225
THE TRANSCENDENTAL ARGUMENT IN KANT'S
Groundwork of the Metaphysic of Morals ROBERT J. BENTON
In the Groundwork of the Metaphysic of Morals, Chapter III, Kant presents an argument that he identifies as a deduction of an a priori synthetic judgment (namely, the categorical imperative) - that is, he presents a transcendental deduction. This argument has been much discussed, but there seems to be no agreement about whether the argument is a valid one. 1 (And it might be suspected that Kant was himself doubtful about the argument's validity, since in the second Critique he seems to treat the same subjectmatter with a totally different argument, apparently even reversing the premises and conclusions.) But before we can decide whether the argument is valid we must know what the argument is. And here it becomes important that the argument is an example of a transcendental argument: Transcendental arguments have recently been the subject of much debate, the primary issue being whether they are even possible. But this debate has been carried on in the absence of a consensus about what a transcendental argument is. r think a significant contribution to this debate can be made by presenting a concrete instance of a transcendental argument, namely, the deduction in the Groundwork. And, conversely, I think that reading this argument as a transcendental argument provides a way of seeing it in a broader perspective by reading it as being concerned with the possibility of there being objective value judgments at all in the field of practice. My aim here is not to give an exhaustive account of the argument (which would require a much more detailed analysis of the text), but only to suggest a way in which the chapter can be read as a coherent argument and thus to provide a basis for further discussion of its goals and its success or failure in achieving them. For that purpose I will make use of two theses concerning transcendental arguments: (1) that a transcendental argument is one that is concerned with problems of establishing a cognitive framework; and (2) that one thing that establishing such a framework involves, for Kant, is exhibiting a priori the conditions of unity and of limitation of faculties that are irreducibly distinct but whose cooperation is required for cognition. 2 The i Both Paton and Wolff consider the deduction to be a failure - see H. J. Paton, The Categorical Imperative, Harper Torchbook, 1967, pp. 244-5 "The Failure of the Deduction"; Robert Paul Wolff, The Autonomy of Reason, Harper Torchbooks, 1973,pp. 210-12. 2 The first thesis was suggested by Patricia Crawford in "Kant's Theory of Philosophical Proof," in Kant Studien 54, 257 (1963). The second was suggested by Dieter Henrich in "The Proof-Structure of Kant's Transcendental Deduction," Review of Metaphysics XXII, 640 (1969). Both of these articles are concerned with relating Kant's transcendental
226
The Journal o f Value Inquiry
application of the first thesis to the argument in the Groundwork means that rather than taking the aim of the argument as being simply the justification of an a priori synthetic judgment (the categorical imperative) - as is generally done - we must take the aim of the argument as being to establish a practical cognitive framework distinct from the theoretical viewpoint but referring to the same world as the theoretical viewpoint. The categorical imperative can then be justified, but only within the framework of practical cognition 9 And the application of the second thesis means that practical cognition, if it exists, like theoretical cognition, must involve two roots: reason and sensibility; but whereas the objective reality of theoretical concepts is demonstrated by intuitions, the objective reality of practical concepts would be demonstrated by will-determinations. (So that, while from the theoretical point of view, phenomena are viewed as possible objects of experience, from the practical viewpoint they would be viewed as possible objects of the will both viewpoints would contain the same phenomena, but the phenomena would be related by different laws.) I think that with this approach we can understand the structure of Kant's argument - a claim that can only be substantiated by our now actually going through the steps of the argument. But I also think that this interpretation is the only one that allows us to understand the original problem of the chapter, since, as I will now try to show, it is only in the context of a possible practical viewpoint that the categorical imperative can be seen as a synthetic j u d g m e n t .
HOW IS THE CATEGORICAL IMPERATIVE SYNTHETIC?
Kant claims, both in the Groundwork and in the second Critique, that the categorical imperative (or the principle that underlies it, the moral principle) is an a priori synthetic proposition; and it is precisely the synthetic character of the imperative that makes a transcendental deduction necessary. In the Groundwork, for example, Kant says the principle of morality is ... a synthetic proposition, namely, "An absolutely good will is one whose maxim can always have as its content itself considered as a universal law"; for we cannot discover this characteristic of its maxims by analyzing the concept of an absolutely good will (447).3 9
form of argument to his understanding of the finitude of the human intellect, a problem that is also discussed, with reference to the second Critique, in my work, "Kant's Second Critique and the Problem of Transcendental Arguements," Martinus Nijhoff, I977. a I use the following convention in referring to Kant's works: Numbers in parentheses refer to the Academy-edition page numbers of the Groundwork of the Metaphysic of Morals; the translations are Paton's with minor alterations. Other references are made as follows: Kr V - Critique of Pure Reason, first or second edition page numbers indicated by A or B. Kp V - Critique of Practical Reason, Academy page numbers. KU - Critique of Judgment, references are made to section numbers. Religion - Religion within the Limits of Reason Alone, page numbers are those of the Harper Torchbook edition of 1960.
Discussions
227
But this formulation of the synthetic character of the principle can be misleading, and in fact it seems untrue: In Chapter I of the Groundwork Kant does just what he says cannot be done - he derives the condition of willing maxims as laws analytically from the concept of an absolutely good will (although the absolutely good will is there thought only problematically). And, even without recapitulating that argument, we can easily show the principle to be analytic (when considered merely speculatively): If we abstract from the question of the existence of such a will and consider it merely as an object of thought, then "an absolutely good will" can only mean "a perfectly rational cause." The reason for that is that the concept of good, for pure speculative reason, is the concept of perfection (see Kr V B 113-114), and the concept of will is the concept of a rational cause (446). But a perfectly rational cause would be a cause that acts solely according to laws of reason - that is, solely according to universal laws. Therefore its maxims could always have as their content themselves considered as universal laws. (In fact, there would be no distinction between laws and maxims for such a will.) So the predicate of the moral principle, considered merely speculatively, can be analytically derived from the subject. Since the moral principle is analytic if we consider it as a purely speculative proposition, Kant's claim that the principle is synthetic must mean that he is considering it not as purely speculative but rather as possible knowledge - as making a claim to objective reality. In other words, the principle as Kant considers it in Chapter 3 must claim some link to existing reality. If the claim to objective reality were a theoretical claim, then the principle would have to be shown to be capable of being given intuitive content. The reason for that is that for Kant all knowledge (as opposed to mere thought) involves two irreducibly distinct roots: rationality and sensibility. And, in the case of theoretical knowledge, it is sensible intuition that gives objective reality to knowledge (whether that knowledge be empirical - experience - in which case the intuition is also empirical, or a priori, in which case the intuition is pure). And the transcendental problem for theoretical knowledge (and that which requires a transcendental argument) is to show how concepts and intuitions - which stem from irreducibly different faculties - are a priori capable of unity in one consciousness. So the objective validity of the categorical imperative could be proved for theoretical reason only by providing the imperative with intuitional content. And ultimately that would mean that the law and the good will would have to be objects of possible experience. But that is clearly impossible, since the categorical imperative presupposes freedom - that is, it presupposes an unconditioned causality, which can never appear in experience. This is the impossibility that prevents an explanation of freedom or the imperative (if by "explanation" we mean giving the cause of a thing) - the impossibility that Kant refers to repeatedly (for example, at 458-9). But this impossibility becomes something of a red herring for the understanding of the argument, since it obscures the fact that the real point of the argument is to show that the theoretical use of reason is not the only one that can have objective
228
The Journal of Value Inquiry
reality. Instead, reason must be shown to have a cognitive function in practice as well. In its practical function reason and its laws acquire objective reality not through intuition but rather through the will. A practical statement has objective reality not by being derived through experience but by containing an actual determination of the will. Practical statements of course contain references to experience, and they cannot contradict experience, but when we consider an assertion as practical we are concerned in the first place with its relation to the will and only secondarily with its relation to intuition. The practical viewpoint is established within the horizon of the "ought" rather than of the "is." So we can now see how the categorical imperative and the moral principle are synthetic: They are practically synthetic, in just the sense in which Kant uses the term "synthetic" in Chapter II of the Groundwork (420n) - that is, they contain a will-determination that is not derivable analytically from some prior will-determination. In the case of the categorical imperative that is clear, since it simply commands a will-determination unconditionally (i.e., not "if you want to be happy do x" but simply "so a c t . . . "). But even in the case of the moral principle we can see this synthetic character: From a practical viewpoint by the "absolutely good" we do not mean perfection. Instead, the absolutely good is that which is a necessary object of the will. So on this view the principle would assert that the property of being able to will one's maxims as at the same time universal laws is a (practically) necessary object of the will. 4 And that is a (practically) synthetic proposition. It is in this sense, then, that the categorical imperative is synthetic namely, that it contains an a priori linking of a determination of the will with a determination of pure reason. This problem of the a priori synthetic corresponds to the problem in the deduction in the first Critique, where the question is how concepts can "apply to" intuitions a priori. In neither case is it a question of proving a particular proposition - rather, it is a question of showing the conditions of the possibility of a cognitive framework. That is, it is a question of showing how it is possible that we have knowledge (practical knowledge in the former case and theoretical knowledge in the latter) at all, granted that knowledge in either case derives from two irreducibly distinct roots. So the claim that the moral principle is synthetic becomes intelligible if we postulate a practical viewpoint - a practical cognitive framework, whose possibility is to be explored by the argument. We now need to see concretely how the argument accomplishes that task.
4 I say practically necessary because theoretically it is not at all necessary that we will our maxims in that way. Nevertheless, Kant's claim is that it is apodictically certain that we ought to so will our maxims, and it is that categorical "ought" that indicates practical necessity.This brings out the more general point that all of our fundamental concepts (like that of necessity)must be rethought if they are to be applied to practical knowledge.
Discussions
229
THE ARGUMENT-STRUCTURE
The argument consists basically of two steps followed by the actual deduction, which combines the two steps into one argument. The first step (446-448) is an analytic argument that operates under an abstraction - that is, it abstracts from the nature of our will as a faculty of desire (as sensibly conditioned - that is, as subject to laws of nature) and considers only will as such (rational causality). It shows that (a) freedom implies the moral law; (b) we must think the will of all rational beings as being free; and (c) therefore we must think will as such as being subject to the moral law. The second step (448-453), also analytic, reinstates what was abstracted from in the first step and considers the will now as a faculty of desire subject to laws of nature. It shows that the argument of step 1 is circular if it is applied to a sensibly conditioned will unless there is a practical cognitive viewpoint distinct from the theoretical. In the first step the possibility of a practical viewpoint was introduced by Kant's abstracting from the requirements of theoretical knowledge and treating the concept of freedom only as an idea necessary for action. But this abstraction does not establish the objective reality of such a viewpoint. That is done in the second step by introducing our elementary awareness of the distinction between phenomena and noumena and arguing that from a practical viewpoint we must represent ourselves as phenomena (and therefore as subject to laws of nature) who nevertheless operate under noumenal laws. The two steps are then brought together and completed in a single deduction (453-455), which establishes the practical viewpoint by showing how it is possible for the practical viewpoint to include the laws of the theoretical viewpoint subordinated to those of the noumenal realm. For a better understanding of how the argument works we now need to look at the actual steps.
STEP
1:
WILL AS SUCH IS SUBJECT TO THE MORAL LAW
The first step has two parts (appearing under the headings "The Concept of Freedom is the Key to Explain Autonomy of the Will" and "Freedom Must be Presupposed as a Property of the Will of all Rational Beings"). In the first part Kant derives the moral law analytically from the concept of a free will, and in the second he derives freedom analytically from the concept of will as such. Taken together these two parts prove the necessity of thinking the moral law as the law of any will. But, as we shall see, Kant can derive the moral law analytically from the concept of a will only by abstracting from an essential feature of the human will (the feature that makes the will an object of experience and which thus gives the concept of will objective reality for us in the first place - cf. Religion 5n), namely, that it is a faculty of desire. So the argument, while it has analytic necessity, does not prove the objective validity of the moral law, which could only be done by taking into account the will's sensible nature.
230
The Journal of Value Inquiry
The abstraction that governs the first step occurs in the first sentence of the chapter, where will is presented simply as rational causality. That is immediately followed by a provisional "definition" of freedom as independence of determination by external causes (he calls this the negative concept of freedom). From these two definitions he shows, in the first part of step 1, that a free will (if it exists) could only be a will subject to the moral law. The argument given here differs only superficially from the one given in the second Critique (Kp V 29 - Section 6 Problem II): The will is always determined by rules. Those rules can determine the will either through their content (heteronomy - determination by alien causes under laws of nature) or through their form (autonomy - determination by reason itself). Heteronomy contradicts the concept of free rational causality. So the law for a free rational cause must be a law of the autonomy of reason - that is, it must be a law that commands the adoption of maxims because of their rational form rather than because of their content. But the rational form of a rule is simply the form of lawfulness as such (that is, reason's thought of a rule is simply the thought of abstract lawfulness, which includes universality and necessity). So the law for a free will is the law of adopting rules as maxims because of their ability to be willed as universal laws - that is, it is the moral law. So if a will is free it is subject to the moral law. But the conjunction of freedom and will as rational causality must itself be justified. Kant does that in the second part of the first step by showing that freedom can be analytically derived from the concept of rational causality. For the moment I will ignore the extra assertion, namely, that "every being who cannot act except under the idea of freedom is by this alone - from a practical point of view really free (488)." Apart from that assertion the argument is analytic. What is taken as given is will as rational cause (it is not assumed that pure reason can be a cause, but only that reason can be a cause at all, which is a fact ascertainable by experience in the case of the human will). And the argument shows that we cannot think reason as a cause at all unless we think it solely as a spontaneously self-determining cause. The argument is closely analogous to the argument in the B-edition deduction of the first Critique where Kant derives the principle of the transcendental unity of apperception. In fact, the relation between the two arguments is more than mere analogy: It is the same unity of apperception that is being analyzed in each case. All theoretical cognitions presuppose the "I think" - the consciousness of the unity of function in the synthesis of a manifold. And this consciousness is of the spontaneity of the synthesis - it is the consciousness that the original unity of apperception is precisely the unity of the functioning of the understanding in its synthesizing, not a unity that could be somehow given from without and of which we could be passively aware. So, likewise, if reason determines the will, the consciousness of that determination can only be of reason's spontaneous activity, because what is in question at this point is not the function of theoretical reason or of practical reason but rather of reason
Discussions
231
as such. Reaon's activity is a pure spontaneity (this and its unity are what distinguishes reason from sensibility, which is always passively determined and always presents a manifold given). So the thought of the determination of the will by reason at all necessarily implies the thought of the determination of the will bypure reason. It is for this reason that Kant can assert . . . we c a n n o t possibly conceive o f a r e a s o n outside in regard to its j u d g m e n t s ; for in that d e t e r m i n a t i o n o f his power o f j u d g m e n t , n o t to m u s t look u p o n itself as t h e a u t h o r o f its o w n fluences (448).
as being consciously directed f r o m case t h e subject w o u l d attribute t h e reason, b u t to a n impulsion. R e a s o n principles i n d e p e n d e n t l y o f alien in-
The conclusion of this argument, therefore, is that a rational cause can only be represented (by us or by itself) as being independent of alien determination - that is, as being free. The two parts of the first step, taken together, would seem to prove that the moral law necessarily applies to any will insofar as it is rational. And the proof should be apodictically certain since it is an analytic argument. But the proof contains a restriction, and that restriction is of decisive importance: The proof is restricted to a "practical viewpoint." In the first instance that means only that we have abstracted from the requirements of theoretical reason - in particular we have abstracted from the requirement that all our concepts be of objects of possible experience. That abstraction is absolutely necessary, since from the theoretical viewpoint the concept of freedom cannot be given intuitive content and is therefore incapable of being experienced. But the theoretical viewpoint is, so far, the only viewpoint we have from which knowledge is possible. Therefore all the results of the first step are pure speculation. And our next task is to show how they can be knowledge not theoretical knowledge, but practical knowledge.
STEP
2: WILL AS FACULTY OF DESIRE CAN BE SUBJECT TO THE MORAL L A W
We have just seen that if the deduction of the moral law is to be carried through Kant must establish a practical cognitive viewpoint. He has already shown the rational necessity of such a viewpoint, but two problems remain: (1) to show that practial reason has a determinate relation to sensibility (not through intuition, but through the faculty of desire) and therefore represents an objectively valid cognitive framework; and (2) to show that the practical viewpoint is capable of being united with the theoretical in one consciousness. Step 2 ("The Interest Attached to the Ideas of Morality") solves the first problem and provides the grounds for dealing with the second. (The second problem is expressly solved only in the third step, where the entire argument is presented synthetically.) The first thing that is required in order to show that the results of the first step are not mere speculation is to show that the concepts dealt with there
232
The Journal of Value lnquiry
have objective reality - that is, that they refer to some content that can be sensibly given (since for Kant concepts alone can never have objective reality). From the first Critique we would expect that the objective reality that needs to be shown is a theoretical objective reality (since the realm of theoretical knowledge is the only realm that has so far been established). And since, as an object of experience, the will is a faculty of desire determinable only in accordance with causal laws of nature by an interest taken in an object, the problem would be to show how the will can take an interest in a law of pure reason. This problem turns out to rest on a misunderstanding, since it is not for theoretical reason that the objective reality of the ideas of morality must be proved. But Kant begins the second step by confronting this misunderstanding. In the first step we saw that the will insofar as it is rational is free and as free is subject to the moral law. If we take that proof as having objective validity for theoretical reason then it would apply immediately to the will as we experience it - namely, to the will as faculty of desire; and the argument would have to be interpreted as having proved that we must take an interest in the moral law. But in that case the abstraction upon which the first step rested would amount to a surreptitious assumption: If we treat the will solely as rational cause and at the same time claim theoretical objective reality for that concept of will, then our initial "definition" is indistinguishable from the premise that the will in question is capable of being determined by pure reason and its laws. That is, by basing our argument on the concept of will simply as rational cause we would effectively have assumed what was to be proved, and our argument would be not only analytic but circular. So Kant is right in claiming that there i s a n apparent circularity in the argument (449, 450, 453), but it arises only if we try to give objective validity to the speculative argument of step 1 by applying it to the will as object of experience - that is, if we insist that the argument be shown to be objectively valid for theoretical reason. There is no circularity if we take the argument as establishing a new cognitive viewpoint, namely, the viewpoint of practical knowledge. And I think that is how we must understand Kant's notorious introduction at this point of the "two standpoints." There are actually three possible standpoints: (1) the standpoint of experience (or, more generally, of theoretical knowledge); (2) the standpoint of pure thought, from which standpoint we think ourselves as members of an intelligible world but have no knowledge; and (3) the standpoint of practical knowledge, from which "we look upon ourselves as belonging to the sensible world and yet to the intelligible world at the same time" (453). More precisely, from the practical standpoint we view ourselves as sensible beings (and therefore subject to laws of nature) who are nevertheless subject to laws of an intelligible world: In the practical viewpoint laws of nature are not omitted but rather subordinated to the moral law (the argument for how that is possible is given only in the third step). The introduction of the practical viewpoint seems at first to be a mere assertion, but in fact Kant gives grounds for it. In the first place, of course,
Discussions
233
the first step has proved apodictically that we must think ourselves as free and as subject to the moral law. That is not yet a cognitive viewpoint, since it lacks a determinate relation to sensibility, but it is a necessity of thought. Second, the thought of an intelligible world (and of ourselves as intelligences) - distinct from the world of appearances and grounding it - is a requirement not only of practical reason but of reason as such. We can have an immediate awareness of the distinction between the phenomenal and the intelligible simply by reflecting on the spontaneity of our own reason (as opposed to sensibility, which is ultimately passive). This distinction was arrived at in the first Critique, although for theoretical reason the thought of an intelligible order merely marks the boundary of knowledge; and, Kant claims, it is available to even the most unsophisticated thinker (451-452). And third, we also can have immediate sensible verification of the validity of the practical viewpoint in the awareness that We do indeed find ourselves able to take an interest in a personal characteristic that carries with it no interest in mere states, but only makes us fit to have a share in such states in the event of their being distributed by reason. That is to say, the mere fact of deserving happiness can by itself interest us even without the motive of getting a share in this happiness (450). This last point needs to be emphasized, even though (or, rather, precisely because) Kant mentions it only in passing, That the practical viewpoint has reality for us as sensible beings (as well as for us as rational beings) is proved by the fact that we can have an immediate awareness of moral obligation (an interest in the worthiness to be happy detached from an interest in happiness itself). So when Kant intitially poses the problem of the second step as being to show how we can take an interest in the moral law he does not mean to raise the question of whether we do take such an interest, but rather to call for a transcendental clarification of the conditions of the possibility of that interest. Moral obligation is here (as in the second Critique) taken as a fact, and the problem is not to show that we are obligated but rather to show how that is possible 5 by showing the possibility of an objectively valid practical cognitive viewpoint. So the second step of the argument shows that the logically necessary thoughts of freedom and the moral law can have objective reality not for theoretical cognition but for practical cognition. It is true that the idea of practical cognition entails the thought of an intelligible order and that the intelligible order is unknowable for theoretical reason, but the thought of noumena is compatible with theoretical reason and is even a necessary limiting idea for theoretical reason. The thought of an intelligible world could not have theoretical objective reality because (by definition) it could not be intuited so it could not (in principle) be shown as having objective 5 It should be kept in mind that "showing how moral obligation is possible" can only mean establishing some other realm of possibility than that of theoretical cognition, since moral obligation is ultimately impossible from the viewpoint of theoretical reason.
234
The Journal of Value Inquiry
reality. But from a practical viewpoint it is not intuition that proves the objective reality of concepts but rather determinations of the will. And from this viewpoint the practical objective reality of the ideas of morality is proved by the fact of our awareness of moral obligation. So both practical reason and practical sensibility demand an objective practical cognitive framework. But the practical cognitive framework consists of laws for the ordering of phenomena and for bringing them to systematic unity, and the phenomena that are included in that framework are the same phenomena that (from the theoretical viewpoint) are objects of experience. That is, there is only one world, even though we must view it as governed by two distinctly different sets of law (namely, moral laws and laws of nature - cf KU Introduction, Section I1). So there remains a question of the unity of that world and of the unity of our consciousness in representing phenomena, which are always subject to natural laws, as being also subject to moral laws. That question is dealt with in the third step, in which Kant finally gives the deduction in a single argument.
STEP
3:
THE DEDUCTION PROPER
The problem of deductions in general is to show how synthetic a priori judgments are possible. The problem of the practical deduction is to show how a pure rational determination of the will can be represented as a priori necessary for our will, which is a faculty of desire. The argument so far has shown that the will, as rational, must be considered as subject to laws of an intelligible realm and ultimately to the moral law as the highest law of that realm. The argument has also shown that the will, as faculty of desire, is subject to laws of the phenomenal realm, that is, to laws of nature. If there is to be a practical cognitive viewpoint it must include both rational determination and sensible conditioning of the will in the unity of one (practical) consciousness. That means that the noumenal determination and phenomenal conditions of the will must be able to be represented as standing in some a priori necessary relation to one another. What that relation is has already been suggested in the second step of the argument: When Kant argues that reflection can reveal to us the distinction between the noumenal and the phenomenal he also points out that the noumenal realm must be represented as the ground of the phenomenal realm (451). This relation between noumena and phenomena is necessary because the way reflection becomes aware of the distinction between the two is by recognizing the necessary element of passivity in our representation of appearances. That means that our representation of appearances includes the representation of ourselves as having-been-determined. But the representation of having-been-determined is the representatioti of a consequence, which necessarily implies a ground. The ground cannot be something that passively receives determinations but must rather itself be active. And the only pure activity that reason can represent to itself is its own activity in
Discussions
235
thinking. So the noumenal realm must be represented as ground of the phenomenal realm. (This argument is simply an amplification of Kant's argument in the first half of the paragraph beginning at 450-451 .)6 So the solution to the problem of the deduction is that the unity of the practical framework (which contains noumenal determinations and sensible conditions under mutually irreducible laws) is possible because the intelligible worm contains the ground of the sensible worm and therefore also of its laws; and so in respect of my will, for which (as belonging entirely to the intelligible
world) it gives laws immediately, it must also be conceived as containing such a ground (453). That means that (at least "so far as reason is concerned" -454) in acting we can and must view the will as it is experienced (namely, as sensibly conditioned) as being subordinate to the will as it is thought under the idea of freedom (namely, as subject to determination solely by the moral law); or, as Kant puts it, we must view the pure will as the "supreme condition" of the will as affected by sensuous desires (454). This result is synthetic in the strong sense since it includes sensible conditions subordinated to pure rational determinations. With this step the deduction is complete: It has been shown that both reason and sensibility demand a practical cognitive framework, and the unity of that framework has been shown to be possible on a priori grounds. So the conclusion is that there must be a practical cognitive framework, and that within that framework the moral law is the highest law.
THE LIMITATIONS OF THE DEDUCTION The remainder of Chapter III is devoted to discussing the limitations of practical philosophy. This discussion is frequently read as being a discussion of the limitations of the proof, and very negative conclusions are drawn about what the proof has accomplished. But in fact what Kant points out is simply that practical knowledge is not theoretical knowledge - the thought of obligation, for example, does not admit of verification in experience. That is certainly an important restriction, but, on the other hand, it is a necessary consequence of the fact that the deduction has established the objective validity of a cognitive framework that is distinct from the theoretical viewpoint - namely, the practical viewpoint. And, what is more, the practical 6 This argument does not involve the fallacy of appealing to the theoretical category of cause/effect to determine the relation between the intelligible and the phenomenal: The categories of cause/effect (or, better, of ground/consequence) before they are schematized are, in the first place, categories of the pure understanding, and as such they are applicable to noumena as readily as to phenomena, although they lack a determinate reference to a definite object and so cannot by themselves supply knowledge. Kant makes this point in the second Critique (Kp V 54).
236
The Journal of Value Inquiry
viewpoint includes the entire world of experience (insofar as it relates to the will) under its laws, so that the deduction, rather than limiting the practical use of reason, actually limits the theoretical use of reason (which views the will as subject only to laws of nature). The fact that these restrictions are generally interpreted so negatively seems to indicate that the preoccupation in our times with the viewpoint of the natural sciences has blinded us to an essential part, not of our experience in Kant's strict sense, but of our total world - namely, to our awareness of our own existence as agents. If the interpretation presented here is correct, then the primary aim of the deduction is to establish (show the conditions of the possibility of) an alternative to experience, namely, the practical viewpoint. In that case, its success or failure has to be judged by the degree to which such a viewpoint is convincingly established. And in these terms I think there is one important failure in the argument - a failure that is corrected only in the second
Critique. Kant is here able to justify the subordination of the sensibly conditioned will to the pure will only "so far as reason is concerned" (454) - that is, only by the pure formal argument that phenomena presuppose and are subordinated to a noumenal ground. What is needed in addition to that argument is one showing that the particular phenomenon in question (the sensibly conditioned will) is able to be subordinated to its noumenal ground. That is, the proof that the pure will is the supreme condition of the sensibly conditioned will needs to be carried through on the level of the sensibly conditioned will itself if the argument is to be complete. That means that the will's sensible conditions (its objects and the pleasure and pain through which they condition the will) need to be shown to be a priori subsumable under the moral law. The reason why Kant did not (and could not) complete the argument in that way is, I think, that at the time of the writing of the Groundwork he was still convinced that "the concepts of pleasure and pain, of the desires and inclinations, etc . . . . are of empirical origin" (Kr 11, B29; cf the passage in the A-edition that this replaces). If that is the case, then those concepts cannot be brought into the argument, since transcendental philosophy must "consist in knowledge wholly a priori" (KrV, B28). I think, however, that this is one of the few cases in which Kant can be clearly shown to have changed his mind: By the time of the writing of the second Critique he had realized that those concepts are capable of being given "definitions" that "consist only of terms belonging to the pure understanding, i.e., categories, which contain nothing empirical" (KpV, 9n). That being the case, it becomes possible for him to carry through the argument grounding the practical viewpoint even on the level of the sensibly conditioned will itself, which he does by showing that the moral law is the supreme condition of all objects of the will (Kp V, Chapter 2) and that it is itself the supreme incentive of the will (KpV, Chapter 3). Aside from this difference, I think there is no essential difference in the proofs of the two arguments (although there are certainly differences in the manner of presentation) - a position that can only be put forth as an assertion here.
237
Discussions CONCLUSIONS
From this summary account of the deduction we can draw a number of conclusions: In the first place, the guiding thesis used to make sense of the argument was that the argument needed to ground not just the moral law but a cognitive framework within which the moral law is the highest law. This distinction is important since it allows us to distinguish in the practical philosophy (as in the theoretical) a level of transcendental argumentation from a level of metaphysical argumentation: The concern of transcendental philosophy would be to establish the conditions of the possibility of a cognitive framework, whereas the concern of metaphysics (in Kant's sense) is to develop the principles of a given cognitive framework and to derive from them knowledge of objects (see KU, Introduction, Section V). (This distinction turns out to be invaluable in dealing with the problem of the "application" of the moral law.) In the second place, this interpretation allows us to avoid reducing the practical viewpoint to the status of a poor imitation of "real" knowledge, one whose inadequacies must be passed over in embarrassed silence for the sake of rescuing morality. If the argument is correct, practical knowledge is in one sense more firmly grounded than (and even subsumes) theoretical knowledge, even though its grounds do not allow of complete insight. And, finally, we can make at least some cautious generalizations about Kant's understanding of a transcendental argument. Two things in particular seem to characterize the deduction: (1) It is essential that the argument is concerned with a cognitive framework rather than with any specific knowledge within that framework; it is the possibility of knowledge at all that is in question, and it is that fact that requires a special kind of argument. (2) Kant's assumptions about the distinct roots of human knowledge are indispensable to the argument. The distinction between mere speculation (mere thinking, which requires only our rational faculties) and knowledge (which also requires sensibility) is essential: Without that distinction we cannot understand why the argument is necessary in the first place (since, as I argued above, we cannot see how the moral law is synthetic), nor can we understand what it means to ground a cognitive framework: Grounding a cognitive framework for Kant has turned out to involve showing how two irreducibly distinct faculties can cooperate in acts of cognition - here it is the faculty of desire and the pure will which are conceptually irreducible and for which grounds of a possible a priori unity must nevertheless be given. The fact that these two features are so central to the argument would indicate that they should be taken into consideration in any discussion of transcendental arguments framed along Kantian lines. New School for Social Research