Introduction

It is widely believed that Kant's practical philosophy requires a ‘constructivist’ account of moral value or some other alternative to realism.Footnote ¹ One powerful reason to think this is that realism seems incompatible with his view of autonomy. If Kant were a realist about moral values, he would hold that such values exist independently of what anyone thinks or decides is the case. This would suggest that, in order to act virtuously, an agent must respond appropriately to something that does not depend on his will. Yet this seems incompatible with the very idea of autonomy. Kantian autonomy is self-legislation, and on the realist picture virtuous conduct is a response to something other than the self – some independently existing value. Is this not a picture of what Kant calls heteronomy?Footnote ²

Not necessarily. Kant seems to think that the very nature of persons makes them unconditionally valuable. He also seems to think that this fact explains (at least partly) our obligation to treat them as ends in themselves. But neither of these beliefs is incompatible with his view of autonomy. It is true that an agent who responds appropriately to the value of persons is guided by something that does not depend on his will, but this does not undermine his autonomy. Or so I shall argue. I begin with a sketch of Kant's conception of autonomy. Then I articulate – without defending – a realist interpretation of his understanding of the value of persons. After that I proceed to the main argument, which is designed to show that this realism is perfectly compatible with autonomy. Since this is the main objective of the essay, I do not enter directly into debates about whether it is correct to read Kant as a moral realist. But I do hope my argument will offer some indirect support for that view. I also hope the main argument will lend some plausibility to the stronger thesis that Kantian autonomy requires some form of realism to the extent that the capacity for self-legislation actually requires the capacity to respond appropriately to something that does not depend on the will (or any volitional procedure, however pure it may be). But this stronger thesis should not be confused with the more modest thesis I will defend directly – namely, that it is a mistake to think Kant's conception of autonomy commits him to constructivism or any other form of anti-realism about moral value.

According to Kant's official definition, autonomy is the property of the will by means of which the will can give itself a law (G 4: 440, 447).Footnote ³ So to say of a person that he is autonomous is to say he is capable of self-legislation. We can understand this claim in at least two different ways. First, we can interpret it as a normative claim about the powers we are capable of exercising in the practical sphere. Taken this way, it means that each person, by virtue of his will, possesses a power in the practical sphere analogous to the powers possessed by a legislative body in the political sphere. Each person has the power to issue practical laws. And this means that each person has a moral authority that cannot be overridden or usurped. Second, we can interpret it as the psychological claim that we have the ability to act on reasons that are not based in inclination. Taken this way, autonomy pertains to motivation – namely, the capacity to be motivated by pure practical reason. It is closely related to the ‘autocracy’ or self-command exhibited by a person who regulates his inclinations according to the demands of morality (MM 6: 383).Footnote ⁴ Andrews Reath argues persuasively that the first sense of autonomy is more central to Kant's ethics than the second (Reath Reference Reath2006: 125–31).Footnote ⁵ It leads to the claim that autonomy is the basis of morality, and it is this thesis that distinguishes Kant's ethics from all other theories (G 4: 433, 441–3; KPV 5: 39–41). So I shall follow Reath in treating autonomy as a normative rather than psychological matter.Footnote ⁶ Hence, when I use words such as ‘capacity’, ‘ability’, and ‘exercise’, I am speaking of a normative capacity and the exercise thereof. It is the will's authority that is at issue, not its psychological force or power.Footnote ⁷

What is it for the will to have this property of autonomy? To summarize crudely a famously difficult line of thought, Kant proposes the following. He begins with the claim that the will (der Wille) is the capacity to guide one's conduct by means of what one takes to follow from a general rule (G 4: 412, 427). When a person guides his conduct this way, he does so on the basis of a ‘maxim’, which is a general policy or principle of volition (G 4: 401n., 421n.). Maxims have no binding force because they are rooted in personal inclinations – the ‘habitual desires’ out of which a person develops an interest in pursuing this or that end (MM 6: 212; A 7: 251) – and such inclinations can change on their own or be actively renounced. For this reason, maxims cannot obligate. They have no force of law – neither for the person who holds the maxim nor for anyone else. Therefore, while an individual who has the capacity to act on the basis of maxims has a will, his will does not necessarily have the property of autonomy because autonomy is the capacity for self-legislation and maxims are not laws.

Kant explains what it is for the will to have the property of autonomy by setting forth the conditions a maxim must meet in order to count as a law. The kind of law that matters in this context is a practical law – as opposed to a law of nature or law of the state. Practical laws impose moral obligations by presenting reasons to act that do not depend on personal inclination. They are what Kant calls categorical imperatives. So the question is: what conditions must a maxim meet in order to be a categorical imperative? For the purposes of this essay, we can look to the first two main formulations of the categorical imperative for the two most important conditions.Footnote ⁸ These formulations of the ‘supreme principle of morality’ (G 4: 392) tell us what a principle of volition must look like in order to carry the force of a practical law – that is, in order to be an imperative that is categorical rather than hypothetical. According to the universal formulation (FUL), we are enjoined to reject any personal principle that recommends exempting ourselves from the moral law. That is, we are enjoined to act only on the basis of maxims that we can coherently will as binding on everyone, not just ourselves (G 4: 421). This condition specifies the ‘form’ a maxim must have in order to be a categorical imperative. According to the humanity formulation of the categorical imperative (FHE), we are enjoined to treat all persons, including ourselves, as ends and never as mere means (G 4: 429). This condition identifies the ‘matter’ a maxim must specify in order to be a categorical imperative. Maxims that meet these two conditions qualify as categorical imperatives, and therefore as practical laws. A will that can propose to itself policies that meet these conditions has the property of autonomy. To have such a will is to have the capacity for self-legislation. And to have this capacity is to have a special kind of authority in practical matters.

We shall look more closely at this second condition on autonomy in the next section, but for now I would like to note briefly two important points about it. First, FHE states that we must treat persons as ends, but certainly not all ends (Zwecke) are purposes (Absichten) to be accomplished or goals to be achieved.Footnote ⁹ When we treat a person as an end, we do not act in order to accomplish or achieve that person; we act for the sake of that person.Footnote ¹⁰ And when we treat that person as an end in himself, we act for the sake of that person without doing this merely for the sake of something else. Treating a person this way contrasts with treating him as an end merely for the sake of a further end, which would amount to treating him as a provisional or instrumental end rather than an end in himself. The second point is that, while it is tempting to think that Kant's conception of autonomy rules out reference to ends, we should prefer an account of Kantian autonomy that includes the role of ends. For this is the only one of these two options that is consistent with his general account of practical rationality, according to which ‘rational nature is distinguished from the rest of nature by this, that it sets itself an end’ (G 4: 437). This idea is, in fact, central to his practical philosophy as a whole. We find him repeating versions of this Groundwork claim in the Critique of Practical Reason (KpV 5: 34) and in the Metaphysics of Morals (MM 6: 385, 395). Indeed, in the latter he goes so far as to define ethics as ‘the system of the ends of pure practical reason’ (MM 6: 381). Surely we do not want to attribute to Kant a conception of autonomy that does not fit his own definition of rationality.Footnote ¹¹

The question of realism naturally arises in the context of Kant's second condition on autonomy (FHE). What is the value of persons, and how should we understand the nature of that value? I would like to propose a realist interpretation of Kant's response to these questions. My own philosophical and hermeneutical intuitions tell me that this proposal has obvious appeal, but I will not try to defend it as the correct interpretation of Kant's views. The main objective is to show that at least one form of realism is perfectly compatible with his conception of autonomy. This is the central claim I wish to defend, since I hope to show that Kant's conception of autonomy in no way requires constructivism or any other form of anti-realism. ‘Moral realism’ labels a wide variety of positions. I will work with the minimal, and hopefully uncontroversial, idea that moral realism is the view that (i) some things have moral value and this value does not depend on what anyone thinks, wants or decides. And (ii) that in order to get things right, morally speaking, we must respond appropriately to this value.Footnote ¹²

There is evidence in Kant's work that he accepts both components of this view. Consider his argument for FHE. This argument is proposed in the context discussed above – namely, the second section of the Groundwork of the Metaphysics of Morals, in which Kant attempts to specify the conditions a principle of volition must meet in order to qualify as a law. In addition to being universalizable, he argues, such a principle must also include reference to an objective end (G 4: 427). An end is objective if it ‘holds for every rational being’ (G 4: 427). Objective ends contrast with the subjective ends we set on the basis of personal inclinations, since these hold only for those who share those inclinations. Why must practical rules refer to ends in the first place? Kant holds that all rational action aims at some end (G 4: 437; MM 6: 392). So a practical rule that did not include, implicitly or explicitly, reference to an end could not guide – much less demand – rational action. But not just any end will do. Kant's intention here is to specify the conditions under which a rule can be a practical law, and we cannot be obligated to set an end that depends on an inclination for the simple reason that no one who lacked that inclination would have a reason to adopt that end as his own. It follows that if one of the conditions under which a rule can be a practical law is that the rule specifies an end, then that end cannot be a subjective end – something a person aims at because of a personal inclination. What, then, could qualify as the right sort of end? Kant introduces his answer in the following famous passage:

Here, then, is his answer: the only thing that qualifies as an objective end is ‘the human being and in general every rational being’. After ruling out other candidates for this status, Kant further specifies what he has in mind:Footnote ¹³

The idea is that, since all rational action has an end, no principle of conduct could be a practical law unless it specified an objective end – something for the sake of which we must act regardless of our personal inclinations. And only one thing can qualify as an objective end: a person. By virtue of this status, persons have an ‘inner worth’ or ‘unconditional value’ that Kant calls ‘dignity’ (Würde, dignitas). Dignity is distinguished from ‘price’ (Preis), which is the sort of value possessed by everything else in the universe insofar as it can be treated by us either as a means or an object of delight (G 4: 434–5).

Kant's claim about what gives persons dignity seems to provide evidence for his commitment to the first component of realism. Persons have dignity because ‘their nature already marks them out as an end in itself’. What is it about their nature that does this? Kant's answer is that a person is an end in itself because a person has the capacity for self-legislation or autonomy (G 4: 434). As he puts it, autonomy is ‘the ground of the dignity of human nature and of every rational nature’ (G 4: 436). This line of thought provides evidence for realism because on one very natural reading Kant is claiming that persons have an ‘unconditional, incomparable worth’ (G 4: 436) called dignity and this dignity does not depend on what anyone happens to think or decide. Instead, it depends on the very nature of persons, namely, their capacity for self-legislation. This capacity elevates persons above everything else in nature (Cf. MM 6: 434–5).Footnote ¹⁴ There is, of course, one sense in which the value of persons is not mind-independent in the way we would associate with a realist commitment to electrons, for example. The value of persons is dependent on the existence of persons, so it is true that if there were no persons, there would be nothing in the world that has dignity. But the dignity of persons does not depend on whether anyone thinks or wants it to be the case that persons do or do not have dignity. And this is the sense in which Kant's line of thought seems committed to realism about the value of persons.Footnote ¹⁵ Kant treats the dignity of persons as a fact, and he explains that fact by appealing to their nature.

The second component of the realist view sketched above says that in order to get things right we must respond appropriately to a value that does not depend on what we think or want. If the dignity of persons should be understood as an objective value, then several passages suggest that Kant accepts this second component of realism as well. Consider, first, FHE itself, which says that we must treat humanity, whether our own or someone else's, as an end in itself, and never as a mere means (G 4: 429). This version of the categorical imperative seems to say that we must respond appropriately to a value that does not depend on what we think or decide is the case. Recall the earlier claim that the ‘nature’ of persons ‘already marks them out as an end in itself’, and then consider the way Kant introduces FHE:

FHE is introduced almost immediately after this line of thought. On the interpretation under consideration, FHE says that we must respond appropriately to the dignity of persons. We must treat persons as ends in themselves because persons are, by their very nature, ends in themselves. By virtue of this fact persons have dignity, an unconditional and incomparable worth, and the only appropriate response to this dignity is what Kant calls respect (G 4: 436). We show respect by conducting ourselves virtuously, and the categorical imperative's job is to help us determine what this entails.

In the Tugendlehre portion of the Metaphysics of Morals, Kant argues that this end is the basis for a system of ethical duties. Ethics, he contends, requires that we guide our conduct by particular maxims. The basis of this requirement is ‘the concept of an end that is also a duty’ (MM 6: 389). In other words, the basis of this requirement is the concept of an obligatory end, an end that we must adopt as that for the sake of which we act. And the only kind of end that can play this role is an objective end, since the subjective ends we have on account of our personal inclinations can never be obligatory. Here Kant assumes that his argument for FHE in the Groundwork has identified the only thing that can be an objective end. The system of duties thus rests on FHE. As he puts it: ‘The supreme principle of the doctrine of virtue is: act in accordance with a maxim of ends that it can be a universal law for everyone to have. – In accordance with this principle a human being is an end for himself as well as for others, and it is not enough that he is not authorized to use either himself or others merely as means … ; it is in itself his duty to make man as such his end’ (MM 6: 395).

The central idea here is that ethical virtue requires us to act for the sake of a particular end. It does not require particular actions; instead it requires that we guide our conduct by policies that treat this objective end as that for the sake of which we act. Kant believes that it is possible to construct a system of ethical duties on this foundation. Because of our own status as persons, we are required to pursue our own perfection. This gives rise to duties to self. We fulfil these duties by cultivating those natural capacities that distinguish us from other animals and by cultivating our capacity for morality. And the existence of other persons requires us to adopt their happiness as one of our own ends. This gives rise to duties to others. We fulfil these duties by showing others an active (‘practical’) form of love and respect – by helping them satisfy their true needs and by refraining from interfering with their own (legally permissible) pursuits.

In sum, persons really do have dignity. And in order to get things right, we must respond appropriately to this value. The only way to do this is to adopt our own perfection and the happiness of others as our ends. Hence, these ends are obligatory rather than discretionary or optional. They are duties of ethical virtue.

This form of realism is modest. Kant says that persons have ‘absolute worth’ (G 4: 428). But to say that they are valuable in this sense is not to say that they have odd or mysterious ontological properties. Instead, to say that persons are unconditionally valuable is to say that there is reason to value them above all else. The fact that human beings are capable of autonomy is a reason to treat them as ends in themselves – to treat them as persons rather than as mere things. I think this is what Kant means when he says that the nature of persons ‘marks them out as an end in itself’ (G 4: 428). The nature of persons gives us a reason to refrain from subordinating them to other pursuits. Furthermore, the existence of such a reason does not depend on what any particular agent happens to think or feel about persons. Instead, the reason to treat persons as ends in themselves applies to everyone. And it is a reason that can be grasped by anyone looking at things from an impartial point of view. It is notoriously difficult to say exactly what kind of conduct is permitted, forbidden and demanded by this thought about the value of persons. But I am inclined to think that these difficulties do not and should not detract from the considerable force of the basic idea.

This modest form of realism does not clash with Kant's explicit claims about moral value in the Critique of Practical Reason. It is true that the view I attribute to Kant says that persons really are unconditionally valuable – independently of what anyone thinks or decides, even according to a purely rational procedure guided by the moral law. And it is also true that in the second chapter of the Analytic, Kant famously argues that the concepts of good and evil are determined by means of the moral law and not the other way around (KpV 5: 64). But there is no inconsistency here. Kant's famous doctrine – his ‘paradox of method’ (KpV 5: 62–3) – pertains to the appraisal of actions. The question is what standard should be used to judge whether a particular maxim is good or evil. And Kant's answer is that the moral law, rather than some independently ascertained conception of goodness, provides the ultimate standard. So if a person wonders about the morality of his maxim, he should consult the categorical imperative instead of wondering whether it would promote some good, the moral value of which cannot be determined without begging the question. Kant's discussion is about the application of moral concepts to maxims. It says nothing about the value of persons, since persons are not ‘objects of practical reason’ (KpV 5: 57). So there is no reason to think these two ideas conflict. The realism under consideration does not deny that the categorical imperative is the ultimate standard of moral judgement.Footnote ¹⁶ It only reminds us that its role as that standard is tied to the fact that it tracks the only thing in nature that has unconditional value, which it does by saying ‘that a rational being, as an end by its nature and hence as an end in itself, must in every maxim serve as the limiting condition of all merely relative and arbitrary ends’ (G 4: 436).Footnote ¹⁷

The realist interpretation of the value of persons combines with Kant's claims about the necessary conditions for self-legislation to yield the following argument about autonomy and moral value. Anyone with a will can act on the basis of maxims. So anyone with a will can propose to himself a principle of conduct and then try to act on the basis of that principle. But he will not be legislating unless his maxim can qualify as a practical law. Among the conditions a maxim must meet in order to qualify are the following. (a) The maxim must be applicable to everyone, and (b) it must include reference to an objective end. The only objective ends are persons. Therefore, a maxim qualifies as a law only when its ‘matter’ is a person. And from this it follows that an agent's will has the property of autonomy (self-legislation) only when he is capable of proposing to himself maxims that specify persons as the end for the sake of which he acts. Furthermore, persons have dignity – a value that does not depend on what anyone thinks or decides. Their ‘nature marks them out as an end in itself’. Therefore, an agent is capable of self-legislation only if he is capable of acting for the sake of the only thing in nature that has an objectively independent value. It seems reasonable to conclude from this that Kantian autonomy does not require us to believe that the value of persons must be conferred by the legislative acts of the rational will.Footnote ¹⁸

One might raise the following objection. The proposal under consideration implies that the value of persons is something to be recognized by the will, not conferred by it. From this it follows that when an agent manifests the proper respect for a person, he is responding to something other than his own will. But if an agent is guided by something other than his own will, he is not self-legislating. This is heteronomy, not autonomy. So Kant cannot be a realist about the value of persons and also claim that autonomy is the basis of morality. Realism undermines the very idea of autonomy. As Andrews Reath puts it, moral requirements ‘cannot be based on values, principles, or ends that are externally imposed on the will’ (Reath 1996: 122). Hence the need for a constructivist account of moral value.

Plausible as it seems, I believe this objection is misguided. The value of persons does not depend on the agent's decisions, but this fact does not threaten his autonomy. The agent who treats persons as ends in themselves recognizes their dignity and acts accordingly. This does not mean that something is being ‘externally imposed’ on his will. To appreciate this point, it is important to avoid taking too literally the distinction between what is internal and what is external to the will. Assuming that the will has no spatial boundaries, it seems fair to say that there is no sense in which anything could be literally inside or outside of it. So the question of whether something is internal or external to the will is not like the question of whether a pair of socks is inside or outside a drawer. Even if the spatial imagery could be interpreted literally, this is not the pertinent question.

When an agent must take his orders from someone else, something is certainly being externally imposed on his will. The problem is not merely that the source of the orders lies outside his will. This fact alone is not what threatens his autonomy. Instead, the threat comes from the fact that he cannot use his own practical reason to determine who or what will have the final say over his conduct. Such impositions come in many forms. A hospital patient might have a regimen of diet and exercise imposed by a doctor, for example. The regimen is externally imposed when the patient regards it as unnecessary or unbearably burdensome. Similarly, a criminal on parole might have to abide by a curfew. Or an employee might have to comply with a company policy he views as unfair, demeaning or just ridiculous. When something is externally imposed on a person's will in this manner, he must decide and act against his will. His will is subject to something unwelcome. And it is easy to see that this undermines or even eliminates his autonomy.Footnote ¹⁹

But when an agent regards a person as an end in itself, he is not taking orders from someone else. For one thing, according to Kant, no one can impose an end on another individual. We can be made to do something, but not for the sake of some end we do not set for ourselves. Only the agent himself can make something his end (MM 6: 381). The very act of setting an end is an ‘act of freedom’ (MM 6: 385). More to the point, when an agent responds appropriately to persons, he treats the dignity of persons as authoritative in the sense that he regards human dignity as an objective reason not to treat persons as merely instrumental to his other purposes. But this does not mean that the value of persons tells him what to do. There are no specific directives or statutes inscribed in the order of things. The mere fact that persons have unconditional value says nothing specific about how to treat them. It only tells us that it would be a mistake to subordinate them to other, less valuable, ends. Therefore, an agent must use his own judgement – his own practical reason – to determine what this implies for his conduct. It follows that the value of persons cannot impose anything in the manner of a superior giving orders.

We see this most clearly if we concentrate on the difference between submitting to an order and acknowledging the fact of a person's dignity. When an agent submits to an order from someone else, the order constrains his will in a particular way. It functions as a surrogate of his own volition. To the extent that he does not resist or question the order, there is a sense in which the agent willingly goes along with it. But going along with an order in this sense does not involve identifying with or endorsing it. Going along in the sense of giving in is not the same thing as going along in the sense of joining in. If I order you to humiliate someone else and you go along with that order even though it repulses you, because you are unwilling to bear the cost of disobeying, my order is an imposition on your will. In this sense, the order undermines your autonomy. The reason why is that my will, rather than your own, is the authority to which you submit. It is true that if I do not literally force your hand, you go along willingly, but only in the sense of giving in, not in the sense of joining in or identifying with the order.Footnote ²⁰

It is also true that something is externally imposed on an agent's will when it causally impinges on his decisions. Consider a sailor whose decision to sail north is overridden by a powerful wind that blows his ship west, or a dieter whose sincere New Year's resolution is overwhelmed by the smells from the kitchen. Both are subject to brute causal factors that interfere with their wills. But the value of persons does not impose on agents’ wills in this manner. An agent who responds appropriately to human dignity is not responding causally, as if to a sensory stimulus. Instead, he makes a decision to act for the sake of a person – either himself or another. And, again, this is ‘an act of freedom on the part of the acting subject, not an effect of nature’ (MM 6: 385).

The general point is that, when an agent responds appropriately to the value of persons, he is adopting a particular end as the ‘matter’ of the maxim that guides his conduct. This involves making a decision about what it is for the sake of which he acts. This decision is not simply the termination of a causal sequence over which he has no control. And it is not a decision that is or even could be made for him by someone else. We must set our own ends. Hence, the virtuous agent who treats persons as ends in themselves does not act against his will. It is true that the value of persons does not depend on his will. So there is some sense in which this value is external to his will. But it does not follow that anything is being imposed on him. On the contrary, it looks as if this cannot be the case, given what it means to set an end. This line of reasoning suggests that the agent's capacity for self-legislation is not impaired or undermined by the fact of human dignity. And this suggests that it is a mistake to think that Kant's view of autonomy commits him to the rejection of moral realism.

Consider a person who acts virtuously by fulfilling her ethical duty to others. Kant argues that we have an ethical duty to promote the happiness of others, and we do this by loving and respecting them. Love and respect, as Kant understands them, are not mere feelings. They are active dispositions of the will, reflected in the policies that an agent adopts. This ‘practical’ sense of respect is ‘to be understood as the maxim of limiting our self-esteem by the dignity of humanity in another person’ (MM 6: 449), a dignity that we are obligated to ‘recognize’ or ‘acknowledge’ (MM 6: 462). Now imagine Emily, the manager of an office in a medium-sized firm. Emily is ambitious, and she would like to become regional manager one day soon. Yet she has the capacity to check her ambition when it threatens to interfere with her ability to respect her co-workers. Some days this is more of a struggle than others. For example, imagine that Emily knows that upper management rewards arrogance, regarding it as necessary for success in a competitive environment. She and Tim are chatting with their boss at the firm's annual holiday party. She is well aware that this is not just a casual conversation. It is an opportunity. And she believes that the best way to capitalize would be to ridicule Tim, who lacks her sharp wit. She has seen this sort of behaviour rewarded by the firm, and she has a strong urge to take advantage of the opportunity to get ahead. But the thought of demeaning Tim for her own gain disturbs her and she fights this urge. It is not that she recognizes in Tim a friend. They barely know each other. Rather, she recognizes that he is a person, and cannot bear the thought of stepping on him to impress her boss. So she resolves to treat Tim as an equal partner in a casual conversation.Footnote ²¹

Emily's conduct might not be terribly heroic, but it is surely virtuous. She manifests the right response to the fact that Tim is a person, just as she is. This is what makes her conduct admirable. She may not have asked herself whether her maxim could be willed as a universal law, but let us assume that it meets the condition set by FUL. It also seems to meet the condition set by FHE. She has refused to treat Tim as an instrument of her ambition, regarding him instead as an end in himself. Therefore, her resolve manifests a maxim that qualifies as a practical law. In checking her ambition ‘by the dignity of another person’, she has engaged in an act of self-legislation, thus confirming and exercising her capacity for autonomy. On the proposed reading of dignity, she has done this by responding correctly to the dignity of her colleague. His dignity is not a product of her or anyone else's will; it is a fact about him that she acknowledges.

Even if it is also true that Emily would have benefited from demeaning Tim, there is no reason to think that the dignity of her colleague has been externally imposed on her will. She may be acting against some of her desires, but she is not acting against her will. Quite the opposite is true. Her act of volition is constituted by the decision to respect Tim. At the risk of sounding overly dramatic, she has adopted his humanity as her end, treating it, rather than career success, as that for the sake of which she acts. So there seems to be no reason to believe that her colleague's objectively independent value undermined her autonomy. On the contrary, her acknowledgement of his dignity is what enabled her to exercise her capacity for self-legislation. Had she not acknowledged that dignity, she would have made a mistake. And this mistake clearly would have undermined her capacity for self-legislation. If she had demeaned her colleague in pursuit of her own ambition, she would have acted on a maxim that could not meet the conditions specified by FUL and FHE. Even if she had been acting rationally, she would not have been acting on the basis of a practical law.

If we consider a similar scenario from a different angle, we can see the general point even more clearly. In the second scenario, Emily is not so virtuous and she has no qualms about ridiculing Tim to get ahead. She is Tim's immediate superior in the firm, and he believes he is likely to benefit from allowing himself to be used as an instrument of her ambition. At the very least, he does not want to get in her way. Being ambitious himself, he finds it tempting to roll over and let himself be ridiculed. But his sense of self-respect checks his ambition, and he resolves to resist Emily's attempts to demean him. He steadies himself to protest or walk away if she will not respond to his polite efforts to deflect her remarks. In this scenario, Tim displays a ‘love of honour’ (Ehrliebe), the virtue opposed to lying, avarice, and servility (MM 6: 420).Footnote ²²

Tim's conduct fulfils his duty of ‘moral self-preservation’ (MM 6: 419). He does not allow himself to be used by Emily to further her ambition – nor by himself to further his own. He thus avoids a morally blameworthy form of self-deprecation.Footnote ²³ On the proposed interpretation of moral value, he has responded appropriately to his dignity. His value depends on his will in the sense that if he did not have a will, he would not be a ‘lawgiving being’, and he would not have this unconditional, inner worth. But his value is not a product of his, or anyone else's, will since it does not depend on what he or anyone else decides is the case. Rather, his nature as a being with the capacity for giving law makes him an end in himself. And his conduct in this scenario manifests a proper respect for that fact. Perhaps there is some sense in which even his own dignity is external to his will, but it does not follow that when he acknowledges this dignity something is being externally imposed upon his will. In fact, it seems quite odd to think otherwise. To think this would be to think that the agent's own dignity, since it is not a product of his will, undermines his autonomy. My own sense is that simply putting the point this way shows that it is false.

According to Robert N. Johnson, ‘orthodox’ Kantians must reject this line of thought (2007: 134). He argues that the will is autonomous only if the rational will is the author of the moral law, and that this implies that moral value must be something the will creates rather than something to which it responds. Johnson further claims that this means that Kantians must resolve Euthyphro-style dilemmas in the same manner as voluntarists. Things have value because they are ‘the objects of rational agency’ (Johnson Reference Johnson2007: 140). They are not the objects of choice because they have value.

I think Johnson is mistaken on both points. Here is how he expresses the first: ‘What autonomy of the will requires is only that the explanation of the authority of the principles governing the will comes from the fact that the will is the source of those principles. And if the reason for you to conform to a law is the fact that you gave that law to yourself, then the reason does not derive from any value, such as the value of your will or your humanity’ (Johnson Reference Johnson2007: 141). But the mere fact that a person gives himself a practical principle does not explain why that principle has authority. Practical principles are laws only when they meet certain conditions.Footnote ²⁴ It is these conditions that specify the constraints on what counts as self-legislation. As we have seen, there are two main such conditions. First, the principle must be applicable to everyone in relevantly similar circumstances. Second, the principle must incorporate an objective end. It is easy to forget that this is the case, given our tendency to think of Kantian ethics as deontological, which can obscure the importance of ends in his moral theory. But it is right there in the heart of the Groundwork. An agent does not legislate unless his principles incorporate an objective end. And nothing other than a person counts as an objective end. Persons are ends in themselves because of ‘their nature’ (G 4: 428). If an agent does not legislate, he does not govern. And if he does not govern, he is not autonomous. Hence, an agent cannot exercise his autonomy unless he manifests the appropriate attitude towards persons, which he does by following policies that treat them as ends in themselves.

Johnson worries that if some independent ‘value is the source of reasons for the pursuits of rational agents, then the authority governing rational agency is external to that agency itself, in the value of the things that are its objects’ (Johnson Reference Johnson2007: 140). This concern seems misplaced. Perhaps it gets at least part of its force from an overly literal interpretation of the difference between what is internal and external to the will. This, at any rate, is how I diagnose my own susceptibility to the worry. But the realist view I have attributed to Kant does not entail that we are governed by the unconditional worth of persons. The pertinent sense of ‘to govern’ is to make and administer rules that regulate people's behaviour. But no rules or directives arise directly from human dignity. Rules arise when we try to determine what maxims to adopt in response to this value. The virtuous agent strives to treat the categorical imperative as the ultimate standard by which to judge all such rules. And he is dispositionally motivated to cultivate the strength it may take to follow the judgements he makes on the basis of that standard. So once we get past the literal concern with the will's spatial boundaries, it is difficult to see how an objectively independent value – understood in the sense I have proposed – could usurp an agent's authority over himself. It is hard to see how such a value could literally govern anyone's behaviour. After all, the dignity of persons does not depend on our will, but we still have to figure out what to do in the face of it.

Since the adoption of this end is one of the conditions for self-legislation, it is one of the conditions for occupying a position of sovereign authority over one's own conduct. And this is why an appropriate response to the independent value of humanity does nothing whatsoever to undermine an agent's autonomy. It is true that the fact that you gave yourself the law is the reason for you to follow that law. But you are not, strictly speaking, giving yourself a law unless you are adopting a maxim that incorporates humanity as your end. And, as we have seen, it is easy to read Kant as saying that this is because the very nature of persons marks them out as the only things in the universe with unconditional worth. The law – the principle that obligates unconditionally – is indeed self-given. It comes from the agent's own will. But it tracks something about human beings that does not.

If this line of thought is correct, then it is a mistake to think that Kantians must grab the voluntarist horn of Euthyphro-style dilemmas. Johnson thinks this is the only view about value that is consistent with the Kantian commitment to autonomy (2007: 140). But this does not seem right. As Rae Langton argues, the Euthyphro question is a question about direction of fit: ‘Does the gods’ love, or my rational choice, fit the value? Or does the value fit the gods’ love, my rational choice? Does the valuing fit the value? Or does the value fit the valuing’ (2007: 14)? On the realist reading of the value of persons, when an agent treats a person as an end in itself, the maxim that guides his action fits the person's value. That value is not a product of his will. It is an independent value to which he responds by adopting it as his end. As we have seen, there is no reason to think that this involves submitting to an external authority. It follows that there is no reason to think that Kantians cannot grab the other horn of Euthyphro-style dilemmas and say that we must respect persons because they have dignity. It is not that they have dignity because we have chosen to respect them.

And it is worth noting that when Kant discusses theological voluntarism and Euthyphro-style questions in his lectures, he seems to reject the voluntarist answer to such questions. There he says:

This sort of remark should be treated with caution. From the fact that Kant rejects theological voluntarism it does not automatically follow that he is a moral realist. But it seems difficult to square these arguments with what Johnson calls ‘orthodox’ Kantianism. As Langton remarks, ‘what goes for God presumably goes for ourselves too’ (2007: 27).Footnote ²⁵ At any rate, the main point I wish to make is that we should not think that the Kantian commitment to autonomy forces us to ascribe to the rational will a power that Kant would not even ascribe to God. Indeed, if Johnson is right to say that ‘there is no room and no need for metaphysical glitter in Kantian ethical theory’, there is reason to be suspicious of the idea that the rational will is the source of objective value (Johnson Reference Johnson2007: 134).Footnote ²⁶