Kant on the Justification of Moral Principles

I The Task of a Deduction

Kant maintains that the categorical imperative is a synthetic judgment a priori (GMS, AA 04: 420; KpV, AA 05: 31, 46). The demand for a justification, or “deduction”, arises from the fact that, for Kant, only in synthetic judgments does the meaning of the predicate term go beyond the meaning of the subject term.^[10] So, only in synthetic judgments do we learn something we did not already know when we know the meaning of the subject term. In order to justify the connection of the predicate term with the subject term, we cannot simply refer to the meaning of those terms. What we need instead is a “third thing”, as Kant puts it, which functions as the justificatory ground of the synthetic judgment. In the case of synthetic judgments a posteriori, this “third thing” is actual experience (KrV, B 193 ff.; cf. B 778). The categorical imperative, however, is not a synthetic judgment a posteriori but rather a synthetic judgment a priori; hence an empirical justification of this judgment via actual experience is in principle impossible. It is well known that for Kant the possibility of philosophy as a science in its own right turns essentially on the question whether synthetic judgments a priori are possible. His answer in the first Critique is that these judgments cannot be justified with reference to actual experience but must instead be justified with reference to possible experience (ibid.). To give a well-known example: only under the condition that all events (changes in time) have a cause is experience (in the sense of determinate empirical knowledge) possible.

However, the categorical imperative is not a theoretical synthetic judgment a priori but rather a synthetic practical judgment a priori (GMS, AA 04: 420). In practical judgments, we determine the subject term not as it is but as it ought to be. Hence a justification with reference to possible experience is ruled out. Moreover, the subject term of a practical judgment is the agent herself rather than an object given to us from elsewhere; practical judgment in its core case is self-determination. The subject term of the categorical imperative is the human agent. Human agents by nature do not necessarily will what is practically universal. But in being guided by the categorical imperative, we demand precisely this; we oblige ourselves only to will in such a way that the principles of our action are practically universal. In judging morally, we demand of ourselves that we always act from universalizable maxims, even though this universalizability is not already contained in our actual volition (cf. GMS, AA 04: 420; 454). In other words, the categorical imperative connects the concept of a finite rational being by means of the copula ‘ought’ with an idea which is beyond any possible experience: the idea of moral perfection, which is agency from universalizable maxims.

Kant’s universalization procedure has been the subject of much criticism. The goal of a deduction in Groundwork III is not to justify this procedure. Kant does not address content scepticism with regard to the moral principle.^[11] Instead, he takes it for granted that the first two sections of the Groundwork have established the universalization procedure as a proper articulation of pure practical cognition (moral reasoning). The deduction takes it for granted that the categorical imperative in its different formulations is the correct analysis of the unconditional practical necessity which, according to Kant, is expressed in our practical judgments. What the deduction seeks to justify, instead, is the synthesis in those practical judgments. By tracing moral obligation back to its origin, it seeks to show why finite rational beings like us are obliged to act under the kind of moral law Kant has already provided an account of. The goal of the deduction is, in other words, to explain why the categorical imperative has binding force for us ‒ or, as Kant puts it, “on what grounds (woher) the moral law is binding” (GMS, AA 04: 450).^[12]

II The Deduction of Moral Obligation

The logical form of the categorical imperative is not the point of contention between the Groundwork and the second Critique. In both writings, Kant holds that the categorical imperative is a synthetic practical judgment a priori (GMS, AA 04: 420; KpV, AA 05: 31). The disagreement lies in the justification of this judgment. The first two sections of the Groundwork only show that those who hold “morality to be something and not a chimerical idea without any truth must also admit the principle of morality [i. e. the moral law] brought forward”. Conversely, those who hold morality to be a mere “chimerical idea” do not need to accept that the moral law has any binding force for us as finite rational beings (GMS, AA 04: 445). “Well-disposed souls” might gladly concede its bindingness (GMS, AA 04: 453), but what about those who put the “absolute necessity” (duty) of morality into doubt (GMS, AA 04: 445)? In the third chapter of the Groundwork, Kant attempts to establish the “third thing”, which functions as a justificatory ground for the syntheticity of our moral judgments. By means of this “third thing”, Kant wants to gain the entitlement to connect a finite will with the idea of a will that necessarily wills what is universalizable.

So far, we have been mainly concerned with the task of the deduction in Groundwork III. We now need to turn to the actual argument Kant delivers in Groundwork III. There is considerable disagreement about each move in his argument. But in order to be in a position to critically adjudicate between Kant’s deduction in Groundwork III and his later account in the second Critique, we need to get a clear grasp of what his argument in Groundwork III consists in. In this section, I would like to lay out, in four argumentative steps, Kant’s deduction as I understand it.

III The Primacy of Moral Cognition

In the second Critique, Kant no longer makes any attempt to explain the possibility of the categorical imperative or, what amounts to the same thing, to give any sort of a deduction (strong or weak). Kant now calls the moral law a “fact of pure reason” (‘KpV, AA 05: 31, 47), and he claims that “something different and quite paradoxical takes the place of this vainly sought deduction of the moral principle, namely that the moral principle itself serves as the principle of the deduction of an inscrutable faculty […], namely the faculty of freedom” (KpV, AA 05: 47). Many have claimed that the main difference between the Groundwork and the second Critique is Kant’s reversal in the order of the justification of freedom.^[30] He no longer attempts, as he did in the Groundwork, to deduce the moral law from some prior cognition of freedom. Instead, the moral law is now that which is prior in cognition and is itself the ratio cognoscendi, the ground of cognition, of the freedom of our capacity of volition (KpV, AA 05: 04). I think that this suggestion goes in the right direction; in claiming that the moral law is a fact of pure reason, Kant is concerned with the order of justification. However, since Kant already attempted in the Groundwork to give a “deduction of the concept of freedom from pure practical reason” (GMS, AA 04: 447), and since practical reason is nothing but the capacity for moral cognition, we need to further specify the reversal claim. I would like to focus on a passage where Kant explicitly explains why he calls the moral law a ‘fact’ of reason:

[The moral law is a fact of reason] because one cannot reason it out from antecedent data of reason, for example, from consciousness of freedom (since this is not antecedently given to us) and because it instead forces itself upon us of itself as a synthetic a priori judgment that is not based on any intuition, either pure or empirical, although it would be analytic if the freedom of the will were presupposed; but for this a positive concept, an intellectual intuition would be required, which certainly cannot be assumed here.^[31]

This passage is very dense. In the following, I want to lay out five different claims that I take Kant to be making here. Bringing them into focus will help us to see more clearly what distinguishes Kant’s argument in the second Critique from his argument in the Groundwork.

IV Exposition Instead of Deduction

Thus far, we have restricted our attention to the meaning of Kant’s phrase ‘fact of pure reason’. But for many, it remains an open question whether Kant is entitled to the claim that the categorical imperative has any binding force. Kant tries to address this sceptical worry in Groundwork III. To put the worry slightly differently, even if it is the case that human beings are guided in action by the moral law, this only answers the quid facti question; the quid juris question is still open. Hence, critics demand a deduction, and they presume that Kant’s initial attempt in the Groundwork was a move in the right direction.^[40] Some claim that Kant never gave up on his weak deduction in the Groundwork.^[41] Others maintain that when Kant came to realize that his deduction in the Groundwork had failed and he found himself unable to improve his argument, he falsely took the mere exposition of the moral law to be a sufficient justification of it.^[42] According to a widespread view, consciousness of the moral law could simply be a historically contingent fact. What right does Kant have to claim that the “reality of the moral law cannot be shown by means of a deduction” because it is “firmly established of itself” (KpV, AA 05: 47, cf. 94)? Or, as he also says, that consciousness of the moral law cannot be “reason[ed] out” (herausvernünfteln) from “antecedent data of reason”, but rather “forces itself on us” (KpV, AA 05: 31)? What is consciousness of the moral law like, such that it allows for non-perceptual knowledge (as a fact of pure reason rather than sensibility)? Or, to put the question in an interpretive context, is Kant right to replace the “deduction” of the categorical imperative provided in the Groundwork with a mere “exposition” of it (KpV, AA 05: 46)?

The demand for a deduction of the moral principle Kant puts forward raises at least three different kinds of worries, which should be made explicit and kept distinct. First, the formal content of the principle requires a derivation (content scepticism). Second, Kant needs to give an argument for why acting from this principle is possible (predeterminist scepticism). Finally, he needs to show how the principle has any binding force ‒ or, to put it in more Kantian terms, how it determines moral agents (normative scepticism). Let me explain what role each of these worries plays in the deduction.

The first worry concerns the so-called “universalization procedure” of the categorical imperative. More precisely, this is the worry about whether the categorical imperative is in fact the proper articulation of the form of our moral judgments. Even though it is far from obvious that the categorical imperative in all its different formulations is indeed a proper articulation of our moral judgments,^[43] I don’t think that this is the question Kant was worried about when he raised the deduction issue in Groundwork III. This issue had already been addressed in the first two sections of the Groundwork. If Kant’s account of moral judgment is correct, he can take it for granted, in the beginning of the third section, that the categorical imperative is the form of practical judgment. Similarly, content scepticism is not at stake when Kant considers the deduction issue in the second Critique. He again takes it for granted that our moral judgments make a claim to strict universal validity. His argument in sections 1–8 is supposed to establish mainly two claims: first, only the formal law of practical reason can hold up to the claim of strict universality; second, the principle of this law is autonomy, which is freedom in its absolute sense.

At this point of the argument in the Groundwork, Kant still thought it necessary to address a different kind of worry. Even if the categorical imperative is an adequate articulation of unconditional practical necessity, it is still an open question whether human beings are subject to its demands. We might call this worry “normative scepticism”. As we saw in section II of this paper, Kant’s strategy is to trace consciousness of the moral law’s bindingness back to consciousness of freedom. He first argues for the analytic connection between freedom and the moral law. He then tries to establish the claim that, as rational beings with a capacity of pure volition and conscious of freedom, we are entitled to view our actions from two standpoints. Since rationality is normative for our sensibility, Kant finally concludes that we consider ourselves obliged to the categorical imperative. We also saw that this argument turns out to be circular, for it presupposes the capacity of pure practical reason, which is known to us only via cognition of the moral law.

In the second Critique, Kant is still convinced that the final part of his response to normative scepticism is correct. He even implicitly refers to what is in all likelihood his argument in Groundwork III: “[I]t has been sufficiently proved elsewhere that freedom, if it is attributed to us, transfers us into an intelligible order of things” (KpV, AA 05: 42). So Kant does not abandon his idea of the two standpoints, which is supposed to explain, as we saw in section II, the imperatival form of the moral law.

However, since he endorses the reversal claim in the second Critique (as we saw in the previous section), he has to give up the first part of his response to normative scepticism. He no longer tries to argue from consciousness of freedom to consciousness of the binding force of the moral law. Instead, he now claims that “[t]he most ordinary attention to oneself confirms that this idea [i. e. the moral law] is really, as it were, the pattern for the determination of our will” (KpV, AA 05: 44). He now simply refers to our self-consciousness of practical universality, and he illustrates this consciousness with the very same examples he used when he introduced the universal law formula in Groundwork II. In other words, the starting point of Kant’s argument in the second Critique is the actuality of consciousness of the moral law. He infers from this actuality that it must be possible and concludes that the problem of the possibility of the moral law (the problem which the deduction is supposed to answer) does not need to be addressed.

Now at this point, one might wonder whether there is room for a fourth kind of scepticism. This time, the scepticism would be psychological, and it would proceed as follows: it might be the case that many people are in fact conscious of the moral law when they determine their will; however, this consciousness could just be a mere historical accident of the particular time we or Kant happen to live in.^[44] Kant would respond to psychological scepticism by pointing out the formal character of the moral law and its a priori origin. The requirements of objective and subjective universality are not historically contingent, but are rather constitutive of cognition as such. Without those requirements, practical as well as theoretical cognition would be impossible. Assuming that Kant can demonstrate how the categorical imperative determines the practical universality of any of the maxims put forward by the sceptic, and therefore how it functions as the condition of the possibility of practical cognition, psychological scepticism is unfounded. If the demand for a deduction arises because Kant’s claim about the fact of pure reason has been taken to give rise to psychological scepticism, then Kant’s analysis of practical cognition in sections 1–8 of the second Critique can be seen as a deduction. The “exposition” of the categorical imperative as an a priori principle of pure practical cognition would be at the same time its “deduction”, i. e. a justification against psychological scepticism. But if we want to stick as closely as possible to Kant’s vocabulary, it would be most appropriate to call Kant’s argument in §§ 1–8 a “metaphysical exposition”. For an exposition is “the clear, though not detailed, representation of that which belongs to a conception; and [it] is metaphysical when it contains that which represents the conception as given a priori”.

The deduction in Groundwork III is not supposed to be a response to psychological scepticism. When Kant worries that the categorical imperative could be a chimerical idea without truth at the end of section II, he is worried that the categorical imperative might not apply to finite rational beings. As we will see in the final section, a fundamental difference between theoretical and practical cognition renders this deduction of the categorical imperative unnecessary.

V Dogmatism Instead of a Critique of Pure Practical Reason

I claim that the demand for a deduction of the categorical imperative only arises because Kant erroneously modelled his argument for the principles of practical cognition in the Groundwork on the deduction of principles of pure theoretical cognition. Hence he tried to provide a “critique of pure practical reason” (GMS, AA 04: 440, 446, my emphasis) ‒ a critique that he deemed unnecessary in the second Critique (KpV, AA 05: 03). So if we want to see why Kant gave up on the deduction, we need to get a better grip on the difference between theoretical and practical cognition.

Judgments of pure theoretical reason need to be restricted to possible experience, for otherwise they are transcendent and do not have any meaning. Accordingly, it is not sufficient to show that the fundamental concepts of pure theoretical reason are thinkable; Kant also needs to show that they map on to the world as constitutive principles of possible experience. In other words, criticizing reason in its pure employment means restricting its employment to possible experience. The principles of practical cognition, by contrast, are not restricted to possible experience, for in practical cognition we do not determine “objects that may be given to reason from elsewhere”. Instead, cognition of the moral law is the self-consciousness of pure practical reason (KpV, AA 05: 29), and, in the paradigm case, this cognition makes the object of its cognition “actual” (wirklich) (KrV, B IX f.). As Kant puts it elsewhere, practical cognition is the “ground of the existence of objects” (KpV, AA 05: 46). So, when Kant claims that practical cognition is concerned with the determination of our faculty of volition (KpV, AA 05: 19), he means not merely that we judge our maxims to be good or evil, but that cognition thereby brings about the action and is thus also causally efficacious. Practical cognition, in its paradigmatic form, requires that the principle of legislation also be the principle of execution.

This is rather abstract, and it is not immediately obvious how this account of practical cognition can explain why the categorical imperative does not require a deduction. It might help to elucidate the claim with one of the examples to which Kant alludes when he claims that we become conscious of the moral law as the determining ground of our volition by simply attending to our own consciousness (see above). Let us consider the following maxim: “When I am in need of money, I will borrow money and promise to pay it back even though I know this will never happen.” Again, this is not the place to respond to content scepticism and to critically assess Kant’s universalization procedure. But Kant’s main idea is that the end of this maxim (borrowing money) could not be achieved if the maxim became a universal law. There wouldn’t be such a thing as borrowing money if it became a universal law that everyone in need of money falsely promised to pay it back. There is nothing problematic about a world without promises as such. But if an agent did have this maxim and wanted to borrow money, she could not bring about her end if it became a universal law. So, from a universal or rational standpoint (which is our own standpoint), this maxim can be neither willed nor even thought without contradiction. Now, a proper maxim is supposed to function as a principle in a practical syllogism. For this principle to be justified, it should hold for every action subsumed under it (objective universality), and all other rational beings should be willing to agree to it (subjective universality). But if the maxim cannot even be thought without contradiction, it cannot function as a sound principle of a practical syllogism and therefore does not hold up to a maxim’s claim to universality. A maxim that cannot be universalized lacks practical necessity and is morally impermissible; it is, in Kant’s own modal terms, “contrary to duty”.

What this is supposed to show is that in practical cognition the merely “problematic […] thought a priori of a possible general legislation” is the determining ground of our will (KpV, AA 05: 31), and therefore of the reality of some object: the good (or evil) action. Through this merely problematic thought, we determine whether a principle counts as a principle of practical cognition. This is not to say that in practical cognition we gain positive insight into the intelligible world, cognizing the true end of moral agency or receiving insight into this end by divine revelation. On the contrary, even in moral cognition it is merely a formal end of which we have cognition a priori: the merely problematic thought of general legislation. Moral cognition would remain empty if our maxims did not contain given desires or, as Kant likes to say, inclinations. For as I pointed out earlier, only when we lay out maxims (which is our way of determining under which conditions we choose to satisfy our desires) do we become conscious of the moral law. And only in having particular desires can our reasoning guide our action. Without inclination, the moral law would be ineffective. It is not as though we must repress all our desires when we act morally; we must only make their realization depend upon their possible universalizability or law-likeness. But in determining those desires, we determine them not as they are but as they ought to be, for practical cognition is volition under the condition of its practical universality. Volition is practically universal if it is valid for all actions that fall under the rule (objective universality) and if it holds for every rational cognizer (subjective universality). Since our inclinations are not already in line with what is practically universal, practical reason issues its demands in the form of imperatives. Now on Kant’s view, in order for the cognition to become practical, it needs to give rise to a feeling which is able to strike down all competing desires. Kant calls this feeling “respect for the moral law” (KpV, AA 05: 73 ff.). Only if reason is in fact able to determine our volition in such a way that the cognition is itself the ground of the object’s existence are we entitled to say that reason is practical and not merely theoretical. For even in technical-practical volition, the rule is always determined by a presupposed desire. Only in moral-practical cognition is the determining ground of our volition not some presupposed desire but a merely problematic thought: the thought of universal legislation.

One might wonder how this can explain why the moral law does not require any deduction, for in Kant, and especially in Kant, there is a fundamental similarity between theoretical and practical cognition. Even theoretical cognition is not something that we are passively saddled with. It is only through the spontaneity of our understanding and reason that theoretical cognition is possible. And one might think that even in practical cognition the matter of our maxims does not (or at least does not entirely) depend on practical reason but is instead given to us from elsewhere. As we have just seen, however, in sharp contrast to theoretical cognition, practical cognition is not concerned with the determination of an independently existing manifold. Through the forms of our practical cognition, the categories of freedom, we do not simply determine what we in fact desire. Instead, in practical cognition we cognize the formal requirements of rational volition, and the cognition of these requirements is, in the paradigm case, also “the ground of the existence of the objects” (KpV, AA 05: 46). We can call this view Kant’s causal efficaciousness view of practical reason and distinguish it from the mere representationalist view. Reason as a causally efficacious capacity shapes the world not only formally but also materially.

These two views lead to two different procedures for justifying the fundamental principles of cognition. For although we can become conscious of the apriority and necessity of the fundamental principles in both practical and theoretical cognition by simply “attending” to our judgments (KpV, AA 05: 30), the fundamental principles of theoretical cognition are still in need of a deduction. In theoretical cognition, Kant holds a representational view of reason. He therefore needs to show that the fundamental principles are not mere “chimera” (Erdichtungen) but do in fact map onto the world. It transpires that their “objective reality” can be “quite easily and evidently proved through examples from the sciences” (KpV, AA 05: 91, 47). The a priori principles of theoretical reason are the condition of the possibility of experience, but they do not bring the object of their cognition into existence. Our theoretical cognition is therefore the cognition not of a divine intellect ‒ an “intuitus originarius” which brings the object of its cognition into existence by cognizing it ‒ but of an “intuitus derivativus” (KrV, B 72). In practical cognition, we are closer to the “intuitus originarius” insofar as the fundamental principle of practical reason does not simply represent given desires but is rather itself the origin of the object of our volition. This does not mean that we are merely finite as theoretical cognizers but divine as practical agents; rather, our practical finitude is constituted by the fact that we are subject to the influence of inclination.

Now since practical cognition is the origin of the object of our willing, and since willing is considered a causal force in nature, reason “proves” the reality of the categorical imperative “through a deed and all ratiocination against its possibility is in vain”. In other words, a deduction of the principles of pure reason is the attempt to answer the question how those principles are possible. If it turns out that the principles of this cognition are already actual, the question of their possibility does not even arise (KpV, AA 05: 03).

I think that what Kant came to realize in the second Critique was that the critical project in his practical philosophy was not, as he had envisioned it in the Groundwork, a critique of pure practical reason but was instead (simply) a critique of practical reason; pure reason in its practical use is not to be criticized, because it is by its very nature transcendent. Kant merely wants to show “that there is pure practical reason […]. If it succeeds in this it has no need to criticize the pure faculty itself in order to see whether reason is merely making a claim in which it presumptuously oversteps itself (as does happen with speculative reason)” (KpV, AA 05: 03). I take this to mean that Kant only needs to show that there is a fundamental practical principle constitutive of pure practical cognition. The question whether and how it relates to an object need not be addressed, for this cognition first brings the object of its cognition into existence. As an a priori cognition of what ought to be, it is by its very nature transcendent. Kant calls “dogmatic” any method that does not “at first examine the capacity or incapacity of reason” (KrV, B 7). I think we are therefore entitled to the rather bold claim that, with respect to pure practical reason, Kant’s method in the second Critique is dogmatic. With this qualification in mind, Kant can again endorse a claim that Erich Adickes dates to between 1769 and 1771: “Pure reason is dogmatic in morals” (Refl, AA 17: 520, R 4362). The third section of the Groundwork was supposed to make a transition from metaphysics, which proceeds in dogmatic fashion (KrV, B 7), to a critique of pure practical reason. Kant abandoned this idea. With respect to pure reason, the Critique of Practical Reason remains metaphysical. The critical project in the second Critique is limited to two tasks: first, Kant needs to fend off all attempts to derive moral principles from given desires; second, he needs to defend his absolute notion of freedom against predeterminist scepticism. It turns out that the self-consciousness of pure practical reason is consciousness of an unconditional requirement of practical volition. This unconditional requirement cannot in principle depend on any given desire, but must instead have its source in reason. As soon as we adopt maxims, we become aware of this requirement. Since ‘ought’ implies ‘can’, we are entitled to infer our absolute freedom from the consciousness of this requirement. This is the only deduction of the fundamental principle of pure practical reason offered by Kant in the second Critique ‒ a deduction of freedom.^[45] The only further argument needed at this point is an argument that will fend off determinist sceptics. This scepticism is theoretical in origin, since it questions the possibility of absolute freedom in the world of appearances. Against this kind of scepticism, Kant attempts to prove not the theoretical actuality of the concept of transcendental freedom but rather its logical possibility. Whenever this scepticism arises, he reminds his readers of his solution to the third antinomy (GMS, AA 04: 455 f.; KpV, AA 05: 30, 48 ff., 94‒106; KU, AA 05: 175; RGV, AA 06: 39 f., 49 f.). However, if the self-consciousness of pure practical reason as the moral law is a cognition, then not only can we ascribe the merely problematic idea of freedom to a finite rational being, but we can even positively determine this idea through our non-perceptual moral cognition, thereby proving ‒ from a practical point of view ‒ the objective reality of freedom. How is consciousness of freedom possible? Through consciousness of the moral law. The famous example of the gallows is supposed to confirm the order of this deduction.

Conclusion

In this paper, I have tried to develop a radical incompatibilist story about the relationship between the Groundwork and the second Critique. This incompatibilism is radical because it concerns the roots of Kant’s approach to the justification of the bindingness of the principles of practical cognition. I don’t think, as standard incompatibilists maintain, that Kant rejected his argument in Groundwork III because he detected a mistaken inference from theoretical freedom to practical freedom, or because he became aware of the fact that his axiological “superiority claim” concerning the priority of the noumenal over the sensible world was unjustified. My analysis of Groundwork III has shown that those particular arguments can be spelled out such that they are perfectly compatible with his view in the second Critique. The radical incompatibilism I want to endorse instead concerns the false belief that practical cognition, like theoretical cognition, requires a critique of pure reason. In the second Critique, Kant revised his argument because he came to realize that, in contrast to synthetic a priori judgments of theoretical cognition, the possibility of synthetic a priori judgments of practical cognition can be derived from the actuality of a “deed”. With respect to pure practical reason, the second Critique proceeds not critically but metaphysically, i. e. dogmatically. Hence Kant abandoned the idea of a deduction of the categorical imperative and with it the idea of a critique of pure practical reason: the very project he had started to undertake in the third section of Groundwork III. I agree with the compatibilist story that the deduction Kant considered impossible both in the Groundwork and in the second Critique is what I have here called a strong deduction ‒ a deduction that takes as its starting point theoretical cognition of transcendental freedom. And I also agree with compatibilists that the deduction he actually delivers in Groundwork III is only a weak deduction, which makes use of a merely problematic conception of transcendental freedom. However, I don’t think that Kant’s argument in Groundwork III succeeds entirely, for Kant moves from our conception of pure practical reason to transcendental freedom, and from there to our consciousness of autonomy; hence, his argument remains question begging. As argued above, the facticity claim in the second Critique, by contrast, can provide a non-question-begging account of moral obligation. Here I agree with the optimistic incompatibilist, who views the argument in the second Critique as an improvement on his argument in the Groundwork. However, more importantly, the radical difference between his approaches in the Groundwork and the second Critique, which the standard incompatibilist account overlooks, concerns the very idea of a critique of pure practical reason, and with it the very nature of pure practical cognition. This is why his attempt at a deduction in Groundwork III and his facticity claim are fundamentally incompatible. And, as I point out in the beginning of this paper, we should for this very reason resist the generality of Kant’s claim in the first Critique to the effect that, for all synthetic judgments a priori, “if not a proof then at least a deduction of the legitimacy of its assertion must unfailingly be supplied” (KrV, B 286).