1

Graham Bird's TRK (Bird Reference Bird2006) is concerned not just with Kant's text, but also with the central opposition between what Bird calls the ‘traditionalist’ and the ‘revolutionary’ accounts of the Critique.Footnote 1 He thinks ‘the former ascribe to Kant an exhaustive idealist dualism of mental states, “ideas”, and transcendent things in themselves and then note the inevitable tensions and contradictions which arise from Kant's apparent attempt to escape from that tradition’ (xii). Bird also characterizes this approach as the attempt to ascribe a ‘phenomenalistic idealism’ to Kant that makes Kant look more or less like Berkeley (8). Its adherents include Peter F. Strawson, Paul Guyer and Karl Ameriks, to mention just three.

But there are also some, perhaps unexpected, villains in the piece, like Adolf Trendelenburg and Erich Adickes. Bird's entire commentary may be characterized as an argument against the traditionalist view and as an attempt to defend the revolutionary account, which regards Kant as explicitly ‘denying that exhaustive division in his complex contrast between empirical and transcendental forms of both idealism and realism’ (xii). The revolutionary account takes comfort from similar attempts by Gerold Prauss, Henry Allison and Robert Pippin (8). Kuno Fischer, who denied that Kant overlooked Trendelenburg's gap, also called the ‘neglected alternative’, is a nineteenth-century revolutionary.

More generally, Bird views Kant as a thinker who opposes ‘windy metaphysics’, and defends a ‘reformed’ version of this discipline that is best described as a ‘metaphysics of experience’, i.e. ‘a descriptive survey and classification of the central, fundamental elements constituting our experience’. While it cannot be completely freed from epistemological and even psychological elements, it is most decidedly not normative and foundational.

There is much to praise in this book. Just like his Kant's Theory of Knowledge: An Outline of one Central Argument in the ‘Critique of Pure Reason’ (Bird Reference Bird1962), from which I learnt a lot when I first began seriously to confront Kant during my undergraduate and early graduate studies between 1969 and 1974 in Canada, this book is a rigorous attempt to come to terms with Kant's text, while at the same time engaging contemporary ideas in order to see in what ways they are related to Kant. The early book deeply influenced my own way of thinking about Kant. It is therefore perhaps not surprising that I very much like the new book as well. Though I am not sure that the exclusivist opposition between traditionalists on the one hand and revolutionaries on the other is always the best device to bring out what is and what is not important in Kant, I am sure that everyone serious about Kant can and will learn from this book. I could say more about the merits of the book, but I am not here to praise, but to comment. And there is much worthy of comment. I could have chosen a particular passage, argument, or part of Bird's commentary or Kant's Critique – and that might have been the wiser course of action – but I have instead decided to take a somewhat closer look at one of the central contrasts in the book, namely the one between Erscheinung and Ding an sich, or between ‘appearance’ and ‘thing in itself’.

There has never been any doubt that these two concepts play a most central role in Kant's Critique. To say that the meaning and import of these terms have been a matter of intense discussion and disagreement ever since the Critique first appeared in 1781 would be an understatement. The import and the value of Kant's Critical philosophy have often been thought to depend precisely on the interpretation of these two terms and their cognates. Many philosophers have argued that Kant's project stands or falls with them, and from the beginning some have dismissed them as not much more than meaningless verbiage. They are also central in the dispute between the ‘revolutionary’ Kantians and the ‘traditionalist’ detractors of Kant. It is therefore hardly surprising that they and the contrast between them also play a central role in Bird's work. In the space allotted to me here, I will, of course, not be able to address either Bird's or Kant's view on these matters in their entirety. I have therefore chosen to address just two different but in my view closely related points about this issue. To do so, I will first briefly present Bird's view of the contrast.

2

Bird starts, most appropriately, from the assumption that ‘any account needs to accommodate Kant's own claims about appearances and things in themselves’ and the most relevant ‘ways of characterizing that distinction’ (41). What are the most relevant views for Bird? There are basically two traditional interpretations, namely the ‘two world view’ and the ‘two aspects’ view, which is also sometimes called the ‘dual aspects theory’.

Bird acknowledges that Kant himself sometimes speaks of the contrast between appearances and things in themselves in terms of two worlds (A249) or two aspects. But he then claims that ‘more often [Kant] speaks of a difference in “points of view” in which we conceive or consider objects either as appearances or as things in themselves’ (41; my emphasis; see also 34, 44 et passim). This, he thinks, is no accident, but indicative of problems inherent in the two traditional interpretations. Whether these were also problems in Kant's eyes, or whether they are more problems for the interpreter of Kant, who has to deal with more than two hundred years of mistaken readings of Kant, never becomes entirely clear. In any case, Bird has two main objections to the two world interpretation.

He objects, first, to its suggestion that we should construe the contrast as involving ‘a commitment to the existence and knowledge of those worlds’ – a wrongful suggestion, since ‘Kant firmly and repeatedly denies any such commitments to things in themselves’ (41). He objects, second, to the presumption that two worlds must be invoked; as he puts it, we ‘need to know why it is necessary to invoke references to possible worlds of which we can in principle have no knowledge’ (42). These are the reasons why the ‘two world’ label seems unacceptable to Bird.

The two aspect view has even more serious problems, according to Bird, for if we speak of different ‘aspects’, we need to say what they are aspects of. Since appearances and things in themselves are usually identified as the aspects, some ‘mysterious third thing’ in which they both inhere must be postulated (41). This involves talk of ‘properties’ as well, which cannot be defended either. In Bird's words, ‘Kant's account of “two points of view” does not talk of “aspects” or “features” and does not draw a distinction between things and properties’ (41).Footnote 2 So, neither of these two labels ‘seems to be acceptable’ (42).

Bird goes even further, claiming that both labels ‘are thoroughly misleading and should be given up’ (42). What remains is the two points of view interpretation: ‘Kant undoubtedly thinks of the basic distinction as one between two points of view, that is between conceiving or considering objects of experience as appearances and as things in themselves’ (42; my emphasis). And ‘a “point of view” ’, for Bird, ‘is a point from which objects, in this case objects of experience, can be considered from one side or another; it is not a feature of objects but a relation between viewer and objects’ (41). Accordingly, Bird constantly speaks of the necessity of ‘taking’ things or objects ‘as appearances’, and contrasts this with ‘taking’ these same things ‘as’ things or objects ‘in themselves’, where ‘taking them as appearances’ means that they are taken as being attached to the senses and ‘taking them as things in themselves’ means that they are taken as related to reason. This basic or fundamental conception is compatible with a two world view and may give rise to it, if only because the two conceptions may be (or, I would say, involve) conceptions of two different worlds. But this is not Kant's own position. For Kant holds that ‘objects of experience are the only objects we can properly identify and know, but that these may be construed or misconstrued in different ways’ (42). Things in themselves are just ‘correlates’ of appearances. They do not add to our knowledge simply because they are unknowable.

Bird notes that, since there are ‘inevitable limitations in talking about things in themselves’, Kant needs to show ‘that there is a real point in such a vocabulary’. He must explain how and why we come to think of things in themselves in the first place. Put differently, he must show how and why the concept of a thing in itself is implied as a correlate of the philosophical conception of a pure sensibility. Bird thinks that there are really two reasons, one external, the other internal to his philosophical enterprise, and he concentrates on the latter. Externally speaking, Kant may have come to this question by way of looking at theories proposed by other philosophers, like Plato's idealism, Locke's representative realism, Leibniz's monadology, and Berkeley's immaterialism – all theories which he clearly knew (42, 100, 154). However, Bird argues that ‘Kant's own motivation’, provided at Bxxvi–xxvii, ultimately has more to do with the fact that the concept of ‘ “appearance” requires things in themselves to be conceivable’.Footnote 3 The passage has been translated as follows:

In other words, Bird thinks that Kant needs the things in themselves as necessary correlates of appearances. Talk of appearances makes talk of things in themselves necessary. Thus he argues in chapter 23 that the very notion of ‘appearance’ depends on the conceivability of things in themselves. Put differently, ‘appearance talk’ crucially depends on ‘thing-in-itself talk’. Bird is adamant about one point. Kant is only committed to ‘the conceivability of things in themselves but not to their actual existence’ (560).Footnote 4 We do not need to postulate the existence of things in themselves, only their conceivability.

If we now ask what talk of things in themselves amounts to, or what it means to ‘take’ the object of experience as a thing in itself, or perhaps even better what Bird's account of the thing in itself is, the answer seems to be:

We may also say that mathematics seems to open up a Platonic realm of knowledge. Kant understands why Platonists in particular came to hold this erroneous view; and his ‘diagnosis of the error leaves a purified, but acceptable concept of a thing in itself or noumenon as only negative, problematic, and limiting’ (553).

3

As we have seen, Bird develops an argument for the distinction between appearances and things in themselves on the basis of what might be called the ‘absurdity or inconceivability claim’ made at Bxxvi–xxvii and elsewhere. We can talk of things in themselves because talk of appearance would be absurd without it. This is a very weak argument, for we may very well ask why we need to talk of empirical things as ‘appearances’ in the first place.

Now, in some sense, appearance talk is innocuous. Kant finds that we

This is the difference between an empirical thing in itself and an empirical appearance. It is a distinction we make every day. Something appears to be one thing, but really turns out to be quite different. Something might look green to us, while it is really not.

No further justification needed.

To make a distinction between two perspectives makes perfect sense in ordinary experience and in science. We can talk about the table in front of us, and we can speak of the perspective I have on it and the perspective you have on it. There is nothing mysterious about this. In fact, even though I have just one perspective on it at any given time, under most conditions I can say with certainty how the table would look from any other perspective I (or you) could take up. From a certain perspective a circular shape will look elliptical, and from some perspectives a square may look like a trapezoid. Here it makes sense to speak of ‘a point from which objects … can be considered from one side or another’.

We can in such contexts say that some characteristics are not features of the objects ‘but a relation between viewer and objects’. But what sense does it make to speak of ‘one side or another’ when we know only two terms in what is essentially a three-term relation (e.g. x appears F to y). One might argue that for perspectival talk to be even possible one needs to be able to identify the different sides as perspectives (or perhaps reducible to perspectives), which, in the transcendental sense, would presuppose knowing something about both the appearance and the thing in itself.

Physicists during the early modern period took perspectival talk a few steps further than common sense. They found that their models of physical objects should not depend on the point of view of the modeller or observer, and they argued that physical description of such objects should be invariant with regard to rotations in space and differences in time. One might say that a perspective-invariant view of physical objects lies at the very basis of modern science and is largely unproblematic.

Yet, this is not the point at issue. Kant himself used the relatively unproblematic contrast between appearances and things in themselves only to introduce another distinction between an appearance and a thing in itself, which he thought was deeper or more fundamental. Thus he went on and asked us to ‘consider this empirical object in a general sense and … ask whether it represents an object in itself’ (A45). In other words, he wants us to consider the empirical thing in itself (or the perspective-invariant physical object) as an appearance in some other sense, and suggested that we may need to distinguish it from a thing in itself in a more fundamental sense. He said: if we consider the empirical thing in this sense, then the ‘question of the relation of the representation to the object is transcendental’. The raindrops in a sun-shower must be considered as appearances in the transcendental sense, and there must therefore be a thing in itself in a transcendental sense.

This is deeply problematic, since the transcendental distinction is not analogous to the empirical distinction. The empirical thing in itself can be known just as well as the empirical appearance. In some important sense, we might even say that what is really known is the empirical thing in itself, not the appearance; and in this case, it is really true that something is properly called a ‘mere appearance’ only after we have understood – or after we know – what the thing in itself is. But the object in the transcendental sense, Kant claims, ‘remains unknown to us’.Footnote 5

So why even dream of postulating it? It may well be true that we cannot speak of ‘appearances’ in the transcendental sense, unless we can also speak of ‘things in themselves’ in the transcendental sense. But why do we need to engage in transcendental talk? What do we lose, if we give it up? Why do we need to assume that there are two sides, one empirically real, the other merely conceivable – and completely unknowable and thus non-identifiable? Kant needs to give an answer to this question. And that answer would be more fundamental than the ‘absurdity argument’.

It appears to me that the objections Bird makes against the two world view apply with equal justice against the two standpoint theory. We ‘need to know why it is necessary to invoke references to’ a possible standpoint that allows us to consider something as a thing of which we can in principle have no knowledge. One might even say that speaking of a standpoint that forces one to take something as a thing in itself suggests a commitment to the existence and knowledge of those things, and then add that ‘Kant firmly and repeatedly denies any such commitments to things in themselves’, which would be contradictory. Indeed, Bird thinks that he needs arguments against such a view. Much of chapter 23 and many other passages present arguments against the claim that we must assume things in themselves as existing. All we need is their conceivability, or so the argument goes.Footnote 6

In any case, I hope to have shown that the inconceivability argument is insufficient as an answer to objections of this sort. One might also say that, being based on ‘mere grammar’, it cannot bear the weight Bird assigns to it. It should perhaps also be observed that ‘absurd conclusion’ (Guyer/Wood have ‘absurd proposition’) as a translation for ungereimter Satz is a bit strong. Literally, it means ‘non-rhyming sentence’, or claim that does not ‘fit’, or what an earlier generation of analytic philosophers would have called ‘queer’ or ‘odd’. ‘Questionable claim’ would have been a better rendering, since it is weaker.Footnote 7

None of these criticisms is necessarily fatal to the perspectival view or contradicts any of the fundamental tenets of Bird's view, though they do show that the supposed ‘simple formal argument’ to the effect that ‘if things in themselves were inconceivable then our empirical concept of “things that appear” would be impossible’ is not as strong as it might appear at first. The thing in itself, in any of the senses given so far, does not play any foundational role at all. It has nothing to do with the possibility of things that appear. I take myself to have shown only that the inconceivability argument or the absurdity claim cannot bear the weight Bird seems to assign to it in some parts of his book.

4

Bird himself suggests that the distinction between appearances and things in themselves in the transcendental sense is the result of Kant's conception of a priori forms of intuition. Indeed, it is, Kant claims, the necessary consequence of transcendental philosophy. To get to Kant's real motivation, we must, therefore, remember that transcendental distinctions arise only within the context of transcendental philosophy and are due to considerations of our ‘a priori concepts of objects in general’ and have rather little to do with general considerations of appearances and their relations to what appears ‘in them’. In the footnote to the claim that ‘we can know of things a priori only what we ourselves have put into them’ at Bxviii, Kant tries to make this clear. He claims that he is engaged in a peculiar experiment, which ‘admit[s] of no test by experiment with their objects (as in natural science)’. Here, he says, we can experiment ‘only with concepts and principles that we assume a priori by arranging the latter so that the same objects can be considered from two different sides, on the one side as objects of the senses and the understanding for experience, and on the other side as objects that are merely thought’. Kant suggests that if we ‘now find that there is agreement with the principle of pure reason when things are considered from this twofold standpoint, but that an unavoidable conflict of reason with itself arises with a single standpoint, then the experiment decides for the correctness of that distinction’. This seems to me to point to Kant's real motivation for the distinction between things in themselves. I will not here address the issue whether a ‘twofold standpoint’ (doppelter Gesichtspunkt) is the same as ‘two standpoints’, and simply grant that this passage supports the perspectival view.

Kant further develops this point in other passages, especially in his discussion of what I have called the ‘boundary principle’ in the Prolegomena. Kant there argues that it is necessary to supplement ‘Hume's principle’ with a principle that Hume quite overlooked, namely the principle that claims that experience cannot be all there is, and that the ‘field of possible experience does not limit itself in the eyes of reason’, but must be formed by something that, though itself unknown, shows that the world of the senses does not exhaust reality (4: 360–1).

Kant finds:

Though ‘the nature of these intelligible beings’ is for us an empty space that cannot be determined in any way, it is important for reason to expand up to the boundary, for in doing so it ‘partakes of real, positive knowledge’, as long as we do not try to go beyond the boundary. Here we cannot ‘think things but only forms of things (zwar Formen zu Dingen, aber keine Dinge selbst)’ (4: 361). Thinking the forms of things is not merely negative, but itself something positive.

Kant speaks in this context of a ‘standpoint of reason’, but this standpoint is clearly determined by the attempt to set the boundaries of the sensible world by a field or realm of intelligible beings that lie completely outside the sensible world. One might also say that what Kant is trying to demolish is the natural attitude, which simply identifies what we see with the things in themselves by ignoring the transcendental perspective.Footnote 8 Whatever the relation between appearances and things in themselves in the transcendental sense may be, it also has to do with this discussion of a boundary principle or the limits of the sensible world. It seems that it is our duty as interpreters of Kant that we should at least try to make sense of claims such as this, which at the very least go to show that talk of two worlds, namely the sensible world as opposed to a non-sensible world, is closely connected for Kant with talk about two standpoints.

I believe this shows that when Kant talks of things in themselves, his main goal is not just to deny that we know things in themselves, but just as importantly to assert that the things we perceive are precisely not the things in themselves. Kant does not just assert that space and time are forms of intuition, but also that things in themselves are most definitely non-spatial and non-temporal, and that therefore they open up a conceptual space that goes beyond what is merely sensed. Though as far as knowledge is concerned this space remains empty for us, we must admit that there is another possible world.

I would consider Kant's defence of the boundary principle to be more fundamental than the inconceivability argument. Since it is essentially concerned with the boundary between the sensible and a non-sensible world, it is also the answer to the question: ‘How, or why, do we come to think of a world of things in themselves?’Footnote 9 In other words, the two worlds theory should not be thrown out so quickly.

When I speak of the ‘two world view’, I do not mean to ascribe to Kant an idealist dualism of mental states, ‘ideas’, and transcendent things in themselves. Nor do I intend to accuse Kant of inevitable tensions and contradictions. Nor do I wish to claim that the perspectival interpretation is false, or that the two world theory is more basic than the two standpoint theory. Rather, I am interested in expounding Kant's own view and making it as plausible as possible. I would like to make sense of such passages as A249, where Kant speaks of ‘the division of objects into phenomena and noumena, [and] thus also a world of the senses and of the understanding (mundus sensibilis & intelligibilis) … in such a way that the difference here would … concern … the difference between how they can originally be given to our cognition’ (A249).Footnote 10

There are many other such passages in the Critique of Pure Reason. Especially the chapters on the Antinomies are full of them. Kant speaks of the Sinnenwelt and appearances, contrasting this world at times with an intelligible world.Footnote 11 This is also true of the Prolegomena.Footnote 12 But most important are the passages in the Groundwork and the Critique of Practical Reason where Kant speaks of the ‘world of sense’, the ‘world of the understanding’, and claims that the world of the understanding forms the basis of the world of sense.Footnote 13

Central in this is the passage in the Groundwork where Kant discusses the supposed circle in the discussion of freedom: ‘We take ourselves as free in the order of efficient causes in order to think ourselves under moral laws in the order of ends; and we afterwards think ourselves as subject to these laws because we have ascribed to ourselves freedom of the will.’ And he then inquires ‘whether we do not take a different standpoint’ in the two cases. Kant's answer is that we do, and that we must view or ‘count’ ourselves as members of two worlds, namely the world of sense and the ‘intellectual world, of which, however [we have] no further cognizance’ (Kant Reference Kant1996; 4: 451). As rational beings, we belong to the ‘intelligible world’, or so Kant claims. Ultimately the transcendental distinction between things in themselves and appearances is the consequence of two other doctrines: the theory of space and time as mere forms of sensible intuition and the doctrine of freedom as moral autonomy.Footnote 14

I cannot here talk about Kant's arguments for this view, but only point out that talk about ‘standpoints’ or perspectives is inseparably bound up for Kant with talk about ‘worlds’. If talk about two worlds is ‘thoroughly misleading and should be given up’, then very significant, even central chunks of his moral philosophy also have to be given up, and this, I would hope, cannot be the result of a close reading of Kant's Critique, for, whatever else this work may have meant to Kant, it was also a preparation for his moral philosophy. So, I would submit that the account of the contrast between appearances and things in themselves be revised to take this into account. Whether this makes a difference to some of Bird's more subtle points about a ‘fourfold distinction’, which involves ‘a complex combination’ of ‘object’, ‘idea’, ‘subjective’ and ‘mind-(in)dependence’ in ‘different transcendental and empirical senses’ (182), I dare not decisively answer here. But it appears to me that it need not.