The Alleged Paradox of Self-Legislation and the Normative Requirements of the Rational Will: Kant’s Derivation of the Categorical Imperative Reconsidered more

The Alleged Paradox of Self-Legislation and the Normative Requirements of the Rational Will: Kant’s Derivation of the Categorical Imperative Reconsidered [Draft – 9.14.2011. Please do not cite.] Bob Robinson Loyola Marymount University I. Introduction Recent scholarship on the German Idealists has done a great deal to clarify their philosophical debts to Kant’s critical philosophy. This work has also led to a reinvigoration of serious interest in their criticisms of it.1 In his survey, German Philosophy 1760-1860: The Legacy of Idealism, Terry Pinkard claims that central to a proper understanding of German Idealism is their preoccupation with issues concerning Kant’s conceptions of freedom and agency, particularly his account of self-legislation. According to Kant, self-legislation is the act of willingly binding oneself to the authority of the moral law, and he thinks it is only if an agent willingly self-imposes the authority of the moral law that she can consider herself as both subject to moral obligations and also autonomous. Pinkard says: The Kantian legacy, by taking normative authority to be self-legislated, to be a product of our spontaneity as it combines itself with our receptivity in the theoretical sphere and to be a product of our autonomy in the practical sphere, raised the issue as to whether that kind of normative authority could itself be secured against further challenge. Kantians had their own answer: this normative authority, although spontaneously generated and therefore self-imposed, is nonetheless that of a “universal self-consciousness,” of the rules binding all rational agents […] Post-Kantians, while at first attempting to hold on to that notion of “universal self-consciousness” […] moved to the idea […] that as selflegislated, such authority, since it involves an inescapable self-consciousness about itself, is itself always subject to challenge.2 1 See, for example, Ameriks 2000, William F. Bristow, Hegel and the Transformation of Philosophical Critique (Oxford: Clarendon Press, 2007), Frederick C. Beiser, German Idealism: The Struggle Against Subjectivism, 17811801 (Cambridge: Harvard University Press, 2002), Robert B. Pippin, Hegel’s Idealism: The Satisfactions of SelfConsciousness (Cambridge: Cambridge University Press, 1989) and Pippin 2008. 2 Pinkard 2002, pp. 358-9. 1 As Pinkard claims, although the Idealists are deeply influenced by Kant’s account of selflegislation, their perception of self-legislation as a paradoxical concept left them dissatisfied with Kant’s moral theory. We can put the argument for the paradox in this way. Kant maintains both that the selflegislation of the moral law is the ground of freedom, and, an agent acts freely if and only if she acts according to the moral law, i.e., for a universally valid reason. From these two propositions it follows that a free agent must possess a universally valid reason for self-legislating the moral law. However, the preservation of freedom requires the self-legislation of that reason, which entails an infinite regress of self-legislated reasons. So, there can be no reason for the selflegislation of the moral law. Consequently, Kant’s concept of self-legislation commits him to the paradoxical claims that an agent must possess a universally valid reason for self-legislating the moral law and yet no such reason can be provided. The central thesis of this paper is that this paradox rests on a misunderstanding of Kant’s conception of self-legislation, namely, it is false that he believes that an agent needs a reason to bind herself to the moral law. But I also defend the claim that the interpretations of Kant’s account of self-legislation offered by the eminent Kant scholars Allen Wood3 and Henry Allison4 are incapable of defending against the paradox, and for the same reason. I use my arguments for the latter thesis to provide greater motivation for the former thesis. Before I begin, let me sketch the primary arguments of the paper. 3 4 Wood 1999. Allison 1990. 2 In the first and second sections of the Groundwork of the Metaphysics of Morals5, Kant offers a “derivation” of the supreme principle of morality. There, Kant argues that no principle save one of an unconditionally binding and obligatory sort is sufficient to ground moral obligations. The derivation is the justification of the claim that if there is such a supreme principle of morality, then it is the categorical imperative or what has become known in the literature as the Formula of Universal Law (hereafter, FUL): an agent ought never to act except in such a way that she could also will that her maxim should become universal law). Both Allen Wood and Allison maintain that Kant’s derivation of FUL contains a “gap,” arguing that the inference from the mere idea of a categorical imperative (or practical law) to FUL is unsupported because Kant surreptitiously sneaks into the derivation his undefended account of selflegislation. It is only if an agent does bind herself to the authority of the moral law that one has reason to think she is also bound by its obligation. Wood argues that FUL is not, in fact, the categorical imperative; instead, the self-legislated principle of morality is what is known as the Formula of Humanity (hereafter, FH): an agent must so act that she uses humanity, whether in her own person or that of another, always at the same time as an end, never merely as a means. Allison claims that the defense of self-legislation is supported in the third part of Groundwork, which contains Kant’s deduction of the moral law. The deduction aims to discharge the conditional that if rational beings possess freedom, then the categorical imperative is universally All translations of Kant’s writings are cited from the Cambridge Edition of the Writings of Immanuel Kant (Cambridge: Cambridge University Press, 1992-). Following standard convention, the pagination for those citations refers to the Academe edition of Kant’s works, Immanuel Kants Schriften, Ausgabe der königlich preussichen Akademie der Wissenschaften (Berlin: W. de Gruyter, 1902-). I shall use the following abbreviations of Kant’s texts throughout this essay: CPR = Critique of Pure Reason (1997a; following convention, citations refer to the A/B editions) Groundwork = Groundwork of the Metaphysics of Morals (1996a) CPrR = Critique of Practical Reason (1996a) Religion = Religion within the Boundaries of Mere Reason (1996b) MM = Metaphysics of Morals (1996a) LE = Lectures on Ethics (1997b) All italics of quoted passages from Kant are his own emphasis unless otherwise noted. 5 3 valid or necessarily binding. On Allison’s view, Kant argues that an agent must acknowledge that she is self-legislating in order to consider herself as possessing freedom, and this claim can then be plugged back into the derivation to preserve the derivation of FUL. Pinkard’s argument for the paradox of self-legislation exploits the views of those who defend the claim that there is a gap in Kant’s derivation of FUL. The question the paradox forces for Kant is why, exactly, does an agent bind herself to the moral law? Let us refer to this question as the ‘question of self-legislation’. Both Wood’s and especially Allison’s positions are, I think, deficient in failing to answer the question of self-legislation. In the first case, I shall argue that the problem for Wood’s view is that Kant appeals to FUL in arguing for FH, and this fact clearly manifests itself in how some philosophers attack FH as a sufficient ground for Kant’s moral theory. So, although Wood attempts to answer the question that drives the paradox, Wood’s attempt at an answer falls short. In the second case, I shall argue that Allison’s interpretation is incapable of answering the question of self-legislation. The problem for Allison is that he thinks that the argument of the deduction, the claim that moral obligation requires that an agent conceive herself as free and self-legislating, is sufficient to answer the question as to why an agent self-obligates herself to the authority of the law. But the necessity of an agent’s conception of herself as free and self-legislating is not identical to her actually binding herself to the moral law, and Allison therefore does not provide a reason for why an agent would bind herself to the authority of the moral law in the first place. In the final section of this paper I argue that the derivation gap, which the paradox exploits, does not exist. Kant thinks that an agent self-legislates or binds herself to the moral law as a requirement of exercising her practical rationality; that is, the moral law is an a priori logical principle to which an agent commits herself in possessing practical rationality. The crucial point 4 that Pinkard, Wood, and Allison fail to notice, then, is that Kant claims that maxims or rules of action possess a legal character. Without explicating the legal character of a maxim, however, it is mystifying why the demand of the categorical imperative is a self-imposed demand. For a maxim to possess a legal character means that whenever an agent adopts a rule of action and acts upon it, she approves of or endorses her rule as a permissible course of action for every other rational agent. That is, she treats her maxim as if it were a universal law or universally valid. Kant is clear that the moral law states nothing more than the logical constraint on practical rationality that the contents of a maxim are fit, or consistent, with it’s actually being capable of becoming a universal law or valid for all agents. Otherwise, if an agent wills a maxim unfit to become a universal law, then she treats her maxim as universally valid when it cannot in principle be universally valid; that is, an agent possesses an inconsistent state of will – she is irrational. So, the moral law is simply normative for practical reason as such, and to say it is self-legislated is to assert, as Kant does, that an agent commits herself to its authority, though she need not respect that authority, by simply exercising her practical reason. No reason is needed to explain the self-legislation of the moral law, for an agent binds herself to the authority of the law in reasoning about her actions. II. The Purported Derivation Gap In Groundwork I, Kant announces the “principle” that is to ground all moral obligations, and which is defended there and in Groundwork II: Since I have deprived the will of every impulse that could arise for it from obeying some law, nothing is left but the conformity of actions as such with universal, which alone is to serve the will as its principle, that is, I ought never to act except in such a way that I could also will that my maxim should become a universal law. Here mere conformity to law as such, without having as its basis come law determined for certain actions, is what serves the will as its principle, and must so serve it, if duty is not to be everywhere an empty delusion and a chimerical concept. (G 4:402) 5 To very briefly summarize, Kant maintains that, if there is an unconditional principle of action, then it must command an action irrespective of an agent’s “material incentives” or empirical grounds for acting. As Wood states Kant’s requirement, the reason an agent possesses for obeying a moral command “must be a reason that is overriding for all rational beings” and which is also “numerically the same”.6 Material incentives are insufficient as a source of obligation because, by their very nature, they can be neither overriding nor numerically the same. Kant’s support for this claim is that, human beings desire many of the same objects, e.g., happiness and personal well-being, and the attainments of those objects, if they are supposed as ultimate ends, sometimes require remarkably diverse courses of action or maxims. For example, you and I both desire to be happy, but, as it turns out, my happiness requires me to adopt a maxim that aims to deprive you of your happiness, while your happiness requires you to adopt a maxim that describes a course of action of resistance to me. Our reasons are not numerically the same, since my reason for acting pertains only to my happiness, and neither of our reasons can be overriding since each obligates contradictory maxims. The problem in making happiness a ground of moral obligation it reduces all obligations to hypothetical imperatives and subjective incentives. If morality is not to be “an empty delusion and a chimerical concept,” then there must be some objective, i.e., universal and necessary, ground of obligation. Kant’s claim is that this ground is to be found in the very concept of obligation or lawfulness. “[O]nly law,” Kant says, “brings with it the concept of an unconditional and objective and hence universally valid necessity, and commands are laws that must be obeyed, that is, must be followed even against inclination” (G 4:416). So, the lawfulness contained in the institution of a reason for acting must be, in terms of its material, 6 Wood 1999, p. 76. 6 conformable to its lawful form. This produces the requirement, which, following Bruce Aune, I refer to as L, that an agent must conform her actions to universal law.7 Kant seems to reason that from this requirement it follows that (FUL) an agent ought never to act except in such a way that she can will that her maxim should become a universal law. The purported fallacy of the derivation – its gap – is that FUL is derivable from L. Bruce Aune expresses surprise at this move, claiming that although it is quite clear that Kant takes them to be the same principle, they are not obviously identical.8 L expresses the most basic requirement of morality while FUL “shows us what must be done to comply with this requirement.”9 The apparent assumption that connects the two, Aune reasons, is that “we conform to universal law […] just when we obey [FUL] and act only on maxims that we can will to be universal laws.”10 What Kant needs to defend, then, is the claim that if an agent can consistently will her maxim of action in accordance with universal law, then such a willing produces an action in conformity with universal law. If this assumption is true, Aune thinks, Kant has a defensible derivation of FUL. But Aune goes on to contend that the assumption fails, and the reason is that “a maxim lacks the logical form of a proper law; what one can will to be a practical law is, at best, the generalization of a maxim.”11 Where a “typical maxim” has the logical form, “For any action A, if A satisfies the condition C, I will do A,” a law has the form, “All A’s are necessarily B.”12 The crucial difference, Aune claims, is that “a maxim is not a fully general statement” because “it contains a specific reference to the agent whose maxim it is.”13 Laws, on the other hand, contain no such 7 8 Aune 1979, p. 29. Aune 1979, p. 29. 9 Aune 1979, p. 31. 10 Aune 1979, p. 30. 11 Aune 1979, p. 88. 12 Aune 1979, p. 24. 13 Aune 1979, p. 24. 7 reference. As such, Kant’s undefended claim fails because there is simply no logical relation between a maxim and a law, a relation required to derive FUL from L. Aune says that “[a]part from telling us that universal laws are freely legislated by rational beings, he tells us very little about them until he has [FUL] in hand”.14 Allison claims that “[t]he difficulty to which Aune points is a real one” because L and FUL express two different requirements. For Allison, there is a relation, but the problem is that the relation is forced. The primary difference between L and FUL is that the latter expresses a “decision procedure for the choice of maxims” while the former expresses the logical form of a law.15 The decision procedure described in FUL does not obviously entail that one’s maxim, when universalized, takes on the formal features of a law because, as Allison states the problem, Kant conflates two different senses of universality. With some appropriate insertions for the sake of clarity, Allison says: The fundamental problem is simply that one cannot move from the claim that [L] all rational agents must regard their principles of action as universalizable in the sense that they are willing to acknowledge that it would be reasonable for every other agent in relatively similar circumstances to adopt the same principles or even that such agents ought to adopt them […] to the desired conclusion that [FUL] the agent ought to be able to will (as a universal law) that every rational agent act on the basis of the principle in question.16 The first sense of universality is the sort required by L, that all actions must conform to universal law, which Allison interprets as that of every rational agent recognizing their reasons as universally binding. The second sense of universality is the sort required by FUL, that it is universally binding that an agent act only on a maxim conformable to universal law. The first indicates the form of law, while the second indicates how the content of one’s principle is to be adequate to that form. It is unclear whether, in fact, the universality of FUL is capable of 14 15 Aune 1979, p. 89. As we shall see, both Allison and Wood follow up on this suggestion. Allison 1990, p. 210. 16 Allison 1999, p. 205-206. 8 overcoming the indexical reference of a maxim; that is, it is not clear that every agent does recognize their maxims as universally binding.17 Wood follows both Aune and Allison in asserting a gap in the derivation. He summarizes the worry about the gap in Kant’s derivation when he asks if FUL follows “merely […] from the idea of a categorical imperative or objectively grounded principle? The answer to this question unfortunately is, No”.18 The purported “fallacy” of Kant’s derivation of the categorical imperative is, Wood claims, due to his unwarranted inference “from the mere concept of a categorical imperative” to “the will of a rational being – what a rational being wills or can consistently will” as a valid criterion for “determining the content of moral laws”.19 Wood suggests that the only way such a derivation could be valid is: if we knew already that this will is the author of objective practical laws, hence that the moral law is a principle of autonomy. Kant does argue for this claim later (G 4:432), and his arguments do not depend in any way on the fallacious derivation of FUL and FLN.20 Stated differently, merely because a coherent conception of moral obligation requires, Kant argues, that no maxim contain self-interest, since such a reason for acting cannot be attributed to all rational agents, Wood objects that it does not follow, as Kant seems to think, that we actually impose this requirement (i.e., that no maxim contain self-interest) upon ourselves. If there is a principle of morality, then an adequate derivation of it must speak to the fact that its requirement is a self-imposed requirement. So, Wood argues, what Kant needs to explain is why the moral law is a requirement that rational beings impose upon themselves through the very nature of their rational will. That is, 17 Admittedly, although Allison appears to follow Aune, he does not make this last point. But it does appear to be the point of his assertion that the gap in Kant’s argument allows for the compatibility between L and rational egoism. See Allison, p. 205 and 211. 18 Wood 1999, p. 81. 19 Ibid. Wood explains further, “[Kant claims] what a rational being wills, or can consistently will, to be a valid criterion for what is consistent with a law objectively valid for rational beings in general” (Ibid.). 20 Ibid. 9 only if agents self-legislate the requirement that their maxims not contain any self-interest does it follow that they are obligated by the moral law. This is the premise that is needed to fill the gap between what moral obligation demands and the consistency of the will as a criterion for determining moral obligations. He explains that Kant “take[s] what a rational being wills, or can consistently will, [FUL] to be a valid criterion for what is consistent with a law objectively valid for rational beings in general.”21 However, the problem, he notes, is that it “does not follow from the mere concept of a categorical imperative [or, practical law] that the will of a rational being – what a rational being wills or can consistently will – has any role at all to play in determining the content of universal laws.”22 Wood suggests that the only way the derivation of FUL could be valid is “if we knew already that this will is the author of objective practical laws, hence that the moral law is a principle of autonomy.”23 Stated differently, merely because a coherent conception of moral obligation requires, Kant argues, that no maxim contain a material incentive as its basis (since the numerically same incentive cannot be attributed to all rational agents), Wood objects that it does not follow that rational agents actually impose the requirement of FUL upon themselves. For Wood, then, the undefended claim in the derivation is Kant’s doctrine of self-legislation. Despite somewhat different formulations of the problem of the derivation of FUL from the analysis of the requirements of moral obligation, what each of Aune, Allison and Wood indicate as potentially closing this gap in Kant’s argument is his account of self-legislation, though not the self-legislation of FUL or at least in a qualified sense. If Kant can establish the claim that it is only if rational beings impose the universality requirement of FUL upon 21 22 Wood 1999, p. 81. Ibid. 23 Ibid. 10 themselves that their maxims will conform with universal laws, then he will have derived FUL from the mere idea of a categorical imperative. III. The Paradox of Self-Legislation Pinkard offers subtly different formulations of the paradox in two distinct locations. Below I reproduce both passages in full: (A) The notion of the “fact of reason” thus boiled down to a restatement of the quasiparadoxical formulation of the authority of the moral law itself, which seems to require a “lawless” agent to give laws to himself on the basis of laws that from one point of view seem to be prior to the legislation and from another point of view seem to be derivative from the legislation itself. The paradox arises from Kant’s demand that, if we are to impose a principle (a maxim, the moral law) on ourselves, then presumably we must have a reason to do so; but, if there was an antecedent reason to adopt that principle, then that reason would not itself be self-imposed; yet for it to be binding on us, it had to be self-imposed.24 (B) Kant had argued that we must practically take ourselves to be self-determining, that what we as agents were “ultimately about” was freedom in this radical sense […] But if the will imposes such a “law” on itself, then it must do so for a reason (or else be lawless); a lawless will, however, cannot be regarded as a free will; hence the will must impose this law on itself for a reason that then cannot itself be self-imposed (since it is required to impose any other reasons). The “paradox” is that we seem to be both required not to have an antecedent reason for the legislation of any basic maxim and to have such a reason.25 In (A) Pinkard represents the paradox in a way that is, I think, a misrepresentation of Kant’s position, and clarifying why (A) is a misrepresentation enables to capture the force of the paradox. Pinkard says that if rational agents give themselves the law, then it seems appropriate to think that prior to giving themselves the law agents are in a state of lawlessness. Yet, the idea of a lawless agent coming to bind itself freely to the law, however, admits of a state in which the law does not obtain, though that very law is, by Kant’s lights, universally valid. Surely, it is 24 25 Pinkard 2002, p. 59. Ibid., p. 226. 11 paradoxical conception to think of oneself as not being subject to the law and then freely, but necessarily, binding oneself to it.26 Robert Stern has persuasively argued that Pinkard’s representation of self-legislation in (A) is not accurate.27 As Stern rightly points, Pinkard presents self-legislation in too “existentialist” a manner, making it a question of “understanding how the authoring of a law by a lawless subject could then bind that subject,” rather than a question of why we must think of the moral law as self-legislated or rationally binding as a result of an agent’s own will.28 Indeed, Kant explains in a footnote that although the law has a subjective source, it is a “law that we impose upon ourselves and yet as necessary in itself,” or, “as imposed upon us by ourselves it is nevertheless a result of our will” (G 4:401n). The necessity that Kant believes is contained in self-legislation is, as Stern sees it, the sort of self-imposed rational requirement that is contained in logical laws. As Stern explains, we, as rational beings, have no choice over which logical laws obtain. Though we can violate such laws through fallacious reasoning, we are rationally obligated to hold ourselves to their dictates.29 For Kant, to act without a reason is impossible, since agents impose reasons upon themselves in acting. But not just any reason is sufficient for acting, in the same way that not just any set of premises is sufficient to entail a valid conclusion. In order for an agent to act in a way that is universally valid, it is necessary that she impose the Pinkard admits to following Robert Pippin on this point (Pinkard 2002, p. 60), who phrases the paradox in the following way: “The idea of a subject, prior to there being a binding law, authoring one and then subjecting itself to it is extremely hard to imagine. It always seems that such a subject could not be imagined doing so unless he were already subject to some sort of law, a law that decreed he ought so to subject himself, making the paradox of this notion of ‘self-subjection’ all the clearer” (Pippin 2000, p. 192). Although in Pippin 2008 he still finds Kant’s view paradoxical, he ultimately relinquishes this formulation of the paradox. 27 This is a contentious point. See, again, Kain 2004. If it is not already clear, I think so-called ‘constructivist’ interpretations of Kant’s moral philosophy are off the mark. 28 Stern 2004, p. 7. For more on the rational necessity of the legislation of the law, see, for example, G 4:401n, where Kant stresses this point; cf. G 4:454, where Kant compares the faculty of freedom as self-legislation to the understanding or faculty of laws, a comparison that implies the necessity of the legislation of the law. 29 Stern 2004, p. 8. 26 12 moral law upon herself as a result of her own reason, and she can violate it if she chooses just like anyone can reason fallaciously.30 In a separate article, however, Pinkard makes clear that the version of the paradox that he endorses is not the one that Stern attacks: Kant himself also saw quite clearly, I think, that the very idea of some kind of existential moment in which we just freely will the moral law with no antecedent reason, or where we simply elect to impose the moral law is itself no more than fantasy.31 So, the point of the paradox seems to be even deeper than Stern suggests. I think this deeper notion is offered in the second half of (A): The paradox arises from Kant’s demand that, if we are to impose a principle (a maxim, the moral law) on ourselves, then presumably we must have a reason to do so; but, if there was an antecedent reason to adopt that principle, then that reason would not itself be self-imposed; yet for it to be binding on us, it had to be self-imposed. We can summarize Pinkard’s claims here in several incisive questions. What reason could there be for the imposition of the moral law, since it is itself a self-imposed rational obligation for consistency in one’s state of will? If there is a reason for the self-imposition of the moral law, doesn’t it follow that that reason (i.e., the reason agents are forced to legislate it) is more authoritative for an agent, or more deserving of self-imposition, than the moral law itself? And, if that reason is self-legislated, what reason does an agent have for legislating it? What these questions indicate is that Kant apparently fails to provide a satisfactory account of why an agent should treat the moral law like a logical law. Stern has helpfully shown how Kant intends it to be understood but not why the moral law must be thought to obtain in the Considering the moral law as an analog to logical laws is not inconsistent with Kant’s view of freedom. The violation of logical laws does not make one more free than if one adheres to them; rather, the violation of logical laws makes one irrational and their thinking unintelligible. Likewise, the moral law defines the intelligibility of freedom in a way very similar to the way in which logical laws define the intelligibility of thinking. For Kant, the moral law qua a priori principle gives rational beings a standard of reasoning that allows them to separate themselves from phenomenal causal necessitation and to determine their actions in accordance with pure reason. 31 Pinkard 2004, p. 27. Pinkard is, at times, not forthright in saying that he does not endorse the version of the paradox that Stern attacks, though I think Pinkard is, as the quotation makes clear, committed to rejecting it. 30 13 way described. On Pinkard’s view it is safe to assume that an agent needs a reason to treat the moral law like a logical law; however, if we offer such a reason it appears that that reason explains why we treat the law as having the authority of a logical law. Now, logical laws need no antecedent reason to explain their authority, but the important question just is why we should treat the moral law as self-legislatively authoritative. According Pinkard, then, Kant has not actually given us a reason to think that moral law possesses the authority that rational agents are supposedly obligated to attribute to it. We have now arrived at the central issue of the “Kantian paradox”, namely, the question as to why rational agents must treat the moral law as authoritative. To understand the paradox thoroughly this question requires more explication. We know that Kant is committed to the thesis that for the acts of a rational being to count as rational, those acts require reasons; those reasons may be either the self-imposed, rationally obligatory sort or they may be dependent upon some antecedent reason. Recall that Kant thinks freedom must be an unconditioned activity, dependent upon no antecedent reason.32 This means that an agent’s initial reason for acting must be a reason that she imposes upon herself – and, such a self-imposed reason, by its nature, has no antecedent reason. While Stern correctly points out that the reason rational agents impose the moral law upon themselves is because it is rationally obligatory (in a manner similar to logical laws), we must be able to explain why this is so. Here, things get tricky for Kant. If there is a reason that explains why the moral law is obligatory, then that reason must be self-imposed. The self-imposition of reasons is required by Kant’s system in two ways: (1) without it, Kant’s theory of freedom lapses into incoherence; (2) since the moral law is unconditional, there can be no antecedent reason for explaining why it is self-imposed. By contrast, in any other case, agents require reasons for their acts. As we saw in the discussion of 32 See Section III above. 14 maxims, Kant argues that rational agency just is to act on reasons. So, if Kant is to consistently apply his theory of rational agency to both the moral law and to all ‘other’ cases, then it seems that we need a reason to impose the moral law upon ourselves (or, what amounts to the same thing, to self-legislate the moral law). Yet, to have recourse to such a reason in relation to the moral law would entail that it is not authoritative; consequently, the very possibility of our freedom is destroyed. The question, then, reduces to the following: what reason do we have to view the moral law as authoritative other than that we must do so in order to consider ourselves as free? While both Allison and Wood take Kant to argue that a coherent conception of morality simply requires rational beings to consider themselves as free beings, the way in which Allison and Wood interpret Kant’s argumentative strategy in Groundwork (and beyond) plays right into Pinkard’s formulations of the paradox. The practical correlate of the theoretical idea of freedom is that of a rational agent obligating herself to standard contained in the categorical imperative or moral law. The Reciprocity Thesis, as a justification of the core of Kant’s moral philosophy, might show that freedom and morality are inextricably connected and that morality requires the idea of self-legislation, but that does not explain how such a concept of self-legislation avoids being non-paradoxical. That is, for the sake of argument we can grant that the self-legislation of the moral law might be required for a coherent conception of moral obligation. Yet, we can also admit without inconsistency that the conception of practical freedom understood as a rational agent obligating herself to the moral law for no reason other than she must for the sake of a coherent conception of morality is, in the end, somewhat unsatisfactory if not wholly so. Now, I am not suggesting that the conclusion is so unsatisfactory as to warrant a modus tollens rejection of Kant’s moral philosophy; in toto, that would be too hasty, as it may turn out that Kant’s 15 arguments are sound. However, it is sufficiently unsatisfactory to motivate one to think that somewhere along the way Kant made some mistakes.33 IV. Plugging the Derivation Gap: FH and The Reciprocity Thesis Although Allison and Wood agree that there is a derivation gap, they take divergent paths to plugging it. According to Wood, Kant only states the derivation of FUL as conditional on the justification of the later arguments for the FH. He believes, and he claims Kant believes this too, that there is no straightforward derivation of FUL from the analysis of the concept of a practical law, and that the heavy-lifting of his moral philosophy is performed by FH. Thus, on Wood’s account, FUL obtains only if FH obtains. According to Allison, Kant argues from the concept of a practical law in conjunction with his doctrine of transcendental freedom to the categorical imperative. This means that the deduction of the moral law in Groundwork III acts as an implicit premise of the derivation. I discuss Wood’s interpretation and then Allison’s. On Wood’s interpretation, the Formula of Autonomy consists in Kant’s claim that rational agents are “to choose only in such a way that the maxims of [their] choice are included as universal law in the same volition” (G 4:440), and it is this claim that explains the selflegislation of the moral law.34 This appears to be right, as Kant says that it is “[i]n accordance with this principle [FA]” that “all maxims are repudiated that are inconsistent with the will’s own giving of universal law” (G 4:431). In other words, it is Kant’s assertion that rational agents obligate themselves to choose to act only on those maxims that are consistent with their own giving of universal law. And there is no doubt that “the idea of the will of every rational being as a will giving universal law” (G 4:431) is of critical importance, for Kant says: 33 Both Pinkard 2002 and Pippin 2008 interpret the post-Kantian Idealists as unsatisfied with this consequence and engaged in the project of resolving the paradox. 34 See Wood 1999, pp. 160-165 passim. 16 If, then, there is to be a supreme practical principle and, with respect to the human will, a categorical imperative, it must be one such that, from the representation of what is necessarily an end for everyone because it is an end in itself, it constitutes an objective principle of the will and thus can serve as a universal practical law. The ground of this principle is: rational nature exists as an end in itself. (G 4:428-9) Boiled down: if there is a supreme practical principle of morality, then its “ground” is that “rational nature exists as an end in itself”. And, it is because rational nature exists as an end in itself that rational beings are obligated to repudiate all maxims that are inconsistent with the will’s own giving of universal law. Or, it is that the rational nature of rational agents exists as an end in itself, Kant thinks, that obliges rational agents to impose the universal law upon themselves. The most prominent interpretation of Kant’s claim about rational nature, which Wood ascribes to, is that that rational agents are rationally obligated to value their own rational nature.35 It is clear that puts a great deal of weight on this claim, as verified by the following passage: But suppose there were something the existence of which in itself has an absolute worth, something which as an end in itself could be a ground of determinate laws; then in it, and in it alone [my emphasis], would lie the ground of a possible categorical imperative, that is, of a practical law. (G 4:428) It is in FA, “and in it alone,” Kant says, that lies the ground of a possible categorical imperative. What Kant means is that the rational obligation an agent has to value her own rational nature is ground of a possible categorical imperative obtaining. Of course, Kant is clear that the moral law obtains because rational agents make it obtain as a result of their own rational will – for Kant, that is just what it means to be a self-legislating being. There must be, then, an entailment relation from the universally valid fact that every agent necessarily values her own rational nature to the self-legislation of the moral law. 35 See Wood 1999, Korsgaard 1996 and 2008, and Hills 2005. 17 However, even if we grant the claim that all rational agents do, of necessity, value their own rational nature, it remains unclear why one’s valuing their own rational nature specifically requires the self-legislation of the categorical imperative. What is the connection between the necessity of self-valuation and the self-legislation of the standard described in the categorical imperative of possessing a consistent state in what one chooses to will? Kant offers the following argument: The human being necessarily represents his own existence in this way [as an end in itself]; so far it is thus a subjective principle of human actions. But every other rational being also represents his existence in this way consequent on just the same rational ground that also holds for me; thus it is as the same an objective principle from which , as a supreme practical ground, it must be possible to derive all laws of the will. (G 4:429) Kant is arguing that if it is universally the case that every rational agent necessarily values their own rational nature, which he asserts is the case, then it is an objectively valid law that is selfimposed by the rational will itself. Thus, it is a priori that every rational values their own rational nature. From this conclusion Kant maintains that an agent is being inconsistent if she values her own rational nature and does not value the rational nature of every other agent because those agents all value their rational natures too. (This appears to be the meaning of his claim: “But every other rational being also represents…”) It appears, then, that we have a law in hand that explains why rational beings must consider themselves as legislators of it: it simply follows from the necessity of the valuation of one’s own rational nature. Notice that although Kant may have established a law of the rational will that is selfimposed, it does not justify that the categorical imperative that is to serve as the criteria for what an agent ought to will is also self-imposed. The objectively valid law that Kant has established says nothing more than that it is a priori that every rational agent necessarily values their own 18 rational nature. But that law is clearly not identical to either “act only in accordance with that maxim through which you can at the same time will that it become universal law” (G 4:421) or “choose only in such a way that the maxims of your choice are included as universal law in the same volition” (G 4:440). So, even if rational agents self-legislate the proposition that every agent necessarily values their own rational nature, it does not follow that they also self-legislate the categorical imperative. For that conditional to be true, we would need to see some connection between the necessity of self-valuation and the categorical imperative. Kant tries to support that conditional by an appeal to consistency in the state of the will: because I represent myself as of intrinsic value, and all others do as well, then I must represent them as having the same intrinsic value. The reason I must do so, Kant claims, is because agents are committed to valuing their rational nature “on the same rational ground”. The idea here is that it is inconsistent for an agent to value her own rational nature and not value the rational nature of another, for both appeal to the same rational ground, namely, their own absolute worth. However, why think that an agent is being inconsistent if she values her own rational nature absolutely and not the rational nature of another? There is a gap between the subjective necessity of self-valuation and the objective necessity of other-valuation. As Robert Pippin insightfully expresses that gap, Kant’s argument shows “at most that I must acknowledge the value of each person’s humanity to him or her, and not that I must value your humanity.”36 Put differently, Kant has merely shown that every rational agent necessarily values her own rational nature, and he has only stated that there is an inconsistency in valuing one’s own rational nature and not 36 Pippin 2008, p. 88. 19 valuing the rational nature of every other. He needs to show why it is inconsistent to not value the rational nature of every other.37 Not only is there still no clear relation between the necessity of self-valuation of one’s rational nature and the self-legislation of the categorical imperative, Kant appeals to the very claim he is supposed to derive from his grounding of FA as the basis of all objective practical laws, namely, the categorical imperative. Kant maintains that it is inconsistent for a rational being to value her own rational nature and not the rational nature of another. It is that inconsistency that is supposed to get us from the a priori proposition that every agent necessarily values their own rational nature to the command that every agent ought to value the rational nature of every other. Recall that the categorical imperative commands nothing more than a consistent state of the will: the only reason an agent ought to possess a consistent state of will is because it is an a priori, self-legislated command of her own rational will. However, it is only if an agent already self-imposes that requirement that she is genuinely obligated to recognize the equal worth of other rational beings, for it is that requirement that generates the inconsistency that Kant appeals to. Thus, the “on the same ground” appeal appears to be nothing more than an underhanded appeal to the categorical imperative.38 Clearly, the central issue here is why the command to remain consistent in one’s will is a self-obligated command. So, it would seem that we are left in a similar, if not worse, situation described by Wood in reference to the fallacious derivation of FUL. It remains unclear why the fact that every rational agent necessarily values her own rational nature requires the selflegislation of the categorical imperative, and, by Kant’s argument, the self-legislation of the 37 In a word, this is a summary of my critique of one of the more promising interpretations of FA by Alison Hills. See Allison Hills, “Rational Nature as the Source of Value,” Kantian Review (2005), vol. 10, no. 1, pp. 60-81. 38 To be fair, adequate treatment of Kant’s arguments for FA, and the interpretations of those arguments that exist in the literature, requires a book-length treatment rather than a few paragraphs of an essay. Nevertheless, the points raised have not, I think, been decisively addressed. 20 categorical imperative is assumed in order to provide the basis of the derivation of all practical laws. As it turns out, Wood thinks, there is good reason to think that Kant never intended any of the arguments for FUL or FA to be successful in the first place. He states: Throughout the First and Second Sections of the Groundwork, Kant repeatedly indicates that his search for the moral law is conditional or hypothetical. He is provisionally assuming that there are a good will, action from duty, and categorical imperatives binding on the finite rational will. He is seeking the principle of morality that follows from these assumptions. He promises to discharge the assumption in the Third Section by arguing that morality is not an illusion and there is indeed a categorical imperative binding on rational beings.39 Of course, if Kant assumes the self-legislation of the consistency standard contained in the categorical imperative and which is used to defend the centrality of FA, then he gets what he wants. Wood maintains that Kant provides the defense of the fact of self-legislation in Section III of Groundwork.40 This line of interpretation is, as Wood makes clear, borrowed from Henry Allison: The basis of Kant’s deduction of the moral law is what Allison calls the ‘Reciprocity Thesis.’ The reciprocity is a mutual entailment between the following two propositions: F: The rational will is free M: The moral law is unconditionally valid for the rational will.41 Admitting a more narrow interpretation than Allison, which is based exclusively on Groundwork, but that is essentially the same as Allison’s Wood argues “that F is a necessary presupposition of rational judgment”; that is, F M: “If the rational will is free, then the moral law is unconditionally valid”.42 On this interpretation, Kant’s moral philosophy is carried by the justification of the Reciprocity Thesis. Why? Because the Reciprocity Thesis shows that agents must recognize that their judgments are rationally self-binding because that is what a coherent 39 40 Wood 1999, p. 171. Wood 1999, p. 171. 41 Wood 1999, p. 171. 42 Wood 1999, p. xxi. Of course, the conditional runs the other way as well: M F. 21 conception of freedom necessitates. So, an agent must regard herself as the author of her obligations, as self-legislating or obligating herself to the categorical imperative, because without such a self-obligation both freedom and morality are chimerical. If the rational will is free as self-legislative, then the moral law is unconditionally valid for the rational will. Importantly, both Allison and Wood admit that there is a “gap” between the idea of the categorical imperative (“Conform your actions to universal law”) and the criterion of the rational will as self-legislating the requirement that it should only will those maxims consistent with universal law (FUL).43 And because Kant’s argument from FA to the derivation of the selflegislation of the categorical imperative is both obscure and ultimately fails, the justification of the Reciprocity Thesis is meant to do all of the heavy lifting for Kant’s moral philosophy (rather than FA, as Wood suggests). In other words, the concept of freedom itself requires that the categorical imperative is self-legislated, for self-legislation just is the practical correlate of the idea of transcendental freedom. Consequently, if agents are free, then they are self-legislative of the moral law; and, if they are self-legislative of the moral law, then they know that they themselves are the grounds of our obligations. To grasp how the Reciprocity Thesis is supposed to do this work, we must note the relation between Kant’s conceptions of self-legislation and freedom. As Pinkard nicely puts it, Kant maintains that the moral law has a claim on us independently of our inclinations and motivates us to act of itself, “while at the same time being something to which each agent and that agent alone binds himself.”44 With regard to the moral law’s claim on us independently of our inclinations, it must be that we can freely act in accordance with the moral law for no reason other than our respect for the law (i.e., in such a case, we are not being influenced by any 43 44 See, Allison 1990, pp. 210-213. Pinkard 2002, p. 51. 22 inclination whatsoever). The ability to act independently of inclination is where Kant’s account of transcendental freedom makes contact with his moral philosophy. Freedom, as it is described in CPR, is an “unconditioned causality” (A448/B476) or “a faculty of determining oneself from oneself, independently of necessitation by sensible impulses” (A534/B562). It is unconditioned because our reason demands the source of the causes and effects that occur within nature, where such a source would be a cause unto itself or unconditioned causality.45 Kant goes on to contend that we must view appearances as causally necessitated. However, because we can distinguish between the realm of phenomena (things as conceived according to the conditions of the possibility of experience) and noumena (things conceived apart from those conditions), we can consider ourselves as possessing a noumenal character not bound by the necessitation of appearances. This character cannot be known, however, since it is beyond the bounds of possible experience. Nevertheless, Kant maintains that the faculty of freedom is appropriately understood as an unconditioned causality, and must be presumed to exist as a practical necessity; without it we could not consider ourselves as morally responsible.46 Because freedom is the ability to act independently of sensible impulses, the very possibility of acting in accordance with the moral law requires that we have the presumed ability to choose against our inclinations. Choosing against our inclinations means to choose to act in a way that is motivated out of pure respect for the lawfulness of the moral law and its unconditional nature. In this sense, the moral law is a necessary condition of our acting freely, for it gives us a reason to act that is not from mere sensibility. Kant then states that the law is sufficient of itself to motivate our choice to act in 45 For more on the demands of reason, see A321/B377-A332/B389; with regard to freedom, see A532/B560A558/B586. See Allison 1990 for a thorough treatment of Kant’s views on freedom. 46 See G Section III. He also makes the much stronger claim that everyone thinks of themselves as having a free will (G 4:455). 23 accordance with it; this is equivalent to acting unconditionally because no subjective inclination factors into the motivation that the law provides. With regard to self-legislation, rational beings must see themselves as bound by the moral law, that they are responsible to it because they make themselves responsible to it. This selfbinding is a natural complement to Kant’s view of freedom, and it is one that, as we shall see, Pinkard ultimately works to his advantage. The state of being responsible to the law must not be externally imposed by some object, since there is no object sufficient to generate the law (for it obligates agents independently of any thing they desire or are inclined toward); this entails that an agent must understand herself as making herself responsible to the law. More clearly stated, the moral law obligates agents to act for universally valid reasons because agents obligate themselves to act in that way. Only such a self-imposed obligation gives them the rational space to choose against their sensible natures. This is, as it were, a requirement of freedom, one which agents must take to be true if they are to think of themselves as having the capacity for selfdetermination. So, since there is no object sufficient to generate the law other than the self of a rational agent, and to be free is for an agent to act spontaneously from herself, agents must consider themselves as generating the obligation of the moral law for the sake of their own freedom. Accordingly Pinkard, perhaps hyperbolically, describes Kant’s view as an agent’s “[k]eeping faith with the [self-legislation of the] moral law, almost as if it were not chosen by [her], all the while recognizing (however implicitly) [herself] as the author of that very law to which [she is] keeping faith.”47 Citing the infamous “fact of reason” passage from CPrR48 Pinkard asserts that, Pinkard 2002, p. 52. See CPrR 5:31-32, where Kant says, “Consciousness of this fundamental law may be called 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 48 47 24 for Kant, agents must believe that there is an unconditional, universally valid moral law because their transcendental freedom requires it.49 And, since moral responsibility depends upon an agent’s ability to do otherwise, she must presume her freedom for the sake of the very existence of morality. Presuming her freedom just means presuming she is self-legislative of the moral law. V. The Normative Robustness of Self-Legislation I will now develop my argument that Allison, Wood, and Pinkard are collectively mistaken in their interpretation of Kant according to which the latter justifies the derivation of the moral law through exclusive appeal to the Reciprocity Thesis. I intend to show how Kant’s conception of self-legislation is both non-paradoxical and sufficient to derive the moral law. An immediate question presents itself: how can Kant derive the moral law from self-legislation, if self-legislation is the very act of imposing the moral law? There is evidence, I believe, that Kant understands maxims as self-legislated, i.e., they are always treated as universally valid laws when they are chosen and acted upon. Kant’s crucial move, then, is to indicate the implicit rational requirement involved in every act of willing: the maxims which rational beings already treat as universal laws ought to actually be capable of becoming universal laws. Thus, the moral law or categorical imperative is nothing other than a rational requirement rational beings impose upon themselves in virtue of imposing maxims upon themselves. Maxims, Endorsement, and the Self-Legislation of the Moral Law proposition 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, as a positive concept, an intellectual intuition would be required, which certainly cannot be assumed here. However, in order to avoid misinterpretation in regarding this law as given, it must be noted carefully that it is not an empirical fact but the sole fact of pure reason which, by it, announces itself as originally lawgiving.” For more on this passage, see Section V below and Ameriks 2002, pp. 69-75. 49 Pinkard 2002, p. 59. 25 The task of this section is to dissolve the paradox. Kant, I believe, employs an argument for the rational necessity of the moral law; this argument neither makes justificatory appeal to the Reciprocity Thesis, nor does it undermine the authority of the moral law by relying on an antecedent reason for its legislation. Instead, the self-legislation of the moral law follows from the self-legislation of maxims. I interpret Kant to claim that, when an agent commits herself to a maxim, she endorses that maxim as reasonable for anyone to adopt. And, by endorsing a maxim as reasonable, she necessarily treats that maxim as if it were a universally valid principle of action, that is, a universal law. It is not necessarily the case, however that just because an agent treats her maxim as universal law that it is also consistent with its being a universal law. Kant argues that, if an agent’s treatment of a law is not consistent with its being a universal law, then she has adopted a rationally inconsistent maxim of action; in this case, she wills a maxim as universal law that cannot in any way actually become universal law. Thus, insofar as an agent desires to be rationally consistent, she ought to act only on those maxims that she can will to become universal laws. As I interpret Kant, the rational consistency the moral law requires of rational beings is itself sufficient to motivate them to obey it, in the same way that the rational requirements of logical laws are sufficient to motivate rational beings to obey them. Our freedom just is the ability to choose to be rationally consistent with regards to practical matters. It is here and only here that Kant’s account of the derivation of the moral law crucially depends upon his doctrine of freedom. From Section I, recall that, for Kant, to be a rational agent is (1) to have a reason for seeking some end and (2) to make a judgment about how best to attain that end; this results in (3) the formation of a rule of action for the attainment of said end based on that reason for acting, and (4) where that reason for acting is incorporated into the maxim in terms of its end. In acting 26 on a maxim one commits oneself to that maxim (i.e., the rule or principle that guides the action), which underlies Kant’s description of maxims as self-imposed rules of action. Given this, I must establish how it is that imposing a rule upon oneself results in the self-imposition of the moral law. By way of beginning, consider some of the shared properties of maxims and laws. Maxims, like the moral law itself, are intrinsically normative entities (e.g., rules or principles). This shared property is evident at Groundwork 4:421n, for example, where Kant contrasts maxims – defined as subjective principles of action (i.e., a principle valid for me) – with that of the moral law – defined as an objective principle (i.e., a principle valid for everyone). To make the contrast more explicit, one might say that a maxim is valid for me (i.e., because of some inclination of mine) whereas the moral law is valid for everyone (i.e., irrespective of inclinations). Moreover, Kant says that the moral law “must carry with it absolute necessity” (G 4:389). That is, it is valid for everyone because it is absolutely necessary. A maxim, then, is merely a rule that describes an action for a particular subject, while a law is a rule that prescribes an action for everyone. Additionally, although maxims are part of the normative structure of rational agency, no particular maxim need be adopted over any other, the moral law, on other hand, is necessarily self-legislated. If this is correct, then, how is it possible to transition from the self-imposition of maxims – an entirely subjective act – to the self-imposition of the moral law – a subjective act that necessarily holds for all rational beings?50 This argumentative strategy is strikingly similar to Kant’s description of the self-legislation of the categories in CPR. He says, “[T]he representation of a universal condition in accordance with which a certain manifold (of whatever kind) can be posited is called a rule, and, if it must be so posited, a law” (A113; Kant’s bolding). Notice that he divides ‘universal conditions’ into two sorts: rules and laws. Where a rule is universal, it is, unlike a law, not also necessary. In a separate paper I intend to develop the commonalities between Kant’s argumentative strategy in the first Critique (i.e., the objective validity of the categories is based on subjective conditions that rational beings import into judgments) and in his moral philosophy as I interpret it (i.e., the objective validity of the moral law is based on subjective conditions of practical judgments or maxims). 50 27 The self-legislation of the moral law can follow from the self-imposition of maxims only if the self-imposition of maxims is somehow itself an act of self-legislation. Put differently, unless rational agents build universality and necessity into their maxims, then self-legislation cannot follow from the self-imposition of maxims. That Kant expresses this view is, I believe, evident. For example, following his claim that our consciousness of the moral law is so “undeniable” that it constitutes a “fact of reason,” he says: The fact mentioned above is undeniable. One need only analyze the judgment that people pass on the lawfulness of their actions in order to find that, whatever inclination may say to the contrary, their reason, incorruptible and self-constrained, always holds the maxim of the will in an action up to the pure will, that is, to itself inasmuch as it regards itself as a priori practical. (CPrR 5:32; my emphasis) Why does Kant say that consciousness of the moral law is undeniable? He suggests that “one need only” analyze how rational agents make judgments about the lawfulness of their actions “in order to find” that their reason “holds the maxim of the will in an action up to the pure will”. Kant explicitly recommends we analyze the lawfulness that agents bestow upon their actions; only then is the moral law undeniable. Apart from this passage, there are several others where Kant either describes maxims as laws or suggests as much. In Groundwork, Kant says, “Thus the principle of every human will as a will giving universal law through all its maxims [durch alle seine Maximen …] would be very well suited to be the categorical imperative” (G 4:432; cf. 4:333). Of particular importance here is the fact that the categorical imperative itself is given as universal law “through all [the will’s] maxims” (my bolding). Notice that Kant does not specify that the categorical imperative is only given in some maxims. He explicitly states that the law is given through all the will’s maxims. Additionally, when Kant says “a will giving universal law,” he does not say, “a will giving the categorical imperative”. He is simply not speaking about the moral law here. Instead, 28 he is referring to the fact that all maxims have the character of being universal laws when an agent wills them. This interpretation is given undeniable support in a passage from Kant’s lectures on ethics, where he says: “A precept is an objective law, by which we ought to act; but a maxim is a subjective law by which we actually do act” (LE 27:1427; my emphasis). But why is just any maxim willed as a law? Consider Kant’s division of laws as such. In the Groundwork, Kant says, “Everything in nature works in accordance with laws” (G 4:412; my emphasis). For Kant, only those beings that are not merely in nature can act independently of natural laws; or, “Only a rational being has the capacity to act in accordance with the representation of laws, that is, in accordance with principles, or has a will” (G 4:412). Kant equates here laws and principles, identifying the possession of a will as the ability of a rational agent to represent to itself the laws or principles that it acts upon. This is in contrast to non-rational, sentient creatures, who act according to laws but are unable to represent that fact to themselves. Kant makes a similar claim in CPrR: “[P]ractical laws refer only to the will, without regard to what is attained by its causality, and one may abstract from this letter (as belonging to the world of sense) so as to have them pure” (CPrR 5:21). The relevant difference here is between laws of sensibility (or heteronomy) and laws of autonomy, or pure laws. The former occur, Kant says, when “the will seeks the law that is determine it anywhere else than in the fitness of its maxims for its own giving of universal law – consequently if, in going beyond itself, it seeks this law in a property of any of its objects” (G 4:440). Maxims (or mere principles) unfit to be universal laws must be, on Kant’s interpretation, laws furnished by one’s sensible nature. Consider yet another passage from Religion within the Boundaries of Mere Reason, where Kant says: 29 [T]hat freedom of the power of choice has the characteristic, entirely peculiar to it, that it cannot be determined to action through any incentive except so far as the human being has incorporated it into his maxim (has made it into a universal rule for himself, according to which he wills to conduct himself). (R 6:23-24; the second set of italics are mine) Here, Kant clearly states his ‘incorporation thesis,’ and, importantly, he attributes a universality to the maxim when the reason is incorporated into it. That is, when a reason is incorporated into a maxim, an agent then makes his maxim into a “universal rule for himself”. Notice the tension here. To make one’s maxim into a universal rule is to eliminate its subjective character. But, Kant immediately qualifies his statement about maxims qua universal rules as being rules that hold only ‘for oneself’ or subjectively. If a rule holds universally, that just means it holds for all subjects and not merely subjectively. The very idea of a subjective law or subjectively held universal rule is oxymoronic. Why think that agents actually do treat their maxims as laws, as objectively valid or universal and necessary? Recall that, according to Kant, when an agent chooses her maxim and acts upon it, she thereby commits herself to it. That is, there is no rescinding one’s reason for acting; once an agent chooses a maxim and acts, she has irrevocably committed herself to her principle of action and reason for acting. Furthermore, recall that for each maxim, an agent judges it to be a reasonable course of action for the end she desires to attain. An agent makes a judgment that a maxim is reasonable, and in virtue of this fact, she incorporates her reason for acting into the maxim. It is the judgment about the reason for acting on a maxim that forces an agent to commit herself to that reason as a reasonable means of attaining that end. Kant’s argument, as I interpret it, alleges that the commitment to a maxim includes not only a judgment about the principle as a reasonable means of attaining an end (i.e., means to ends are not selected arbitrarily), but it also includes a tacit endorsement, judgment of estimation, or approval of that 30 reason for adopting the principle.51 To judge a principle as reasonable therefore includes a judgment of estimation, endorsement, or approval of the reason for adopting the principle and the principle as a means of attaining the end. As a result, an agent’s approval of her reason for acting is equivalent to asserting that all other agents ought also to approve of one’s maxim of action for the very reason incorporated within it. From a prima facie perspective, of course, it may seem counterintuitive that in adopting a maxim an agent actually wills it to be universal and necessary, or demands that all other agents approve of her reason for acting and principle of action. After all, though I may sometimes consider my reason for acting and principle of action as being of a quality that it ought to be adopted by everyone else, it seems implausible that I always treat all of my reasons and principles as universal and necessary. What would prompt me to recognize that my judgments about my actions are universally and necessarily binding? Indeed, it may be that I never recognize any of my judgments about my actions as binding. As I interpret Kant, what makes an agent’s judgments about her actions universally and necessarily binding is the fact that she cannot help but approve of her own actions, and this act of approval forces an agent to authorize her reasons for others. When an agent acts on a maxim, not only does she commit herself to the approval of her reason for acting and her principle of action for herself, she is implicitly committed to the approval of her reason and course of action for everyone. Although some reason and principle are mine insofar as my acting on it depends upon my reason for doing so (and, in this sense is subjective), it is not an intrinsically private reason because it can be, by my lights, legitimately acted on by anyone that could offer that same reason The theory of endorsement that I am attributing to Kant is not that which Korsgaard attributes to him (see Korsgaard 1996, pp. 49-130). What she calls “reflective endorsement” is the means rational being use to distance themselves from their inclinations by considering them as reasons for acting. I do not think that my interpretation is at odds with hers in essential respects; however, if my interpretation is correct, then endorsement is, for Kant, more primary and does more work than Korsgaard thinks. 51 31 for said action. This point is crucial: it would be entirely inconsistent of me to disapprove of an action that any other person chooses to undertake if they share my reason for acting. I am thus forced to approve of my reason as worthy of esteem, in addition to any chosen course of action that might satisfy that reason, for anyone that shares my reason for acting. Here, one might object that there are simply are no reasons for acting that are the same. What individuates and privatizes a reason for acting is that said reason is possessed by me. It is the possessive aspect of a reason that then allows an agent to, for example, approve of one’s own act of telling a lie in a situation and disapprove of another agent’s act of lying in a similar or identical situation. To say that a reason is mine (and only mine) is to say that it is not, and cannot be, your reason; moreover, it is for me to not recognize your reasons as binding. To counter this objection, consider again the acts of approval and disapproval. If I act in order to bring about my happiness and you act to bring about your happiness, the propositional contents of our reason for acting are identical (i.e., happiness), and the reasons are, therefore, only individuated according to their being indexed to different subjects (i.e., ‘my happiness’ and ‘your happiness’). But that indexical tie is broken in my implicit or explicit act of approving or disapproving of your reason. If I approve of your act, I have publicized my reason; that is, I have extended it to you the approbative nature of happiness as a justificatory condition for your act.52 If I disapprove of your reason for nothing more than the indexical tie of the reason’s content, I have nevertheless publicly authorized my reason (i.e., my happiness). The reason for this publicity is that I have committed myself to the proposition that the only genuine and justified reason you could have to act is whatever reason I impose upon you; that is, in this case, my happiness ought to be your reason for acting (or refraining from acting). What these Incidentally, Kant claims that “all material principles, which place the determining ground of choice in the pleasure or displeasure to be felt in the reality of some object, are wholly of the same kind as they belong without exception to the principle of self-love or one’s own happiness” (CPrR 5:22). 52 32 considerations speak to is that I cannot but help to publicize or universalize my reason’s for my actions whenever I approve of them or disapprove of the reasons of another.53 Thus, in every imposition of a maxim, an agent cannot but approve of her own reason for acting and her course of action. And, to approve of her reason for acting is to also approve of any other agent’s appeal to that reason as justification for undertaking some reason-satisfying course of action. If I disapprove of this same reason for anyone else to act upon, then I am inconsistent: I approve and disapprove of the very same reason. This is what Kant means in saying that a maxim is a universal rule that an agent makes for herself when she incorporates her reason for acting into the maxim: agents cannot but help universalize their maxims. If the argument above is sound, Kant has, at most, only established that agents treat their maxims with universality but not necessity. For an agent to treat her maxim with both universality and necessity, she must will it to hold for all other agents, and in a way that eliminates all contingent, subjective conditions. There is, however, a definite worry here. To eliminate all subjective conditions appears to demand that, whenever an agent finds herself in some situation and performs some act, she obligates all other agents to follow her rule if those agents find themselves in the same situation. The reason this obligation seems too strong is due to the fact that even if two agents share the same reason for acting, it not clear that they can act on the same rule. For example, suppose I have a friend, Barney, who has become virtually incapable of telling lies because past experience has shown to him that such acts weigh too heavy on his conscience and make him unhappy. As for myself, I can lie without hesitancy because I do not have such weights on my conscience. After having told a lie about something of significance, I reflect on the fact that Barney, if he were in the same situation as myself, probably 53 This is a similar, but not identical, argumentative strategy to Korsgaard 1996, pp. 132-145. I do not take my argument here to be decisive. As I stated at the outset, my goal in this paper is not to defend (my interpretation of) Kant’s arguments against all objections, but merely to show how I understand the arguments to work. 33 wouldn’t have been able to bring himself to tell the lie. It is that “subjective condition” about him that makes me except him from the obligation of having to tell the lie in the same situation in which I told the lie. It is precisely because agents recognize the differences between one another that they may not consider their reasons for acting and courses of action as holding for all other agents. There are two points to make. First, this ‘happiness principle’ as a ground of motivation for actions cannot withstand scrutiny. In the example given our shared reason requires contradictory acts, and in other situations it may require contradictory and conflicting acts, e.g., two agents who are obligated by their shared reason to perform acts that have conflicting ends. When I endorse my reason for acting, my approval of that reason entails the very problematic consequence that I might be obligated to approve of courses of action that are in direct conflict with my reason for acting, i.e., my happiness. This is obviously true when we consider that, perhaps, there is some person who’s happiness involves depriving me of my happiness. Does this consideration entail that I necessitate my maxim? In a certain sense it does suggest this. Because I can recognize that the happiness of others might conflict with mine, if I want to keep my happiness as my ground for acting, then I am obligated to disapprove of any act that does not cohere with my happiness. Those acts that violate my reason for acting must be disapproved because they are in violation of my criterion of approval (i.e., my happiness). So, I do not necessitate my act of lying per se; rather, I necessitate the act (and any act, for that matter) insofar as it is consistent with my reason for acting. What this suggests is that whenever an agent gives herself a reason for acting, she necessitates an entire system of ends that is consistent with that reason. I propose this is what Kant means when he says, “The concept of every rational being as one who must regard himself as giving universal law through all [my emphasis] the 34 maxims of his will, so as to appraise himself and his actions from this point of view, leads to a very fruitful concept dependent upon it, namely that of a kingdom of ends” (G 4:433). Given the universalization requirement that I have claimed is central to the selfimposition of a maxim, I interpret Kant to be claiming that his requirement implies the universalization of binding nature of the maxim. When an agent binds herself to her reasons and principle, she binds all other agents to her reason for acting, and to her course of action, insofar as those other agents are obligated (by the former agent) to recognize that her reason for acting justifies her course of action. In short, the self-imposition of a maxim is for an agent to demand the approval of her reason for acting and subsequent course of action; that is, it ought to be approved of. This is not to say that she requires that all other agents act in the same way; that is too demanding of a claim, since not all agents share the same roles or enter into the same situations. Rather, she demands that all agents approve of her reason for acting and the course of action undertaken to satisfy that reason (regardless of whether they ever find themselves in said situation or act in that way). So, even if agents do not share the same reason for acting, they each command the approval of their own reason for acting and the course of action they intend to undertake to satisfy that reason. Consequently, an agent cannot consistently disapprove of any other agent’s reason for acting and course of action if their reason is shared. What is more, agents demand the approval of their reasons for acting and the course of action they undertake to satisfy that reason. In this sense, agents necessitate their maxims. From the above considerations, we can see that, for Kant, for an agent to adopt a maxim is to endorse it; to endorse a maxim is for her to approve of it as a satisfactory course of action in relation to her reason for acting; to endorse a maxim is, therefore, for her to approve of her reason for acting; the act of approval of a reason is for an agent to make that reason universally 35 available to act upon; furthermore, the act of approval is to demand that everyone act for that reason. To approve of one’s reason for acting is to publicize one’s to endorse a maxim is to treat it as if it were a principle of action that everyone ought to approve of. That is, in adopting a maxim a rational being legislates it as a universal law such that everyone ought to act according to that principle if they possess the reason for acting and fined themselves in the same situation. This tacit obligation follows from the claim that to endorse one’s own maxim just is to assert that the reason for adopting it is a universally valid reason; that is, that any agent ought to recognize the maxim as a reasonable and worthy course of action for attaining that end. The view, then, that I ultimately mean to attribute to Kant is the following: in adopting a maxim, an agent always already universalizes it and demands its approval through her endorsement of it.54 Uncovering the implicit universalizing act of approval in the adoption of a maxim is, I believe, what Kant takes to be the real breakthrough in his moral theory. Though previous moral theorists had correctly recognized “that the human being is bound to laws by his duty,” they failed to discover the principle of morality because “it never occurred to them that he [i.e., the human being] is subject only to laws given by himself but still universal and that he is bound only to act in conformity with his own will” (G 4:432). Using this passage to justify my interpretation requires some explanation. First, Kant’s use of “laws” here describes neither the moral law nor laws willed to be in conformity with the moral law. The use of the plural “laws” rules out reference to the moral law. Furthermore, the moral law is, according to Kant, a specific kind of law, namely, “the sole law that the will of every rational being imposes upon itself […] without having to put underneath it some incentive or interest as a basis” (G 4:444). In other words, the It might be objected that one can act for a reason and yet not approve of their own reasons for acting. Suppose I think that lying in order to save my job is wrong, i.e., that it is not worthy of approval, and yet I perform the act of lying in order to save my job. I think the obvious response here is that I am simply acting irrationally. I am performing an act that I think should not be approved of, yet implicitly approving it in the very act. 54 36 moral law is a law without subjective interest, or a law not based on any particular reason that an agent possesses exclusively. Lastly, he cannot be describing the lawfulness that maxims possess when willed in conformity with the moral law. To see why, notice that he says “laws given by himself but still universal” (my bolding), and adds that agents are “bound” to act in conformity with their own wills. In the case of the former, by their very nature laws are universal, so Kant clearly intends the qualification to highlight a feature other than their universality. What he means to emphasize, I believe, is that maxims are subjectively valid rules which are always treated as laws. In the case of the latter, agents are so bound because the will is, as we already seen Kant say, the faculty of determining agents in accordance with the representation of laws. In short, the feature Kant has in mind is that agents will the universality and necessity of their maxims in an act of endorsement. Kant’s breakthrough, then, amounts to the claim that, although agents act for a reason and on principles that are specific to that reason, that specificity does not entail that said reason and principle fail to be as universally valid entities. While an agent may not know she is treating her reason for acting as universally valid, Kant maintains that she implicitly does so in endorsing her reason for acting, and therefore, her principle of action. Furthermore, an agent does not justify those items only for herself or for some set of other agents; she justifies them for anyone who could share them. Given that inclinations are entirely contingent entities, i.e., they shift and change due to internal reflection or external reasons altogether, anyone could share another person’s reasons for acting whether or not they do so in fact. Therefore, according to Kant, an agent treats her subjective principle of action as if it were a universal law. What is the same, she treats her principle of action as one that anyone could adopt for the reason it was acted upon, and it ought to be approved of because of that reason. 37 Kant’s next step involves showing that not all self-legislated laws are capable of becoming universal laws, regardless of how we treat them. Consider Kant’s well-known declaration that ‘ought implies can’. If it is true that, ‘Acting to secure one’s happiness is to be approved, and, if an act of lying about reckless negligence in order to keep one’s job satisfies one’s desire to be happy, then said act in that situation ought to be approved,’ then I am also implicitly committed to the claim that everyone can, in principle, approve of my action and adopt it as their own. But, as I understand Kant’s reasoning, if it is not possible that everyone could (at least in principle) act on that maxim, then that maxim ought not be approved. So, the task remains to determine how it is possible that everyone could act on some specific maxim. Kant’s strategy, I believe, is quite simple: if an agent is already treating her maxim as a universal law of action, then it must be capable of actually being a universal law of action. As it turns out, not all agents can adopt just any maxim without such an adoption resulting in incoherence.55 Again, consider the previously employed example: if I will that everyone adopt my maxim (‘I shall lie about my reckless negligence in order to keep my job’, then I consequently will that everyone ought to adopt my maxim if they share my reason, and that, therefore, everyone can adopt my maxim if they share my reason. So we are assuming that everyone shares my reason for acting and adopts my principle. However, if it was universally the case that whenever anyone self-imposed my maxim (‘I shall lie about my reckless negligence in order to keep my job’), then it would be universally known all acts due to reckless negligence are to be followed by a lie. The problem that arises, of course, is that one’s employer would know that a lie is being told and act appropriately. The end one desires to achieve in lying is thwarted when the principle of action is consciously and reflectively examined in regards to its already existing and implicitly imposed universality. In such a world – a world that tacitly trails one’s self55 For more on the universalization tests, see Wood 1999, pp. 84-89. 38 imposition of a maxim – the means of attaining my end is simply not possible. Accordingly, not everyone can adopt my maxim so it ought not be adopted as a matter of rational necessity. On the view I have presented, Kant argues that because one judges their maxim to be universally valid, for the sake of rational consistency one’s maxim must actually be universally valid. What this means is that, as rational beings, to be rationally consistent we require ourselves to “act only in accordance with that maxim through which you can at the same time will that it become universal law” (G 4:421). The requirement of the moral law is, therefore, not only a moral requirement of how we are to act; it is in fact a moral requirement that follows from what is essentially a rational requirement that follows from the fact, according to Kant, that we always treat our maxims as laws despite whether it is coherent that such laws can actually be laws. The rational requirement of the categorical imperative is not unlike the logical imperative of holding a consistent position free of self-contradictions.56 What the moral law requires is that (a) “you can at the same time” (my bolding) or “in the same volition” (G 4:447) endorse a maxim as a universal law and (b) its content is appropriate for becoming a universal law. (Why would Kant so frequently use the locution “at the same time” if he were not referring to the universality latent in every willing of a maxim?) Of course, we do not take rational requirements like consistency to be imposed from a source external to our own reason, and, in the exact same way, Kant does not understand the rational requirement of the moral law to be found in a source external to our own reason. Hence, Kant refers to a maxim’s possessing a “fitness”57 to be universal law in addition to using the language of conformity. Maxims are already treated as universal laws when chosen and endorsed. So, the appropriate question that one must ask oneself is whether one’s maxim possesses “fitness” to become a universal law – whether the matter or content of the 56 57 Or, to continue with Stern’s example, if one desires to be rational one will heed logical laws. See G 4:444. 39 maxim conforms to its antecedently imposed form. If an agent adopts a maxim unfit to become universal law, then that agent is adopting an inconsistent rationale for their acting in such and such a way. They are endorsing a maxim and tacitly committing themselves to the claim that everyone ought to, and therefore, could act on that maxim if they so possessed a reason for doing so, when, in fact, not everyone can do so given that reason. There is simply no better way to state Kant’s point than that one is being rationally inconsistent: it is making a demand of everyone that not everyone can possibly satisfy. The moral law is, as it were, always already endorsed in acting on a maxim, binding “us a priori and unconditionally by our own reason” because we implicitly legislate it in endorsing a maxim (MM 6:227). Kant is thus led to make what is perhaps his most well-known statement about self-legislation: In accordance with this principle all maxims are repudiated that are inconsistent with the will’s own giving of universal law. Hence the will is not merely subject to the law but subject to it in such a way that it must be viewed as also giving the law to itself and just because of this as first subject to the law (of which it can regard itself as author [Urheber]). (G 4:431) There are some things to note about this passage. The first part of it is essential, for the inconsistency is not between the maxim, on the one hand, and the universal law, on the other. The inconsistency occurs between the maxim and the will’s giving of universal law (note: not the universal law, but the will’s giving of it); the inconsistency, we see, is a byproduct of a maxim’s being treated as a universal law when it is endorsed even though its content is not capable of being instantiated as a universalized law. For this reason an agent is first subject to the law, because she subjects herself to the law in the necessity of the universalization of a maxim in endorsing it. Yet, she also binds herself to the law as an act of reason because the agent 40 herself performs the act independently of any subjective inclination when she makes a judgment about her principle and endorses it. It is worth noting that the German Urheber, which Kant uses for “author,” also means “originator” or “initiator,” the latter of which is a particularly felicitous term for my interpretation of self-legislation. On my view, rational agents initiate the law in imposing upon themselves their maxims of action. They do not create it, but load it into their practical reasoning. The standard translation of Urheber as “author” also connotes ‘endorsement,’ in the sense that one ‘authorizes’ another to speak on their behalf. Indeed, this connotation and my interpretation of ‘authorship’ as implicit commitment to the lawfulness of all of one’s maxims can be seen to be particularly useful in interpreting what is perhaps the most difficult passage regarding self-legislation that Kant offers. In MM, he says: A (morally practical) law is a proposition that contains a categorical imperative (a command). The one who commands (imperans) through a law is the legislator (legislator). He is the author [Urheber] (autor) of the obligation in accordance with the law, but not always the author [Urheber] of the law. In the latter case, the law would be a positive (contingent) and arbitrary law. The law which obligates us a priori and unconditionally by our own reason can also be expressed as proceeding from the will of a supreme legislator, i.e., one that has only rights and no duties (hence from the divine will), but this only signifies the idea of a moral being whose will is a law for everyone, without his being thought of as the author of it.58 This is a confusing passage for a couple of reasons. First, Kant refers to arbitrary and contingent laws, a seemingly contradictory idea. Second, he makes the bizarre distinction between a selflegislator as the author of both (a) the obligation in accordance with the law, but not necessarily (b) of the law. How can one be an author of the obligation in accordance with the law but not be the author of the law itself? Kant even suggests that a supreme legislator is not the author of the law, but, like us finite legislators, he is only the author of the obligation in accordance with the law. How does one even begin to make sense of this passage? 58 Cf. LE, 27:283. 41 First, notice that Kant is not discussing the moral law until he says, “The law which obligates us a priori…” Prior to that, he is discussing morally practical laws as propositions that “contain” a categorical imperative or a “command”. If he were talking about the moral law, it would be more appropriate for him to identify the law with the categorical imperative. Instead, he is merely discussing moral laws and imperatives as such. This is helpful for understanding his bizarre statement about the contingency and arbitrariness of some laws. Unless we assume that Kant is speaking of maxims unfit to be universal laws, I am not sure we can make sense of his statement, for, as I have explained above, all maxims are laws, and only those maxims that are unfit to be universal laws are contingent and arbitrary because they are produced by subjective interests and incentives, which is what makes those laws arbitrary. Now we can make sense of the bizarre distinction. We are not the originators (or authors) of these maxims because they are produced through subjective interests and incentives; that is, externally to our wills. However, all legislated laws contain imperatives, including those that are unfit to be universal laws. In this sense, we are the originators (or authors) of the obligation in accordance with the law, for we produce this obligation in imposing laws upon ourselves as maxim of action. Because we are rational beings, we recognize that the contents of those willed laws (maxims) must be consistent with their form as laws, unless we want to will an inconsistency. This just is to legislate the moral law. Thus, we legislate the (moral) law even when those laws (maxims) we adopt are not fit to be laws. Finally, this explains why a supreme legislator or divine will provides the best expression for the law that obligates us unconditionally of our own reason: all of his self-legislated maxims are at the same time fit to be universal laws. But, as Kant notes, this supreme legislator possesses a will that is a law for everyone, insofar as he only wills laws, 42 but he himself does not originate the obligation of the law, for the obligation of the law is a selfimposed obligation in the nature of our very reason. At this point sufficient work has been done to show that the Kantian paradox is built upon a misinterpretation of Kant’s account of self-legislation, and that Kant’s strategy of justifying the ‘undeniable’ nature of the moral law is not exclusively through an appeal to his account of freedom. By way of review, Pinkard had argued that, for Kant, agents always act for a reason. Indeed, this is true. He then argued that, as rational beings, we require a reason for the selfimposition of the moral law. Yet, if he is right, then the paradox emerges: if we have an antecedent reason, then that reason is more authoritative than the law itself; and, if we lack a reason, then we are performing an act that is apparently contrary to our rational nature. Like many other scholars, Pinkard believes that Kant does not have a direct argument for the selflegislation of the moral law. Instead, Pinkard, like Allison and Wood, believe that Kant justifies the self-legislation of the moral law by appeal to his account of freedom; that is, thinking of ourselves as morally responsible requires freedom, and freedom requires the self-legislation of the moral law. What Pinkard, Allison, and Wood have failed to take into account, however, is that Kant believes that the moral law is self-legislated through the self-imposition of maxims. In judging maxims as reasonable and then acting on them, agents commit themselves to endorsing their reasons for acting. Kant understands an endorsement of a maxim is to assert that it is adopted for a universally valid reason; that is, every agent ought to, and can, act on that principle for the reason endorsed. It is because we necessarily treat our maxims in this way that the moral law becomes a self-imposed requirement. The content of the moral law is nothing other than the rational requirement to be consistent in what we will. Because we always treat our maxims as universal laws in imposing them upon ourselves, the moral law arises from that part of our 43 rational nature that forces us to question whether what we treat as a universal law (the maxim) is capable of actually becoming universal law. So, insofar as we are rational, we impose rational requirements upon ourselves; and, insofar as we always will our maxims as universal laws, we will the rational requirement of the moral law, namely, that the contents of our maxims, which we treat as universal laws, actually be capable of becoming universal laws. Of course, we need not obey that requirement any more than we need to obey the laws of logic. I can make invalid arguments to my hearts content, but only if I desire to be rational do I seek out valid arguments. Likewise, I can endorse any maxim as universal law, but only if I desire to be rational do I seek to ensure that my maxim can actually be a universal law. Nothing compels us to obey the law other than our desire to exercise our rationality in accordance with its own demands for consistency in what we already will. But, it is surely the case that many of us do desire to be rational for no other reason than that we impose that requirement upon ourselves.59 References Allison, Henry E. (1990). Kant’s Theory of Freedom. Cambridge: Cambridge University Press. Ameriks, Karl. (2000) Kant and the Fate of Autonomy: Problems in the Appropriation of the Critical Philosophy. Cambridge: Cambridge University Press. Johnson, Robert. (2008) “Kant’s Moral Philosophy,” Stanford Encyclopedia of Philosophy. Kant, Immanuel. (1996a) Practical Philosophy. Mary J. Gregor, trans. & ed. Cambridge: Cambridge University Pres. ––––. (1996b) Religion and Rational Theology. Allen W. Wood & George Di Giovanni, trans. & ed. Cambridge: Cambridge University Press. ––––. (1997a) Critique of Pure Reason. Paul Guyer & Allen W. Wood, trans. Cambridge: Cambridge University Press. 59 I am indebted to Professor T. Allan Hillman for his comments on this essay. 44 ––––. (1997b) Lectures on Ethics. Peter Heath & J. B. Schneewind, ed., Peter Heath, trans. Cambridge: Cambridge University Press. Kain, Patrick. (2004) “Self-Legislation in Kant’s Moral Philosophy,” Archiv fur Geschichte der Philosophie, vol. 86, no. 3, pp. 257-306 Kerstein, Samuel. (2002) Kant’s Search for the Supreme Principle of Morality. Cambridge: Cambridge University Press. Korsgaard, Christine. (1996a) Creating the Kingdom of Ends. Cambridge: Cambridge University Press. Korsgaard, Christine. (1996b) The Sources of Normativity. Cambridge: Cambridge University Press. Pinkard, Terry. (2002) German Philosophy 1760-1860: The Legacy of Idealism. Cambridge: Cambridge University Press. Pinkard, Terry. (2004) “Response to Stern and Snow,” Bulletin of the Hegel Society of Great Britain, vol. 49/50. Pippin, Robert B. (2000) “Hegel’s Practical Philosophy: The Realization of Freedom,” The Cambridge Companion to German Idealism, ed. Karl Ameriks. Cambridge: Cambridge University Press. ––––. (2008) Hegel’s Practical Philosophy: Rational Agency as Ethical Life. Cambridge: Cambridge University Press. Stern, Robert. (2004) “Pinkard on German Idealism,” Bulletin of the Hegel Society of Great Britain, vol. 49/50, 7 Wood, Allen W. (1999) Kant’s Ethical Thought. Cambridge: Cambridge University Press. 45