II. Duties of Virtue and the Obligation to Be Virtuous

In the “Metaphysical Foundations of the Doctrine of Virtue,” the second half of The Metaphysics of Morals, Kant distinguishes between specific duties of virtue (Tugendpflichten), such as the duty to develop one's own talents or to be beneficent to others, and a general obligation of virtue or obligation to be virtuous (Tugendverpflichtung).Footnote ¹ He expresses this distinction by characterizing the latter as “formal” and the former as “material,” and he infers the singularity of the general obligation to be virtuous from its formality, and the multiplicity of the specific duties of virtue from their materiality. Kant uses the distinction between “formal” and “material” throughout his philosophy for many purposes, so it is not immediately clear what this means; but as I have already indicated, it could suggest that the obligation to be virtuous is a vacuous addition to our specific duties of virtue. So we must first see what Kant means by this distinction, and then explain why it does not imply the vacuity of the concept of a general obligation to be virtuous.

We can begin to see what Kant means when he first makes the distinction between material and formal duties, and the inference to the plurality of the former and the singularity of the latter, in section II of the introduction to the “Doctrine of Virtue.” He starts by distinguishing what he here calls “ethical obligations” (a general term that does not yet distinguish between specific duties of virtue and the obligation to be virtuous) from the juridical obligations or duties of right that have already been discussed in the first half of The Metaphysics of Morals, the “Doctrine of Right.” He makes this distinction by means of the criterion that duties of right are accompanied with rights on the part of specific persons to coerce their fulfillment,Footnote ² while ethical obligations do not permit coercive enforcement:

But following his dash, Kant then introduces the distinction within the class of noncoercively enforceable ethical obligations with which I am concerned:

The specific duties of virtue are material and plural because they are duties that are also ends, or duties to have and pursue ends: in the first instance, the two comprehensive ends of one's own perfection and the happiness of others,Footnote ⁴ and then the more specific ends that comprise those two comprehensive ends, such as the duties to preserve and cultivate one's own physical and mental capacities (in the first case) and the duties to practice beneficence, gratitude, and sympathy toward others while refraining from arrogance, defamation, and ridicule toward them (in the second case). The ethical obligation that is not one of these material and specific duties, however—or what seems to be the duty to have a virtuous disposition—is formal in the sense that it concerns not what is to be done but, so to speak, how it is to be done, that is, the kind of motivation from which the more specific ethical duties are to be fulfilled. This formal obligation is singular rather than plural because there is only one way to have a truly virtuous disposition, namely, to act “from duty,” or out of respect for the moral law itself.Footnote ⁵ Kant emphasizes this point by describing the categorical imperative that is to be the content of virtuous motivation as itself a formal principle: “The concept of duty stands in immediate relation to a law (even if I abstract from all ends, as the matter of the law). The formal principle of duty, in the categorical imperative ‘So act that the maxim of your action could become a universal law’ already indicates this. Ethics adds only that this principle is to be thought as the law of your own will and not of will in general.”Footnote ⁶ It is something about the will in general that requires us to adopt the specific material ends that are also duties (namely, self-perfection and the happiness of others), but ethics adds the obligation that we be virtuous by making the moral law itself our fundamental motivation. This obligation seems to be formal precisely in the sense that it does not require any further specific actions (“matter”) of us, but requires only that we fulfill our specific ethical duties from one sort of motivation rather than another. As we will eventually see, however, this impression is misleading.

In a later section of the introduction to the “Doctrine of Virtue” in which he discusses “Concepts Preliminary to the Division of the Doctrine of Virtue,” Kant also presents the contrast between the plural duties of virtue and the singular obligation to be virtuous as a contrast between material and formal duties, although here he uses his all-purpose distinction between the formal and the materialFootnote ⁷ in two different ways. He first uses it to distinguish all ethical obligations of any kind from duties of right, and only then uses it to contrast the specific duties of virtue with the general obligation to be virtuous. Thus:

Here Kant claims that ethical obligations, unlike duties of right, cannot permissibly be coercively enforced, because they are duties to adopt certain ends or policies, such as being beneficent, and therefore they do not specify specific types of actions that must always be performed or omitted, as fulfilling contracts may always be required or expropriating property without consent may always be prohibited. Instead, ethical obligations leave open to contextual determination when and how they are to be fulfilled: for example, to whom and in what ways one should be beneficent, since one obviously cannot be beneficent to everybody and in all ways. The crucial difference between duties of right and ethical obligations, in other words, arises from the fact that the former are themselves in a sense formal duties—they arise from the necessity of preserving equal spheres for freedom of external action for all persons regardless of what anyone's particular ends areFootnote ⁹—while the latter are material duties—they are duties to adopt and promote particular ends. The relative determinacy of the formal duties of justice (that is, the relative specificity of what it takes to preserve equally maximal spheres of freedom of action) may be contrasted with the relative indeterminacy of the material duties of ethics (that is, the relative indeterminacy of what it takes to realize such ends as self-perfection and the happiness of others). And this contrast is at least part of what, in turn, explains why duties of justice are enforceable through coercion, while duties of virtue are not—for specificity is a necessary condition of enforceability. (But it may not be a sufficient condition for coercive enforceability: a moral justification for the coercive enforcement of a duty is always necessary, even if what is required by such a duty is specific enough for the attempt to enforce it coercively to make sense in the first place.)Footnote ¹⁰

This argument may seem only to strengthen the impression that all ethical obligations are reducible to the material obligations of duties of virtue. However, in going on to make a third point, Kant now reuses the contrast between formal and material in a different sense, in order to draw the contrast between specific duties of virtue and the general obligation to be virtuous in the same way as he had in section II of the introduction to the “Doctrine of Virtue”:

Again, the obligation to be virtuous is necessarily singular because it is the obligation to act out of a certain motivation, namely, respect for “law as such” or for the moral law, and there is only one such law, while the moral law itself gives rise to a number of specific duties to adopt ends that cannot be coercively enforced. Kant here emphasizes the difference between the general obligation to be virtuous and the specific duties of virtue by remarking that the general obligation to be virtuous could, and in some sense even should, be the motivation for fulfilling duties of right. Nonetheless, of course, while the outward fulfillment of the latter can be coercively enforced, the satisfaction of the requirement to have this inward motivation cannot be: the “virtuous disposition … extends to duties of right as well although they cannot, because of this, be called duties of virtue.” Kant had also made this point in the general introduction to the whole Metaphysics of Morals. Footnote ¹²

Thus, Kant distinguishes between the general obligation to be virtuous (as the formal obligation to make respect for the moral law one's fundamental motivation) and the specific duties of virtue (as the material obligations to make one's own perfection and the happiness of others one's mandatory ends, and the even more specific requirements for implementing these goals). But this distinction only seems to force upon us the question whether it makes any sense to speak of the former as itself a duty or obligation, that is, whether it makes sense to say that one has a duty to fulfill one's duties in a specific way, out of a specific motivation. It might well seem that it must be either redundant or potentially infinitely regressive to say simply that one has a duty to fulfill one's duties. That is, it might well seem that to describe something—whether juridical, such as the requirement to fulfill one's contracts, or ethical, such as the requirement to be beneficent—as a duty is already to say that one is obliged to fulfill it, and that it would either add nothing or else add a potentially infinite repetition to say that one has a duty to fulfill one's duties (that is, that one has an obligation to fulfill one's obligations). To call them duties is already to say that their fulfillment is mandatory, and there seems nothing more to say. And if there is anything more to say, why not keep saying it?—that is, why not say that if one has a duty to fulfill one's duties, then one has a duty to fulfill one's duty to fulfill one's duties, and so on?

To be sure, Kant is not simply saying that one has a general duty to fulfill one's specific duties; rather, he is saying that one has a general obligation to fulfill one's specific duties in a particular way, namely, out of a certain motivation, out of one's respect for the moral law. But is even that the sort of thing that can sensibly be called a duty or an obligation? One might object to this, first, out of a certain general picture of motivation and action. That is, one might suppose that it makes sense to talk of duties where there is room to talk of choice, and that there is room to talk of choice with regard to our actions, but that our motivations are just states that we have, not states that we choose: motivations are the states on which our choices of actions are based. Thus, one might argue that it makes no sense to say that we have a duty to have a certain motivation, because we cannot choose to have one motivation or another. Kant himself does not limit choice in this way, however: at least by the time he published The Metaphysics of Morals in 1797, he had, in his 1793 work Religion within the Boundaries of Mere Reason, straightened out earlier confusions in his view about the freedom of choice. In the latter work, he says that our most basic—“radical”—free choice is precisely the choice of whether to make the moral law or self-love our most fundamental maxim and motivation, that is, whether to adopt the principle always to make the principle of self-love subordinate to the principle of morality (and therefore to act out of self-love only when that is consistent with morality) or, vice versa, to make the principle of morality subordinate to the principle of self-love and to act morally only when that is consistent with acting out of self-love.Footnote ¹³ Kant clearly incorporated into The Metaphysics of Morals the fundamental distinction that he had to make in order to arrive at this view that we freely choose whether or not to make the moral law our fundamental motivation. This is the distinction between Wille or will simpliciter as an essentially cognitive faculty that offers us the moral law, and Willkür as the faculty of choice or “elective will,” the practical faculty that makes the choice between the moral law and self-love.Footnote ¹⁴ So when he makes the distinction between the general obligation to be virtuous and the specific duties of virtue in this work, Kant does think that we can choose whether or not to make the moral law our fundamental motivation; thus, he thinks that we can choose to adopt specific ends (such as being beneficent to others) out of respect for morality as such or, for example, out of our sense of prudence. Thus, Kant himself would not think it an objection to the idea of a general duty to be virtuous that this duty would make sense only if we had a freedom to choose our motivation that we do not have; in his view, we do have the freedom to choose our fundamental motivation.

However, it might still be objected that Kant's own conception of “moral worth” conflicts with his idea of a general obligation to be virtuous. In the famous opening section of the Groundwork for the Metaphysics of Morals, Kant appeals to what are supposed to be common-sense notions of the good will and duty, “the good will though under certain subjective limitations and hindrances,”Footnote ¹⁵ in order to arrive at an initial formulation of the categorical imperative that will then be confirmed and amplified from more philosophical premises in the second section and, finally, proved to be binding for us human beings in the third. His strategy is to find indubitable examples of moral worth (not empirical examples, to be sure, but thought experiments), that is, cases in which it is beyond doubt that there is good will or that duty has been done, in order to establish the constraints on any possible candidate for the moral law or categorical imperative.Footnote ¹⁶ (The categorical imperative is the moral law as it presents itself to creatures, like we human beings, who do not automatically follow it because they have inclinations that compete with the demands of morality. Should there be any purely rational beings who do not have such contra-moral inclinations, the moral law would not present itself to them in the form of an imperative.) Kant's argument from the concept of duty to the formulation of the moral law then proceeds in three steps.

His analysis begins from the premise that although agents who are motivated solely by inclination can still, at least under fortunate circumstances, act in conformity with duty, only in cases in which agents have no inclination in favor of the actions that duty requires (or in which those very agents who were previously moved by inclination have now lost their inclination but still do what duty requires) can we be sure that they are acting from duty alone and thus that their actions have what Kant calls “moral worth.”Footnote ¹⁷ The point of this is to show that if we can be sure that an agent demonstrates moral worth only when he acts without inclination, then it must be possible to act on the moral law without inclination, and therefore the moral law cannot be grounded on inclination.Footnote ¹⁸ Kant then draws the consequences of this premise in the two further steps of his analysis of duty, inferring, first, that if moral worth (and therefore the moral law) is independent of inclination, then the moral law also cannot concern the objects of inclination and, indeed, cannot concern the objects of our actions at all,Footnote ¹⁹ and, second, that moral worth can lie only in acting from respect for the moral law itself,Footnote ²⁰ which, since it cannot concern inclinations or their objects, can only prescribe that our maxims have the form of law as such, or be universalizable.Footnote ²¹ The relevance of all this to our present concern is simply that, in this analysis, Kant presents “moral worth” or acting from duty as something different from and more estimable than merely acting in conformity with duty: acting in conformity with duty regardless of motive “deserves praise and encouragement but not esteem,”Footnote ²² while only acting from duty merits esteem. And this suggests that even though we can use the absence of any inclination from these indubitable cases of moral worth to determine the content of the moral law (because obviously agents who clearly have moral worth must be acting in accordance with the moral law), there is something supererogatory about actually acting from respect for the moral law. The agents who merely act in conformity with the law and who merit praise and encouragement for so doing are, after all, acting in accordance with the law. Thus, it seems as if they are doing their duty, and as if acting from duty (that is, more precisely, from respect for the moral law) is doing more than one's duty and deserves “esteem” precisely because it is so. Thus, it would seem, acting in conformity with the moral law is what duty requires, and acting from duty deserves esteem just because it is more than what duty requires; and so it seems that if to have a virtuous disposition is to be motivated to act from respect for the moral law itself, then to have such a disposition must be more than what duty requires; therefore, it seems that there cannot be a duty to have a virtuous disposition.

Kant suggests a similar objection to the idea of a general obligation to be virtuous in the introduction to the “Doctrine of Virtue” itself, by means of a contrast between “merit” and “demerit.” Kant says that the fulfillment of imperfect duties “is merit (meritum) … but failure to fulfill them is not in itself culpability (demeritum) … but rather mere deficiency in moral worth … unless the subject should make it his principle not to comply with such duties”;Footnote ²³ and he includes the general obligation to be virtuous as well as the specific duties of virtue among the imperfect duties. In fact, he immediately illustrates this claim with the general obligation to be virtuous rather than with a specific duty of virtue: “It is only the strength of one's resolution, in the first case, that is properly called virtue (virtus); one's weakness, in the second case, is not so much vice (vitium) as rather mere want of virtue, lack of moral strength (defectus moralis).”Footnote ²⁴ Since there is no actual demerit in not being virtuous (unless one has made it a principle not to do one's duty) but there is special merit in being virtuous (that is, in fulfilling one's particular duties with a virtuous disposition), it seems again as if fulfilling one's duties is one thing and doing so virtuously is another, additional thing, above and beyond duty, and therefore not part of duty.

These reasons are internal to Kant's own account of moral worth and merit, and seem to make his idea that we have a general obligation to be virtuous in addition to our specific duties of virtue problematic. Can his idea be saved? In fact, the ground for an affirmative answer to this question has already been suggested in the last quotation from Kant: by virtue or a virtuous disposition in general, Kant does not just mean being motivated by respect for the moral law alone, but rather the strength of one's resolution to be so motivated. This is something that can be developed and, indeed, that one can choose to develop, as well as something the development of which is, for real-world human beings, a necessary condition for the fulfillment of their particular duties in the absence of external, coercive constraints. Thus, it is a necessary condition for the fulfillment of their particular duties of virtue, and, for that reason, it is something that can be demanded of them as itself a duty. Kant makes it clear in two key passages that virtue is not just being motivated by the moral law but rather the strength to be so motivated in the presence of natural inclinations that may be an obstacle to being so motivated in every human being. By means of this strength we may constrain ourselves to the fulfillment of our particular duties. In the first passage, Kant writes:

This passage makes it clear that virtue in general lies not merely in being motivated by the moral law in addition to acting in conformity with its mandates, but in having the strength to constrain oneself to have this motivation, a strength that is a necessary condition in us human beings for acting morally because of the obstacles that our natural inclinations can and so often do present. Kant reiterates this conception of virtue in a second passage, under the rubric “On Virtue in General”:

Again, Kant suggests that virtue is not just being motivated by the moral law, thereby possessing the ground of moral worth and merit that goes beyond duty; rather, virtue is the strength of will that is necessary in order to fulfill duty in the face of the obstacles to so doing that are natural to human beings. Thus, virtue is, in this way, a necessary condition for the fulfillment of duty, not something above and beyond duty, and, for this reason, it is something that can be demanded in the name of or as part of duty. For this reason, it seems to make sense after all to say that we have an obligation or duty to be virtuous as well as specific duties of virtue.

No sooner has he led us to this conclusion, however, than Kant seems to undermine it by adding that “Virtue itself, or possession of it, is not a duty (for then one would have to be put under obligation to duties); rather, it commands and accompanies its command with a moral constraint (a constraint possible in accordance with laws of duty).”Footnote ²⁷ Here Kant seems to hold that it cannot be a duty to have what is a necessary condition for the fulfillment of particular duties, and perhaps that such a supposition would generate an infinite regress after all. So now it seems as if we are back where we began, and must ask again whether Kant's idea of a general obligation to be virtuous can be saved from what is apparently his own direct criticism of any such idea.

I will argue that Kant himself finally shows us how to defuse this criticism by arguing that even if it may not make sense to say that we have a duty to have a disposition that is a necessary condition for any specific obligation, it makes perfectly good sense to say that we have a duty to cultivate and strengthen a natural disposition that is a necessary condition to the performance of our particular obligations. On Kant's view, this is how the general obligation to be virtuous should be understood: as an obligation to cultivate and strengthen our natural tendency to be moved by the moral law.Footnote ²⁸ This calls for acts that are different from the specific acts required to fulfill particular duties of virtue, and thus the obligation to be virtuous is not, in the end, merely formal and is not reducible without remainder to the duties of virtue. Before I make that argument, however, I want briefly to locate Kant's conception of obligation historically, and by so doing to suggest that there is historical precedent for his conception of a duty to cultivate what is a necessary condition for the fulfillment of more particular duties.Footnote ²⁹

III. Two Conceptions of Duty

On one conception of duty with which Kant was obviously familiar, a duty is simply an action required by a law. Thus, in his German Ethics (1720), Christian Wolff wrote: “By duty [Pflicht] we understand an action that is in accordance with a law…. No law is without obligatoriness [Verbindlichkeit].”Footnote ³⁰ In a textbook three decades later (1750), Johann Stephan Pütter and Gottfried Achenwall defined moral obligation as whatever actions the will should determine to do in accordance with the concepts of good and evil: “That obligates [obligat] in the most general sense which connects with a spontaneous act a resulting good or evil….The connection of a motive with a free action is called a moral obligation [Obligatio].”Footnote ³¹ These definitions suggest that whatever is required by a law (in Wolff's apparently deontological formulation) or by a moral conception of good and evil (in Pütter and Achenwall's explicitly teleological formulation) is a duty, and that it would make no sense to look beyond the particular requirements of the moral law or moral ends for a general duty to conform to moral law or seek moral ends. (I say Wolff's “apparently deontological” formulation, because his larger argument is that our duties are to perfect the mental, physical, and external conditions of ourselves and others—as well as to worship God—and thus his moral philosophy, as a whole, is as teleological as they come.)Footnote ³² To look for such a general duty would indeed risk generating an infinite regress of duties which does not need to get started in the first place.

A more complex conception of duty, however, and one to which Kant seems to be responding, is to be found in the widely influential, somewhat earlier work of Samuel Pufendorf, whose Whole Duty of Man and Citizen (1673) remained influential throughout Europe for a century.Footnote ³³ Pufendorf makes two moves that are helpful to understanding Kant's conception of duty. First, he not only defines a duty as “human action in conformity with the commands of law on the ground of obligation,”Footnote ³⁴ but goes on to define a law as “a decree by which a superior obliges one who is subject to him to conform his actions to the superior's prescript.”Footnote ³⁵ Thus, he implies that a duty or an obligation is not merely prescribed by a law, or an act required by a law, but is also “introduced into a man's mind by a superior.” To this last inference he immediately adds that a superior is “one who has not only the strength to inflict some injury on the recalcitrant, but also just cause to require us to curtail the liberty of our will at his discretion.”Footnote ³⁶ There are thus two elements in Pufendorf's conception of duty: a duty is not merely an action required of agents in accordance with a law for the existence of which there is a good reason; it is also an action required of agents by an authority that has the power to enforce that law. On Pufendorf's account, the primary source of duties is God, who has created us to live sociably with one another and has legislated the laws of nature with the good reason that they are what it is necessary for us to follow in order to live sociably with one another. The secondary source of duties is the civil authority that transforms natural law into civil law and adds whatever further obligations are necessary in order to maintain civil authority itself.

A second aspect of Pufendorf's conception of duty that will turn out to be important to Kant is Pufendorf's conception of the nature of duty to oneself. Pufendorf employs the tripartite division of duties into duties to God, duties to self, and duties to others—a division that would remain canonical in German moral philosophy throughout the eighteenth century.Footnote ³⁷ “Perhaps the duties imposed on man by natural law are most conveniently divided in accordance with the objects on which those duties are to be exercised,” Pufendorf writes. “On these lines they form three principal divisions. The first teaches, on the basis of right reason alone, how one should behave towards God; the second towards oneself; the third towards other men.”Footnote ³⁸ Our duty to other men is obviously to live in accordance with the rules that make it possible to live with them sociably (to which we have a natural disposition) by keeping our unsociable impulses (which we also naturally have) in check. Our duty to God is the direct duty to worship him in gratitude for our creation, but we also have an indirect duty to fear him in order to better maintain our society with other men:

Pufendorf also interprets our duty to ourselves as an indirect duty to cultivate those powers of our mind and body by means of which we may successfully fulfill our duties to others and to God (in the latter case, since our duty to God is already itself, at least in part, an indirect duty, our duty to ourselves is, at least in part, a doubly indirect duty):

The duty that we have with regard to our mind or soul is, in turn, divided into “the right Formation of the Mind and the Heart,” that is, the duty to form true opinions about God and society, and the duty to regulate “the Dispositions of our Minds; in reducing and conforming them to the Dictates of right Reason … and in one word, in getting ourselves possest [sic] of all those Qualities which are necessary for us to lead an honest and a sociable Life”Footnote ⁴¹ and “To gain the Mastery over our Passions.”Footnote ⁴² The duty that we have with regard to our body is to “strengthen and preserve the powers of the body with appropriate food and exercise”; “We should not weaken them by intemperance in food or drink or by unseasonable and unnecessary toil or by any other means,”Footnote ⁴³ so that our bodies will be fit instruments for the accomplishment of our moral objectives with regard to others in society.

The contents of Kant's own enumeration of our duties to ourselves in the “Doctrine of Virtue” is obviously deeply indebted to Pufendorf, but this is not the point that I want to emphasize. Rather, I want to draw attention to two more general points. First, Kant's account of duty clearly includes Pufendorf's requirement that there is a duty only when there is not just a good reason for the performance or omission that is enjoined, but also a proper authority for the enforcement of that performance or omission. However, Kant changes Pufendorf's conception of the primary superior for the enforcement of duty from God to our own reason and the strength of our own will. Kant begins his “Discussion of the Concept of a Doctrine of Virtue” with the Pufendorfian premise that “The very concept of duty is already the concept of a necessitation (constraint) of free choice through the law,” but he fundamentally alters the theological foundation of Pufendorf's theory of natural law when he continues that “Since the human being is still a free (moral) being, when the concept of duty concerns the internal determination of his will (the incentive) the constraint that the concept of duty contains can only be self-constraint (through the representation of the law alone); for only so can that necessitation (even if it is external) be united with the freedom of his choice.”Footnote ⁴⁴ For Kant, the idea that our duty to live sociably with other men is imposed upon us by the will of God, and enforced by the strength of God, is the very picture of heteronomy. For Kant, for whom autonomy is the ultimate moral value and the source of all moral obligations, the requirements that we preserve our own freedom and cultivate the means for its successful exercise, and that we live on terms of equal freedom with others, can only be the contents of the moral law suggested by our own reason. Moreover, the freedom that is the ultimate aim of all morality can be imposed upon us only by the strength of our own will if it is not to contradict itself—if it is not to become an imposed freedom that is no longer freedom. For all the similarities in both form and content between Kant's moral philosophy and the Pufendorfian paradigm, this is a fundamental revolution.

The sea-change between Kant's ethics of self-imposed and self-enforced obligation and Pufendorf's theory can also be seen in Kant's revision of Pufendorf's conception of the relation between divine and civil authority. For Pufendorf, our primary duty is the duty to live sociably with others, a duty in accordance with natural law that is primarily enforced by our fear of God, and secondarily enforced by civil authorities in accordance with civil law (but under the authority of God). For Kant, civil or juridical law is typically enforced through the external, aversive constraints of a public penal code, which works upon and is satisfied with our merely prudential motivation for obedience. Yet it is also possible and morally estimable for us to obey the dictates of civil law out of the purely moral motivation of respect for the moral law itself, which is, after all, the source of the requirement that we preserve maximally equal spheres of external freedom in our interactions with others, as well as of the non–coercively enforceable duties to ourselves and others that comprise ethics: “All that ethics teaches is that if the incentive which juridical lawgiving connects with … duty, namely external constraint, were absent, the idea of duty by itself would be sufficient as an incentive,” and indeed there is a “proof of virtue” in the observation of juridical legislation only if we “do it even where no coercion may be applied.”Footnote ⁴⁵ Kant does not see the observation of legal requirements out of fear of the consequences of disobeying them as undermining our autonomy in the way that he sees fulfilling moral requirements generally only out of fear of divine punishment or hope for divine rewards as the epitomy of heteronomy, because he views the institution and maintenance of the juridical authority of the state as itself a product of our freedom. However, he does view the motivation to fulfill even juridical legislation out of our respect for the moral law alone as the fullest expression of our freedom.

My first general point, then, has been that Kant accepts from Pufendorf the requirement that duty must always involve the possibility of constraint by an appropriate authority, but he radically transforms Pufendorf's conception of duty by making the ultimate authority for duty our own reason, enforced by the strength of our own will. My second general point is that Kant accepts from Pufendorf the idea of indirect duty that structures Pufendorf's conception of duty to oneself and even, to an extent, his conception of duty to God. To confine ourselves to the former case, Pufendorf, as we saw, conceives of duty to self as the duty to develop those capacities and powers in oneself that will enable one to satisfy one's duties to others in society. On his account, then, we can have a duty to ensure the existence of the necessary conditions to satisfy other duties, and there is nothing redundant or infinitely regressive in such an idea. We have a duty to develop those capacities in ourselves that are necessary to enable us to fulfill our duties to others, and that is not redundant, because there is a substantive difference in content between the duties to self and the duties to others. Nor is it infinitely regressive, because there is no further duty to fulfill the duties to self that are the necessary conditions for being able to fulfill duties to others, and so on. What I now want to suggest is that Kant uses this structure in his own account of the relation between the general obligation to be virtuous and the specific duties of virtue: the development of the strength of will that is required by the former is the necessary condition of fulfilling the latter (and even of fulfilling the duties of right, if there are cases in which external constraint is either not available or not necessary).

By drawing this parallel, I do not want to suggest that Kant reduces all duties to self to instruments for fulfilling duties to others, either other humans or God, as Pufendorf does. For Kant, we have a duty to treat humanity in our own selves always as an end and never merely as a means—a duty that is on exactly the same plane as our duty to treat humanity in others that way—and we would have that duty to ourselves even if, per impossibile, there were no other human beings whom we could affect by our actions.Footnote ⁴⁶ I only want to suggest that Kant exploits the idea of an indirect duty to realize the necessary condition of satisfying other duties that Pufendorf suggests in his account of duties to self. But before we can see how the details of this strategy work, we must face the threat left hanging at the end of the previous section, the suggestion by Kant himself that the possession of “virtue itself” cannot actually be a duty.Footnote ⁴⁷

IV. The Obligation to Be Virtuous as an Obligation to Strengthen One's Resolve

As I noted earlier, in Religion within the Boundaries of Mere Reason (written before The Metaphysics of Morals), Kant had abandoned the view of the Groundwork that the moral law is the causal law of the noumenal selfFootnote ⁴⁸ and committed himself to the alternative view that we have a free, noumenal choice whether to subordinate the principle of self-love to the moral law or to subordinate the moral law to the principle of self-love. Freely choosing to make the moral law superior to the principle of self-love is a necessary condition for fulfilling any particular duties when doing so is not prescribed by self-love. The whole point of instituting a penal code is to make sure that self-love usually will recommend fulfilling duties of right even if commitment to the moral law does not. Thus, in this case, a virtuous disposition (a commitment to be motivated by the moral law above all else) cannot be demanded as a necessary condition for the fulfillment of duties of right, and, from the point of view of right, a virtuous disposition can indeed be regarded as supererogatory, something whose presence is meritorious but whose absence brings no demerit. But it is not appropriate or even possible to enforce fulfillment of duties of virtue by coercive means. It is not possible because coercion can only bring about the performance of particular actions, not the adoption of an end; and it is neither appropriate nor possible because the duty to adopt an end is the duty to choose an end freely. In the case of duties of virtue, then, firm commitment to the superiority of the moral law over the principle of self-love is a necessary condition for the fulfillment of those duties when self-love does not itself prescribe it, which will sometimes or often be the case. Thus, it seems that a virtuous disposition (commitment to the moral law) can be demanded as the necessary condition of fulfillment of the particular duties of virtue, and is not supererogatory at all, but can be considered a duty, the obligation to be virtuous. This could be understood on the model of the two-tiered structure of direct and indirect duty that Kant could have adopted from Pufendorf.

Thus, Kant could have countered his own objection by arguing that “Virtue itself, or the possession of it, is not a duty,” but this is not what he does argue. Rather, what he does say (four pages after making this remark) is that “Virtue is always in progress and yet always starts from the beginning.—It is always in progress because, considered objectively, it is an ideal and unattainable, while yet constant approximation to it is a duty. That it always starts from the beginning has a subjective basis in human nature.”Footnote ⁴⁹ In other words, while it may make no sense to say that to be virtuous is a duty, it makes sense to say that to strive to make progress toward becoming virtuous is a duty. How does this help save Kant from an objection to the idea that a necessary condition of a duty is also a duty?

One thing this remark could be taken to mean is that to be perfectly virtuous cannot be considered a duty, because it is simply not possible for human beings to have a holy will (that is, to be moved only by the moral law and not even be moved by self-love) and because agents have no obligation to attempt to do that which is impossible for them to do. But humans can always have purer motivation than they do, and thus can always make progress toward having perfectly virtuous dispositions, so it is possible for “constant approximation” to virtue to be a duty. However, although Kant makes that point later in the body of the “Doctrine of Virtue,”Footnote ⁵⁰ and we will return to it, that does not seem to be his point here. Rather, his point seems to be something like this. On the one hand, the disposition to make the moral law our fundamental maxim is the necessary condition of the possibility of constraining ourselves to fulfill our particular duties of virtue, so it may not seem to make sense to think of that disposition as itself the object of a duty, precisely because it is the condition of the possibility of being put under an obligation. Unless we had such a disposition, it would not be possible for us to recognize any claims of duty at all, so it does not make sense to say that we have a duty to have that disposition itself—that would start us off on an infinite regress. On the other hand, although we do have a natural disposition to do our duty that makes it possible for us to recognize the claim of (and fulfill) particular duties, that is not our only natural disposition. We also have the disposition toward self-love, and thus our natural disposition to be moral always needs to be strengthened, and our disposition toward self-love weakened and controlled. At the noumenal level, we may simply be able to choose to subordinate the principle of self-love to the principle of morality rather than the contrary, but what doing this requires of us at the phenomenal level is that we constantly strive to strengthen our natural disposition in favor of morality at the expense of our natural disposition in favor of self-love. Thus, even though it would make no sense to say that we would have a duty to have a disposition in favor of morality (a virtuous disposition) if we did not already have one—because without such a disposition, duties can make no claim on us—it makes perfectly good sense to say that we have a duty to strengthen our natural disposition to morality, because that very disposition itself allows us not only to recognize the claim of duty in general but also to recognize that we may always need to strengthen our moral resolve in order to be able to do our duty in particular cases where self-love tempts us to do otherwise. Thus, the general obligation to be virtuous can be perfectly well understood as a duty to strengthen our natural disposition to be moral.Footnote ⁵¹

Kant suggests precisely such a pattern of reasoning in his discussion of “The Mind's Receptivity to Concepts of Duty as Such.” In this section of the introduction to the “Doctrine of Virtue,” preceding by several pages the remark that “constant approximation” to virtue is a duty, Kant writes:

Love of one's neighbor and self-esteem are being treated here as particular natural dispositions that make it possible for one to fulfill particular duties of virtue toward others and toward oneself; let us leave them aside and focus on the dispositions toward moral feeling and conscience, which seem to be general conditions for recognizing and fulfilling any claims of duty in the absence of external constraint. Kant's claim is that it makes no sense to say that we have a duty to have these dispositions, because without them we could not be receptive to the concept of duty at all; that is, we could not recognize and be motivated to fulfill the claims of duty. His idea must be that although at the noumenal level we simply choose to prioritize morality over self-love, at the phenomenal level our recognition of the superiority of morality manifests itself in our moral feeling and conscience outweighing our equally natural disposition to self-love; moral feeling and conscience are the natural means through which our free choice to be made moral is made effective on our actions, which of course take place in the phenomenal world.Footnote ⁵³ But Kant does not think that we simply make the noumenal choice to be moral and then, as if by magic, our moral feeling and conscience automatically outweigh any contrary natural dispositions. On the contrary, he thinks that our noumenal choice to be moral leads us to take steps to strengthen our moral feeling and conscience so that they will be effective in particular cases of action. In other words, the first manifestation of our recognition of duty in general will be the recognition that we have the duty to strengthen our moral feeling and conscience. Here is how Kant puts the point in the case of moral feeling:

Kant does not worry that any human being is simply lacking in moral feeling. “No human being is entirely without moral feeling, for were he completely lacking in receptivity to it he would be morally dead,” that is, a mere animal. (Perhaps this is a merely analytical point: if you do not have moral feeling, it would not be possible for you to be moved by the thought of duty, and you would not count as a moral agent—we would have to deal with you under some other conception, as a nonperson who should nevertheless be treated humanely.) Kant believes the same about conscience, which he defines as “practical reason holding the human being's duty before him for his acquittal or condemnation in every case that comes under a law,” and which I think we should understand as the natural disposition that presents particular situations to our empirical selves as cases calling for moral decision.Footnote ⁵⁵ Kant writes: “So too, conscience is not something that can be acquired, and we have no duty to provide ourselves with one; rather, every human being, as a moral being, has a conscience within him originally,” but “The duty here is only to cultivate one's conscience, to sharpen one's attentiveness to the voice of the inner judge and to use every means to obtain a hearing for it (hence the duty is only indirect).”Footnote ⁵⁶ Again, while it may make no sense to say that we have a duty to have a predisposition without which we could not have duties at all, it makes perfectly good sense to say that we have a duty to strengthen a natural predisposition through which we can make our choice to be moral in particular cases effective. Kant explicitly says that this duty is “indirect” in the case of conscience: what he most obviously means by this is that conscience is what allows us to recognize that we are in particular circumstances in which a duty needs to be fulfilled, while moral feeling is that which provides the impetus to fulfill it, or transmits the impetus of the noumenal choice to fulfill it into the phenomenal world. But both the duty to strengthen moral feeling and the duty to strengthen conscience can also be regarded as indirect duties in the sense previously suggested; that is, they can be regarded as duties because they are the means to the fulfillment of more particular duties, and indeed the only means to the fulfillment of particular duties of virtue where self-love would do otherwise.

So the indirect duties to strengthen moral feeling and conscience suggest a model for the indirect general obligation to be virtuous. Indeed, I would suggest that the duties to strengthen moral feeling and conscience comprise the general obligation to be virtuous, or comprise a large part of it. That is, what we have to do in order to make progress toward being virtuous, which is how we are now understanding the general obligation to be virtuous, is constantly take steps to strengthen our natural disposition to have moral feeling and conscience and by that means to “fight” against the “brood of dispositions opposing the law.”Footnote ⁵⁷ It is by strengthening these dispositions that we come to “govern ourselves” and gain control over our feelings and inclinations, and even more so over “affects” or “precipitate and rash” feelings and “passions” or “sensible desire[s]” that have become “lasting inclination[s].”Footnote ⁵⁸ Kant says that “Virtue Requires, in the First Place, Governing Oneself”:

This is the conclusion of the section that immediately precedes the one in which Kant states that constant approximation to virtue is a duty. So the positive command to govern ourselves—the general obligation to be virtuous—can only be understood as the command to constantly strengthen our self-government, and this is, in turn, accomplished in large part by strengthening our natural dispositions to moral feeling and conscience, which are the means by which we counter the attractions of affects and passions.

I say “in large part” because Kant amplifies the steps that we must take in order to fulfill our general obligation to be virtuous a little bit in the main body of the “Doctrine of Virtue.”Footnote ⁶⁰ In a characteristic display of his love for architectonic organization of his arguments, Kant divides the body of the “Doctrine of Virtue” into two parts, the first concerning duties of virtue to the self and the second concerning duties of virtue to others;Footnote ⁶¹ he then divides Part I (but not Part II) into two books, the first concerning perfect duties to oneself and the second concerning imperfect duties to oneself; and he finally divides each of these into two chapters, the first concerning perfect or imperfect duties to oneself respectively as “an animal being,” and the second concerning those duties to oneself “merely as a moral being.”Footnote ⁶² The perfect duties to oneself as an animal being are the prohibitions against suicide, self-abuse, and self-stupefaction,Footnote ⁶³ that is, duties to preserve one's body and its efficacy as an instrument of one's free choices. The imperfect duties to oneself as an animal being are the duties to perfect (or adopt the end of perfecting) one's natural powers of “spirit” or reason, “soul” or other mental powers (such as memory and imagination) that can be put to work by reason, and “body” or physical strength and skill, all of which are powers by means of which one's free choices can be made efficacious.Footnote ⁶⁴ One's perfect duties to oneself as a moral being are twofold. First, Kant enumerates prohibitions against lying, avarice, and servility; the prohibition of these vices might seem to be duties toward others, but Kant treats them as duties toward oneself because violating them would undermine one's power to communicate with others, one's power to use nonhuman things for reasonable purposes, and one's dignity in the eyes of others.Footnote ⁶⁵ Second, and what is relevant to our present argument, he lists the “Human Being's Duty to Himself as His Own Innate Judge”Footnote ⁶⁶ as “the First Command of All Duties to Oneself,” namely, the duty to “‘know (scrutinize, fathom) yourself,’ not in terms of your natural perfection … but rather in terms of your moral perfection in relation to your duty.”Footnote ⁶⁷ Kant then finally adds an agent's imperfect duty to himself as a moral being, the duty “to increase his moral perfection, that is, for a moral purpose only.”Footnote ⁶⁸

What Kant actually says about our duties to ourselves as moral beings within this complicated framework is quite simple. First, he says: “Every human being has a conscience and finds himself observed, threatened, and in general kept in awe (respect coupled with fear) by an internal judge; and this authority watching over the law in him is not something that he himself (voluntarily) makes, but something incorporated in his being.”Footnote ⁶⁹ Kant does not need to add that our duty with regard to this faculty, which is already present within us and by its nature commands our attention, can only be to strengthen it, because he has already said this in the introduction to the “Doctrine of Virtue.” This first perfect duty to oneself as a moral being thus does not add anything to the analysis of what is necessary in order to fulfill one's general obligation to be virtuous that has already been provided in the introduction, and its inclusion here among particular duties of virtue may be momentarily confusing but is harmless.

Second, under the command to know oneself, not in terms of one's natural abilities (and thus what sorts of projects one could prudently undertake) but with respect to one's “moral perfection in relation to our duty,” Kant adds that one needs to know one's heart. That is, one needs to know one's good and evil tendencies, which is necessary in order “first to remove the obstacle within (an evil will actually present in him) and then to develop the original predisposition to a good will within him, which can never be lost.”Footnote ⁷⁰ Kant adds that moral cognition of oneself will “dispel fanatical contempt for oneself as a human being (for the whole human race)”; that is, it will dispel the belief that all human beings are necessarily evil, for this moral self-knowledge will also reveal the natural predisposition to a good will. Kant then infers that “Impartiality in appraising oneself in comparison with the law, and sincerity in acknowledging to oneself one's inner moral worth or lack of worth are duties to oneself that follow directly from this first command to cognize oneself.”Footnote ⁷¹ Here Kant does add something to the requirements of strengthening moral sense and conscience as the constituents of the general obligation to be virtuous that have already been provided in the introduction, but what he adds is something that we might suppose is implicit in them, since knowing what we need to do in order to strengthen our moral feeling and conscience and to weaken other feelings and voices within ourselves surely presupposes knowing what states those feelings are currently in and what our current degree of moral worth or lack of moral worth is. (Nonetheless, as Kant famously asserts in the Groundwork, we can never be certain of our real degree of moral worth and are constantly tempted to overrate itFootnote ⁷²—all the more reason why we have a duty constantly to strengthen our moral self-knowledge.) Kant does not explicitly state that we must have some natural predisposition to appraise ourselves in comparison with the law and to acknowledge our inner moral worth or lack thereof, and that our actual duty with regard to moral self-appraisal can only be to cultivate and strengthen it, but perhaps that is self-evident.

Finally, under the rubric of our imperfect duty to ourselves as moral beings, Kant does not actually add any new requirement, but only makes explicit that a human will is never a holy will (a will that is simply not tempted by inclination), so that our duty to strive after moral perfection “remains only a progress from one perfection to another,”Footnote ⁷³ and thus: “This duty to oneself is a narrow and perfect one in terms of its quality; but it is wide and imperfect in terms of its degree, because of the frailty (fragilitas) of human nature.”Footnote ⁷⁴ That is, we do not have a choice whether or not to strive to be virtuous in this situation or that, as we might have a choice whether to cultivate this potential talent or that one; rather, we are always under a general obligation to improve our virtue. But while there is no limit to how much we can strengthen our virtue, there is also nothing that will ever count as completing the task of strengthening our virtue.Footnote ⁷⁵

What it takes to cultivate and strengthen our general resolve to fulfill our particular virtues, even when we are not externally constrained to do so, has thus already been largely spelled out in the introduction to the “Doctrine of Virtue,” and Kant has not really confused the general obligation to be virtuous with the particular duties of virtue that he discusses in the body of the text. What he has done in the introduction is to show that it does make sense to impose upon ourselves a general obligation to cultivate and strengthen our natural disposition to be moral by cultivating and strengthening our natural tendencies to moral feeling and conscience, which both require cultivating moral self-knowledge. He has also shown this general obligation to be one freely imposed on us by our own reason and thus to be an autonomous form of self-constraint rather than a heteronomous form of external constraint. By so doing, Kant has made sense of a general obligation to be virtuous as well as of particular duties of virtue, and any confusions in his account are quite superficial.