« Prev A. Two Worlds or One? Next »

A. Two Worlds or One?

What is to be said for this understanding of Kant? Hermeneutical obstacles of formidable proportions loom. First, how are we to think about this distinction between the noumena and the phenomena, the things in themselves and the things for us? Unfortunately, the commentators are not of one mind. There is a huge interpretative watershed, a continental divide, between two fundamentally different interpretations or basic pictures of what Kant had in mind, each with several variations when it comes to detail. According to the first and 11more traditional picture, Kant held that there are two realms of objects, two fundamentally different kinds of things. These are the phenomena, on the one hand, and the noumena, on the other; the things in themselves and the things für uns. (These two distinctions don’t exactly coincide in Kant; the ways in which they don’t aren’t relevant to our present inquiry.) On the one hand, on this picture, there are tables and chairs, horses and cows, stars and planets, the oak tree in your backyard, just as we ordinarily think. These things really exist and are really there. They are phenomenally real, real parts of the world of experience. But they are also transcendentally ideal: that is, they are not part of the world as it is independent of human experience. On the other hand, there are the noumena, which are transcendentally real. These are the things as they are in themselves; they do not depend for their existence or character upon human beings or human experience. These two realms are disjoint: none of the phenomenal objects is a noumenon, and none of the noumenal objects is a phenomenon. Here are a couple of passages supporting this interpretation:

Now we must bear in mind that the concept of appearances, as limited by the Transcendental Aesthetic, already of itself establishes the objective reality of noumena and justifies the division of objects into phaenomena and noumena and so of the world into a world of the sense and a world of the understanding (mundus sensibilis et intelligibilis) and indeed in such manner that the distinction does not refer merely to the logical form of our knowledge of one and the same thing, according as it is indistinct or distinct, but to the difference in the manner in which the two worlds can be first given to our knowledge, and in conformity with this difference, to the manner in which they are in themselves generically distinct from one another. (A249, Kant’s emphasis)

Appearances are the sole objects which can be given to us immediately, and that in them which relates immediately to the object is called intuition. Appearances are not things in themselves; they are only representations, which in turn have their object—an object which cannot be intuited by us, and which may, therefore, be named the non-empirical, that is, transcendental object = x. (A109)

These phenomena are objects, objects that exist in space and time. The noumena, by contrast, are neither temporal nor spatial; space and time are forms of our intuition rather than realities that characterize the things in themselves. Noumena and phenomena, therefore, are distinct.

Still further, we have experience only of the phenomena, not of the noumena:

We have sufficiently proved in the Transcendental Aesthetic that everything intuited in space or time, and therefore all objects of any experience possible to us, are nothing but appearances, that is, mere representations,


Further still, the phenomena, the world of stars and planets, trees and animals, depends on us for existence. The above passage continues:

which in the manner in which they are represented, as extended beings, or as series of alternations have no independent existence outside our thoughts. (A491, B519)


That nature should direct itself according to our subject ground of apperception, and should indeed depend upon it in respect of its conformity to law, sounds very strange and absurd. But when we consider that this nature is not a thing in itself but is merely an aggregate of appearances, so many representations of the mind, . . . (A114)

Now to assert in this manner, that all these appearances, and consequently all objects with which we can occupy ourselves, are one and all in me, that is are determinations of my identical self, is only another way of saying that there must be a complete unity of them in one and the same apperception. (A129)

This is the more traditional way of understanding Kant, the way Kant was taken by his great successors. To put it briefly and all too baldly, there are two realms of objects; our experience is only of one realm, the realm of phenomena, which themselves depend on us for their existence; if we should go out of existence, so would they. That is because the phenomenal realm is somehow constructed by us out of the given, the data, the raw material of experience. The noumenal realm, however, is not thus dependent on us but is also such that we have no intuition, no direct experience of it. Finally, there is nevertheless a connection between the two worlds in that something like a causal transaction between the noumena and the transcendental ego (itself a noumenon) produces in us the given out of which we construct the phenomenal world.

Call the above the two-world picture; this has been the dominant interpretation. There has always been another basic interpretation of Kant, however, one that more recently has perhaps achieved majority status. According to this other picture, there really aren’t two worlds after all, a world of phenomena and underlying it another world of noumena. There is only one world and only one kind of object, but there are (at least) two ways of thinking about or considering this one world. All objects are really noumenal objects, and talk about the phenomena is just a picturesque way of talking about how the noumena, the only things there are, appear to us. The phenomena-noumena distinction is not between two kinds of objects but, rather, between how the things are in themselves and how they appear to us.

So, for example, Graham Bird:


Such phrases [e.g., ‘transcendental objects and empirical objects’] should be understood to refer not to two different kinds of entity, but instead to two different ways of talking about one and the same thing.1818   Kant’s Theory of Knowledge (New York: Humanities Press, 1962), p. 37.

And Michael Devitt:

It is tempting to equate an appearance with the foundationalist’s sense datum, taking the thing-in-itself as the unknowable external cause of this mental entity. Kant’s writing often encourages this temptation. Nevertheless, scholars seem generally agreed—and have convinced me—that this two-worlds interpretation is wrong. What Kant intends is the following influential, but rather mysterious, one world view.

An Appearance is not a mental sense datum, but an external object as we know it. In contrast the thing-in-itself is the object independent of our knowledge of it; it is not a second object, and does not, indeed could not, cause an appearance. . . . 1919   Realism and Truth (Princeton: Princeton University Press, 1984), p. 59. See also D. P. Dryer, Kant’s Solution for Verification in Metaphysics (Toronto: University of Toronto Press, 1966), chapter 11, section vi; H. E. Matthews, “Strawson on Transcendental Idealism,” Philosophical Quarterly 19 (1969), pp. 204–220; Henry Allison, Kant’s Transcendental Idealism (New Haven: Yale University Press, 1983). I am indebted, for these references, to Karl Ameriks (“Recent Work on Kant’s Theoretical Philosophy,” American Philosophical Quarterly 19 [1982], and “Kantian Idealism Today,” History of Philosophy Quarterly 9 [1992]) and to James Van Cleve, Problems from Kant (New York: Oxford University Press, 1999).

Although this second picture is perhaps now the majority opinion, it seems a bit difficult to reconcile it with Kant’s own view that his thought constituted a revolution—his famous second Copernican revolution.2020   “Hitherto it has been assumed that all our knowledge must conform to objects. But also attempts to extend our knowledge of objects by establishing something in regard to them a priori, by means of concepts, have, on this assumption, ended in failure. We must therefore make trial whether we may not have more success in the tasks of metaphysics, if we suppose that objects must conform to our knowledge. We should then be proceeding precisely on the lines of Copernicus’ primary hypothesis.” (Bxvii) After all, much of this second picture would be accepted even by such staunch prerevolutionaries as Aristotle and Aquinas. Both would agree that there is or can be a difference between the world (or any less impressive object) as it is in itself and the world as it appears to us; this is to admit no more than that we can be mistaken about the world or things in the world, and of course Aristotle and Aquinas would hardly deny that. Both would agree to something much stronger: that the world might have many properties of which we have no conception, so that our way of thinking about the world, the properties we ascribe to it, are not necessarily all and only the 14properties it has. For Aquinas or any other theist, this would be close to a truism: God, obviously enough, has many properties we don’t know about, and presumably many of which we could not so much as form a conception. The essential elements of the one-world view seem perhaps a bit too uncontroversial, at least with respect to Kant’s predecessors, to constitute a revolution, Copernican or otherwise.

According to Merold Westphal:

Finally, all twelve categories insofar as they constitute the world of human experience and are not merely formal features of judgment, are schematized with an essential reference to time. Thus the object and property that would disappear from the world in the absence of human knowers are not object and property per se, but substance and accident as defined by human temporality. Similarly, the truth and falsity that would disappear derive from the categories of reality and negation as essentially linked to our experience of time. Thus we are back to the tautology that in the absence of human cognition the world as apprehended by human minds would disappear.2121   “In Defense of the Thing in Itself,” p. 170.

That does, indeed, seem to be a tautology, or at least a trivially necessary truth; we could add that in the absence of bovine cognition, the world as apprehended by bovine minds would disappear. But how could Kant think of this as constituting a revolution, one according to which objects must conform to our minds (rather than, as previously thought, our minds to objects) if we are to have knowledge? Could a tautology constitute a revolution?

1. The One-World Picture and Reference to the Noumena

Our main interest here does not lie in trying to resolve the question of what Kant intended: that is perhaps necessarily beyond our powers. Instead, we are looking to see if there is good reason, either given by Kant or constructible from materials given by him, for the conclusion that our concepts do not apply to God. And how does the difference between these two interpretations of Kant bear on this question? Consider the second picture first, and note that on this picture, if our concepts apply to anything, they apply to the Dinge, those being the only things there are. Similarly, if we manage to refer to and think about anything at all, we succeed in referring to and thinking about the Dinge, because they are all that there is. So how could it be that the categories and our other concepts do not apply to them?


Well, what is it for a concept to apply to something, for something to fall under a concept? Consider the concept being wise. That concept applies to something (a thing falls under that concept) only if that thing is wise, only if, that is, it has the property of being wise. Properties and concepts are thus correlative. I have the concept being wise only if I grasp, apprehend, understand the property being wise. I have the concept being a prime number if and only if I grasp or apprehend the property being a prime number. For each property or attribute of which I have a grasp, I have a concept. Of course there are properties of which I have no concept. Small children often lack the concept of being a philosopher; that is to say, they have no grasp of the property being a philosopher. Large philosophers often lack the concept of being a quark; that is to say, they have no grasp of the property being a quark. No doubt there are properties none of us human beings grasps.

One further familiar fact about properties and concepts: they have negations or complements. There is the property being red; there is also its complement, which, naturally enough, is being unred, not being red. There is the property of being wise but also the property of being unwise, failing to be wise. So if one of my concepts (e.g., being wise) does not apply to a thing, then the complement of that concept (being nonwise, not being wise) does apply to it.

Perhaps you want to point out that this way of putting the matter presupposes that there are negative properties, such properties as being nonred, being unwise, and the like; you might object that in fact there are only positive properties, not negative ones. (You might also object to disjunctive and conjunctive properties.) This is no place to try to settle that issue. Clearly, there is the concept of a thing’s failing to be wise (I know what it is for a thing not to be wise), even if there is no negative property nonwisdom. So if you object to negative properties, say that a thing falls under the concept nonwisdom just in case it does not fall under the concept wisdom; more generally, for any property P, a thing falls under the concept P if and only if it has the property P; it falls under the concept not-P if and only if it does not fall under the concept P.

Given this elementary lore about concepts and properties, how could it be that the categories and our other concepts do not apply to the Dinge? Take the categories first—the category of causality, for example. What would it mean to say that this category does not apply to the Dinge? So far as I can see, what this would mean is that the noumena do not stand in causal relations to each other or anything else. Consider the property stands in causal relation to something; if the category of causality does not apply to the noumena, then it must be that none of them has that property. So our concept standing in causal relation to something wouldn’t apply to things as they are in themselves. It follows, however, that the complement of that category or concept 16would apply to things as they are in themselves: each of them would be such that it does not stand in a causal relation to anything else. The same would go for our other concepts. On this way of thinking of the matter, our ‘positive’ concepts, you might say, do not apply to things as they are in themselves, which is really to say that there is no positive property we grasp that characterizes a thing as it is in itself. As it stands, however, this needs more work: there are problems about this distinction between positive and negative properties. There are also problems of other sorts: what about such positive properties as being self-identical, for example? Are we to suppose the Dinge are not self-identical?

Well, perhaps these matters can be straightened out. (See chapter 2, p. 48.) For present purposes, what we need to see is that on this way of thinking, it would not really be the case that our concepts fail to apply to God in such a way that we cannot refer to and think about him. What would follow, given that he is a noumenon (of course, in this way of thinking, everything is a noumenon), is that God would not have any of the positive properties of which we have a grasp. It would not be the case that we couldn’t refer to God and predicate properties of him: we could perfectly well do so, but we would be mistaken if we predicated of him a positive property of which we have a grasp. Thus we would make a mistake if we said that God is wise, or good, or powerful, or loving. That would be because nothing is wise, good, powerful, loving, and the like. (On the one-world picture, the Dinge are all there is; so if positive properties can’t be ascribed to the Dinge, they can’t be ascribed to anything.) Here there would be nothing at all special about God; what holds for him also holds for everything else. But those theologians who suggest that Kant showed we cannot refer to and think about God presumably believe that Kant showed there was a special problem about God; they don’t think that what Kant really showed is that we can’t talk or think about anything. As Kaufman puts it in the passage I quoted above (p. 4), “The central problem of theological discourse, not shared with any other ‘language game,’ is the meaning of the term ‘God.’ ” So Kant, taken this way, doesn’t fill this particular bill; it doesn’t give us a relevant way of seeing that our concepts do not apply to God.

2. The Two-World Picture and Reference to the Noumena

Now suppose we consider the other main interpretation of Kant: the two-world picture. This is the more traditional way to understand Kant and still, perhaps, deserves the nod. (Here I am not interested in which picture most accurately represents Kant, but whether Kant, taken any plausible way, gives support to the idea that we cannot 17refer to and think about God.2222   Of course, I do not mean to suggest that the one- and two-world pictures as I present them are the only possible (or actual) interpretations of Kant; clearly, there are various complications and extensions of each. What I claim is that none of them offers aid and comfort to the claim that our concepts do not apply to God.) On this picture, there are two disjoint realms: phenomena and noumena, the Dinge and the things of experience. To add another quotation:

Accordingly, that which is in space and time is an appearance; it is not anything in itself, but consists merely of representations, which, if not given in us—that is to say, in perception—are nowhere to be met with. (A494, B522)

Now when we think about the application of our concepts to the noumena, we see that this two-world picture divides into two subpictures.

(a) The Moderate Subpicture. On the one way of thinking, (some of) our concepts apply to the things in themselves; we can think about them and refer to them, all right, but we can’t have any knowledge of them. When we think about them, predicate properties of them, what we have is just speculation, mere transcendental schein, and we deceive ourselves if we think we have more. Our knowledge doesn’t extend beyond experience; hence, it does not extend to the realm of the things in themselves. This would explain that bewildering variety and proliferation of metaphysical views Kant found so shocking. The reason, fundamentally, is really that all the metaphysicians have been just guessing, whatever their pretensions to apodictic conclusions and conclusive certainty. Our reason can’t operate in the rarefied atmosphere of the noumena, and the result of trying to do so is a mere beating of wings against the void.

Of course Kant also represents his own work in the Critique of Pure Reason as knowledge and as certain and conclusive. And in that Critique he seems to tell us a fair amount about the Dinge: that they are not in space and time, that the world of experience is (in part) a result of a ‘causal transaction’2323   We need the scare quotes because Kant’s official view is that the concept of causality doesn’t apply to the Dinge. between the Dinge and the transcendental ego, and that the latter has no intellectual intuition into the former. So the picture isn’t wholly coherent. Coherent or not, however, this picture doesn’t even suggest that we cannot think about and predicate properties of God. What it suggests, instead, is that when we do, we are not on the sure path of knowledge but on some much more hazardous climber’s trail of mere opinion. So the moderate 18subpicture, too, gives no aid and comfort to the claim that our concepts do not apply to God.

(b) The Radical Subpicture. There is a more striking version of the two-world picture, however, on which we do get the result that we can neither refer to God nor predicate properties of him (call it ‘the radical subpicture’). On both versions of the two-world picture, the appearances are distinct from the things in themselves. The appearances are objects; they exist; they are empirically real. But they are also transcendentally ideal. And what this means, in part, is that they depend for their existence on us (on the transcendental ego[s]) and our cognitive activity. We ourselves are both noumena and phenomena: there is both a noumenal self and an empirical self. The things in themselves somehow impinge on us (taken as transcendental ego), causing experience in us; there is a productive interaction between the transcendental ego and the Dinge (the other Dinge, since the transcendental ego is itself a noumenon), the result of which is experience, the manifold of experience.

As it is initially given to us, this manifold of experience is a blooming, buzzing confusion with no structure. Perhaps it contains among other things what Kant calls ‘representations’ (Vorstellungen); these are of more than one kind, but among them might be phenomenal qualia, something like sense data, or Humean impressions and ideas. The manifold must be ‘worked up’ (Kant’s term) and synthesized by the application of the categories and other concepts. Thus we impose structure and form on it, and in so doing we construct the phenomena, the appearances. So the phenomena, the things für uns, are constructed out of the manifold of experience.

Well, how do we do a thing like that? How do we construct a phenomenon (a horse, let’s say) from the manifold of experience? At this point, the radical subpicture diverges from the more pedestrian version of the two-world picture, for on the radical subpicture, we construct objects by applying concepts (representations, Vorstellungen) to the manifold. The world of appearance gets constructed by virtue of our synthesizing the manifold, which proceeds by way of our applying concepts—both the categories and other concepts—to the manifold. We can’t perceive or in some other way witness this construction; Kant says we are largely unconscious of the activity whereby we structure the manifold and construct the phenomena. Still, it proceeds by way of the application of concepts to the blooming buzzing manifold of experience.

This would require a way of thinking about concepts and their function that is very different from the way of thinking about them I outlined above (a way according to which a concept is fundamentally a grasp of a property). And Kant suggests a different way of thinking of 19concepts: he sometimes calls them rules. Kant says that the understanding is the faculty of concepts; it is the source of our concepts. But he also says of the understanding, “We may now characterize it as the faculty of rules. . . . Sensibility gives us forms (of intuition) but understanding gives us rules” (A126, Kant’s emphasis). And he goes on to say,

Rules, so far as they are objective . . . are called laws. Although we learn many laws through experience, they are only special determinations of still higher laws, and the highest of these, under which the others all stand, issue a priori from the understanding itself. They are not borrowed from experience; on the contrary, they have to confer upon appearances their conformity to law, and so to make experience possible. Thus the understanding is something more than a power of formulating rules through comparison of appearances; it is itself the lawgiver of nature. (A127)

I don’t for a moment pretend that this passage or others that could be cited are easy to interpret. Still, the passage does seem to suggest that concepts are rules and rules are laws. What sort of rules and what sort of laws? Perhaps they are rules for synthesizing the manifold, rules for constructing the phenomena. This is the heart of the radical subpicture. Again, I don’t mean to suggest that this is Kant’s view, but some of what he says suggests it. (Some of what he says also suggests that it is false; that is part of his charm.) For example: “What is first given to us is appearance. When combined with consciousness, it is called perception. . . . ”

Interpretative difficulties abound; the basic idea, however, is that concepts are rules, rules for the synthesis of the manifold and the construction of phenomena. (They are also laws, laws whereby the phenomena are constructed from the manifold of experience.) These rules apply to portions or bits of experience and, by way of their application, the phenomena are constructed. A rule of this sort perhaps specifies that certain portions of the manifold are to be combined or ‘thought together’ as an object. So, for example, consider your concept of a horse: it instructs you to associate, think together a variety of representations, a variety of items of experience, thus unifying that bit of the manifold into an empirical object: a horse. It is a rule which would say something like: think that particular congeries of representations together as a unity.

Now again, I don’t mean to claim that this is a coherent picture or a coherent way of thinking about concepts; on the contrary, I believe that it is not. But note that if it is coherent, then (at least if all of our concepts have this function2424   As Karl Ameriks (private communication) reminded me, Kant’s metaphysical deduction certainly seems intended to reveal concepts which are rules for judgments of any sort, whether limited to items of experience or not. and only this function) our concepts will not apply to the noumena. Consider the concept being a horse. 20Understood this way, this concept is a rule for constructing phenomenal objects out of the manifold of experience. Of course it does not apply to the noumena: it cannot be used to construct an object out of them; they are not given to us (experience, the manifold, is what is given to us), and in any event they aren’t the sorts of things out of which phenomenal objects could be constructed. So it isn’t just that the concept being a horse does not apply to the Dinge in the sense that none of them, as it happens, is a horse (all are nonhorses), for then the complement of that concept—being a nonhorse—would apply. But that concept doesn’t apply either: it, too, is a rule for constructing objects from the manifold. It is another way of unifying, synthesizing the manifold. So thought of, a concept could no more apply to the Dinge than a horse could be a number.

On the radical subpicture, therefore, our concepts surely wouldn’t apply to God, if there were such a person. For God would be a noumenon. God would not be something we have constructed by applying concepts to the manifold of experience (God has created us; we have not constructed him.) So, on the radical subpicture, we can’t refer to, think about, or predicate properties of God.

This way of thinking clearly displays a deep incoherence: on this picture, Kant holds that the Dinge stand in a causal or interactive relationship with us, taken as transcendental ego(s);2525   How many of those transcendental egos are there, anyway? Like many questions of Kantian exegesis, this question is vexed. Indeed, on the radical subpicture, it is more than vexed. If the category of number doesn’t apply to the noumena, then there is presumably no number n, finite or infinite, such that the right answer to the question “How many of those transcendental egos are there?” is n. and he also says that they are not in space and time. But on the radical subpicture, Kant (at least if his intellectual equipment is like that of the rest of us) should not be able to refer to the Dinge at all, or even speculate that there might be such things. He certainly shouldn’t be able to refer to them and attribute to them the properties of being atemporal and aspatial, or the property of affecting the transcendental ego(s), thereby producing experience in them. He shouldn’t be able to refer to us (i.e., us transcendental egos), claiming that we don’t have the sort of godlike intellectual intuition into reality that would be required if we were to have synthetic a priori knowledge of the world as it is in itself. (On this picture, we might say, Kant’s thought founders on the fact that the picture requires that he have knowledge the picture denies him.) If this picture were really correct, the noumena would have to drop out altogether, so that all that there is is what has been structured or made by us. The idea that there might be reality beyond what we ourselves have constructed out of experience would not be so much as thinkable.2626   In addition, of course, there is the problem that it takes a great deal of effort to believe that we are really responsible for the existence of sun, moon, and stars, not to mention dinosaurs and other things, that (as we think) existed long before there were any human beings.

« Prev A. Two Worlds or One? Next »
VIEWNAME is workSection