The Radical Unknowability of the Thing in Itself

 

by Stephen Palmquist (stevepq@hkbu.edu.hk)

 

 


So the question whether something is a cognizable entity or not, is a question which touches, not the possibility of the things them­selves, but the possibility of our knowledge of them. [Kt7:467]

 

A. Transcendental Arguments for the Thing in Itself

 

      Few commentators (if any) would question Schrader's poignant obser­vation that 'the doctrine of the thing in itself presents the single greatest stumbling-block in the Kantian philosophy' [S5:49]. Understanding what Kant meant by the doctrine-i.e., the role it plays both in his overall System and in his transcendental idealism-can help prevent it from being discarded 'as a per­versity' [49], inasmuch as it can be interpreted in such a way that it makes quite good sense [see VI.2]. Yet even the most coherent interpretation could not prevent the philosopher who demands knowledge from 'stumbling' over it; for, accord­ing to Kant, the thing in itself is by definition unknow­able. In V.3 we saw, however, that there is one alternative to faith as the ultimate jus­tification for its employment which, if successful, would satisfy even the most persistent skeptic: viz., to justify the thing in itself by constructing a valid transcendental argument for the necessity of its existence. Since any ap­peal to faith would thereby be rendered superflu­ous, we must now examine more carefully the possibility of realizing this goal. For, although Kant him­self did not believe he required such a transcendental argu­ment, it may be pos­sible to reconstruct his System on the basis of a slightly different presup­posi­tion, such as that the thing in itself can, in fact, be proved to exist and to have certain knowable characteristics. Hence, in this Appendix I shall analyze the logical consistency of an affirmative answer to the meta­critical question: Is the thing in itself knowable? Our answer to this question will inevitably determine to a large extent how we should approach the task of interpreting the elements of Kant's Critical philosophy (as in Part Three), so it is important to deal with it seriously.

 

     Walker's recent arguments in favor of the knowability of the thing in itself can serve as a sounding board for our discussion of this issue.[1] He insists in W3:3 that 'even the most dedicated Kantian must admit that ... there is no reason in principle why transcendental arguments should not establish conclusions about how the world must really be, and not just about the phe­nomenal world ... Here Kant was simply mistaken.' He frequently alludes to the same position in W2 [e.g., 122-3,126,134]; but his detailed elaboration and defense of just what this view entails occurs in W4, in which he argues that 'Kant never saw ... that there is no general reason why truths established [by transcenden­tal arguments] should always have to be read into the world: instead they may be conditions which must hold independently of us [and thus, must be true of the thing in itself], but without which experience would be impossible for us' [20]. After analyzing the implications of Kant's notion of 'synthesis', he concludes:

 

All we can infer is that things in themselves must exist and must (therefore) have properties; and, no doubt, that some of them are subjects of experience. This much knowledge about the noumenal world is indispensable for tran­scendental idealism, if it is to retain its transcendental character and keep itself distinct from idealism of a more radical, but less satisfactory, kind. [27]

 

      Walker's position is enigmatic, however, inasmuch as it is difficult to understand how anyone can both suggest that the thing in itself is to any ex­tent knowable (in the Kantian sense of the word [see below]) and also unequivocally call oneself a Kantian. In order to bring to light the discrep­ancy between these two views it is necessary neither to presuppose an exhaus­tive understanding of the intricacies of Kant's doctrine of the thing in itself [see VI.2], nor to discuss thoroughly the well-worn issue of the nature and limita­tions of transcen­den­tal arguments. For the inadequacy of the position Walker represents can be established independently of a detailed knowledge of the elements of Kant's System. Accordingly, I will first examine the validity of the claims Walker believes would be most likely to constitute knowledge of the thing in itself, and I will then attempt to determine why he regards his revisionary interpreta­tion of Kant's doctrine as so obviously compatible with the re­main­der of Kant's philosophy. My goal will be to demonstrate that one cannot, in fact, both claim to be a 'dedicated Kantian' and maintain that the thing in itself is knowable 'in principle'-if, that is, one wishes to remain self-consistent.

 

B. The Status of the Four Basic Knowledge-Claims

 

      The four most eligible candidates for election to the honorable status of 'knowledge of the thing in itself' can be expressed in the following proposi­tions:

 

      (1)  The thing in itself exists.

      (2)  I am a thing (or self) in itself.

      (3)  The thing in itself has spatio-temporal properties.

(4)  The thing in itself has properties conforming to the principles of pure understanding.

 

All other possible knowledge-claims could be subsumed under one or more of these fundamental assertions. Therefore, it will suffice for our purposes to examine the merits of each of these four in turn.

 

      Before we can determine whether or not (1) constitutes knowledge of the thing in itself, we must take careful notice of what the word 'knowledge' (Erkenntnis) means for Kant. He uses this word in a number of different ways [see IV.1-2]; but he is clearly referring to empirical knowledge when he limits knowledge to phenomenal objects. And he repeated­ly defines empirical knowl­edge as consisting of judgments in which 'concepts' and 'intuitions' are syn­the­sized with each other [see VII.3.A and Kt22:209-10]. This is the point, for example, of his famous claim that 'Thoughts without content [i.e., con­cepts without intuitions] are empty, intuitions without concepts are blind' [Kt1: 75]. Kant's various transcendental arguments, however, are intended to yield knowledge of the 'synthetic a priori' conditions for the possibility of ex­peri­ence (i.e., of empirical knowledge); and it is apparently this transcendental type of knowledge which Walker thinks can in principle be gained in regards to the thing in itself. (Other kinds of transcendental knowledge, such as the type Kant thinks is manifested in Euclidean geometry, can be ignored at this point [but see Pq14].) Like empirical knowledge, transcendental knowledge must also be connected with intuition, otherwise it could not be 'synthetic' [see IV.2-3; Kt21:164; S7:171]. A transcendental argument can make this connection, and thus yield knowledge, only if the concepts with which it deals are 'schema­tized'-i.e., only if they are made to be 'homogeneous on the one hand with the cate­gory, and on the other with the appearance' [Kt1:177]. Therefore, if such an argument is to provide knowledge of the thing in itself, it must syn­thesize an intuition-or at least reveal how an intuition can be synthesized-with the con­cept of the thing in itself.[2]

 

      A transcendental argument for the existence of the thing in itself would maintain that experience is possible only if something real, to which this concept refers, can be said to 'exist' (in some loose sense of the word [cf. Ap. VI]), and that, since experience undoubtedly is possible, the thing in itself must exist. That Kant would support some such argument is evident from the fact that what little argument he does give in support of the thing in itself tends to take the form of a transcendental-type argument. Any doubt about this was dispelled with the help of Chipman in V.3. So we can agree with both Kant and Walker that something along the lines of a transcendental argu­ment can provide a good reason for presupposing the reality of the thing in itself. But we cannot agree with Walker when he parts company with Kant by in­ferring from this that a bit of knowledge about the thing in itself can there­by be gained. For, in spite of his use of transcendental-type argumenta­tion on its behalf, Kant does not believe he has (or could gain) any knowl­edge of the thing in itself. On the contrary, he always stresses that the concept 'merely characterizes a something in so far as it should be distinguished from objects of the senses', and of which therefore 'we cannot have the least knowl­edge': 'for us [it] remains entirely unknowable' [Kt22:209; cf. Kt1:45].

 

      But if Kant does not think a transcendental argument (or indeed, anything resembling one) can yield knowledge of the thing in itself, then what is such an argument supposed to prove when used to support this concept? Does it at­tempt to prove that the thing in itself is a necessary condition for the possi­bility of experience in the same sense as Kant's three 'analogies' (also proved by transcendental arguments [Kt1:218-65]) are supposed to be? I think not; for the analogies-and indeed, all his categorial principles-constitute, or de­fine, what ex­peri­ence is, whereas the thing in itself has to do with what expe­ri­ence is not. Unknowability is a covert, yet analytic ingredient in the concept of the thing in itself:[3] to talk about the thing in itself which is knowable is like talking about the 'red' which is not a color. So the most any argument can prove conclusively about the thing in itself is that it is indi­rectly related to the possibility of experience-that is, that the thing in it­self is the neces­sary starting point which must be presupposed if transcenden­tal re­flection is to be possible.[4] The difference in arguing about the status of the principles is that they are related to intuition (via the schematism of the cate­gories), whereas the thing in itself by definition cannot be related in any posi­tive sense to intuition, at least of the human type [see Kt1:72,306-9 and VII.2.A].

 

      Rather than proving the objective validity of the thing in itself, Kant in­tends his arguments to prove the rational coherence, and thus to secure the subjective validity, of presupposing the 'existence' of the thing in itself. He clarifies the difference between these two types of truth in Kt1:850, where he says only a judgment which is both objective­ly and subjectively valid consti­tutes knowledge (Wissen), while one which is merely subjectively valid con­stitutes belief (Glauben). Such belief, however, is by no means groundless, as we have seen in Chapter V. At present it will suffice to recall that such faith, which is required for the acceptance of the unknowable thing in itself, is simi­lar to-though not the same as-that required for the acceptance of the ideas of reason, and employed by Kant throughout systemp. Walker himself gives some attention in W2 to the role of belief, but he limits his discussion to empirical belief-i.e., to beliefs about what we can exper­i­ence [122-5]. As a result, he inevitably concludes, reference to belief 'is not good enough. The argument is not transcendental' [122].

 

      Therefore, although Walker is right when he says Kant admits 'cer­tain con­clusions about the independent character of things in themselves can be established by transcendental[-type] arguments' [W4:20], he is nevertheless mistaken to think such conclusions can be regarded in any Kantian sense as constituting knowledge. For a transcendental argument would be no more ca­pable of providing knowledge of the thing in itself than an ontological argu­ment could provide knowledge of God [see X.2 and note V.7]: in both cases the most such arguments could accomplish with respect to their unknowable subject matter would be to encourage rational faith in the reality of something corresponding to the concept, in spite of its lack of intuitive content.

 

      Having now demonstrated the incoherence of the claim that the Kantian must affirm that (1) can in principle constitute knowledge, it hardly seems necessary to apply our findings to the other three propositions; for if I am right in ascribing an analytic status to the proposition 'The thing in itself is unknowable',[5] then all propositions claiming to convey knowledge of the thing in itself are ipso facto meaningless, if not false. Nevertheless, a brief account of the other three will point up the radically unkantian nature of such claims.

 

      That Walker would insist that something like (2) can in principle consti­tute knowledge of the thing in itself can be unambiguously infer­red from his statements quoted at the beginning of this Appendix [s.e. W4:27; cf. W2:123, 133-4]. This contrasts sharply, however, with Kant's repeated warn­ing that 'we know our own subject only as appearance, not as it is in itself' [Kt1: 156e.a.]. (The only apparent exception to this-the self-awareness pro­duced by 'the synthetic original unity of apper­ception'-is said by Kant to be 'a thought, not an intuition' [157], so it does not constitute knowledge. This point, which Kant himself empha­sizes, seems to be missed by Walker when he states that in this passage 'The flood-gates are opened; the principle that I can have no knowledge regarding things in themselves has been breached' [W2:134].) Kant would want to add, of course, that some sort of transcendent self must be pos­tulated in thought in order to help bring 'systematic unity' [Kt1:596] into the structure of one's overall philosophy; but he would not interpret this to mean such a self as a real object is a synthetic a priori con­di­tion for the possibility of empirical knowledge (i.e., is tran­scenden­tally knowable). The existence of a transcendent self is neither em­piri­cally nor transcendentally knowable, because it can never be anything for us but a non-intuitable concept in which we may or may not believe.

 

      That Walker would insist that something like (3) and (4) can constitute knowledge of the thing in itself is implied in his claim, quoted above, that 'there is no general reason why truths established [by transcendental argu­ments] should always have to be read into the world ...': for he thinks space, time and the categories could, in principle, be proved to be properties inhering in the thing in itself, rather than a priori forms of knowledge inhering in the knowing subject.[6] In both cases Walker is closer to Kant's position than it might seem at this point; yet in both he allows one step further to be taken than Kant would allow.

 

      Concerning (3) Kant admits, as Allison points out, 'that space and time have objective as well as subjective grounds, and that their ulti­mate objective grounds are things in themselves', which, however, 'are not themselves in space and time and are therefore unknowable' [A7:93; cf. Kt22:207]. If space and time are grounded objectively in the thing in itself, then the latter must be conceived of as having some­thing analogous to the spatio-temporal proper­ties of phenomenal objects. But to refer to this unknown something with the word 'properties', as Walker does [W4:27q.a.], is misleading, if not contra­dic­tory, since the word 'properties' can be meaningful to us only if it refers to an object with intuitive content-which it does not when it refers to the thing in itself.[7] (Walker's contention that 'the conception of things in them­selves without any properties is simply absurd' [23] seems to result from a misin­terpretation of Kant's theory of perception, which we shall discuss in the next section.) Thus, although Kant would never agree that (3) can consti­tute knowledge, he might affirm the conceptual consistency and subjective va­lidity of the less presumptuous assertion:

 

(3') The thing in itself has something analogous to the spatio-temporal properties of phenomenal objects.

 

      Similarly, with respect to (4), Kant does not deny that we can use the categories to think about the thing in itself: what he denies is that we can ap­ply them as principles in such a way as to produce knowledge. To suppose we can apply them in the latter sense would be to ignore Kant's view that the categories can produce knowledge only when used in connection with intu­ition and applied to an empirical object as represented. Thus, whereas Walker would presumably affirm the possibi­lity of knowing (4) to be true (as long as a sufficiently rigorous tran­scendental argument were formulated on its behalf), Kant would affirm no more than the conceptual consistency and subjective validity of:

 

(4') Thinking about the nature of the thing in itself makes use of our cat­egories of thought (but produces no knowledge).

 

      None of the four basic propositions cited at the beginning of this section has proved to be a legitimate candidate for knowledge of the thing in itself, unless we are prepared to make a radical departure from the meaning Kant gives to his terms. Yet there is no need to take this unkantian alternative, for as Findlay says, 'in such cases we do not, according to Kant's usage, have knowledge, only a rational presupposi­tion of knowledge. There is no reason to reform Kant's usage, provided we clearly understand it' [F3:8]. The quandary which has yet to be resolved is why a 'Kantian' philosopher such as Walker should insist on the legitimacy of an obviously unkantian position. Therefore, to this problem I shall now turn.

 

C. Three Common Interpretive Errors

 

      The claims argued against in the previous section would appear to be legitimate to an interpreter who first commits one or more of three common er­rors. The first is an equivocal use of the word 'knowledge'; the second is a misapplication of transcendental arguments; and the third is a (covertly) unkantian rendition of Kant's theory of percep­tion. As we shall see, each of these stems from a failure to distinguish clearly between Kant's transcendental and empirical perspectives. Although we have already looked briefly at the first two errors, a closer look at the extent of Walker's commitment to each is in order here, inasmuch as he is by no means alone in his tendencies.

 

      A popular way of defining 'knowledge' nowadays is in terms of 'justified true belief' [see e.g., F4:1.10; K4:35; M8:56; W18:516n]. Helpful though this may be in many contexts, it can be a dangerous stumbling block to a clear understanding of Kant's philosophy. For when Kant de­fines the thing in itself as unknowable, he does not intend to preclude justified true belief in its reality. On the contrary, he is affirming that knowledge in this sense-as the subjective conviction that one has sufficient rational reasons for believing cer­tain proposi­tions to express truth about the way things are-is not only pos­sible, but indispens­able to the Critical philosopher. When he says he must 'deny knowledge, in order to make room for faith' [Kt1:xxx], he is acknowl­edg­ing the necessity of differentiating between what is known through the synthesis of intuition and conception and what cannot possibly be known (because it cannot be intuited [265]), but neverthe­less can become an object of justified true belief.

 

      When Walker refers to the 'knowledge about the noumenal world [which] is indispensable for transcendental idealism' [W4:27] he may unintentionally be using 'knowledge' in its popular contemporary sense, in which case Kant would be largely in agreement with him: (1), (2), (3') and (4') are the justified true beliefs which Kant would regard as being indispensable to his transcen­dental idealism. Only in this sense could Walker be consistent (though not entirely accurate) in saying: 'Officially all that Kant allows we can know about things in themselves is that they are somehow the source of the data we receive in intuition' [18e.a.]. For if 'know' in this context is intended in its strict Kantian sense, then Walker's claim is simply not true: as we have seen, the thing in itself is a matter of rational belief, not knowledge. Indeed, the only actual knowledge Kant would regard as indispensable for tran­scen­den­tal idealism would be the transcendental knowledge of the neces­sity of the 'principles of pure understanding' for empirical knowledge (i.e., the neces­sity of synthesizing the pure intuitions of space and time with the categories [see VII.3.A]). So Walker's supposition that he has shown Kant's limitation of knowledge in its strict sense to be untenable would in this case simply re­veal an equivocation in Walker's use of Kant's terms. This may well be the full explanation of why he makes this supposition;[8] but I suspect it only scratches the surface. For the second of the above-mentioned errors seems to take us closer to Walker's main intentions.

 

      Anyone who holds a position such as Walker's, yet is not misled by the various meanings of the word 'knowledge', must believe some sort of intu­itive data concerning the thing in itself is actually available (perhaps by using transcendental arguments). For only in this case would one's use of the word 'knowledge' in reference to the claims we are con­sidering be consistent with Kant's usage.[9] But what could it mean for a transcendental argument to con­nect the concept of the thing in itself with a corresponding intuition? The thing in itself is by definition not intuitable. Or, if it is associated with that which is presented to us in intuition-as some interpreters would suggest-then it would in so doing become an appearance for us [see VI.2]. In the pre­vious section I have already questioned the legitimacy of extending the appli­cation of transcendental arguments to the thing in itself by pointing out that Kant limited their use to the proof of transcendental (as opposed to transcen­dent) truths. Transcendental arguments cannot prove anything about 'condi­tions which must hold independently of us' as Walker seems to think they can [W4:20], because by 'independent of us' Kant would mean 'independent of our knowledge'; and if we suddenly knew about such con­di­tions, they would no longer be independent (in Kant's transcenden­tal sense). The only way one could continue to maintain the legitimacy of such an ex­tension of the scope of transcendental arguments would be by regarding the thing in itself-as well as the process of perception in general-in a radically unkantian manner. This third suggestion takes us, I believe, to the very heart of Walker's divergence from Kant.

 

      Walker reconstructs Kant's theory of perception and of the role of the thing in itself by analyzing 'the character of the given' in Kant's system [W4: 23]. In particular, he investigates 'the relationship that may obtain be­tween the properties of things in themselves and the quali­ties we ascribe to things' [24]. But throughout his discussion he treats the thing in itself in much the same way as Kant treats the appearance when viewing it from the empirical perspective. (Kant's use of the term 'appearance' is thoroughly discussed in VI.2 and VII.2.A; it refers to 'the undetermined object of an em­pirical intuition' [Kt1:34], which is 'outside us' when regarded from an em­pir­ical per­spective [A373] even though it is 'only in us' when regarded from a tran­scen­dental perspec­tive [59], and which is the root cause of our most prim­itive awareness of the material sources of our empirical knowledge [207].) In one of his clearest explanations of the 'given data' with which the transcen­den­tal philosopher deals, Kant says: '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. But these appearances are not things in them­selves; they are only representations' [A108-9]. In sharp contrast to this warn­ing, Walker's reconstruction of Kant's theory, as I shall demonstrate, ig­nores the part played by such appear­ances and instead treats their 'given data' as a prop­erty of the thing in itself.

 

      Walker points out, quite correctly, that 'the data given to us in intuition must possess a character of their own' [W4:18]. But he then intimates that this 'character of their own', which has a 'substantive role to play in deter­min­ing the character of the [empirical] world of appearances', is derived from 'things in themselves' [21], not from appearances. Later he makes the same point rather more explicitly: 'given the ways of ordering [or synthesis] that we do use, what the world of appearances turns out to be like depends on the character of things in themselves' [23]. The vital role of appearances as such in perception is wholly ignored, to the extent that he considers synthesis-which for Kant concerns a manifold of intuited appearances-to be concerned directly with the thing in itself. But what he is really doing, I suggest, is redefining the thing in itself so that it takes over the role Kant assigns to the appearance in his theory of perception. This explains why Walker thinks it 'is simply absurd' to conceive of 'things in themselves without any properties' [23]: for Kant the appearance does have properties, and these are what we map 'into our own quality space' [24] through empirical synthesis. Walker, how­ever, declares that empirical synthesis is 'guided by' the properties of the thing in itself, and that it 'can be thought as governed by a function which maps the intrinsic properties things have in themselves into our own quality space' [24].

 

      The reason Walker himself gives for putting forward this view is com­mendable: he urges that unless 'some relationship does obtain'-i.e., can be known to obtain-between the thing in itself and objects in the phenomenal world, there will be nothing to prevent 'the given element' in perception from being 'dropped out as otiose' [W4:25], as in 'more radical' forms of ideal­ism. What he misunderstands is that this 'given element' is for Kant the ap­pear­ance: Kant would agree that some relationship must obtain between it and the object as empirically known (the phenomenon); but this frees the thing in itself to fulfill quite a distinct function as the rationally presupposed starting point for the transcendental perspective. Thus Kant would agree that we have good reasons to believe, as Walker says, that 'What happens in the world of appearances is dictated not by ourselves alone but by the character of the an sich' [25]; but that this is true can never be a matter of knowledge for us-not even transcendental knowledge. What could be a matter of knowledge-and it is this well-known Kantian view which I suspect Walker is actually leaning towards in the above suggestion-is that empirical knowledge must be com­posed not only of the form of experience (i.e., space, time and the categories), but also of its matter (i.e., appearances).

 

      If this criticism of Walker's interpretation of Kant is accurate, then such an approach actually turns out not to extend the scope of transcendental argu­ments at all: it merely affirms (covertly) that they can be applied in just the way Kant applies them (viz., as yielding conclusions about empirical objects). The trouble is that in so doing it inevitably uses Kantian terms in unkantian ways-most notably, 'thing in itself' in place of 'appearance'. Perhaps in the end this muddled reconstruction of Kant's theory of perception is responsible for Walker's willingness to declare that various (patently unkantian) asser­tions must be affirmed by 'even the most dedicated Kantian'.

 

D. Caveat and Conclusion

 

      The foregoing criticism of Walker's interpretation of Kant has had to be rather selective, and hence has not been entirely fair to his gener­al approach. Walker readily admits, for example, that although the thing in itself should 'in principle' be regarded as knowable, he him­self is unable to construct a transcendental argument which can demon­strate the validity of such knowl­edge. Thus, it could even be that when he says 'in principle' he means some­thing like 'from God's perspective', in which case his position would be (for us mortals) irrefutable. But it seems unlikely that this is his intention, since he makes it fairly clear that he simply wants to acknowledge that Strawson and others have come closer than Kant thought possible to estab­lishing the valid­ity of certain claims, and that, however 'improbable' it might be [W2: 126], there is nothing which necessarily prevents some future philoso­pher or logi­cian from going even further. In any case, the bulk of Walker's inter­pre­ta­tion suffers surprisingly little from the views criticized above. The main reason for concentrating on them here has been to clarify the nature of Kant's most basic presupposition, and in so doing, to point up the danger of reject­ing and/or revising his theories without first adopting such presupposi­tions as one's own, so that the precise meanings he gives to his many techni­cal terms can be accurately understood. But a sec­ondary reason has been to purge Walker's approach of those features whose incon­sis­tency renders them unable to fit coherently into his otherwise helpful inter­pretation of Kant's philoso­phy. My focus on these features, therefore, is ultimately intended to render his interpretation more tenable, rather than to dis­pense with it com­pletely.

 

      One does not have to be an expert on the formulation and/or criticism of transcendental arguments to see their inapplicability to the thing in itself. For even a partial understanding of their form and the scope of their applica­tion is sufficient to reveal that for an expert transcendental logician to try to extend his proofs to cover that which is transcendent is as thoroughly unkan­tian as it would be for him to try to prove by transcendental arguments that, for exam­ple, the logical law of noncontradiction is true, or that God exists. As I have tried to demonstrate, transcendental arguments can be applied properly only to prove that certain principles instantiated in empirical objects are in fact 'synthetic a priori conditions for the possibility of experience'. So the only way to maintain that such arguments can be extended to cover transcen­dent reality is to contend that we grasp transcendent reality itself in empirical knowledge; yet this entails a denial of the validity of 'transcendental ideal­ism', and with it, the label 'dedicated Kantian'.

 

     The essence of my criticism of such an approach, at least when it is con­nected in any way with Kant's philosophy, is that it requires the acceptance of two incompatible positions. It tries to allow for some form of direct, know­able contact with transcendent reality, and yet to defend transcendental ideal­ism (which itself is the view that nothing we know empirically is 'ultimate' in this transcendental sense). To render these positions compatible would in­volve at the very least the clarifi­ca­tion of those views which lack coherence even if they are not regarded as Kantian: viz., an equivocal use of the word 'knowledge', an extension of the limits Kant puts on transcendental argu­ments, and the association of a theory of perception such as Walker's with the name 'transcendental idealism'. But even if somehow they were to be made compatible, I cannot see how anyone holding them could legitimately be regarded as a Kantian. For the only Kantian answer to the question 'Is the thing in itself knowable?' is that it must be radically unknowable. Indeed, anyone who wishes to insist otherwise is likely to find a friendlier atmo­sphere in the rationalism of (for instance) the Leibnizian camp, or in the empiricism of (for instance) the more recent phenomenalist camp, than in the strange borderlands occupied by the transcendental idealism of the Kantian camp.

 


 [1] Other attempts to make sense of the notion of the thing in itself, such as Prauss'  stress on the phrase 'Dinge an sich selbst betrachtet' [see e.g., P13:225; cf. note V.2]orHintikka'sreconstructionintermsof'thestructureofour"language-games"ofseekingandfinding'[H17:88],mighthelptorenderitmoreacceptable;buttheyattempt more to correct than to justify Kant's presup­position of it in the more traditional, 'reified' way of inter­preting it [see Ap. VI]. As we shall see in the third section of this Appendix, the same holds true for Walker's reconstruction as well.

 [2]  Kant stresses the conceptual nature of the thing in itself when he says the Aes­thetic in Kt1 distinguishes 'between the knowledge of things as appearances and the conception of them according to what they are as things in themselves' [Kt22: 208e.a.]. Merely thinking about the thing in itself, of course, cannot pro­vide us with an intuition of it.

 [3] See IV.2 and VI.2. The unobvious nature of the analytic connection between un­knowa­bility and the thing in itself explains why Kant repeat­ed­ly emphasizes its ne­cessity. And its analytic nature explains why he never bothers to support it with a rigorous argument: his primary concern is with synthetic, not with ana­lytic, judgments [see A9:319-20 and P4:154].

 [4]  I call such an argument a 'transcendental-type argument', rather than simply a 'transcendental argument', because the latter technically deals not with the possi­bility of transcendental reflection, but only with the possibility of empirical knowledge.

 [5]  This proposition is not a 'judgment' in Kant's technical sense, for it does not connect an intuition with a concept [see VII.3.A and note IV.13].

 [6] See also W2:130. Although he regards this as a reasonable possibility, Walker him­self would not wish to describe the thing in itself as being in space and time [W3:25-6]. He does suggest, however, that Kant's proofs of the objective va­lidity of the principles might need to be extended to apply to the thing in itself as well as to the world of appearances if they are to succeed [26].

 [7] See above, note 2. When Kant says (arguing against Eberhard) that an in­tu­i­tion 'does not provide us with the properties which pertain to it as it is in itself'[Kt22: 209-10e.a.], he is not implying that there might be some other way for us to fix a ref­erence for such properties, and thus make them knowable (e.g., by ex­amining the simple constituents of appearances á la Eberhard, or by perfecting a transcen­dental argument á la Walker); on the contrary, he is implying that the question as to whether or not the thing in itself has 'properties' can for us be a mat­ter only for speculation, not knowledge.

            Walker comes close to this position when he says 'there must be some­thing in the things in themselves which corresponds to redness and is respon­sible for the similarity of our recurrent sensations of red' [W3:19]; but he gives the impres­sion that the 'must' means something more for him than 'must be believed to be', which iswhatitwouldmeanforKantinthiscon­text.AreviewofKant'sargu­ments    in the Postulates [Kt1:265-74] would re­veal the thoroughly unkantian na­ture of Walker's claim in W3:18e.a.: 'There are also conditions which the noume­nal world must satisfy if experience is to be possible for us.'

 [8]  If it is, then much of the remainder of this section is superfluous, inasmuch as Kant, Walker and I would all agree on everything except how loosely the word 'knowledge' should be used in Critical philosophy.

 [9]  The same could be said for Walker's use of the word 'must' [see above, note 7].

  

http://www.hkbu.edu.hk/%7Eppp/hline.gif

 

Send comments to the author: StevePq@hkbu.edu.hk

Back to the Table of Contents for this book.

Back to the listing of Steve Palmquist's published books.

Back to the main map of Steve Palmquist's web site.