption, in two ways. First, there is the question of whether I really know that the cup is cylindrical, given that I might be under an illusion about this; skepticism undermines the claim to knowledge. The belief that the cup is cylindrical purports to record an objective property of the object, which is not logically entailed by any subjective evidence I have, such as that it looks cylindrical to me. Thus, the belief is not justified by anything immediately available to my mind, according to traditional skepticism, and hence is not a case of knowledge. On the other hand, and second, there is a problem about the claim that I know the color of the object, because it is not true that the cup is really blue—the cup is not objectively blue and hence this cannot be known.[1] Maybe there is something bluish going on in my mind, but the cup itself is not objectively blue, so the truth condition for knowledge isn’t satisfied. I might know that I am experiencing something blue, but it isn’t the cup—that thing existing in the mind-independent objective world. The justification condition is met just as well as in the shape case, for what that is worth, but the truth condition fails. So, I know neither that the cup is cylindrical nor that it is blue, but for different reasons. In the one case, the attribution is too objective to be known, because it exceeds my subjective evidence; in the other case, the attribution is too subjective to be known, since objects are not objectively colored. I cannot know the objective properties of things because of the gap between subjective and objective; and I cannot know that things have subjectively defined properties because they don’t, not really, not in themselves.

There is a way out of this pair of problems that has seemed attractive to many: bring shapes closer to the mind and move colors closer to external objects. Thus, we might maintain that shapes are dispositions to appear a certain way to subjects of experience, and colors are likewise dispositions of objects so to appear. We make the properties span the divide between subject and object. Then, we will have closed the epistemic gap between the objective nature of shape and the subjective experience of shape; and also linked the subjective nature of color to the external object. We will be able reliably to infer shape from appearance of shape, since shape is just a propensity to appear a certain way; and we will also link color to objects outside the mind, since colors just are dispositions of objects to produce sensory impressions. We will therefore have straddled the divide, making room for both sorts of knowledge. We put shapes partly in the mind, to which we have reliable access; and we put colors partly in the object, thus ensuring that our belief that objects are colored is true. We thereby save commonsense epistemology. This is a clever idea, and it satisfies a felt need, but it is really not plausible—it is a rickety contraption that fails to do justice to the facts. For, first, the real shapes of objects are not dispositions to appear thus and so: such a disposition could exist and yet the object not have the property in question, or the object could have a shape property that fails to generate the alleged disposition. Shapes characterize objective reality no matter how they may be disposed to affect the mind; they would be there even if there were no minds to interact with them. So, the epistemic gap remains; and it is not merely the product of some weird hyperbolic skepticism—it might be the result of serious science, as in certain physical theories (non-Euclidian geometry etc.). In the case of colors, the problem is that the appearance of color doesn’t square with the alleged metaphysics of color—colors don’t look like dispositions. Moreover, color perception and color belief actually attribute color to external objects in much the same way they attribute shape, i.e., non-dispositionally. It looks to us as if objects have intrinsic color properties that are not dependent on our minds. But they don’t actually have such properties, according to the secondary quality conception, so error is built into our beliefs about them. Accordingly, we don’t know them to have properties they really (truly) have. They are more subjective than our common sense suggests; so, we are in error about them. I don’t really know that my cup is cylindrical and blue, after all.

The problem arises because knowledge purports to be objective. It aspires to reach beyond itself into objective reality (same for perception and belief). It doesn’t purport to be merely subjective (I am speaking of knowledge of external objects not of properties of the mind itself, recognized as such). We don’t commonsensically think we know only what is going on inside us; we think we know facts about objects outside us. But this raises two objections: first, we are too epistemically ambitious, thinking we know more about reality than we have good reason to suppose; and second, we mistake the subjective for the objective, as when we project colors onto external objects. The aspiration to knowledge encourages us to objectivize, but this exceeds what we have a right to believe—we overdo our epistemic credentials. We underestimate the epistemic remoteness of primary qualities, and we overestimate the proximity of secondary qualities to external reality. We think the former are closer to our minds than they are, and we think the latter are closer to objects outside our minds than they are. So, we are more sanguine about our knowledge of these things than we ought to be. Upon examination, we see that we don’t really know the primary (objective) qualities of things, and we don’t know the true location of secondary qualities in the universe. Common sense is wrong about the status of both kinds of putative knowledge. We thus need to revise our ordinary beliefs about these things. What we should say is that we don’t know the objective reality of the primary qualities of nature, as a matter of our common sense (advanced science might know it better); and we don’t know that objects are colored, as they seem to common sense to be–we are under an illusion about that.

This has broader implications for the nature of knowledge in general. It sounds reasonable to separate human knowledge into two piles: absolute knowledge and relative knowledge. Some of our knowledge concerns reality as it exists independently of any sentient being, while some concerns matters of human interest and sensibility. But there are two objections to this kind of division: the absolute nature of reality is too far removed from our ordinary knowledge-acquiring abilities to count as genuinely known, though it undoubtedly exists; and the relative reality that reflects human nature is not conceived as such by our ordinary common sense—we tend to think it is more objective than it really is. We think our parochial representations are more absolute than they are, because knowledge aspires to generality—we over-objectivize because we take ourselves to be knowing beings in good standing. We are doubly overconfident epistemically: we think we know more about mind-independent reality than we do, and we think our merely relative knowledge is more robust than it is—more absolute, unconditional, and general. This is not mere dogmatism; it reflects the very nature of knowledge as objectively aspirational. The concept of knowledge itself makes us expect more generality than what naturally comes to us; we are suckers for the universal. To fully acknowledge the parochial nature of many of our beliefs would be to retreat from the enterprise of acquiring knowledge. The concept is too ambitious to fit what we are actually capable of achieving. Modesty recommends abandoning the concept, or scare-quoting it, or substantially qualifying it. You can’t, for example, keep on saying you know various moral truths when your official theory is that morality is an historically contingent relative human practice. You can’t have it both ways. There is thus no easy way out of the dilemma I have sketched: either claims to human knowledge are too ambitious, being concerned with the nature of reality absolutely considered; or they are falsely modest, because all knowledge aspires to universality in some measure. No one wants to say, “Oh, this is just something that we know; you may know something quite different”.  Knowledge and universality go together. Knowledge is undermined by relativity, but absoluteness is asking an awful lot. Both ideas are quixotic, intrinsically difficult to satisfy, arguably impossible of achievement.[2]

[1] I am assuming the classic distinction between primary and secondary qualities in this paper. I am concerned with the epistemological implications of this distinction, particularly as regards knowledge. For some background, see my The Subjective View (1983) in which I argue that primary qualities are subject to skepticism while secondary qualities are not. In that book I was too confident of the dispositional theory of color as consistent with common sense.

[2] If you find this paper hard to follow in places, don’t blame yourself; this is difficult stuff. It may help if I say that the figure of Bernard Williams lurks in the background.

1 reply
  1. Giulio Katis
    Giulio Katis says:

    What about the objectivity of the degree to which some particular subjective knowledge is shared, or not (across cultures, times, even perhaps species)? Could there be any rational claims to this form of knowledge?


Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.